Trace Elements in Magnetite from the Pagoni Rachi Porphyry Prospect, NE Greece: Implications for Ore Genesis and Exploration

Magnetite is a common accessory phase in various types of ore deposits. Its trace element content has proven to have critical implications regarding petrogenesis and as guides in the exploration for ore deposits in general. In this study we use LA-ICP-MS (laser ablation-inductively coupled plasma-mass spectrometry) analyses of trace elements to chemically characterize magnetite from the Pagoni Rachi Cu–Mo–Re–Au porphyry-style prospect, Thrace, northern Greece. Igneous magnetite mostly occurs as euhedral grains, which are commonly replaced by hematite in fresh to propylitic-altered granodiorite porphyry, whereas, hydrothermal magnetite forms narrow veinlets or is disseminated in sodic/potassic-calcic altered (albite + K-feldspar + actinolite + biotite + chlorite) granodiorite porphyry. Magnetite is commonly associated with chalcopyrite and pyrite and locally exhibits martitization. Laser ablation ICP-MS analyses of hydrothermal magnetite yielded elevated concentrations in several trace elements (e.g., V, Pb, W, Mo, Ta, Zn, Cu, and Nb) whereas Ti, Cr, Ni, and Sn display higher concentration in its magmatic counterpart. A noteworthy enrichment in Mo, Pb, and Zn is an unusual feature of hydrothermal magnetite from Pagoni Rachi. High Si, Al, and Ca values in a few analyses of hydrothermal magnetite imply the presence of submicroscopic or nano-inclusions (e.g., chlorite, and titanite). The trace element patterns of the hydrothermal magnetite and especially the decrease in its Ti content reflect an evolution from the magmatic towards the hydrothermal conditions under decreasing temperatures, which is consistent with findings from analogous porphyry-style deposits elsewhere.

. Geological map of NE Greece (after Voudouris et al., [30]; Mavrogonatos et al., [31]). The white diamond marks the location of the study area. Subsequent hydrothermal alteration of the magmatic and/or the sedimentary rocks led to the formation of significant metallic mineralization, mainly in the Rhodope (Thrace), and to a lesser extent the Serbo-Macedonian (Chalkidiki) domains [39,42,44,45].
The Northern Rhodope Domain, which occupies most of the NE part of Greece (Figure 1), consists of (a) a structurally lower high-grade gneissic basement derived from Permo-Carboniferous protoliths, subdivided into the Arda, Biala Reka-Kechros, and Kesebir-Kardamos migmatitic domes; (b) an intermediate unit of high-metamorphic grade basement rocks, whose protoliths yield two age intervals (Neoproterozoic to Ordovician and Permo-Carboniferous to Early Cretaceous); and (c) a structurally upper and low metamorphic grade unit known as Circum-Rhodope Belt, including several sub-units (e.g., Drymos-Melia, Makri) as well as the Evros ophiolite [33][34][35].
Exhumation of the deep-seated lithologies in the Rhodope Massif took place at  Ma [26,35,36] and the second, at~35 Ma [35]. Activation of major detachment faults resulted in the exhumation of deep lithologies, thus forming the metamorphic core complexes, and in the opening of several E-W trending, structurally-controlled basins [25,37]. Syn-extensional sedimentation resulted in the formation of thick clastic rock sequences (transgressive conglomerates, sandstones/claystones and limestones). Contemporaneous asthenospheric upwelling triggered by crustal thinning, generated magmatic rocks of Eocene to Miocene age, with a pronounced post-subduction signature [38][39][40]. These magmatic rocks are characterized by calc-alkaline to shoshonitic and ultra-potassic affinities, and display a mafic to felsic compositional range [38,[41][42][43].
Subsequent hydrothermal alteration of the magmatic and/or the sedimentary rocks led to the formation of significant metallic mineralization, mainly in the Rhodope (Thrace), and to a lesser extent the Serbo-Macedonian (Chalkidiki) domains [39,42,44,45].

Materials and Methods
Ten polished thin sections of magnetite-bearing samples and host rocks were studied by optical and a JEOL JSM 5600 (JEOL Ltd., Tokyo, Japan) scanning electron microscope (SEM) equipped with back-scattered imaging capabilities, at the Department of Mineralogy and Petrology of the University of Athens, Greece. Major and minor elements quantitative analyses on magnetite from three representative samples were carried out at the "Eugen F. Stumpfl" Laboratory, University of Leoben, Austria, using a Jeol Superprobe JXA 8200 (JEOL Ltd., Tokyo, Japan) wavelength-dispersive electron microprobe (WDS). Analytical conditions were as follows: 20 kV accelerating voltage, 10 nA beam current, and 1 µm beam diameter. Peak and backgrounds counting times were 20 s and 10 s for major and 40 s and 20 s for minor and trace elements, respectively. The MgKα, AlKα, SiKα, CaLα, TiKα, VKα, CrKα, FeKα, MnLα, NiLα, and ZnKα X-ray lines were used, with respect to the following standards: corundum for Al, wollastonite for Si and Ca, rutile for Ti, synthetic V for V, chromite for Mg and Cr, natural pyrite for Fe, rhodonite for Mn, NiS for Ni, and natural sphalerite for Zn. Corrections were applied using the PAP online program [59]. Laser ablation ICP-MS analyses of the same samples were conducted at the Institute of Mineralogy, University of Münster, Germany, with a pulsed 193 nm ArF excimer laser (Analyte G2, Photon Machines). A repetition rate of 10 Hz and an energy of~4 J/cm 2 were used throughout the entire session. The beam spot diameter was set to 35 µm. Trace element analysis has been carried out with an Element 2 mass spectrometer (ThermoFisher). Forward power was 1250 W and reflected power <1 W, gas flow rates were 1.2 L/min for He carrier gas, 0.9 L/m and 1.2 L/m for the Ar-auxiliary and sample gas, respectively. The argon cooling gas flow rate was set to 16 L/min. Before starting an analysis, the system was calibrated to a NIST 612 reference glass measuring 139 La, 232 Th, and 232 Th 16 O to get stable signals and high sensitivity, as well as low oxide rates ( 232 Th 16 O/ 232 Th < 0.1%) during ablation. A total of 26 elements ( 25 Mg, 27 Al, 29 Si, 43 Ca, 4 5Sc, 49 Ti, 51 V, 53 Cr, 55 Mn, 59 Co, 60 Ni, 63 Cu, 66 Zn, 69 Ga, 85  were quantitatively analyzed using the NIST 612 glass as an external standard while stoichiometric Fe (97% wt.% FeO) was the internal standard, since total concentration of minor elements is <3 wt.% (cf. Milani et al. [60]). Run time of the single ablation pattern was 60 s (20 s for background and 40 s for peak after switching the laser on). Concentrations of measured elements were calculated using the Glitter software [61,62]. Standard reference glasses BHVO-2G and BIR1-G were analyzed in order to monitor for precision and accuracy during the course of this study. The results from these glasses match the published range of concentrations (standard deviation in the range of 5%), given in the GeoReM database (version 23, [63]).

Geology and Description of the Pagoni Rachi Prospect
The Pagoni Rachi Cu-Mo-Re-Au porphyry prospect lies ca. 18 km NNE of Alexandroupolis. First described by Arikas [64], the Pagoni Rachi porphyry system occurs in an Oligocene granodiorite porphyritic body, which intrudes a Tertiary volcanosedimentary sequence (Figures 1 and 2). This sequence consists of conglomerates, marls, sandstones, and volcaniclastic material (e.g., tuffs) of Priambonian age. It is in turn overlain by volcanic (lava domes and flows) and/or pyroclastic rocks. The northern part of the study area is occupied by a large quartz monzodiorite intrusion, which crosscuts both the volcano-sedimentary sequence and the granodiorite porphyry. The monzodiorite exhibits a medium-grained to porphyritic texture, consisting of plagioclase and various amounts of pyroxene, amphibole, biotite, quartz, and opaques. Radiometric dating for the monzodiorite yielded ages of 31.9 ± 0.5 Ma (Rb/Sr on biotite, [38]). Recently, Perkins et al. [65] provided further geochronological data for the magmatic intrusion in the Kirki area (32.05 ± 0.07 Ma, U-Pb on zircon).  Late, NNW-trending microgranite porphyry dikes ( Figure 2) crosscut the aforementioned magmatic rocks and comprise the final intrusive stage in the area [51]. These dikes display a porphyritic texture and are composed of coarse-grained resorbed quartz and feldspar phenocrysts; the latter were commonly replaced by sericite, due to pervasive phyllic alteration.
The granodiorite porphyry that hosts the porphyry-style Cu-Mo-Re-Au mineralization is composed of plagioclase, amphibole, biotite, titanite, and scarce resorbed quartz phenocrysts (quartz-eyes), set in a microlitic matrix of quartz, orthoclase, plagioclase, and abundant opaque minerals, mainly magnetite. Mineralization occurs in a very dense quartz stockwork and is associated with extensive hydrothermal alteration of the granodiorite host ( Figure 2b). Late-stage, N-and NW-trending epithermal-style veins crosscut the porphyry mineralization and are associated with argillic (quartz-sericite-carbonates), and, rarely, advanced-argillic (quartz-kaolinite-sericite) alteration selvages, leading to partial telescoping of the porphyry system.
Porphyry-style mineralization has been attributed to four paragenetic stages [51]: (a) early Late, NNW-trending microgranite porphyry dikes ( Figure 2) crosscut the aforementioned magmatic rocks and comprise the final intrusive stage in the area [51]. These dikes display a porphyritic texture and are composed of coarse-grained resorbed quartz and feldspar phenocrysts; the latter were commonly replaced by sericite, due to pervasive phyllic alteration.
The granodiorite porphyry that hosts the porphyry-style Cu-Mo-Re-Au mineralization is composed of plagioclase, amphibole, biotite, titanite, and scarce resorbed quartz phenocrysts (quartz-eyes), set in a microlitic matrix of quartz, orthoclase, plagioclase, and abundant opaque minerals, mainly magnetite. Mineralization occurs in a very dense quartz stockwork and is associated with extensive hydrothermal alteration of the granodiorite host ( Figure 2b). Late-stage, N-and NW-trending epithermal-style veins crosscut the porphyry mineralization and are associated with argillic (quartz-sericite-carbonates), and, rarely, advanced-argillic (quartz-kaolinite-sericite) alteration selvages, leading to partial telescoping of the porphyry system.
Porphyry-style mineralization has been attributed to four paragenetic stages [51]: (a) early sodic/potassic-calcic alteration associated with magnetite + bornite + chalcopyrite + molybdenite + native gold + quartz veins (M-type veins,) and contemporaneous, distal propylitic alteration; (b) sodic/potassic alteration with pyrite + chalcopyrite + molybdenite + quartz veins (B-type); (c) sericitic alteration with pyrite + chalcopyrite + molybdenite veins (D-type) that also contain native gold along with an important suite of Ag-, Bi-,Te-, and Se-bearing minerals; and (d) argillic alteration with base and precious metal epithermal-style veins (E-type) rich in Au, Ag, and Te. Mineralogical similarities of these veins suggest that they may share a genetic relationship with the D-type veins. Bulk ore chemical analyses of surface samples showed that the system is significantly enriched in Cu (up to 0.5 wt.%), Mo (up to 2000 ppm), Au (up to 5 ppm), and Re (up to 20 ppm). The extreme rhenium enrichment of the Pagoni Rachi system reflects the presence of the rare sulfide rheniite and Re-rich molybdenite [50,51].

Modes of Magnetite Occurrence (Petrography)
Following the textural and morphological criteria proposed by McQueen and Cross [66] and Nadoll et al. [14], magnetite in the Pagoni Rachi porphyry-style mineralization occurs in two distinct types: The first type, hereafter called as magmatic magnetite, occurs in fresh to propyliticaly altered granodiorite porphyry (Figure 3a-c). Primary magnetite (Figure 4a-d) usually forms relatively large (up to 0.2 cm), euhedral to subhedral crystals (octahedral, cubo-octahedral habits) that are randomly scattered in the matrix of the rock, and which exhibit a characteristic porphyritic texture ( Figure 4). It comprises the most abundant opaque accessory phase (approx. 5% modal) and is associated with the primary silicate minerals of the granodiorite. Commonly propylitic alteration is present; in this case magnetite is set in a matrix of relict primary minerals replaced by variable amounts of secondary epidote, albite, actinolite, chlorite, and calcite (Figure 4a,b).
In such cases, magnetite is commonly replaced by hematite (martitization, Figure 4b) in the periphery of the crystals or along fractures. Rare exsolution lamellae (trellis-type) of ilmenite, typical for magmatic magnetite (cf. Wen et al. [17]) were identified (Figure 4d), at places also replaced by hematite.
The second type of magnetite is secondary (hereafter called "hydrothermal") and (Figures 3c-f and 5) is only found in the central part of the studied porphyry mineralization and is associated with sodic-potassic/calcic alteration. Its modal abundance (more than 10%) is significantly higher compared to its magmatic counterpart in the fresh to propylitic-altered granodiorite.
(up to 0.2 cm), euhedral to subhedral crystals (octahedral, cubo-octahedral habits) that are randomly scattered in the matrix of the rock, and which exhibit a characteristic porphyritic texture ( Figure 4). It comprises the most abundant opaque accessory phase (approx. 5% modal) and is associated with the primary silicate minerals of the granodiorite. Commonly propylitic alteration is present; in this case magnetite is set in a matrix of relict primary minerals replaced by variable amounts of secondary epidote, albite, actinolite, chlorite, and calcite (Figure 4a  In such cases, magnetite is commonly replaced by hematite (martitization, Figure 4b) in the periphery of the crystals or along fractures. Rare exsolution lamellae (trellis-type) of ilmenite, typical for magmatic magnetite (cf. Wen et al. [17]) were identified (Figure 4d), at places also replaced by hematite.  The second type of magnetite is secondary (hereafter called "hydrothermal") and (Figure 3c-f and Figure 5) is only found in the central part of the studied porphyry mineralization and is associated with sodic-potassic/calcic alteration. Its modal abundance (more than 10%) is significantly higher compared to its magmatic counterpart in the fresh to propylitic-altered granodiorite.
Hydrothermal magnetite forms a complex network of dark-colored veinlets set in a greenish to brown-colored matrix of secondary minerals. Moreover, magnetite may also be present in granular wavy to locally straight-sided quartz veins (A-type), which lack alteration selvages. Rarely, aggregates of magnetite occur in the sodic-potassic/calcic altered matrix.
This style of alteration (sodic-potassic/calcic) has totally obliterated the igneous porphyritic texture of the granodiorite and resulted in its complete recrystallization. It is composed of variable amounts of secondary albite, orthoclase, biotite, actinolite, and chlorite. The veinlets (M-type) are commonly very narrow (less than 0.5 cm), irregular in shape and discontinuous, but there are also cases where straight-sided veinlets were identified.
Magnetite in the veins is commonly anhedral, has a grain size that does not exceed a few tens of microns, and may contain (sub-) microscopic inclusions of silicates (e.g., titanite) and/or sulfides (e.g., chalcopyrite). It is associated with variable amounts of orthoclase, biotite, albite, chlorite, actinolite, and quartz (Figure 5a-b). Minor chalcopyrite and pyrite accompany magnetite in the veins (Figure 5c-e) along with minor amounts of bornite, gold, and molybdenite. In places, replacement features occurred (martitization) whereas exsolution lamellae were absent. Hydrothermal magnetite forms a complex network of dark-colored veinlets set in a greenish to brown-colored matrix of secondary minerals. Moreover, magnetite may also be present in granular wavy to locally straight-sided quartz veins (A-type), which lack alteration selvages. Rarely, aggregates of magnetite occur in the sodic-potassic/calcic altered matrix.
This style of alteration (sodic-potassic/calcic) has totally obliterated the igneous porphyritic texture of the granodiorite and resulted in its complete recrystallization. It is composed of variable amounts of secondary albite, orthoclase, biotite, actinolite, and chlorite. The veinlets (M-type) are commonly very narrow (less than 0.5 cm), irregular in shape and discontinuous, but there are also cases where straight-sided veinlets were identified.
Magnetite in the veins is commonly anhedral, has a grain size that does not exceed a few tens of microns, and may contain (sub-) microscopic inclusions of silicates (e.g., titanite) and/or sulfides (e.g., chalcopyrite). It is associated with variable amounts of orthoclase, biotite, albite, chlorite, actinolite, and quartz (Figure 5a-b). Minor chalcopyrite and pyrite accompany magnetite in the veins (Figure 5c-e) along with minor amounts of bornite, gold, and molybdenite. In places, replacement features occurred (martitization) whereas exsolution lamellae were absent.

Chemical Composition of Magnetite
Analytical data from both magmatic and hydrothermal magnetite are shown in Tables 1 and 2 for the Electron Probe Microanalyses (EPMA) and LA-ICP-MS analyses, respectively. Chemical differences between the two types of magnetite are displayed in covariation plots of Figure 6.
Electron probe micro-analyses of magmatic magnetite show that Ti was the major impurity, reaching up to 2.77 wt.% TiO2 (average 1.41 wt.%). Despite this relatively broad range, no zoned structure was observed. Silica (up to 0.52 wt.%) and Ca (up to 0.34 wt.% CaO) were the next most abundant elements. The majority of other divalent (e.g., Mg and Mn) and trivalent (e.g., Al and Cr) cations that could be incorporated into the structure of magnetite were mostly very close to, or below detection limits, with the exception of V, with values of up to 0.28 wt.% V2O3.
On the other hand, hydrothermal magnetite contained significantly less TiO2 (up to 0.78 wt.% and 1.13 wt.% in the two samples respectively) compared to the magmatic one, but had relatively elevated contents of Mg, Mn, and V (values up to 0.88 wt.% MgO, 0.32 wt.% MnO, and 0.33 wt.%

Chemical Composition of Magnetite
Analytical data from both magmatic and hydrothermal magnetite are shown in Tables 1 and 2 for the Electron Probe Microanalyses (EPMA) and LA-ICP-MS analyses, respectively. Chemical differences between the two types of magnetite are displayed in covariation plots of Figure 6.
Laser ablation-inductively coupled plasma mass-spectrometry analyses yielded comparable results to those obtained through EPMA: Ti values in the two hydrothermal magnetite-bearing samples that reach up to 4852 ppm and 11,517 ppm Ti (average values 3202 ppm and 2419 ppm respectively), and were much smaller compared to those in magmatic magnetite, which contained up to 22,545 ppm (average 9915 ppm). Alumina, Si, and Ca values are up to 7845 ppm (average 4112 ppm), 8521 ppm (average 991 ppm), and 4125 ppm (average 740 ppm) respectively. It must be noted that the maximum values in these elements corresponded to the same analytical spots, which might reflect the presence of sub-millimetric or nano-scale silicate inclusions at subsurface during ablation (cf. Deditius et al. [67]).
The average Mg content of the two types of magnetite was approximately the same (612 ppm for magmatic magnetite and 451 ppm and 550 ppm for the two hydrothermal magnetite samples. Maximum values were much higher (up to 2678 ppm in the magmatic sample and 6896 ppm and 3613 ppm in the two hydrothermal samples. Vanadium was enriched in the hydrothermal magnetite. Maximum values ranged between 1968 ppm (average 1622 ppm) and 3079 ppm (average 1077 ppm) in the two samples, whereas maximum values in the magmatic magnetite was 770 ppm (average 395 ppm).
A similar trend was the case for most trace elements ( Table 2): hydrothermal magnetite was enriched in high-field strength elements (e.g., Nb, Hf, Ta, and W) and metals (e.g., Cu, Zn, and Mo).
On the other hand, some elements were preferably enriched in the magmatic magnetite: the Co of magmatic magnetite was slightly higher (up to 27 ppm) than that of the hydrothermal ones (up to 21 ppm). Other enriched elements were Cr (up to 454 ppm in magmatic; and up to 28 ppm in hydrothermal magnetite respectively), Ni (up to 31 ppm in magmatic and up to 21 ppm in hydrothermal magnetite), and Sn (up to 99 ppm in magmatic and up to 41 ppm in hydrothermal magnetite).
Differences in the concentrations observed between EPM and LA-ICP-MS analyses for elements like Ti, Mg, Cr, Ca, and Si was attributed to the variation of the beam-size of the two methods, especially when exsolved phases or micro-inclusions were present in the magnetite. Furthermore, elevated contents acquired by LA-ICP-MS analyses (e.g., values of Si, Ca, Ti, and Al) might suggest that at least some of the very high concentrations was due to unavoidable submicroscopic or nano-scale inclusions of silicates.

Discussion
Magnetite is a common mineral in porphyry deposits. Magnetite of magmatic origin is commonly disseminated in the unaltered intrusive rocks or partly as a relict in propylitic alteration that surrounds the mineralized center [14,18]. The composition of magnetite seems to be a function of several factors including temperature, f S 2 and/or f O 2 , cooling rate, melt composition, and element partitioning between co-precipitating phases [1,2,12,13,16,68]. Available experimental data on igneous magnetite have shown that its composition depends mainly on the temperature of crystallization, the f S 2 and f O 2 conditions, and buffering between melt and the host rock [69][70][71]. For instance, an increase in oxygen fugacity would reduce the solubility of Ti in magnetite and consequently cause oxy-exsolution of ilmenite (cf. Wen et al. [17]).
In contrast, hydrothermal magnetite is usually common in early, high-temperature stages associated with potassic alteration assemblages at the cores of the systems [13][14][15]18]. Here, elemental availability in the hydrothermal fluid seems to be a critical component that determines its composition, combined with temperature, sulfur and oxygen fugacities, and fluid-rock interaction [13]. Although present in the majority of the porphyry-style deposits, there seems to be a greater abundance of magnetite in gold-rich deposits, a feature that could be attributed to the generation of such deposits by more oxidizing magmas [72,73]. Iron in the porphyry environment is usually thought to be transported as a chloride complex in a magmatic vapor (FeCl 2 ) and magnetite is considered to precipitate as a result of volatile exsolution at relatively high temperature and pressure conditions [74].

Fingerprinting Magmatic and Hydrothermal Processes
The Ti versus Ni/Cr discrimination diagram of Dare et al. [1] distinguishes the two types of magnetite in the Pagoni Rachi prospect (Figure 6a), as magmatic magnetite has high Ti and low Ni/Cr values compared to its hydrothermal counterpart. Furthermore, the Ti versus Al plot, proposed by Canil et al. [15], also distinguishes the studied magnetite, but mostly based on Ti content, since their Al content is relatively constant (Figure 6b) in both magmatic and hydrothermal magnetite.
The significant enrichment of V in magnetite of the Pagoni Rachi prospect is a feature previously described from other porphyry deposits elsewhere (e.g., [14,15,18]). Vanadium in the Pagoni Rachi is present in both types of magnetite, but its concentration is higher in the hydrothermal grains (Figure 6c). Vanadium is characterized by significant mobility even in low temperature hydrothermal fluids [75]. Its presence, commonly is in the trivalent state, since it has a similar ionic radius to Fe 3+ [69], is a function of oxygen fugacity of the hydrothermal fluid as well as of the presence of any other co-existing minerals with magnetite, which can also incorporate V 3+ (e.g., ilmenite and biotite). Since V is enriched in the hydrothermal type of magnetite from Pagoni Rachi, its availability in the hydrothermal fluid could be strengthened by dissolution of V-rich magmatic magnetite. This procedure could result in reprecipitation of newly-formed hydrothermal magnetite that has partly inherited features of its magmatic counterpart, as described by Wen et al. [17]. The two types of magnetite in the Pagoni Rachi prospect are partly discriminated in the plot of Ca + Al + Mn versus Ti + V plot (Figure 6d), although it should be noted here that composition of magnetite regardless of its origin in the prospect overlaps the genetic fields defined by Dupuis and Beaudoin [7] and Nadoll and Koenig [76]. Figure 7 graphically represents the chemical composition of magmatic and hydrothermal magnetite analyzed during the present study. In Figure 7a, the average composition of the two varieties of magnetite from Pagoni Rachi, normalized to bulk composition of the continental crust, plot along with the compositional range of high-temperature hydrothermal magnetite of Dare et al., [1]. With the exception of Ti, Sn, and Cr, most elements are enriched in the hydrothermal magnetite. The concentrations of Si and Ca from both types of magnetite in the Pagoni Rachi prospect are higher than those reported by Dare et al. [1] for hydrothermal magnetite. In general, magmatic magnetite is depleted in both elements, since they are considered to be incompatible, when igneous magnetite crystallizes from silicate melts [12]. magmatic magnetite. This procedure could result in reprecipitation of newly-formed hydrothermal magnetite that has partly inherited features of its magmatic counterpart, as described by Wen et al. [17]. The two types of magnetite in the Pagoni Rachi prospect are partly discriminated in the plot of Ca + Al + Mn versus Ti + V plot (Figure 6d), although it should be noted here that composition of magnetite regardless of its origin in the prospect overlaps the genetic fields defined by Dupuis and Beaudoin [7] and Nadoll and Koenig [76].  plot along with the compositional range of high-temperature hydrothermal magnetite of Dare et al., [1]. With the exception of Ti, Sn, and Cr, most elements are enriched in the hydrothermal magnetite. The concentrations of Si and Ca from both types of magnetite in the Pagoni Rachi prospect are higher than those reported by Dare et al. [1] for hydrothermal magnetite. In general, magmatic magnetite is depleted in both elements, since they are considered to be incompatible, when igneous magnetite crystallizes from silicate melts [12].  On the other hand, hydrothermal magnetite may carry significant amounts of Si and Ca, as both elements are highly mobile during hydrothermal alteration. Silicon-and Ca-rich magnetite have been reported from several deposits, commonly skarns (e.g., [78]). In the case of the Pagoni Rachi porphyry prospect, significant Si-Ca mobility during alteration has been well documented, as hydrothermal magnetite is accompanied by Ca-rich mineral assemblages (e.g., hydrothermal actinolite, Voudouris et al. [50,51]). Moreover, very high values of these two elements, when they correlate along high Ti values ( Figure 8) are likely due to the presence of nano-scale inclusions of titanite (cf. Deditius et al. [67]). been reported from several deposits, commonly skarns (e.g., [78]). In the case of the Pagoni Rachi porphyry prospect, significant Si-Ca mobility during alteration has been well documented, as hydrothermal magnetite is accompanied by Ca-rich mineral assemblages (e.g., hydrothermal actinolite, Voudouris et al. [50,51]). Moreover, very high values of these two elements, when they correlate along high Ti values (Figure 8) are likely due to the presence of nano-scale inclusions of titanite (cf. Deditius et al. [67]). Magnesium and less commonly Mn can be enriched in hydrothermal magnetite, due to extensive fluid-rock interactions [79,80]. Such interactions enriched both elements in the studied magnetite and could easily be explained by the mafic component of the hosting granodioritic intrusion.
With the exception of Zr and Hf, which displayed relatively constant concentrations both in magmatic and hydrothermal magnetite, other high-field strength elements were preferably enriched in the hydrothermal magnetite ( Figure 7b). As most of these elements were incompatible in magnetite, they were relatively immobile and tend to be incorporated in late magmatic magnetite. Their solubility in hydrothermal fluids is restricted [81][82][83]. However, the enrichment of elements like W, Mo, Ta, and Nb in the hydrothermal magnetite could indicate that it precipitated from a hydrothermal fluid that derived from a felsic intrusion and contained a significant amount of such elements. Large variation in their concentration could imply that they were incorporated in the magnetite grains following an inhomogeneous distribution.
The copper content of magmatic magnetite was very low, but a significant enrichment was present in the hydrothermal variety. Copper-rich magnetite is common in many porphyry-style deposits, with extreme values of a few thousands of ppm (e.g., magnetite from Mt Polley porphyry deposit, [18]), but its origin is still a matter of debate, as the presence of nano-scale inclusions of copper sulfides cannot be ruled out (e.g., [84]).
Lead enrichment is an important feature of magnetite from Pagoni Rachi. Minor amounts of Pb were present in the magmatic magnetite, and could be explained by its weakly compatible behavior (DPb = 1.4 in magmatic magnetite). The Pb content rose significantly in the hydrothermal type of magnetite: given the fact that Pb is a chalcophile element, and is preferably incorporated in coexisting Fe-Cu sulfides (e.g., [1]), its presence in hydrothermal magnetite suggests enrichment of this element in the parental hydrothermal fluid, that is exsolved from a felsic magmatic source, as also described by Huang et al., [16]. Magnesium and less commonly Mn can be enriched in hydrothermal magnetite, due to extensive fluid-rock interactions [79,80]. Such interactions enriched both elements in the studied magnetite and could easily be explained by the mafic component of the hosting granodioritic intrusion.
With the exception of Zr and Hf, which displayed relatively constant concentrations both in magmatic and hydrothermal magnetite, other high-field strength elements were preferably enriched in the hydrothermal magnetite ( Figure 7b). As most of these elements were incompatible in magnetite, they were relatively immobile and tend to be incorporated in late magmatic magnetite. Their solubility in hydrothermal fluids is restricted [81][82][83]. However, the enrichment of elements like W, Mo, Ta, and Nb in the hydrothermal magnetite could indicate that it precipitated from a hydrothermal fluid that derived from a felsic intrusion and contained a significant amount of such elements. Large variation in their concentration could imply that they were incorporated in the magnetite grains following an inhomogeneous distribution.
The copper content of magmatic magnetite was very low, but a significant enrichment was present in the hydrothermal variety. Copper-rich magnetite is common in many porphyry-style deposits, with extreme values of a few thousands of ppm (e.g., magnetite from Mt Polley porphyry deposit, [18]), but its origin is still a matter of debate, as the presence of nano-scale inclusions of copper sulfides cannot be ruled out (e.g., [84]).
Lead enrichment is an important feature of magnetite from Pagoni Rachi. Minor amounts of Pb were present in the magmatic magnetite, and could be explained by its weakly compatible behavior (D Pb = 1.4 in magmatic magnetite). The Pb content rose significantly in the hydrothermal type of magnetite: given the fact that Pb is a chalcophile element, and is preferably incorporated in coexisting Fe-Cu sulfides (e.g., [1]), its presence in hydrothermal magnetite suggests enrichment of this element in the parental hydrothermal fluid, that is exsolved from a felsic magmatic source, as also described by Huang et al., [16].
Another notable feature of the studied magnetite is their enrichment in Sn, which is obvious in both types of magnetite but is larger in the igneous magnetite. Analogous enrichment of Sn in porphyry-related magnetite was reported previously from the Endako porphyry-Mo deposit [85], which is hosted in granitoids, which are notably more felsic compared to the granodiorite hosting the Pagoni Rachi prospect.

Genetic Considerations-Implications for Exploration
Although hosted in an intermediate magmatic body with high-K calc-alkaline affinity, the Pagoni Rachi porphyry prospect was characterized by a number of features that point towards a genetic relation to a more evolved, felsic intrusive phase. This inquiry has already been stated by Voudouris et al. [51], who discussed the possibility of a genetic relationship between the porphyry system and a buried granitic intrusion similar to that exposed at the nearby St. Philippos deposit. Furthermore, the presence of fluorite as a gangue mineral in the quartz porphyry veins of the Pagoni Rachi [51], as well as the enrichment of F and Cl in secondary biotite (C. Mavrogonatos unpublished data) from the sodic/potassic-calcic alteration of the system, also suggest a strong affinity with a halogen-enriched hydrothermal fluid derived from a felsic magma.
High-field strength elements, which are enriched in secondary magnetite, suggest elevated mobility during hydrothermal alteration. Such enrichment has been described previously from fluoride-rich magmatic systems, especially of alkaline character [86]. Moreover, as has been shown by experimental studies, even the presence of a minor amount of base metals like Mn and Zn in hydrothermal magnetite, reflects an elevated concentration of these elements in chloride-rich hydrothermal fluids, which suggest an enrichment caused by fluid-rock interactions [78,79].
Significant porphyry-style mineralization hosted in microgranite, has already been described from the Thrace metallogenic district, thus documenting a genetic link between porphyry-style prospects and the microgranitic magmatism (e.g., the Maronia porphyry Cu-Mo-Re-Au [53] and the Aisymi-Leptokarya porphyry Mo [45,54] systems). The existence of microgranitic bodies in the Pagoni Rachi area, as well as the presence of F-rich phases in advanced argillic alteration assemblages in epithermal-style mineralization [31,48,49], locally superimposed on the porphyry-style ores in the broad area, is also compatible with a deep, large-scale microgranitic intrusion, which could (at least partly) control the alteration styles and contribute to the metal endowment of the Pagoni Rachi porphyry prospect.
Further studies of trace elements in magnetite from other deposits and prospects in the district will increase our knowledge concerning their distribution during ore formation and will help to determine a trace element pattern that could be used as a vector towards ore-grade mineralized centers in the Thrace metallogenic district.

Conclusions
Magnetite occurs as two modes of occurrence in the Pagoni Rachi porphyry Cu-Mo-Re-Au prospect. The first, recognized as magmatic-type, is present in fresh to propylitic-altered granodiorite porphyry, peripherally to the mineralized center of the system. The second type of magnetite, which is of hydrothermal origin, is found in the core of the system, associated with sodic/potassic-calcic alteration of the host rock. Magmatic magnetite forms subhedral to euhedral grains, with sizes up to 0.2 cm. In places, rare trellis-type exsolution lamellae of ilmenite may occur. Hydrothermal magnetite occurs as narrow veinlets (M-type) or within sinuous quartz veins (A-type), forming fine, anhedral grains. It is associated with various amounts of secondary biotite, K-feldspar, and/or albite, actinolite, and chlorite.
Trace element analyses revealed significant differences in the composition of the two types of magnetite. Magmatic magnetite from Pagoni Rachi, similar to magmatic magnetite from intrusions related to ore deposits elsewhere, was characterized by elevated concentrations of Ti, Cr, Ni, and Sn. In contrast, high-field strength elements and metals were preferably enriched in the hydrothermal magnetite. In addition, enrichment in Zn, Pb, W, Ta, and Mo was a distinctive feature of the Pagoni Rachi hydrothermal magnetite. Very high values of Si, Ca, Mg, and Al might indicate submicroscopic and/or nano-scale inclusions in hydrothermal magnetite.
The trace elements pattern of the hydrothermal magnetite could be further tested as an exploration tool towards ore-grade mineralization in other fertile intrusions in NE Greece, which share similar genetic features to Pagoni Rachi (e.g., causative intrusion and metal endowment, similar alteration assemblages, etc.).