Iron Isotope Fractionation during Skarn Cu-Fe Mineralization

: Fe isotopes have been applied to the petrogenesis of ore deposits. However, the behavior of iron isotopes in the mineralization of porphyry-skarn deposits is still poorly understood. In this study, we report the Fe isotopes of ore mineral separations (magnetite, pyrite, chalcopyrite and pyrrhotite) from two different skarn deposits, i.e., the Tonglvshan Cu-Fe skarn deposit developed in an oxidized hydrothermal system and the Anqing Cu skarn deposit developed in a reduced hydrothermal system. In both deposits, the Fe isotopes of calculated equilibrium ﬂuids are lighter than those of the intrusions responsible for the skarn and porphyry mineralization, corroborating the “light-Fe ﬂuid” hypothesis. Interestingly, chalcopyrite in the oxidized-Tonglvshan skarn deposit has lighter Fe than chalcopyrite in the reduced-Anqing skarn deposit, which is best understood as the result of the prior precipitation of magnetite (heavy Fe) from the ore ﬂuid in the oxidized-Tonglvshan systems and the prior precipitation of pyrrhotite (light Fe) from the ore ﬂuid in the reduced-Anqing system. The δ 56 Fe for pyrite shows an inverse correlation with δ 56 Fe of magnetite in the Tonglvshan. In both deposits, the Fe isotope fractionation between chalcopyrite and pyrite is offset from the equilibrium line at 350 ◦ C and lies between the FeS-chalcopyrite equilibrium line and pyrite-chalcopyrite equilibrium line at 350 ◦ C. These observations are consistent with the FeS pathway towards pyrite formation. That is, Fe isotopes fractionation during pyrite formation depends on a path from the initial FeS-ﬂuid equilibrium towards the pyrite-ﬂuid equilibrium due to the increasing extent of Fe isotopic exchange with ﬂuids. This ﬁnding, together with the data from other deposits, allows us to propose that the pathway effect of pyrite formation in the Porphyry-skarn deposit mineralization is the dominant mechanism that controls Fe isotope characteristics.

Traditionally, light stable isotopes (e.g., H, O, S) have been used to study sources and processes of mineralization [18][19][20], contributing to the understanding of the petrogenesis of ore deposits, but cannot provide more direct information since these elements are not ore-forming metals [21,22].

Anqing Cu skarn Deposit
The Anqing Cu deposit is a typical skarn ore deposit in the Yueshan metallogenic belt [60], located within the contact between the dioritic intrusions and the Low Triassic Yueshan Formation (limestone) (Figure 4). A previous study divided the mineralization

Anqing Cu skarn Deposit
The Anqing Cu deposit is a typical skarn ore deposit in the Yueshan metallogenic belt [60], located within the contact between the dioritic intrusions and the Low Triassic Yueshan Formation (limestone) (Figure 4). A previous study divided the mineralization into four stages: (I) garnet-diopside skarn stage, (II) oxides stage, (III) quartz-sulfide stage and (IV) quartz-carbonate stage [60]. The fluid inclusion work suggests the temperature of 200-375 • C at the quartz-sulfide stage [60].
The main mineral assembles in the Anqing Cu deposit are pyrite, chalcopyrite and pyrrhotite, similar to reduced porphyry-skarn deposit ( Figure 5; see Table 2 for petrography) [56]. The anhedral and subhedral pyrrhotite is replaced and cut by anhedral chalcopyrite, indicating that chalcopyrite formed later than pyrrhotite (Figure 5a-d). Pyrite is subhedral and replaces the disseminated pyrrhotite, suggesting that the pyrite formed later than pyrrhotite (Figure 5f). Hence, the paragenetic sequence may be as follows: pyrrhotite → pyrite/chalcopyrite. Table 2. Petrography of the Anqing skarn ore samples.
Minerals 2021, 11, x FOR PEER REVIEW 6 phy) [56]. The anhedral and subhedral pyrrhotite is replaced and cut by anhedral cha pyrite, indicating that chalcopyrite formed later than pyrrhotite (Figure 5a-d). Pyrit subhedral and replaces the disseminated pyrrhotite, suggesting that the pyrite form later than pyrrhotite ( Figure 5f). Hence, the paragenetic sequence may be as follows: p rhotite → pyrite/chalcopyrite. Table 2. Petrography of the Anqing skarn ore samples.

Analytical Method
Fe isotope compositions of the ore mineral separates were analyzed in the Laboratory of Ocean Lithosphere and Mantle Dynamics, Institute of Oceanology, Chinese Academy of Sciences (IOCAS), Qingdao, China. The skarn ore samples were crushed using a corundum jaw crusher. The separation was done by using a magnetic separator to separate magnetite and pyrrhotite from non-magnetic minerals. Individual mineral phases were handpicked under a binocular microscope [36]. We have altogether obtained mineral separates of garnet, magnetite, pyrite, chalcopyrite and pyrrhotite from the massive skarn ore samples.

Analytical Method
Fe isotope compositions of the ore mineral separates were analyzed in the Laboratory of Ocean Lithosphere and Mantle Dynamics, Institute of Oceanology, Chinese Academy of Sciences (IOCAS), Qingdao, China. The skarn ore samples were crushed using a corundum jaw crusher. The separation was done by using a magnetic separator to separate magnetite and pyrrhotite from non-magnetic minerals. Individual mineral phases were handpicked under a binocular microscope [36]. We have altogether obtained mineral separates of garnet, magnetite, pyrite, chalcopyrite and pyrrhotite from the massive skarn ore samples.
For Fe isotopes analysis, about 50 mg of mineral samples were digested in HNO3-HCl-HF (1 mL H 2 [(N 3 O 8 )Cl] and 0.5 mL HF) at~190 • C for 15 h and then re-dissolved with the distilled 3 mol L −1 HNO 3 for two hours until complete dissolution after evaporation. Finally, the sample were dissolved in the 1 mL 9 mol L −1 HCl for chromatographic separa-tion for Fe. We used a column filled with 1 mL Bio-Rad AG-MP-1 M resin (200-400 mesh) to purify the Fe element in a HCl medium following the procedures in Ref. [62]. After total purification, the eluted Fe solutions were analyzed using ICP-OES to ensure purity and full recovery. Prior to analysis, each sample was doped with GSB Ni standard (an ultrapure single elemental standard solution from the China Iron and Steel Research Institute) with Ni:Fe = 1.4:1 to monitor the instrumental mass bias during the analysis using a Nu Plasma MC-ICP-MS with wet nebulization in medium resolution (a mass resolution > 8000). Mass bias fractionation was corrected by 60 Ni/ 58 Ni similar to [63], with the 58 Fe interference on 58 Ni corrected for based on 56 Fe. Each sample was analyzed five times and every two sample solutions were further bracketed with 14 ppm GSB Fe standard solution that was also doped with the GSB Ni solution with Ni:Fe ratio of 1.4:1 (a substitution of IRMM-014; δ 56 Fe IRMM-014 = δ 56 Fe GSB + 0.729; δ 57 Fe IRMM-014 = δ 57 Fe GSB + 1.073) [64]. Iron isotopic composition is expressed in δ-notation and normalized to IRMM-014 value: where i refers to mass 56 or 57. The δ 56 Fe values of the USGS standard GSP-2, BCR-2, AGV-2 and BHVO-2 were 0.13 ± 0.04‰, 0.07 ± 0.04‰, 0.09 ± 0.04‰, 0.11 ± 0.02‰, respectively, which are consistent with the literature values within error [64][65][66][67]. Instrumental and analytical details are given in [62].

Tonglvshan Cu-Fe Skarn Deposit
In the Tonglvshan Cu-Fe skarn deposit, the δ 56 Fe values vary in the order of Mt~Grt > Py > Cp > Sid~Bn (Table 3; Figure 7). The δ 56 Fe values for mineral separates from TLS4-7 are generally lighter than those of mineral separates from other seven samples (Table 3), which may reflect a two-stage magmatic-hydrothermal event [53]. One magmatichydrothermal event may be related to the samples from ore body III, whose ore-forming fluid has a heavier Fe isotope composition. The other event is possibly concerned with the TLS4-7, as the ore-forming fluid has a lighter Fe isotope composition.
The magnetite has the heaviest Fe isotopic composition, which is consistent with its high 56Fe β-factor [43]. Therefore, we considered that the magnetite may have reached Fe iso-topic equilibrium. Only one co-existing chalcopyrite and bornite were analyzed on sample (TLS4-7) to show the Fe isotope contrast between chalcopyrite and bornite. This observation (∆ 56 Fe Cp-Bn ≈ 0.31 ± 0.05‰) is consistent with those in the literature [24]. The siderite has the lightest Fe (δ 56 Fe = −1.16 ± 0.03‰), which is consistent with the prediction that ferrous carbonates preferentially incorporate the light Fe isotope [32,34].
However, according to theoretical and experimental studies [42,43,68,69], pyrite is expected to have the heaviest Fe isotopes among co-existing minerals in equilibrium because pyrite has the highest 56 Fe β-factor. There is a mainstream view to explain the 'light pyrite', whose rapid precipitation may preserve the isotopic composition of FeS precursor with light Fe, as the result of kinetic control [44,[48][49][50]. Nevertheless, with the continuous evolution of hydrothermal fluids, equilibrium fractionation will also cause a great change in the Fe isotope composition of pyrite, which may not need to invoke kinetic fractionation (Section 5.4; see below).

Anqing Cu Skarn Deposit
In the Anqing Cu-Fe skarn deposit, pyrrhotite has the lightest Fe isotope composition, which is consistent with the results from the other deposits [24,32,33] and the experimental Fe isotope fractionation factor [46]. Chalcopyrite has the heaviest Fe isotope composition. Except for one chalcopyrite-pyrrhotite mineral pair, the other co-existing chalcopyritepyrrhotite have a good positive 56 Fe correlation (Figure 8c), indicating that chalcopyrite and pyrrhotite may reach Fe isotopic equilibrium. The variation of Fe isotopic composition of pyrite is the largest, which may be explained by the Rayleigh fractionation or kinetic fractionation, if any. Detail discussions are in Section 5.4.

The Pathway Effects for Pyrites Formation
In nature, pyrite can be formed through multiple pathways. The detailed reactions are as follows: Fe 2+ (aq) + H 2 S = FeS (s) + 2H + (aq) The first precipitate is not pyrite, but the unstable mineral FeS (mackinawite) [50,70]. Then, dissolution of FeS s produces FeS aq , which forms pyrite through the H 2 S pathway or polysulfide pathway [71][72][73]: A previous study [31] on pyrite of the Tongshankou porphyry deposit suggests that there is no obvious kinetic Fe isotopes fractionation for Reactions (3) and (4). Experimental studies show that the primary rapidly precipitated pyrite would inherit the Fe isotopic composition of the intermediate FeS phase that has a lighter Fe isotopic composition in equilibrium with the fluid [44]. However, as the reaction proceeds, the Fe isotope fractionation during pyrite formation moves towards the pyrite-fluid equilibrium with an increasing extent of the Fe isotopic exchange [31,44]. The Fe isotopic composition of the fluid is up to the extent of Fe isotopic exchange between pyrite and the fluid. The greater the extent of Fe isotopic exchange between pyrite and fluid is, the heavier the Fe isotopic composition of the pyrite and the lighter the Fe isotopic composition of the fluid will be, and vice versa. The study [31] found that the Fe isotopic composition of pyrite and chalcopyrite mineral pairs in the Tongshankou porphyry deposits are negatively correlated, which was interpreted to reflect pyrite formation through the FeS pathway in combination with S isotope data. The wide Fe isotope range of pyrite may be related to different extents of reactions during pyrite precipitation from an initial FeS-fluid equilibrium towards pyritefluid equilibrium, governing the Fe isotopic composition of ore-forming fluid, which in turn affects the chalcopyrite. In brief, the FeS-pathway effect is an important mechanism controlling the Fe isotope fractionation.
In the Tonglvshan skarn deposit, there is an interesting phenomenon that the Fe isotopic composition of magnetite and pyrite has a good negative correlation (Figure 8a) despite a limited Fe isotope variation. These magnetite crystals are characterized by abundant porous pits and are mostly replaced by sulfides (Figure 3), which may be caused by the dissolution-reprecipitation process (DRP) [57][58][59]. Hu et al. [57] suggest that DRP may be important in the formation of hydrothermal magnetite. During such a process, the primary magnetite may exchange the Fe isotopes with the fluid whose Fe isotope composition is governed by the pyrite formation pathways [31,44]. Heavy-Fe pyrite causes the fluid to be Fe-light, which in turn causes the reprecipitation magnetite to be Fe-light. The light-Fe pyrite causes the reprecipitation magnetite to be Fe-heavy, finally resulting in the negative correlation between the magnetite and pyrite. The Fe isotope fractionation between chalcopyrite and pyrite is offset from the equilibrium line at 350 • C and lies between the FeS-chalcopyrite equilibrium line and pyrite-chalcopyrite equilibrium line at 350 • C (Figure 8b). This is best understood as the FeS-pathway effect. The extent of Fe isotope exchange between the pyrite and the fluid controls the Fe isotope composition of the fluid, further affecting the chalcopyrite.
In the Anqing skarn deposit, the Fe isotope fractionation of chalcopyrite-pyrite lying between the FeS-chalcopyrite equilibrium line and pyrite-chalcopyrite equilibrium line at 350 • C may display the role of the FeS-pathway effect (Figure 8b). Except for one chalcopyrite-pyrrhotite pair (sample AQ-2), the other co-existing chalcopyrite-pyrrhotite pairs have a good positive 56 Fe correlation (Figure 8c), indicating that chalcopyrite and pyrrhotite may reach Fe isotopic equilibrium at 300 • C-350 • C.
In summary, the FeS pathway of pyrite formation in the skarn deposit may be a common mechanism causing a change in the isotopes' composition of the fluid, which in turn affects the Fe isotopic composition of the fluid and other co-existing minerals.

What Are the Controlling Variables That Govern the Fe Isotope Variation in Co-Exiting Phases in a Magmatic-Hydrothermal Fluid System?
With the continued cooling and crystallization of the magma, the magmatic-hydrothermal fluids are released, resulting in ore deposit formation [32,33,74,75]. During this process, Fe is transported as Fe 2+ -chloride complexes [76,77]. There exist two main views about the Fe isotope fractionation during the fluid exsolution. One is the "light fluid" hypothesis, mainly confirmed by several studies [21][22][23][24][25]27,31], while a recent study [32] considered that the redox state actually governs the fluid Fe isotope composition. Another study [33] proposed that the oxygen and sulfur redox state of ore fluids also have an influence on isotope values of mineral assemblages because of the presence or absence of pyrrhotite (Figures 9 and 10a). about the Fe isotope fractionation during the fluid exsolution. One is the "light fluid" hypothesis, mainly confirmed by several studies [21][22][23][24][25]27,31], while a recent study [32] considered that the redox state actually governs the fluid Fe isotope composition. Another study [33] proposed that the oxygen and sulfur redox state of ore fluids also have an influence on isotope values of mineral assemblages because of the presence or absence of pyrrhotite (Figures 9 and 10a). . Temperature (°C) vs. oxygen fugacity (logfO2) diagrams [27,33,56]. The blue path shows the mineral precipitation sequence of the oxidized porphyry-skarn Cu deposit (e.g., Tonglvshan skarn deposit) and the red path represents the path of the reduced porphyry-skarn Cu deposit (e.g., Anqing skarn deposit).

Figure 9.
Temperature ( • C) vs. oxygen fugacity (logf O 2 ) diagrams [27,33,56]. The blue path shows the mineral precipitation sequence of the oxidized porphyry-skarn Cu deposit (e.g., Tonglvshan skarn deposit) and the red path represents the path of the reduced porphyry-skarn Cu deposit (e.g., Anqing skarn deposit). Figure 9. Temperature (°C) vs. oxygen fugacity (logfO2) diagrams [27,33,56]. The blue path shows the mineral precipitation sequence of the oxidized porphyry-skarn Cu deposit (e.g., Tonglvshan skarn deposit) and the red path represents the path of the reduced porphyry-skarn Cu deposit (e.g., Anqing skarn deposit). A recent study [27] summarizes several magmatic-hydrothermal ore deposits and proposes to use the Fe isotopic composition of chalcopyrite (δ 56 Fe chalcopyrite~− 0.1‰) as a limiting value (Figure 10b) to judge the redox conditions of deposits. Using this criterion, the Tonglvshan skarn deposit would be regarded as representing an oxidized hydrothermal system and the Anqing skarn deposit would reflect a reduced system. Combined with the phase diagram ( Figure 9) and petrographic analysis (Figures 3 and 5) in the oxidized system, magnetite is the mineral that precipitates first, while in the reductive system, pyrrhotite precipitates first. Hence, magnetite from the oxidized-Tonglvshan skarn deposit is used to calculate the equilibrium fluid Fe isotope composition (δ 56 Fe fluid = δ 56 Fe mt + ∆ 56 Fe Mt-fluid , the ∆ 56 Fe Mt-fluid = 0.6‰ at 350 • C), obtaining the heaviest δ 56 Fe~−0.393 ± 0.029‰ [43,69], which is consistent with Ref. [41]. In the reduced-Anqing skarn deposit, we chose pyrrhotite to calculate the equilibrium fluid Fe isotope composition (δ 56 Fe fluid = δ 56 Fe Po + ∆ 56 Fe Po-fluid , the ∆ 56 Fe Po-fluid = −1.12‰ at 350 • C) [46] and obtained the lightest δ 56 Fe~0.045 ± 0.057‰, which is similar to the results in [41], within a rate of error. The calculated Fe isotope compositions of the mineralization fluids for both deposits are lighter than the stock intrusions associated with skarns and also porphyry mineralization [23,24,27,32,33], which is in support of the "light fluid" hypothesis.
The chalcopyrite in the oxidized-Tonglvshan skarn deposit is lighter than that from the reduced-Anqing skarn deposit, which is best understood as relating to the redox state of ore fluids because of earlier precipitated heavy Fe magnetite from the oxidized-Tonglvshan system and because of earlier precipitated light Fe pyrrhotite from the reduced-Anqing system (Table 3, Figures 7,9 and 10) [27,32,33]. In the Tonglvshan skarn deposit, the deposited magnetite causes the fluid to become lighter, and the subsequently precipitated chalcopyrite thus has a lighter Fe isotopic composition, while in the Anqing skarn deposit, pyrrhotite incorporates light Fe isotopic composition, causing the fluid to become heavier-Fe and causes the chalcopyrite to have a heavier Fe isotopic composition ( Figure 10). Moreover, the wide range of δ 56 Fe in pyrrhotite and chalcopyrite from the reduced-Anqing skarn deposit may be contributed to Rayleigh fractionation of pyrrhotite during the hydrothermal evolution (Figure 10a) [27]. The overlap of the Fe isotopic composition of pyrite in the two deposits may be related to the formation pathway of pyrite and the relative timing of the fluid evolution (Figure 8).
In Section 3 above, we discussed the redox state of the ore fluid that governs the Fe isotopes composition of precipitated minerals [33]. In the Tonglvshan deposit, precipitation of magnetite with heavy Fe will cause the fluid to become lighter, which will impart lighter Fe in subsequently precipitated chalcopyrite from this fluid. In the Anqing deposit, precipitation of pyrrhotite with light Fe can cause the fluid to become heavier ( Figure 10) [27], which is expected to have great influence on the latter precipitated chalcopyrite. This is important. If we assume an initial δ 56 Fe value of 0‰ for fluid, magnetite precipitation can deplete the heavier Fe, resulting in light Fe in the liquid, readily reaching a value of δ 56 Fe =~−0.3‰ ( Figure 10). Likewise, precipitation of pyrrhotite can deplete the light Fe, resulting in heavy Fe in the liquid, readily reaching a value of δ 56 Fe =~0.3‰ ( Figure 10).
As mentioned in Section 1, we suggest equilibrium fractionation will also cause a great change in the Fe isotopic composition of pyrite (Figure 11a). With the FeS pathway mechanisms, i.e., the transient FeS-fluid equilibrium followed by pyrite-fluid equilibrium, we can readily explain the large Fe isotope compositional variation of pyrites in skarn deposits ( Figure 11). The δ 56 Fe for pyrites can be simply contributed to the Rayleigh fractionation model by the combination of pyrite-fluid and FeS-fluid equilibrium fractionations ( Figure 11). Our model shows that the pyrite-fluid equilibrium fractionations result in the light Fe isotopes boundary of the pyrite (Figure 11a,b), while the heavy Fe isotopes boundary of the pyrite is caused by the FeS-fluid equilibrium fractionations (Figure 11c,d). With all the observations and discussion above considered, we advocate that that the FeS-pathway effects on pyrite formation may be a common mechanism controlling the fluid Fe isotope composition in both oxidized (i.e., Tonglvshan) and reduced (i.e., Anqing) skarn ore-forming systems. tions ( Figure 11). Our model shows that the pyrite-fluid equilibrium fractionations result in the light Fe isotopes boundary of the pyrite (Figure 11a,b), while the heavy Fe isotopes boundary of the pyrite is caused by the FeS-fluid equilibrium fractionations (Figure 11c,d). With all the observations and discussion above considered, we advocate that that the FeSpathway effects on pyrite formation may be a common mechanism controlling the fluid Fe isotope composition in both oxidized (i.e., Tonglvshan) and reduced (i.e., Anqing) skarn ore-forming systems. Figure 11. Rayleigh fractionation modeling showing that δ 56 Fe (‰) variations in the residual fluid(red solid line), instantaneous(blue solid line ) and cumulative pyrite(blue dashed line) during pyrite precipitation. Instantaneous pyrite represents single pyrite formed at every instant and cumulative pyrite means bulk pyrite at every instant. The initial fluid Fe isotope composition for the oxidized systems is set to be −0.3‰ (a,c) and for the reduced system is + 0.3‰ (b,d). (a) Equilibrium fractionation between pyrite and fluid in the oxidized system; (b) Equilibrium fractionation between pyrite and fluid in the reduced system; (c) Equilibrium fractionation between FeS and fluid in the oxidized system and (d) Equilibrium fractionation between FeS and fluid in the reduced system. Equilibrium fractionation factors at 350 °C between pyrite and fluid is 0.99‰ from [44], while between the FeS and fluid is −0.31‰, calculated from [43,69]. The Rayleigh fractionation equations are as follows: δ 56 Fefluid = δ 56 Fefluid (initial) + Δ 56 Femineral-fluid × ln(F); δ 56 Femineral (instantaneous) = δ 56 Fefluid + Δ 56 Femineral-fluid; δ 56 Femineral (cumulative) = [δ 56 Fefluid (initial) − (1-F) × δ 56 Fefluid]/F; F is the Fe mass fraction in the remaining fluid. The yellow line stands for the Fe isotope compositional range of pyrite from the oxidized system and the green line represents the same for the reduced system [22,23]. Figure 11. Rayleigh fractionation modeling showing that δ 56 Fe (‰) variations in the residual fluid(red solid line), instantaneous(blue solid line ) and cumulative pyrite(blue dashed line) during pyrite precipitation. Instantaneous pyrite represents single pyrite formed at every instant and cumulative pyrite means bulk pyrite at every instant. The initial fluid Fe isotope composition for the oxidized systems is set to be −0.3‰ (a,c) and for the reduced system is + 0.3‰ (b,d). (a) Equilibrium fractionation between pyrite and fluid in the oxidized system; (b) Equilibrium fractionation between pyrite and fluid in the reduced system; (c) Equilibrium fractionation between FeS and fluid in the oxidized system and (d) Equilibrium fractionation between FeS and fluid in the reduced system. Equilibrium fractionation factors at 350 • C between pyrite and fluid is 0.99‰ from [44], while between the FeS and fluid is −0.31‰, calculated from [43,69]. The Rayleigh fractionation equations are as follows: δ 56 Fe fluid = δ 56 Fe fluid (initial) + ∆ 56 Fe mineral-fluid × ln(F); δ 56 Fe mineral (instantaneous) = δ 56 Fe fluid + ∆ 56 Fe mineral-fluid ; δ 56 Fe mineral (cumulative) = [δ 56 Fe fluid (initial) − (1 − F) × δ 56 Fe fluid ]/F; F is the Fe mass fraction in the remaining fluid. The yellow line stands for the Fe isotope compositional range of pyrite from the oxidized system and the green line represents the same for the reduced system [22,23].

Conclusions
This work, based on Fe isotope data on skarn ore mineral separates and petrography, led to the following important conclusions.

1.
We used magnetite from the oxidized-Tonglvshan skarn deposit and pyrrhotite from the reduced-Anqing skarn deposit to calculate the equilibrium Fe isotope compositions of the fluids, respectively. The calculated heaviest Fe isotope of the equilibrium fluid for the oxidized-Tonglvshan skarn deposit and the lightest Fe isotope of the equilibrium fluid for the reduced-Anqing skarn deposits are both lighter than the stock intrusions associated with skarn and also porphyry mineralization, supporting the "light fluid" hypothesis for granitoid magmatic fluids. Moreover, the chalcopyrite Fe isotope in the oxidized-Tonglvshan skarn deposit is lighter than that from the reduced-Anqing skarn deposit, which is controlled by the prior precipitation of magnetite (heavy Fe) from the ore fluid in the oxidized-Tonglvshan systems and the prior precipitation of pyrrhotite (light Fe) from the ore fluid in the reduced-Anqing system.

2.
The δ 56 Fe for pyrite shows an inverse correlation with δ 56 Fe for the magnetite in Tonglvshan ore samples. In both deposits, the Fe isotope fractionation between chalcopyrite and pyrite is offset from the equilibrium line at 350 • C and lies between the FeS-chalcopyrite equilibrium line and pyrite-chalcopyrite equilibrium line at 350 • C. These observations are best understood as the FeS pathway towards pyrite formation. That is, the initial FeS-fluid equilibrium Fe isotope fractionation is a critical step for continued pyrite-fluid equilibrium fractionation via increased extent of Fe isotopic exchange. With data from other ore deposits considered altogether, we advocate that the pathway effect on pyrite formation in the skarn deposit mineralization is important in controlling the Fe isotope fractionation.

Data Availability Statement:
The data presented in this study are available in the text.