The Sulfide/Silicate Coefficients of Nd and Sm: Geochemical “Fingerprints” for the Syn- and Epigenetic Cu-Ni-(PGE) Ores in the NE Fennoscandian Shield

: One of the current directions of the Sm-Nd isotope systematics development is a dating of the ore process using sulfide minerals. Yet, the issue of the existence of rare earth elements (REE) in sulfides is still a matter for discussion. Sulfides from ore-bearing rocks of Proterozoic (2.53–1.98 Ga) Cu-Ni and platinum group elements (PGE) deposits of the Fennoscandian Shield were studied. It is found that the most probable source of REE in sulfide minerals from Cu-Ni-PGE complexes could be submicronic fluid inclusions, which are trapped at the mineral crystallization stage. In such a case, fluid or melt inclusions are specimens of the syngenetic parental melt, from which the base mineral formed, and these reflect a composition of the parental fluid. The mineral–rock partition coefficients for Nd and Sm can be used as “fingerprints” for individual deposits, and these are iso-tope-geochemical indicators of the ore-caused fluid that is syngenetic to sulfide. Moreover, the D Nd /D Sm ratio for various sulfide minerals can be used as a prospective geochemical tool for reconstructing a mineral formation sequence in ore complexes. On the other hand, differences in isotope compositions of sulfide neodymium could be markers of some ore-caused fluids and related to certain generations of sulfide minerals.


Introduction
The samarium-neodymium system is one of the most demanded and informative isotope-geochronological tools for studying geological objects. It is successfully used for a wide age range, i.e., from the earliest point of Precambrian to Phanerozoic, and it provides reliable geochronological and geochemical data for different rock types. With accumulating knowledge on the behavior of REE in various geological processes, there is a matter of enhancing the Sm-Nd method capacity by using new minerals-geochronometers. One of the prospective areas is a dating of an ore process by the Sm-Nd method using sulfide minerals. A successful implementation of this approach on key ore objects of the Baltic Shield allowed defining major time points of ore formation of Proterozoic Cu-Ni-PGE complexes and geochronologically supports the conclusions that the ore process is syn-or epigenetic. The obtained Sm-Nd data are well-correlated with available U-Pb zircon and baddeleyite datings, which confirms the accuracy of the obtained results [1][2][3][4][5].
Our new data indicate a possible use of sulfides to determine the age of ore formation in mafic-ultramafic rocks, which allows extending the set of minerals suitable for the Sm-Nd method. The measured neodymium and samarium concentrations in the sulfides might be related to fluid or melt inclusions that reflect a composition of the parental orebearing melt. The determined coefficients of the sulfide/rock for Nd and Sm could be used as "fingerprints" for individual deposits, and they are some kinds of isotope-geochemical indicators of the ore-caused fluid that is syngenetic to the sulfide. Aside from that, the DNd/DSm ratio for various sulfide minerals can be used as a prospective geochemical tool for reconstructing a sequence of mineral formation in ore complexes.

Geological Settings
The Fennoscandian Shield hosts the vast Palaeoproterozoic East Scandinavian Mafic Large Igneous Province. Its current remnants cover about 1,000,000 km 2 . The shield basement formed as a mature Archaean granulite and gneiss-migmatite crust 2550 Ma ago. It is exposed in the Kola-Lapland-Karelia Craton. The main structural features of the East Scandinavian mafic Large Igneous Province and its Pd-Pt and Ni-Cu-PGE deposits are described in [27]. The exposed part of the shield extends beneath the sedimentary cover toward the northern Russian Platform as a vast Palaeoproterozoic Baltic-Mid-Russia wide arc-intracontinental orogen [28].
The long Early Palaeoproterozoic (2530-2400 Ma) geological history of the East Scandinavian mafic Large Igneous Province comprises several stages. They are separated by breaks in sedimentation and magmatic activity often marked by uplift erosion and the deposition of conglomerates. The Sumian stage (2550-2400 Ma) is crucial for the metallogeny of Pd-Pt ores. It can be related to the emplacement of high-Mg and high-Si boninite-like and anorthosite magmas [29,30]. The ore-bearing intrusions were emplaced in the Kola Belt (2530-2450 Ma) and in the Fenno-Karelian Belt (2450-2400 Ma) [2,3,5].
Recently, the Baltic Shield has been defined as the PGE-bearing East Scandinavian mafic Large Igneous Province of plume nature [2], or the Baltic Large Igneous Province with igneous rocks rich in Mg and Si [31], or the Kola-Lapland-Karelian plume province [32]. These Early Paleoproterozoic geological settings fill a substantial gap in the understanding of geological events and Pd-Pt and Ni-Cu metallogeny of the Late Neoarchaean-Early Palaeoproterozoic transitional period in the Earth's evolution (2.7-2.2 Ga ago). In classic metallogenic summaries [33][34][35], this period is characterized by the Stillwater, Great Dike of Zimbabwe, Bushveld, and Sudbury ore-bearing complexes. However, their geological setting cannot be coordinated in space and time with regional geological frameworks.

Pilgujärvi 1.98 Ga Deposit (Pechenga Area)
The Pilgujärvi Cu-Ni deposit occurs in the central part of the Eastern ore field [60,61]. The genesis and distribution of ore bodies in the Pilgujärvi deposit is primarily related to the internal structure and tectonics of the main differentiated gabbro-peridotite massif notwithstanding the fact that this deposit accommodates more than a dozen other smaller nickel-bearing intrusive bodies with a similar composition. The internal structure of the massif demonstrates a primary banding of the composing rocks and ores. The ore body is mostly made up of crude impregnated ores (95%) with minor rich impregnated breccia and massive sulfide ores (ca. 5%). The constant confinement of the deposits to the peridotitic bottom parts of the differentiated intrusives, the sheet-like shape and occurrence of the majority of the ore bodies, and ore structures and textures imply a syngenetic nature of the overwhelming ore mass with respect to the mother rock. This is a reason why the ore-forming process or the sulfide accumulation and segregation shall be considered as a phenomenon that accompanied the intrusive emplacement, differentiation, and chilling of the nickel-bearing magma at the turn of 1980 Ma [42].

Portimo 2.44 Ga Complex (Ahmavaara Intrusion)
The Portimo Complex accommodates the Ahmavaara, Narkaus, and Sukhanko-Konttijärvi intrusions with a crystallization age of 2.44 Ga, which presumably associate with a coeval dike swarm known as the Portimo dikes [51]. This complex is supposed to have formed at the expense of two various magmas with the early magma being richer in МgО, Сr, and LRЕЕ than the later one. Both magmas contained low ТiO2 (<0.5 wt %) and belong to the boninite series [62].

Monchegorsk 2.5 Ga Ore Field
The Monchegorsk ore field includes rocks of the ore-bearing layered Monchepluton, gabbro-anorthosite massifs of the Main Ridge (Monchetundra and Chunatundra), small intrusions of the Ostrovskoy massif and Yarva-Varaka, and dikes cutting through the Monchepluton [38,63,64]. The complex had been forming during a long period of time in the interval of 2.51-2.46 Ga [36,37,64].
At the modern erosional truncation, the Monchepluton has an arched shape and consists of two branches (or chambers). The northwestern branch is represented by the Nittis-Kumuzh'ya-Travyanaya (or NKT) massifs while the nearly east-west-trending one is composed of the Sopcha-Nyud-Poaz and Vurechuaivench massifs. The complex is made up of dunites, harzburgites, orthopyroxenites (NKT), orthopyroxenites (Sopcha), norites (Nyud), and gabbronorites (Poaz and Vurechuaivench) that form a single syngenetic rock series. It is distinctly differentiated both vertically and horizontally, which is generally expressed in the reduced rock basicity from the bottom up and from west to east [38,63,64].

Fedorovo-Pansky 2.5 Ga Layered Complex
The Fedorovo-Pansky layered complex occurs in the central part of the Kola Peninsula at a border between the Early-Proterozoic volcano-sedimentary rift series and Archean basement gneisses. The Fedorovo-Pansky complex mostly comprises gabbronorites with varying proportions of mafic minerals and different structural features. From the bottom up, the layered sequence is as follows [2,46]: Marginal Zone (50-100 m) of plagioclase-amphibole schists with relicts of massive fine-grained norite and gabbronorite, which are referred to as chilled margin rocks; Taxitic Zone (30-300 m), which contains an ore-bearing gabbronoritic matrix (2485 Ma) and early xenoliths of plagioclase-bearing pyroxenite and norite (2526-2516 Ma). Syngenetic and magmatic ores are represented by Cu and Ni sulfides with Pt, Pd, and Au, as well as Pt and Pd sulfides, bismuth-tellurides, and arsenides; Norite Zone (50-200 m) with cumulus interlayers of harzburgite and plagioclase-bearing pyroxenite that includes an intergranular injection Cu-Ni-PGE mineralization in the lower part.

Sulfide Geochronology of Cu-Ni-PGE Ore Deposits-Results of Previous Studies
For more than a decade, we have studied the main Cu-Ni-PGE deposits in the northeastern part of the Fennoscandian Shield to determine the age of their ore formation. The main results are shown in Figure 2. It can be seen that sulfide minerals together with rockforming minerals form good quality isochrones. This made it possible to determine the age of ore genesis for the main Cu-Ni-PGE ore deposits of the Fennoscandian Shield. In general, the Sm-Nd system of sulfide minerals is becoming a promising tool for studying ore processes. Along with syngenetic sulfides, young ages of redeposited and metamorphosed ores were also obtained for the Ahmavaara deposits (Finland) and the Monchegorsk region (Nyud-2). Further studies of the distribution of REEs in rocks and sulfides and their comparison led to important conclusions about isotopic fingerprints for syngenetic and epigenetic ores. However, there are serious limitations and problems associated with the use of sulfides as geochronometers for the ore process (see Results and Discussion).

Samples and Methods
Sulfides from Paleoproterozoic rocks (2.53-1.98 Ga) of Cu-Ni-PGE deposits of the Fennoscandian Shield were studied. The isotope research was carried out in the Collective Use Centre of the Kola Science Centre RAS (Apatity, Russia). First, the samples were prepared by crushing; then, minerals were separated using heavy liquids, and mineral fractions were selected under binocular microscope.

Sm-Nd Analytical Methods
In order to define concentrations of Sm and Nd, the sample was mixed with a compound tracer 149 Sm/ 150 Nd prior to dissolution. Then, it was diluted with a mixture of HF + HNO3 (or +HClO4) in Teflon sample bottles at a temperature of 100 °C until complete dissolution. Further extraction of Sm and Nd was carried out using standard procedures with two-stage ion-exchange and extraction-chromatographic separation using ion-exchange tar "Dowex" 50 × 8 in chromatographic columns employing 2.3 N and 4.5 N HCl as an eluent. The separated Sm and Nd fractions were transferred into a nitrate form, whereupon the samples (preparations) were ready for mass-spectrometric analysis.
The isotope Nd composition and Sm and Nd contents were measured at a 7-channel solid-phase mass-spectrometer Finnigan-MAT 262 (RPQ) in a static double-band mode, using Ta+Re filaments. A mean value of 143 Nd/ 144 Nd ratio in a JNdi-1 standard was 0.512081 ± 13 (N = 11) in the test period. An error in 147 Sm/ 144 Nd in ratios was 0.3% (2σ), which is a mean value for 7 measurements in a BСR-2 standard [65]. An error in estimation of isotope Nd composition in an individual analysis was up to 0.01% for minerals with low Sm and Nd contents. The blank intralaboratory contamination was 0.3 ng in Nd and 0.06 ng in Sm. The accuracy of estimation of Sm and Nd contents was ±0.5%. Isotope ratios were normalized per 146 Nd/ 144 Nd = 0.7219 and then recalculated for 143 Nd/ 144 Nd in JNdi-1 = 0.512115 [66].

Study of Sulfide Mineral Texture
Sulfides were studied using back-scattered electrons with a high-performance LEO 1450 scanning electron microscope. Analyses were carried to study possible inclusions in sulfides with considerable concentrations of REE that might have distorted results of Sm-Nd dating [67]. A preliminary study of sulfide minerals using BSE allowed finding small silicate microinclusions (Figure 3), which cannot have a substantial impact on the results of the Sm-Nd analysis. Moreover, to provide a higher study accuracy, the sulfides were sampled manually under a binocular microscope to discard minerals with visible microinclusions.

ICP-MS
To define REE in samples with no preliminary separation and concentration, reference values of REE concentrations in the GSO 2463 Standard (apatite), sulfide from the Talnakh deposit, and international standard samples of the Centre of Petrographic and Geochemical Research (Nancy, France) were reproduced using the ELAN 9000 DRC-e (Perkin Elmer, Waltham, MA, USA) quadrupole mass-spectrometer in Tananaev Institute of Chemistry KSC RAS (Apatity, Russia). The samples were separated under the conditions provided in [67].
The analysis was based on the following parameters: plasma power of 1300-1350 W; sprayer gas flow (high-clean Ar) within 0.75-1.0 L/min −1 ; ion lens voltage <11 V; and level of doubly charged and oxide ions <2.5%. A geological sample weighing up to 100 mg in the polystyrene hermetically sealed test tube mixed with distilled acids (HNO3, HF, HCl 5 mL each) was exposed to water bath at a temperature of 50-60 °С until fully dissolved. No HCl was added in the course of opening of sulfide minerals. When opening the sample, we have registered an increased pressure of acid and nitrogen oxide vapor that suppressed the volatility of the components in the sample. The chilled sample was mixed with 0.1 mL H2O2, and the dissolved sample was diluted with 2% HNO3. The level of total REE content in the blank sample was <0.5 ppb. This blank sample qualifies the level of analytical accuracy and limit of element detection. Since the samples have yielded high concentrations of those elements that may cause a matrix effect and ion interference, the calibration curves were plotted with interfering agent added to the blank calibration solution. The multicomponent standard solutions by Perkin Elmer (Multi-element ICP-MS Calibration Std, Waltham, MA, USA) were used. The sample itself was selected as an interfering agent. The amount of the interfering agent was chosen so that the macrocomponent concentration after mixing exceeded the REE concentration in the calibration solution by a factor of 100. The approximation linearity of the REE correction curves is ≥99.99%. Spectral superimposition was recognized by ELAN 9000 DRC-e MS software (Perkin Elmer, Waltham, MA, USA) and adjusted by the introduction of correction equations into the analytic program defined with reference to a natural abundance of REE isotopes. The blank sample solution free of the interfering agent was used to analyze the solution of the opened sample.

Results and Discussion
A total of 34 monofractions of sulfide minerals from Paleoproterozoic-layered complexes of the Fennoscandian Shield were studied using mass spectrometry (Table 1).  In general, the observed variations in values are quite wide, but the DNd/DSm ratio is more stable. Most of the DNd/DSm values are within the 0.70-1.70 range. Only for redeposited or metamorphosed ores, this ratio is sharply increased and varies from 2.30 to 7.70. At the same time, on the diagram in DNd-DSm coordinates (Figure 4), sulfide points form a distinct trend, reflecting the relationship between the Nd and Sm partition coefficients. The sulfide figurative points from the redeposited Ahmavaara ores, the massive Pilgujärvi ores, and the metamorphized ores of the Nyud-2 deposit are located above the trend line. The greater mobility of neodymium as compared to samarium during superimposed processes probably affects the redistribution of Nd by rock and sulfide, causing an increase in the partition coefficient. Thus, due to the epigenetic process, the DNd/DSm ratio tends to increase. There are a few hypotheses on existing REE in sulfide minerals: isomorphic substitution of major cations [11,68]; silicate microinclusions [9,69]; defects of mineral crystalline lattice [70]; sorption on mineral surfaces [13]; and fluid inclusions in sulfides with the inherited REE composition from the ore-bearing melt [21][22][23].
Despite the fact that finding REE in sulfide minerals is contentious, different researchers are sufficiently confident to suggest that the type of REE partition in sulfides is inherited from the ore-bearing fluid, impacting the formation of sulfide minerals. Our studies ( Figure 5, Table 3) on the partition of REE in the sulfides from ores of the Pilgujärvi deposit (Pechenga) and the Penikat intrusion (Finland) revealed that the type of REE partition in sulfides repeats the REE partition in the rock [71].  Analogous patterns of REE partition are detected in sulfides and many hydrothermal deposits, for which a role of fluids in ore genesis is defining [19,23]. Therefore, the most probable source of REE in the studied sulfides is heterogeneous fluid or melt inclusions occurring as submicronic bubbles. Additional arguments in favor of this suggestion are studies on silicate microinclusions in sulfides as a possible REE source. It was determined that the REE composition of silicate microinclusions is part of the overall REE balance of the fluid from which the sulfide crystallized. Thus, the total REE composition in a mineral can be taken as a reflection of the ore-forming fluid composition [9,69]. From a practical point of view, in terms of the Sm-Nd method, this means that the position of the sulfide points on the isochron will be controlled by different distribution coefficients of Nd and Sm, which is associated with different fluid inclusions. However, these inclusions are syngenetic to sulfide and rock-forming minerals. Therefore, sulfides can be used to determine the age of ore genesis, along with rock-forming minerals.
One of the issues on REE behavior in magmatic systems of Cu-Ni-PGE deposits is to determine a possibility that there are REE-bearing fluids at temperatures of the magmatic process. A hydrothermal process is characterized by much lower temperatures (generally lower than 400-450 °С) compared to temperatures for the crystallization of magmatic sulfides of Cu-Ni-PGE complexes, i.e., more than 1000 °С. Yet, studies on heterophase fluid inclusions in rocks of the Merensky reef (Bushveld) justify a presence of the ore-bearing fluid at temperatures of 900 °С and higher [72,73]. This means that already at temperatures of sulfide formation, ore fluid exists that can be captured by sulfide in the form of submicron inclusions.
Taking into account the arguments presented, the most probable source of REE in sulfide minerals from Cu-Ni-PGE complexes can be submicronic fluid inclusions, which are trapped at the early stage of mineral crystallization. In this case, fluid or melt inclusions are specimens of the syngenetic parental melt, from which the base mineral crystallized, and reflect a composition of the parental fluid. Thus, the determined coefficients of the mineral/rock partition for Nd and Sm will characterize a genesis of ore matter of a certain deposit and define isotope indicators (fingerprints) of the ore process. Moreover, they provide isotope-geochemical data on the ore-forming fluid and allow dating the ore process directly. The DNd/DSm = 1.4 ratio obtained for all of the minerals is well-correlated with the interval of values obtained in the work [74], where a dependence of partition coefficients on FeO content and temperature is shown based on an experiment with sulfides under reduction conditions. It is found that if the temperature decreases from 2100 to 1400 °С at 1.5 GPa, the DNd/DSm ratio tends to be in the range of 1.3-1.5. For natural sulfides, it can underlie different partition coefficients depending on the temperature occurring when cooling and their sequential formation.
E.g., for massive syngenetic ores of Ahmavaara, the DNd/DSm ratio is, on average, 1.3, while for the redeposited ore, the DNd/DSm ratio increases up to 4.8 (Table 1). It can indicate a significant hydrothermal impact leading to an increase in neodymium mobility and its migration of diffusion into fluid inclusions of the sulfide (Figure 4). The consequence of this is a relative accumulation of neodymium as compared to samarium and the increasing DNd/DSm ratio. By contrast, for ore gabbronorites of the Penikat intrusion and the Fedorovo-Pansky layered complex, the DNd/DSm ratio shows a rather narrow range of values, i.e., 1.2-1.3 (Table 1). Along with similar Sm-Nd (sulfide) and U-Pb ages, it indicates a syngenetic genesis of the sulfides and rocks of the complexes, which is testified by geological and mineralogical observations. Apart from it, a trend of the increasing DNd/DSm ratio in the pyrite-chalcopyrite-pyrrhotite-pentlandite sequence was determined ( Figure 6, Table 2). Figure 6. DNd/DSm ratio trend in the pyrite-chalcopyrite-pyrrhotite-pentlandite sequence. Higher values of this ratio are found for epigenetic or hydrothermal ores. This may be due to the hydrothermal processes of sulfide redeposition or the influence of a late fluid with different geochemical characteristics. An increase in the DNd/DSm ratio is also likely associated with a melt temperature decreasing and can be used to reconstruct the sequence of sulfide crystallization. Dotted arrow indicates an estimated decreasing trend in temperature. This offers prospects for reconstructing a possible sequence of sulfide crystallization and defining conditions for ore formation in intrusive complexes of different age and genesis. Using these new data, it seems intuitively reasonable to implement the obtained results for analyzing other deposits, e.g., three stages of mineral formation with the pyritechalcopyrite-pyrrhotite-pentlandite sequence were defined for the Kun-Manye Cu-Ni deposit (Russia) [75]. For the Irankuh Zn-Pb deposit (Iran), a pyrite-chalcopyrite sequence of ore mineral formation is also defined [24]. The analysis of the DNd/DSm ratios for sulfides of various genesis shows that values of DNd/DSm increase for minerals of later processes, which correspond to ore redeposition or hydrothermal impact. Moreover, different isotope compositions of sulfide neodymium can be indicators of ore-caused fluids and related to certain generations of sulfide minerals, as shown for the Bathtub intrusion, Duluth Complex [76]. Yet, the issues related to the numerous episodes of hydrothermal processes or influence of some fluids with various isotope characteristics during the ore formation are being extensively studied [76][77][78][79].

Conclusions
In general, the proposed approach considerably extends the available tool set of the Sm-Nd method. Apart from isotope-geological data, the results of the conducted study provide additional geochemical data on ore genesis processes at sulfide mineral crystallization stages, as well as at the following stages of hydrothermal impact.
Therefore: (1) The DNd and DSm partition coefficients were determined for various sulfide types, which is consistent with the experimental data; (2) The most probable REE source of sulfides is an ore-forming fluid appearing at the very stage of formation of sulfides from the melt at temperatures above 900 °С; (3) Prospects for defining a possible sequence of sulfide crystallization and reconstructing ore formation conditions in intrusive complexes of different ages were studied; (4) Differences in isotope compositions of sulfide neodymium could be indicators of some ore-caused fluids and related to certain generations of sulfide minerals.