Main characteristics and genesis of the Vale de Pães skarn (Cuba-Vidigueira, Ossa Morena Zone, Portugal)

The Vale de Pães (Cuba-Vidigueira) mineralisation is composed of magnetite ± sulphides and hosted in a Pre-Variscan metamorphic sequence intruded by igneous rocks belonging to the Beja Igneous Complex. Its mineral and chemical features are compatible with a zoned Fe-skarn: Mg-rich (Fo + Di90, oxidised) and Ca-rich (Grs + Di81-39, oxidised or relatively reduced). In the Fe-Mg skarn, magnetite deposition occurred along with the anhydrous mineral assemblage at ≈ 600 °C; sulphides precipitated from the retrograde stage onset (≤ 550 °C) and during the hydrated and carbonate phases formation period (< 420 °C). In the Fe-Ca skarn, magnetite precipitated during the retrograde stage (< 550 °C) together with the hydrated mineral association, and was followed by sulphides at ≈ 400°C. The mineralising process involved moderate-high salinity fluids and was controlled by variations in redox potential and pH.


Introduction
The Vale de Pães (Cuba-Vidigueira) mineralisation is part of a set of small ore deposits included in the Magnetite-Zinc Belt of the Ossa-Morena Zone (e.g., Carvalho, 1976, Oliveira 1986).It was recognized during exploration surveys carried out by the Serviço de Fomento Mineiro (SFM) since 1958.They included geological mapping (1:25000 scale), magnetometry and gravimetry surveys, and 25 drillholes designed to intersect possible mineralised bodies suggested by magnetic anomalies.The hidden Vale de Pães deposit extends in depth up to 180 m and contains ≈ 9 Mt with average grades of Main characteristics and genesis of the Vale de Pães skarn (Cuba-Vidigueira, Ossa Morena Zone, Portugal) Principales características y génesis del skarn de Vale de Pães (Cuba-Vidigueira, Ossa Morena Zone, Portugal) R. Salgueiro 1 , C. Inverno 2 , A. Mateus 3 ABSTRACT The Vale de Pães (Cuba-Vidigueira) mineralisation is composed of magnetite ± sulphides and hosted in a Pre-Variscan metamorphic sequence intruded by igneous rocks belonging to the Beja Igneous Complex.Its mineral and chemical features are compatible with a zoned Fe-skarn: oxidised) and  , oxidised or relatively reduced).In the Fe-Mg skarn, magnetite deposition occurred along with the anhydrous mineral assemblage at ≈ 600 °C; sulphides precipitated from the retrograde stage onset (≤ 550 °C) and during the hydrated and carbonate phases formation period (< 420 °C).In the Fe-Ca skarn, magnetite precipitated during the retrograde stage (< 550 °C) together with the hydrated mineral association, and was followed by sulphides at ≈ 400°C.The mineralising process involved moderate-high salinity fluids and was controlled by variations in redox potential and pH.
42 wt% Fe, 19 wt% SiO 2 and 0.6-5.2wt% S (Carvalho, 1976, Oliveira 1986, Carvalho & Oliveira, 1992).Comprehensive examination of samples collected in one of the drill-holes performed in this area (SD42) allowed to recognise for the first time features that are crucial to the characterisation of these iron ores.The observed mineral assemblages, typical of skarn type deposits, include several calcsilicate minerals, magnetite and sulphides; this deposit typology is likewise compatible with the geological framework in which the Vale de Pães mineralisation developed.The present work intends to report and discuss these data, characterising the mineralised domains and the processes involved in the genesis of the Vale de Pães ore-forming system.

Sampling and analytical procedures
Representative samples were selected from the SD42 drill-hole core after comprehensive logging.Rock samples were analysed for major elements (Si, Ti, Al, Fe, Mg, Mn, Ca, Na, K, P) and some trace elements (Ba, Rb, S, Sr, Nb, Zr, Y, Sn) by Xray-fluorescence spectrometry (XRF) performed at the ex-Instituto Geológico e Mineiro (IGM) Laboratory (Portugal); they were also analysed for Au by atomic absorption (AA) and for Cd and B by Coupled Plasma-Atomic Emission Spectrometry (DCP-AES) in the same laboratory; the estimated precision analytical errors were usually < 5-10% for major and trace elements.Complementary analyses were obtained at the Activation Laboratories Ltd (Canada) using the research grade package for Au + 48 elements that combines Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES; including Ni, V, Cu, Pb, Zn, Bi, Mo, Be, Ag) and Instrumental Neutron Activation Analyses (INAA; including La, Ce, Nd, Sm, Eu, Tb, Yb, Lu Cs, U, Th, Ta, Hf, Cr, Co Sc, W, Br, Ir, Au, Hg, As, Se, Sb); in terms of precision, the estimated analytical errors range from 5-10% on the measured content.
Electron microprobe analyses of pyroxene and olivine were made with a JEOL JCXA 733-CG routinely operated with an acceleration potential of 15 kV and a beam current of 25 nA; the different mineral elements were analysed for 20s each; the analytical quality was controlled through measuring of natural and synthetic mineral standards before, during and after each work period; the analytical errors were less than 2% for the analysed elements.

Geological setting
In the Vale de Pães area, different igneous rocks (gabbros, diorites, granites and porphyries) belonging to the Beja Igneous Complex (BIC) intrude a pre-Variscan metamorphic suite containing marbles, amphibolic gneisses, chlorite-rich schists, amphibolites and metavolcanic (quartz-feldspar) rocks (Fig. 1A); contact domains between the two rocksuites are traced by undifferentiated hornfels and amphibolites (e.g., Piçarra et al., 1992).The SD42 drill-hole is located in a contact domain and the recovered drill-hole core reveals a sequence of (volcanic, subvolcanic or intrusive) rocks that were metamorphosed in greenschist to amphibolite facies conditions (Salgueiro, in prep).Meta-felsic intercalations are present from 26 to 166.5 m depth, the thickest horizon occurring between 160 and 166.5 m (Fig. 1B).BIC intrusive rocks form different kinds of veins, veinlets and pseudo-intercalations, the main one identified between 185 and 192.04 m depth; quartz-monzonite (21.58-22.70 and 67.80-69.50 m) and granite (54.21-56.00m) mineral assemblages appear to dominate in these BIC rocks (Fig. 1B).It is also noteworthy the presence of a meta-dolostone horizon between 178 and 185 m depth.

Petrography, mineralogy and relative chronology of deposition
The Fe-Mg skarn matrix is mostly composed of forsterite (Fo 67 ) and diopside (Di ≈90 ) arranged in a coarse-grained granoblastic texture (Fig. 2A, B and Table 1).These minerals, tracing the prograde stage of skarn development (Fig. 3), are surrounded by Mgt I (predominant) and pyrrhotite (Po) I. Mgt I forms massive but heterogeneously fractured aggregates that include (sub)euhedral grains usually displaying spinel (Spl) exsolutions; its deposition took place during the final steps of the prograde stage.The shallower mineralised domains are characterised by the predominance of massive Mgt showing lesser fracturing, as well as minor re-crystallisation.Po I occurs as massive, irregular, variably fractured aggre-gates, developing complex inter-granular relations with Mgt I that suggest close periods of deposition or reflect imbalance/balance conditions established during Mgt I re-crystallisation (in the retrograding stage).Po I deposition precedes the development of hydrous phyllosilicates belonging to the serpentine group (Serp), as well as a fracturing event (affecting this sulphide and Mgt I), possibly related to the volume increase in consequence of serpentinisation.Pyrite (Py) I forms, usually, idiomorphic grains and develops textural relationships with other mineral phases of dubious interpretation; nonetheless, it appears to be related to the local early change of Po I, further recorded by Py II + Mgt II exsolutions and, sometimes, marcasite (Mrc).Chalcopyrite (Ccp)  ments.A carbonatisation event experienced by the Fe-Mg skarn took place after Mgt I re-crystallisation, leading to carbonate deposition in veins or differential replacement of the matrix-forming minerals before development of the second generation of sulphides (Py II and Po II).Hydration processes, such as serpentinisation, are synchronous or run shortly after carbonatisation, and play an important role on the heterogeneous replacement of pyroxene and olivine relics; amphibole (Amph) and phlogopite (?), although rare, complete the mineral paragenesis developed during the retrograde stage.The chart in Fig. 3 summarises the paragenetic information concerning the Fe-Mg skarn.
Contents of some redox sensitive elements (RSE), such as Cr, Mn, Co, Ni, As, Sb, Th and U, were normalised to the North American Shale Composite (NASC, Gromet et al., 1984); V values were normalised to the Marine Shale of Ruhr (Degens et al., 1958) and those of Mo to the Recent Sediments (Wedepohl, 1974).Other reference elements, such as Ti, Pb, Zn, Cu and Cd, were included in this approach, normalising their contents relatively to NASC (Ti) and to Standard Shales (Turekian & Wedepohl, 1961).Plots of the normalised RSE con-R.Salgueiro (Boynton, 1984)

normalised REE patterns. B) RSE and other reference elements normalized patterns (see text for information about the standards used). Samples of the Vale de
Pães Fe-Ca (SD42/95) and SD42/155 and SD42/156.46).Concentrations below the detection limit of the analytical method used were not considered.

Discussion
Considering the matrix-forming minerals present in mineralised domains, critical information can be inferred by considering the oxidation state reflected by mineral paragenesis developed in the prograde stage (Einaudi et al., 1981, Einaudi & Burt, 1982).Indeed, according to Newberry (1983), the compositions displayed by the grossularite relics in the Vale de Pães Fe-Ca skarn (SD42/95) correspond to those generated in equilibrium with reducing fluids.Additionally, the analysed pyroxenes show significant oscillation in hedenbergitic contents, besides their tendency to increase in the Fe-Ca skarn domains; in the latter, Di 39 compositions are known in sample SD42/95, suggesting further media reduction (e.g., Einaudi et al., 1981).The mineral paragenesis tracing the final step of the prograde stage experienced by the Fe-Mg skarn (Fo + Di ≈90 + Mgt), suggests that the upper limit of a(O 2 ) was closer to that ruled by the Mgt/Hem buffer.In transition to the retrograde stage, redox media conditions and Fe availability allowed the precipitation of sulphides (Po I and Py I?, which persisted after the carbonisation process), favouring the development of Mgt-Hc exolutions as well.The redox conditions at this stage were also enough to cause the instability of olivine, leading to the generation of phyllosilicates (serpentine group) and Mgt II.The presence of Fe oxides in late veins indicates that a(O 2 ) conditions near the buffer Hem/Mgt were achieved during the final steps of the retrograde stage.
The distribution of RSE contents testifies enrichments in Cu, Cd, Zn, Co, Ni and Mo (Fig. 5B), consistent with an increase of sulphides (particularly abundant in the Fe-Mg skarn), also agreeing with the prevalence of reducing conditions during the retrograde stage.
The incipient Mn enrichment (a minor component in Mgt and Spl) is interpreted as reflecting the limited availability of this metal, determined by the original concentration in protolith and/or influenced by its solubility in metasomatic fluids.In fact, the differential mobility of chemical elements is a determinant factor for chemical zoning of many mineralising systems (including those of skarn type) and, in the case of Mn, the enrichment trend is focused on domains far from the source magma (Meinert, 1992(Meinert, , 1997)).According to the available data, Mn contents are higher in Fe-Ca skarn intervals [incorporated in pyroxenes (Table 1) and in Mgt, Salgueiro, in prep.],hence suggesting the possibility of a position further away from the magmatic source compared to the Fe-Mg skarn position; this inference must be, however, unequivocally demonstrated with comprehensive analytical studies.
Pyroxenes in the Vale de Pães Fe-Mg skarn are mainly diopsidic (Di ≈90 ) and show Al 2 O 3 contents (average ≈ 1:00 wt%) similar to specimens that are typical of magmatic skarns according to Zharikov (1970).However, the fayalitic component of the olivine included in the (anhydrous) prograde mineral paragenesis (Fa 33 ) is significantly above the values reported by Zharikov (1970) for olivines in those magmatic skarns (i.e., Fa 5-15 ).These data, taken together with the period of Mgt precipitation, can be interpreted as a consequence of the compositional characteristics of the magmatic component involved in the prograde stage, which is consistent with the possibility of placing the Fe-Mg skarn near the magma source.Based on experimental data for the Fo + Di + Spl equilibrium (e.g., Aleksandrov, 1998), early stages of the Fe-Mg skarn development must have occurred at temperatures ≈ 600 °C.This inference is compatible with temperature values calculated for plagioclase (with evidence of re-crystallisation) -edenite (late) pairs included in host rocks (≤ 600 °C) (Salgueiro, in prep), tracing, quite possibly, the heat peak related to the emplacement/cooling of the magmatic intrusion that triggered the development of the ore-forming system.
The Fe-Mg skarn shows ΣREE low values, which can be interpreted as an inherited geochemical characteristic from the protolith subsequently replaced, although the hydrothermal mobility of REE may contribute, at least partly, to explain those totals and some of the features of their normalised concentra-tion patterns.Indeed, the relative impoverishment of LREE in the Fe-Mg skarn may be related to the circulation of aqueous-carbonic fluids in reducing conditions like those typically involved in carbonatisation processes.Likewise, the positive Ce anomaly in sample SD42/126.30(Fe-Mg skarn) may reflect anoxic conditions, mainly achieved during the deposition of sulphides (quite significant in this sample) in the course of the retrograde stage.Finally, the negative Eu anomalies (Fig. 5B) are indicative of significant interaction with moderately oxygenated hydrothermal fluids (e.g., Michard & Albarède, 1986).
Following closely the approach proposed by Einaudi et al. (1981) and considering XCO 2 ≤ 0.1, the precipitation of carbonates should have occurred at temperatures below 550 °C, which is compatible with the fact that carbonatisation took place after Mgt I re-crystallisation.Olivine hydration (serpentinisation) in the Fe-Mg skarn should trace a new stage in the evolution of the system, being almost synchronous to the carbonate deposition; despite the serpentine mineral group analysed (Salgueiro, in prep) do not present ideal stoichiometry due to subsequent changes, their deposition must have occurred at temperatures below 420 °C, as reported by Greenwood (1967b, in Einaudi et al., 1981) for similar situations characterised by low partial pressure of volatile and XCO 2 < 0.05.Equivalent temperature (less than ≈ 420 °C) conditions can be deduced for the deposition of Py II, Po II and Ccp taking into account their position on the paragenetic sequence.The development of Po II and Py II crypto-crystalline aggregates is interpreted as evidence of abrupt temperature drop following the relative media reduction, pH decrease and a(S 2 ) increase.
The formation of prehnite in Fe-Ca skarn suggests that the late stages of hydration took place under conditions of low fluid pressure, which, according to the approach outlined by Einaudi et al. (1981) for similar conditions, should have occurred at temperatures below 400 °C.In this context, the scarcity of Mrc as a result of Po destabilisation, suggests that the decrease in pH during the retrograde stage (also found in iron-magnesian skarn) was not enough to favour the development of this sulphide.Additionally, the relative rarity of Ccp (particularly in the Fe-Ca skarn) indicates relative paucity of Cu in the system since there are no physical-chemical reasons that prevent Ccp deposition.Moreover, in this geochemical environment, it is expected that the process had involved mineralising fluids with moderate to high salinity, which is consistent with the relative enrichment in Na reported by the development of mg-hastingsite/pargasite in the Fe-Ca skarn and amphiboles with significant edenitic component in the host rocks (Salgueiro, in prep).
Considering several skarn studies (e.g., Einaudi et al., 1981, Meinert, 1992, 1997, and Pons et al., 2009), the spatially coexistence of two skarn types, as interpreted for the Vale de Pães deposit, may be the result of: 1) an overlapping of iron-calcic skarn on ironmagnesian skarn; 2) heterogeneous fluid/rock ratio due to different protolith features; 3) multiple magmatic source fluids and variable interaction with (different?) host protholith; and 4) differences in progression of equilibrium reactions within the metasomatic front.The geological, mineralogical and geochemical data herein documented for the Vale de Pães deposit, suggest, as a first approach that, with the exception of the first hypothesis, all the others are possible, either plain or modified.However, these hypotheses must be comprehensively examined in detailed studies.

Conclusions
During formation of the prograde Fe-Mg and Fe-Ca skarn mineral parageneses only slight a(O 2 ) variations took place; these mineral parageneses are generally oxidised, but in Fe-Ca skarn a relatively reduced media can be inferred.The physico-chemical conditions (decrease of pH, temperature and increase in the a(S 2 )) achieved during the retrogradation stage were crucial for sulphide deposition in the Fe-Mg skarn, influencing also the observed chemical zonation.In particularly, the circulation of aqueous-carbonic fluids during the late carbonatisation event experienced by the Fe-Mg skarn must have influenced REE mobilization.
Some relevant features (such as the whole-rock Mn and Al contents, and Fa component of olivine), together with the period of Mgt precipitation, suggest that the Fe-Mg skarn is positioned relatively close to the magma source.

Fig
Fig. 1.-A) Geological map of the area surrounding the Vale de Pães deposit (adapted from Serviço de Fomento Mineiro, 1980 approx.) and B) schematic representation of the SD42 drill-hole core log (adapted from Serviço de Fomento Mineiro, 1968).