Fe-, Fe,Mn- and Fe,Mg-chlorite: a genetic linkage to W, (Cu,Mo) mineralization in the magmatic-hydrothermal system at Borralha, northern Portugal

ABSTRACT A genetic linkage between W, (Cu, Mo)-mineralization and chlorite minerals, and the discrimination of different mineralization events in the magmatic-hydrothermal system of Borralha, northern Portugal, is discussed on the basis of textural relationships, crystal chemistry and stable isotopic data obtained from chlorite. Chlorite minerals were identified in assemblages with quartz, feldspars, tungstates and sulfides. X-ray diffraction studies of selected chlorite minerals shows a trioctahedral structural type. Electron probe micro-analyses identified four different compositions and associations: (1) Fe,Mn-chlorite with scheelite I; (2) Fe-chlorite with wolframite + scheelite II ± sulfide; (3) Fe,Mg-chlorite with molybdenite + bismuthinite; and (4) Mg,Fe-chlorite with chalcopyrite. The composition of Fe-chlorite (Al3.01Fe3+0.25Fe2+7.95Mn0.26Mg0.19)11.66(Si5.44Al2.56)8O20(OH)8 corresponds to daphnite and Fe,Mn-chlorite (Al2.69Fe3+0.02 Fe2+7.54Mn1.08Mg0.62)11.89(Si5.31Al2.68)4O20(OH)8 to a mixed composition between daphnite and amesite. The Fe,Mg-chlorite (Al2.89Fe3+0.24Fe2+6.42Mn0.21Mg2.08)11.84 (Si5.31Al2.79)8F0.31O20(OH)8 corresponds to ripidolite and Mg,Fe-chlorite (Al2.63Fe3+0.37Fe2+1.72Mn0.01Mg6.40Ca0.26)11.39(Si6.02Al1.98)8O20(OH)8 to pychnochlorite. Chlorite geothermometry estimates a temperature for Fe,Mn-chlorite (scheelite I) from 400°C to 500°C, for Fe-chlorite (Mn-rich wolframite + scheelite II ± sulfide) from 250 to 350°C, for Fe,Mg-chlorite (Mo-mineralization) from 200°C to 250°C and for Mg,Fe-chlorite at ~150°C. Oxygen isotopes (V-SMOW) yielded values of +3.8 (1σ) (Fe-chlorite), +6.91 (1σ) (Fe,Mn-chlorite) and +1.5 (1σ) (Fe,Mg-chlorite). The calculated δ18OF of Fe- and Fe,Mg-chlorite is ~+3.75 (1σ) and +1.45 (1σ) for the mineralizing fluid, whereas for Fe,Mn-chlorite it is +8.17 (1σ). The δ18O data obtained from quartz in W- and Mo-mineralization yielded values of +12.6 and +11.4 (1σ), whereas for adularia δ18O is about +10 (1σ). These estimates allow us to conclude that the Fe,Mn-chlorite crystallized from a magmatic-hydrothermal fluid, whereas the Fe- and Fe,Mg-chlorite quartz and adularia resulted from a mixed contribution between meteoric and magmatic-hydrothermal fluid.


Introduction
THE metallogenic processes intimately associated with a variety of late-to post-magmatic events are sometimes poorly understood in detail due to a wide compositional diversity of the fluids involved in metal transport and deposition, where the late silicate phases may be an indicator of a distinct stage of mineralization. One of the most common late silicate phases is chlorite. It is frequently found either in hydrothermal alteration haloes that host ores or accompanying the ore-forming phases. The chlorite represents an alteration product following the breakdown of different primary minerals as the H + or Mg 2+ ,Fe 2+ -metasomatism proceeds. These reactions are usually related to propylitic alteration (i.e. actinolite + epidote + chlorite or epidote + chlorite + calcite or chlorite + calcite + hematite assemblages). The reactions are quite common and well developed in a wide range of ore-forming systems, including iron-oxide copper gold, porphyry Cu, Cu-Mo or Cu-Au, epithermal Au-Ag, and retrograde alteration stages of skarn-type deposits (e.g. Lowell and Guilbert, 1970;Gustafson and Hunt, 1975;Meinert, 1992;Sillitoe, 2000Sillitoe, , 2010Cooke et al., 2014). The chemical composition of chlorites from the distal fringes of hydrothermal envelopes is distinct from that characteristic of metamorphism.
The wide range of crystal chemistry and structure displayed by the chlorite minerals group may also be used as a tool to determine the direction towards the mineralized area, to estimate the temperature under which the ore process took place (e.g. Cathelineau and Nieva, 1985;Walshe, 1986;Inoue et al., 2010) or to estimate the heat flux from a magmatic-hydrothermal centre (Sillitoe, 2010). Recently, the relationships 'chlorite vs. metal assemblages' were revisited and explored to use chlorite as a proxy for detecting porphyry ore deposits (Wilkinson et al., 2015).
Dispersion or remobilization of metals in longlived magmatic hydrothermal systems occurs via circulation of compositionally diverse fluids under a relatively large range of temperatures, generating useful environmental markers when ore-forming minerals are deposited with other phases such as chlorite. When precipitated directly from the hydrothermal fluid, chlorite displays a chemical composition that reflects many of the constraints imposed by the fluid chemistry during mineralization. For example, a genetic linkage between Mg, Fe-chlorite or Fe,Mg-chlorite and Cu-Au mineralization (Dora and Randive, 2015) or a Pb-Zn, Bi, Ag vein-type ore deposit (Inoue et al., 2010) has been established. Beside the relative enrichment of Fe or Mg in the octahedral sheet in chlorite, the systematic presence of measurable amounts of Mn has been tested as a new pathfinder for chlorite developed during an ore-forming processes. For instance, Fe,Mg,Mn-chlorite has been identified in Pb-Zn, Bi, Ag ore veins from the Toyoha geothermal system, Japan (Inoue et al., 2010), and Fe,Mn,Mg-chlorite was recognized in the Zn-Pb, Ag vein-type ore epithermal system in NE Chile (Chinchilla et al., 2016).
The main goal of the present contribution is to characterize the optical, structural and crystal chemistry of chlorite minerals associated with tungstate and sulfide mineralization from the W, (Cu,Mo) ore deposit of Borralha (Portugal). The relationships between the development of chlorite minerals and different ore-forming stages as well as discrimination of different mineralization events based on stable isotopic studies of chlorite minerals are addressed.

Geology
The Variscan orogeny is an obduction-collision orogenic event (e.g. Ribeiro et al. 1990Ribeiro et al. , 2007Matte, 1991), where three main deformation phases (D1, D2 and D3) have been documented in northern Portugal (Ribeiro, 1974;Noronha et al., 1981). The maximum crustal thickening was achieved by the end of D1/D2 phases, whereas the D3 phase (intra-Westphalian age) relates to the final stages of the continental collision process. Most of the granite intrusions and the associated thermal metamorphic peak are coeval with the D3 phase (e.g. Ferreira et al., 1987;Dias and Ribeiro, 1995;Noronha et al., 2000;Ribeiro et al., 2007). The syn-to late-orogenic magmatic activity generated voluminous granitoid batholiths, often with contrasting compositions, and the emplacement was, in most cases, controlled by major Variscan structures.
The W, (Cu,Mo) deposit of Borralha is located in the northwestern Iberian Peninsula near the boundary between the Central Iberian Zone and the Galicia Trás-os-Montes subzone (Fig. 1), where different synorogenic Variscan granites intrude Palaeozoic metasedimentary rocks to form twomica peraluminous (syn-D3) granites and biotiterich (syn-D3) granites. The ore deposit occurs at the contact between metasedimentary formations (Silurian?) and the syn-D3 porphyritic Borralha biotite granite (∼315 Ma) and a syn-D3 two-mica S260 granite (Fig. 2) assumed to be related to a concealed post-D3 granite body (i.e. the Gerês granite).
Petrogenetic and petrophysical studies of the post-D3 Gerês granite show that the magma rose relatively high in the crust and settled at upper crustal levels (6-7 km), generating a temperature gradient which triggered contact metamorphism at 200 MPa and 500-600°C which was sufficient to promote a vigorous hydrothermal convection cell (Noronha and Ribeiro, 1983;Noronha et al., 2000).
The Borralha ore deposit is an important tungsten deposit of Portugal second only to Panasqueira. It was mined from 1903 to 1985 for wolframite, scheelite and argentiferous chalcopyrite. The mineralization occurs in quartz veins and in the two breccia pipes known as Santa Helena and Venise. The Santa Helena breccia occurs in outcrop (Fig. 2), whereas the Venise breccia was identified underground during mining. Rock fragments of variable size, cemented with mineralized
K-Ar dating of muscovite associated with molybdenite from the Venise breccia provided an age of 280 ± 5 Ma (Portugal-Ferreira and Noronha, unpublished data). K-Ar dating of feldspar from episyenitic rocks of the Gerês granite (Carris facies) yielded an age of 273 ± 11 Ma (Jaques et al., 2016), whereas Re-Os dating of molybdenites from Carris quartz veins placed the ore-forming stage at 279.5 ± 1.7 (Moura et al., 2014).

Analytical techniques
Optical microscopy Thin polished sections were prepared for optical microscopy in transmitted and reflected light. Textural relationships between chlorite minerals and opaque minerals (i.e. tungstates and sulfides) were examined using a Nikon E400 polarized transmitted-reflected light microscope using thin polished sections (30 µm).

X-ray diffraction
Chlorite fractions (<2 µm) were separated from bulk-rock mineralized pulps by centrifugation. The oriented specimens of <2 µm fractions prepared on glass slides were analyzed in an air-dried (AD) condition and saturated in ethylene-glycol (EG) vapour for 24 h. Structural characterization of the <2 µm oriented clay-fraction specimens was performed using a Philips X'Pert diffractometer with CuKα radiation and a monochromator, working at 50 kV and 30 mA. A scanning speed of 0.05°2θ/min for the range 2-65°2θ and a counting time of 5 s/step were used. The interlayer nature, the polytype and d 060 values of the chlorite minerals were examined using randomly oriented specimens. The XRD patterns of these specimens were acquired using a RINTUltima + diffractometer with a Cu X-ray tube, Ni-filter, and silicon strip X-ray detector (Rigaku D/teX Ultra). A measurement condition of 0.25°divergence slit, 10 mm mask confining the beam width, and 8 mm antiscatter slit was used. The XRD pattern was obtained with a scanning speed of 0.10°2θ/min at a counting time of 20 s/step over the range 2-75°2θ.

Oxygen isotopes
Fe-, Fe,Mn-and Fe,Mg-chlorite samples (previously analysed by XRD) were selected for oxygen isotope analyses carried out at the Laboratory for Stable Isotope Science, University of Western Ontario (London, Ontario, Canada). Oxygen isotopic analysis of quartz samples from the W-and Momineralization were analysed at the University of Salamanca (Spain). A conventional procedure was used for the extraction of oxygen from chlorite and quartz with BrF 5 and quantitative conversion to CO 2 prior to analysis by stable-gas mass spectrometry (Clayton and Mayeda, 1963). The δ 18 O values are reported in ‰ relative to Vienna-Standard Mean Ocean Water (V-SMOW) and normalized to the S263 NBS-30 reference (+5.17‰ vs. V-SMOW) and NBS-28 (+9.34‰ vs. V-SMOW). The reproducibility of δ 18 O measurements is ±0.2‰ (1σ). The following isotope fractionation equations were used here between the mineral and water system: the chlorite-H 2 O (Wenner and Taylor, 1971), the quartz-H 2 O (Clayton et al., 1972), and albite-H 2 O (Matsuhisa et al., 1979).

Macroscopic observations and optical microscopy
Five distinct assemblages of chlorite minerals were identified in the W, (Cu,Mo) ore deposit of Borralha ( Fig. 3): (1) quartz + albite + adularia + Fe,Mg-chlorite; (2) Fe,Mn-chlorite + scheelite; (3) Fe-chlorite + wolframite + scheelite; (4) Fe-chlorite + chalcopyrite + sphalerite ± pyrite + secondary muscovite (+ sericite); (5) Fe,Mg-chlorite + molybdenite + bismuthinite ± adularia + quartz. Mg,Fechlorite with chalcopyrite was observed by optical and electron microscopy. Chlorite minerals observed in transmitted light were recognized easily by their green pleochroism, varying from pale green for some massive aggregates to dark green with a basal cleavage and anomalous birefringence in cross-polarized light. Under the microscope, Fe,Mn-chlorite is characterized by strong green pleochroism with an atypical dark orange interference (Fig. 4a,b), which is in accordance with octahedral occupancy by Mn 2+ (Bailey, 1988). The elongation sign is opposite to the optical sign and the optical plane is almost parallel to (010). The Fe,Mn-chlorite aggregates are larger (800-1500 µm) than those of the Fe-chlorite. Fe-chlorite (Fig. 4c,d) shows flaky micrometric aggregates with greenish pleochroism and a much lower birefringence compared to Fe,Mn-chlorite. Small flaky aggregates of Mg,Fe-chlorite (Fig. 4e, f ) with chalcopyrite show a randomly oriented grey to pale green pleochroism and small sheet aggregates, <100 µm in size.

Structural data
Oriented specimens of Fe-, Fe,Mn-and Fe,Mgchlorite were examined both air-dried (AD) and following treatment with ethylene-glycol (EG) vapour (Supplementary material, see below). XRD patterns of the three samples studied show the (00l ) reflections distinctive of chlorite minerals, where the 002 peak at 7 Å is sharper than the 001 peak at 14 Å and the 001 and 003 peaks show approximately equal intensities. This is a typical structural feature for Fe-chlorite, where a large number of Fe atoms are distributed between the two octahedral sites. The different relative intensities of the 00l reflection planes may be used to measure the total metals (e.g. Ni, Co, Cr, Mn) incorporated between the silicate and hydroxide octahedral sheets (Brown and Brindley, 1980). Greater intensities of the 00l reflections in Fe,Mn-chlorite are due to the presence of Mn (measured by EMPA) in the octahedral sheet. Increasing the amounts of Fe and Mn in the chlorite structure causes a decrease in the 001, 003 and 005 reflections relative to the 002 and 003 reflections. The normal 001 peak at 14.2 Å occurs at 14.60 Å which is typical of clinochlore (daphnite), or at 14.3 Å for pennantite. This is explained by the presence of the metals in the hydroxide layer and, to a lesser extent, by the presence of a few tri-smectite layers interstratified with tri-chlorite. There is a marked effect of Fe symmetry on the intensity ratio of the 001 and 003 peaks and approximately equal intensities indicate a symmetrical Fe substitution. The 001 and 005 reflections are weak relative to the 003 reflection due to the presence of Fe in the hydroxide layer. After EG treatment the peak positions did not change, but the intensity of the 00l peaks decreased. This probably indicates that expandable layers such as those of smectite do not occur in these chlorites although changes in the oxidation state may complicate this issue.

Textural relationships: chlorite vs. mineralization
The areas of the sections examined by means of optical microscopy were selected for detailed subsequent analysis by electron microscopy for to better understand the textural relationships between the chlorite minerals and the oxides or sulfides. Several textural relationships were examined, in particular those involving Fe,Mn-chlorite and scheelite (Fig. 5a), Fe-chlorite and wolframite and scheelite (Fig. 5b), Fe-chlorite and wolframite ( Fig. 5c), Fe-chlorite vs. wolframite and Mg,Fechlorite vs. chalcopyrite (Fig. 5d), Fe-chlorite + secondary muscovite + chalcopyrite + wolframite (Fig. 5e), and Fe,Mg-chlorite along the cleavage planes of molybdenite (Fig. 5f ).

S265
The Fe 3+ calculated is present in minor amounts. It is assumed here that Fe 3+ replaces only Al(VI) in the chlorite structure. Small amounts of Zn 2+ , Ca 2+ , Na + and K + were identified in Fe,Mn-chlorite. The oxidized chlorites are distinguished by the Si contents of 5.6-6.2 atoms per formula unit (apfu), whereas for the non-oxidized chlorites the Si contents range from 5 to 7 apfu for a maximum of 8 apfu (Deer et al., 1976). The Fe/(Fe + Mg) ratio shows an average value of 0.92 (Supplementary material). Compositionally, the Fe,Mn-chlorite falls between daphnite and amesite (see Fig. 11) in the diagram of R 2+ vs. Si 4+ (Wiewióra and Weiss, 1990).

S266
state. The chlorite composition corresponds to daphnite. These compositions are similar to those of Fe-chlorite occurring with chalcopyrite and secondary muscovite (sericite) (Figs 3d and 5e).

S267
clear and it is considered to be a typical feature of this Fe,Mg-chlorite. The Fe/(Fe + Mg) average ratio is ∼0.76 (Supplementary material) and the chlorite composition corresponds to that of ripidolite.
X-ray maps of chlorite minerals vs. tungstate and sulfide minerals X-ray maps illustrate the Fe and Mn distribution and support the notion of chlorite crystallization and metal precipitation directly from the fluid, as discussed below.
Fe,Mn-chlorite The area illustrated in the inset of Fig. 6 shows Fe, Mn-chlorite associated with scheelite grains. The Fe,Mn and Mg distributions are distinct and reflect the incorporation of Fe followed by Mn. The X-ray maps show the W and Ca distributions in the scheelite (Fig. 7, inset) and the EPMA of the

Fe-chlorite
The area illustrated is a veinlet infill of Fe-chlorite aggregates associated with wolframite (Fig. 8). The X-ray map shows that the Fe distribution in the chlorite aggregates is clearly defined, Mg is limited to the edges of the chlorite aggregates and Mn is confined to the central part of the chlorite (Fig. 8,  inset). The EMPA analyses of the wolframite (  Noronha (1983). In contrast to the wolframite from Panasqueira which is enriched in the ferberite component (FeWO 4 ), the hübnerite (MnWO 4 ) component of wolframite is more significant at Borralha (Fig. 9) and as well as in other Iberian tungsten deposits (Neiva, 2008).

Mg,Fe-chlorite
The areas illustrated correspond to chlorite veins close to chalcopyrite. The back-scattered image (Fig. 10, inset) shows aggregates of secondary muscovite (sericite) which may be a greisen-type alteration, and Fe-chlorite associated with wolframite and chalcopyrite. Particular attention was given to the X-ray map of Mg distribution, where small chlorite aggregates associated with chalcopyrite are enriched in Mg (shown in red) at the expense of Fe.

Chlorite geothermometry
Chlorite shows a wide compositional range because it is sensitive to P-T and redox conditions, as well as the bulk-rock and fluid composition (Walshe, 1986;Cathelineau, 1988;Vidal et al., 2001Vidal et al., , 2005Vidal et al., , 2006Inoue et al., 2009;Bourdelle et al., 2013;Lanari et al., 2014). Here we used the geothermometry model of Bourdelle et al. (2013) and Bourdelle and Cathelineau (2015) which considers all the iron to be Fe 2+ . Those authors justified this approach by S269 the fact that, accounting for Fe 3+ in the structural formula, fixed numbers of oxygens reduce the number of cations, implying an increase in octahedral vacancies. That in turn creates an indirect link between the Fe 3+ content and the temperature. The chlorite temperature estimates are controlled in the first place by Tschermak substitution, where Fe 3+ has a subordinate role that may be disregarded as a first approximation. The mineral compositions were plotted in the Si 4+ vs. R 2+ diagram (Wiewióra and Weiss, 1990), and four distinct geothermometric fields were (Fig. 11): (1) Fe,Mn-chlorite falls into a field with T ranging from 400 to 500°C; (2) Fe-chlorite plots in a field where T varies from 250 to 350°C; (3) Fe,Mg-chlorite corresponds to a field with T of ∼200-250°C; and (4) Mg,Fe-chlorite falls in a field with T of ∼150°C.

Stable isotope data
Oxygen isotopic measurements of the Fe-, Fe,Mn-, Fe,Mg-chlorites, quartz from W-and Momineralization (Santa Helena and Venise breccia pipes), adularia from the quartz + Fe,Mg-chlorite + molybdenite assemblage and the oxygen isotopic fractionations between chlorite, quartz, adularia and H 2 O related to W-and sulfide mineralization are shown in Table 3. These data are compared with the isotopic data for quartz from the Gerês lateorogenic granitic massif (Jaques et al., 2016) and from the Panasqueira deposit (Polya et al., 2000). The δ 18 O isotope (V-SMOW) of the chlorites yielded values of +1.5 (1σ) for (Fe,Mg-chlorite), +3.8 (1σ) for (Fe-chlorite) and +6.91 (1σ) for (Fe, Mn-chlorite). The isotopic fractionation for chlorite and H 2 O calculated with the bond-type approach depends on the composition which should be known before interpreting the isotopic data (Cole, 1985(Cole, , 1994Savin and Lee, 1988). Oxygen isotopic exchange experiments in the chlorite-H 2 O system (Wenner and Taylor, 1971) were calculated assuming a temperature of 200-500°C, inferred from fluid-inclusion studies, and deduced on the basis of chlorite geothermometry according to the procedure described by Bourdelle et al. (2013). The δ 18 O F (fluid, F ) oxygen isotopic fractionation for the chlorite-H 2 O pair (Wenner and Taylor, 1971), assuming a temperature of 300°C, yield values (Table 3)

Discussion
Previous studies of the Borralha W, (Cu,Mo) ore deposit have demonstrated a clear relationship between mineralization and chlorite formation, with ripidolite and daphnite intimately associated with the evolving stage of sulfide deposition (Noronha, 1983(Noronha, , 1984aMateus, 1985). The progression from tungstate to sulfide deposition coupled with chlorite crystallization identified in this W, (Cu,Mo) ore system, records an unequivocal connection with the chemical evolution of mineralizing fluids along with the declining P-T conditions under which the ore-forming processes took place. The following mineral assemblages are well documented: Fe,Mn-and Fe-chlorite crystallize during the W-ore stage, where a Cu contribution was indicated by the presence of chalcopyrite, whereas Fe,Mg-and Mg,Fe-chlorite throughout the sulfide ore stage is characterized by molybdenite, bismuthinite and pyrite assemblages.
Wolframite crystallized from fluids with slightly acid to neutral pH, where WO 4 2− is easily complexed with Fe 2+ , Mn 2+ and Ca 2+ . Note that stability of the Fe 2+ end-member (ferberite) is much more sensitive to the large variation of f o 2 than is the Mn 2+ end-member wolframite (hübnerite); the composition of ferberite might be affected strongly by the oxidation state of the environment (Hsu, 1976). Furthermore, tungsten may be transported in various aqueous solutions under different conditions where the dominant species above 400°C is tungstic acid, whereas HWO 4 − and WO 4 2− , and the alkali-tungstate ion pairs are the most stable species at lower temperatures (Higgins, 1985;Wood and Vlassopoulos, 1989).
Experimental studies on the stability of H 4 [Si (W 3 O 10 ) 4 ].2H 2 O at high P-T conditions and variable pH in the presence of Fe 2+ , Mn 2+ and Ca 2+ have confirmed the separation of silicotungstic acid under the effect of solution neutralization (Gundlach, 1967) according to the reaction: Two distinct tungstate vs. silicate assemblages are demonstrated by this work: one consists of scheelite and Fe,Mn-chlorite, whereas the other involves Fe-rich wolframite crystallized in more acidic conditions than did Mn-rich wolframite and scheelite, and the presence of scheelite associated with Mn-rich wolframite, indicate that the minerals are not contemporary. Mn-rich wolframite is associated with a late-evolving W-mineralization stage possibly triggered by rising alkalinity of the fluid as a result of an increase in S activity and the Ca 2+ availability of the system. This can be correlated with a late alkaline metasomatic reaction which is very common in geochemical settings surrounding the development of W-ore deposits (Wood and Samson, 2000).
The result of [SiO 4 ] 4− and [WO 4 ] 2− separation is direct crystallization of scheelite (I) from the magmatic-hydrothermal fluid and Fe,Mn-chlorite followed by Mn-rich wolframite and scheelite (II) together with Fe-chlorite from the hydrothermalmeteoric fluid. The observed Mn-rich wolframite + scheelite (II) + chalcopyrite assemblage confirms a mixed Mn-rich wolframite-sulfide evolving stage, where scheelite replaced Mn-rich wolframite (seen clearly during optical microscopy) and sulfide minerals precipitated according to the reaction: The common association of tungstate with some sulfide minerals indicates that sulfur is commonly present and plays a role in the stability of wolframite whereas the stability fields of ferberite (FeWO 4 ) and hübnerite (MnWO 4 ) do not depend

S272
on the presence of oxygen and sulfur (Hsu, 1975). Crystallization of Mg,Fe-chlorite associated with chalcopyrite ( Fig. 5d) in the Santa Helena breccia, in a mineral assemblage comprising Fe-chlorite, secondary fine-grained muscovite (sericite), Mn-bearing wolframite and scheelite, is particularly interesting. The suite of mineralogical reactions should be expected to begin with the silicates followed by the development of the tungstate minerals. The presence of the chalcopyrite + Mg, Fe-chlorite assemblage suggests that Fe from wolframite was consumed during the crystallization of chalcopyrite, whereas the Mg,Fe-chlorite is a byproduct after the formation of secondary muscovite from biotite according to the reaction: In this case, the Mg,Fe-chlorite is not a suitable guide to sulfide mineralization. Crystallization of Fe,Mg-chlorite associated with molybdenite in the Venise breccia pipe is clearly related to different geochemical conditions imposed by some chemical rejuvenation of hydrothermal fluids, expressed by an increase in f s 2 and a pH variation from alkaline to neutral pH. Molybdenite, in contrast to ferberite or hübnerite, crystallized in a more reduced environment according to the reaction: This reaction generates more protons in the fluid and produces conditions favourable to the destabilization of primary biotite, and enriching the fluid in Mg 2+ and Fe 2+ . Releasing this fluid through the cleavage planes of molybdenite, contributed to the later crystallization of Fe,Mg-chlorite in the system (Fig. 5f ). The distribution of Fe,Mg-chlorite is widespread and often occurs together with hydrothermal quartz, adularia or secondary albite (Fig. 3a) which means that the fluid was widely distributed throughout in the Venise breccia pipe. The Fe-and Fe,Mn-chlorite is described here associated with W-mineralization for the first time in the magmatic-hydrothermal Borralha system.

S273
The structure and chemistry of the chlorites reveal important details of the conditions of crystallization. The XRD data confirmed that there are no expandable phases in the Fe-, Fe,Mn-or Fe,Mgchlorite structure, thus making them suitable for geothermometry. The presence of metals (i.e. Mn) in the chlorite structure affected the 002/001 intensity ratios. Large amounts of Fe [Fe/(Fe + Mg + Mn) = 0.4-0.95] and Mn [Mn/(Fe* + Mg) > 0.05] in chlorites are diagnostic of the chemical environments during ore formation (Inoue et al., 2010).
The δ 18 O F (+6.86 and +8.16 1σ) values calculated (Wenner and Taylor, 1971), responsible for Fe,Mn-chlorite crystallization in the temperature range of 300-400°C, show that the Fe,Mnchlorite and scheelite (stage I) were likely to have been deposited from magmatic-hydrothermal fluids. The oxygen isotopic results from Fe,Mnchlorite are very close to those reported by Borsevski et al. (1979) for chlorite from Sn-ore deposits in the Transbaikal (Serlovaia Gora and Hapceranga) Russia, where measured values of +6.1 and + 4.4‰ are due to equilibration with a hydrothermal fluid with an isotopic signature of ∼+6.05‰.
The temperature estimation of tungsten and sulfide deposition (stage II) based on fluidinclusion studies of quartz from the Borralha deposit ranges from 270 to 350°C for tungsten associated with aqueous-carbonic fluids with an average salinity of 10 wt.% eq. NaCl; and from 250 to 300°C for chalcopyrite associated with aqueous fluids (Noronha, 1984b). Oxygen isotopic results from Fe-and Fe,Mg-chlorite (stage II and III) confirm the same isotopic association between chlorite and the aqueous fluid.

S275
isotopic data obtained on quartz from W (Santa Helena breccia) mineralization yielded a value of +12.6 (1σ), in equilibrium with a calculated δ 18 O F of +5.67 (1σ). Such a δ 18 O F value, quite close to a magmatic-hydrothermal signature, would suggest that quartz and Fe-chlorite formation could be related to a mineral-water system with a mixed contribution from meteoric and magmatichydrothermal fluids during the late tungstensulfide (II) stage.  (Matsuhisa et al., 1979) in the case of adularia yielded a value of +1.75 (1σ) and for the quartz-H 2 O pair ∼+4.45 (1σ). The δ 18 O values for waters that deposited the latest assemblage of quartz-Fe, Mg-chlorite-adularia range from +0.50 to +4.45 (1σ). Oxygen isotope thermometry of early quartz and adularia indicate equilibrium temperatures of ∼220°C (Matsuhisa et al., 1979).
The isotopic signatures can be compared with the δ 18 O data obtained from quartz (episyenitic rocks) from the post-D 3 Hercynian granite of Gerês, which revealed values of ∼+10.97‰ ± 1.9‰, in equilibrium with an aqueous fluid where the calculated δ 18 O F is ∼+5.16‰ (Jaques et al., 2016). δ 18 O values of 10 and 11‰ have been determined for other granites in the Portuguese Hercynian rocks (Reavy et al., 1991).
Magmatic quartz from the Panasqueira (Portugal) W-ore, has a high δ 18 O value of ∼+13‰ and the δ 18 O value of primary magmatic waters is +11 ± 1‰ (Kelly and Rye, 1979). The fluids responsible for ore formation exhibited relatively constant fluid inclusion homogenization temperatures (254-260°C) with salinities of 7.4-8.7 wt.% NaCl equivalent and a calculated fluid δ 18 O with values from 3.8 to 4.4‰ (Polya et al., 2000). This means, that the ore-forming stage fluid can be explained by isotopic exchange of meteoric waters equilibrated with granite at ∼350°C (Polya, 1989).

Conclusions
Chlorite minerals may indicate enriched, and economically more favourable, hydrothermal systems. The samples studied show a genetic link between Fe-, Fe,Mn-and Fe,Mg-chlorite and W-and Mo-mineralization supported by textural relationships, geothermometry and isotopic signatures. The composition of chlorite was influenced directly by thermal gradients developed during different stages of the mineralizing process in the Borralha magmatic-hydrothermal system. The identification of a possible hydrothermal alteration halo was not possible, but the chlorite phases associated with W-and Cu,Mo-mineralization are an excellent indicator which help to identify the direction in which the mineralized area may be found and a means of estimating the crystallization temperature of chlorite and tungstate or sulfide.