Geothermobarometry of very low-grade metamorphic pelites of the Vendian–Early Cambrian Puncoviscana Formation (NW Argentina)

: The Vendian–Early Cambrian Puncoviscana Formation is a pelite-greywacke turbidite sequence affected by polyphase deformation cropping out extensively in the Cordillera Oriental of northwestern Argentina. Previous X-ray diffraction and analytical high-resolution TEM studies on southern locations found anchizonal grade and proposed medium–high pressure metamorphism followed by higher thermal conditions. We have determined the chemical composition of mica and chlorite with EDX on small uncontaminated areas selected using SEM in backscattered electron mode. The application of mica-chlorite geothermobarometry to mineral grains defining two different foliations has produced consistent pressure/temperature results based on the intersection of six reactions: 9 kbar/250 (cid:2) C peak conditions were followed by an isothermal decompression to 3 kbar and a successive increase of temperature to nearly 350 (cid:2) C at 1.5 kbar. The shales from the overlying Meso´n and Santa Victoria Cambro–Ordovician groups do not contain chlorite, thus chlorite-mica geothermobarometry could not be applied. According to the range of KI CIS obtained (0.41–0.68 D (cid:2) 2 y ) and their mineralogical assemblage (illite (cid:3) kaolinite (cid:3) corrensite þ Qtz þ Pl (cid:3) Kfs), these rocks have not surpassed diagenetic conditions, which implies a minimum temperature difference with the Puncoviscana Formation of 100 (cid:2) C, consistent with the sharp angular unconformity between the two units.


Introduction
In clastic sequences affected by very low-to low-grade metamorphism, it is difficult to assess peak pressure-temperature (P-T) conditions and, even more so, the reconstruction of P-T-time (t) paths.Rough estimates of metamorphic temperatures can be achieved with the Ku ¨bler Index (KI), but more precise determinations are problematical essentially for two reasons: a lack of reliable index minerals, but also because thermodynamic equilibrium is seldom reached under these conditions.Several thermobarometric indicators based on white mica and chlorite compositions have been proposed in recent decades because it has been experimentally proven they depend on the P-T conditions predominating during their formation.Hence, metamorphic pressures are frequently estimated from the Si contents of phengites (Massone & Schreyer, 1987;Massone & Szpurka, 1997), and temperatures from the IV Al content of chlorites in sub-greenschist (,400 C) conditions (Cathelineau, 1988;Jowett, 1991;de Caritat et al., 1993).Such thermobarometers have been heavily used mainly due to their simplicity, but also because both minerals are widespread in low-grade metapelites.However, P and T estimates obtained with these methods frequently show a wide scatter (de Caritat et al., 1993) reflecting the lack of thermodynamic equilibrium.Moreover, T estimates based on the IV Al content of chlorites have received considerable criticism (Hillier & Velde, 1991;de Caritat et al., 1993;Lo ´pez Munguira et al., 2002).The thermodynamic-based chlorite geothermometer proposed by Walshe (1986) was little used perhaps because it requires more complex calculations.More recently, Vidal et al. (2001) deduced a solid-solution model for aluminous (Si , 3 apfu) chlorites found in metapelites over a wide range of P-T conditions.The chlorite solution model of these authors consists of a four-thermodynamic-component (Mg-amesite, clinochlore, daphnite, and Mg-sudoite) that accounts for the Tschermak, Fe-Mg, and di/trioctahedral substitutions observed in nature.The strength of this model stems from the fact that it was calibrated with independent sets of published experiments conducted in the MASH and FMASH systems and in a large set of natural data involving chlorite þ quartz AE (carpholite or chloritoid) assemblages.Later, Inoue et al. (2009) developed a semi-empirical geothermometer based on four endmembers (corundophillite, chamosite, sudoite and Al-free chlorite) that could also be applied to chlorites showing Si contents .3 apfu; however it requires the assessment of Fe þ3 content in chlorite, which cannot be accomplished by SEM or EMPA but, e.g., by Mo ¨ssbauer spectroscopy.According to Inoue et al. (2009) the determination of the Fe þ3 /Fe þ2 is essential in estimating the temperature of formation of low-T chlorites.Nevertheless, for chlorites formed at intermediate temperatures (.250 C), Inoue et al. (2009) and Lecle `re et al. (2012) found a good correlation between temperatures estimated using Inoue's and Vidal et al. (2001)'s geothermometers.Furthermore, Parra et al. (2002) derived a solid solution model for K-white micas based (like the chlorite model) on phase-equilibrium experiments conducted in different systems and also on a large dataset for natural assemblages covering a wide range of P and T conditions.This model comprises seven thermodynamic components: muscovite, Fe 2þ -Al-celadonite, Mg-Al-celadonite, annite, phlogopite, pyrophyllite, and paragonite, accounting for the Tschermak, Fe-Mg, di/ trioctahedral, illitic, and paragonitic substitutions observed in natural micas.Recently, Nieto et al. (2010) presented data indicating that illitic substitution in dioctahedral micas does not correspond to true vacancies but instead with H 3 O þ partially replacing K in interlayer sites.
The application of thermobarometers in anchizonal siliciclastic rocks has to cope with the lack of thermodynamic equilibrium shown by strong compositional heterogeneities in dioctahedral mica and chlorite even at distances above a few millimetres, probably arising from the co-existence of several mineral generations that did not re-equilibrate during the successive events.In order to overcome the problem of short-scale compositional heterogeneities, Vidal & Parra (2000) suggested basing P-T estimates on the composition of phengite-chlorite pairs corresponding to different microsites at the micrometre scale, as their investigation demonstrated that short-range equilibrium between these two minerals might be approached and preserved.Thus, if each clay mineral generation occurs in different microstructural sites, local chlorite-phengite assemblages can be identified and analysed using scanning electron microscope (SEM) or electron microprobe (EMPA) and their P-T formation conditions be deduced through the thermobarometer proposed by Vidal & Parra (2000).This approach was successfully applied to assemblages forming at temperatures well above 400 C (Vidal & Parra, 2000), but their use for very lowgrade metapelites is not yet well established.
The aim of this work is to test the suitability of the thermobarometer based on chlorite-phengite local equilibria in anchizonal metapelites in order to constrain their P-T path and to compare the results with evidence from the phengite geobarometer, Ku ¨bler index values, and microstructural analysis.We have studied the Vendian-Early Cambrian Puncoviscana Formation (Turner, 1960), which is mainly composed of a pelite-greywacke turbidite sequence affected by polyphase deformation (Mon & Hongn, 1991, 1996).Based on 'illite crystallinity' values Toselli & Toselli (1982) and Toselli (1990) roughly estimated low-anchizone to epizone grade for Puncoviscana Formation; whereas Willner et al. (1990) postulated that the metamorphic grade and the deformation of this unit increase from north to south.More recent studies indicate that this unit was subjected to anchizone-grade metamorphism (Do Campo & Nieto, 2003;Do Campo et al., 2005).The Puncoviscana Formation was defined in the northern portion of the Sierra de Santa Victoria; however, most studies carried out in recent decades deal with localities farther south than 24 (Toselli & Toselli, 1982;Toselli, 1990;Do Campo, 1999;Do Campo et al., 2003;Do Campo & Ribeiro Guevara, 2005, and others), and no detailed study of its metamorphic evolution has been carried out on its northern portion.Given that understanding its metamorphic evolution is crucial for tracing the Vendian-Early Cambrian evolution of the western margin of Gondwana, we decided to perform a detailed mineralogical and petrological study focusing on the northern outcrops of the Puncoviscana Formation.
Furthermore, in order to compare the degree of postdepositional transformation attained by the Puncoviscana Formation and the overlying Cambro-Ordovician units cropping out in the same region, we studied samples of fine-grained levels from the Lizoite and Campanario formations (Meso ´n Group, Cambrian) and the Santa Victoria Group (Ordovician).
Standard petrographic analyses were carried out on all samples to determine lithology, general composition, and primary and metamorphic foliations.In addition, XRD analyses were performed for whole rocks and clay fractions, including measuring KI.Afterwards, representative samples were chosen for detailed study at the SEM scale and EDX microanalyses.Thermobarometric calculations were performed with the TWQ (1.0) software package (Berman, 1991) for well-constrained chlorite-phengite pairs.
A better knowledge of the metamorphic evolution of the Puncoviscana Formation and the Meso ´n and Santa Victoria Groups outlined from mineralogical data would also provide a constraint for further geodynamic models of the area.

Regional geology
The Sierra de Santa Victoria belongs to the Cordillera Oriental morphostructural region, which is bordered by the Puna region to the west and the Sierras Subandinas to the east.The Puncoviscana Formation (Turner, 1960) crops out extensively in the Cordillera Oriental and to a lesser extent on the eastern border of Puna (Salta and Jujuy provinces), constituting their basements (Fig. 1).It is mainly composed of a pelite-greywacke turbidite sequence, but in a few localities outside the study area it by Adams et al., 2011) in the Ampujaco River area were interpreted as contemporary volcanic sources at a late Early Cambrian Puncoviscana depocentre.
Ku ¨bler index values for several localities south of 24 S Lat, in the Lules-Puncoviscana and Choromoro belt (Fig. 1), are in the range of medium anchizone up to the anchizone-epizone boundary (0.36-0.25 D2y according to Do Campo, 1999;Do Campo & Nieto, 2003).Prograde biotite was only detected in several metapelites from the Sierra del Cobre area; however, its formation was probably the result of the strong metamorphic overprint produced by Ordovician magmatic bodies (Hongn et al., 2001a).
The Lower-Middle Cambrian succession of the Meso ´n Group comprises the Lizoite, Campanario, and Chahualmayoc formations (Turner, 1960).The basal section of the Lizoite Formation consists of conglomerates with overlying pink quartzites and sandstones with crossbedding stratification; the Campanario Formation is a succession of predominantly fine-grained sandstones with interbedded shales; and the Chahualmayoc Formation comprises coarse-grained quartzites (Rubiolo et al., 2001;Astini, 2003).
The Meso ´n Group is, in turn, unconformably covered by the Upper Cambrian-Lower Ordovician Santa Victoria Group (Turner, 1960).This group consists of the Santa Rosita and the Acoite formations, comprising a thick shale succession with interbedded grey sandstones (Rubiolo et al., 2001;Astini, 2003;Buatois & Mangano, 2003).

Analytical methods
Fifteen metapelites and meta-arenites samples from Puncoviscana Formation were collected along a transect approximately perpendicular to the trend of fold axis along the San Isidro River valley (Table 1), 6 km southeast of the village of San Isidro (22 44'1.7''S Lat, 65 15'00'' W Long, Salta province) (Fig. 1), where these rocks are particularly well exposed.Aditionally, six samples from the Campanario and Lizoite formations (Meso ´n Group) and the Santa Victoria Group were collected from a small outcrop in the village of San Isidro, in the vicinity of Miyuyoc and somewhat farther south in the Quebrada de Humahuaca (22 52' S Lat,65 17' W Long and 23 22' S Lat,65 20' W Long,respectively).
Samples for clay-mineral X-ray diffraction (XRD) analysis were prepared following the recommendations of Moore & Reynolds (1997).The ,2 mm fraction was separated for 21 pelitic samples from the Puncoviscana Formation, Meso ´n Group, and Santa Victoria Group.Clay-mineral composition was established by comparing air-dried (AD) and ethylene-glycol-solvated (EG) oriented mounts.The XRD analyses were performed with Philips PW1050 (INGEIS) and X-Pert Pro (Departamento de Fı ´sico Quı ´mica-UNC) diffractometers, employing CuKa radiation, from 3 to 40 2y in scanning mode.Semiquantification of clay-mineral phases was made using mineral intensity factors (MIF) and the recommendations of Moore & Reynolds (1997).The Ku ¨bler index (KI) was measured in oriented mounts and CIS values (Crystallinity Index Standard, Warr & Rice, 1994) were established from the regression equation for the Philips PW1050 diffractometer: y ¼ 1.2175x þ 0.0833 (R 2 ¼ 0.975).The accepted boundaries between the diagenetic zone and anchizone is at present KI ¼ 0.42 D 2y, and for the anchizone to epizone limit, it is KI ¼ 0.25 D 2y.The white mica b parameter was measured in rock slices orientated perpendicular to the main foliation, and the quartz (211) reflection positioned at 1.541 A ˚was used as internal standard (Guidotti & Sassi, 1986).The petrographic and microstructural characterization of metapelites and meta-arenites was carried out by optical microscopy, then five metapelites and meta-silts were analysed by means of scanning electron microscopy (Zeiss DSM950 SEM and Variable Pressure SEM) at the Scientific Instrument Centre, University of Granada.Textural analyses were performed on carbon-coated polished samples employing backscattered electron images (BSE).The chemical composition of phyllosilicates was established through EDX analysis on areas recognized as homogeneous in the BSE images, employing 20 kV, a beam current of 1-2 nA and a counting time of 100 s.The following compounds were used as calibration standards: albite (Na), orthoclase (K), periclase (Mg), wollastonite (Si), and synthetic oxides: Al 2 O 3 (Al), Fe 2 O 3 (Fe) and MnTiO 3 (Ti and Mn).Data were reduced using a conventional ZAF routine.The EMPA was not employed in this study because most chlorite and mica laths are thinner than 5 mm and phyllosilicate intergrowths are normal under the optical microscope.The BSE images obtained under the SEM allow easy selection of very small, contamination-free areas for analysis.Abad et al. (2003a) demonstrated that SEM-EDX analyses obtained under the same conditions as EMPA produced equivalent results.These authors found very small differences between the two techniques, well below their precision level, for all the major elements involved in the chemical composition of the micas.Hence, EDX analyses can be employed with confidence for thermobarometric purposes, provided that all the methodological requirements usually accepted for EMPA quantitative analyses are also followed in the SEM, in particular a careful calibration with real standards and preparation of polished samples.
Atomic concentration ratios were converted into formulae according to stoichiometry (number of oxygens in the theoretical formulae of minerals).The structural formulae of dioctahedral micas were calculated considering 22 negative charges and a Fe 3þ /(Fe 3þ þ Fe 2þ ) ratio equal to 0.70 as all samples contain hematite and ilmenite as accessory iron-bearing phases (cf.Guidotti et al., 1994b).Mica formulae with Mg 2þ contents higher than (Si -3) were interpreted as contaminated with chlorite.Chlorite formulae were calculated considering 28 negative charges, whereas Fe 3þ /(Fe 3þ þ Fe 2þ ) ratios were initially considered as equal to 0.10 (cf.Vidal et al., 2005).Small amounts of K þ , Na þ , and Ca þþ identified in some chlorites from the Puncoviscana Formation were interpreted as indicative of mica contamination.Structural formulae from corrensite from the Santa Victoria Group were calculated considering 50 negative charges (28 þ 22) and Fe ¼ Fe 2þ .
Phengite-chlorite thermobarometric estimations (Vidal et al., 2001) were carried out for five pelites from the Puncoviscana Formation applying the TWQ 1.0 software modified by Vidal et al. (2001) and its associated database JUN92.Atom site distribution of chlorite and phengite to input in the TWQ software were calculated following the solid-solution models from Vidal & Parra (2000); Vidal et al. (2001); Vidal et al., (2005Vidal et al., ( , 2006) ) and Parra et al. (2002), which included a re-calculation of Fe 3þ .The association of each particular grain to a certain microstructure (S 1 or S 2 ) was established at the time of performing the analysis, although grains that do not depict an orientation with respect to primary or secondary foliations were also analysed in order to avoid biases.Only analyses of chlorite-phengite pairs corresponding to the same microstructure were employed to perform thermobarometric calculations.

Petrography and microstructural analysis
The Puncoviscana Formation mainly comprises greenish phyllites alternating with decimetric levels of meta-silts and -arenites.Metasilt-arenites are characterized by variably developed rough-spaced slaty S 1 cleavage (Passchier & Trouw, 2005) subparallel to S 0 (Fig. 2a).Phyllosilicaterich cleavage domains (P domains) contain blastic phengite and chlorite flakes, whereas microlithons (quartz-rich, Q domains) are mainly composed of rotated quartz (with subgrains), feldspar, lithic fragments, and interleaved Chl/ Ms grains, most with wavy extinction (Fig. 2b).Pressure shadows around quartz grains, mainly composed of quartz and chlorite, and deformed grains of detrital mica are frequently observed.The S 0 primary foliation, sub-parallel to S 1, is preserved and outlined by alternating centimetric bands of slightly different grain size.In some coarsegrained bands, incipient normal grading could be recognized.
Metapelites are mainly composed of detrital and recrystallized quartz and albite plus neoformed white mica and chlorite.They record an S 1 zonal to continuous spaced metamorphic foliation defined by granulometric-compositional layering (Fig. 2c, d, and 3a) and by the preferred orientation of chlorite and mica flakes (Ms 1 -Chl 1 grains, Fig. 3b), but also by other recrystallized minerals such as quartz, scarce Kfs, and opaque minerals.As accessory minerals, zircon, apatite, ilmenite, hematite, iron-titanium oxides, and monazite are present.An S 2 crenulation cleavage, oblique to S 0 -S 1 , was also identified in several metapelites (Fig. 2c, d and f).Quartz-rich bands (microlithons) related to S 1 appear frequently as lenses or slightly folded within phyllosilicate-rich bands.Inside the microlithons, some large grains of chlorite and muscovite, more than 100 mm long (Fig. 3b, c), as well as some interleaved phyllosilicate grains (Chl/Mica, Fig. 3d) oriented parallel to S 1 are interpreted as detrital textural sites whose composition could have been altered by the metamorphic process (Giorgetti et al., 1997).Muscovite frequently occurs as elongated flakes from less than 10-60 mm in length, although scarce larger grains are also present.Chlorite has a more variable morphology, with common flakes but also rounded (Fig. 3e) and irregular grains of a size similar to mica grains.Features indicative of reorientation and deformation, such as kinking and microfolding (Fig. 4a) in mica and chlorite crystals, as well as quartz pressure shadows (Figs.3f and 4b), were observed at the optical microscope and SEM scales.
The S 2 crenulation cleavage is mainly developed in the primary phyllosilicate-rich bands, and is defined by an associated blastesis of Ms 2 -Chl 2 oriented parallel to the flanks of the folds (Fig. 4c, d).
Additionally, in the metapelites PU08-3 and PU08-14, bent grains were identified corresponding, respectively, to a Na-rich trioctahedral mica with deficiency in the interlayer charge that, following the IMA nomenclature of micas (Rieder et al., 1998), should be named as wonesite, and interleaved chlorite/Na-mica (Fig 4a ; PU 08-3 8/4 and PU08-14A 6/6, Table 2).Likewise, in the metapelite PU08-6, an inhomogeneous grain was identified that (according to EDX) probably corresponds to interleaved chlorite and wonesite (Fig. 4e; PU08-6 7/4 and 7/5 Table 2).The two zones of the grain differ mainly in their Mg and Fe contents, explaining the difference in contrast observed in the BSE image.They probably represent detrital ferromagnesian grains transformed during metamorphism.
Shales of both the Meso ´n and the Santa Victoria groups record an occasional S 1 secondary foliation, oblique to the millimetric sedimentary layering S 0 , with alternating light quartz-rich bands and dark bands with higher phyllosilicate contents (Figs. 5 and 4f).They are mainly composed of fine-grained white mica, chlorite, quartz, and opaque minerals.

X-ray diffraction -Ku ¨bler index
The phyllosilicates identified in all Puncoviscana metapelites collected along the San Isidro River were chlorite and illite-muscovite; no change in the basal reflection was observed in EG traces (Fig. 6a).Furthermore, KI (CIS) values were in the range of 0.31-0.23D2y (Table 1).The medium anchizonal to epizonal grade indicated by   1999;Do Campo & Nieto, 2003;Do Campo et al., 2005).
In the Campanario Formation, illite þ kaolinite þ Qtz þ Pl AE Kfs were identified through XRD (Fig. 6b), whereas in the only level analysed for the Lizoite Formation illite þ Qtz þ Kfs þ Pl were recognized.The KI (CIS) values determined for the Meso ´n Group vary between 0.41 and 0.51 (n ¼ 4), showing conditions between deep diagenesis and low anchizone (Table 1).Only two samples from the Santa Victoria Group were analysed by XRD: the sample from the Quebrada de Humahuaca comprises    1, Fig. 6c).The last sample was also studied by SEM, where a secondary foliation (S 1 ) oblique to the sedimentary layering (S 0 ) was identified.
White mica b parameter values for the Puncoviscana Formation ranged between 9.035 and 9.050 (n ¼ 7, mean ¼ 9.042), whereas two samples from the Campanario Formation and the Santa Victoria Group gave values of 9.025 and 9.012, respectively (Table 1).

Composition of phyllosilicates
More than 20 phengite analyses were obtained for each sample analysed within the Puncoviscana Formation (Table 3).Uncontaminated analyses of chlorite (Table 2) were more difficult to obtain as many grains seem to contain interleaved lamellae of mica not distinguishable under the SEM.

Dioctahedral micas
Table 3 shows representative analyses of dioctahedral micas.In the Na þ K vs. Si diagram (Fig. 7a), the more prominent substitutions correspond to the Tschermak vector, whereas illitic substitution is almost absent in most of the grains, apart from some micas from samples PU 08-14.Generally, analyses show a large variation in the Fe þ Mg contents of dioctahedral micas within each sample.In the (Fe þMg) vs. Si diagram (Fig. 7b), most of the analyses plot above the line for ideal Tschermak substitution ((Mg, Fe þ2 ) VI , Si IV ¼ Al VI , Al IV ), indicating that, besides Tschermak substitution, ferrimuscovite/ferriceladonite (Fe þ3 substituting for Al) components are also present.Guidotti et al. (1994b) demonstrated that, even in medium redox parageneses, containing ilmenite þ magnetite, the Fe þ3 /Fe tot ratio is over 0.60.In most of the samples, ilmenite was identified by SEM, and hematite by XRD, which also indicate oxidizing conditions.Taking this into account, the structural formulae in Table 3 were calculated assuming Fe þ3 /Fe tot ¼ 0.70.
In most of the micas, Na/(Na þ K) is lower than 0.05, although in a few cases this ratio reaches values in the range of 0.05-0.09,indicating minor paragonitic substitution (Guidotti, 1984;Guidotti et al., 1994a).
Therefore, with the exception of sample PU08-14, dioctahedral micas evidence a minimal illitic component, with Tschermak and ferrimuscovite/ferriceladonite substitutions being more relevant, what is charactheristic of true metamorphic phengites.
The frequency distribution of the Si contents of phengites (Fig. 8) is unimodal and approximately normal Gaussian, although it depicts positive asymmetry.As illitic substitution is minimal in the micas (Na þ K .0.87 apfu), all the analyses in Table 3 could be included in the graph because an Si excess of over 3 is basically related with Tschermak substitution.It is worth noting that, performing a statistical analysis of variance (ANOVA test), the null hypothesis that the average Si contents of the five samples are equal could be rejected with a 0.05 level of confidence.On the other hand, if the ANOVA test is done excluding micas from sample PU08-14, the Si contents of micas from the other four samples could be considered part of the same population.The Si contents of phengite from the Puncoviscana Formation at the San Isidro River site ranges from 3.05 to 3.48; these results are very similar to the ones obtained previously for southern localities from the Lules-Puncoviscana belt (3.06-3.50)by Do Campo & Nieto (2003).
Nonetheless, micas associated with each metamorphic foliation (S 1 and S 2 ) depict dissimilar compositional ranges.Namely, mica grains oriented parallel to sub-parallel to S 1 give a wide range of Si contents from 3.05 to 3.48 (Table 3).On the other hand, blasts of micas associated with the S 2 foliation show a narrower range of 3.19-3.32(Table 3).However, the number of analyses is notably lower in the second case for two reasons: S 2 crenulation cleavage was only clearly developed in metapelites PU08-3 and PU08-6, and in both samples micas related with the second foliation only represent a small proportion of the total.
Micas from the Santa Victoria Group (Table 4a) are characterized by a higher illitic content and a lower Tschermak substitution than those from the Puncoviscana Formation (Fig. 7c, d).The micas with interlayer charge .0.75 show Si contents between 3.02 and 3.32, and Fe þ Mg average contents of 0.28 (n ¼ 26).

Chlorite
Table 2 shows representative chlorite analyses.Analyses with evidence of contamination, exhibiting Na CaO contents of over 0.6 wt% were not included in the table nor plotted.In the Fe þ Mg vs Si diagram (Hillier & Velde, 1991), chlorites from samples PU08-3, PU08-4, PU08-5, and PU08-6 plot close to the line corresponding   to full octahedral occupancy (Fig. 9a).In addition, these chlorites present VI Al , IV Al.In contrast, chlorites from sample PU08-14 are more siliceous ( VI Al .IV Al) and depict slightly lower occupancies as well as less total (Fe tot þ Mg).In the five samples, Fe tot /(Fe tot þ Mg) varies by less than 0.1, although sample PU08-14 shows a slightly wider range of variation than the others.Moreover, chlorites from sample PU08-14 are distinct from the rest in their lower Fe tot /(Fe tot þ Mg) ratio, as shown by a Al IV -1 versus Fe tot /(Fe tot þ Mg) plot (Fig. 9b).The same plot indicates no correlation between tetrahedral substitution and (Fe tot þ Mg) contents; the variation in Fe tot /(Fe tot þ Mg) is clearly less than the change in Al IV .Chlorite grains aligned parallel to sub-parallel to S 1 depict a wide range of Si contents between 2.63 and 2.90, whereas chlorite associated with the S 2 crenulation cleavage shows a narrower range from 2.56 to 2.69.However, the more siliceous chlorites (Si .2.69) correspond exclusively to sample PU08-14, as mentioned above.
For the metapelite corresponding to the Santa Victoria Group, EDX analysis corroborates the occurrence of corrensite (Table 4b) and the absence of chlorite, in agreement with XRD analysis.

P-T estimates based on chlorite-phengite local equilibrium
The P-T conditions for the paragenesis Phe þ Chl þ Qtz þ H 2 O found in the Puncoviscana Formation were estimated considering the following end-members: muscovite, pyrophyllite, phlogopite, Al-and Fe-celadonite, sudoite, clinochlore, daphnite, Mg-and Fe-amesite, quartz, and water in the KFMASH system applying the TWQ software modified by Vidal & Parra (2000) and Vidal et al. (2005) to phengite-chlorite pairs linked to S 1 and S 2 .For S 1 two independent reactions were obtained in samples PU08-4, PU08-5 and PU08-6, considering muscovite, pyrophyllite, Fe-celadonite, Mg-celadonite, sudoite, clinochlore, quartz, and water, with 15 chlorite-mica pair intersection points between $225 and 315 C and $5.5 and 10 kbar (Fig. 10).Considering the same end-members, two independent reactions were obtained for S 2 in sample PU08-3, with 3 chlorite-mica pair intersection points between $275 and 350 C and $0.7 and 3 kbar (Fig. 11).No clear results were obtained for the phengite-chlorite pairs associated with S 1 in the rest of the samples analysed, nor for pairs linked to S 2 in the other sample in which this foliation is conspicuous (PU08-6).
The temperature estimated through the phengite-chlorite thermobarometer for blasts associated with S 1 agree with the one suggested by KI, as KI values for Puncoviscana phyllites in the study area largely correspond to high anchizone and epizone.The anchizone-epizone boundary (KI of 0.25) was correlated by Merriman & Peacor (1999) with temperatures of around 300 C.

Discussion
The two foliations (S 1 and S 2 ) identified in the metapelites of the Puncoviscana Formation, each one associated with a concomitant blastesis of mica and chlorite, allow defining two metamorphic deformational events, M 1 -D 1 and M 2 -D 2 .
In all phyllites of the Puncoviscana Formation from the San Isidro-Iruya area, dioctahedral mica with a wide range of Tschermak substitutions co-exist, as shown by Fe þ Mg as well as by Si contents.These variations could not arise from the disturbing effect of detrital white K-mica because SEM evidence indicates they are absent or represent a minor component.Most of the phengites with the highest Si contents correspond to M 1 -D 1 , but most of the micas associated with M 2 -D 2 show lower Si contents, with nearly all of the values in the range of 3.20-3.30apfu.
These compositional variations suggest that dioctahedral micas of individual slates crystallized at different pressures in response to changing metamorphic conditions.The decompression path of the sequence during the M 1 -D 1 event is recorded in chlorite-phengite pairs from three of the slates (PU08-4, PU08-5 and PU08-6, Fig. 12).Therefore, the lack of local thermodynamic equilibrium for chlorite-phengite pairs associated with S 1 foliation in several samples could be due to the progressive crystallization and compositional change of phyllosilicates through varying P-T metamorphic conditions during the M 1 -D 1 event.
According to the thermobarometric results the M 2 -D 2 event occurred at higher tempeature than the M 1 -D 1 event , probably overprinting previous mica and chlorite grains.This circumstance could explain the lack of thermodynamic equilibrium (even at microsites) that prevents reliable P-T estimates for M 1 -D 1 through the chlorite-phengite thermobarometer in some of the samples.
Two features are consistent with an initial intermediateto high-pressure/low-temperature event (M 1 -D 1 ) followed by a lower-pressure overprint (M 2 -D 2 ): (1) the higher  Geothermobarometry of low-grade rocks, Puncoviscana Formation (NW Argentina) Tschermak substitution shown by micas associated to S 1 compared with the ones associated to S 2 , and (2) the occurrence of inhomogeneous mica flakes depicting a phengitic core contrasting with a rim closer to end-member muscovite (Table 3, PU08-5: 3/4 and 3/6).Similarly zoned white Kmica packets with phengitic cores and muscovitic rims were described in a TEM-AEM and electron microprobe (EMP) study of anchizonal shales from the Franciscan Complex (Diablo Range, California) by Dalla-Torre et al. (1996).These authors concluded that the metamorphic evolution of the Franciscan Complex rocks included a high-pressure/ low-temperature (HP/LT) event related to a subductionzone tectonic regime, followed by a lower-pressure overprint possibly at higher temperatures than the HP/LT event.(2003b) concluded that this variability is a consequence of the history of mica crystallization throughout a decompression process of the Talas Ala-Tau rocks.We interpret similarly the wide range of Tschermak substitution shown by Puncoviscana phyllites at the single sample level.These results suggest that the application of the thermobarometer based on chlorite-phengite local equilibria in metapelites affected by polyphase deformation is not a straightforward tool, even when their post-depositional evolution is close to the anchizone-epizone boundary.In fact, consistent results for the M 1 and M 2 events, showing the attainment of local thermodynamic equilibrium for phyllosilicates associated respectively with S 1 and S 2 foliations, were only obtained in certain samples.In the case of metapelites recording S 2 foliation, equilibrium would have been promoted by the slight increase in temperature taking place from M 1 to M 2 (Fig. 12) that is characteristic of a clockwise P-T-t metamorphic path.Such a metamorphic evolution coincides with the P-T-t path established for Puncoviscana Formation metapelites in southern and southwestern outcrops (Do Campo & Nieto, 2003).
The maximum P values deduced herein for the M 1 -D 1 event affecting Puncoviscana metapelites (between 8 and 9 kbar) in the San Isidro-Iruya area are not far from the maximum pressure of 7 kbar derived by Do Campo & Nieto (2003) for southern localities according to the geobarometer based on the Si content of phengites (Massone & Schreyer, 1987).However, other barometric approaches reduce the deduced pressure obtained from Si values of 3.4 and 3.5 apfu to 4 kbar (Massone & Szpurka, 1997).The values obtained for the white mica b parameter (9.035-9.050)also imply intermediate/high-pressure Fig. 11.P-T conditions estimated for phengite-chlorite pairs linked to S 2 (PU08-3) applying the TWQ software modified by Vidal & Parra (2001).Local equilibrium in several pairs from sample PU08-3 is represented by the intersection of the black lines, corresponding to reactions between mica and chlorite components in the KFMASH system.References for the reactions are the same that in Fig. 10.facies series (Guidotti & Sassi, 1986).Since phengites associated to S 1 foliation clearly represent the majority of the micas in Puncoviscana metapelites, the b parameter should mostly indicate pressure conditions during the M 1 event as this parameter is related to the average compositions of the different mica grains within an individual sample (Abad et al., 2003b).
It is worth noting that the metapelites of the Puncoviscana Formation cropping out in the area of the San Isidro River have very similar clay-mineral assemblages (chlorite-phengite) and stronger metamorphic foliations than equivalent rocks from southern localities considered in previous studies (Do Campo, 1999;Do Campo & Nieto, 2003).These results demonstrate that the metamorphic grade of this unit throughout the Eastern Cordillera, particularly in the Lules-Puncoviscana Belt, is quite constant between 22 30' and 26 S, even though equivalent units cropping out farther south in the Sierras Pampeanas show medium-to highgrade metamorphism.
The geothermal gradient inferred for Puncoviscana Formation phyllites, as well as their clockwise P-T-t metamorphic path presented herein, represent key constraints to geodynamic models for the Neoproterozoic-Early Cambrian evolution of western Gondwana.Although the low geothermal gradient may be compatible with a forearc tectonic setting, the geochemistry of these rocks does not coincide with such a hypothesis (Do Campo & Ribeiro Guevara, 2005).On the other hand, although clockwise P-T-t metamorphic paths are typical of collisional orogens, they are characterized by normal to high geothermal gradients.Therefore, we consider that more studies will be needed to achieve a reliable geodynamic model of the western Gondwana margin for Neoproterozoic-Early Cambrian times.
The metamorphic grade attained for the Cambro-Ordovician successions during their post-depositional evolution in the study region is markedly lower than the high anchizone-epizone determined for the underlying Puncoviscana Formation.The KI (CIS) values obtained, corresponding to late diagenesis to low anchizone, are compatible with the occurrence of kaolinite, the corrensite identified in the levels of the Santa Victoria Group cropping out at Miyuyoc, and the higher degree of illitic substitution in the micas of the same sample.Merriman & Frey (1999) proposed temperatures close to 200 C for the late diagenesis-anchizone boundary.Furthermore, micas from the Santa Victoria Group are characterized by lower Tschermak substitution, indicative of crystallization under a lower P/T ratio, in agreement with the lower values obtained for the b parameter, as well as for the weakly developed S 1 foliation.
Consequently, this study demonstrates a difference of at least 100 C between the temperatures reached by the Meso ´n Group and the Puncoviscana Formation, indicating independent post-depositional histories.This is consistent with the sharp angular unconformity between the two units, which would have exhumed .10 km of metamorphic rocks before the Upper Cambrian.The outlined metamorphic evolution of the Puncoviscana Formation and overlying units will allow geodynamic models for the area to be improved.

Conclusions
The Vidal & Parra (2000) geothermobarometer was applied to very low-grade pelites of the Puncoviscana Formation, using chemical formulae obtained with EDX analyses of areas selected on BSE images under SEM; it yielded consistent P-T conditions for their decompression path.Micas and chlorites began to crystallize at around 8-9 kbar and 240-300 C and evolved through isothermal decompression to less than 6 kbar, with development of a second foliation at 2-3 kbar and a slight increase in temperature to 350 C at 1.5 kbar.This P-T-t evolution, based on chemical data, which includes a rise in temperature at the end of the decompression path, is nicely consistent with that predicted by Do Campo & Nieto (2003) in southern locations of Puncoviscana using XRD, HRTEM, and AEM data.
The overlying Cambro-Ordovician Meso ´n and Santa Victoria Groups present clearly lower P-T conditions, in the diagenetic field, which indicates a difference of at least 100 C, consistent with the sharp angular unconformity between the two units.
Thermobarometric calculations, according to Vidal & Parra (2001), on chemical data of mica-chlorite pairs obtained by SEM has proven to be a powerful tool to establish very low-grade metamorphic conditions and their evolution over time.
Abad et al. (2003b) also described two generations of micas in anchizonal slates of Talas Ala-Tau consisting of an intergrowth of phengites and muscovites in phyllosilicate stacks.Based on the textural relations, the authors inferred that the phengitic laths corresponded to an earlier episode, whereas the growth of muscovitic laths must have been later during decompression.Like the Puncoviscana metapelites, the micas of Talas Ala-Tau slates show a wide range of phengitic contents at the individual sample level.Abad et al.

Fig. 12 .
Fig. 12. P-T diagram showing the metamorphic facies and the localization of equilibrium points (Fig. 10 and 11) from pairs associated to S 1 (light grey circles) and S 2 (white circles) foliations.The P-T area indicated for the Meso ´n (M) and Santa Victoria (SV) Groups, showed for comparison, is only estimated from equivalent values of KI (CIS), white mica b parameter and clay mineral composition.

Table 1 .
Whitney & Evans (2010)ion of the clay fraction based on XRD results.KI (CIS scale) and b parameter are also shown.Abbreviations of minerals according toWhitney & Evans (2010).
eschweizerbart_xxx the KI data is in agreement with evidence from previous XRD and TEM studies in southern localities (Do Campo,

Table 4 .
Chemical composition (as anhydrous oxide wt%) and structural formulae for dioctahedral micas and corrensite from the Santa Victoria Group.Corrensite analyses were normalized for 50 negative charges (28 þ 22) per formulae unit and an Fe 3þ /Fe 2þ þ Fe 3þ ratio equal to 0.