Dating low-grade metamorphism and deformation of the Espinhaço Supergroup in the Chapada Diamantina ( Bahia , NE Brazil ) : a K / Ar fine-fraction study

Manuscrito ID 30062. Recebido em: 14/05/2014. Aprovado em: 09/06/2014. ABSTRACT: This study focuses on the northernmost part of the Mesoproterozoic Espinhaço Supergroup that crops out in the Chapada Diamantina. The fine-fraction K/Ar dating obtained on slightly metamorphosed sediments of the siliciclastic Espinhaço Supergroup shows a polyphase deformation history that corresponds to the Brasiliano (Pan-African) orogenic cycle. The isotopic results are interpreted to indicate three age domains coincident with three structurally different domains. Constrained by the Kübler Index (‘illite crystallinity’) and illite polytypism, the thermal conditions generated during the tectonic activity show a gradual trend from the craton margins to the interior from epizonal to diagenetic. The northern Chapada Diamantina is situated in the foreland of the Riacho do Pontal belt and comprises the sediments of the Espinhaço Supergroup northeast of the Irecê basin. The K/Ar ages for < 2 μm illite fractions range between 645 and 621 Ma [mean 637±9 Ma (2σ)] and for < 0.2 μm fraction range between 625 and 603 Ma [mean 614±9 Ma (2σ)]. Samples from the central Chapada Diamantina east of the Irecê basin are not affected by a Brasiliano deformation event and therefore, the N-S-trending structures are assumed to be older. The deformation of the southern Chapada Diamantina was established in conjunction with the formation of the Araçuai orogenesis and the inversion and reactivation of the Paramirim impactogen. The last stage of deformation in this area is recorded by the K/Ar fine-fraction dating between 470 and 460 Ma.


INTRODUCTION
The study area Chapada Diamantina (Bahia, NE Brazil) represents the northern part of the wide Espinhaço basin (extending from Minas Gerais to Piaui, along the São Francisco Craton, > 1,100 km) (Fig. 1).In this study the focus will be on the northernmost part of the mesoproterozoic Espinhaço Supergroup, which crops out in the Chapada Diamantina.
Here, the sediment succession mainly consists of quartzites and metapelites and is unconformably overlying the basement rocks of the São Francisco Craton (Pedreira & De Waele 2008 and references therein).After an early stage of rifting, the Espinhaço Supergroup underwent metamorphism and deformation during the amalgamation of the western Gondwana paleocontinent during the so-called neoproterozoic Brasiliano-orogenic cycle between 650 and 550 Ma (Schobbenhaus 1996;Danderfer 2000).This collision was responsible for the present structural configuration of the mesoproterozoic sediment sequence, including a km-scale folding with dominating NNW, in parts SW-trending axes (Alkmim et al. 2006).Alvarenga and Dardenne (1978) showed that extensively aligned faults, mostly NNE-SSW, may represent basement structures reactivated during sedimentation.
Because fairly little is currently known about Mesoproterozoic events within the São Francisco Craton, mainly because of lack of reliable age determination, it is not surprising that a possible folding of the Espinhaço Supergroup before the Neoproterozoic Brasiliano orogeny is controversial (Fuck et al. 2008;Cordani et al. 2010).However, the foreland fold and thrust belt recorded in the Irecê basin, in the northwestern part of the Chapada Diamantina, is characterized by E-W structures and is unconformably overlying the NNW-SSE folding observed in the Espinhaço strata.The structural polarity of the latter is due to the active margin of the São Francisco Craton, and can be attributed to the amalgamation of Gondwana at about 600 -550 Ma (Cordani et al. 2010).
Previous efforts to date the deformation were made on adjacent units but not for the Espinhaço Supergroup (in the Chapada Diamantina) itself.Cordani et al. (1992) reported K/Ar cooling ages in minerals separated from the graniteorthogneiss complex Lagoa Real (southwest of the Chapada Diamantina), which fall within the interval of 570 -500 Ma, indicating the influence of a late Brasiliano tectonothermal cycle.Investigations on undeformed carbonates of the Una Group were carried out by Misi and Veizer (1998), suggesting that the sedimentation (670 -600 Ma) of the Saltire Formation (Una Group, Irecê Basin) and of the correlative units of the Bambuí Group (São Francisco Basin) was coeval with the onset of the Brasiliano tectonic cycle.
The present study aims to elucidate the deformation history of the Chapada Diamantina and its supposed relation to an orogenic cycle, either Pan-African/Brasiliano or Grenvillian.Besides structural and tectonic analyses, the K/Ar fine-fraction dating is an important tool for reconstruction of the deformation history of very low grade metamorphic rocks.This method has proved to be a very useful technique dating illites formed by prograde metamorphic reactions during the deformation (Ahrendt et al. 1978;Clauer et al. 1995Clauer et al. , 1997)).In particular, metapelitic rocks proved to be very suitable because of their rheological properties that make them highly sensitive even to minor variations in the regional stress field.The cleavage planes on which the newly crystallized illites are expected are in this context of particular importance and will be used in an attempt to shed some light on the final lowgrade deformation events.
The age of synkinematically newly grown illites was determined using K/Ar geochronology on fine mineral fractions.This isotopic method was combined with measurements of the illite crystallinity (expressed as Kübler Index) and polytype quantification to unravel the degree of metamorphism and a possibly polyphase thermal overprinting in the Chapada Diamantina.The presented results are considered within the frame of the regional tectonic acting on the southwestern margin of Gondwana.

GEOLOGICAL SETTING
The sediments of the Chapada Diamantina comprise the northwestern part of the São Francisco Craton and are the focal point of this study.The eastern part of the São Francisco Craton underwent rifting during Mesoproterozoic times, leading to the development of the wide Espinhaço basin and the Paramirim valley.What at first began as failed rift resulted in the formation of sag basins that were filled with sandstones, conglomerates, fine-grained rocks and limestones, summarized as Espinhaço Supergroup (Pedreira & De Waele 2008 and references therein).The onset of sedimentation, determined by U-Pb analyses of detrital zircons, yielded ages between 1.8 and 1.76 Ga (Turpin et al. 1988;Machado et al. 1989;Cordani et al. 1992;Babinski 1993).This approximately 2-km-thick sediment succession crops out in three domains namely Chapada Diamantina, Northern-and Southern Espinhaço in which all are related to the orographic system Espinhaço Range.In the Chapada Diamantina, the Espinhaço Supergroup has been divided into three groups that comprise six formations; from base to top, Rio dos Remédios (Ouricuri do Ouro Fm.), Paraguaçu (Mangabeira Fm., Guiné Fm.) and Chapada Diamantina group (Tombador Fm., Caboclo Fm and Morro do Chapéu Fm.) (Inda & Barbosa 1978;Pedreia 1994;Pedreia & Rocha 2004).
The Chapada Diamantina domain, situated in the central part of the Bahia state, is limited to the north and south by Brasiliano orogenic belts that circumscribe the foreland

Structural lineaments
The structural lineaments (folds) that cross the Mesoproterozoic sediment cover of the Chapada Diamantina are evidence enough to show that the sediments were affected by different deformation events that are linked to the surrounding fold and thrust belts.The NNW-SSE-trending Paramirim impactogen divides the São Francisco Craton in two geological domains: on the western side the Mesoand Neoproterozoic sediments of the Espinhaço Supergroup and Bambuí Group, and on the eastern side the equivalent sediment sequences of the Chapada Diamantina and Una Group.The Chapada Diamantina can structurally be divided in at least three sections, which show as a far field response to the intense deformation that took place along the craton margin, NNW-SSE-trending fold axes as in the southwest, NNE-SSW-trending fold axes as in the central part, or WSW-ENE-trending fold axes to the north.Those folded and re-folded structures prove an earlier polyphasic deformation of the Espinhaço Supergroup, which is until now supposedly attributed to a Brasiliano orogeny.According to Torquato and Fogaça (1981) the deformation in the Chapada Diamantina increases progressively from north to south and does not exceed the greenschist facies based on illite crystallinities from the Tombador formation.Whereas, a deformation under greenschist facies conditions is recorded for the Morro do Chapéu formation (Brito Neves et al. 2012).

MATERIALS AND METHODS
We collected fresh metapelitic material from 10 sample localities in the Chapada Diamantina, which was subjected to K/Ar and X-ray diffraction (XRD) analyses to define the age of the deformation under lower greenschist facies conditions (Tab. 1, Fig. 2).The samples are covering a N-S section of roughly 200 km.
All samples were cleaned of weathered crusts and subsequently, gently crushed to avoid artificial fragmentation of the original grain components by a shatter mill and passed through a 63-mm sieve.The size fractions (< 2 and 2 -6 mm) were separated by differential settling in distilled water according to Stoke's law (Atterberg method).Enrichment of the grain size fraction < 0.2 µm was carried out by centrifugation, following the procedure described in Wemmer (1991).The size fractions were collected using a Millipore filter.Subsequently, the concentrates were subjected to isotope measurements and ■ The full-width at half maximum (FWHM), also known as Kübler Index (KI), of the illite 001 peak (Kübler 1967).The illite crystallinity is a function of the grain size and increases with increasing grain size.The determination of the illite crystallinity is performed on glass plates, according to Weber (1972), with "thin" texture compounds prepared using 1.5 -2.5 mg/cm².Samples were scanned from 7 to 10° 2Ө, using a step size of 0.01° 2Ө and an integration time of 4 s.The illite crystallinity was determined using a computer program (Idefix) developed at the Geosciences Center of the University of Göttingen by Ullemeyer.Limits between the fields of diagenesis, achi-, and epizone are given by 0.420° ∆2Ө CuKα and 0.250° ∆2Ө CuKα, respectively (Kübler 1967;Marheine 1997).■ The mineral content to verify the presence of illite and to rule out the presence of other K-bearing minerals that may influence the age data.To identify the mineral content, we performed a stepscan (0.02° 2Ө) in the range from 4 to 70° 2Ө.■ Polytype quantification.We prepared the size fraction as random powder mounts using the sideload method described and recommended in Moore and Reynolds (1997) and Grathoff et al. (1998).The concentration of different polytypes is determined by measuring the area/ intensity of polytype-specific peaks (in case of 2M 1 of five specific peaks) and rationing them against the area/ intensity of a peak at 2.58 Å, which is common at both polytypes (Grathoff & Moore 1996).These methods represent a significant improvement compared to earlier techniques that relied on the measurement of only one specific 2M 1 peak such as from Hower et al. (1963), summarized in Grathoff and Moore (1996).Pevear (1992) and Grathoff and Moore (1996) used different polytypes to distinguish diagenetic from detrital illite and suggested methods for extrapolating to the ages of diagenetic and detrital illites in shales.By plotting the K/Ar ages of at least three size fractions of illite against the percentage of illite that is detrital (2M 1 ), extrapolation to zero detrital illite gives the age of diagenesis, and extrapolation to zero authigenic/diagenetic illite gives the age of the detrital illite.It is noteworthy that based on XRD analysis, high-temperature 2M 1 illite cannot be distinguished from muscovite, which is also a 2M 1 polytype.
The samples were scanned from 16 to 44° 2Ө in the step-scanning mode with a step size of 0.05° and a counting time of at least 30 s/step to gain a high peak-to-background resolution.We prevented the problem of peak interferences with other clay minerals (kaolinite) by heating the sample to 550°C for 1 h, which caused dehydroxylation of kaolinite.
The age determination analyses proceeded in two stages, an analysis of potassium and an analysis of argon.The potassium (K 2 O) was analyzed in duplicates by atomic absorption spectrophotometer using a PerkinElmer 5100 PC.The samples were dissolved in a mixture of HF and HNO 3 followed by addition of CsCl suspension (12.5%) as an internal standard and ionization buffer for alkali metals.The pooled error of the duplicate potassium determinations in samples and standards is more than 2%.For Ar analyses, the samples were pre-heated at high vacuum at 120°C for at least 24 h to reduce the amount of atmospheric argon added to the mineral surfaces during handling.The argon isotopic composition was analyzed in a Pyrex glass extraction and purification line coupled to a VG 1200 C noble-gas mass spectrometer operating at static mode.The amount of radiogenic 40 Ar was determined by isotope dilution method using a highly enriched 38 Ar spike (Schuhmacher 1975), which was calibrated against the international biotite standard HD-B1 (Fuhrmann et al. 1987).The extracted gases were purified in a multistage process by Ti-and SORB-ACs getters.The reproducibility of Ar isotopic results was controlled by repetitive analysis of the standard HD-B1.The obtained values were considered to be satisfying and no corrections were applied to the raw data.The K/Ar ages were calculated with the usual decay constants (Steiger & Jäger 1977), the overall error being < 1% (2σ).

XRD mineralogy
The mineral content of metapelites in the Espinhaço Supergroup is dominated by illite, kaolinite and quartz with minor amounts of hematite, pyrophyllite and gypsum (Tab.2).However, there is no indication that any other potassiumor calcium-bearing mineral disturbs the isotopic system.

Illite crystallinity index
To identify retrograde grown mixed-layered minerals, we compared the KI values of oriented air-dried and glycolated samples.As the data yield identical values, the existence of smectite-type minerals can be excluded.Differences between both values are not significant.The KI of all analyzed samples varies approximately 0.136 to 0.458 ∆° 2Ө CuKα.Almost all obtained KI values are indicative of epimetamorphic conditions of approximately 300 -350°C (KI < 0.250 ∆° 2Ө CuKα; Tab.3), unless they result from the occurrence of detrital micas.Only two samples deviate from those, in case of sample BBB 11-09 anchi-metamorphic (< 2 µm) conditions and for sample BBB 13-09 diagenetic (< 0.2 µm) conditions can be identified.In spite of the small number of samples, a decreasing KI value from the interior basin to the margin can be observed, which corresponds to an increase in the stacking ordering and higher metamorphic conditions (Fig. 3).Hence, as the major part of the samples show epimetamorphic conditions, the age data should be interpreted as cooling ages during retrograde metamorphism.Any detrital memory in the investigated fine fractions was certainly removed.It can be noted that increasing metamorphic conditions (decreasing KI values) do not correlate with the stratigraphic position within the succession.

Polytype
The (060) and hkl XRD peaks of randomly oriented powder samples were used to distinguish different illite polytypes of the metapelitic samples either detrital 2M 1 or different generations of authigenic 1M d and 1M polytypes.The samples always reflect a mixture of the three polytypes with varying amounts depending on the grain size fractions.A general trend shows that the percentage of 2M 1 illite decreases as particle size decreases.Correspondingly, the amount of 1M d polytype decreases with increase in grain size.Most samples contain variable amounts of 1M d /1M and 2M 1 illites (Tab.3), whereas 2M 1 illite is generally the main component even in the < 0.2 µm fraction.This is not the case for samples BBB 11-09 and BBB 13-09, which show 1M d /1M illite as dominant polytype in all three size fractions.The < 0.2 µm fraction is relatively enriched in most recently grown 1M illite, which crystallized in the final phase of metamorphism (Lee et al. 1989).Samples from the northern and southern Chapada Diamantina show a 2M 1 proportion from 50 to 90%.As there was insufficient material for sample BBB 34-09 no data were acquired.Although the content of obtained polytypes correlates with the KI, and hence, with the intensity of deformation, there is no correlation between the KI values and the stratigraphic position of the samples

K/Ar age data
The 10 analyzed samples from metapelitic rocks (total 29 grain size fractions) provide age values ranging from 1,265.4 ± 18 Ma for BBB 11-09 (2 -6 µm) to approximately 447.8 ± 6.9 Ma for BBB 33-09 (< 0.2 µm).The results are presented in Tab. 3 and Fig. 4. As there was relatively little material for BBB 11-09, no < 0.2 µm fraction could be separated.Radiogenic 40 Ar content ranges between 92.29 and 99.38%, indicating reliable analytical conditions for all analyses.The potassium content in the studied samples ranges from 1.61 to 9.11 K 2 O%.This large scatter results from mixing with other finegrained minerals in the fine mineral fractions.The measured age data change with particle size and polytype.Three age groups are found, depending on their location within the Espinhaço Basin and Chapada Diamantina: (1) The northern Chapada Diamantina yield ages in the < 2 µm fraction to be approximately 582.9 ± 8.7 to 644.9± 9.6 Ma for the CHD-R-C3, CHD-R-D3 and CHD-R-D4 samples.The recorded distribution of different grain size fractions in these samples overlaps within errors.

DISCUSSION
The following section reviews the most important approaches to interpret fine-fraction data: The crucial aspect, however, is the temperature, which reaches at the maximum of the metamorphism.On the one hand, it controls the neocrystallization of illite and its crystallinity, on the other hand, exceeding closing temperature leads to a resetting of the mineral ages.The understanding of the closing behavior of small illite crystals has a fundamental influence on the interpretation.The importance of the effective diffusion radius on the closure temperature for the Ar system has been shown by a large number of reports (e.g.Dahl 1996;Villa 1998;Hodges 2003).These studies focused on white micas with grain sizes > 200 µm.Hunziker et al. (1986) indicated a closure temperature interval of 260 ± 30°C for the mica fractions < 2 µm, whereas Wemmer and Ahrendt (1997) reported K/Ar ages on fine-grained white mica (sericite < 2 µm) that did not behave as open systems, confirming that even temperatures approximately 275°C did not reset the Ar-system.Therefore, the closure temperature of fine-grained mica has to be estimated somewhere between 275 and 350°C for grain size fractions < 2 µm (Wemmer & Ahrendt 1997).However, this value is significantly lower than the closure temperature interval 350 ± 50°C for coarse-grained micas recommended by Purdy and Jäger (1976) and McDougall and Harrison (1999).Attention should also be given to the case of a potential contamination by other K-bearing phases.Potassium feldspar, for example, can make the age appear younger due to the lower closing temperature (Fitz-Gerald & Harrison 1993).Otherwise, Zwingmann et al. (2010) showed that K/Ar ages of fault gouge samples from the Swiss Alps are independent of K-feldspar content.This suggests that the fine-grained feldspar contaminant was isotopically reset during deformation approximately at the same time when illite growth occurred.Focus should be given on the problem that might result due to a mixture of different generations of illite growth.In general, the smallest grain size fractions (< 0.2 µm) are the least contaminated, and therefore, closest to the age of the youngest deformation event and should represent the most recently grown illites.Similar increases in both the K/Ar age data and the 2M 1 content with increasing grain size, representing earlier illite-forming events, have been reported in many previous studies (e.g.Hower et al. 1963;Grathoff & Moore 1996;Clauer et al. 1997;Grathoff et al. 1998).A highly simplified scheme to interpret K/Ar ages in various scenarios of mixing different polytypes is given in Marheine (1997).In principal, in low-grade sedimentary rocks, it can be assumed that the 2M 1 polytype is detrital whereas the 1M and 1M d polytype are authigenic products formed during diagenesis to anchi-metamorphic conditions during subsequent burial (e.g.Grathoff & Moore 1996).On the basis of XRD, the 2M 1 illite cannot be distinguished from detrital micas (also a 2M 1 polytype).Only the combined application of these independent results (age data, KI values and polytype proportions) allows differentiating between both.Type 2M 1 illite forms at temperatures approximately 200 -250°C (Velde 1965), and if the diagenetic reaction from 1M d to 1M and  finally to 2M 1 from laboratory experiments of Velde (1965) also applies for metapelitic rocks, high concentration of 2M 1 illite can be used as an additional indicator for temperature (>>200°).1M/1M d are thought to form at significantly lower temperatures, below 200°C (Velde 1965).2M 1 illite should be expected in shales because its source is the weathering mica flakes of the host rock.
The results of K/Ar geochronology (Tab.3) of metapelitic rocks show that three domains can be distinguished: northern, central and southern Chapada Diamantina.These domains differ in their age and structural configuration.The latest Brasiliano overprint was indentified in the southern Chapada Diamantina, the oldest deformation age in the northern Chapada Diamantina.
The northern Chapada Diamantina comprises the SW-NE-trending Mesoproterozoic sediments of the Espinhaço Supergroup northeast of the Irecê basin.The investigated samples of this area (CHD-R-D3 and CHD-R-D4) yield an average age of 632.7 ± 9.5 Ma in the < 2 µm fraction, indicating a maximum age for the latest deformation and peak metamorphism.The final growth of illites is dated at 614 ± 9.1 Ma (< 0.2 µm fraction).As all samples yield KI values characteristic for epimetamorphic conditions, it is to be expected that all authigenic prograde-formed illites completely transformed from 1M d /1M to 2M 1 .Still more evidence for the onset of deformation under epizonal conditions is supported by the high amount of 2M 1 illites, which even in the smallest grain size fraction (< 0.2 µm) exceeds 60%.Therefore, the occurrence of 1M and 1M d illite can be explained by retrograde metamorphism and growth during decreasing temperatures.Hence, the data can at best be described as cooling ages, which implies that the < 2 µm fraction should be closer to the peak metamorphism than the < 0.2 µm fraction.As the 1M/1M d content is relatively low, it is indicative of a rapid single-phase cooling history.Thus, it is clear that the age of the authigenic illite is also the age of the last deformation event.However, it remains questionable whether the slightly younger age of the sample CHD-R-C3 of 582.9 ± 8.7 Ma (< 2 µm fraction) can be attributed to local tectonic activities or an alteration/ regional fluid migration effect.However, rejuvenation due to potassium-bearing mineral phase can be excluded from the XRD data.
The K/Ar fine-fraction ages are synchronous with a widespread Precambrian deformation event located between the northern margin of the São Francisco Craton and the Pernambuco-Alagoas massif (approximately 200 -250 km to the north of CHD-R-D4) that culminated in the final assembly of West Gondwana.This event resulted among other structural elements in SW-trending fold axes and built up the Riacho do Pontal belt.Brito Neves et al. (1980Neves et al. ( , 1999) ) described that age data approximately 630 Ma, which have been inferred for the main phase of folding, whereas post-tectonic plutons are dated between 580 and 540 Ma (Rb-Sr; Jardim de Sá 1994).Within the northern Chapada Diamantina, the deformation and thermal stress were completed at that point, while apparently at least 100 km to the north in the Riacho do Pontal belt the collision continued.The supposed age of 630 Ma for the major peak of metamorphism could be confirmed with unpublished U-Pb and Rb-Sr data (Brito Neves, in preparation) and is identical to our data as well as the structural alignment of SW-trending fold axes in the northern Chapada Diamantina.
The results derived from illite polytype quantification and age dating indicate that post-tectonic plutonism, which occurred in the Borborema province, did not affect the Mesoproterozoic Espinhaço Supergroup of the northern Chapada Diamantina.This appears also to be in good agreement with the observation of a northward increasing of deformation made by Inda and Barbosa (1978).
A resetting of the isotopic and magnetic systems that affected the Neoproterozoic sediments of the Irecê basin and the mesoproterozoic Caboclo formation at approximately 520 Ma can be attributed to a regional-scale fluid migration and mineralization as a repercussion of the Brasiliano orogenic cycle (Cunha et al. 2003;Trindade et al. 2004;Texeira et al. 2010) and is, therefore, negligible in this context.
The central Chapada Diamantina is bounded by the Irecê basin to the west and the Jacobina belt to the east.The samples BBB 11-09 and BBB 13-09 can be assigned to this region.BBB 11-09 of the Caboclo formation yield ages of 1,089.8 ± 16.3 (< 2 µm fraction) and 1,265.4± 18 Ma (2 -6 µm fraction).The interpretation of this data is affected by the lack of the < 0.2 µm fraction, which is most sensitive for thermal impulses.Owing to the large age difference within the two size fractions, it can be assumed that these ages do not correspond to the total resetting of the isotopic clock.Therefore, these are meaningless mixed ages of detrital and authigenic minerals.This is supported by the KI values indicative of anchizonal metamorphic conditions and the low content of 2M 1 illite, giving evidence for, if at all, weak deformation intensity.The high 1M/1M d content of 60% is probably due to a weak retrograde overprint/alteration involving new growth of low-temperature clay minerals.Thus, two conclusions can be drawn: first, the deformation was insufficient to reset the detrital age of the illites, and second, the results cannot be assigned to a Brasiliano deformation event as it was identified in the northern and southern Chapada Diamantina.This seems plausible because the sediments of the Morro do Chapéu and Ventura area are covering the archaic rocks of the Gavião/Lençois block, which is considered to be geologically stable.BBB 13-09 was investigated in more detail and shows a relatively large time period from 616 ± 9.2 Ma in the smallest size fraction to 730.6 ± 10.9 Ma in the 2 -6 µm size fraction.The age of the < 0.2 µm fraction corresponds to the deformation of the northern Chapada Diamantina, indicating that at least the smallest size fraction adapt to the last low-grade thermal event.The decrease in temperature is followed by a decrease in deformation intensity as can be observed by KI and illite polytypism.The apparent diagenetic and detrital ages of the illites using linear extrapolation, read from Fig. 5, are 616 -610 and 870 Ma.We say apparent because these are calculated ages for the two extreme cases: 100% 1M/1M d and 100% 2M 1 .The upper limitation for the apparent diagenetic age is given by the < 0.2 µm age because even the smallest size fraction is not completely free from 2M 1 illite.The extrapolated "detrital" component age for BBB 13-09 yielded an age of approximately 860 Ma.As this age is significantly less than the stratigraphic age (> 1,200 Ma); it clearly does not reflect the real sedimentary detritus.It is apparent that the coarser size fractions contain varying components of two illite polytype populations, an older 2M 1 polytype (probably detrital muscovite that gives an averaging age of the formation of the 2M 1 polytype in their pre-depositional provenance areas) and a younger synkinematic 1M illite.However, the detrital component reflects a total reset that is not related to the visible deformation but presumably related to a thermal event (> 350°C).This event can be related to an episode of rifting that affected the interior of the São Francisco Craton and began probably approximately 900 Ma (Pedrosa-Soares et al. 2001;Tack et al. 2001).
Taking all the results into account, it is likely that the sediments in this region were less strongly affected by the collision at the northern craton margin than the samples belonging to the northern Chapada Diamantina discussed above, and in addition, there are no indications of intra-plate tectonic episodes of "Grenvillian-type age" that triggered new growth of illite.
The southern Chapada Diamantina comprises the area situated to the south of the Irecê basin with mainly NNW-SSE-trending fold axes from where samples BBB 29-09, BBB 32-09, BBB 33-09, BBB 34-09 and BBB 38-09 were obtained.Samples BBB 32-09, BBB 33-09 and BBB 34-09 are of particular interest as they were taken directly from the hinge zone.Hinge zones are characterized by a pronounced cleavage in which new growth of synkinematic minerals is expected and, therefore, the deformation is highly preserved.The last illite growth, and therefore the last deformation of the southern Espinhaço Supergroup, occurred between 460.4 ± 6.9 and 447.8 ± 6.9 Ma (< 0.2 µm fraction).The < 2 µm fraction recorded slightly older ages in the range of 470.1 ± 7 to 464.9 ± 6.9 Ma.These samples show overlapping ages within errors, suggesting the absence of detrital contamination and a synchronous deformation history independent of the stratigraphic position (Ouricuri do Ouro Fm -Guiné Fm.).As the KI data yield epizonal conditions, it can be assumed that the fine fractions are composed of recessed detrital micas and retrograde-formed authigenic illites.The shown polytype data, high 2M 1 and low 1M/1M d content, in combination with the age data suggest a short-lived deformation event approximately 470 Ma and a single-phase rapid-cooling path.The slightly younger age of the sample BBB 33-09 is reflected in the lower 2M 1 content, probably caused by a local alteration effect.Thus, more emphasis should be given to the ages of samples BBB 32-09 and BBB 34-09 as maximum ages for the final metamorphism.The slightly older ages of samples BBB 38-09 and BBB 29-09 are a combination of the regional distribution and structural position.Both samples were not taken from hinge zones and are therefore less deformed.
The large synclines and anticlines at the eastern margin of the Paramirim impactogen, extending along hundreds of kilometers with NNW-SSE trends, are related to the formation of the Araçuaí belt, which is now identified as a late stage Brasiliano event approximately 470 Ma.In addition, this interpretation is supported by both Danderfer and Dardenne (2002) and Cruz and Alkmim (2006) who postulated that the Paramirim impactogen underwent partial inversion during the Brasiliano orogeny.In the southern two thirds of the impactogen, this inversion reactivated pre-existing normal faults and transfer faults to generate NNW-trending basement-involved thrust faults linked by strike slip shear zones (Alkmim et al. 2006).As the NNW-trending faults and folds of the Paramirim impactogen, which formed during the inversion, overprint the E-W-trending thrust and folds of the Rio Padro Salient, they are probably younger (Cruz & Alkmim 2006).This seems to be reasonable regarding the latest post-collisional activity approximately 490 Ma within the Araçuaí orogeny (Pedrosa-Soares et al. 2008).
By comparison, the age data obtained in this study are slightly younger than K/Ar hornblende cooling ages of the São Timoteo Granite in the Lagoa Real complex (ca. 100 km in the southwest of localities of samples BBB 32-09, BBB 33-09 and BBB 34-09) dated at 510 -490 Ma by Cordani et al. (1992).Turpin et al. (1988) interpreted discordant zircon U/Pb ages approximately 500 Ma as the age of metamorphism of the Lagoa Real complex.This coincides with that reported by Pimentel et al. (unpublished data) on zircon and sphene grains of the same complex that yielded discordant ages with an upper intercept at 1,740 Ma and a lower intercept at 490 Ma.The slightly lower fine-fraction ages (approximately 30 Ma), found in this study, can be explained due to the higher stratigraphic position and the lower closing temperature.Therefore, they are attributable to the thermal event that also affected the Lagoa Real complex and provide evidence for a wide-ranging deformation event.On the basis of the K/Ar fine-fraction data, a reactivation and inversion of the Paramirim impactogen and folding after 460 Ma can be excluded.

CONCLUSION
The K/Ar ages presented here suggest that the portion of the Espinhaço Supergroup, which crops out in the Chapada Diamantina, underwent metamorphism and deformation in various areas at different times, leading to the formation of distinct structural features.The deformation of the northern Chapada Diamantina occurred in the course of the collision between the northern margin of the São Francisco Craton and the Pernambuco-Alagoas massif.This event resulted in SW-trending fold axes and built up the Riacho do Pontal belt.K-Ar fine-fraction dating yield an age of 630 Ma for this event, which is coeval to the peak metamorphism in the Riacho do Pontal belt.Till now, no statement can be made on the N-S/NNW-SSE-trending axes of the sediments in the area east of the Irecê basin.The deformation of the southern Chapada Diamantina was established in conjunction with the emergence of the Araçuai orogen and the inversion and reactivation of the Paramirim impactogen.The final deformation phase is recorded by the K/Ar fine-fraction dating in the range between 470 and 460 Ma.
Figure 1.(A) Map of Brazil with highlighted São Francisco Craton.(B) Geological map of the São Francisco Craton and surrounding belts.(C) The working area is marked by the black rectangular.

Figure 2 .
Figure 2. Regional digital elevation model showing the location of the samples analyzed in this study.

Table 1 .
(Moore & Reynolds 1997localities and stratigraphic position XRD analyses.The XRD analyses were carried out using a PW 1800 Phillips diffractometer (CuKα radiation, at 45 kV and 40 mA conditions).Oriented clay aggregates were prepared onto glass slides(Moore & Reynolds 1997) and analyzed by XRD both after airdrying and after ethylene glycol treatment.From the resulting XRD patterns, we determined:

Table 2 .
Semi-quantitative results of X-ray diffraction analyses