Introduction

Cratons and orogens reflect two distinct types of lithospheres, characterized by their different genesis and by their response to posterior tectonic history. Orogenic lithosphere is generally easily reactivated and reworked, while the cold and rigid cratonic lithosphere is more resilient to major tectonic events1,2. In Brazil, cratons and intervening orogens compose the current passive Atlantic margin, where they are shown to have different responses to the Cretaceous rifting of the South Atlantic Ocean and later tectonic stresses3,4. Initial rifting during West Gondwana break-up eventually separated the South American plate (Brazil) from its (West) African counterpart. The two conjugate Atlantic passive margins further remained in distal positions with respect to tectonically active regions5. Considering the large dataset of apatite-fission track (AFT) thermochronology from the Brazilian passive margin, it was possible to verify that, for example, samples from the São Francisco craton (SFC) mainly yield AFT ages older than the Early Cretaceous, i.e. pre-rift (Fig. 1), whereas, in the adjacent orogens, i.e. Araçuaí belt (AB) and Borborema Province, AFT ages tend to be younger than the Early Cretaceous, i.e. rift to post-rift (Fig. 1). These results point to a first-order influence of lithosphere composition and properties with respect to basement exhumation4.

Figure 1
figure 1

Geological context of the study area. (a) The Araçuaí-West Congo Orogen (AWCO) and adjacent São Francisco-Congo craton in the context of West Gondwana. (b) Published AFT ages4. SFC São Francisco craton, AB Araçuaí belt, BP Borborema province, C-A Camamu/Almada basin, J Jequiqinhonha basin. (c) Simplified geologic map of the northern AB and the high grade/granitic core of the AWCO7. IZ Interference zone, RP Rio Pardo salient. Cities A Araçuaí, M Macaúbas. The maps were created using Corel Draw Graphics Suite 2018 (http://www.coreldraw.com).

Full size image

Orogenic phases may rework cratonic lithosphere at its borders, mainly by thrusting and folding of the cratonic marginal terranes or by partial inversion of aulacogen structures in the craton. While the SFC remained weakly deformed, the northern half of the AB, in eastern Brazil, was generated by the inversion of pre-existent NNW-trending rift structures and nucleation of new shear zones by WSW–ENE compression during the Ediacaran–Cambrian orogeny6,7. The objective of our study is to gain deeper insights into the exhumation of the basement in an area of transition between the SFC and the AB, during subsequent Phanerozoic tectonic events and, thus, verify how different types of lithosphere exert control on later basement exhumation.

Geological background

The São Francisco craton (SFC) consists of an Archean-Paleoproterozoic basement older than 1.8 Ga, with Mesoproterozoic to Cambrian cover units8(Fig. 2). To the southeast of the craton, the Araçuaí belt (AB) involving a basement older than 1.8 Ga and Proterozoic-Cambrian cover complexes, encompasses two large scale features: (i) a large arcuate salient, the Rio Pardo Salient, and (ii) a NNW-trending so-called interference zone (IZ in the Fig. 1) with a subparallel oriented Proterozoic intracontinental rift structure, the Paramirim Aulacogen8 (Figs. 1 and 2). The interference zone is characterized by the inversion of the normal faults of the aulacogen phase and nucleation of thrusts, folds, and dextral shear zones. The AB, as a whole, is a segment of the wider Araçuaí-West Congo Orogen (AWCO) (Fig. 1), which developed along the margin of the SFC and the Congo Craton during the Ediacaran to Early Cambrian (9 and references therein). The AWCO evolved due to the closure of an embayment, i.e. the terminal branch of the Adamastor ocean, carved into the São Francisco-Congo paleocontinent, and eventually resulted in the assembly of West Gondwana9.

Figure 2
figure 2

(a) Geological map of the study area (modified from8) with indication of sample locations, central AFT ages (in Ma), Mean Track Lengths (MTL) (in μm), and MTL standard deviation. (b) Expected t-T paths from thermal history modelling (QTQt10; individual data in Supplementary S.1 online), grouped by thermal history trend. The map and charts were created using Corel Draw Graphics Suite 2018 (http://www.coreldraw.com) and ArcGIS 10.4.1 (https://www.esri.com).

Full size image

The SFC and the AB remained confined in the interior of West Gondwana from the early Paleozoic until the rifting and break-up of this supercontinent around 130 Ma5. The separation between South America and Africa took place with the opening of the South Atlantic Ocean and resulted in the shifting of the regional base level of our study area towards the present-day coast. Currently, the Contas River drains the area to the east, while the São Francisco River drains it to the north (Fig. 3). Erosion-resistant quartzite ridges from the Espinhaço Supergroup (Figs. 2 and 3) form the N-S oriented watershed that reaches up to 1400 m altitude. Diffuse and thin (meters thick) Cenozoic deposits are present in the main river channels and laterized plateaus, mainly in the Conquista Plateau (Fig. 3). This plateau is also relatively elevated to around 1000 m and partially overlaps the Rio Pardo salient structures, where it delimits the the Contas River Basin to the south.

Figure 3
figure 3

Topographic and bathymetric map of the study area and adjacent passive margin. Source of the digital elevation model was USGS (https://earthexplorer.usgs.gov). Source of hydrography data was ANA (https://dadosabertos.ana.gov.br/). The map was created using, ArcGIS 10.4.1 (https://www.esri.com) and Corel Draw Graphics Suite 2018 (http://www.coreldraw.com).

Full size image

Apatite fission-track thermochronology: method and results

Thirty-two basement samples (gneiss, granites and migmatites) were collected from the basement of the AB–SFC boundary (Fig. 2) to perform low-temperature thermochronology with the AFT method. In order to explore the effects of northwards decreasing deformation, we collected closely interspaced samples in an N–S transect, crossing the limit orogen/craton. Another transect was sampled in the Rio Pardo salient structure. Sample lithology, locations, and elevations are found in Table 1.

Table 1 Summary of AFT analyses with sample location (in World Geodetic System -WGS84), elevation (Z) and lithology.
Full size table

The AFT method is a low-temperature dating technique based on the accumulation of mineral lattice damage, i.e. ‘fission tracks’, generated by the spontaneous fission of 238U11. These tracks are preserved in the apatite lattice on geological time scales at temperatures lower than c.120 °C, i.e. upper crustal temperatures12. AFT ages are hence cooling ages registering the time since the fission tracks became thermally stable in the apatite crystals. After etching to reveal the natural or spontaneous tracks by optical microscopy, the observed track density per unit area is a measure for the AFT age11. At temperatures between c.120–60 °C, fission tracks are able to accumulate in the apatite lattice but are subject to track length shortening due to thermal annealing or lattice restoration. Hence, a track length distribution of a sample is an indicator of the thermal history experienced by the apatite and its host rock12,13.

Apatite grains were concentrated using standard procedures, hand-picked and embedded in epoxy resin14. Mounts with c.120 apatite grains were etched for 20 s in 5.5 M HNO3 solution at 21 °C to reveal spontaneous fission tracks15. In this study we applied the external detector (ED) approach using thermal neutron irradiation16. U-free mica (Goodfellow, clear ruby) was attached as ED on top of each sample and age standard (Durango and Fish Canyon Tuff) mount17. IRMM-540 dosimeter glasses were used for monitoring the thermal neutron fluence. The packages were irradiated at the Belgian Nuclear Research Centre (SCK, Mol) using the Belgian Reactor 1 (BR1) facility (Channel X26;18). After irradiation, the ED was etched using 40% HF for 40 min at 21 °C in order to reveal induced fission tracks.

For each sample, 20 or more apatite grains were analysed. Fission track density was measured using a motorized Nikon Eclipse Ni-E microscope with a DS-Ri2 camera attached, at a 1000 × magnification. Central age calculation was performed using “IsoplotR”19 with an overall weighted mean zeta value of 330.6 ± 3.9 a·cm2 (Analyst AF) based on multiple Durango and Fish Canyon Tuff apatite age standards and the IRMM-540 dosimeter glass20. All samples pass the χ2 test, indicating single age populations. In 23 samples, where it was possible to achieve a representative AFT length–frequency histogram (n > 50), Markov Chain Monte Carlo (MCMC) inverse modeling was performed using the QTQt software10. Except for the present-day temperature (25 ± 15 °C), no time–temperature constraints were added. The Ketcham annealing model21, with Dpar as the kinetic parameter was used. The data is summarized in Table 1.

Despite the long-wavelength topography between 515 and 816 m elevation, and the low difference of erodibility of the sampled rocks (gneiss/migmatites), the AFT central ages vary substantially between 102 ± 14 Ma and 392 ± 21 Ma (Fig. 2), showing no correlation with Dpar (see Supplementary S.7 online). The mean track length (MTL) values are short to intermediate (9.9 to 12.8 µm) with mainly unimodal distributions (see Supplementary S.1 online), indicating long residence in the apatite partial annealing zone (APAZ, c.120–60 °C). Track lengths and their angle with the c-axis show no evident correlation (see Supplementary S.8 and S.9 online). Neither is this the case for the MTL vs. AFT central age plot (Fig. 4). From the thermal history modelling, three main t–T trends can be distinguished (Fig. 2):

  1. 1.

    continuous, undisturbed, and slow cooling from the Paleozoic onwards; identified in the northernmost samples (TJ 21–26);

  2. 2.

    Paleozoic to Mesozoic fast to moderate cooling with subsequent, late c.50 °C cooling during the Eocene–present, as evidenced by samples from the Rio Pardo salient zone (TJ 09–16 and TJ 28–34); and

  3. 3.

    continuous and slow cooling from the Paleozoic to the Cenozoic, followed by c.35 °C of rapid cooling during the Eocene–present, as observed for the southernmost samples (TJ02–09) and samples geographically between the Rio Pardo salient and sample site TJ 21 (TJ17–20).

Figure 4
figure 4

(a) Track length vs. angle with c-axis from confined tracks of samples TJ03, TJ12, and TJ21 (each representing one of the thermal history model groups: (i–iii). (b) MTL (μm) vs. AFT age (Ma) or “boomerang” plot12.

Full size image

Discussion

The AFT dataset distinctively reveals differential basement cooling and hence exhumation patterns. Samples from the northernmost area, close to Riacho de Santana city (TJ 21–26; models i in Fig. 2), do not show any important cooling during the Meso-Cenozoic. According to the models, slow and protracted cooling through the APAZ brought these samples (TJ 21–26) close to the surface (< 60 °C) even before the Jurassic, indicating that Cretaceous Atlantic rift and post-rift thermal events are lacking or exerted limited influence in this area. These samples (TJ 21–26) are from a previously identified cratonic segment of the Paramirim Aulacogen22, confirming no or very low degrees of tectonic rejuvenation of this lithospheric segment. The results are also in accordance with previous data from the southwestern SFC (Fig. 1), where AFT ages are older than the Jurassic4, and data from cratonic areas close to the Atlantic margin, where several samples are also older and do not seem affected by Cretaceous rifting or posterior events23 (Fig. 1). Areas with no clear signal in the AFT system from the rifting phase have indeed been associated to cratonic rheology (e.g.4,24,25). The rifting architecture varies greatly depending on the crustal thickness, the structural fabric and directions, the lithospheric rheology, and potential magmatic activity (e.g.26,27,28). With respect to our study region, we can conclude that the cratonic region of the Paramirim Aulacogen inherits the rigidity from the SFC. In this case, epirogenic uplift during the opening of the South Atlantic was mitigated as well as the erosional response to this process, resulting in an almost stable thermal structure.

Contrary to the craton, our data in the AB indicate relatively fast exhumation during the Eocene–present (models ii and iii in Fig. 2), with AFT ages not older than the Jurassic. Hence, confirming that our thermochronology data present a good correlation between the delimitation of the SFC and the AB. During the Cenozoic, the study area was embedded in the relatively stable South America platform, far from plate boundaries. The observed basement cooling must most probably be the consequence of erosional exhumation as indicated by the sediment supply to the adjacent basins. In the Eocene, the Jequitinhonha, Almada and Camamu passive margin basins, which are connected by river systems to our study area (Fig. 3), started their regressive phase. Progradation wedges, made up of coarse-grained sandstones, platform carbonates and distal mudstones, accumulated until the recent and reach 500 m in thickness29,30,31,32. Onshore, forming the typical regional tableland geomorphology, siliciclastic continental-to-shallow–water marine sediments of the Barreiras Formation were deposited from the Upper Oligocene onwards33. Thereby, we suggest that during the Eocene to present, our study area was heterogeneously eroded, likely partially contributing as a source of sediments to the above-mentioned deposits.

Previously published thermochronological data also indicates that the AB was indeed deeply exhumed during the Cenozoic (e.g.4,14,23) and, thus, our data further underscores the tracing of the belt (Fig. 2). Additionally, the Rio Pardo Salient seems one of the main structures in concentrating deformation since samples from this domain (TJ 11–16 and TJ 30–34) exhibit c.50 °C of rapid cooling during the Eocene to recent (models ii in Fig. 2). This salient presents a relatively high topography (c. 1000 m) in the Conquista Plateau region (Fig. 3), which may be a product of the reactivations from the Eocene. The formation of the Rio Pardo Salient resulted from the closure of the depocenter of the Neoproterozoic Macaúbas Basin, precursor to the AWCO collision34. The structure prior to the salient deformation was probably a major fault of the Macaúbas basin, deeply rooted through the crust. Analogue modelling has demonstrated that master faults are, indeed, more likely to concentrate stress and localize vertical displacements under compressional stress fields35. It is also important to note that the Rio Pardo Salient area (where the TJ 30–34 samples are located) limits the outcrops of the Macaúbas Basin fill (Fig. 2), supporting the location of the master fault.

Other authors also identified Neogene denudation pulses in northeastern Brazil using low-temperature thermochronology36,37, indicating that the event was widespread. Although climate has already been proposed as a main driving force for intensifying erosional denudation and hence basement cooling37, our results indicate that the denudation was far more heterogeneous. Evidence of brittle tectonics, related with ongoing ENE–WSW-oriented compression38, has been identified within the Cenozoic deposits (e.g.39) and challenged the concept of relative tectonic inactivity of passive margins. In line with the ideas of36, far-field partitioning of contemporaneous intraplate stress from the Andean collision zone seems to be the most plausible driving mechanism for the compressional tectonic reactivation of the intracontinental region, mainly during the Incaic and Quechua phases (Fig. 4). This final rapid cooling event most probably (partially) erased evidence from previous thermal events, including signals from West Gondwana break-up around 130 Ma. This latter event can however still be observed in three of our thermal history models (TJ 14–16).

Although in some areas it is easy to distinguish craton vs. orogen behavior, the transition between them is not always clearly traceable. In general, stress and resulting deformation decrease from the collisional zone to the plate’s interior. In our data, we could identify an intermediate t–T path (models iii in Sect. 3; Fig. 2) with c.35 °C of Eocene–present cooling (from 60 °C to surface temperature). In this way, it is possible to use the decrease of recent basement cooling as documented by AFT data, to trace the transition from orogenic to cratonic lithosphere. In the main N–S transect (Fig. 2), this intermediate zone is about 40 km in S–N direction and was probably affected by AWCO orogeny but with less penetrative deformation. In the area where samples TJ 17–20 were collected, close to Guanambi city, a set of dextral reverse shear zones40 seems to be related to this intermediate zone and then part of the orogen area. Based on our observations, we propose an adjustment of the AB boundaries within the Paramirim Aulacogen (Fig. 2). We also suggest the compartmentalization of the orogen (Fig. 5) in weak zones (e.g. Rio Pardo Salient) and the area of decreasing rigidity (i.e. transitional zone) until finally the craton s.s. is reached. Remarkably, the heterogeneous basement cooling patterns in our study area, likely resulting from the same stress regime, reinforces the concept that some lithospheric segments are more easily deformed than others.

Figure 5
figure 5

(a) AFT age distribution map of the study area identifying zones of tectonic reactivation with basement exhumation and stability. (b) Representative models for each zone and indication of the phases of Andean orogeny and West Gondwana break-up. The map was created using, ArcGIS 10.4.1 (https://www.esri.com) and Corel Draw Graphics Suite 2018 (http://www.coreldraw.com).

Full size image

Conclusions

AFT data from the São Francisco craton and adjoining Araçuaí belt in the Paramirim Aulacogen area (eastern Brazil) elucidate the differential behavior of the cratonic and non-cratonic lithosphere during the Phanerozoic exhumation of this region. To the north of our study area, thermal history modelling of the basement rocks exhibits slow and protracted cooling during the Phanerozoic, consistent with the rigid cratonic lithosphere of the São Francisco Craton. Samples from the Araçuaí belt, in the interior of the Paramirim aulacogen, display reactivation during the Cenozoic, mainly between the Eocene to present, reflecting its weakened lithosphere, inherited from the Ediacaran–Cambrian collision. An intermediate zone is identified, and it is considered mostly part of the Araçuaí Belt but with less penetrative deformation as to the orogen proper. The thermochronological data proved to be highly useful in determining the decreasing magnitude of reactivation along the craton—orogen boundary and can be used as a tool to trace and distinguish cratonic areas weakened by later deformation events.