Crustal growth in the northeast portion of the Rhyacian Bacajá domain, SE Amazonian craton, based on U-Pb, Lu-Hf, and Sm-Nd data

Abstract The Bacajá domain, southeastern Amazonian craton, comprises Mesoarchaean and Siderian terrains reworked during the Transamazonian cycle. Combined analyses of zircon LA-ICP-MS U-Pb and Lu-Hf with whole-rock Sm-Nd from the northeast portion of this domain made it possible to propose an evolutionary sequence between ca. 2.60 and 2.06 Ga. Gneisses with an igneous protolith age of 2630±15 Ma show negative signatures (ɛHft = -0.3 to -1.7 and ɛNdt = -3.08 to -2.98) and a Mesoarchaean formation (Hf-TDMC and Nd-TDM model ages range from 3.0 to 3.2 Ga). Rhyacian granite genesis lasted about 40 million years (2.10–2.06 Ga) and was divided into two magmatic periods. The first is represented by deformed granitoids with zircons yielding crystallization ages between 2.10 and 2.09 Ga and model ages (Hf-Nd) at about 2.5 Ga. The second event is represented by granitoids with preserved magmatic texture, crystallization ages of 2.06 Ga, and Siderian model ages (Hf-Nd) of around 2.3 Ga. The overall Hf isotopic analyses of this Rhyacian granite genesis exhibit a spread of ɛHft values between 1.8 and -2.9, which show a probably underestimated mantle-derived contribution in this period.


INTRODUCTION
The northern portion of the Amazonian craton comprises a Paleoproterozoic mobile belt reworked during the Transamazonian cycle (Tassinari and Macambira 2004).The southeastern part of this mobile belt, known as the Bacajá domain (BJD) (Fig. 1), is considered a key area for understanding crustal growth of this region.Such domain contains Siderian registers, scarce through all the Amazonian craton; it borders the Carajás block that is not affected by the Transamazonian cycle and shows several magmatic and deformational episodes that are still not very well characterized.
The BJD comprises a large volume of deformed granitoids and, to a lesser extent, granulites, gneisses, and metavolcano-sedimentary sequences.Isotopic data were mainly obtained from whole-rock Sm-Nd and Pb-evaporation on zircon (e.g.Barros et al. 2007, Vasquez et al. 2008a, Macambira et al. 2009, Macambira and Ricci 2013).These isotopic data, structural trends, lithostratigraphy, and geophysical features have made possible to reveal that the BJD is a Mesoarchaean to Siderian (3.00-2.30Ga) terrain reworked during the Rhyacian orogenies and linked to the Transamazonian cycle (2.26-2.05Ga) (Vasquez et al. 2008a, Macambira et al. 2009).This cycle was an important rock-forming event in the South American Platform, and it is detailed in Rhyacian stages on the BJD.
At least three distinct stages are inferred for the Rhyacian orogenies in the BJD.Vasquez et al. (2008b) and Macambira et al. (2009) have revealed that the oldest stage of ca.2.2 Ga comprises granitoids with ɛNd (t) significantly negative and Nd-T DM predominantly Mesoarchaean.They were formed in continental arcs at the margin of an Archaean continent.At about 2.1 Ga, granitogenesis appears as the main event, correlated with other regions, as in the Guiana Shield (e.g.Vanderhaeghe et al. 1998).These granitoids were emplaced in a continental margin linked to the collision climax of this cycle (Vasquez et al. 2008a).The final stage occurred between 2.09 and 2.06 Ga, and it is represented by granitoids and charnockites, some with preserved igneous textures (Barros et al. 2007, Macambira andRicci 2013).ɛNd (t) spread from positive to negative values, and model ages (Nd) vary from Neoarchaean to Siderian.This final magmatic event is interpreted as the product of mixing processes (juvenile and crustal components).
In this study, we present the first data of integrated U-Pb, Lu-Hf, and Sm-Nd isotope data obtained on zircon and whole rock of Archaean gneisses, Rhyacian granitoids, and granulites from the northeastern portion of the BJD.The purpose of the study was to elucidate crustal growth from this domain.
Crustal growth in the northeast portion of the Rhyacian Bacajá domain, SE Amazonian craton, based on U-Pb, Lu-Hf, and Sm-Nd data It involves a discussion of the Archaean records, features from the Rhyacian granitogenesis, related with the Transamazonian cycle, and the event of high-grade metamorphism that affected that region.
The Aruanã Complex occurs in the northern portion of the BJD, which is crossed by important sets of NE-SW strikeslip faults and NW-SE foliations, identified through aero-geophysical surveys carried out by the Geological Survey of Brazil (CPRM) (Fig. 2).This unit mainly consists of orthogranulites with subordinate mobilized granitic material, and granitic pegmatites (Macambira and Ricci 2013).Barros and Besser (2015) identified banded metagranitoids and orthogneisses in the amphibolite facies.Although the mineral assemblage of the rocks from this complex would appear to indicate the amphibolite facies, the granoblastic texture suggests a higher degree of metamorphism (granulitic facies), as proposed by Macambira and Ricci (2013).These authors also presented geochemical data that point to an enrichment in large-ion lithophile elements (LILE) when compared with high-field strength element (HFSE) contents, indicating calcic-alkaline magmatism generated in a continental margin.Besser (2012) also obtained calcic-alkaline affinity and proposed a volcanic arc or sin-collisional environment for these rocks.Vasquez et al. (2008c) obtained an age of 2606 ± 4 Ma, interpreted as the igneous protolith crystallization age (Pb-evaporation in zircon method).Near Pacajá town, mobilized granitic materials were dated by Macambira and Ricci (2013) by the same method.Zircon crystals show a low Th/U ratio and yielded an age of 2.1 Ga, interpreted as related to a possible metamorphic event.
To the west, Barros and Besser (2015) analyzed a metagranitoid that yielded an age of 2.6 Ga by zircon Pb-evaporation.
The Bacajaí Complex is composed of enderbites, charno-enderbites, and granitoids crossed by NE-SW strike-slip faults and NW-SE foliations.Mesoperthitic intergrowth and well-developed allanite crystals suggest that they are deeply emplaced (Ricci 2006).Geochemical data point to an intermediate composition from monzogranite to granodiorite calcic-alkaline of medium-to-high K, sodic, metaluminous affinity, and enrichment in LILE and light rare-earth elements (LREE).They are interpreted as orogenic bodies emplaced at the climax of the continental collision of the Transamazonian cycle (Macambira and Ricci 2013).In the central portion of the BJD, charnockites analyzed by U-Pb SHRIMP and Pb-evaporation methods yielded zircon ages between 2114 and 2094 Ma (Faraco et al. 2005, Monteiro 2006).Macambira and Ricci (2013) obtained a mean age of 2090 ± 6 Ma (Pb-evaporation in zircon), interpreted as being the crystallization age of this suite.
The Arapari Intrusive Suite is composed of charnockites, charno-enderbites, and granitoids with NW-SE foliations (Fig. 2).Macambira and Ricci (2013) described a texture variation from porphyritic to equigranular, banded to preserved rocks, deeply emplaced in similar conditions as the Bacajaí Complex.Those authors confirmed a crustal contribution on an active continental margin observed by a calcic-alkaline magmatism enriched in LILE and LREE.Pb-evaporation in zircon yielded crystallization ages of 2088 ± 2 Ma, 2069 ± 2 Ma, and 2059 ± 4 Ma.Macambira and Ricci (2013) suggested an interpretation that this unit was formed by several magmatic pulses.To the west, Pb-evaporation and U-Pb ages in zircon show values between 2086 and 2070 Ma (Santos 2003, Vasquez et al. 2008a, Macambira et al. 2009), Nd-T DM ages of about 2.47 Ga and εNd (t) between -2.40 and -3.12 (Vasquez 2006).
The João Jorge intrusive suite is the youngest lithostratigraphic unit.It mainly consists of monzogranites and syenogranites that occur as NW-SE batholiths and stocks.Geochemical data show enrichment in LILE/LREE and depletion in HFSE and high rareearth elements.They have affinity with post-collisional granites and preserved signature of volcanic arcs (Macambira and Ricci 2013).Pb-evaporation on the zircon method yielded ages of ca.

Zircon preparation and SEM images
After fieldwork (Macambira and Ricci 2013), zircon crystals were first concentrated by traditional conventional methods (crushing/grinding, granulometric, magnetic, and water/ alcohol concentrations).Then, scanning electron microscopy (Zeiss SEM LS15 Scanning Electron Microscope) was performed at the laboratory of the CPRM, Belém (LAMIN-BE), using high vacuum (3.0-1.5 x 10 -5 mPa) and tungsten filament mode with an EVO15RHS CL detector set at 13-14 kV, 10 Na input current.Cathodoluminescence (CL) images of zircon crystals were obtained using a LEO-ZEISS 1430 scanning electron microscope in the Microanalysis Laboratory (LabMev) at Universidade Federal do Pará (UFPA), Belém.All zircons are within the fraction 175-125 μm, and the images were used for choosing the most suitable sites for analysis, avoiding inclusions, and keeping away from metamict and highly fractured sites.

4/18
Zircon U-Pb dating LA-ICP-MS U-Pb isotopic analyses were carried out using a multi-collector Neptune Thermo Finnigan mass spectrometer coupled with a Nd:YAG LSX-213 G2 CETAC laser microprobe in the Isotopic Geology Laboratory (Pará-Iso) at UFPA.The analytical sequence alternates sample collection with GJ-1 reference zircon (608 ± 1.5 Ma; Jackson et al. 2004) for fractionation and correction of mass discrimination.The Plesovice (337 Ma; Slama et al. 2008)
The raw data file from the mass spectrometer was retrieved into an Excel spreadsheet for processing in order to calculate the corrected isotopic ratios for each point (blank analysis was also taken into account).The standard procedure cycle consists of the analysis of blank, GJ-1, MudTank, sample (at least 10 zircon crystals), and finishes with the same initial sequence that begins a new cycle for another sample.Calculations of ɛHf used a decay constant of 1.867 × 10 -11 years -1 (Scherer et al. 2001, Soderlund et al. 2004), present-day ratio of 176 Lu/ 177 Hf of 0.0336, and a ratio of 176 Hf/ 177 Hf of 0.282785 for the chondritic uniform reservoir (CHUR) (Bouvier et al. 2008).The 176 Lu/ 177 Hf of 0.0388 and 176 Hf/ 177 Hf of 0.28325 were used for depleted mantle (DM) (Andersen et al. 2009), and for crustal model ages (T DM C ), a 176 Lu/ 177 Hf of 0.015 was assumed as a continental crust average value (Griffin et al. 2004).As for the U-Pb analysis, instrumental parameters, spectrometer, and laser settings are all also in accordance with the parameters set by Milhomem Neto and Lafon (2019), who established a routine in this laboratory for those methods.

Sm-Nd whole rock
For the Sm-Nd analyses, approximately 100 mg of rock powder were dissolved with a mixture of HF and HNO 3 and a mixed 150 Nd-149 Sm spike.REE were separated by cation exchange chromatography (Dowex 50WX-8 resin) using HCl and HNO 3 .Then, Sm and Nd were separated from the other REE by anion exchange chromatography (Dowex AG1-X4 resin) using a mixture of HNO 3 and methanol.Isotopic ratios were measured on a Finnigan MAT 262 thermo-ionization mass spectrometer.Nd is normalized to a 146 Nd/ 144 Nd ratio of 0.7219.Total procedural blanks during the analysis were lower than 0.1%, and values obtained for the La Jolla and BCR-1 reference materials were in agreement with the literature.The decay constant used was 6.54 × 10 -12 years -1 and the chondritic values used to calculate ɛNd were 143 Nd/ 144 Nd = 0.512638 and 147 Sm/ 144 Nd = 0.1967.The crustal residence ages were calculated using DePaolo's (1988) model for DM.

U-Pb and Lu-Hf in zircon
Zircon crystals from six samples from the northeastern region of the BJD were dated.Among the results, listed in Table 1 and represented in the concordia diagrams, the red ellipses were not used in the elaboration of the discordia line to obtain the age of the upper intercept.Discordant points in the concordia diagram or, less frequently, those that diverged from the average age were discarded.Coherent concordant U-Pb ages or, in most cases, the age of the upper intercept of each sample were used to calculate the Lu-Hf parameters.Spots of 50 μm for the analyses of Lu-Hf were targeted at or near the U-Pb analytical point.
The zircon crystals are brown in color and elongated, prismatic, and euhedral to subhedral in shape.CL images show fading patchy zoning (Fig. 4D1) and luminescent apatite inclusions (Fig. 4D3).The U-Pb analyses were performed on 20 crystals, 6 of which were used to obtain a discordia, whose upper intercept yielded the crystallization age of the igneous protolith of 2630 ± 15 Ma (MSWD = 0.44, Fig. 5A).Nine concordant to subconcordant zircon crystals yielded negative values from -0.3 to -1.7 of ɛHf (2.63 Ga) and Hf-T DM C model age range between 3. 2 and  3.1 Ga (Table 2).
Zircon crystals are mainly subhedral.CL images show convoluted zoning (Fig. 6A4), weak or no internal zoning (Fig. 6C9), and rims of the grains with greater luminescence that surround all crystals (Fig. 6C9).A total of 23 crystals were analyzed using the U-Pb method, and it was possible to individualize two different concordant ages.The lower age was 2086 ± 4 Ma (MSDW < 1, n = 4), while the higher age was 2120 ± 4 Ma (MSDW < 1, n = 6), without overlapping (taking the errors into account) (Fig. 5B).These two different ages obtained in the U-Pb analyses are supported by the Lu-Hf results.Seven U-Pb concordant and three subconcordant crystals from this sample were analyzed by this method.The younger group yielded exclusively negative values for ɛHf (t = 2.09 Ga) , ranging between -1.2 and -3.2, whereas the older group yielded mainly positive values, and some slightly negative values for ɛHf (t = 2.12 Ga) , ranging from 2.5 to -0.8.The Hf-T DM C model ages are also different, 2.9-2.7 Ga and 2.7-2.5 Ga for the youngest and oldest groups, respectively (Table 2).
Zircon crystals range from euhedral to subhedral, and elongated to prismatic in shape.There are many features of zoning loss (Figs.7B2, 7B3, and 7C2) and metamictization (Figs.7A1, 7A4, and 7C4).There is a distinct zone where luminescence is more intense at or around the rims.From the analyses of 20 zircon crystals, an age of 2103 ± 21 Ma (MSWD = 1.3, Fig. 5C) at the upper intercept was obtained.Ten zircon crystals were analyzed for Lu-Hf: two U-Pb concordant (99%), four subconcordant (98-95%), and five (< 95%) discordant.The values of ɛHf (t = 2.1Ga) are predominantly negative, ranging from 1.5 to -2.2, while the ages of model Hf-T DM C vary between 2.8 and 2.6 Ga (Table 2).
The zircon crystals are brown in color and prismatic, and euhedral to subhedral in shape.CL images show conspicuous oscillatory zoning (Fig. 8), and some crystals exhibit rims with a strong luminescence and more shaped (darker) areas.On the upper intercept, the most concordant crystals yielded a  crystallization age of 2094 ± 11 Ma (MSWD = 0.95, Fig. 5D).The ɛHf (t = 2.09 Ga) from 10 zircon crystals showed scattered results, ranging from positive to negative values of 1.7 to -1.3.The model Hf-T DM C age varied between 2.7 and 2.6 Ga (Table 2).

Syenogranite PR-165 (Arapari Intrusive Suite)
Sample PR-165 shows a porphyroclastic texture.Mortar texture, perthitic, antiperthitic, and myrmekite intergrowth are also quite common.Megacrystals (> 3 cm) are mainly of poikilitic microcline (60%) and subordinate plagioclase (10%).Quartz (25%) crystals are elongated and show a ribbon texture; locally, they are recrystallized and present a granoblastic texture.Igneous textures are partially preserved, such as the subhedral megacrystals (feldspar), which have been very little recrystallized or not at all.In addition, the typical Carlsbad twinning is well preserved and even observable with the naked eye.Mylonitization is more pronounced, and a protomylonitic foliation is identified, mainly by oriented biotite (5%).The matrix is comminuted but still keeps its preserved and rotated porphyroclasts (Fig. 3E).Accessory minerals are hornblende and zircon.
Zircon crystals from sample PR-165 are brown in color and subhedral in shape.CL images show well-developed oscillatory zoning, many inclusions, and low luminescence (Fig. 9).Most crystals from this sample show high values of common Pb, which were not considered in the calculation.However, an upper intercept was obtained with an age of 2080 ± 16 Ma (MSWD = 0.93, Fig. 5E) from eight analytical points (ƒ 206 < 0.02).Nine zircon crystals from this sample were subjected to analysis by the Lu-Hf method.The values of ɛHf (t = 2.08 Ga) vary between 1.1 and -2.9, and the Hf-T DM C model age varies between 2.8 and 2.6 Ga (Table 2).
Zircon populations are brown in color, euhedral to subhedral, and prismatic in shape.Oscillatory zoning is common, and some grains have different features, such as areas with high luminescence (Fig. 10D9).Zircon analyses form a scattered array on the concordia diagram.However, based on 11 more consistent results, it was possible to obtain a crystallization age at the upper intercept of 2062 ± 22 Ma (MSWD = 1.1, Fig. 5F).Nine zircon crystals produced predominantly positive values of ɛHf (t = 2.06Ga) , varying between 1.8 and -0.3, and the Hf-T DM C model age ranges from 2.7 to 2.5 Ga.
The εNd (t) values obtained are all negative (-0.88 to -9.68), and the ranges of the Nd-T DM model ages vary between 2.35 and 3.05 Ga.From these data, it was possible to distinguish three groups.The first one is represented by Neoarchaean gneisses (A), followed by strongly deformed charnockitic rocks and Rhyacian granitoids (B), and the last group is represented by weakly deformed granitoids (C).In summary, considering the characterized rock, age of crystallization, lithostratigraphic unit, and Nd isotopic data, the samples of the study area were divided into the following three groups:

Crustal growth and Hf versus Nd tracers
Unraveling crustal evolution relies on isotope studies from whole-rock samples (more conventional, with widespread data) and single-grain zircon (which presents recent advances in analytical capabilities and a growing data acquisition on the Amazonian craton).In this study, samples were analyzed with both methods, for comparison and to elucidate the growth and differentiation of the continental crust.
The first and more obvious difference from both methods is that whole-rock samples show an overall dominant negative value (ɛNd (t) ) (Fig. 11) for rocks through all timespans analyzed.Single-grain zircon samples, however, comprise a more distinct 12/18 pattern.Archaean registers are dominantly negative (ɛHf (t) ), and Rhyacian registers show a group of disperse values varying from negative to positive and another group (younger) with dominantly positive values (Fig. 12).Although the Lu-Hf single-grained zircon method has limitations, some authors (e.g., Scherer et al. 2007) concluded that they are more resistant to disturbance than Sm-Nd in whole rock and therefore have the potential to decipher the crustal growth history of highly metamorphosed terranes such as the BJD.From this comparison, Hf tracer on zircon enables deconvolution of intragrain isotope heterogeneity that might reflect complex growth histories (i.e., metamorphic overgrowths and inherited cores; unfortunately, in this work it was not possible to make such a distinction) or changes in the isotope composition of the melt from which the zircon precipitated in response to magma mixing or crustal assimilation in a more precise resolution than whole rock (this is corroborated with this work).Therefore, the crustal evolution for this region relies upon the zircon, yet the comparison with the whole rock is an important discussion for a better understanding of prior interpretations based on such data.

Archaean registers
Orthogneisses are strongly banded and show augen structure.NE-SW faults and foliations are identified by a new aero-geophysical project (CPRM 2016) that were formed prior to the Transamazonian cycle (Rhyacian magmatism covered this Archaean basement without continuity of such features) (Fig. 2).They yielded an igneous protolith age of 2.63 Ga (EM-161A) and a signature of a reworked crust (Hf-tracer).Those orthogneisses are correlated to the charnoenderbitic gneiss dated at around 2.60 Ga (Table 1) included in the Aruanã Complex (Vasquez et al. 2008c).Geochemical data of this unit point to a calcic-alkaline magmatism signature, suggesting a generation of volcanic arc in a continental margin (Macambira and Ricci 2013).Such data are, however, contrasting with the juvenile gneisses of the Manelão area (Macambira et al. 2009).Thus, two groups are possible to distinguish in the Neoarchaean: the first of 2.67 Ga is linked to island arc with the addition of juvenile crust, followed by a second of 2.63-2.60Ga continental margin magmatism with crustal components.

Rhyacian magmatism and metamorphism
Rhyacian magmatism lasted ca.40 Ma (2.10-2.06Ga) and is divided into two magmatic periods.The first is represented by deformed granitoids included in the Bacajaí Complex (samples EM-100 and PR-170) that have crystallization ages between 2.1 and 2.09 Ga and show a significant spread in the Figure 12.Evolution diagram, εHf (t) versus age (Ma) for the BJD studied units.The dotted lines represent crustal evolution trends, calculated using a 176 Lu/ 177 Hf of 0.015 for the average continental crust (Griffin et al. 2002(Griffin et al. , 2004)).This diagram highlights the two distinct groups formed by the zircons from the Orthogranulite PR-143.

13/18
values of ɛHf (t) from negative to positive (some zircon crystals with ɛHf (t) = 0, Fig. 12).The fact that some zircon crystals have a precise chondritic (ɛHf (t) = 0) or slightly positive Hf composition at the time of crystallization is a solid evidence that they derive from mixed juvenile and crustal sources.Most granites have isotope signatures that preclude a direct mantle ancestry (for basic principles of partial melt), yet wholerock samples show an overall negative signature for them as for the Rhyacian magmatism in this region.Thus, this points to a possible underestimated recognition of a new, mantle-derived material source (although mixed) in this magmatism.
At this same period, the orthogranulite (PR-143) yielded two concordant ages: 2.12 Ga and 2.09 Ga.Zircon crystals of these samples have irregular, chaotic zoning (Figs.6D8  and 6D3) and zoning loss (Figs.6A4 and 6C9), that according to Rubatto (2017), most of these features indicate metamorphic conditions of granulitic or anatexis facies.They also show a Th/U ratio variation of 0.32-0.91(Table 4), out of some Th/U ratios range for metamorphic zircon < 0.1, although exceptions do exist as Vavra et al. (1996) attested.Most occurrences of metamorphic zircon with Th/U > 0.1 are from samples of high or ultra-high temperatures (> 900°C) (Vavra et al. 1996, Schaltegger et al. 1999, Moller et al. 2003, Kelly and Harley 2005), and this possibly indicates a highgrade temperature for the PR-143 sample.Orthopyroxene assemblages and the granoblastic texture also support a high metamorphic grade for this sample (Fig. 3).Otherwise, metamorphic zircon forms more easily in high-grade metamorphic rocks, where they commonly consist of overgrowths on inherited or detritic magmatic nuclei (Pidgeon et al. 2000).However, CL images obtained from the studied zircons do not distinguish metamorphic or igneous domains (Fig. 6), only a bright narrow rim not possible to analyze (< 25 μm).
Although one can see that both ages of 2120 ± 4 Ma (n = 6) and 2086 ± 4 Ma (n = 4) were obtained from mainly concordant analytical points, many zircons are discordant (Fig. 5).Such discordance is possibly related to losses of lead due to metamorphic events and/or metamictization process.Moreover, the two distinct Hf signatures observed in this sample corroborate the existence of two separated ages.The highest age exhibits ɛHf (2.12Ga) = 2.5 to -0.8; Hf-T DM C = 2.7 to 2.5 Ga, and the lowest age shows ɛHf (2.09Ga) = -1.2 to -3.2; Hf-T DM C = 2.9 to 2.7 Ga.Taking these data into account, an alternative more plausible interpretation is that two-zircon crystallization moments are related to Zr-saturation pulses in the magmatic chamber.At an earlier stage, zircon from a first magma pulse was crystallized (at 2.1 Ga with a juvenile signature) and was then progressively contaminated or mixed with crustal components in a second pulse (at 2.09 Ga with a negative signature).Because only a single sample from orthogranulite with such features was studied, further data are needed for confirmation.Thus, this could be an example of a sample that could distinguish with more precision the two different sources (the juvenile and the mixed one).
The last magmatism period described is represented by granitoids formed at 2.06 Ga that show preserved magmatic texture.These granitoids are included in the João Jorge intrusive suites (EM-55) and yielded ages of 2062 ± 22 Ma, slightly negative-to-positive Hf-Nd signatures, and model ages (Nd-Hf-T DM ) that are mainly Siderian (ca.2.3 Ga).Geochemical data presented by Macambira and Ricci (2013) for this unit is characteristic of an A-type (anorogenic) magmatism.Thus, those data suggest a less contaminated/mixed source for these rocks, probably related to the final stage of the Transamazonian cycle (the third stage, post-collisional), as already proposed by Vasquez et al. (2008b).

CONCLUSION
The northeastern portion of the BJD has a large Archaean register comprised in the Aruanã Complex.Orthogneisses with conspicuous banded structure were recognized, and zircon from them yielded ca.2.6 Ga.Hf-Nd data suggest a reworked crust with a Mesoarchaean source.These data show two distinct coeval Mesoarchaean magmatism occurrences in the BJD.One in the central portion of the BJD with a juvenile signature located in a WNW-ESE transcurrent shear zone, and the other in the northern portion with a reworked signature and NE-SW faults and foliations.
The Rhyacian granitogenesis is dominant in the BJD and lasted about 40 million years (approximately 2.10-2.06Ga).Deformed granitoids comprise metatonalites, monzogranites, and syenogranites with ca.2.1 Ga.The overall Hf isotopic results show a spread on the values of ɛHf (t) , between 1.8 and -2.9, which indicates a mainly mantle-derived magma that assimilated older crust.Hf-T DM C and Nd-T DM model ages for the rocks from this and prior studies indicate a main period of mantle extraction and crust formation in the Neoarchaean (between 2.8 and 2.5 Ga), but younger Siderian sources were identified (Nd-T DM of 2.45 Ga; EM-115).This stage on the Transamazonian cycle probably has an underestimated observed mantle-derived contribution for crustal growth, which becomes evident with modern single-grain zircon analysis.The same samples analyzed by whole-rock methods show an overall negative signature, as in many other previous studies.The orthogranulite zircons described with chaotic and irregular zoning exhibit two concordant ages and distinct Hf signatures that could represent distinct moments in the magmatic chamber, perhaps detailing the magmatic evolution.The highest age (2.1 Ga) is the crystallization, with a juvenile signature, and the lowest age (2.09 Ga) has contributions of crustal components.On the contrary, João Jorge suite, represented by preserved granodiorites formed at 2.06 Ga with an Hf-Nd model age of ca.2.3 Ga, probably represents a late stage (post-collision stage) within the Transamazonian cycle.

Figure 5 .
Figure 5. Concordia diagrams for zircon U-Pb results.Red ellipses were disregarded for the age calculation.7/18

Figure 6 .Figure 7 .
Figure 6.CL images of representative zircon grains of PR-143 orthogranulite.Circles indicate spot locations, with the small ones being U-Pb and the large ones being Lu-Hf.

Figure 8 .
Figure 8. CL images of representative zircon grains from PR-170 monzogranite from the Bacajaí Complex.Circles indicate spot locations, with the small ones being U-Pb and the large ones being Lu-Hf.

Figure 9 .
Figure 9. CL images of representative PR-165 syenogranite zircon grains from the Arapari Intrusive Suite.Circles indicate spot locations, the small ones being U-Pb and the large ones being Lu-Hf.

Figure 10 .
Figure 10.CL images of representative EM-55 granodiorite zircon grains from the João Jorge Intrusive Suite.Circles indicate spot locations, with the small ones being U-Pb, and the large ones being Lu-Hf.11/18

Figure 11 .
Figure 11.Diagram εNd (t) versus age (Ma) showing the evolution trends for the study samples.Previous data from Macambira et al. (2009).
standard zircon was also used as secondary reference material and obtained the weighted mean 206 Pb/ 238 U age of 340.4 ± 3.6 Ma (n = 11, MSWD = 0.063).Isotopic ratios ( 206 Pb/ 238 U, 207 Pb/ 235 U, and 207 Pb/ 206 Pb) and ages were calculated from the background-corrected values and 204 Hg interference over 204 Pb.For the correction of common lead contribution, the terrestrial Pb evolution model over time of Stacey and

Table 2 .
Lu-Hf isotopic data on zircon from the NE Bacajá domain.

Table 4 .
U-Pb isotopic data in zircon from the northeastern Bacajá domain.

Table 4 .
Continuation.fraction of the nonradiogenic 206 Pb in the analyzed zircon spot, where ƒ 206 = [ 206 Pb/ 204 Pb]c/[ 206 Pb/ 204 Pb]s (c = common; s = sample); c the Th/U ratio is calculated relative to GJ-1 reference zircon; d Rho is the error correlation defined as the quotient of propagated errors of 206 Pb/ 238 U and 207 Pb/ 235 U; Continue... a Identification of the zircon sample; b e corrected for mass-bias by normalizing to GJ1 reference zircon and common Pb using the model proposed by Stacey and Kramers (1975); f degree of concordance = ( 206 Pb/ 238 U*100 age)/( 207 Pb/ 206 Pb age).