U-Pb zircon ages of rocks from the Amazonas Territory of Colombia and their bearing on the tectonic history of the NW sector of the Amazonian Craton

Manuscript ID: 20150012. Received in: 06/29/2015. Approved in: 10/08/2015. ABSTRACT: Here we report the results of several U-Pb zircon ages, made to generate an integrated history for the Rio Negro-Juruena tectonic province, for the northwestern part of the Amazonian Craton. This region comprises granitoid rocks, described as calc-alkaline syntectonic gneisses, granites and migmatites, affected by medium level amphibolite facies metamorphism. The new measurements, with the available Rb-Sr and K-Ar ages, indicate the formation of these rocks within a series of essentially juvenile magmatic arcs, that are closely related with subduction. Sm-Nd analyses indicate that all samples, regardless of their zircon ages, yielded TDM model ages roughly between 1.9 and 2.2 Ga, suggesting the absence of a much older source material. In the northeastern part (areas of Puerto Inírida and San Carlos), the Atabapo belt comprises rocks formed within a period of about 60 Ma, from 1800 to 1740 Ma. In the southwestern region, including the towns of Mitú and Iauaretê, the granitoids formed in the Vaupés belt between 1580 and 1520 Ma. Finally, the available K-Ar measurements indicate the onset of the Nickerie-K’Mudku intraplate heating event, with temperature above 300oC within the entire region at 1200 – 1300 Ma.


INTRODUCTION
A first attempt towards a geochronological reconnaissance study of the Amazonian Craton was made in the late 1970's (Cordani et al. 1979), after comprehensive geological mapping through the RADAM program of the Brazilian government.In contrast from the previous fixistic tectonic models for that cratonic area, these authors adopted a mobilistic interpretation.Some proterozoic tectonic provinces were envisaged, growing successively around an ancient nucleus located in the central part of the craton.One of them, including the Amazonian region of eastern Colombia, SW Venezuela and NW Brazil, was named Rio Negro-Juruena tectonic province.
At the time of the initial work, only Rb-Sr and K-Ar ages were available.Cordani et al. (1979) reported ages between 1700 and 1500 Ma for the tectonic events within that province.A later increase of available rock ages in Amazonia permitted marked progress in understanding the region (see e.g, comprehensive reviews by Teixeira et al. (1989), Tassinari (1996), Tassinari and Macambira (1999) and Cordani et al. (2000).
With the increasing use of U-Pb zircon ages in recent years, some alternative interpretations for the tectonic evolution of the Amazonian Craton were presented (e.g.Santos et al. 2000, Santos 2003, Cordani and Teixeira 2007).Boundaries of the tectonic domains were altered, and names were changed, but the mobilistic frame was always maintained.Based on Nd isotopic work, Cordani and Teixeira (2007) suggested that the tectonic evolution of the SW half of the Amazonian Craton was accretionary.They proposed that the craton grew by the stacking of successive magmatic arcs originating from subduction zones, from 2000 to1500 Ma.Within the Rio Negro-Juruena province, unequivocal evidence of continental material older than 2000 Ma has not been found.
Figure 1 illustrates the interpretation given by Cordani et al. (2000), with the location of the study area.The first geochronological ages in that region were determined by Pinson et al. (1962), who dated the alkaline rocks of São José del Guaviare by the K-Ar method, back to an ordovician age (about 480 Ma).During the 1970's and the 1980's, the Geological Survey of Colombia, INGEOMINAS, made an extensive reconnaissance geological mapping of the country (PRORADAM 1979).Priem et al. (1982) performed a relevant geochronological study of that area, primarily using Rb-Sr and K-Ar methods.Around the same time, important reconnaissance works were carried out in Brazil and Venezuela by the respective geological surveys, to assess the potential of the region for mineral exploration.
Our original focus in this work was the poorly known region of the Amazonas Territory of Colombia, located in the NW part of the Rio Negro-Juruena Province of Cordani et al. (1979).We later enlarged the area of interest to involve large parts of SW Venezuela and NW Brazil.In these countries, a few important geochronological works by Fernandes et al. (1976), Pinheiro et al. (1976), Gaudette and Olszewski (1985), and Barrios et al. (1985 and1986) made several age determinations for the existent governmental mapping projects.Our consolidated study area (Fig. 2) includes the international boundaries of Colombia with Venezuela and Brazil.The terrain of the entire territory was completely shaped by erosion, and the present land surface is a widespread peneplain.Granitic and gneissic rocks, deformed or not, and frequently migmatitic, are largely predominant.In a general way, these regional rocks were formed by tectono-magmatic and metamorphic processes related to medium-grade metamorphic environments.Supracrustal rocks and intraplate volcanic-sedimentary sequences occur in restricted areas.
During the 1980's, Priem et al. (1982) and Gaudette and Olszewski (1985) produced a few U-Pb zircon analyses, using conventional TIMS studies.Important geochronological papers were later produced, including work by Tassinari et al. (1996), using SHRIMP, and by Sato and Tassinari (1997), using Sm-Nd model ages.Tassinari (1996) prepared a complete synthesis of all available geochronological ages in the Brazilian territory.In the last 10 to 20 years, additional U-Pb zircon work has been performed, such as studies by Santos et al. (2000), Santos (2003), Almeida (2006), and, more recently, Ibañez et al. (2011), who presented a few U-Pb zircon determinations by LA-ICP-MS.
The aim of the present work was to produce a tentative integrated picture of the tectonic evolution of the NW part of the Amazonian Craton.Following the analyses of about 20 samples of granitoid rocks from the INGEOMINAS collection, we report the results of geochronological and isotopic measurements using accurate U-Pb zircon ages, measured by the SHRIMP or LA-ICP-MS methods.Despite the reasonable amount of available geochronological data, this project remains a very preliminary reconnaissance work, because the covered basement area exceeds 200,000 km 2 and only around 30 precise U-Pb zircon ages exist.This work includes the results of several new Sm-Nd isotopic measurements and some new Rb-Sr analyses which were included in the available isochron diagrams constructed from works of Priem et al. (1982) and others.Our regional interpretations consider all available geochronological controls, by K-Ar and Rb-Sr methods.As more than 100 ages by Rb-Sr dating are available for the study area, we tried to select the most significant ages in terms of interpretative results.Figure 3 shows the location of 97 rock samples with Rb-Sr determinations.Their ages are described in a series of Rb-Sr isochron diagrams (Fig. 9) and will be properly discussed later.Overall, these data have helped us to obtain a suitable perspective for interpreting the tectonic history of the region.

ANALYTICAL METHODS
All K-Ar and most Rb-Sr dates considered in this work were obtained during the 1970's and 1980's from laboratories in Amsterdam (Priem et al. 1982) and São Paulo (Barrios et al. 1985, Fernandes et al. 1976, Pinheiro et al. 1976).Analytical procedures can be found in the indicated references.Andrade-Santos (2010) obtained 10 additional Rb-Sr whole-rock measurements at the São Paulo laboratory (CPGeo-USP).Analytical procedures are the same as indicated by Tassinari (1996).The instrument used was a Finnegan-MAT 262, with five Faraday collectors operated in a static way.Table 1 provides the analytical data for the 97 Rb-Sr measurements considered in this work.
A few Sm-Nd measurements, reported by Sato and Tassinari (1997), were produced at the CPGeo-USP, by using the same instrument used for the Rb-Sr measurements.Sixteen measurements, reported by Andrade-Santos (2010), were later made in the same laboratory by employing 149 Sm and 150 Nd spikes, as well as elemental separation by AG50WX8 and LN Spec resins (Sato et al. 1995).All Sm-Nd analyses available for the studied area are included in Table 2.
Several U-Pb zircon ages, obtained by TIMS, were available from the works of Priem et al. (1982) and Gaudette and Olszewski (1985).Santos et al. (2000) and Ibañez et al. (2011) produced additional U-Pb ages by Pb evaporation, SHRIMP or LA-ICP-MS.All them are displayed in Table 3. Analytical procedures are provided in the respective references.
For U-Pb dating in this work, zircon grains were extracted by standard crushing, milling, sieving (0.150 -0.063 µm), Wilfley table, Franz and heavy liquid techniques.Extracted grains were set in an epoxy disk and polished to reveal half sections.Reflected, transmitted and cathodoluminescence (CL) images were obtained by SEM and XMAX CL detectors.
Eight U-Pb dates were obtained by SHRIMP II at the GeoLab-IGc-USP (Sato et al. 2014).Mounts were gold-coated, and dating was performed with the standard Temora2 for age reference and SL13 for uranium composition.Acquisition was obtained by following the procedure described in Williams (1998).Individual ages were determined from six successive MS scans.Correction for common Pb was made based on the measured 204 Pb.The typical error component for 206 Pb/ 238 U ratios is less than 2%.Data were reduced by using SQUID 1.06.Concordia diagrams were plotted with ISOPLOT 4 (Ludwig 2009).Analytical results are included in Annex I. Seven U-Pb zircon ages were determined by a Neptune ICP-MS instrument coupled with an excimer laser ablation system.Khan titanite standard was utilized for mass bias correction and the GJ standard was utilized for zircon.Residual common Pb was corrected by using the terrestrial composition reported by Stacey and Kramer (1975).Analytical results are included in Annex II.However, these U-Pb LA-ICP-MS analyses were done in 2009, at the very beginning of the use of our instrument, when we were still dealing with the calibration of it.Now we do not have the complete knowledge of the analytical conditions of that time, when the software was set up for detrital zircon and was not optimized for crystallization ages.The Concordia diagrams of Fig. 7 are rather odd.Several measurements indicate reverse discordance, which may have been caused by inadequate adjustment of the detectors, or to some inadequacy of the 204 Pb correction, when the 204 peak may have been not properly stabilized, or perhaps to some Hg interference not detected.We are keeping the diagrams as they were produced in 2009, and the rather imprecise calculated ages are only used in this paper as indicators for the regional interpretation.Barrios et al. 1985;3 -Priem et al. 1982;4 -Fernandes et al. 1976;5 -Barrios 1983; 6 -This work.(e.g.zircon 10.1 of Fig. 4A).This rock is a paragneiss; however, only 13 zircons were dated, which is too few to analyse the statistical significance of the detrital zircons distribution.The Concordia diagram (Fig. 5A) shows that many of the analytical points are discordant.are euhedral and well preserved, presenting a prismatic habit.The length to width ratios range from 0.5:1 to 4:1, and lengths range from 0.12 to 0.37 µm.CL images reveal a complex to well-developed oscillatory zoning (e.g.. zircons 7.1 and 1.1 of Fig. 4B.The Concordia diagram (Fig. 5B) indicates a good-quality crystallization age of 1775.3 ± 7.7 Ma (MSWD = 1.2, n = 16 ) for the protolith.

PR-3141 -Biotite gneiss -Caño Cuaubén, near Puerto Inírida, Colombia
Sample PR-3141 is a fine-grained foliated biotite gneiss, very likely an orthogneiss, with a granolepidoblastic structure.Zircons of this rock are euhedral with a prismatic to sub-rounded habit.They range in length from 0.150 to 0.290 µm, with length to width ratios from 2:1 to 3:1.CL images reveal a complex to well-developed oscillatory zoning in most of the grains (e.g.zircons 6.1 and 9.1 in Fig. 4C).The Concordia diagram (Fig. 5C) shows a few discordant grains, but 15 grains near the Concordia, yield an age of 1501.0 ± 9.5 Ma (MSWD = 1.08, n = 15), which can be attributed to magmatic crystallization.

EP-2 -Biotite-gneiss -Caquetá River, near Araracuara, Colombia
Sample EP-2 is a probable biotite-muscovite orthogneiss, with a fine-grained granoblastic structure.Zircons of this sample are euhedral to subhedral, mostly with a prismatic habit.They range in size from 0.1 to 0.3 µm and have length to width ratios from 1:1 to 3:1.CL images reveal mostly oscillatory zoning (e.g.zircon 2.1 of Fig. 4D).Some points are discordant in the Concordia diagram (Fig. 5D), but a group of 10 grains located very close to the Concordia indicate an age of 1721.0 ± 9.6 Ma (MSWD = 1.8, n = 10) which is attributed to the crystallization of the igneous protolith.PR-3001 -Biotite-chlorite gneiss -Caño Cuduyarí, Vaupés River, near Mitú, Colombia Sample PR-3001 is a coarse-grained biotite-chlorite gneiss, possibly an orthogneiss, with a granolepidoblastic structure.Zircons are euhedral to subhedral, and most have prismatic habit.They range in size from 110 to 400 µm and have length to width ratios from 1:1 to 3:1.CL images reveal mostly complex and well-developed oscillatory zoning (e.g.zircons 2.1 and 5.1 in Fig. 4F).In the Concordia diagram (Fig. 5F) zircons are mostly concordant.Age calculation of 12 selected zircon grains indicates a crystallization age of 1769 ± 33 Ma (MSWD = 1.9, n = 12).
The following U-Pb dates, obtained by WS, were reported in a preliminary form by Andrade-Santos (2010).Figure 2 shows the locations of the samples.Figs.4A to 4G provides CL images of some selected zircons, and Figs.5A to 5G the resulting Concordia diagrams.

AH-1231 -Monzogranite -Serrania Mitu, Colombia
Sample AH-1231 is a monzogranite with a medium to coarse-grained faneritic structure.Zircons are euhedral to subhedral, with a prismatic habit.They range in size from 200 to 400 µm and have length to width ratios from 1.5:1 to 2:1.CL images reveal mostly complex sector and oscillatory zoning (e.g.zircons 8 and 10 in Fig. 6B).Analyses of 23 selected zircons reveal concordance and a crystallization age of 1510 ± 26 Ma (MSWD = 0.15, Fig. 7B).

AH-1248 -Paragneiss -Caño Chaquita, Atabapo River, near Puerto Inírida
Sample AH-1248 is a paragneiss with a fine-grained structure and some muscovite.Zircons of this sample are euhedral, mostly with a prismatic habit.They range in size from 100 to 230 µm and have length to width ratios from 2:1 to 6:1.CL images reveal mostly oscillatory zoning (Fig. 6C).Analyses of 37 selected zircons reveal that most are discordant.Those close to the Concordia were plotted (Fig. 7C), with most being located between 1120 and 1550 Ma.One of the grains is located at about 650 Ma and its age and tectonic significance must be investigated further.

J-42 -Paragneiss -Mitú, Colombia
Sample J-42 is a paragneiss characterized by a medium to coarse-grained texture with some centimetric K-feldspar phenocrysts.Zircons of this rock are subhedral and medium rounded.They range in length from 120 to 260 µm, with length to width ratios from 1.5:1 to 2:1.CL images reveal sector and oscillatory zoning in most grains (e.g.zircons 13 and 9, Fig. 6D).Twenty-five detrital zircons were analyzed.Several grains are not far from the Concordia (Fig. 7D), located between 700 and 1900 Ma.Possible Neoproterozoic sources have not been identified in the area, and this metasedimentary unit must be investigated further.

J-98 -Monzogranite -Caño Nabuquén, Inírida River
Sample J-98 is a monzogranite comprised of mediumto coarse-grained faneritic rock.Zircons from this rock are mostly euhedral with a prismatic habit, sizes ranging from 100 to 320 µm, and length to width ratios from 1.6:1 to 3:1.CL images reveal oscillatory zoning in most grains with well preserved cores (e.g.zircon 2 in Fig. 6E).Twenty zircons were analysed, and a few of them are discordant.Fifteen grains close to the Concordia indicate a crystallization age of 1752 ± 21 MA (MSWD = 0.13, Fig. 7E) J-159 -Tonalite -Serrania de Naquén, Guainia River, Colombia Sample J-159 is a tonalite characterized by a medium to coarse-grained faneritic texture, with some muscovite.Zircons are mostly euhedral and well preserved, with a prismatic habit.Length to width ratios range from 1:1 to 3:1, and lengths range from 120 to 310 µm.CL images reveal a complex to well-developed oscillatory zoning (Fig. 6F).Analyses of 26 zircons revealed that most are concordant, with a crystallization age of 1770 ± 40 Ma (MSWD = 0.21, Fig. 7F).

PR-3228 -Paragneiss -Rio Mesai, Yarí, N of Araracuara, Colombia
Sample PR-3228 is a biotite gneiss, with microcline and some chlorite, characterized by a fine to medium granoblastic structure.Zircons of this rock are euhedral to subhedral with a prismatic habit.They range in length from 70 to 250 µm with length to width ratios from 1,3:1 to 4:1.
CL images reveal a complex oscillatory zoning in most of the grains (e.g.zircon 3, Fig. 6G) and some complex sector zoning.Thirty-eight zircons were analyzed and plotted in Fig. 7G.Although most of the zircons are discordant, nine grains are nearly concordant and yield detrital ages between 1800 and 1300 Ma.Older grains were not found.

SAMARIUM-NEODYMIUM MEASUREMENTS
Twenty-two Sm-Nd model ages are currently available for the study region (see Fig. 2 for sample locations and Tab. 2 for the analytical data).Twelve of these samples (described above) were also dated by the U-Pb zircon method.
Most samples yield very similar paleoproterozoic T DM model ages (ca.1.9 -2.2 Ga, Tab. 2) although their U-Pb zircon ages varied within the 1800 -1500 MA interval.Their calculated ε Nd(TDM) values were also similar (positive values of 3.0 -3.5), suggesting formation from the same juvenile source material.Given that the ε Nd(T1) values of the granitoid rocks are near zero or slightly negative, the possible presence of much older source material is improbable.This finding reinforces the idea of accretion though subduction during the Proterozoic, as well as the presence of juvenile magmatic arcs in this part of the Rio Negro-Juruena province.
There is marked similarity in all trends in the Nd isotopic evolution diagram for the 11 granitoid rocks dated by the U-Pb zircon method (Fig. 8).Rocks with ages in the 1800 -1750 Ma range exhibit ε Nd(T1) values close to zero, but 2 younger rocks (J-84 and PR-3141, ages around 1550 Ma) have moderately negative values.We suggest that these younger granitic rocks could have  originated from the melting or complete reworking of the accretionary crustal material that formed about 200 to 250 Ma earlier.

General remarks
A total of 97 granites, gneisses and migmatites within the studied region were analysed by the Rb-Sr method and selected for further interpretation (Fig. 3, Tab.1).Some of the analyses were performed recently by the CPGeo-USP, but most of them were taken from the literature (Gaudette and Olszewski 1985, Priem et al.1982, Barrios et al.1985, Fernandes et al.1976, Pinheiro et al. 1976).An interpretative exercise was made to verify the possible temporal relationship among granitoid rocks located close enough to have been subjected to the same geological history.Potentially related samples were identified in eight areas (different colors in Fig. 3, Ventuari River, Atabapo River, Negro-Casiquiare Rivers, Puerto Inírida, San Carlos, Mitu-Iauaretê, Caquetá River and São Gabriel + Içana River).
Analytical points of the potentially related samples were included in isochron diagrams (Fig. 9A-H), each of which obtained in different works from different laboratories, with different equipment and precision levels.To "normalize" the calculations, we fixed the same values to the experimental errors of the calculated 87 Rb/ 86 Sr (3%) and 87 Sr/ 86 Sr (0.25%) results.This approach should minimize the preference for precise analytical results during the Isoplot calculations.As the granitoid samples in each diagram are not strictly cogenetic, the calculated isochron ages should be viewed as rough approximations for interpreting the overall tectonic history.Analytical points in all diagrams (Fig. 9) show reasonable alignments.As the calculated best-fit lines could be broadly interpreted as "isochron ages", with probable geological significance, they are referred as "reference isochrons".We speculate that these lines indicate, for each area, a regional event of Sr homogenization for the wholerock system.(Venezuela).These rocks are located along a reference isochron of 1837 ± 87 Ma, with a low initial 87 Sr/ 86 Sr ratio of 0.7021 (Fig. 9A).Analysis of the same rocks by U-Pb zircon (TIMS) revealed concordant values of 1859 Ma at Minicia and 1823 Ma at Macabana (Tab.3).Several authors (Gaudette and Olszewski 1985, Priem et al. 1982, Barrios et al.1985), collected 17 samples of granitoid rocks along the Atabapo River (Tab.1).Their isochron diagram (Fig. 9B) indicated an apparent age of 1749 ± 92 Ma, with an initial 87 Sr/ 86 Sr ratio of 0.7073.A similar age span (1787 ± 53 Ma), with an initial 87 Sr/ 86 Sr ratio of 0.7025 (Fig. 9C), was obtained for samples from the Negro and Casiquiare Rivers, by Priem et al. (1982), Gaudette and Olszewski (1985), and the present work (Tab.1).Four isolated samples of granitoid rocks collected along the Inírida and Guainia Rivers yielded precise U-Pb zircon ages in the same range (Fig. 2, Tab.3): J98(1772 ± 15 Ma), J-127 (1772 ± 4 Ma), J-159 (1785 ± 6 Ma) and J-199 (1796 ± 4 Ma).Considering the isolated U-Pb zircon ages and the Rb-Sr reference isochrones (Fig. 9B and 9C), we provisionally name this granitoid region of the Rio Negro-Juruena province as the "Atabapo belt".We propose that a series of orogenic pulses, lasting at least 60 Ma, from 1800 to 1740 Ma, in the late Paleoproterozoic (Statherian), was responsible for the development of this belt in the NE part of the study area.

Atabapo belt
We constructed another isochron diagram (Fig. 9D) from data of six samples of a non-deformed granite (Priem et al. 1982), and one sample of the present work (Tab.1), obtained within the same corner near Puerto Inírida.The apparent age of these samples was 1476 ± 68 Ma, with an initial 87 Sr/ 86 Sr ratio of 0.7064.One sample of the present work (Tab. 1) and seven previously obtained granitic samples (Gaudette and Olszewski 1985, Priem et al. 1982, Barrios et al. 1985), all collected near the town of San Carlos at the Negro River, yielded an age of 1521 ± 52 Ma, with an initial 87 Sr/ 86 Sr ratio of 0.7051 (Fig. 9E).Three isolated granitic samples within the same region were dated by the U-Pb method (PR-3141: 1500 ± 9 Ma, J-84: 1507 ± 19 Ma, and PRA-2: 1480 ± 70 Ma; Tab. 3).The results confirm the intrusive age of some granitic intrusions into the Atabapo belt at about 1500 Ma, within the Mesoproterozoic (Calymmian).Priem et al. (1982), Pinheiro et al. (1976) and Santos (2003) obtained and dated 27 samples of granitoid rocks from central area of the study region, between the villages of Mitú, Colombia, and Iauaretê, Brazil.A few samples from the same area were analysed by us (Tab.1).A reference isochron (Fig. 9F) seems to indicate a mesoproterozoic ( calymmian) Sr homogenization event at 1529 ± 43 Ma with an initial 87 Sr/ 86 Sr ratio of 0.7067.Within the same area, there are 8 U-Pb zircon ages (see Fig. 2 and Tab.3).Four of them are clearly older, and have statherian apparent ages (AH-1213A: 1746 ± 8 Ma; PR-3001: 1740 ± 5 Ma; J-36: 1739 ± 38 Ma; HB-667: 1778 ± 4 Ma), whereas the 4 other are within the same age range as the Rb-Sr isochron (AH-1231: 1555 ± 7 Ma; AH-1216: 1574 ± 10 Ma; PA-SP-22: 1521 ± 13 Ma; PRA-4: 1552 ± 34 Ma).In this region, the younger granitoid rocks are described as calc-alkaline syntectonic gneisses and migmatites affected by medium-level amphibolite facies metamorphism (Priem et al. 1982, Santos 2003).Thus, we may conclude that the results of the Rb-Sr systematics indicate an episode of Sr isotopic homogenization of mesoproterozoic age related to that specific orogenic pulse.

Vaupés belt
We provisionally name this area "Vaupés belt".We postulate that the belt developed from a series of orogenic pulses in the Calymmian, with duration of at least 60 Ma, between 1580 and1520 Ma.Older statherian granitoid rocks in the region (U-Pb zircon age ~ 1750 Ma) may be considered as basement inliers.These rocks must have been involved within the younger calymmian metamorphism.A few younger and undeformed granitic rocks, whose points lie within the same reference isochron, could represent post-tectonic granitic batholiths.
Four samples by Priem et al. (1982) and one by us were obtained from the SW of the study area, near Araracuara.The reference isochron with these samples indicated an apparent age of 1557 ± 41 Ma, with an initial 87 Sr/ 86 Sr ratio of 0.7050 that was fixed in the age calculation (Fig. 9G).Older ages in the same area were obtained by Ibañez et al. (2011) from two syenogranitic gneisses (PR-3215: 1756 ± 8 Ma; J-263: 1732 ± 17 Ma) and sample EP-2 from this work 1725 ± 10 MA (Fig. 2 and Tab.3).As was the case in the central Mitú-Iauaretê region, these older rocks could represent statherian basement rocks within younger mesoproterozoic metamorphic gneisses.However, for the same region, along the Apaporis river, the same authors encountered three other rocks with younger ages (CRJ-19: 1593 ± 6 Ma; PR-3092: 1578 ± 27 Ma; AH-1419: 1530± 21 Ma), reinforcing the idea of the existence of a Vaupés tectonic-metamorphic belt of calymmian age.

Regional thermal evolution
The initial Sr 87 /Sr 86 ratios encountered in the reference isochrones of Figs 9A to 9G (0.702 -0.708), are considered relatively low and indicate an important participation of juvenile material.This possibility is supported by the Sm-Nd systematics.The only exception to the relatively low initial ratios was observed for samples from the Brazilian region along the Içana River and near the town of São Gabriel da Cachoeira (Fig. 9H).In this case, the reference isochron of 11 points yielded an apparent age of 1203 ± 58 Ma with a very high initial 87 Sr/ 86 Sr close to 0.72.This age, much younger than what was encountered for most domains within the study area, should be related to a very strong regional heating that imposed Sr isotopic redistribution in the whole-rock samples.This condition seems to have been localized to an elongated area separating the Atabapo and Vaupés belts (Fig. 3), affected respectively by metamorphic belts with statherian and calymmian age.About 100 K-Ar measurements, predominantly from micas, were made a few decades ago in rock samples from the study area.Complete analytical data of these measurements and sample locations can be found in Priem et al. (1982) for Colombia, Barrios et al. (1985) for Venezuela and Tassinari (1996) for Brazil.A histogram of all available apparent K-Ar ages (Fig. 10) reveals that the ages are concentrated within the 1200 -1400 Ma interval.This result reflects regional heating above 350 -400 o C, which affected the entire territory of Figure 2 and beyond, covering much of the Amazonian Craton.This regional heating was first observed in the 1960's (Priem et al. 1971) and is called the Nickerie thermo-tectonic episode in Suriname and K'Mudku in the Guyana Republic (Gibbs, Barron 1993).Cordani et al. (2010) attempted a comprehensive review and tentative interpretation of this major intraplate heating episode, whose duration may have been of 100-to-200 Ma.It is clear that this mesoproterozoic thermal episode was pervasive and widespread in the Amazonian Craton.

General considerations
Robust geochronological tools are required when synthesizing the tectonic evolution of a very large region where basic geologic information is quite rare.Here, we employ four radiometric methods, each with its own interpretative value and tectonic significance.U-Pb measurements, either SHRIMP, ICP or TIMS, are essential for any geochronological work, because they produce significant punctual ages.However, alone, these measurements do not provide an entire geological history.They may indicate several magmatic events localized in time, but not their integration into a complete regional tectonic evolution.Rb-Sr whole-rock isochrones are less precise, but they indicate the timing of relevant episodes of Sr isotopic homogenization, related to medium-to high-grade metamorphic episodes.Sm-Nd model ages give insight into the type of the regional tectonic processes (e.g.intraplate or subduction-related, accretionary or collisional, juvenile or reworked).Finally, K-Ar ages, especially of micas, are related to the final cooling of the region, usually with respect to the principal episode of cratonization, or, alternatively, to some episodes of major intraplate crustal heating above 350 -400 o C.
According to Cordani and Teixeira (2007), the Rio Negro-Juruena tectonic province of the Amazonian Craton (1.78 -1.55 Ma) was formed by continued soft-collision/ accretion processes driven by subduction, which produce a very large "basement" with the predominance of granitoid rocks, many of them with a juvenile-like Nd isotopic signature.Clear evidence of archean or paleoproterozoic basement has not yet been found in this region.This province is considered to be basically accretionary, formed from the complex juxtaposition of tectonic units, including intra-oceanic material, but also containing Cordilleran-type granites, collisional-type belts, volcanic-sedimentary basins, as well as post-tectonic and anorogenic-type complexes.

Summary of the tectonic history
Considering the geochronological pattern encountered in the Atabapo belt, we agree with Gaudette and Olszewski (1985), Barrios et al. (1985) and Cordani et al. (2000) that the possible NE boundary of the Rio Negro-Juruena province with the older Ventuari-Tapajós province would be located close to or along the Atabapo River (Fig. 2 and 3).The U-Pb zircon SHRIMP measurements indicate the formation of a series of statherian magmatic arcs in that region, in which juvenile and possibly intra-oceanic material predominates.Closely related to subduction, these magmatic arcs piled up by soft-collision episodes and successive stacking from SW to NE, encompassing a period of about 60 Ma.In the SW region, comprising the Mitú-Iauaretê and Caquetá River areas, calymmian granitic and gneissic rocks formed between 1580 and 1500 Ma within the Vaupés belt.These calymmian ages can be confirmed by the U-Pb zircon ICP-MS ages obtained by Ibañez et al. (2011) on granitic rocks collected along the Apoporis River (see Fig. 2), well inside the younger belt.Results of the Sm-Nd systematics for rocks of the Vaupés belt indicate the presence of substantial juvenile material.An important time-gap of 150-to-200 Ma exists between the youngest rocks of the Atabapo belt (1740 Ma) and the oldest rocks of the Vaupés belt (1580 Ma).The latter are products of a second orogenic pulse within the same Rio Negro-Juruena tectonic province.Thus, there is ample time for cratonization of the first series of tectonic belts before the stacking of the second series of possibly accretionary belts in Mesoproterozoic time.At the NE corner, a cratonized 1740 -1800 Ma basement was intruded by granitic batholiths at 1550 Ma, which may correspond to the reflection of the orogenic pulse occurring at that time to the SW.However, at the SW corner, the second 1580 -1500 Ma orogenic pulse includes parts of possibly retrogressed basement inliers with ages of about 1750 Ma.Gorayeb et al. (2005) showed that the NW part of the Rio Negro-Juruena province continues to the SE below the Solimões sedimentary basin, where its basement presents a few U-Pb zircon evaporation ages in the 1800 -1550 Ma range.Geochronological control in the SE half of the province, in Mato Grosso, Brazil, shows that the ages of the granitoid rocks decrease from NNE to SSW.Near the boundary with the older Ventuari-Tapajós province, in the region of Alta Floresta, Santos (2003) encountered ages of 1780 Ma in the São Romão and São Pedro granites.On the other side, near the border with the Rondonian-San Ignacio belt, in the Alto Jauru region of Mato Grosso, Geraldes et al. (2001) reported the age of the Cachoeirinha magmatic arc as 1590 Ma.We found a comparable age pattern for the NW part of the Amazonian Craton (Fig. 3), where calc-alkaline orogenic type rocks yielded ages of 1800-to-500 Ma, decreasing from the NE (Atabapo belt) to the SW (Vaupés belt).

Thermal history
Finally, a peculiar but remarkable aspect of the tectonic evolution of this area is the widespread Nickerie-K'Mudku intraplate mid-proterozoic regional heating.This phenomenon affected all the rock units of the study area, from about 1400-to-1200 Ma (see histogram of Fig. 10).This regional heating episode, with temperatures exceeding 300 o C, was uniform and affected the entire crust of the study area.The episode affected very large parts of the Amazonian Craton, and its duration may have been on the order of 100 -200 Ma.All the rock ages determined by U-Pb zircon measurements are much older (1800 -1500 Ma).Regardless of their ages, all rocks were affected in the same way.Therefore, none of the K-Ar apparent ages likely represents a primary magmatic age of the corresponding dated rock.Considering the Rb-Sr systematics of the region, the Rb-Sr reference isochron age (1200 ± 60 Ma) and the high initial 87 Sr/ 86 Sr ratio (~ near 0.720; Fig. 3 and  9G), we can make some additional speculations from the K-Ar ages.Rocks in the area of São Gabriel da Cachoeira and the Içana River, in NW Brazil, are well within the Rio Negro-Juruena province, and very likely have primary ages between 1800 and 1500 MA, as suggested by the U-Pb zircon ages.However, the rocks were perhaps heated to as high as 600 o C, which would be necessary to produce the observed widespread Sr isotopic homogenization in the whole-rock systems.This very high heating event seems to have been restricted to the territory located more or less between the Atabapo and Vaupés belts (see Fig. 3).

CONCLUSIONS
From the currently available data, we suggest the following possible tectonic-thermal history for the overall region: 1. Formation of the first orogenic pulse of the province, the Atabapo belt, with stacking of magmatic arcs of the Atabapo-Negro-Casiquiare region against the cratonic area formed by the Ventuari-Tapajós continent, at 1800 -1740 Ma. 2. Formation of the second orogenic pulse related to the Vaupés belt, with stacking of the Mitú-Iauretê and Caquetá magmatic arcs against the already cratonized area of the first pulse, at 1580 -1500 Ma. 3. Onset of the Nickerie-K'Mudku intraplate regional heating to above 300 o C within the entire region at 1200 -1300 Ma, but attaining 600 o C in the belt separating the first (Atabapo) and second (Vaupés) orogenic pulses.4. The probable existence of younger metasedimentary units such as the paragneisses J-36, J-42, AH-1248 and PR-3228, which contain young detrital zircons of meso and neoproterozoic age, should be investigated.
Figure 2. NW Corner of the Amazonian Craton.Location of samples analysed by the U-Pb and Sm-Nd methods.

Figure 2 Figure 3 .
Figure 2 illustrates the study area and location of samples with U-Pb zircon ages, 15 of which were produced in this work.Cathodoluminescence (CL) images of some selected zircons dated by SHRIMP are shown in Figs.4A to 4H.The resulting Concordia diagrams are provided in Figs.5A to 5H.J-36 -Muscovite-chlorite paragneiss -Vaupés River, near Mitú, Colombia Sample J-36 is a fine-grained muscovite-chlorite paragneiss, with a granolepidoblastic structure and centimetric porphyroblasts of plagioclase.Zircons from this rock range in size from 70 to 220 µm and exhibit length to width ratios from 2:1 to 3:1.CL images reveal oscillatory zoning

Table 1 .
Rb-Sr analytical data for the rock samples included in Fig.3.
Only 5 of them are relatively close to the Concordia, with ages between 1800 and 1000 Ma.Zircons 4.1 (6/38 age of 2094 ± 17 Ma) and 13.1 (6/38 age of 2089 ± 16 Ma) indicate the possible existence of much older sources.

Table 3 .
U-Pb ages already available for the region, obtained by means of different methods.
Loios for zircon s eparation.We thank geologist Guilherme Andrade Santos for his aid during the preliminary phase of the work.W.R. Van Schmus and M. Ibañez-Mejia are acknowledged for their much appreciated and important revision of the original version of our manuscript.INGEOMINAS of Colombia provided the samples and valuable geological information for this research.Financial support was received from FAPESP through grant 2013/12754-0 to UGC.Pb zircon ages from the Amazonas Territory of Colombia Annex I. Isotope ratio data for samples analysed by LA-ICP-MS method.Pb zircon ages from the Amazonas Territory of Colombia Brazilian Journal of Geology, 46(Suppl 1): 5-35, June 2016 U-Continue... 29 Brazilian Journal of Geology, 46(Suppl 1): 5-35, June 2016 Umberto Giuseppe Cordani et al. 30 Brazilian Journal of Geology, 46(Suppl 1): 5-35, June 2016 U-31 Brazilian Journal of Geology, 46(Suppl 1): 5-35, June 2016 Umberto Giuseppe Cordani et al.