Tectonic significance of the Meso- to Neoarchean complexes in the basement of the southern Brasília Orogen

Accepted Manuscript Tectonic significance of the Meso- to Neoarchean complexes in the basement of the southern Brasília Orogen Caue Rodrigues Cioffi, Mario da Costa Campos Neto, Andreas Möller, Brenda Chung Rocha PII: DOI: Reference:

S0301-9268(16)30137-1 http://dx.doi.org/10.1016/j.precamres.2016.10.009 PRECAM 4601

To appear in:

Precambrian Research

Received Date: Revised Date: Accepted Date:

11 May 2016 3 October 2016 16 October 2016

Please cite this article as: C. Rodrigues Cioffi, M. da Costa Campos Neto, A. Möller, B. Chung Rocha, Tectonic significance of the Meso- to Neoarchean complexes in the basement of the southern Brasília Orogen, Precambrian Research (2016), doi: http://dx.doi.org/10.1016/j.precamres.2016.10.009

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

Tectonic significance of the Meso- to Neoarchean complexes in the basement of

2

the southern Brasília Orogen

3 4

Caue Rodrigues Cioffia.b; Mario da Costa Campos Netoa; Andreas Möllerb; Brenda Chung Rochaa,b

5

a

6

b

Instituto de Geociências, Universidade de São Paulo, São Paulo, SP, Brazil. Department of Geology, The University of Kansas, Lawrence, KS, USA

7 8

Abstract

9 10

The late Archean was a period of important changes in geodynamic processes and magmatism

11

style. This period seems to mark the time when crustal reworking processes start to

12

predominate over new continental crust generation and is most likely related to the beginning

13

of continental collision and “modern-style” plate tectonics. This study reports a new dataset of

14

zircon U-Pb and Hf isotopes, whole-rock geochemistry and Nd isotopes from Meso- to

15

Neoarchean complexes within the basement of the Neoproterozoic southern Brasília Orogen,

16

SE Brazil. The data provide important insights into the petrogenesis of Mesoarchean TTG suites

17

and their implications for tectonic settings of Archean continental crust generation. Isotopic

18

and geochemical data constrain the timing and nature of the Neoarchean transition from TTG-

19

type to high-K granitic magmatism in the studied area and we discuss the involvement of these

20

complexes in the southern São Francisco paleocontinent assembly. A well-defined period of

21

TTG-type magmatism between 3.00 and 2.96 Ga is identified, with mostly suprachondritic εHf(t)

22

values, between 0 and +5.1, associated with average two-stage model ages between 3.2 and

23

3.3 Ga. Whole-rock Nd analyses yield TDM ages between 3.2 and 3.4 Ga. These TTGs are

24

interpreted as juvenile additions to the continental crust most likely generated by partial

25

melting of a hydrous mafic crust. The variable trace element compositions of the analyzed TTG

26

samples indicate partial melting at different depths at the same time. These results strongly

27

support the idea of a non-unique tectonic setting of Archean continental crust generation. An

28

additional period of Neoarchean high-K granitic magmatism at ca. 2.76 Ga is interpreted to

29

record the transition from TTG-type to high-K granitoid magmatism in the studied area. This

30

Neoarchean magmatism is associated with less radiogenic isotopic signatures with mostly

31

negative εHf(t) values. The isotopic and geochemical signatures suggest that reworking of the

32

Mesoarchean TTG crust, most likely in a collisional setting, was the main mechanism for the

33

generation of the Neoarchean granitic suite. This supports the idea that the Neoarchean

34

transition from TTGs to granites was most likely a result of the beginning of continental

35

collision processes.

36

1. Introduction

37

The Archean Eon has been considered a major period of continental crust generation.

38

Models for the growth of the continental crust based on Hf isotopes indicate that ca. 50 to 75

39

% of the volume of the continental crust was extracted from the mantle during this time

40

(Belousova et al., 2010; Dhuime et al., 2012; Roberts and Spencer, 2014). However, large parts

41

of this Archean continental crust were involved in later tectonic events and its present-day

42

record is very fragmented (e.g. Bleeker, 2003). According to deWit (1998), well-preserved

43

Archean continental crust represents less than 8% of the present-day exposed continental

44

surface area. This makes it challenging to understand the mechanisms and settings of Archean

45

continental crust generation and it also makes every Archean terrane a valuable piece of

46

information about the early Earth´s continental crust evolution.

47

Archean juvenile felsic crust is mostly composed of low-K granitoid gneisses of the

48

tonalite-trondhjemite-granodiorite (TTG) series (e.g. Moyen, 2011; Moyen and Martin, 2012;

49

Laurent et al., 2014). TTGs are key elements to understand the processes and sites of Archean

50

continental crust generation. These rocks are generally interpreted as a result of partial

51

melting of hydrous metamafic rocks, but there is an ongoing debate about the depth of

52

melting and geodynamic sites of TTG magma generation (e.g. Foley, 2002; Rapp, 2003; Bédard

53

et al, 2006; Moyen and Stevens, 2006; Moyen, 2011; Moyen and Martin, 2012). Regional

54

studies have shown that the TTG series is not homogenous with respect to trace element

55

compositions, especially regarding Sr and HREE concentrations (e.g. Halla et al., 2009; Almeida

56

et al., 2011). These elements are sensitive to the pressure of melting and some studies have

57

therefore suggested the existence of low- and high-pressure TTG subseries (Halla et al., 2009;

58

Moyen, 2011). These different TTG subtypes were formed in a wide range of pressure

59

conditions from ca. 10 to 25 kbar (Moyen, 2011 and references therein) and thus coeval

60

occurrences of low- and high-pressure TTGs would be strong arguments in favor of a non-

61

unique tectonic setting of Archean continental crust generation.

62

Another striking feature of the Archean record is the occurrence of a late-Archean

63

transition from juvenile TTG-type to high-K granitic magmatism (e.g. Sylvester, 1994; Romano

64

et al., 2013; Laurent et al., 2014; Farina et al., 2015). Late-Archean granites have been

65

described in every well-studied Archean terrain worldwide and are generally interpreted as a

66

result of differentiation of the preexisting juvenile TTG crust most likely in collisional settings.

67

Based on that, this transition has been interpreted as a result of the emergence of continental

68

collision tectonics and to mark the time since when crustal recycling processes predominated

69

over juvenile additions (Sylvester, 1994; Laurent et al., 2014). This transition has so far been

70

documented for different times in different Archean terranes worldwide, from ca. 2.95 to 2.55

71

Ga (e.g. Laurent et al., 2014). The need for a global approach to determine its timing and

72

tectonic significance requires to determine the age and nature of this transition in the world’s

73

less studied Archean terranes, including those cryptic ones occuring as basement complexes

74

within younger orogens, such as the one studied here.

75

The combination of zircon U-Pb and Hf isotope techniques is a powerful tool to

76

understand the mechanisms of crust formation and evolution in Archean terranes. This

77

approach is very useful to identify different crustal domains within cratons that could have

78

been amalgamated by horizontal tectonics (Griffin et al., 2004; Zeh et al., 2009; Geng et al.,

79

2012). However, as demonstrated by Laurent and Zeh (2015), without petrological and

80

geochemical information about the zircon-hosted rocks, these techniques are not fully capable

81

of deciphering Archean geodynamic processes. In this contribution, a complete dataset

82

including zircon U-Pb and Hf isotopes as well as whole-rock geochemistry and Nd isotopes is

83

used to constrain the tectonic evolution of the Archean complexes in the basement of the

84

southern Brasília Orogen. These poorly studied Archean complexes have been interpreted as

85

parts of the São Francisco paleocontinent strongly reworked during the development of the

86

Neoproterozoic southern Brasília Orogen. The presented dataset allows us to discuss the origin

87

of the Mesoarchean TTG suites and their implications for tectonic settings of Archean

88

continental crust generation as well as the timing and nature of the transition from TTG to

89

high-K granitic magmatism in the studied area. The new data also give clues on how these

90

crustal segments were involved in the assembly of the São Francisco paleocontinent.

91 92

2. Geological setting

93

The Neoproterozoic Brasília Orogen (Dardenne, 2000) surrounds the western and

94

southern edges of the São Francisco craton and its southernmost portion has been interpreted

95

as a product of an Ediacaran (ca. 630 Ma) collision between the active margin of the

96

Paranapanema plate and the passive margin of the São Francisco paleocontinent (Brito Neves

97

et al., 1999; Campos Neto, 2000; Trouw et al., 2000). This part of the orogen comprises a thick-

98

skinned nappe stack with southwest-dipping tectonic wedge and transport towards east-

99

northeast (Campos Neto and Caby, 1999, Campos Neto et al., 2011; Trouw et al., 2000, 2013).

100

Two main tectonic domains are identified within this nappe system (Fig. 1): (1) an active

101

margin domain related to the Paranapanema plate, consisting of a magmatic arc unit (the

102

Socorro-Guaxupé Nappe) (Campos Neto and Caby, 2000; Janasi, 2002) and active margin-

103

related metasedimentary units (the Andrelândia Nappe System) (Campos Neto et al., 2010,

104

2011) and (2) a passive margin domain related to the São Francisco paleocontinent that

105

comprises basement orthogneisses (Fetter et al., 2001; Cioffi et al., 2016) and passive margin-

106

related metasedimentary units (the São Vicente Complex and the Carrancas and Lima Duarte

107

nappes) (Rocha, 2011; Westin and Campos Neto, 2013; Westin et al., 2016).

108

The basement orthogneisses can be divided into two main tectonic domains: (1) the

109

Paleoproterozoic Pouso Alegre Complex (Cioffi et al., 2016) and (2) the Archean complexes

110

(the Amparo, Serra Negra, Heliodora-Minduri and Mantiqueira complexes) (Fetter et al., 2001;

111

Tassinari and Nutman, 2001; Peternel, 2005) (Fig. 1). The Pouso Alegre Complex orthogneisses

112

have igneous crystallization ages between 2.15 and 2.08 Ga associated with juvenile Nd and Hf

113

signatures (Fetter et al., 2001; Campos Neto et al., 2011; Cioffi et al., 2016). The Pouso Alegre

114

Complex has been interpreted as a continuation of the cratonic Paleoproterozoic Mineiro Belt

115

arc system (e.g. Ávila et al., 2010; Seixas et al., 2013; Barbosa et al., 2015; Teixeira et al., 2015)

116

underneath the southern Brasília Orogen (Cioffi et al., 2016). Assuming that hypothesis is

117

correct, this at least 350 km long, NE-SW trending arc system, could be originally lying in-

118

between the Archean crust of the São Francisco craton and the Archean complexes in the

119

basement of the southern Brasília Orogen.

120

Published data from the Archean complexes in the southern São Francisco craton show

121

U-Pb zircon crystallization ages spanning from ca. 3.22 to 2.72 Ga with a gap between ca. 3.20

122

and 2.93 Ga (e.g. Lana et al., 2013; Farina et al., 2015). Other previous data from the Archean

123

complexes in the basement of the southern Brasília Orogen are very scarce and not fully

124

published (conference abstracts of: Fetter et al., 2001; Tassinari and Nutman, 2001;

125

unpublished thesis of: Peternel, 2005; Santos, 2014), indicating igneous crystallization ages

126

between 3.02 and 2.75 Ga with TDM ages from 3.3 to 3.0 Ga. This could mean that the ages of

127

granitoid crystallization in the southern Brasilia Orogen are different from the Sao Francisco

128

craton, but this has to be confirmed by a more detailed study.

129 130

3. Geology of the Archean complexes in the basement of the southern Brasília Orogen

131

The Archean basement of the southern Brasília Orogen consists of three orthogneiss

132

complexes, the Amparo, Serra Negra and Heliodora-Minduri complexes, which are described

133

below. These complexes occur as large antiforms within the southern Brasília Orogen (Figs. 1

134

and 2) and representative samples of these three complexes were selected for analyses (Fig. 2)

135

(Table 1). The Amparo Complex (Ebert, 1968) is located in the western part of the southern

136

Brasília Orogen (Figs. 1 and 2) and mainly consists of layered migmatitic orthogneisses of

137

tonalitic to granitic compositions with centimeter- to meter-scale amphibolite layers parallel to

138

the main foliation (Fig. 3a). The Amparo Complex tonalitic rocks (samples A4, A9I, A9K) are

139

medium-grained, dark-gray, migmatitic biotite-hornblende orthogneisses (Fig. 3c) with a color

140

index of ca. 12-15%. Peritectic hornblende crystals up to 2 cm length are common within

141

tonalitic leucosomes (Fig. 3c). Common accessory minerals of tonalitic samples include

142

chlorite, titanite, apatite, opaque minerals, zircon and allanite. Amparo Complex granodioritic

143

rocks (samples A5, A9B) are fine- to medium-grained migmatitic biotite orthogneisses with a

144

color index of ca. 5-7%. Common accessory minerals of granodioritic samples are chlorite,

145

titanite, apatite, allanite and zircon. Granitic rocks (sample C22) are pinkish-gray, fine- to

146

medium-grained, migmatitic biotite orthogneisses with a color index of ca. 5% (Fig. 3b).

147

Common accessory minerals of granitic samples are: chlorite, allanite, opaque minerals,

148

epidote, apatite and zircon. All of the Amparo Complex samples were taken from bulk

149

migmatites that consist of unsegregated neosome and small-scale leucosomes (up to 5mm

150

thick), too thin to be sampled individually.

151

The Serra Negra Complex occurs in the western portion of the southern Brasília Orogen

152

in close association with the Amparo Complex (Fig. 2). The main lithotype is a homogeneous

153

medium-grained, gray, biotite orthogneiss of granodioritic composition with a color index

154

between 5-7% (sample C20) (Fig. 3d). Accessory minerals include titanite, apatite, allanite,

155

amphibole, epidote, opaque minerals and zircon. The relationships between the Amparo and

156

Serra Negra complexes are not well understood. However, because of the more complex

157

deformation patterns observed in the Amparo Complex, the Serra Negra Complex has been

158

considered intrusive into the Amparo Complex (Campos Neto et al., 2011). The Heliodora-

159

Minduri Complex is located in the central and eastern portions of the southern Brasília Orogen

160

(Figs. 1 and 2). The main lithotype is a light-gray, fine- to medium-grained, biotite orthogneiss

161

of trondhjemitic composition, with a color index of ca. 5% (sample C37) (Fig. 3e). Common

162

accessory minerals are apatite, epidote and zircon.

163 164 165 166

4. Analytical methods Samples were collected at the sites shown in Figure 2. A data summary including coordinates of sampling sites is presented in Table 1.

167 168

4.1 Zircon U-Pb geochronology

169

For U-Pb analyses, zircon grains were extracted from crushed whole-rock samples using

170

heavy-mineral separation techniques that include a disk mill, Wilfley table, FrantzTM

171

isodynamic magnetic separator and heavy liquids (bromoform and methylene iodide). Zircon

172

grains were then handpicked, mounted in epoxy resin discs and polished to half width.

173

Cathodoluminescence (CL) images of four samples (A9I, A9K, C20, C37) were obtained using a

174

scanning electron microscope (SEM) at the Microscopy and Analytical Imaging Laboratory

175

(MAI), The University of Kansas. CL images of sample C22 were acquired with a SEM at the

176

Geochronological Research Center (CPGeo) of the Universidade de São Paulo. U-Pb analyses of

177

all samples were obtained by laser ablation-inductively coupled plasma-mass spectrometry

178

(LA-ICP-MS). Four samples (A9I, A9K, C20, C37) were analyzed at Department of Geology, The

179

University of Kansas, using a Thermo Scientific Element2 ICP-MS attached to a Photon

180

Machines Analyte.G2 193 nm ArF excimer laser ablation system. The laser was used to ablate

181

20 µm circular spots and was set to 2.2 J cm-2 fluency at a 10 Hz repetition rate. The ablated

182

material was carried to the ICP in He gas. Elemental fractionation, downhole fractionation and

183

calibration drift were corrected by bracketing measurements of unknowns with the GJ1

184

reference material (Jackson et al., 2004) and data reduction using the VizualAge data reduction

185

scheme (Petrus and Kamber, 2011) for the IOLITE software package (Paton et al., 2011). The

186

analyses were performed in two analytical sessions. During the first one (samples A9I, A9K,

187

C37), the secondary standard Plešovice (Sláma et al., 2008) yielded a weighted mean 206Pb/238U

188

date of 336.1 ± 1.0 Ma (2σ) (n=26; MSWD=0.96), in good agreement with the age determined

189

by TIMS (337.13 ± 0.37 Ma; Sláma et al., 2008). During the second analytical session (sample

190

C20) the secondary standard Plešovice yielded a weighted mean 206Pb/238U date of 342.1 ± 1.9

191

Ma (2σ) (n=10; MSWD=1.7). No common lead correction was made for these two analytical

192

sessions. Sample C22 was analyzed at CPGeo-USP using a Thermo Scientific Neptune multi-

193

collector ICP-MS attached to a Photon Machines Analyte.193 nm ArF laser ablation system.

194

The laser was used to ablate 32 µm circular spots at a repetition rate of 6 Hz, and He was used

195

as the carrier gas. The GJ1 reference material was used as the primary standard, and

196

corrections were made using an in-house spreadsheet. The data from this sample were

197

corrected for common lead based on

198

Concordia plots and weighted mean U-Pb dates were derived using Isoplot (Ludwig, 2003). The

199

U-Pb data is shown in Online Supplementary Table S1.

204

Pb and the model of Stacey and Kramers (1975).

200 201

4.2 Zircon Lu-Hf isotopes

202

Zircon Lu-Hf analyses were performed on four samples (A9K, C20, C22, C37). The

203

analyses were carried out at the CPGeo-USP using a Neptune multi-collector ICP-MS attached

204

to a Photon Machines Analyte.193 nm ArF laser ablation system. The laser was used to ablate

205

47 µm circular spots that were placed on the same internal domains dated by U-Pb technique.

206

The isotopes

207

simultaneously. Hf isotope ratios were corrected for mass bias assuming a

208

0.7325. Yb isotope ratios were corrected from mass bias assuming a

209

1.123456. The mass behavior of Lu was assumed to follow that of Yb. The interference of 176Lu

210

on

211

ratio of 0.026549. The interference of

212

ratio of 0.786956. During the course of analyses, the zircon standard GJ1 yielded a weighted

213

average

214

were conducted based on the 176Lu decay constant of 1.867 x 10-11 a-1 (Söderlund et al., 2004)

215

and the present-day chondritic ratios of

216

(Bouvier et al., 2008). The present-day depleted mantle ratios of

217

176

218

Hf analyses are shown in Online Supplementary Table S2.

176

171

Yb,

173

Yb,

175

Lu,

177

Hf,

178

Hf,

179

Hf, and

176

176

Yb on

176

(Hf+Yb+Lu) were collected 179

173

Hf/177Hf ratio of

Yb/171Yb ratio of

Lu isotope and using a

176

Hf was corrected by assuming a

176

Hf was corrected by measuring interference-free

175

176

Lu/175Lu

Yb/173Yb

Hf/177Hf ratio of 0.282031 ± 0.000017 (2σ) (n=12). The calculations of εHf values

176

Hf/177Hf = 0.282785 and 176

176

Lu/177Hf = 0.0336

Hf/177Hf = 0.283225 and

Lu/177Hf = 0.038512 (Vervoort and Blichert-Toft, 1999) were adopted. The results of the Lu-

219 220 221

4.3 Whole-rock geochemistry

222

For whole-rock geochemical analyses, unweathered samples were crushed in a steel-jaw

223

crusher and then ground to powder with an agate mill. Major element compositions of eight

224

whole-rock samples were determined by X-ray fluorescence (XRF) spectrometry after lithium

225

metaborate/tetraborate fusion. Four analyses (A9I, A9K, C20, C22) were carried out the

226

Geoanalitica Core Research Center, Universidade de São Paulo, following the protocol

227

described in Mori et al. (1999) and four samples (A4, A5, A9B, C37) were analyzed at the ACME

228

Analytical Laboratories, Vancouver. For trace element analyses the powdered samples were

229

dissolved by acid (HF+HNO3) in Parr bombs for five days. Trace element concentrations of

230

seven samples were acquired by inductively coupled plasma mass spectrometry (ICP-MS) using

231

a Perkin Elmer Plasma Quadrupole MS Elan 6100DRC at the Geoanalitica Core Research

232

Center, Universidade de São Paulo (see Navarro et al., 2002, 2008 for further details). The

233

results of whole-rock geochemical analyses are shown in Online Supplementary Table S3.

234 235 236

4.4 Whole-rock Nd isotopes

237

Five whole-rock samples were selected for Nd isotopic analyses. The Nd isotopic

238

compositions of four samples (A9I, A9K, C20, C22) were determined using a Neptune multi-

239

collector ICP-MS at the Geochronological Research Center (CPGeo) of the Universidade de São

240

Paulo. The powdered samples were dissolved in acid (HF+HNO3) and the elements of interest

241

were separated in ion exchange columns following the protocol described in Sato et al. (1995).

242

During the period of analyses, the JNdi standard (Geological Survey of Japan; Tanaka et al.,

243

2000;

244

0.000005 (1σ). One sample (C37) was analyzed by thermal ionization mass spectrometry using

245

a Finnigan MAT 262 at the Geochronological Laboratory of the Universidade de Brasília,

246

following the protocol described by Gioia and Pimentel (2000). Uncertainties for

247

are assumed to be better than ±0.005% based on repeated analyses of the USGS standards

248

BHVO-1 and BCR-1. The Nd isotopic ratios of all analyzed samples were mass bias corrected

249

assuming a

250

concentrations determined by ICP-MS. The results of the Nd isotopic analyses are shown in

251

Online Supplementary Table S4.

143

Nd/144Nd = 0.512115 ± 0.000007) yielded an average 143Nd/144Nd value of 0.512097 ±

146

Nd/144Nd ratio of 0.7219. The

147

143

Nd/144Nd

Sm/ 144Nd ratios were calculated from the

252 253

5. Results

254

5.1 Whole-rock geochemistry

255

Regarding major element concentrations, tonalitic samples from the Amparo Complex

256

have the lowest SiO2 contents among the analyzed samples, ranging from 63 to 68 wt% (Fig. 4).

257

The granodioritic samples from the Amparo and Serra Negra complexes and the trondhjemitic

258

sample from the Heliodora-Minduri Complex have SiO2 contents that are intermediate

259

between the tonalitic and granitic samples, ranging from 70 to 73 wt%. The Amparo Complex

260

granitic sample shows the highest SiO2 concentration of 74 wt%. There is a well-defined

261

negative correlation between increasing SiO2 and CaO, MgO, FeOt, TiO2 contents (Fig. 4). In

262

general, Na2O and K2O contents show no variation with increasing SiO2. The exception is the

263

granitic sample from the Amparo Complex that shows a considerably higher K2O concentration

264

than the other samples, clearly falling off the main trend (Fig. 4). As shown in the diagrams of

265

Fig. 5, all samples can be classified as TTGs, with the exception of sample C22 from the Amparo

266

Complex, which plots on the biotite granite field.

267

All analyzed samples, with the exception of the Amparo Complex granitic sample (C22),

268

have trace element concentrations and patterns that are typical of TTGs, including negative Nb

269

and Ti anomalies in primitive-mantle normalized diagrams and well-defined correlation

270

between the Nb, Y and heavy rare earth elements (HREE) concentrations (Fig. 6). The Amparo

271

Complex tonalitic samples have the lowest light rare earth elements (LREE) concentrations and

272

La/YbN ratios among the analyzed samples (Figs. 6a, b). These samples show LREE

273

concentrations that are similar to the high-pressure TTGs of Moyen and Martin (2012) (Fig. 6a,

274

b). However, they also show high Y and HREE concentrations and slightly negative Eu

275

anomalies in chondrite-normalized REE patterns, which are typical features of low-pressure

276

TTGs (Fig. 6b). The Amparo and Serra Negra complexes granodioritic samples have trace

277

element concentrations and patterns that are similar to the average low-pressure TTG of

278

Moyen and Martin (2012) (Figs. 6c, d). The main differences between these samples and the

279

average low-pressure TTG are their higher average Y and HREE concentrations and more

280

pronounced negative Eu anomaly in chondrite-normalized REE patterns (Fig. 6d). The

281

Heliodora-Minduri Complex trondjhemitic sample has trace element concentrations and

282

patterns that are very similar to the high-pressure TTGs of Moyen and Martin (2012) (Figs. 6e,

283

f). This sample shows strongly negative Nb and slightly positive Sr anomalies in primitive

284

mantle-normalized multi-element diagrams. It also has low Y and HREE concentrations and a

285

strongly fractionated chondrite-normalized REE pattern with a slightly positive Eu anomaly

286

(Fig. 6f). The Amparo Complex granitic sample differs from the TTG-type samples in its much

287

higher Rb, Ba, Th and LREE concentrations (Figs. 6g, h). This sample is also characterized by low

288

HREE and Y concentrations, resulting in trace element patterns that are more fractionated

289

than the average late-Archean granite of Laurent et al. (2014) (Figs. 6g, h).

290

Figure 7 shows binary trace element diagrams with fields of the three different TTG

291

groups and the potassic group from Moyen (2011). Because of the low Sr concentrations and

292

high Y and Yb contents, the Amparo Complex tonalitic and granodioritic samples and the Serra

293

Negra granodioritic sample plot within or close to the low-pressure TTG field. The Heliodora

294

Complex trondhjemitic sample has a high Sr concentration and low Y and Yb contents and thus

295

falls in the high-pressure TTG field. The Amparo Complex granitic sample has low Sr, Y and Yb

296

concentrations and high La/Yb ratio and tends to plot close to the potassic field.

297 298

5.2 Zircon U-Pb geochronology and Hf isotopes

299

5.2.1 Sample A9I (tonalitic migmatite) (Amparo Complex)

300

Sample A9I is a medium-grained, dark-gray migmatitic biotite-hornblende orthogneiss of

301

tonalitic composition (Fig. 3c). The sample corresponds to the bulk migmatite including

302

unsegregated neosome and small-scale (up to 5mm thick) leucosomes. Zircons show a large

303

variety of external morphologies from stubby to elongated with aspect ratios from 1.5:1 to 4:1.

304

Most of the analyzed zircons display well-defined oscillatory zoned cores and narrow (5-20

305

µm) bright rims visible in CL images. Some stubby grains also show sector zoning (Fig. 8a). The

306

Th/U are generally high and vary from 0.14 to 0.73. Seventy U-Pb spots were analyzed and

307

excluding four analyses that show a common lead component, the remaining sixty-six spots

308

define a discordia line with an upper intercept at 3002.4 ± 9.7 Ma (MSWD = 1.2) (Fig. 9a). The

309

similar internal structures, Th/U ratios and dates demonstrate the existence of one single

310

zircon population that is interpreted to be related to the crystallization of the igneous

311

protolith. No hafnium isotope data is provided for this sample.

312 313

5.2.2 Sample A9K (tonalitic migmatite) (Amparo Complex)

314

Sample A9K is a fine- to medium-grained gray migmatitic biotite orthogneiss with a small

315

amount of hornblende (ca. 2-3 vol. %). The analyzed sample corresponds to the bulk migmatite

316

including unsegregated neosome and small-scale (up to 5mm thick) leucosomes. Zircon show

317

various morphologies from stubby to elongated, with aspect ratios from 1.5:1 to 4:1, but show

318

simple internal structures with well-defined oscillatory zoning and narrow (1-10µm) bright rims

319

in CL images (Fig. 8b). In general, Th/U ratios vary from 0.21 to 0.66 and only two discordant

320

spots yield lower values of 0.06 and 0.11. Excluding one U-Pb analysis, that show a different

321

lead-loss trend, the remaining fifty-nine analyses define a discordia line with an upper

322

intercept at 3000.9 ± 8.7 Ma (MSWD=1.05) (Fig. 9b). The analyzed zircons yield

323

values that range from 0.28087 to 0.28100, corresponding to suprachondritic εHf(t) values from

324

+0.6 to +5.1 (Fig. 11). The similar internal structures, Th/U ratios, 176Hf/177Hf(t) ratios and dates

325

clearly demonstrate the existence of one single zircon population. Based on that, the obtained

326

upper intercept is interpreted as the crystallization age of the igneous protolith. These zircons

327

yield a weighted average εHf value at the time of intrusion of +3.5 ± 0.9. Two-stage model

328

ages, projected back from zircon crystallization ages assuming a typical mafic crust 176Lu/177Hf

329

value of 0.022, are between 3.0 and 3.4 Ga with an average of 3.20 ± 0.12 Ga.

176

Hf/177Hf(t)

330 331

5.2.3 Sample C20 (granodioritic orthogneiss) (Serra Negra Complex)

332

Sample C20 is a medium-grained biotite orthogneiss of granodioritic composition (Fig.

333

3d). Two distinct zircon morphological populations are recognized. The first is constituted by

334

ca. 100-150 µm long grains with aspect ratios of ca. 2:1 to 3:1 and the second comprises ca. 75

335

to 100 µm long, stubby grains with aspect ratios of ca. 1.5:1 to 2:1. Both populations display

336

oscillatory and/or sector zoned cores and narrow (5-10 µm) bright rims in CL images (Fig. 8c).

337

In general, the Th/U ratios of the first population vary from 0.11 to 0.95 with only two spots

338

yielding lower values. The U-Pb analyses of grains from the first population show a large

339

spread towards a Neoproterozoic date (lead-loss or mixing of age components) and forty-two

340

analyzed spots define a discordia line with an upper intercept at 2962 ± 11 Ma and a lower

341

intercept at 613 ± 13 Ma (MSWD=1.02) (Fig. 9c). Zircon grains from this population concordant

342

at the upper intercept yield

343

Amparo Complex tonalitic sample (A9K), ranging from 0.28087 to 0.28100, corresponding to

344

chondritic to suprachondritic εHf(t) values from 0.0 to +4.4 (Fig. 11). The five analyzed grains of

345

the second population show Th/U ratios between 0.33 and 0.56 and yield older

346

dates, with a weighted mean

347

grains also have lower

348

population, ranging from 0.28083 to 0.28085. The U-Pb dates and Hf isotope ratios clearly

349

demonstrate that this is a different zircon population, which is interpreted as reflecting zircon

350

inheritance. Based on that and on the typical magmatic Th/U ratios of the predominant zircon

351

population, the upper intercept at 2962 ± 11 Ma is interpreted as the crystallization age of the

352

igneous protolith. These magmatic zircons yield a weighted average εHf value at the time of

353

intrusion of +2.3 ± 1.6. Two-stage model ages, projected back from zircon crystallization ages

354

assuming a typical mafic crust 176Lu/177Hf value of 0.022, are between 3.1 and 3.5 Ga with an

355

average of 3.31 ± 0.16 Ga.

176

176

Hf/177Hf(t) values very similar to the ones obtained from the

207

207

Pb/206Pb

Pb/206Pb age of 3190 ± 14 Ma (MSWD=0.36) (Fig. 9c). These

Hf/177Hf(t) values than those obtained from the main zircon

356 357

5.2.4 Sample C37 (trondhjemitic orthogneiss) (Heliodora Complex)

358

Sample C37 is a fine- to medium-grained leucocratic biotite orthogneiss of trondhjemitic

359

composition (Fig. 3e). The zircon grains from this sample have variable external morphologies

360

from oval to elongated grains, with aspect ratios from 2:1 to 4:1. In CL images all analyzed

361

zircon grains show oscillatory zoned cores and narrow (<5 µm) bright rims (Fig. 8d). Th/U ratios

362

vary, in general, from 0.06 to 0.68 with only one discordant spot yielding a lower value of 0.02.

363

Seventy-three zircon U-Pb spots were analyzed. Excluding nine spots that have a common-lead

364

component or a different trend of lead-loss, the remaining sixty-four analyses define a

365

discordia line with an upper intercept at 2957 ± 14 Ma (MSWD=1.2) (Fig. 9d). These zircons

366

show

367

0.28101 that correspond to εHf(t) values from +0.2 to +4.6. The similar internal structures,

368

dates, Th/U ratios and

369

zircon population. Thus, the obtained upper intercept at 2957 ± 14 Ma is interpreted as the

370

crystallization age of the igneous protolith. This magmatic zircon population yields a weighted

371

average εHf value at the time of intrusion of +2.6 ± 1.0. Two-stage model ages, projected back

372

from zircon crystallization ages assuming a typical mafic crust

373

between 3.1 and 3.5 Ga with an average of 3.26 ± 0.12 Ga.

374

176

Hf/177Hf(t) values that are similar to the other TTG samples ranging from 0.28089 to

176

Hf/177Hf(t) ratios are strong evidence of the existence of one single

176

Lu/177Hf value of 0.022, are

375

5.2.5 Sample C22 (granitic migmatite) (Amparo Complex)

376

Sample C22 is a fine- to medium-grained pinkish-gray migmatitic biotite-chlorite

377

orthogneiss of granitic composition (Fig. 3b). Zircon grains from this sample are subhedral ca.

378

100-175 µm long with aspect ratios from ca. 2:1 to 3:1. Most of the grains have oscillatory-

379

zoned CL-bright cores and CL-dark rims (Fig. 8e). The least discordant analyses on these bright

380

cores yield a weighted mean

381

This date is within uncertainty identical to the concordia age of 2765 ± 15 Ma (2σ) (MSWD=1.2;

382

Probability=0.24) obtained using the seven least discordant analyses on these CL-bright cores.

383

These bright cores yield

384

values from -6.8 to +2.6. Despite of the large spread on

385

structures and U-Pb dates clearly demonstrate that all these cores belong to the same zircon

386

population. Thus, the obtained Neoarchean date is interpreted as the crystallization age of the

387

igneous protolith. One spot on a CL-dark rim and two spots on CL-dark cores yield slightly

388

discordant Paleoproterozoic dates with a weighted mean

389

(n=3; MSWD=0.15) (Figs. 8e and 10). The Th/U ratios of these Paleoproterozoic domains (Th/U

390

= 0.11-0.14) are lower than those from the CL-bright Archean domains (Th/U = 0.43-1.37).

391

While the dataset from the Paleoproterozoic zircon domains is very small, the relatively low

392

discordance of these analyses associated with different internal textures and Th/U ratios than

393

those from Archean domains suggest recrystallization of zircon around ca. 2.0 Ga. Few

394

analyses spread between the two populations can be the result of either lead-loss or mixing of

395

different domains.

176

207

Pb/206Pb date of 2759 ± 13 Ma (n=21; MSWD=1.16) (Fig. 10).

Hf/177Hf(t) ratios from 0.28080 to 0.28105, corresponding to εHf(t) 176

207

Hf/177Hf(t) ratios, the internal

Pb/206Pb date of 2028 ± 33 Ma

396 397

5.3 Whole-rock Nd isotopes

398

Whole-rock Sm-Nd analyses were performed in five samples (A9K, A9I, C20, C22 and

399

C37). Four of these samples are Mesoarchean (2.96 - 3.00 Ga) orthogneisses with TTG affinities

400

from the Amparo, Serra Negra and Heliodora-Minduri complexes (samples A9K, A9I, C20 and

401

C37). These Mesoarchean samples, with only slightly negative to moderately positive εHf(t)

402

values, yield slightly negative εNd(t) values between -1.2 to -2.5. Nd model ages based on the

403

depleted mantle model of DePaolo (1981) range from 3.2 to 3.4 Ga (Fig. 12). The Neoarchean

404

(2.76 Ga) granitic sample of the Amparo Complex yields a εNd(t) value of -2.5 that overlaps the

405

Nd evolution lines of the Amparo Complex tonalitic samples (Fig. 12).

406 407

6. Discussion

408

6.1. Petrogenesis of the Mesoarchean TTG suites: implications for tectonic settings of

409

Archean felsic crust generation

410

The current consensus in the literature is that TTG magmas were generated from

411

hydrous basaltic sources (e.g. Moyen and Martin, 2012; Martin et al., 2014; Laurent et al.,

412

2014). These magmas were formed either by partial melting of basaltic rocks (e.g. Foley et al.,

413

2002; Rapp et al., 2003; Moyen and Stevens, 2006) or fractional crystallization of basaltic

414

magmas (e.g. Jagoutz et al., 2013). The TTG series is rather diverse, especially regarding Sr,

415

HREE and Y concentrations. Experimental studies (e.g. Moyen and Stevens, 2006) have shown

416

that partial melting of basaltic compositions at pressures around 10-12kbar generates a

417

residuum with amphibole-plagioclase-pyroxene and little garnet. At higher pressures (ca.

418

20kbar), the residuum is eclogitic with large amounts of garnet and no plagioclase. Based on

419

these results, several studies (e.g. Halla et al., 2009; Almeida et al., 2011; Moyen, 2011; Moyen

420

and Martin, 2012) have interpreted the diversity of the TTG series as a result of melt

421

generation at variable depths and Moyen (2011) proposed the existence of three TTG

422

subseries: the low-, medium- and high-pressure subseries.

423

The Mesoarchean TTG samples from the Amparo and Serra Negra complexes have

424

crystallization ages of 3.00 and 2.96 Ga, respectively, and geochemical signatures that

425

resemble those of low-pressure TTGs of Moyen (2011) (Figs. 6 and 7). These samples are

426

characterized by high HREE and Y contents and Sr values lower than 400 ppm. They also show

427

moderately negative Eu anomalies (Eu/Eu*=0.60-0.86) (Fig. 6). All these features are evidence

428

of melt generation from garnet-poor and plagioclase-rich sources and indicate melt generation

429

at maximum pressures of ca. 10-12kbar (Moyen and Stevens, 2006; Moyen, 2011; Moyen and

430

Martin, 2012). On the other hand, sample C37 from the Heliodora-Minduri Complex, which

431

also has a crystallization age of 2.96 Ga, shows geochemical signatures typical of high-pressure

432

TTGs with low HREE and high Sr contents (Fig. 6 and 7). This sample displays a highly

433

fractionated REE pattern (La/YbN = 35) and a positive Eu anomaly (Eu/Eu*=1.07) (Fig. 6). These

434

characteristics suggest a garnet-rich eclogitic source, what implies melt generation at

435

pressures of at least ca. 20 kbar (e.g. Moyen and Stevens, 2006; Moyen, 2011; Moyen and

436

Martin, 2012). Therefore, the present study defines a temporally and spatially related

437

occurrence of low- and high-pressure Mesoarchean TTGs, a fact that has important

438

implications for the tectonic settings of Archean continental crust generation.

439

As pointed out by Moyen (2011), many studies on the origins of TTGs have proposed

440

one unique setting for Archean continental crust generation, being either subduction-related

441

(e.g. Foley et al., 2002; Rapp et al., 2003) or intra-plate (e.g. Smithies, 2000; Bédard, 2006).

442

However, simultaneous and spatially related occurrences of low- and high-pressure TTGs, as

443

described here, requires models that can account for TTG magma generation at different

444

pressures/depths at the same time. These requirements are most likely met by hybrid models,

445

such as the ones proposed by Halla et al. (2009) and Almeida et al. (2011) to explain the coeval

446

occurrence of low- and high-pressure TTGs in the Fennoscandian Shield and the Amazon

447

Craton, respectively. In the model of Almeida et al. (2011), hot subduction underneath an

448

oceanic plateau could explain the simultaneous generation of low- and high-pressure TTGs in

449

the Carajás province. The low-pressure TTGs were most likely generated by partial melting at

450

the base of a thickened oceanic crust whereas the high-pressure TTGs were derived from slab

451

melting. Even though we cannot precisely discriminate the tectonic settings of the studied TTG

452

suites based on the presented data, our study indicates simultaneous melt generation at

453

variable depths, most likely in different tectonic settings, and strongly supports the hypothesis

454

of a non-unique setting of Archean felsic crust generation.

455

A combination of geochemical, geochronological and isotopic data allows the definition

456

of a simplified petrogenetic model for the studied TTG samples. This model starts with mafic

457

crust generation from mantle sources between ca. 3.4 and 3.2 Ga. This mafic, most likely

458

basaltic crust, was then partially melted at variable pressures (ca. 10 to 20 kbar) at ca. 3.00-

459

2.96 Ga. These partial melting processes most likely took place at the base of oceanic plateaus

460

and in intermittent subduction zones, generating the low- and high-pressure TTG suites,

461

respectively. The inherited zircon grains found in the Serra Negra Complex sample could

462

indicate the existence of an older TTG crust with age of ca. 3.19 Ga, that was assimilated

463

during the 3.00-2.96 Ga event. However, we cannot define the source of these inherited zircon

464

grains based only on the ages and Hf isotopic data, because both mafic and TTG crust tend to

465

show juvenile signatures. These zircons could also have been generated from mafic sources

466

and assimilated by the TTG magmas. If that is the case, the 3.19 Ga date would represent the

467

mafic crust generation age, in agreement with the Nd and Hf isotopic data. Therefore, without

468

additional information about the sources of these inherited zircons (i.e. zircon trace element

469

compositions), a distinction between these two hypothetical scenarios is not feasible at

470

present.

471 472

6.2 Neoarchean granitic magmatism: insights into late-Archean geodynamic changes

473

A late-Archean transition from TTGs to high-K granitoids has been recognized in several

474

cratonic areas around the globe, including the Amazonian, Dharwar, Kaapvaal, North China,

475

Pilbara, and São Francisco cratons, among others (e.g. Almeida et al., 2013; Romano et al.,

476

2013; Laurent et al., 2014; Farina et al., 2015). As pointed out by Laurent et al. (2014), this

477

transition has occurred at different times in different cratonic areas (from ca. 2.95 Ga in the

478

Pilbara craton to ca. 2.55 Ga in the Dharwar and North China cratons). The Neoarchean

479

granites usually show geochemical signatures similar to those of Phanerozoic orogenic

480

granites, a fact that led Sylvester (1994) to the conclusion that these rocks were most likely a

481

result of crustal differentiation in collisional settings. In accordance with Sylvester´s

482

hypothesis, Laurent et al. (2014) have interpreted the late-Archean granites as a result of

483

crustal recycling processes and attributed the late-Archean geodynamic changes to the

484

beginning of “modern-style” plate tectonics.

485

The presented data clearly show the existence of a late-Archean transition from TTG-

486

type to high-K granitic magmatism in the studied Archean complexes. As described earlier and

487

shown in Figs. 4 and 5, the Amparo Complex granitic sample, with a crystallization age of ca.

488

2.76 Ga, has a composition that is distinct from those of the Mesoarchean TTGs and similar to

489

worldwide late-Archean granites. Fetter et al., (2001) and Santos (2014) obtained similar

490

igneous crystallization ages of ca. 2.77 Ga from orthogneisses of the Amparo and Heliodora-

491

Minduri complexes, respectively. We cannot determine the duration of this granitic magmatic

492

event based on these few samples and therefore we cannot precisely determine the position

493

of this transition in time. However, based on worldwide datasets of late-Archean granites

494

(Laurent et al., 2014 and references therein) we can assume that this granitic magmatic event

495

most likely took place during a period shorter than 0.15 Ga. Therefore, we conclude that the

496

transition from TTGs to high-K granitoids in the studied area happened between ca. 2.90 and

497

2.75 Ga.

498

As discussed above, the major element signatures of the Amparo Complex granitic

499

sample are typical of late-Archean granites (Fig. 5). These granites have been described in

500

Archean cratons worldwide and in most of the cases interpreted as a result of partial melting

501

of an older felsic crust composed of TTGs and metasedimentary rocks (Laurent et al., 2014 and

502

references therein). Based on the A/CNK ratio of 1.02 of the Amparo Complex granitic sample

503

(C22), we can rule out the participation of large volumes of metapelitic sources and conclude

504

that this sample was most likely generated by reworking of TTGs ± metagraywackes. Regarding

505

trace element compositions, the Amparo Complex granitic sample is slightly enriched in Ba, Sr,

506

Th, LREE and strongly depleted in HREE and Y compared to the average late-Archean granite

507

from Laurent et al., 2014 (Fig. 6g, h). These trace element signatures cannot be explained by

508

simple batch melting of the TTG samples, alternatively they can be the result of garnet

509

retention in the source, indicating melting at higher pressure than those of the late-Archean

510

granites from Laurent et al. (2014). However, we have to take into account that this sample is a

511

strongly deformed migmatite (Fig. 3b) and therefore any interpretation based only on whole-

512

rock geochemistry is risky, as they could have been partially modified during metamorphism

513

and deformation.

514

The Amparo Complex granitic sample (C22) has a slightly lower εNd(t) value than the TTG

515

suites and overlaps the Nd evolution lines of the Amparo Complex TTG samples (Fig. 11). The

516

re-calculated zircon 176Hf/177Hf ratios of the granitic sample at 3.0 Ga, assuming a typical TTG

517

crust 176Lu/177Hf value of 0.0022 (Laurent and Zeh, 2015) from the crystallization age to 3.0 Ga,

518

yield a weighted average 176Hf/177Hf ratio of 0.280921 ± 0.000051. This value is within error the

519

weighted average

520

0.280941 ± 0.000045 obtained from the TTG samples from the Amparo, Heliodora and Serra

521

Negra complexes, respectively. The larger scatter in the granitic sample

522

be explained as a result of disequilibrium melting of the older TTG crust (e.g. Tang et al., 2014;

523

Laurent and Zeh, 2015). Thus, the whole-rock Nd and zircon Hf isotopic compositions suggest

524

that reworking of the Mesoarchean TTG crust was most likely a major mechanism for the

525

generation of the Neoarchean granitic suite.

176

Hf/177Hf(t) ratios of 0.280946 ± 0.000026, 0.280950 ± 0.000027 and 176

Hf/177Hf ratios can

526

The presented data support the notion that the Neoarchean is globally an important

527

period for geochemical and tectonic change (Laurent et al., 2014 and references therein). The

528

Mesoarchean (3.00-2.96 Ga) TTGs described here are interpreted to have been generated

529

from partial melting of oceanic crust and thus represent juvenile additions to the continental

530

crust. On the other hand, the Neoarchean (2.76 Ga) granitic sample has geochemical and

531

isotopic signatures typical for continental crust recycling and resemble those of collisional

532

granites. Therefore, the presented data support the hypothesis that the late-Archean

533

transition from TTGs to high-K granitoids is most likely a result of the beginning of continental

534

collision processes (e.g. Sylvester, 1994; Laurent et al., 2014). It is likely that before the late-

535

Archean the oceanic crust was thicker, hotter and more buoyant, thus not allowing the

536

coherent subduction necessary to drive “modern-style” plate tectonics and complete Wilson

537

cycles (e.g. Sizova et al, 2010; Moyen and van Hunen, 2012). During the late Archean, the

538

oceanic crust became colder and thinner, allowing long-lived subduction. Moreover, at this

539

time the continental blocks became large enough to collide with each other and undergo

540

significant continental crust differentiation (Laurent et al., 2014).

541 542

6.3 Regional Implications

543

Four main periods of magmatism have been recognized in the Archean crust of the

544

southern São Francisco craton (e.g. Teixeira et al., 2000; Lana et al., 2013; Farina et al., 2015):

545

(1) the Santa Barbara event (ca. 3230 – 3200 Ma), (2) the Rio das Velhas I event (ca. 2930 –

546

2850 Ma), (3) the Rio das Velhas II event (ca. 2800 – 2760 Ma) and (4) the Mamona event (ca.

547

2760 – 2680 Ma). Therefore, the Mesoarchean (ca. 2960 – 3000 Ma) igneous crystallization

548

ages presented in this study lie within a major gap of magmatic activity with respect to the

549

southern São Francisco craton (Fig. 13). We interpret this fact to indicate that these complexes

550

were exotic to the southern São Francisco craton Archean crust at least during the

551

Mesoarchean. The question that remains unanswered and that will be discussed in this section

552

is when these different Archean domains were welded to each other, becoming part of the

553

southern São Francisco paleocontinent.

554

Based on the geological setting and available data, two most likely hypotheses are

555

proposed in this section. The first is that these different domains have been accreted to each

556

other during the Neoarchean. This hypothesis is supported by the similar ages of high-K

557

granitoid magmatism in both domains. The Neoarchean granitic sample has almost the same

558

age as the beginning of the “Mamona” event (Fig. 13) that has been described as the transition

559

from low-K to high-K granitoid magmatism in the southern São Francisco craton (Romano et

560

al., 2013; Farina et al., 2015). This sample has geochemical and isotopic signatures that are

561

typical of continental crust recycling, a fact that reinforces the idea of a Neoarchean

562

continental collision. However, we have to take into account that the late-Archean transition

563

from TTGs to high-K granitoids was a global-scale event and that different cratonic areas, for

564

example the Dharwar and North China cratons, can share the same period of late-Archean

565

high-K magmatism (e.g. Laurent et al., 2014). Therefore, the similar ages of high-K magmatism

566

in both domains are not definitive arguments in favor of a Neoarchean accretion.

567

The other possible hypothesis is that these different Archean domains were accreted to

568

each other during the Paleoproterozoic. This hypothesis rests on the interpretation that the

569

Paleoproterozoic Pouso Alegre Complex is the orogenic counterpart of the Mineiro Belt arc

570

system (Fig. 1) (Cioffi et al., 2016). If this interpretation is correct, this at least 350 km long,

571

Paleoproterozoic arc-system would have been originally situated in-between the described

572

Archean complexes and the Archean crust of the southern São Francisco craton (Fig. 1). This

573

evidence together with the different ages of TTG magmatism in both domains could suggest

574

that the described Archean complexes were accreted to the southern São Francisco

575

paleocontinent after the development of the Pouso Alegre Complex / Mineiro Belt arc system

576

between 2.35 and 2.08 Ga. This is in agreement with the reworking age of ca. 2.03 Ga found in

577

the Amparo Complex granitic sample (C22) that would represent the timing of this

578

hypothetical accretion. The main problem with this hypothesis is that most of the original

579

tectonic scenario was overprinted by the Neoproterozoic orogenic events. These events were

580

responsible for major tectonic transport towards east-northeast, which is most likely the

581

reason why parts of these Archean complexes, especially in the central part of the orogen, are

582

located underneath the Paleoproterozoic Pouso Alegre Complex (Figs. 1 and 2). Therefore, any

583

interpretation based on the present-day geometric position is uncertain and without

584

additional information we cannot confirm this hypothesis. We highly encourage future studies

585

in the cratonic area where the original tectonic scenario is better preserved and the Mineiro

586

Belt clearly separates different Archean domains (Fig. 1). The presence of similar age patterns

587

in the cratonic area would be a strong argument in favor of a Paleoproterozoic accretion.

588 589

7. Conclusions

590

The data provided in this study lead us to the following conclusions:

591

- The Archean complexes in the basement of the southern Brasília Orogen show a well-defined

592

period of Mesoarchean TTG-type magmatism between 3.00 and 2.96 Ga.

593

- This Mesoarchean TTG magmatism is juvenile and was most likely a result of partial melting

594

of an older mafic crust extracted from the mantle between ca. 3.4 and 3.2 Ga.

595

- These partial melting processes took place at different depths at the same time, supporting

596

the hypothesis of non-unique tectonic settings of Archean continental crust generation.

597

- Neoarchean granitic magmatism at ca. 2.76 Ga records the transition from TTG-type to high-K

598

granitic magmatism in the studied area.

599

- This Neoarchean high-K magmatism shows less radiogenic isotopic signatures and is most

600

likely a result of reworking of the Mesoarchean TTG crust in a collisional setting.

601

- The reported Mesoarchean ages fall within a major gap of magmatic activity with respect to

602

the southern portion of the São Francisco craton. This suggests that the studied Archean

603

complexes were exotic to the São Francisco paleocontinent at least during the Mesoarchean.

604

However, the time when these different terranes became amalgamated is still an open

605

question.

606 607

Acknowledgements

608

This research was supported by FAPESP (grant 2013/13530-8). C.R. Cioffi is thankful to CAPES

609

and FAPESP (grants 2012/24933-3; 2014/05881-8) for the PhD scholarships. Rafael Bittencourt

610

Lima and Renato Moraes are acknowledged for their help during field work, Heather Shinogle

611

for assistance with SEM image acquisition and Vasco Loios for support during zircon

612

separation. The manuscript greatly benefited from insightful criticisms and suggestions by

613

Oscar Laurent and an anonymous reviewer. Editorial handling by Randall Parrish is

614

appreciated.

615

References

616

Almeida, J.A.C., Dall’Agnol, R., Oliveira, M.A., Macambira, M.J.B., Pimentel, M.M., Rämö, O.T.,

617

Guimarães, F.V., Leite, A.A.S., 2011. Zircon geochronology, geochemistry and origin of the TTG

618

suites of the Rio Maria granite-greenstone terrane: Implications for the growth of the Archean

619

crust of the Carajás Province, Brazil. Precambrian Research 187, 201-221.

620

Almeida, J.A.C., Dall’Agnol, R., Leite, A.A.S., 2013. Geochemistry and zircon geochronology of

621

the Archean granites suites of the Rio Maria granite-greenstone terrane, Carajás Province,

622

Brazil. Journal of South American Earth Sciences 42, 103-126.

623

Ávila, C.A., Teixeira, W., Cordani, U.G., Moura, C.A.V., Pereira, R.M., 2010. Rhyacian (2.23-2.20

624

Ga) juvenile accretion in the Southern São Francisco craton, Brazil: Geochemical and isotopic

625

evidence from the Serrinha magmatic suite, Mineiro belt. Journal of South American Earth

626

Sciences 29, 464-482.

627

Barbosa, N.S, Teixeira, W., Ávila, C.A., Montecinos, P.M., Bongiolo, E.M., 2015. 2.17 – 2.10 Ga

628

plutonic episodes in the Mineiro belt, São Francisco Craton, Brazil: U-Pb ages, geochemical

629

constraints and tectonics. Precambrian Research 270, 204-225.

630

Bédard, J.H., 2006. A catalytic delamination-driven model for coupled genesis of Archean crust

631

and sub-continental lithospheric mantle. Geochimica et Cosmochimica Acta 70, 1188-1214.

632

Belousova, E.A., Kostitsyn, Y.A., Griffin, W.L., Begg, G.C., O´Reilly, S.Y., Pearson, N.J., 2010. The

633

growth of continental crust: Constraints from zircon Hf-isotopes data. Lithos 119, 457-466.

634

Bleeker, W., 2003. The late Archean record: a puzzle in ca. 35 pieces. Lithos 71, 99-134.

635

Bouvier, A., Vervoort, J.D., Patchett, P.J., 2008. The Lu-Hf and Sm-Nd isotopic compositions of

636

CHUR: Constraints from unequilibrated chondrites and implications for the bulk composition of

637

terrestrial planets. Earth and Planetary Science Letters 273, 48-57.

638

Boynton, W. V. (1983). Cosmochemistry of the rare earth elements. Geochemistry of the rare

639

earth elements: meteorite studies. In: P. Henderson (Ed.), Rare Earth Element Geochemistry,

640

Elsevier (1984), pp. 63–114

641

Brito Neves, B.B., Campos Neto, M.C., Fuck, R.A., 1999. From Rodinia to western Gondwana:

642

an approach to the Brasiliano-pan African cycle and orogenic collage. Episodes 22, 155-166.

643

Campos Neto, M.C., 2000. Orogenic Systems from southwestern Gondwana: an approach to

644

Brasiliano-Pan African Cycle and orogenic collage in southeastern Brazil. In: Cordani, U.G.,

645

Milani, E.J., Thomaz Filho, A., Campos, D.A. (Eds.), Tectonic Evolution of South America. 31th

646

International Geological Congress. Rio de Janeiro, Brazil, pp. 335-365.

647

Campos Neto, M.C., Caby, R., 1999. Neoproterozoic high-pressure metamorphism and tectonic

648

constraint from the nappe system south of the São Francisco Craton, southeast Brazil.

649

Precambrian Research 97, 3-26.

650

Campos Neto, M.C., Caby, R., 2000. Lower crust extrusion and terrane accretion in the

651

Neoproterozoic nappes of southeast Brazil. Tectonics 19, 669-687.

652

Campos Neto, M.C., Cioffi, C.R., Moraes, R., Motta, R.G., Siga Jr., O., Basei, M.A.S., 2010.

653

Structural and metamorphic control on the exhumation of high-P granulites: The Carvalhos

654

Klippe example, from the oriental Andrelândia Nappe System, southern portion of the Brasília

655

Orogen, Brazil. Precambrian Research 180, 125-142.

656

Campos Neto, M.C., Basei, M.A.S., Janasi, V.A., Moraes, R., 2011. Orogen migration and

657

tectonic setting of the Andrelândia Nappe System: An Ediacaran western Gondwana collage,

658

south São Francisco craton. Journal of South American Earth Sciences 32, 393-406.

659

Cioffi, C.R., Campos Neto, M.C., Möller, A., Rocha, B.C., 2016. Paleoproterozoic continental

660

crust generation events at 2.15 and 2.08 Ga in the basement of the southern Brasília Orogen,

661

SE Brazil. Precambrian Research 275, 176-196.

662

Dardenne, M.A., 2000. The Brasília Fold Belt. In: Cordani, U.G., Milani, E.J., Thomaz Filho, A.,

663

Campos, D.A. (Eds.), Tectonic Evolution of South America. 31th International Geological

664

Congress. Rio de Janeiro, Brazil, pp. 231-263.

665

DePaolo, D.J., 1981. Neodymium isotopes in the Colorado Front Range and crust-mantle

666

evolution in the Proterozoic. Nature 291, 193-196.

667

de Wit, M.J., 1998. On Archean granites, greenstones, craton and tectonics: does the evidence

668

demand a verdict? Precambrian Research 91, 181-226.

669

Dhuime, B., Hawkesworth, C., Cawood, P.A., Storey, C.D., 2012. A change in the geodynamic of

670

Continental Growth 3 Billion Years Ago. Science 335, 1334-1336.

671

Ebert, H., 1968. Ocorrências de Fácies Granulíticas no Sul de Minas Gerais e áreas adjacentes.

672

Em dependências da estruturas orogênica: hipóteses sobre sua origem. Anais da Acadêmia

673

Brasileira de Ciências 40, 215-229.

674

Farina, F., Capucine, A., Lana, C., 2015. The Neoarchean transition between medium- and high-

675

K granitoids: Clues from the southern São Francisco Craton (Brazil). Precambrian Research 266,

676

375-394.

677

Fetter, A.H., Hackspacker, P.C., Ebert, H.D., Dantas, E.L., Costa, A.C.D., 2001. New Sm/Nd and

678

U/Pb geochronological constraints on the Archean to Neoproterozoic evolution of the Amparo

679

basement complex of the central Ribeira belt, southeastern Brazil. 3rd South American

680

Symposium on Isotope Geology (Extended Abstracts, CD-ROM).

681

Foley, S., Tiepolo, M., Vannucci, R., 2002. Growth of early continental crust controlled by

682

melting of amphibolite in subduction zones. Nature 417, 837-840.

683

Frost, B.R., Barnes, C.G., Collins, W.J., Arculus, R.J., Ellis, D.J., Frost, C.D., 2001. A geochemical

684

classification for granitic rocks. Journal of Petrology 42, 2033-2048.

685

Geng, Y., Du, L., Ren, L., 2012. Growth and reworking of the early Precambrian continental

686

crust in the North China Craton: Constraints from zircon Hf isotopes. Gondwana Research 21,

687

517-529.

688

Gioia, S.M.C.L., Pimentel, M.M., 2000. The Sm-Nd method in the geochronology laboratory of

689

the University of Brasília. Anais da Academia Brasileira de Ciências 72, 219-245.

690

Griffin, W.L., Belousova, E.A., Shee, S.R., Pearson, P.J., O’Reilly, S.Y., 2004. Archean crustal

691

evolution in the northern Yilgarn Craton: U-Pb and Hf-isotope evidence from detrital zircons.

692

Precambrian Research 131, 231-282.

693

Halla, J., van Hunen, J., Heilimo, E., Hölttä, P., 2009. Geochemical and numerical constraints on

694

Neoarchean plate tectonics. Precambrian Research 174, 155-162.

695

Jackson, S.E., Pearson, N.J., Griffin, W.L., Belousova, E.A., 2004. The application of laser

696

ablation-inductively coupled plasma-mass spectrometry to in-situ U-Pb zircon geochronology.

697

Chemical Geology 211, 47-69.

698

Jagoutz, O., Schimdt, M.W., Enggist, A., Burg, J.-P., Hamid, D., Hussain, S., 2013. TTG-type

699

plutonic rocks formed in a modern arc batholith by hydrous fractionation in the lower arc

700

crust. Contributions to mineralogy and petrology 166, 1099-118.

701

Janasi, V., 2002. Elemental and Sr-Nd isotope geochemistry of two Neoproterozoic mangerite

702

suites in SE Brazil: implications for the origin of the mangerite-charnockite-granite series.

703

Precambrian Research 119, 301-327.

704

Lana, C., Alkmim, F.F., Armstrong, R., Scholz, R., Romano, R., Nalini Jr., H.A., 2013. The ancestry

705

and magmatic evolution of the Archaean TTG rocks of the Quadrilátero Ferrífero province,

706

southeast Brazil. Precambrian Research 231, 157-173.

707 708

Laurent, O., Martin, H., Moyen, J.F., Doucelance, R., 2014. The diversity and evolution of late-

709

Archean granitoids: Evidence for the onset of “modern style” plate tectonics between 3.0 and

710

2.5 Ga. Lithos 205, 208-235.

711

Laurent, O., Zeh, A., 2015. A linear Hf isotope-age array despite different granitoid sources and

712

complex Archean geodynamics: Example from the Pietersburg block (South Africa). Earth and

713

Planetary Science Letters 430, 326-338.

714

Ludwig, K.R., 2003. Isoplot/Ex 3.00: A geochronological toolkit for Microsoft Excel. Berkeley

715

Geochronology Center Special Publication, 4.

716

Martin, H., Moyen J.-F., Guitreau, M., Blichert-Toft, J., Pennec J.-L., 2014. Why Archean TTG

717

cannot be generated by MORB melting in subduction zones. Lithos 198-199, 1-13.

718

McDonough, W.F., Sun, S.S., 1995. The composition of the earth. Chemical Geology 120, 223-

719

253.

720

Morais, S.N., 1999a. Programa Levantamentos Geológicos Básicos do Brasil: Integração

721

Geológica da Folha Campinas. (Escala) 1:250.000 SF-23-Y-A. Estados de São Paulo e Minas

722

Gerais (Nota Explicativa) – São Paulo – CPRM (26pp.).

723

Morais, S.N., 1999b. Programa Levantamentos Geológicos Básicos do Brasil: Integração

724

Geológica da Folha Guaratinguetá. (Escala) 1:250.000 SF-23-Y-B. Estados de São Paulo e Minas

725

Gerais (Nota Explicativa) – São Paulo – CPRM (28pp.).

726

Mori, P.E., Reeves, S., Correia, C.T., Haukka, M., 1999. Development of a fused glass disc XRF

727

facility and comparison with the pressed powder pellet technique at Instituto de Geociências,

728

Universidade de São Paulo. Revista brasileira de Geociências 29, 441-446.

729

Moyen J.-F., Stevens, G., 2006. Experimental constraints on TTG petrogenesis: Implications for

730

Archean Geodynamics. In: Benn, K., Mareschal, J.-C., Condie, K.C. (Eds.), Archean geodynamics

731

and environments. Monographs. AGU, pp. 149–178.

732

Moyen, J.-F., 2011. The composite Archean grey gneisses: Petrological significance, and

733

evidence for a non-unique tectonic setting for Archean crustal growth. Lithos 123, 21-36.

734

Moyen, J.-F., Martin, H., 2012. Forty years of TTG research. Lithos 148, 312-336.

735

Moyen, J.F., van Hunen, J., 2012. Short-term episodicity of Archaean plate tectonics. Geology

736

40, 451–454.

737

Navarro, M.S., Ulbrich, H.H., Andrade, S., Janasi, V.A., 2002. Adaptation of ICP-OES routine

738

determination techniques for the analysis of rare earth elements by chromatographic

739

separation in geological materials: Tests with reference materials and granitic rocks. Journal of

740

Alloy and Compounds 344, 40-45.

741

Navarro, M.S., Andrade, S., Ulbrich, H., Gomes, C.B., Girardi, V.A.V., 2008. The direct

742

determination of rare Earth elements in basaltic and related rocks using ICP-MS: Testing the

743

efficiency of microwave oven sample decomposition procedures. Geostandards and

744

Geoanalytical Research 32, 167-180.

745

O´Connor, J.T., 1965. A classification for quartz-rich igneous rocks based on feldspar ratios. U.S.

746

Geological Survey Professional Paper 525, 79-84.

747

Paton, C., Hellstrom, J., Paul, B., Woodhead, J., Hergt, J., 2011. Iolite:Freeware for the

748

visualisation and processing of mass spectrometry data. Journal of Analytical Atomic

749

Spectrometry 26, 2508-2518.

750

Perrota, M.M., 1991. A Faixa Alto Rio Grande na região de São Gonçalo do Sapucaí, MG.

751

Unpublished Master’s dissertation, IGc-USP, (158pp.).

752

Peternel, R., 2005. A zona de superposição entre as Faixas Brasília e Ribeira na região entre

753

Caxambu e Pedralva, sul de Minas Gerais. Unpublished PhD Thesis. Instituto de Geociências –

754

UFRJ, (257pp.).

755

Petrus, J.A., Kamber, B.S., 2012. VizualAge: A Novel Approach to Laser Ablation ICP-MS U-Pb

756

Geochronology Data Reduction. Geostandards and Geoanalytical Research 36, 247-270.

757

Rapp, R. P., Shimizu, N., Norman, M. D., 2003. Growth of early continental crust by partial

758

melting of eclogite. Nature 425, 605-609.

759

Roberts, N.MW., Spencer, C.J., 2015. The zircon archive of continental formation through time.

760

In: Roberts, N.M.W., Van Kranendonk, M., Parman, S., Shirey, S., Clift, P.D. (Eds.), Continent

761

Formation Through Time. Geological Society, London, Special Publications, 389, 197–225.

762

Rocha, B.C. 2011. Evolução metamórfica dos metassedimentos da Nappe Lima Duarte e rochas

763

associadas do Complexo Mantiqueira. Unpublished Master’s dissertation, IGc-USP, (201pp.).

764

Romano, R., Lana, C., Alkmim, F.F., Stevens, G., Armstrong, R., 2013. Stabilization of the

765

Southern portion of the São Francisco craton, SE Brazil, through a long-lived period of potassic

766

magmatism. Precambrian Research 224, 143-159.

767

Santos, C.A., 2014. Geologia, petrografia e geocronologia dos gnaisses e rochas associadas na

768

região entre Carrancas, Minduri e Luminárias (MG). Unpublished Master’s dissertation, IGc-

769

USP, (57pp.).

770

Sato, K., Tassinari, C.C.G., Kawashita, K., Petronilho, L., 1995. O método geocronológico Sm-Nd

771

no IG/USP e suas aplicações. Anais da Academia Brasileira de Ciências 67, 315-336.

772

Seixas, L.A.R., Bardintzeff, J-M., Stevenson, R., Bonin, B., 2013. Petrology of the high –Mg

773

tonalites and dioritic enclaves of the ca. 2130 Ma Alto Maranhão suite: Evidence for a major

774

juvenile crustal addition event during the Rhyacian orogenesis, Mineiro Belt, southeast Brazil.

775

Precambrian Research 238, 18-41.

776

Sizova, E., Gerya, T., Brown, M., Perchuk, L.L., 2010. Subduction styles in the Precambrian:

777

insight from numerical experiments. Lithos 116, 209–229.

778

Sláma, J., Košler, J., Condon, D.J., Crowley, J.L., Gerdes, A., Hanchar, J.M., Horstwood, S.A.,

779

Morris, G.A., Nasdala, L., Norberg, N., Schaltegger, U., Schoene, B., Tubrett, M.N., Whitehouse,

780

M.J., 2008. Plešovice zircon – A new natural reference material for U-Pb and Hf isotopic

781

microanalysis. Chemical Geology 249, 1-35.

782

Smithies, R.H., 2000. The Archaean tonalite-trondhjemite-granodiorite (TTG) series is not an

783

analogue of Cenozoic adakite. Earth and Planetary Science Letters 182, 115–125.

784

Söderlund, U., Patchett, P.J., Vervoort, J.D., Isachsen, C.E., 2004. The

785

determined by Lu-Hf and U-Pb isotope systematics of Precambrian mafic intrusions. Earth and

786

Planetary Sciences Letters 219, 311-324.

787

Stacey, J.S., Kramers, J.D., 1975. Approximation of terrestrial lead isotope evolution by two-

788

stage model. Earth and Planetary Science Letters 26, 207-221.

789

Sylvester, P.J., 1994. Archaean granite plutons. In: Condie, K.C. (Ed.), Archaean Crustal

790

Evolution. Developments in Precambrian Geology, vol. 11. Elsevier, Amsterdam, pp. 261–314.

791

Tanaka, T; Togashi, S., Kamioka, H., Amakawa, H., Kagami, H., Hamamoto, T., Yuhara, M.,

792

Orihashi, Y., Yoneda, S., Shimizu, H., Kunimaru, T., Takahashi, K., Yanagi, T., Nakano, T.,

793

Fujimaki, H., Shinjo, R., Asahara, Y., Tanimizu, M., Dragusanu, C., 2000. JNdi-1: a neodymium

794

isotopic reference in consistency with LaJolla neodymium. Chemical Geology 168, 279-281.

795

Tang, M., Wang, X-L., Shu, X-J., Wang, D., Yang, T., Gopon, P., 2014. Hafnium isotopic

796

heterogeneity in zircons from granitic rocks: Geochemical evaluation and modeling of “zircon

797

effect” in crustal anataxis. Earth and Planetary Science Letters 389, 188-199.

798

Tassinari, C.C.G., Nutman, A.P., 2001. Archean and Proterozoic multiple tectonothermal events

799

recorded by gneisses in the Amparo region, São Paulo state, Brazil. 3rd South American

800

Symposium on Isotope Geology (Extended Abstracts, CD-ROM).

801

Teixeira, W., Sabatè, P., Barbosa, J., Noce, C.M., Carneiro, M.A., 2000. Archean and

802

Paleoproterozoic evolution of the São Francisco Craton, Brazil. In: Cordani, U.G., Milani, E.J.,

803

Thomaz Filho, A., Campos, D.A. (Eds.), Tectonic Evolution of South America. 31th International

804

Geological Congress. Rio de Janeiro, Brazil, pp. 101-137.

805

Teixeira, W., Ávila, C.A., Dussin, I.A., Corrêa Neto, A.V., Bongiolo, E.M., Santos, J.O., Barbosa,

806

N.S., 2015. A juvenile accretion episode (2.35-2.32 Ga) in the Mineiro Belt and its role to the

807

Minas accretionary orogeny: Zircon U-Pb-Hf and geochemical evidences. Precambrian

808

Research 256, 148-169.

176

Lu decay constant

809

Trouw, R.A.J., Heilbron, M., Ribeiro, A., Paciullo, F., Valeriano, C.M., Almeida, J.C.H.,

810

Tupinambá, M., Andreis, R.R., 2000. The central segment of Ribeira belt. In: Cordani, U.G.,

811

Milani, E.J., Thomaz Filho, A., Campos, D.A. (Eds.), Tectonic Evolution of South America. 31th

812

International Geological Congress. Rio de Janeiro, Brazil, pp. 287-310.

813

Trouw, R.A.J., Nunes, R.P.M., Castro, E.M.O., Trouw, C.C., Matos, G.C., 2008. Nota explicativa

814

das Folhas Varginha (SF.23-V-D-VI) e Itajubá (SF.23-Y-B-III). Programa Geologia do Brasil.

815

Contrato CPRM-UFRJ N° 067/PR/05. (99pp).

816

Trouw, R.A.J., Peternel, R., Ribeiro, A., Heilbron, M., Vinagre, R., Duffles, P., Trouw, C.C.,

817

Fontainha, M., Kussama, H.H., 2013. A new interpretation for the interference zone between

818

the southern Brasília belt and the central Ribeira belt, SE Brazil. Journal of South American

819

Earth Sciences 48, 43-57.

820

Vervoort, J.D., Blichert-Toft, J., 1999. Evolution of the depleted mantle: Hf isotopic evidence

821

from juvenile rocks through time. Geochimica et Cosmochimica Acta 63, 533-556.

822

Westin, A., Campos Neto, M.C., 2013. Provenance and tectonic setting of the external nappe of

823

the Southern Brasília Orogen. Journal of South American Earth Sciences 48, 220-239.

824

Westin, A., Campos Neto, M.C., Hawkesworth, C., Cawood, P., Dhuime, B., Delavault, H., A.,

825

2016. Paleoproterozoic intra-arc basin associated with a juvenile source in the southern

826

Brasília Orogen: using U-Pb ages and Hf-Nd isotopic analyses in provenance studies of

827

complexes areas. Precambrian Research 276, 178-193.

828

Zeh, A., Gerdes, A., Barton, J.M., 2009. Archean accretion and crustal evolution of the Kalahari

829

Craton – the zircon age and hf isotope record of granitic rocks from Barbeton / Swaziland to

830

the Francistown Arc. Journal of Petrology 50, 933-966.

831 832 833

Figure captions

834

Figure 1. (a) Schematic reconstruction of part of western Gondwana with location of figure 1b.

835

1 - Phanerozoic basins; 2 – Proterozoic cover sequences; 3 – Neoproterozoic orogens; 4 –

836

Cratonic basement. (b) Tectonic map of the southern São Francisco craton and southern

837

Brasília Orogen with location of the studied area.

838

839

Figure 2. Geological map of the studied area (compiled and modified from Perrotta, 1991;

840

Morais, 1999a, b; Peternel, 2005; Trouw et al., 2008; Cioffi et al., 2016) with location of the

841

analyzed samples. Schematic cross-section from NW to SE is oriented along the line A-B.

842 843

Figure 3. (a) Layered migmatitic orthogneisses of the Amparo Complex with 1-5 cm thick

844

stromatic leucosomes and amphibolite layers parallel to the main foliation. Note the complex

845

folding patterns visible on the left side of the photo. (b) Dark-gray biotite-hornblende

846

migmatitic orthogneiss of tonalitic composition of the Amparo Complex (sample A9I) with up

847

to 2 cm large peritectic hornblende crystals within leucosomes. (c) Pinkish-gray migmatitic

848

biotite orthogneiss of granitic composition of the Amparo Complex (Sample C22). (d) Hand

849

specimen of the Serra Negra Complex granodioritic gneiss (sample C20). (e) Hand specimen of

850

the Heliodora-Minduri Complex trondhjemitic gneiss (sample C37).

851 852

Figure 4. Harker-type diagrams (SiO2 vs. major elements) for the analyzed samples.

853 854

Figure 5. (a) Na2O+K2O-CaO (MALI) vs. SiO2 diagram proposed by Frost et al. (2001) with the

855

TTG, sanukitoids and biotite granites fields from Laurent et al. (2014), (b) Al2O3 /

856

(CaO+Na2O+K2O) (molar) vs. K2O / Na2O diagram with TTG, sanukitoids and biotite granites

857

fields from Laurent et al. (2014), (c) Normative Ab-An-Or triangle (O´Connor, 1965) with the

858

TTG field of Moyen and Martin (2012), (d) Ternary classification diagram for late-Archean

859

granitoids proposed by Laurent et al., (2014): 2 x A/CNK (molar Al2O3 / CaO+Na2O+K2O); 2 x

860

(FeOt + MgO) wt% x (Sr+Ba) wt% (=FSMB).

861 862

Figure 6. Primitive mantle normalized multielement diagrams and chondrite normalized REE

863

patterns. (a,b) Amparo Complex tonalitic samples, (c,d) Amparo and Serra Negra complexes

864

granodioritic samples, (e,f) Heliodora-Minduri Complex trondhjemitic sample, (g,h) Amparo

865

Complex granitic sample. Primitive mantle values from McDonough and Sun (1995). Chondrite

866

values from Boynton (1983). Average compositions of the low- and high-pressure TTGs from

867

Moyen and Martin (2012). Average composition of the Biotite-, two-mica granites from

868

Laurent et al. (2014).

869

870

Figure 7. Binary trace element diagrams with fields of the three TTG groups and the potassic

871

group of Moyen (2011): high-pressure TTG group (HP TTG); medium-pressure TTG group (MP

872

TTG); low-pressure TTG group (LP TTG). (a) Sr vs. SiO2, (b) Ce / Sr vs. Y, (c) Sr / Y vs. Y; La / Yb vs.

873

Yb.

874 875

Figure 8. Representative cathodoluminescence (CL) images of zircon grains with analyzed spots

876

indicated by open circles (grain numbers within parenthesis). U-Pb results are shown as

877

207

Pb/206Pb dates, with 2σ errors. Lu-Hf analyses spots are indicated by dashed circles.

878 879

Figure 9. Concordia diagrams for zircon U-Pb LA-ICP-MS analyses. Error ellipses are 2σ.

880

Intercepts are quoted at 95% confidence level. (a) Sample A9I (Amparo Complex) – Upper

881

intercept at 3002.4 ± 9.7 Ma (n = 66; MSWD = 1.2), (b) Sample A9K (Amparo Complex) – Upper

882

intercept at 3000.9 ± 8.7 Ma (n = 59; MSWD = 1.05), (c) Sample C20 (Serra Negra Complex) –

883

Intercepts at 613 ± 13 Ma and 2962 ± 11 Ma (n = 42; MSWD = 1.02), (d) Sample C37

884

(Heliodora-Minduri Complex) – Upper intercept at 2957 ± 14 Ma (n = 64; MSWD = 1.2).

885 886

Figure 10. Concordia diagram for zircon U-Pb LA-ICP-MS analyses of sample C22 (Amparo

887

Complex). Error ellipses are 2σ. The main zircon population yields a weighted mean 207Pb/206Pb

888

date of 2759 ± 13 Ma (n = 21; MSWD = 1.16; Probability = 0.28). Three slightly discordant

889

grains yield a weighted mean 207Pb/206Pb date of 2028 ± 33 Ma (n=3; MSWD=0.15).

890 891

Figure 11. εHf(t) versus age diagram for all analyzed samples, plotted at the corresponding U-Pb

892

ages of each analyzed spot. Depleted mantle line calculated for the model proposed by

893

Vervoort and Blichert-Toft (1999).

894 895

Figure 12. Nd evolution diagram for analyzed samples. The isotopic evolution line is for the

896

DePaolo (1981) depleted mantle model.

897

898

Figure 13. U-Pb zircon ages of the Archean complexes in the basement of the southern Brasília

899

Orogen. Samples marked with asterisk are from Fetter et al. (2001). All remaining samples are

900

from this study. Age intervals of magmatic events in the southern São Francisco craton are

901

from Lana et al. (2013) and Farina et al. (2015) (RVI = Rio das Velhas I; RVII = Rio das Velhas II;

902

SB = Santa Barbara).

903 904

Table captions

905

Table 1. Data summary.

906 907

Online supplementary material captions

908

Online supplementary table S1. U-Pb LA-ICP-MS zircon data.

909

Online supplementary table S2. Lu-Hf LA-ICP-MS zircon data.

910

Online supplementary table S3. Whole-rock major (wt%) and trace element (ppm) data.

911

Online supplementary table S4. Sm-Nd whole-rock isotope data.

912

913 914

915 916

917 918

919 920

921 922

925 926

927 928

931 932

933 934

935 936

937 938 Samp le

Coordinate s

Rock Type

Geologi cal Unit

Geochemi cal parameter s SiO2 -

Igneous crystallizat ion age (Ma)

Dating echniq ue

εNd( t)

TD

avera ge εHf(t)

M

(G a)

avera ge

K2 O/Na2O

Hf mode l age (Ga)a A4

References

S 22°43'56.11 ''/W 46°45'51.56 " S 22°44'03.89 ''/W 46°45'54.25 " S 22°43'17.87 ''/W 46°46'56.95 " S 22°43'17.87 ''/W 46°46'56.95 " S 22°43'17.87 ''/W 46°46'56.95 " S 22°35'51.86 ''/W 46°41'11.43 " S 22°37'45.13 ''/W 46°43'19.43 " S 22°03'46.20 ''/W 45°27'46.57 "

tonalitic orthogneis s

Amparo Comple x

67 - 0.39

This stud y

granodiorit ic orthogneis s

Amparo Comple x

73 - 0.45

This stud y

granodiorit ic orthogneis s

Amparo Comple x

73 - 0.38

This stud y

tonalitic orthogneis s

Amparo Comple x

63 - 0.30

3002 ± 10

tonalitic orthogneis s

Amparo Comple x

68 - 0.30

3001 ± 9

granodiorit ic orthogneis s

Serra Negra Comple x

70 - 0.49

2962 ± 11

granitic orthogneis s

Amparo Comple x

74 - 1.29

2759 ± 13

throndhje mitic orthogneis s

72 - 0.22

2957 ± 14

H587

S 22°43.164'/ W 46°46.582'

H601

S 23°06.323'/ W 46°50.567'

throndhje mitic orthogneis s orthogneis s

Heliodor aMinduri Comple x Amparo Comple x

A5

A9B

A9I

A9K

C20

C22

C37

Amparo Comple x

zircon U-Pb (LAICPMS) zircon U-Pb (LAICPMS) zircon U-Pb (LAICPMS) zircon U-Pb (LAICPMS) zircon U-Pb (LAICPMS)

-1.2

3.4 1

-1.6

3.4 0

+3.5

3.14

This stud y

-2.5

3.3 9

+2.3

3.21

This stud y

-2.5

3.0 4

-2.8

3.36

This stud y

-1.8

3.2 2

+2.6

3.17

This stud y

3024 ± 9

zircon U-Pb (IDTIMS)

-1.2

3.2 8

[2]

2772 ± 26

zircon U-Pb (IDTIMS)

-1.5

3.0 2

[2]

This stud y

[2] Fetter et al., 2001

939 940 941 942 943

a

Two-stage model ages were projected back from crystallization ages assuming a mean crustal value for

176

Lu/

177

Hf = 0.015

944 Highlights 945 946 Mesoarchean TTG-type magmatism at 2.96-3.00 Ga. 947 Coeval occurrence of low- and high-pressure TTGs. 948 Neoarchean granitic magmatism at 2.76 Ga records the transition from TTGs to granites. 949 The granitic magmatism is most likely a result of reworking of a Mesoarchean TTG crust. 950 951 952 953 954