Fe isotopic fractionation during the magmatic–hydrothermal stage of granitic magmatism

Journal Pre-proof Fe isotopic fractionation during the magmatic–hydrothermal stage of granitic magmatism De-Hong Du, Weiqiang Li, Xiao-Lei Wang, Xu-Jie Shu, Tao Yang, Tao Sun PII:

S0024-4937(19)30424-4

DOI:

https://doi.org/10.1016/j.lithos.2019.105265

Reference:

LITHOS 105265

To appear in:

LITHOS

Received Date: 28 July 2019 Revised Date:

25 October 2019

Accepted Date: 25 October 2019

Please cite this article as: Du, D.-H., Li, W., Wang, X.-L., Shu, X.-J., Yang, T., Sun, T., Fe isotopic fractionation during the magmatic–hydrothermal stage of granitic magmatism, LITHOS, https:// doi.org/10.1016/j.lithos.2019.105265. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. 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. © 2019 Elsevier B.V. All rights reserved.

Abstract Felsic intrusive magmatic systems are more complex than volcanic systems as they involve prolonged magmatic processes in addition to potentially multiple phases of magmatic–hydrothermal activity within a single pluton. In particular, the competition between exsolved magmatic fluids and crystallizing minerals for elements and isotopes during the magmatic–hydrothermal stages of evolution of these systems remains unclear despite fluid exsolution being one of the key steps in ore formation. This

study

focuses

on

a

suite

of

representative

Late

Jurassic

W–Sn

mineralization-related granites within the Nanling Range of southern China and investigates the behavior of Fe isotopes during magmatic differentiation and magmatic–hydrothermal processes recorded within these granites. The granites in the study area have variable δ56Fe values (relative to IRMM-014) ranging from 0.12‰ ± 0.04‰ to 0.42‰ ± 0.01‰. These values do not correlate with bulk-rock geochemical indicators of fluid exsolution (e.g., Zr/Hf, Nb/Ta, K/Rb, and F/Cl ratios), qualitatively indicating that fluid exsolution has an insignificant effect on Fe isotopic fractionation. We further quantified the relative contributions of fluid exsolution and crystal fractionation on bulk-rock Fe isotopic variability using a Rayleigh fractionation model. This modeling indicates that fluid exsolution can only generate a <0.07‰ increase in bulk-rock δ56Fe values of the Nanling granites even in the most fluid-abundant scenario. This means that Fe isotopic fractionation is controlled mainly by fractional crystallization during the magmatic–hydrothermal stages of the evolution of granitic magmas. This further suggests that Fe isotopes could be used to study the magmatic differentiation of felsic magma chambers, including those within fluid-rich

mineralizing magmatic systems.

1

Fe isotopic fractionation during the

2

magmatic–hydrothermal stage of granitic magmatism

3 4

De-Hong Dua, Weiqiang Lia, Xiao-Lei Wanga,*, Xu-Jie Shua, b, Tao Yanga, Tao

5

Suna

6 7

a

8

Nanjing University, Nanjing 210046, China

9

b

China Geological Survey, Nanjing Center, Nanjing 210016, China

*

E-mail: [email protected] (XL Wang).

State Key Laboratory for Mineral Deposits Research, School of Earth Sciences and Engineering,

10 11

1

12

Abstract

13

Felsic intrusive magmatic systems are more complex than volcanic systems as they

14

involve prolonged magmatic processes in addition to potentially multiple phases of

15

magmatic–hydrothermal activity within a single pluton. In particular, the competition

16

between exsolved magmatic fluids and crystallizing minerals for elements and

17

isotopes during the magmatic–hydrothermal stages of evolution of these systems

18

remains unclear despite fluid exsolution being one of the key steps in ore formation.

19

This

20

mineralization-related granites within the Nanling Range of southern China and

21

investigates the behavior of Fe isotopes during magmatic differentiation and

22

magmatic–hydrothermal processes recorded within these granites. The granites in the

23

study area have variable δ56Fe values (relative to IRMM-014) ranging from 0.12‰ ±

24

0.04‰ to 0.42‰ ± 0.01‰. These values do not correlate with bulk-rock geochemical

25

indicators of fluid exsolution (e.g., Zr/Hf, Nb/Ta, K/Rb, and F/Cl ratios), qualitatively

26

indicating that fluid exsolution has an insignificant effect on Fe isotopic fractionation.

27

We further quantified the relative contributions of fluid exsolution and crystal

28

fractionation on bulk-rock Fe isotopic variability using a Rayleigh fractionation model.

29

This modeling indicates that fluid exsolution can only generate a <0.07‰ increase in

30

bulk-rock δ56Fe values of the Nanling granites even in the most fluid-abundant

31

scenario. This means that Fe isotopic fractionation is controlled mainly by fractional

32

crystallization during the magmatic–hydrothermal stages of the evolution of granitic

33

magmas. This further suggests that Fe isotopes could be used to study the magmatic

34

differentiation of felsic magma chambers, including those within fluid-rich

study

focuses

on

a

suite

of

2

representative

Late

Jurassic

W–Sn

35

mineralizing magmatic systems.

36 37

Keywords: Fe isotopes; magmatic–hydrothermal processes; granite; isotopic

38

fractionation; fractional crystallization.

39

3

40

1. Introduction

41

Basaltic to andesitic igneous rocks generally show small variations in Fe isotopic

42

compositions (e.g., Heimann et al., 2008; Teng et al., 2013), whereas high-silica (>70

43

wt.% SiO2) intrusive and volcanic rocks have highly variable Fe isotopic

44

compositions (~0.1‰–0.6‰ in δ56Fe where δ57Fe = 1.4750 × δ56Fe; e.g., Albarède

45

and Beard, 2004; Heimann et al., 2008; Telus et al., 2012; Zambardi et al., 2014;

46

Foden et al., 2015; Du et al., 2017; Xia et al., 2017). Granitoids can crystallize at very

47

low temperatures (470°C–560°C; Ackerson et al., 2018) and therefore can yield a

48

lengthy composite record of magmatic differentiation. Granites also represent an

49

advanced stage of the evolution of Earth’s continental crust (e.g., Lipman and

50

Bachmann, 2015; Wu et al., 2017), meaning that Fe isotopes can provide insights into

51

the compositional evolution of continental crust as well as the differentiation of felsic

52

magmas. However, the origin of Fe isotopic fractionation in granites remains

53

controversial. For example, Poitrasson and Freydier (2005) and Heimann et al. (2008)

54

suggested that the exsolution of isotopically light Fe-bearing fluids can explain the

55

heavy δ56Fe values of high-silica granites. In contrast, Sossi et al. (2012) and Foden et

56

al. (2015) suggested that mineral–melt equilibrium determines the evolution of Fe

57

isotopic compositions, indicating that the systematically heavy δ56Fe values of A-type

58

granites relative to I-type granites reflect the removal of mineral assemblages

59

crystallized under different oxidation states. Telus et al. (2012) also suggested that

60

fluid exsolution is responsible for Fe isotopic fractionation recorded in some

61

pegmatites and highly differentiated granites with high δ66Zn values, and discussed

62

the contribution of fractional crystallization to the δ56Fe variations observed in 4

63

granites with unfractionated δ66Zn values. In addition, thermal diffusion within the

64

magma chamber may explain the presence of samples with heavy Fe isotope

65

compositions within the Torres del Paine igneous complex and Cedar Butte volcanics

66

(Gajos et al., 2016; Zambardi et al., 2014). However, the expected positive

67

correlations between Fe and other metal stable isotopes (e.g., Li, Zn, Mg) caused by

68

thermal diffusion are not found in analyses of whole-rock granitic samples (e.g.,

69

Schuessler et al., 2009; Telus et al., 2012; Xia et al., 2017).

70

In terms of felsic volcanic rocks, Du et al. (2017) suggested that fractional

71

crystallization was the dominant control of Fe isotopic fractionation during the

72

evolution of associated magmas. However, the complexity of fluid exsolution in

73

intrusive rocks means that the effect of this process on Fe isotopic compositions is

74

still not well constrained. Fluid exsolution is a critical process during the

75

emplacement and evolution of granitic magmas related to ore deposits (e.g., Audétat

76

et al., 2008; Heinrich and Candela, 2014). Granitic plutons commonly undergo

77

protracted periods of fluid–melt interaction from the beginning of fluid saturation in a

78

crystallizing melt to final late-stage magmatic solidification, with these stages

79

collectively termed magmatic–hydrothermal processes (e.g., Candela and Holland,

80

1986; Shinohara, 1994; Audétat et al., 2000). Fluid exsolution from magmas can be

81

triggered by a pressure drop during magma ascent and emplacement (first boiling)

82

and/or mineral crystallization during magma chamber evolution (second boiling; e.g.,

83

Burnham and Ohmoto, 1980; Baker and Alletti, 2012). Volatile exsolution from

84

emplaced plutons is generally controlled by isobaric crystallization (Candela, 1997; 5

85

Audétat et al., 2000). As crystallization proceeds, the H2O content of the residual melt

86

increases because the amount of aqueous fluid stored in the hydrous minerals being

87

precipitated by the magma is very low (<0.6 wt.% H2O; Caricchi and Blundy, 2015).

88

Fluid saturation is followed by the migration of iron from the melt into the exsolved

89

fluid and crystalline minerals, leading to competition for Fe between these two phases.

90

The analysis of fluid and melt inclusions within minerals from felsic intrusive rocks

91

provides evidence of preferential partitioning of Fe into fluids (Davidson and

92

Kamenetsky, 2007; Zajacz et al., 2008). Wawryk and Foden (2015) investigated the

93

Fe isotopic fractionation recorded within the Renison Sn–W deposit and identified a

94

close relationship between the Fe isotopic compositions of cooling plutons and

95

associated hydrothermal sulfide ores. This suggests that the Fe isotopic compositions

96

of granitic magmas during magmatic–hydrothermal processes may be affected by the

97

removal of Fe into fluids, especially within granites that contain low concentrations of

98

residual Fe. However, despite the large body of published research on Fe isotopes in

99

various igneous systems, details of the relative contributions of the two competing

100

processes that control Fe isotopic fractionation during the magmatic–hydrothermal

101

stages of the evolution of these systems remain unclear.

102

Highly differentiated granites were emplaced within the Nanling Range of

103

southern China during the Late Jurassic. A large number of these granites are

104

mineralized as a result of large-scale magmatic–hydrothermal processes (e.g., Shu et

105

al., 2011; Chen et al., 2013; Wu et al., 2017). These granites provide an excellent

106

opportunity to investigate in detail the behavior of Fe isotopes during 6

107

magmatic–hydrothermal processes. Here, we present new Fe isotopic data for these

108

granites as well as major- and trace-element geochemical data and results of

109

quantitative numerical modeling. The new data presented here indicate that fluid

110

exsolution is not a major control on Fe isotopic variations, even within late-stage

111

magmatic–hydrothermal granitic systems.

112 113

2. Sample descriptions

114

The samples analyzed during this study were collected from the western part of

115

the Nanling Range, one of the most important areas of W–Sn mineralization in

116

southern China (Fig. 1). This area contains voluminous Mesozoic (especially Late

117

Jurassic) granitoids associated with extensive W–Sn mineralization (Zhou et al., 2006;

118

Chen et al., 2013). A total of 17 samples were obtained from 4 representative plutons

119

in the western part of the Nanling Range, including the Qitianling (155 ± 2 to 152 ± 1

120

Ma), Qianlishan (155 ± 2 to 150.9 ± 5.4 Ma), Laiziling (151.8 ± 2.5), and

121

Yaogangxian (152 ± 2 Ma) plutons (Shu et al., 2011; Shu, et al., 2013; Wang et al.,

122

2014). All of these plutons are associated with W–Sn and other base-metal (e.g., Pb,

123

Zn, Mo, and Bi) mineralization (Fig. 1; Chen et al., 2013). These granites are

124

classified as aluminous A2-type biotite granites and are thought to have formed from

125

magmas generated by partial melting of Neoproterozoic crust with minor

126

contributions from mantle-derived material (e.g., Shu et al., 2011; Zhu et al., 2011;

127

Zhao et al., 2012; Chen et al., 2014; Li et al., 2018). They have similar Nd isotope

128

ratios (–7.6 < εNd(t) < –5.5 for the Qitianling granites, Zhao et al., 2012; –7.59 < εNd(t) 7

129

< –6.35 for the Qianlishan granites, Mao et al., 1995; –6.3 < εNd(t) < –6.0 for the

130

Laiziling granites, Li et al., 2018). Although the Qitianling granites are generally less

131

evolved than the granites within the other three plutons, they all show similar

132

geochemical trends (e.g., major and trace elements; Cao et al., 2018). This suggests

133

that these more evolved granites might be derived from the magmas that formed the

134

Qitianling granite or alternatively from source protoliths that are compositionally

135

similar to the Qitianling granite (Zhu et al., 2011).

136

The samples analyzed during this study were fresh and free of mineralization and

137

associated alteration. The mineral assemblages within these plutons are dominated by

138

quartz, plagioclase, K-feldspar, and biotite, with minor muscovite and amphibole

139

(Table 1) and accessory magnetite, zircon, apatite, and ilmenite. The samples used for

140

Fe isotopic analysis have variable SiO2 (66–77 wt.%), Li (15–893 ppm), Rb

141

(250–1633 ppm), F (520–9420 ppm), Ba (5–541 ppm), Sr (4.5–181.6 ppm), Zr

142

(34.9–228.9 ppm), and Cl (60–410 ppm) concentrations (Table S1 in the

143

supplementary material; Shu, 2014). These samples also have very variable Zr/Hf

144

(37–11), Nb/Ta (10–1), and K/Rb (189–26) ratios, and some samples from the

145

Qianlishan and Laiziling plutons have rare-earth-element (REE) patterns that contain

146

tetrad effects (Shu, 2014; Table S1). These features are indicative of evolution from a

147

purely

148

magmatic–hydrothermal stage dominated by crystal–melt–fluid interaction (e.g., Zhao

149

et al., 2011; Zhu et al., 2011; Chen et al., 2014). This transition is also recorded by

150

zircons and inclusions within quartz in these plutons. Previous research indicated that

magmatic

system

dominated

8

by

crystal–melt

interaction

to

a

151

zircons within these Late Jurassic granites generally appear dark during

152

cathodoluminescence (CL) imaging although some have cores that appear bright

153

during CL. The dark-CL zircons or zircon domains are thought to have formed during

154

the later magmatic fluid-enriched stages of magmatism (Shu et al., 2011; Wang et al.,

155

2014). In addition, Shan et al. (2011) determined that the cores of individual quartz

156

crystals contain only silicate melt inclusions whereas the mantles of the same grains

157

contain co-existing melt and fluid inclusions, suggesting that these crystals record

158

fluid saturation during the later stages of magma-chamber fractional crystallization.

159

Therefore, the large variations in elemental and Fe isotopic compositions within these

160

granites are most likely to represent a complete history of the magmatic and

161

magmatic–hydrothermal evolution of the magmas that formed these intrusions.

162 163

3. Analytical techniques

164

The Fe isotopic analysis of whole-rock samples was undertaken at the State Key

165

Laboratory for Mineral Deposits Research, Nanjing University, China, using the

166

chemical purification and analytical procedures outlined by Du et al. (2017) and Ye et

167

al. (2017). In brief, ~60 mg of sample powder was digested in a 2:1:1 mixture of

168

concentrated HCl–HNO3–HF before Fe was separated from matrix elements using

169

ion-exchange chemistry with a chloride-form AG MP-1 anion-exchange resin. The

170

resulting purified Fe solutions were analyzed using multicollector–inductively

171

coupled plasma–mass spectrometry (MC–ICP–MS) employing a Thermo Fisher

172

Scientific Neptune Plus instrument with moderate resolution and used in the 9

173

wet-plasma mode. The resulting sensitivity was 6–10 V/ppm for 56Fe, with a 2 ppm Fe

174

solution being used during analysis. A standard–sample–standard bracketing method

175

was used to correct for mass bias and instrument drift, and the external precision of

176

the resulting Fe isotopic analyses was ± 0.06‰ (2 standard deviations (SD)) for δ56Fe

177

(Du et al., 2017). Analyses of the DTS-2b, BHVO-2, and BCR-2 standards, which

178

were purified and measured together with unknowns during this study, yielded δ56Fe

179

values of 0.06 ± 0.08‰ (n = 3, 2SD), 0.13 ± 0.05‰ (n = 9, 2SD), and 0.11 ± 0.08‰

180

(n = 9, 2SD; Du et al., 2017), respectively, consistent with recommended values

181

(Craddock and Dauphas, 2011; Liu et al., 2014). Each sample was analyzed in

182

triplicate, Fe isotopic data are reported in δ56Fe notation relative to the IRMM-014

183

international standard, and all uncertainties are given as ±2SD.

184 185

4. Results

186

The Fe isotopic compositions of the Late Jurassic granites within the study area

187

are given in Table 1 and are shown in Fig. 2. Bulk-rock δ56Fe values for the samples

188

analyzed during this study range from 0.12‰ to 0.42‰. The Qianlishan granites have

189

the highest δ56Fe values (0.29‰ ± 0.05‰ to 0.42‰ ± 0.01‰) whereas the other

190

samples have a relatively narrow range of δ56Fe values (0.12‰ ± 0.04‰ to 0.27‰ ±

191

0.03‰). In detail, the Qitianling granites have δ56Fe values ranging from 0.12‰ ±

192

0.04‰ to 0.24‰ ± 0.02‰ and the Laiziling granites have δ56Fe values that range

193

from 0.20‰ ± 0.06‰ to 0.27‰ ± 0.06‰. In general, all of the samples analyzed in

194

this study have compositions that follow the trends (e.g., δ56Fe vs SiO2 and δ56Fe vs 10

195

Fe2O3t; Fig. 2a, b) defined by felsic rocks in previous studies (Poitrasson and Freydier,

196

2005; Heimann et al., 2008; Schuessler et al., 2009; Telus et al., 2012; Zambardi et al.,

197

2014; Foden et al., 2015; Gajos et al., 2016; He et al., 2017; Xia et al., 2017; Du et al.,

198

2017). In addition, the δ56Fe values of the Qitianling and Qianlishan granites correlate

199

with other indices of magmatic differentiation, including Sr, and Ba contents and

200

Ba/Rb ratios (Fig. 2d-f).

201 202

5. Discussion

203

5.1. Partial melting and wall-rock assimilation effects

204

Previous research indicates that Fe isotopic fractionation can occur during crustal

205

anatexis (Telus et al., 2012), with Fe isotopes fractionating by an average of 0.09‰ ±

206

0.06‰ between leucosome and melanosome (∆56FeL–M; Xu et al., 2017). The fact that

207

heavy Fe isotopes preferentially partition into Fe3+ sites relative to Fe2+ sites (Dauphas

208

et al., 2014; Young et al., 2015) means that the slightly higher incompatibility of Fe3+

209

over Fe2+ can yield melts with heavier Fe isotopic compositions during low-degree

210

partial melting. However, there is no positive correlation between ∆56FeL–M and

211

Fe3+/ΣFe ratios in migmatites (Xu et al., 2017), suggesting that the preferential

212

partitioning of isotopically heavy Fe3+ during partial melting may only have a weak

213

effect on leucosomes. In addition, this effect (of the order of 0.1‰ or less) is small

214

relative to the observed Fe isotopic variability (0.12‰ ± 0.04‰ to 0.42‰ ± 0.01‰)

215

within samples of granites formed during similar or high degrees of partial melting.

216

Wall-rock assimilation is common in granitic systems and could therefore also 11

217

modify the Fe isotopic compositions of granitic magmas. However, this process is in

218

essence analogous to partial melting at shallow crustal levels. In addition, the

219

relatively high and homogeneous Nd isotopic ratios of the Qitianling (–7.6 < εNd(t) <

220

–5.5; Zhao et al., 2012), Qianlishan (–7.59 < εNd(t) < –6.35; Mao et al., 1995), and

221

Laiziling (–6.3 < εNd(t) < –6.0; Li et al., 2018) granites suggest that the magmas that

222

formed these granites did not undergo significant contamination by interaction with

223

the Triassic–Cambrian country rocks (εNd(t) = –15.1 to –10; Yu and Shen, 2007) in the

224

Nanling region. This means that magmatic processes including partial melting,

225

assimilation at shallow levels, and the mixing of different batches of granite melts

226

derived from similar sources cannot account for the observed Fe isotopic variations

227

within rocks in the study area.

228 229

5.2. Fe isotopic fractionation during magmatic–hydrothermal processes

230

The δ56Fe values of the Late Jurassic granites negatively correlate with indices of

231

magmatic differentiation (e.g., Fe2O3t, Sr, Ba, and Ba/Rb; Fig. 2), suggesting that

232

low-δ56Fe components were removed during the evolution of these magmas. Previous

233

research indicates that numerous Fe-bearing silicate minerals (e.g., olivine, pyroxene,

234

amphibole, and biotite) and some Fe–Ti oxides (e.g., ilmenite) are enriched in light Fe

235

isotopes relative to whole-rock or melt compositions (Heimann et al., 2008; Shahar et

236

al., 2008; Schuessler et al., 2009; Sossi et al., 2012; Telus et al., 2012; Dauphas et al.,

237

2014). However, theoretical calculations suggest that Fe-bearing aqueous fluids are

238

also isotopically lighter than silicate melts (Schauble et al., 2001; Dauphas et al., 12

239

2014), indicating that the removal of both aqueous fluid and minerals can drive

240

evolving melts towards heavier Fe isotopic compositions. It should also be noted that

241

Fe, Na, Ba, and Sr are fluid-mobile (Webster et al., 1989; Bai and Koster van Groos,

242

1999; Zajacz et al., 2008) and highly compatible in the major rock-forming minerals

243

(e.g., biotite, K-feldspar, and plagioclase) of granites that form during mineral–melt

244

partitioning of felsic magmas (e.g., Ewart and Griffin, 1994). This means it is difficult

245

to discriminate between the effects of fluid exsolution and fractional crystallization on

246

Fe isotopic compositions during magmatic–hydrothermal processes using Fe2O3t,

247

Na2O, Sr, and Ba versus δ56Fe diagrams alone.

248

Dimensionless Zr/Hf, Nb/Ta, and K/Rb ratios and the REE tetrad effect can

249

provide evidence of fluid exsolution during magmatic–hydrothermal processes (Fig.

250

3). The elements within each of these elemental pairs/combinations behave similarly

251

in minerals but differently in fluids. For example, the Zr/Hf ratios of igneous rocks

252

remain near-chondritic (26–46) during magmatic processes (e.g., zircon fractionation;

253

Bea et al., 2006) but decrease significantly (<26) during fluid–magma interaction

254

because Zr preferentially partitions into F-bearing fluids over Hf (Bau, 1996; Louvel

255

et al., 2014). However, there is no apparent co-variation in the δ56Fe and Zr/Hf values

256

of the Late Jurassic Nanling granites. The Nb/Ta ratios of the granites also decrease as

257

a result of the fractionation of mica and/or amphibole (Stepanov et al., 2014).

258

However, Ballouard et al. (2016) suggested that fractional crystallization alone is not

259

sufficient to reproduce the wide range of Nb/Ta ratios in granitic rocks and proposed

260

that samples that record interactions with fluids have Nb/Ta values of <5. There is a 13

261

clear correlation between δ56Fe and Nb/Ta values during the early fractional

262

crystallization (Nb/Ta > 5) of the Qitianling granites, providing evidence of the

263

fractional crystallization of high Nb/Ta and low δ56Fe minerals (e.g., biotite and

264

amphibole; Stepanov et al., 2014; Fig. 3b). In addition, K/Rb ratios of <100 are

265

thought to reflect interaction between granitic melts and fluid phases (Clarke, 1992).

266

The high K/Rb ratios of the Qitianling granites (Fig. 3c; K/Rb > 100) provide

267

evidence of fractional crystallization. However, the δ56Fe values of the granites in the

268

study area do not correlate with their Nb/Ta and K/Rb ratios for samples associated

269

with the magmatic–hydrothermal stages of evolution (i.e., Nb/Ta < 5, K/Rb < 100; Fig

270

3b, c). Moreover, the REE tetrad effect is thought to reflect the activities of fluids

271

exsolved from magmas (e.g., Bau, 1996; Irber, 1999), but recent experiments indicate

272

that the fractional crystallization of monazite and xenotime can also generate tetrad

273

effects (Duc-Tin and Keppler, 2015). The Nd and Y concentrations within the

274

Nanling granites increase sharply with decreasing concentrations of P2O5 in highly

275

fractionated samples (P2O5 < 0.05 wt.%; Fig. 3e, f), precluding the fractional

276

crystallization of monazite and xenotime and indicating that the REE tetrad effect

277

must have been produced by fluid exsolution. However, samples from the study area

278

with clear REE tetrad effects (TE1,3 > 1.1; Fig. 3d) do not display systematic

279

differences in Fe isotopic compositions relative to samples that are free of the tetrad

280

effect (TE1,3 < 1.1; Fig. 3d). Although each of these ratios may be also influenced by

281

specific mineral–melt

282

including zircon-affecting Zr/Hf (Bea et al., 2006), mica- and titanite-affecting Nb/Ta

partitioning during

14

magmatic–hydrothermal

processes,

283

(Stepanov et al., 2014), and K-feldspar-affecting K/Rb ratios, a combination of these

284

parameters provides a useful indicator of the presence of a magmatic–hydrothermal

285

stage in the evolution of granite magmas. This also indicates that the absence of a

286

correlation between these multiple geochemical indicators and δ56Fe values suggests

287

that fluid exsolution has a limited effect on granite Fe isotopic compositions.

288

We also used the fluid-mobile elements F and Cl to further evaluate the isotopic

289

effects of fluid exsolution. Removal of minerals from the evolving melt leaves F/Cl

290

ratios unchanged or slightly lower because F is slightly more compatible than Cl in

291

volatile-bearing minerals (e.g., biotite, amphibole, and apatite). The DF/DCl (where D

292

is the mineral–melt partition coefficient) is ~1.9 for biotite (Icenhower and London,

293

1997), ~54 for amphibole (Iveson et al., 2017), and ~8.5 for apatite (Webster et al.,

294

2009). However, F and Cl behave the opposite way during fluid–melt partitioning

295

relative to mineral–melt partitioning in that Cl partitions more strongly into aqueous

296

fluids than granitic melts (average DCl = 37) while F is typically concentrated in

297

granitic melts relative to aqueous fluids (Webster and Holloway, 1990). This means

298

that fluid exsolution will significantly increase the F/Cl ratio of the residual melt,

299

meaning that a positive correlation between F/Cl and δ56Fe values is expected. The

300

Qitianling granites data indicate that F/Cl ratios initially decrease with decreasing

301

Fe2O3t (5.28–2.19 wt.%; Fig. 4a), suggesting that the magma at this stage was not

302

fluid-saturated. δ56Fe values show a slight increase in the evolving melt (from 0.12‰

303

± 0.04‰ to 0.20‰ ± 0.06‰, Fig. 4b) during the fractional crystallization of

304

volatile-bearing minerals (e.g., biotite, amphibole). However, an abrupt increase in 15

305

F/Cl occurs in samples containing less than ~2 wt.% Fe2O3t (Fig. 4a), with this change

306

being coincident with fluid exsolution. The fluid saturation of the melt causes a

307

significant increase in F/Cl ratios from 5.8 to 43.1 but there is no resolvable

308

associated increase in δ56Fe values (from 0.20‰ ± 0.06‰ to 0.24‰ ± 0.02‰; Fig.

309

4b). It should be noted that the two high F/Cl samples from the Qitianling pluton

310

record only minor volatile loss, as indicated by their moderately high Zr/Hf, Nb/Ta,

311

and K/Rb ratios, and minimally developed REE tetrad effect (Fig. 3). This reflects the

312

remarkable sensitivity of F/Cl ratios to the extent of fluid exsolution compared to the

313

other indicators outlined above. The samples from the other plutons record more

314

differentiation than the Qitianling granites (as evidenced by their lower Sr, Ba, and Fe

315

contents) and more extensive fluid exsolution (e.g., lower Zr/Hf and Nb/Ta ratios, and

316

more variable F/Cl ratios) but do not have a positive correlation between F/Cl and

317

δ56Fe values (Fig. 4b). This again indicates that volatile exsolution does not have a

318

significant effect on Fe isotopic fractionation.

319 320

5.3. Quantitative assessment of Fe isotopic fractionation in a crystal–melt–fluid

321

system

322

5.3.1. Fractionation model and parameters

323

A quantitative assessment of the fractionation of Fe isotopes within a

324

crystal–melt–fluid system is needed to better understand the behavior of Fe isotopes

325

during magmatic–hydrothermal processes. Here, we model Fe isotopic fractionation

326

in a crystal–melt–fluid system by initially quantifying the Fe2O3t contents of residual 16

327

melts. The Fe2O3t content of a coexisting crystal–melt–fluid system has been

328

described using a fluid-saturated mass-balance equation for isobaric fractional

329

crystallization (Spera et al., 2007). This modeling indicates that Fe2O3t concentration

330

within a residual melt (
 ) after water saturation can be determined as follows:

331

   1−  Ф2 +  −1
 =
 ( )

332

where

 

(1)



333


 = content of Fe2O3t in the melt at water saturation,

334

 = mass fraction of melt,

335

 = mass fraction of melt at water saturation,

336

  = bulk crystal–melt partition coefficient,

337

  = bulk fluid–melt partition coefficient,

338

Ф2 = negative value of the solubility of water in the melt.

339

We assume that the magma underwent Rayleigh-type fractional crystallization

340

before water saturation, indicating that
 can be calculated as follows:

341


 =
0  ( −1)

342

where
0 is the initial content of Fe2O3t in the melt.

(2)

343

We reduced the uncertainties involved in the evaluation of oxygen fugacity and

344

mineral assemblages during magmatic processes by estimating   values for bulk

345

iron using highly incompatible elements. Uranium concentrations within the samples

346

from Qitianling pluton increase by a factor of 3–5 as Fe2O3t decreases from 6.1 to 1.4

347

wt.%, indicating that U is highly incompatible (Fig. 5a). This is consistent with the

348

presence of a dominant mineral assemblage (biotite, plagioclase, K-feldspar, quartz, 17

349

and amphibole) within the Qitianling pluton that contains minerals that cannot

350

accommodate U (e.g., Mahood and Hildreth, 1983; Nash and Crecraft, 1985). We

351

estimate the   values for Fe2O3t by plotting  versus Fe2O3t relative to the

352

parental composition represented by the most primitive sample within the Qitianling

353

pluton (sample c48-2; U = 7.1 ppm; Fe2O3t = 6.06 wt.%; Bai et al., 2005), where 

354

is estimated by inverting the enrichment factor of an incompatible element (

355

for an ideal incompatible element; e.g., U). These calculations indicate that the Fe2O3t

356

values of the Qitianling granites can be reproduced by fractional crystallization with a

357

  value of 1.8 following a Rayleigh fractionation model (Fig. 5b).



0

1

=



358

The next step in this modeling is to constrain the Fe fluid–melt partition

359

coefficient (   ). Measured   values based on coexisting melt and fluid

360

inclusions have a large range (0.42–13.32; Zajacz et al., 2008). In addition, the

361

fluid–melt partition coefficient (  ) of Fe is a positive function of the concentration

362

of Cl in the aqueous phase as Fe is predominantly transported as a chloride complex

363

within fluid phases (e.g., Chou and Eugster, 1977; Zajacz et al., 2008).

364

Pressure–volume–temperature–composition (PVTX) phase relations in the NaCl–H2O

365

system indicate that a magma-derived supercritical fluid would separate into a

366

low-salinity (low NaCl) aqueous vapor phase and a high-salinity (high NaCl) brine

367

phase under shallow crust pressures (Driesner and Heinrich, 2007). However, the ratio

368

of vapor to saline exsolved from a pluton is less well constrained. Here we use an

369

average   value of 4.2 for the bulk aqueous fluid, based on the dataset of Zajacz

370

et al. (2008). This is done for convenience of calculation and to enable the 18

of

Fe2O3t

371

determination

and

372

magmatic–hydrothermal processes.

δ56Fe

variations

in

residual

melts

during

373

The Qitianling granitic melt has a water solubility ( −Ф  ) value at an

374

intermediate pressure of 3.6 kbar (estimated using the Al-in-hornblende barometer;

375

Zhao et al., 2005) of ~8.4 wt.% based on the empirical equations of Holtz et al. (2001)

376

and Liu et al. (2005). This defines the upper limit of H2O content within the magmatic

377

systems considered in this study. In addition, previous research on melt inclusions

378

within phenocrysts indicated that the majority of silicic volcanic rocks contain 3–6 wt.%

379

volatiles (Wallace et al., 2015; Du et al., 2017). Consequently, we used a wide range

380

of initial H2O contents (1–6 wt.%) during our modeling of granitic melts. We assume

381

H2O is 100% incompatible with respect to the crystallizing mineral assemblage

382

because the amount of H2O stored in hydrous minerals is rather low (<0.6 wt.% for

383

felsic igneous rocks; Caricchi and Blundy, 2015). This means that the H2O content of

384

the residual melt continuously increases as crystallization proceeds until the residual

385

melt reaches water saturation (Fig. 6a). This enables the calculation of the mass

386

fraction of melt at water saturation ( ), with these values generally being dependent

387

on the initial magma H2O content (Fig. 6a). The water saturation of the evolving melt

388

is followed by the continuous removal of aqueous fluid from the melt as

389

crystallization proceeds.

390

With  constrained, the Fe isotopic fractionation of a crystal–melt–fluid system

391

can be modeled using a Rayleigh fractionation equation as follows:

392

56 =56 0–

56

 − !"#/( !"# + %&') × ln(  ) 19

(3)

393

where  represents the Fe fraction in the remaining melt that can be calculated as


56

394

 = ×( 0 ). A

395

with this value being based on the observed differences between the mean Fe isotopic

396

compositions of whole-rock samples and associated common Fe-bearing minerals

397

(

398

2012; Du et al., 2017) within the Qitianling granites and other granitic rocks. The Fe

399

isotope fractionation factor between melts and fluids (

400

106/T2) was calculated by Du et al. (2017) based on a theoretically calculated value of

401




56

56

− !"# value of 0.1‰ was used prior to fluid saturation,

+ℎ- !- . − !"#, e.g., amphibole and biotite; Heimann et al., 2008; Telus et al.,

56

 − %&' = 0.15‰ ×

#/0& − %&' of 0.28‰ × 106/T2 (Heimann et al., 2008) and a naturally 56

#/0& −  of 0.13‰ × 106/T2 (Red Hill intrusion; Sossi

402

observed value of

403

et al., 2012) where the theoretically calculated

404

consistent with the experimental value of 0.29‰ at 800°C (Sossi et al., 2017). A

405

56

 − %&'

12

345678987 : ;<=9> value is

value of 0.14‰ was estimated using a

56

 − %&'

406

fractionation factor of 0.15‰ × 106/T2 (Du et al., 2017) for the crystallization

407

temperatures of the Nanling granites during magmatic–hydrothermal processes

408

(including the Qianlishan, Laiziling, and Yaogangxian plutons, where TTi-in-zrn =

409

717°C–801°C; Wang et al., 2014). This means that

410

were weighted after fluid saturation according to the ratio of minerals to fluid and Fe

411

content as follows:

412

56

−( !"#+ %&')=? !"# ·

56

56

 − ( !"# + %&') values

− !"#+? %&' ·

56

− %&' (4)

413

where ? !"# and ? %&' are the mass fraction of Fe in crystals and fluid,

414

respectively, with ? !"# + ? %&' = 1. The Fe content of the fractionated crystals 20

415

and the exsolved fluid can be calculated by CBCDE84< = B3 · C37<8 and C;<=9> =

416

;3 · C37<8 , respectively. The mass ratio of fractional crystals to exsolved fluid is

417

92/8, as the crystallization of every 1 wt.% of melt requires the release of 0.08 wt.%

418

fluid to maintain a water content of 8.4 wt.% (i.e., water solubility) in the residual

419

melt. This yields a bulk fractionation factor

420

0.107‰. The granitic samples from the study area have magnetite as the dominant

421

Fe–Ti oxide phase, and the removal of magnetite from a melt results in lower 56

422

56

 − ( !"# + %&') value of

− !"# values. As such, we have evaluated the contribution of fluid

423

exsolution to the Fe isotopic fractionation of the evolving melt based on differences

424

56 0-  between 56   and   , as follows: 56

425

56 0-  ′ = 56   –  

(5)

56

426

where    is the Fe isotopic composition of the melt produced by the

427

exsolution of Fe-bearing aqueous fluids and 56 0-   refers to the iron isotopic

428

composition of the melt produced by the exsolution of fluids free of Fe. This means

429

that any variations in

430

contribution of fluid exsolution to the Fe isotopic fractionation of the evolving melt

431

(

432

0.1‰, which is identical to the value expected for undifferentiated silicic igneous

433

rocks (e.g., Heimann et al., 2008; Foden et al., 2015).

56

56

 − !"# values do not strongly affect the calculated

′ ). This model assumes for simplicity that the initial melt has a δ56Fe value of

434 435 436

5.3.2. Modeling results and discussion Before discussing the modeling results of Fe isotopes, we first examine the 21

437

capacity of fluids for Fe transport. The mass of Fe transferred from silicate melt to

438

fluids varies with initial water contents (Fig. 6b). Although Fe is more compatible in

439

fluids than in bulk minerals (  >  ), the extraction of Fe from melts into

440

aqueous fluids is less efficient than mineral crystallization as a result of the low

441

water/mineral ratio (~8/92) of the products generated by the evolution of melts during

442

magmatic–hydrothermal processes (e.g., Candela and Holland, 1986; Cline and

443

Bodnar, 1991; Audétat et al., 2000). This means that the maximum difference in

444

Fe2O3t between hydrous and anhydrous melts does not exceed 0.4 wt.% even if the

445

initial water content of the melt is 6 wt.%. This in turn means that melt Fe2O3t

446

variations are largely controlled by the fractional crystallization of Fe-bearing

447

minerals.

448

The aqueous fluid saturation of an evolving melt causes

56

′ values to

449

increase with decreasing mass fraction of melt ( ) and increasing initial H2O

450

contents, indicating that both initial H2O contents and mass fraction of melt ( )

451

affect the Fe isotopic composition of the residual melt. Fluids exsolved from evolving

452

melts only produce resolvable and/or detectable effects on the Fe isotopic

453

fractionation of granitic magmas during high degrees of differentiation. For example,

454

a granitic melt with a 6 wt.% initial fluid content has

455

0.06‰ unless  < 20 wt.%; similar results are obtained for granitic melts with 1

456

wt.% initial fluid contents that need  values of <4 wt.% (Fig. 6c). Previous

457

research based on melt inclusions within phenocrysts indicate that the majority of

458

silicic volcanic rocks contain 3–6 wt.% volatiles (Wallace et al., 2015; Du et al., 2017). 22

56

′ values that do not exceed

459

This means that the exsolution of fluids from evolving melts can only have a limited

460

impact on the Fe isotopic compositions of the majority of granitic plutons that have

461

low initial volatile contents. Resolvable fractionation of Fe isotopes associated with

462

fluid exsolution (i.e., >0.06‰) is therefore only expected in highly hydrous and

463

extremely differentiated granites (e.g.,  < 10 wt.% in Fig. 6c).

464

The Rayleigh fractionation model outlined in section 5.3.1 allows a quantitative

465

estimation of the effect of fluid exsolution on bulk-rock δ56Fe values during

466

magmatic–hydrothermal processes recorded within the Nanling granites. Fluid

467

exsolution during the magmatic evolution of these granites can be constrained by F/Cl

468

ratios, which are generally constant during fractional crystallization but significantly

469

increase after fluid saturation. F/Cl ratios decrease slightly with decreasing 

470

values between 100 wt.% and ~40 wt.% before increasing sharply at  < 40 wt.%

471

(Fig. 7a). Fluid exsolution occurred at 60 wt.% magma solidification ( = 40 wt.%)

472

within the Qianlishan granites, and at 70 wt.% ( = 30 wt.%) within the Qitianling

473

granites. This enables the determination of the contribution of fluid exsolution to the

474

Fe isotopic fractionation recorded in the Nanling granites from the beginning of fluid

475

exsolution ( = 40 wt.%) to the final differentiation of the magma ( = ~10 wt.%;

476

sample SZY-10). The contribution of fluid exsolution to the δ56Fe values of the

477

residual melt increases with increasing crystallization (Fig. 7b). However, the

478

maximum shift in the δ56Fe values of the residual melt does not exceed 0.07‰ even at

479

up to 90 wt.% magma solidification (as shown in the shaded area in Fig. 7b). This

480

indicates that fluid exsolution has a negligible effect on Fe isotopic variations, with 23

481

fractional crystallization being the dominant controlling factor in the fractionation of

482

Fe isotopes during the magmatic–hydrothermal evolution of granites, which is

483

consistent with the results of previous research (e.g., Sossi et al., 2012; Telus et al.,

484

2012; Foden et al., 2015).

485

The relative importance of the fractionation of Fe-bearing mineral phases on Fe

486

isotopic fractionation in the Nanling granites can be assessed using Rayleigh

487

fractionation models. Biotite, magnetite, and amphibole (the latter only in low-silica

488

granites with SiO2 < 69 wt.%) are the major Fe-bearing minerals within the granites

489

from the study area (Table 1), with the Qitianling granites containing nearly pure

490

magnetite (Wang et al., 2013). The fractionation factor between nearly pure magnetite

491

and melt, based on a case study from the Red Hill area, is positive (∆56FeMag–melt =

492

0.13‰ × 106/T2; Sossi et al., 2012), indicating that the removal of magnetite would

493

lead to the depletion of heavy Fe isotopes within the residual melt (Fig. 8). Although

494

no experimental or theoretical estimates of the fractionation factors between silicate

495

minerals (biotite and amphibole) and melts have been calculated to date, a suggested

496

value of –0.1‰ for a biotite–melt fractionation factor was determined using average

497

∆56FeBt–whole rock values (Du et al., 2017). Here, we estimate the fractionation factor

498

between amphibole and melt in a similar way using the δ56Fe values for amphibole

499

separates and corresponding whole-rock Fe isotopic compositions. The amphibole

500

separates have δ56Fe values of –0.00‰ to 0.09‰ (Heimann et al., 2008; Telus et al.,

501

2012), which are slightly below the values of the corresponding whole-rock samples

502

(∆56FeAmp–Whole rock = 0.01‰ to –0.20‰, with an average of –0.08‰). This indicates 24

503

that the removal of biotite and amphibole could drive the evolving melt towards

504

heavier Fe isotopic compositions (Fig. 8). An integrated bulk mineral–melt Fe

505

isotopic fractionation factor can be estimated using Equation 3, demonstrating that the

506

granites from the Nanling Range can be produced by fractional crystallization with

507

mineral–melt fractionation factors ranging from –0.04‰ to –0.08‰ (Fig. 8), which

508

are higher than the ∆56FeBt–whole rock but lower than the ∆56FeMag–melt values for these

509

granites. This indicates that Fe isotopic fractionation in the Nanling granites is the

510

concomitant effect of the separation of silicates (biotite and amphibole) and magnetite,

511

with the silicates being the main mineral control on the Fe isotopic fractionation

512

trends recorded by these granites.

513 514

5.4. Comparison of Fe isotopic compositions among rock types

515

Fluid exsolution from crystallizing magmas is an essential step in the formation of

516

magma-related Cu, W, Sn, and Mo mineralization (e.g., Hedenquist and Lowenstern,

517

1994; Candela, 1997; Audétat et al., 2000). Fluid saturation enables the transportation

518

of metals and other fluid-mobile elements away from the crystallizing magma by the

519

exsolving aqueous fluids. This means that ore-related granites may record

520

significantly more fluid–melt/rock interaction relative to barren granites. For example,

521

the majority of barren granites have high Zr/Hf and Nb/Ta ratios (26 < Zr/Hf < 46,

522

and 5 < Nb/Ta < 16), whereas granites associated with ore deposits (e.g., W–Sn and

523

Ta–Cs–Li–Nb–Be–Sn–W) have low Zr/Hf (<26) and Nb/Ta (<5) ratios (Ballouard et

524

al., 2016). Previous research indicates that isotopically light Fe is preferentially 25

525

incorporated into fluids relative to high-silica melts (e.g., Schauble et al., 2001;

526

Dauphas et al., 2014). This means that isotopically light Fe is continuously removed

527

from crystallizing magmas by aqueous fluids, meaning that ore-related granites

528

should have heavier Fe isotopic compositions within systems that have undergone

529

significant fluid exsolution. However, all of the mineralization-related granites in the

530

study area and numerous other examples from the literature (W–Sn and Mo

531

deposit-related; Heimann et al., 2008; Foden et al., 2015; Wawryk and Foden, 2015)

532

generally follow the barren-granite trend (Poitrasson et al., 2005; Heimann et al., 2008;

533

Telus et al., 2012; Foden et al., 2015; Gajos et al., 2016; He et al., 2017) in a

534

δ56Fe–SiO2/Fe2O3t diagram (Fig. 9a, b), indicating a lack of significant difference

535

between the Fe isotopic compositions of barren and mineralized granites.

536

The effects of fluid exsolution on the δ56Fe values of plutonic rocks can be further

537

assessed by comparing the Fe isotopic compositions of plutonic rocks with the

538

compositions of volcanic rocks. Previous research on fluid-mobile elements and

539

associated isotopes (e.g., Li and δ7Li, Schuessler et al., 2009; Zn and δ66Zn, Xia et al.,

540

2017) and Rayleigh fractionation modeling (Du et al., 2017) suggests that volcanic

541

degassing (fluid exsolution) has only a minor effect on whole-rock δ56Fe values of

542

volcanic rocks (especially rhyolites) relative to crystal fractionation. Plutonic rocks

543

are expected to have heavier Fe isotopic compositions relative to volcanic rocks if

544

fluid exsolution had a significant role within the intrusive magmatic systems.

545

However, published Fe isotopic data for 194 plutonic and 106 volcanic samples (SiO2

546

60–78 wt.%) indicate that differentiated (e.g., SiO2 > 74 wt.%, Fe2O3t < 3 wt.%; Fig. 26

547

9b) plutonic rocks have overall slightly lighter Fe isotopic compositions than similarly

548

differentiated volcanic rocks, with ∆56Fevolcanic-plutonic values of 0.034‰ to 0.062‰

549

(average of 0.05‰), although the two sets of values are consistent within uncertainties

550

(given as 2SE in Fig. 9d). In addition, there is no resolvable difference between these

551

two rock types in less differentiated samples (Fe2O3t > 3 wt.%; Fig. 9b). The more

552

remarkable differences in δ56Fe values between volcanic and plutonic rocks in

553

Fe2O3t–δ56Fe diagrams than in SiO2–δ56Fe diagrams reflect the greater sensitivity of

554

Fe2O3t to the extent of differentiation relative to SiO2 content for highly differentiated

555

rocks. This means that the very small differences in δ56Fe values between volcanic

556

and plutonic rocks are not caused by fluid exsolution but rather are caused by

557

variations in the extent of fractional crystallization.

558 559

6. Conclusions

560

This study presents the results of an investigation into the behavior of Fe isotopes

561

during the magmatic–hydrothermal evolution of four Late Jurassic granites in the

562

western Nanling Range of southern China. These rocks have highly variable δ56Fe

563

values (from 0.12‰ ± 0.04‰ to 0.42‰ ± 0.01‰) that positively correlate with

564

geochemical indicators of magma differentiation. The relative contributions of

565

fractional crystallization and fluid exsolution to the whole-rock δ56Fe values of the

566

samples were assessed using geochemical proxies and Rayleigh fractionation

567

modeling. The results of geochemistry and modeling indicate that fluid exsolution is

568

only a minor control (<0.07‰) on the Fe isotope variability of the Nanling granites, 27

569

barring some extremely differentiated granites (magma solidification >90 wt.%) in

570

which Fe isotopic variations caused by fluid exsolution may be resolvable. This

571

means that crystal fractionation is the dominant control on Fe isotope fractionation

572

during the magmatic–hydrothermal stages of granitic magmatism, indicating that Fe

573

isotopes could be used to robustly identify cumulate and fractional crystallization

574

processes within silicic reservoirs.

575 576

Acknowledgments:

577

This work was supported by the State Key R&D Project of China (2016YFC0600203)

578

and Dengfeng Project of Nanjing University to XLW, and National Natural Science

579

Foundation of China (No. 41622301) to WL. We thank Craig Lundstrom and two

580

anonymous reviewers for their constructive criticisms on a previous version of this

581

manuscript. This version of the manuscript benefits from constructive reviews from

582

John Foden and an anonymous reviewer. We also thank Dr. Xian-Hua Li for his

583

editorial handling and constructive comments.

584 585

References

586

Ackerson, M.R., Mysen, B.O., Tailby, N.D., Watson, E.B., 2018. Low-temperature

587

crystallization of granites and the implications for crustal magmatism. Nature 559,

588

94–97.

589 590

Albarède, F., Beard, B., 2004. Analytical methods for non-traditional isotopes. Rev. Mineral. Geochem. 55, 113–152. 28

591

Audétat, A., Günther, D., Heinrich, C.A., 2000. Magmatic-hydrothermal evolution in

592

a fractionating granite: A microchemical study of the Sn-W-F-mineralized Mole

593

Granite (Australia). Geochim. Cosmochim. Acta 64, 3373–3393.

594

Audétat, A., Pettke, T., Heinrich, C.A., Bodnar, R.J., 2008. Special paper: The

595

composition of magmatic-hydrothermal fluids in barren and mineralized

596

intrusions. Econ. Geol. 103, 877–908.

597

Bai, D.Y., Chen, J.C., Ma, T.Q., Wang, X.H., 2005. Geochemical characteristics and

598

setting of Qitianling A-type granitic pluton in southeast Hunan. Acta Petrol.

599

Mineral. 24, 255–272 (in Chinese with English abstract).

600

Bai, T.B., Koster van Groos, A., 1999. The distribution of Na, K, Rb, Sr, Al, Ge, Cu,

601

W, Mo, La, and Ce between granitic melts and coexisting aqueous

602

fluids. Geochim. Cosmochim. Acta 63, 1117–1131.

603

Baker, D.R., Alletti, M., 2012. Fluid saturation and volatile partitioning between melts

604

and hydrous fluids in crustal magmatic systems: The contribution of experimental

605

measurements and solubility models. Earth-Sci. Rev. 114, 298–324.

606

Ballouard, C., Poujol, M., Boulvais, P., Branquet, Y., Tartese, R., Vigneresse, J.L.,

607

2016. Nb-Ta fractionation in peraluminous granites: A marker of the

608

magmatic-hydrothermal transition. Geology 44, 231–234.

609

Bau, M., 1996. Controls on the fractionation of isovalent trace elements in magmatic

610

and aqueous systems: evidence from Y/Ho, Zr/Hf, and lanthanide tetrad

611

effect. Contrib. Mineral. Petrol. 123, 323–333.

612

Bea, F., Montero, P., Ortega, M., 2006. A LA–ICP–MS evaluation of Zr reservoirs in 29

613

common crustal rocks: Implications for Zr and Hf geochemistry, and

614

zircon-forming processes. Can. Mineral. 44, 693–714.

615

Burnham, C.W., Ohmoto, H., 1980. Late-stage processes of felsic magmatism, in:

616

Ishihara, S., Takenouchi (Eds.), Granitic Magmatism and Related Mineralization.

617

Mining Geol. Spec. Issue, pp. 1–13.

618 619

Candela, P.A., 1997. A review of shallow, ore-related granites: textures, volatiles, and ore metals. J. Petro. 38, 1619–1633.

620

Candela, P.A., Holland, H.D., 1986. A mass transfer model for copper and

621

molybdenum in magmatic hydrothermal systems: the origin of porphyry-type ore

622

deposits. Econ. Geol. 81, 1–19.

623

Cao, J., Yang, X., Du, J., Wu, Q., Kong, H., Li, H., Qiu, W., Xi, X., Gong, Y., Zhao,

624

H., 2018. Formation and geodynamic implication of the Early Yanshanian granites

625

associated with W–Sn mineralization in the Nanling range, South China: An

626

overview. Int. Geol. Rev. 60, 1744–1771.

627 628

Caricchi, L., Blundy, J., 2015. Experimental petrology of monotonous intermediate magmas. Geol. Soc. London, Spec. Publ. 422, 105–130.

629

Chen, B., Ma, X., Wang, Z., 2014. Origin of the fluorine-rich highly differentiated

630

granites from the Qianlishan composite plutons (South China) and implications

631

for polymetallic mineralization. J. Asian Earth Sci. 93, 301–314.

632

Chen, J., Wang, R., Zhu, J., Lu, J., Ma D., 2013. Multiple-aged granitoids and related

633

tungsten-tin mineralization in the Nanling Range, South China. Sci. China Earth

634

Sci. 56, 2045–2055. 30

635 636 637 638 639 640 641 642

Chou, I.M., Eugster, H.P., 1977. Solubility of magnetite in supercritical chloride solutions. Am. J. Sci. 277, 1296–1314. Clarke, D.B., 1992. The mineralogy of peraluminous granites: A review. Can. Mineral. 19, 3–17. Cline, J.S., Bodnar, R.J., 1991. Can economic porphyry copper mineralization be generated by a typical calc-alkaline melt ?. J. Geophys. Res. 96, 8113–8126. Craddock, P.R., Dauphas, N., 2011. Iron isotopic compositions of geological reference materials and chondrites. Geostand. Geoanal. Res. 35, 101–123.

643

Dauphas, N., Roskosz, M., Alp, E.E., Neuville, D.R., Hu, M.Y., Sio, C.K., Tissot,

644

F.L.H., Zhao, J., Tissandier, L., Médard, E., Cordier, C., 2014. Magma redox and

645

structural controls on iron isotope variations in Earth's mantle and crust. Earth

646

Planet. Sci. Lett. 398, 127–140.

647

Davidson, P., Kamenetsky, V.S., 2007. Primary aqueous fluids in rhyolitic magmas:

648

Melt inclusion evidence for pre-and post-trapping exsolution. Chem. Geol. 237,

649

372–383.

650

Deng, X.G., Li, X.H., Liu, Y.M., Huang, G.F., Hou, M.S., 2005. Geochemical

651

characteristics of Qitianling granites and their implications for mineralization.

652

Acta Petrol. Mineral. 24, 93–102 (in Chinese with English abstract).

653

Driesner, T., Heinrich, C.A., 2007. The system H2O–NaCl. Part I: Correlation

654

formulae for phase relations in temperature–pressure–composition space from 0 to

655

1000 °C, 0 to 5000bar, and 0 to 1 XNaCl. Geochim. Cosmochim. Acta 71,

656

4880–4901. 31

657

Du, D.H., Wang, X.L., Yang, T., Chen, X., Li, J.Y., Li, W., 2017. Origin of heavy Fe

658

isotope

659

perspective. Geochim. Cosmochim. Acta 218, 58–72.

660 661 662 663 664 665

compositions

in

high-silica

igneous

rocks:

A

rhyolite

Duc-Tin, Q., Keppler, H., 2015. Monazite and xenotime solubility in granitic melts and the origin of the lanthanide tetrad effect. Contrib. Mineral. Petrol. 169, 8. Ewart, A., Griffin, W.L., 1994. Application of proton-microprobe data to trace-element partitioning in volcanic rocks. Chem. Geol. 117, 251–284. Foden, J., Sossi, P.A., Wawryk, C.M., 2015. Fe isotopes and the contrasting petro-genesis of A-, I-and S-type granite. Lithos 212–215, 32–44.

666

Gajos, N.A., Lundstrom, C.C., Taylor, A.H., 2016. Spatially controlled Fe and Si

667

isotope variations: an alternative view on the formation of the Torres del Paine

668

pluton. Contrib. Mineral. Petrol. 11, 1–20.

669

He, Y., Wu, H., Ke, S., Liu, S.A., Wang, Q., 2017. Iron isotopic compositions of

670

adakitic and non-adakitic granitic magmas: magma compositional control and

671

subtle residual garnet effect. Geochim. Cosmochim. Acta 203, 89–102.

672 673 674

Hedenquist, J.W., Lowenstern, J.B., 1994. The role of magmas in the formation of hydrothermal ore deposits. Nature 370, 519–527. Heimann, A., Beard, B.L., Johnson, C.M., 2008. The role of volatile exsolutions and 56

Fe/54Fe ratios in siliceous

675

sub-solidus fluid/rock interactions in producing high

676

igneous rocks. Geochim. Cosmochim. Acta 72, 4379–4396.

677

Heinrich, C.A., Candela, P.A., 2014. Fluids and ore formation in the Earth's crust,

678

in: Holland, H.D., Turekian, K.K (Eds.), Treatise on Geochemistry (Second 32

679

Edition). Elsevier, Oxford, pp. 1–28.

680

Holtz, F., Johannes, W., Tamic, N., Behrens, H., 2001. Maximum and minimum water

681

contents of granitic melts generated in the crust: a reevaluation and implications.

682

Lithos 56, 1–14.

683

Icenhower, J.P., London, D., 1997. Partitioning of fluorine and chlorine between

684

biotite and granitic melt: experimental calibration at 200 MPa H2O. Contrib.

685

Mineral. Petrol. 127, 17–29.

686

Irber, W., 1999. The lanthanide tetrad effect and its correlation with K/Rb, Eu/Eu∗,

687

Sr/Eu, Y/Ho, and Zr/Hf of evolving peraluminous granite suites. Geochim.

688

Cosmochim. Acta 63, 489–508.

689

Iveson, A.A., Webster, J.D., Rowe, M.C., Neill, O.K., 2017. Major element and

690

halogen (F, Cl) mineral-melt-fluid partitioning in hydrous rhyodacitic melts at

691

shallow crustal conditions. J. Petrol. 58, 2465–2492.

692

Li, H., Wu, J.H., Evans, N.J., Jiang, W.C., Zhou, Z.K., 2018. Zircon geochronology

693

and geochemistry of the Xianghualing A-type granitic rocks: insights into

694

multi-stage Sn-polymetallic mineralization in south china. Lithos 312-313, 1–20.

695

Lipman, P.W., Bachmann, O., 2015. Ignimbrites to batholiths: Integrating perspectives

696

from geological, geophysical, and geochronological data. Geosphere, 11,

697

705–743.

698

Liu, S.A., Li, D., Li, S., Teng, F.Z., Ke, S., He, Y., Lu, Y., 2014. High-precision copper

699

and iron isotope analysis of igneous rock standards by MC-ICP-MS. J. Anal. At.

700

Spectrom. 29, 122–133.

701

Liu, Y., Zhang, Y., Behrens, H., 2005. Solubility of H2O in rhyolitic melts at low

702

pressures and a new empirical model for mixed H2O-CO2 solubility in rhyolitic 33

703

melts. J. Volcanol. Geotherm. Res. 143, 219–235.

704

Louvel, M., Sanchez-Valle, C., Malfait, W.J., Cardon, H., Testemale, D., Hazemann,

705

J.L., 2014. Constraints on the mobilization of Zr in magmatic-hydrothermal

706

processes in subduction zones from in situ fluid-melt partitioning experiments.

707

Am. Mineral. 99, 1616–1625.

708 709

Mahood, G., Hildreth, W., 1983. Large partition coefficients for trace elements in high-silica rhyolites. Geochim. Cosmochim. Acta 47, 11–30.

710

Mao, J.W., Li, H.Y., Pei, R.F., 1995. Geology and Geochemistry of the Qianlishan

711

granite and its relationship to polymetallic tungsten mineralization. Miner. Depos.

712

14, 12–25 (in Chinese with English abstract).

713 714 715 716

Nash, W.P., Crecraft, H.R., 1985. Partition coefficients for trace elements in silicic magmas. Geochim. Cosmochim. Acta 49, 2309–2322. Poitrasson, F., Freydier, R., 2005. Heavy iron isotope composition of granites deter-mined by high resolution MC-ICP-MS. Chem. Geol. 222, 123–147.

717

Schauble, E.A., Rossman, G.R., Taylor, Jr.H.P., 2001. Theoretical estimates of

718

equilibrium Fe-isotope fractionations from vibrational spectroscopy. Geochim.

719

Cosmochim. Acta 65, 2487–2497.

720

Schuessler, J.A., Schoenberg, R., Sigmarsson, O., 2009. Iron and lithium isotope

721

systematics of the Hekla volcano, Iceland – evidence for Fe isotope fractionation

722

during magma differentiation. Chem. Geol. 258, 78–91.

723

Shahar, A., Young, E.D., Manning, C.E., 2008. Equilibrium high-temperature Fe

724

isotope fractionation between

fayalite and

725

calibration. Earth Planet. Sci. Lett. 268, 330–338. 34

magnetite: an

experimental

726

Shan, Q., Liao, S.P., Lu, H.Z., Li, J.K., Yang, W.B., Luo, Y., 2011. Fluid inclusion

727

records from the magmatic-hydrothermal stage: a case study of Qitianling granite

728

pluton. Acta Petrol. Sin. 27, 1511–1520 (in Chinese with English abstract).

729

Shinohara, H., 1994. Exsolution of immiscible vapor and liquid phases from a

730

crystallizing silicate melt: Implications for chlorine and metal transport. Geochim.

731

Cosmochim. Acta 58, 5215–5221.

732

Shu, X. J., Wang, X.L., Sun, T., Chen, W.F., Shen, W.Z., 2013. Crustal formation in

733

the Nanling Range, South China Block: Hf isotope evidence of zircons from

734

Phanerozoic granitoids. J. Asian Earth Sci. 74, 210–224.

735

Shu, X.J., 2014. Petrogenesis and crustal evolution of the Mesozoic granites from

736

Nanling, South China. Ph. D. thesis, Nanjing University (in Chinese with English

737

abstract).

738

Shu, X.J., Wang, X.L., Sun T., Xu X., Dai, M.N., 2011. Trace elements, U-Pb ages

739

and Hf isotopes of zircons from Mesozoic granites in the western Nanling Range,

740

South China: implications for petrogenesis and W-Sn mineralization. Lithos 127,

741

468–482.

742

Sossi, P.A., Foden, J.D., Halverson, G.P., 2012. Redox-controlled iron isotope

743

fractionation during magmatic differentiation: an example from the Red Hill

744

intrusion, S. Tasmania. Contrib. Mineral. Petrol. 164, 757–772.

745

Sossi, P.A., O’Neill, H.S.C., 2017. The effect of bonding environment on iron isotope

746

fractionation between minerals at high temperature. Geochim. Cosmochim.

747

Acta 196, 121–143. 35

748

Spera, F.J., Bohrson, W.A., Till, C.B., Ghiorso, M.S., 2007. Partitioning of trace

749

elements among coexisting crystals, melt, and supercritical fluid during isobaric

750

crystallization and melting. Am. Mineral. 92, 1881–1898.

751

Stepanov, A., Mavrogenes, J.A., Meffre, S., Davidson, P., 2014. The key role of mica

752

during igneous concentration of tantalum. Contrib. Mineral. Petrol. 167, 1009.

753

Telus, M., Dauphas, N., Moynier, F., Tissot, F.L.H., Teng, F.Z., Nabelek, P.I.,

754

Craddock, P.R., Groat, L.A., 2012. Iron, zinc, magnesium and uranium isotopic

755

fractionation during continental crust differentiation: the tale from migmatites,

756

granitoids, and pegmatites. Geochim. Cosmochim. Acta 97, 247–265.

757 758

Teng, F.Z., Dauphas, N., Huang, S., Marty, B., 2013. Iron isotopic systematics of oceanic basalts. Geochim. Cosmochim. Acta 107, 12–26.

759

Wallace, P.J., Plank, T., Edmonds, M., Hauri, E.H., 2015. Volatile in magmas, in:

760

Sigurdsson, H., Houghton, B., McNutt, S., Rymer, H., Stix, J. (Eds.), The

761

encyclopedia of volcanoes. Elsevier, pp. 163–183.

762

Wang, R.C., Xie, L., Chen, J., Yu, A., Wang, L., Lu, J., Zhu, J. 2013. Tin-carrier

763

minerals in metaluminous granites of the western Nanling Range (southern China):

764

Constraints on processes of tin mineralization in oxidized granites. J. Asian Earth

765

Sci. 74, 361–372.

766

Wang, X.L., Coble, M.A., Valley, J.W., Shu, X.J., Kitajima, K., Spicuzza, M.J., Sun,

767

T., 2014. Influence of radiation damage on Late Jurassic zircon from southern

768

China: Evidence from in situ measurements of oxygen isotopes, laser Raman,

769

U–Pb ages, and trace elements. Chem. Geol. 389, 122–136. 36

770

Wang, X.W., Wang, X.D., Liu, J.Q., Chang, H.L., 2004. Relationship of Qitianling

771

granite to Sn mineralization in Hunan Province. Geol. Sci. Technol. Inf. 23, 1–12

772

(in Chinese with English abstract).

773

Wawryk, C.M., Foden, J.D., 2015. Fe-isotope fractionation in magmatic-hydrothermal

774

mineral deposits: a case study from the Renison Sn–W deposit, Tasmania.

775

Geochim. Cosmochim. Acta 150, 285–298.

776

Webster, J.D., Holloway, J.R., 1990. Partitioning of F and Cl between magmatic

777

hydrothermal fluids and highly evolved granitic magmas. Geol. Soc. Am. Spec.

778

Pap. 246, 21–34.

779

Webster, J.D., Holloway, J.R., Hervig, R.L., 1989. Partitioning of lithophile trace

780

elements between H2O and H2O+CO2 fluids and topaz rhyolite melt. Econ.

781

Geol. 84, 116–134.

782

Webster, J.D., Tappen, C.M., Mandeville, C.W., 2009. Partitioning behavior of

783

chlorine and fluorine in the system apatite–melt–fluid. II: Felsic silicate systems at

784

200 MPa. Geochim. Cosmochim. Acta 73, 559–581.

785 786

Wu, F., Liu, X., Ji, W., Wang, J., Yang, L., 2017. Highly fractionated granites: Recognition and research. Sci. China Earth Sci. 60, 1201–1219.

787

Xia, Y., Li, S., Huang, F., 2017. Iron and zinc isotope fractionation during magmatism

788

in the continental crust: evidence from bimodal volcanic rocks from Hailar basin,

789

NE China. Geochim. Cosmochim. Acta 213, 35–46.

790

Xie, L., Wang, R.C., Chen, J., Zhu, J.C., 2010. Mineralogical evidence for magmatic

791

and hydrothermal processes in the Qitianling oxidized tin-bearing granite (Hunan, 37

792

South China): EMP and (MC)-LA-ICPMS investigations of three types of

793

titanite. Chem. Geol. 276, 53–68.

794

Xu, L.J., He, Y., Wang, S.J., Wu, H., Li, S., 2017. Iron isotope fractionation during

795

crustal anatexis: Constraints from migmatites from the Dabie orogen, Central

796

China. Lithos 284, 171–179.

797

Ye, H., Wu, C.Z., Yang, T., Santosh, M., Yao, X.Z., Gao, B.F., Wang, X.L., Li, W.,

798

2017. Updating the Geologic Barcodes for South China: Discovery of Late

799

Archean Banded Iron Formations in the Yangtze Craton. Sci. Rep. 7, 15082.

800

Young, E.D., Manning, C.E., Schauble, E.A., Shahar, A., Macris, C.A., Lazar, C.,

801

Jordan, M., 2015. High-temperature equilibrium isotope fractionation of

802

non-traditional stable isotopes: Experiments, theory, and applications. Chem.

803

Geol. 395, 176–195.

804

Yu, J.H., Shen, W.Z., 2007. Compositions and ages of the basement metamorphic

805

rocks from Nanling Range, in: Zhou, X.M. (Eds), Origin of the late Mesozoic

806

granites in the Nanling Range and lithospheric geodynamics. Science Publishing

807

House, Beijing, pp. 23–48.

808

Zajacz, Z., Halter, W.E., Pettke, T., Guillong, M., 2008. Determination of fluid/melt

809

partition coefficients by LA-ICPMS analysis of co-existing fluid and silicate melt

810

inclusions: controls on element partitioning. Geochim. Cosmochim. Acta 72,

811

2169–2197.

812

Zambardi, T., Lundstrom, C.C., Li, X., McCurry, M., 2014. Fe and Si isotope

813

variations at Cedar Butte volcano: insight into magmatic differentiation. Earth 38

814

Planet. Sci. Lett. 405, 169–179.

815

Zhao, K.D., Jiang, S.Y., Jiang, Y.H., Wang, R.C., 2005. Mineral chemistry of the

816

Qitianling granitoid and the Furong tin ore deposit in Hunan Province, South

817

China: implication for the genesis of granite and related tin mineralization. Eur. J.

818

Mineral. 17, 635–648.

819

Zhao, K.D., Jiang, S.Y., Nakamura, E., Moriguti, T., Palmer, M.R., Yang, S.Y., Dai,

820

B.Z., Jiang, Y.H., 2011. Fluid-rock interaction in the Qitianling granite and

821

associated tin deposits, South China: evidence from boron and oxygen

822

isotopes. Ore Geol. Rev. 43, 243–248.

823

Zhao, K.D., Jiang, S.Y., Yang, S.Y., Dai, B.Z., Lu, J.J., 2012. Mineral chemistry, trace

824

elements and Sr-Nd-Hf isotope geochemistry and petrogenesis of Cailing and

825

Furong granites and mafic enclaves from the Qitianling batholith in the Shi-Hang

826

zone, South China. Gondwana Res. 22, 310–324.

827

Zhou, X., Sun, T., Shen, W., Shu, L., Niu, Y., 2006. Petrogenesis of Mesozoic

828

granitoids and volcanic rocks in South China: a response to tectonic

829

evolution. Episodes, 29, 26–33.

830

Zhu, J.C., Wang, R.C., Lu, J.J., Zhang, H., Zhang, W.L., Xie, L., Zhang, R.Q., 2011.

831

Fractionation, evolution, Petrogenesis and mineralization of Laiziling granite pluton,

832

southern Hunan province. Geol. J. China Univ. 17, 381–392 (in Chinese with English

833

abstract).

834 835

39

836

Figure captions

837

Fig. 1. Simplified geological map showing the location of the studied granites within

838

the Nanling area of southern China (modified after Wang et al., 2014). Three granitic

839

samples from the Qianlishan pluton (sample names starting as SZY-) were collected

840

from underground pit.

841 842

Fig. 2. Granite Fe isotopic composition as a function of (a) SiO2, (b) Fe2O3t (total iron

843

content), (c) Na2O, (d) Sr, (e) Ba, and (f) Ba/Rb. Error bars indicate ±2SD, and the

844

shaded area represents the range of δ56Fe values for mafic to intermediate igneous

845

rocks (0.09 ± 0.08‰; Heimann et al., 2008). Literature data are from Poitrasson and

846

Freydier (2005), Heimann et al. (2008), Schuessler et al. (2009), Telus et al. (2012),

847

Zambardi et al. (2014), Foden et al. (2015), Gajos et al. (2016), He et al. (2017), Xia

848

et al. (2017), and Du et al. (2017).

849 850

Fig. 3. Diagrams showing variations in whole-rock Fe isotopic compositions versus

851

indicators of fluid exsolution in the form of (a) Zr/Hf, (b) Nb/Ta, and (c) K/Rb ratios,

852

and (d) TE1,3 (Degree of the REEs tetrad effect, Irber, 1999). Shaded areas are

853

indicative of rocks recording interaction with aqueous fluids (K/Rb < 100: Clarke,

854

1992; Zr/Hf < 26: Bau, 1996; Na/Ta < 5, Ballouard et al., 2016; TE1,3 >1.1: Irber,

855

1999) and the two outliers from the Qitianling granites with TE1,3 values > 1.1 are

856

caused by the presence of positive Ce anomalies. (e) Plot of whole-rock P2O5 versus

857

Nd for the granitic samples. (f) Plot of P2O5 versus Y for the granitic samples. Grey 40

858

squares, diamonds and triangles in Fig 3e and Fig. 3f represent granites from Laiziling,

859

Qitianling and Qianlishan plutons, respectively (Table S1of the supplements; Shu,

860

2014).

861 862

Fig. 4. (a) Plot of Fe2O3t versus F/Cl for the Nanling granitic samples. (b) Diagram

863

showing variation in F/Cl versus δ56Fe. Grey diamonds and triangles in Fig 4a

864

represent granites from Qitianling and Qianlishan plutons, respectively (Table S1of

865

the supplements; Shu, 2014).

866 867

Fig. 5. (a) Diagram showing variations in whole-rock Fe2O3t versus U for the

868

Qitianling granites. Whole-rock data are from Wang et al. (2004), Bai et al. (2005),

869

Deng et al. (2005), Xie et al. (2010), Zhao et al. (2012), and Shu et al. (2014). The

870

blue line indicates a liquid line of descent calculated by fractional crystallization

871

modeling using bulk crystal–melt partition coefficients of 1.8 and 0 for Fe and U,

872

respectively (estimated using Fig. 5b). Crosses indicate 10 wt.% fractionation steps.

873

(b) Diagram showing variations in Fe2O3t versus melt fraction  relative to parental

874

composition (red star; U = 7.1 ppm, Fe2O3t = 6.06 wt.%).  values were estimated

875

using the inverse of the enrichment factor of the incompatible element U. Average

876

Fe2O3t values (red diamonds) were calculated for every 0.1 variation in  values.

877

Blue curve shows the best-fit line for the Qitianling data with a partition coefficient

878

between bulk minerals and melt of 1.8.

879 41

880

Fig. 6. Diagrams showing the predicted compositional evolution of residual melts

881

within magmas with different initial H2O contents. (a)  versus H2O diagram

882

showing the concentration of H2O in the residual melt. H2O is regarded as a perfectly

883

incompatible component, and water solubility was estimated from the empirical

884

equation proposed by Holtz et al. (2001). (b) Diagram showing the seven different

885

Fe2O3t curves obtained for seven scenarios involving different initial H2O contents.

886

Red stars show the mass fraction of melt at water saturation with different initial H2O

887

contents. (c) Diagram showing differences between 56   (fluid containing Fe)

888

and 56 0-   (fluid without containing Fe) values expressed as

889

56

against  values (see text for details).

′ plotted

890 891

Fig. 7. (a) Diagram showing variations in F/Cl versus melt fraction ( ) for the

892

Nanling granites. Whole-rock geochemical data are from Shu (2014) (Table S1 of the

893

supplements). (b) Diagram showing the effect of fluid exsolution (

894

elevating δ56Fe values during magmatic–hydrothermal processes within the

895

crystallizing Nanling granites ( ; details of calculations are given in the text).

56

′ ) in

896 897

Fig. 8. (a) Rayleigh fractionation models for bulk-rock Fe2O3t (wt%) and Fe isotopes

898

(δ56Fe, ‰). Black line shows the predicted paths of the melt expected from magnetite

899

crystallization and removal, using the Sossi et al. (2012) fractionation factor. The

900

green line illustrates the effect of biotite crystallization and separation, and the

901

fractionation factor was determined as –0.1‰ based on average ∆56FeBt–whole rock (Du 42

902

et al., 2017). The two orange lines show the visual fit for the Nanling granites with

903

bulk mineral-melt fractionation factor ranging from –0.08‰ to –0.04‰, with Fe2O3t

904

mineral-melt partition coefficient of 1.8 (as shown in Fig. 5). The initial magma is

905

assumed to have a δ56Fe of 0.1‰ and 6.06 wt% Fe2O3t.

906 907

Fig. 9. Diagrams comparing the Fe isotopic compositions of ore-related and barren

908

granites (a–b) and volcanic and plutonic rocks (c–d). Data are from Poitrasson and

909

Freydier (2005), Heimann et al. (2008), Schuessler et al. (2009), Zambardi et al.

910

(2014), Sossi et al. (2012), Telus et al. (2012), Foden et al. (2015), Gajos et al. (2016),

911

He et al. (2017), Xia et al. (2017), and Du et al. (2017). The data in (c–d) are average

912

δ56Fe values calculated for every 2 wt.% variation in SiO2 and 1 wt.% in Fe2O3t,

913

meaning that the average value at SiO2 = 61 wt.% was calculated from published

914

δ56Fe values for granites with SiO2 contents of 60–62 wt.%. Uncertainties in δ56Fe are

915

given as 2SE values.

43

113°00′

112°25′

113°40′

North China

Chenzhou n

e

Jia

Changsha

C

h at

ay

si

0

Z-Є

10km

25°50′

Shizhuyuan W-Sn-MoBi polymetallic ore deposit

O ro g e n

ng

Y

tz

na

g an

07QL-02 07QL-03-2 07QL-05-2

a Xintianling W polymetallic ore deposit

Qianlishan

NN

Yaogangxian

Qitianling

D-T 1

25°35′ FR-05 FR-08

Yaogangxian W-Pb-Zn polymetallic ore deposit

07QTL-01

Xianghualing W-Sn-Pb-Zn polymetallic ore deposit

YGX-01

Z-Є

FR-01 BLS-3 AY-05

Baiyunxian W-Pb-Zn polymetallic ore deposit

YiZhang

Z-Є Laiziling

Z-Є

LZL-02 LZL-04 LZL-07 LZL-09

Furong Sn ore deposit

Sinian-Cambrian strata Late Jurassic granites

D-T 1

T 3 -E Devonian to Lower Triassic strata Metal deposit

Fig. 1

25°20′

T 3 -E

Upper Triassic to Paleocene strata Sample location

0.5

0.5

a

0.4

b

Qianlishan Yaogangxian

0.4

Qitianling

0.3

Literature data

δ 5 6 Fe (‰)

δ 5 6 Fe (‰)

Laiziling

0.3

0.2

0.2

0.1

0.1

0.0

0.0 65

68

71

74

77

0

80

1

2

0.5

3

4

5

6

Fe 2 O 3 t (wt%)

SiO 2 (wt%) 0.5

c

0.4

d

Qianlishan Yaogangxian

0.4

Qitianling

δ 5 6 Fe (‰)

δ 5 6 Fe (‰)

Laiziling

0.3

0.2

0.1

0.3

0.2

0.1

0.0 0.0

0.0 1.0

2.0

3.0

4.0

0

5.0

50

100

0.5

150

200

Sr (ppm)

Na 2 O (wt%) 0.5

e

f

Qianlishan Yaogangxian

0.4

0.4

Qitianling

δ 5 6 Fe (‰)

δ 5 6 Fe (‰)

Laiziling

0.3

0.2

0.3

0.2

0.1

0.1

0.0 0

100

200

300

400

500

600

0.0 0.004

Ba (ppm)

0.040

0.400

Ba/Rb

Fig. 2

4.000

0.5

0.5

a

b

Qianlishan Yaogangxian

0.4

0.4

Qitianling

δ Fe (‰)

0.3

56

56

δ Fe (‰)

Laiziling

0.2

0.1

0.3

0.2

0.1

Fluid exsolution

0.0

Fluid exsolution

0.0 10

15

20

25

30

35

40

0

3

6

0.5

9

12

Nb/Ta

Zr/Hf 0.5

c

d

Qianlishan Yaogangxian

0.4

0.4

Qitianling

δ Fe (‰)

56

56

δ Fe (‰)

Laiziling

0.3

0.2

0.3

0.2 outlier

0.1

0.1

Fluid exsolution

Fluid exsolution

0.0 0

40

80

120

160

0.0

200

0.9

1.0

K/Rb

80

Nd (ppm)

f

Qianlishan

70

e zit n na tio mo para se

60

Yaogangxian

350

Qitianling

300

Laiziling

50 40 30

250

e tim n no tio xe para se

200 150 100

20

50

10 0 0.00

1.2

400

e

Y (ppm)

90

1.1

TE 1 , 3

0 0. 05

0. 10

0.15

0.20

0. 25

P 2 O 5 (wt%)

0. 00

0.05

0. 10

0.15

P 2 O 5 (wt%)

Fig. 3

0.20

0. 25

180

a

160

Yaogangxian

120

Qitianling Laiziling

Flui

100

d ex

80

solu

F/Cl

Qianlishan

140

tion

60

Frac tiona l crys talliz ation

40 20 0 0

1

2

3

4

5

6

Fe 2 O 3 t (wt%) 0.5

b

Qianlishan Yaogangxian

0.4

Qitianling

δ 5 6 Fe (‰)

Laiziling

0.3

Flu id exs olu tio n

0.2

0.1 Fractional crystallization 0.0 3

30

F/Cl

Fig. 4

300

45

a

40

U (ppm)

35 30 25 20 15 10 5 0 0.5

1.5

2.5

3.5

4.5

5.5

6.5

t

Fe 2 O 3 (wt%) 7

b

Fe 2 O 3 t (wt.%)

6 5 4 3 2

D Fmei n - m e l t =1.8

1 0 10

20

30

40

50

60

F m (wt%)

Fig. 5

70

80

90

100

H 2 O m e l t (wt.%)

9 8

a

7

H 2 O saturation: 8.4 wt.%

6 5

initial H 2 O content:

4

1 (wt%) 2 (wt%)

3

3 (wt%) 4 (wt%)

2

5 (wt%) 6 (wt%)

H 2 O content in residual melt

3.6 kbar

1 1

10

100

7.00

Fe 2 O 3 t (wt.%)

b

36

48

60

72

24 s

F m =12 wt%

Fe 2 O 3 t content in residual melt

0.70

initial H 2 O content: 0 (wt%) 1 (wt%) 2 (wt%) 3 (wt%) 4 (wt%) 5 (wt%) 6 (wt%)

0.07 1

10

100

0.21 initial H 2 O content:

56

Δ Fe’ (‰)

0.18

1 (wt%) 2 (wt%)

0.15

3 (wt%) 4 (wt%)

0.12

5 (wt%) 6 (wt%)

0.09 0.06 0.03

c

0.00 1

5

10

20

F m (wt%)

Fig. 6

40

100

0.08

160

a

140

Timing of fluid saturation

Yaogangxian

120

Δ 5 6 Fe’ (‰)

F/Cl

80

0.06

Qitianling

100

Laiziling Fluid exsolution

60 40

Qianlishan Qitianling

b

Qianlishan

Fra ctio nal cry sta lliz

atio n

60

90

Range of contributions of fluid exsolution to Fe isotope fractionation in Nanling granites 0.04

0.02

20 0 10

20

30

40

50

70

80

100

0.00 10

15

20

25

F m (wt%)

F m (wt%)

Fig. 7

30

35

40

0.5 Qianlishan

56

∆ Fe Bt-melt = -0.10‰

Yaogangxian

0.4

Qitianling

δ 5 6 Fe (‰)

Laiziling

0.3 56

∆ Fe min-melt = -0.08‰

0.2 56

∆ Fe min-melt = -0.04‰

0.1

56

Sossi et al. (2012) ∆ Fe Mt-melt @ 800℃

0.0 0

1

2

3 t

Fe 2 O 3 (wt%)

Fig. 8

4

5

6

0.6

0.6

a

b 0.5

δ 5 6 Fe (‰)

granites

0.4

ore-related granites

0.3

56

δ Fe (‰)

0.5

0.4 0.3

0.2

0.2

0.1

0.1

0.0 60

0.0 62

64

66

68

70

72

74

76

78

0

80

1

2

Plutonic

0.40

Volcanic

0.35

0.30

δ Fe (‰)

0.25 0.20

56

δ 5 6 Fe (‰)

5

6

0.45

c

0.35

0.15

d

Δ 5 6 Fe v o l c a n i c - p l u t o n i c (‰)

0.30 0.25 0.20 0.15

0.10

0.10

0.05

0.05

0.00 60

4 t

0.45 0.40

3

Fe 2 O 3 (wt%)

SiO 2 (wt%)

0.00 62

64

66

68

70

72

74

76

78

0

1

2

3 t

SiO 2 (wt%)

Fe 2 O 3 (wt%)

Fig. 9

4

5

6

Table 1 Whole-rock iron isotopic compositions of granites in this study Rock type

Sample

Rock-forming minerals

δ56Fe (‰)

δ57Fe (‰)

± 2SD

± 2SD

SiO2

TiO2

Al2O3

Fe2O3

MnO

MgO

CaO

Na2O

K2O

P2O5

LOI

Major element compositions (wt%)

Qianlishan pluton granite

07QL-02

6% Bt + 30% Qtz + 20% Pl + 44% Kfs

0.29 ± 0.05

0.46 ± 0.15

74.28

0.20

13.09

1.89

0.03

0.17

1.39

3.25

4.94

0.06

0.67

granite

07QL-03-2

4% Bt + 35% Qtz + 21% Pl + 40% Kfs

0.42 ± 0.01

0.55 ± 0.04

73.80

0.19

13.40

0.88

0.01

0.12

1.73

2.75

5.79

0.05

1.01

granite

07QL-05-2

3% Bt + 30% Qtz + 20% Pl + 47% Kfs

0.31 ± 0.03

0.48 ± 0.15

75.86

0.06

13.21

1.15

0.03

0.00

0.51

4.02

4.36

0.02

0.87

granite

SZY-06

5% Mus + 2% Bt + 35% Qtz + 18% Pl + 40% Kfs

0.40 ± 0.07

0.54 ± 0.03

75.52

0.07

12.97

1.03

0.03

0.06

0.75

3.23

4.79

0.02

1.16

granite

SZY-10

1% Mus + 2% Bt + 35% Qtz + 15% Pl + 47% Kfs

0.37 ± 0.06

0.54 ± 0.11

74.86

0.06

13.73

0.75

0.03

0.00

0.47

3.49

5.68

0.01

0.83

granite

SZY-13

3% Mus + 2% Bt + 38% Qtz + 20% Pl + 37% Kfs

0.38 ± 0.06

0.57 ± 0.09

76.19

0.06

13.88

0.93

0.02

0.14

0.17

1.28

5.54

0.02

1.64

Qitianling pluton granite

07QTL-01

5% Amp + 10% Bt + 25% Qtz + 25% Pl + 35% Kfs

0.12 ± 0.04

0.18 ± 0.06

66.59

0.68

14.11

5.28

0.09

0.82

2.86

3.30

4.75

0.23

1.07

granite

AY05

3% Amp + 8% Bt + 25% Qtz + 24% Pl + 40% Kfs

0.17 ± 0.06

0.32 ± 0.15

68.12

0.81

13.69

4.94

0.09

1.07

2.64

3.12

4.64

0.22

0.80

granite

FR01

8% Bt + 30% Qtz + 20% Pl + 42% Kfs

0.19 ± 0.02

0.24 ± 0.04

71.96

0.34

13.76

2.19

0.04

0.28

1.80

3.10

5.24

0.08

0.92

granite

FR05

3% Bt + 32% Qtz + 20% Pl + 45% Kfs

0.24 ± 0.02

0.34 ± 0.05

75.72

0.22

12.02

1.93

0.04

0.15

0.74

2.85

4.89

0.06

1.23

granite

FR08

3% Bt + 35% Qtz + 17% Pl + 45% Kfs

0.24 ± 0.06

0.38 ± 0.06

76.96

0.21

11.59

1.71

0.03

0.12

0.86

2.64

4.76

0.06

1.25

granite

BLS-3

2% Bt + 35% Qtz + 15% Pl + 48% Kfs

0.20 ± 0.06

0.26 ± 0.07

75.94

0.19

11.78

1.45

0.03

0.06

0.75

2.46

5.95

0.03

1.13

Laiziling pluton granite

LZL-02

2% Mus + 35% Qtz + 13% Pl + 50% Kfs

0.20 ± 0.06

0.39 ± 0.11

74.74

0.07

13.81

1.34

0.08

0.00

0.46

3.81

4.31

0.02

1.03

granite

LZL-04

3%Bt + 36% Qtz + 15% Pl + 46% Kfs

0.24 ± 0.06

0.34 ± 0.07

74.18

0.07

13.85

1.10

0.03

0.03

0.49

4.39

4.72

0.02

0.70

granite

LZL-07

5%Bt + 35% Qtz + 18% Pl + 42% Kfs

0.27 ± 0.08

0.38 ± 0.15

75.90

0.09

12.59

1.35

0.04

0.17

0.65

3.47

4.55

0.02

1.00

granite

LZL-09

3%Bt + 35% Qtz + 22% Pl + 40% Kfs

0.24 ± 0.04

0.37 ± 0.14

74.75

0.07

13.62

0.95

0.03

0.02

0.39

3.66

5.93

0.02

0.58

2%Bt + 3 wt% Mus + 32% Qtz + 20% Pl + 43% Kfs

0.27 ± 0.03

0.43 ±0.05

75.17

0.07

13.25

0.92

0.10

0.00

0.62

4.00

4.32

0.02

1.01

Yaogangxian pluton granite

YGX-01

Note: The major element data are from Shu (2014). Amp = amphibole, Bt = biotite, Mus = muscovite, Tur = tourmaline, Qtz = quartz, Kfs = K-feldspar, Pl = plagioclase.

Highlights 1. The effect of fluid exsolution on Fe isotope fractionation remains unclear. 2. Fluid exsolution contributes little to Fe isotopic fractionation in granites. 3. Fractional crystallization is the main cause of Fe isotope fractionation. 4. Iron isotopes can be used to trace magmatic differentiation of felsic magma chamber.

Declaration of interests ; The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: