Sulfides, Native Metals, and Associated Trace

8 Sulfides, native metals, and associated trace minerals of the Skaergaard 9 Intrusion, Greenland: Evidence for late hydrothermal fluids 10 11 Ben Wernette1†, Peishu Li1, and Alan Boudreau1 12 1Division of Earth and Ocean Sciences, Duke University, Durham, N.C., 27705, USA 13 †Corresponding Author: [email protected] 14 15

24 Abstract

25 Sulfide assemblages, precious metals, transition metal alloys, and associated accessory

26 phases were characterized throughout the Skaergaard Intrusion to better constrain the sulfide

27 saturation history of the intrusion and the role of late magmatic volatiles in modifying the

28 Skaergaard metal budget and distribution. Sulfides in and below the Middle Zone of the layered

29 series of the intrusion are readily replaced by low-Ti magnetite; in thin section, the ratio of the

30 (low-Ti magnetite mode)/(sulfide mode), indicating oxidation of sulfides, reaches maximum

31 values in samples of the Lower Zone of the Layered Series. Sulfide assemblages below the

32 Middle Zone are typically accompanied by minor biotite, apatite, and rare calcite as well as trace

33 compositionally distinctive clinopyroxene and orthopyroxene. The occurrence of Ag, Au, Pt, Cu

34 and metal alloys outside of the Middle Zone is further evidence of the Skaergaard Intrusion

35 parental magma being S-poor.

36 Native Ag, commonly accompanied by trace amounts of Cl, occurs both in and below the

37 Middle Zone. Evidence of coexisting precious metal + brine assemblages exists where native

38 metals are accompanied by sylvite ± halite and Ag is accompanied by Ag-halides. Ag

39 occurrences in the Middle Zone are of irregular morphology with trace Cl ± S ± calcite. Further

40 evidence supportive of a metal + brine assemblage is observed where Ag + quartz are found in

41 an apparent open fluid inclusion, in clinopyroxene, consisting of Na, Si, Cl, Ca, K, and S.

42 Ag is used to model the behavior of precious and transition metals in the presence of an

43 exsolving fluid phase. Numerical modeling suggests that, in a sulfide bearing system, residual

44 Ag concentrations and concentrations in the exsolved fluid are most affected at the point where

45 sulfide is lost to a separating volatile fluid phase. It is suggested that, owing to the low S nature

46 of the Skaergaard system, fractional crystallization and early fluid saturation produced

47 enrichment of Ag, with other precious and transition metals, in the interstitial silicate liquid

48 much higher than normal due to delayed sulfide saturation. As this interstitial liquid evolved, Ag

49 was lost to an exsolved volatile phase of high salinity and migrated upward. A similar process

50 likely occurred for Au and other elements with high affinities for Cl.

52 1. Introduction

53 In layered mafic intrusions (LMIs), evidence for the existence of volatile phases is

54 preserved in fluid inclusions, apatite composition, and halogen geochemistry (e.g., Boudreau and

55 McCallum 1989; Hanley et al. 2008). Of debate, however, is whether these fluids play a

56 significant role in the processes responsible for generating, or modifying, the stratiform

57 platinum-group element (PGE)-sulfide ore bodies for which LMIs are known. Conventionally,

58 these deposits are thought to form as the result of gravitationally driven downward movement of

59 immiscible sulfide liquid droplets in the magma chamber. Large sulfide liquid/silicate liquid

60 partition coefficients (e.g., Mungall and Brenan 2014) and large silicate liquid/sulfide liquid

61 mass ratios (R-factor, Campbell and Naldrett 1979) allow for the extreme enrichment of

62 transition and noble metals as well as PGE in immiscible sulfide droplets. Experimental results

63 confirm that these are reasonable assumptions (Fleet et al. 1993).

64 Conversely, others have suggested that stratiform mineralization is the result of upward

65 moving exsolved fluid, rich in halides, allowing for the efficient transport of transition and

66 precious metals vertically through an igneous body (e.g., Boudreau and McCallum 1992).

67 Theoretical work (Shinohara 1994) supports this model and experimental work performed on

68 rhyolitic and granitic systems (Candela and Holland 1984; Simon et al. 2008; Frank et al. 2011)

69 and theoretical work done on LMIs (Boudreau and McCallum 1992; Meurer et al. 1999) suggest

70 the importance of exsolved halide-bearing volatiles. Together these end-member models are

71 referred to as the “orthomagmatic” and “hydromagmatic” models, respectively.

72 Complicating our understanding of LMIs is that the larger intrusions (i.e., Bushveld and

73 Stillwater Complex) show evidence of multiple magma injections and prolonged cooling

74 histories for which the preservation of original igneous textures and chemistry is uncertain. It is

75 generally agreed that the small Paleogene Skaergaard intrusion of southeast Greenland cooled

76 and crystallized as a closed system (Holness et al. 2007). Of both scientific and commercial

77 interest is the fact that the Skaergaard intrusion contains a zone enriched in Au and PGEs known

78 as the Platinova Reef (Andersen et al. 1998; Holwell and Keays, 2014; Godel et al. 2014;

79 Nielsen et al. 2015). It is because of this that the Skaergaard intrusion provides a unique

80 opportunity to investigate how metalliferous LMIs evolve through time. In this study, we

81 examine sulfide assemblages and their associated phases and characterize the occurrence of

82 precious and transition metals outside of the Platinova Reef to better understand the role of

83 exsolved volatiles in the distribution and modification of metals in LMIs and the sulfide

84 saturation history of the Skaergaard Intrusion.

86 2. Geologic Background and Summary of Previous Studies

87 2.1 Skaergaard Intrusion

88 For extensive reviews, readers are directed to the early studies of Wager and Deer (e.g.,

89 Wager and Deer 1939) and the more recent summary of Nielsen (2004) and references therein.

90 Briefly, the Skaergaard Complex is a ~ 55 Ma (Brooks and Gleadow 1977; Hirschman et al.

91 1997) intrusion located in southeast Greenland (Fig. 1). The intrusion formed during the

92 Paleogene rifting of Greenland from Eurasia (Nielsen 1975). The Complex is hosted

93 uncomformably in Archean gneiss and, from bottom to top, is composed of the Layered and

94 Upper Border Series. At its margin, the Complex is composed of the Marginal Border Series .

95 The Layered Series is thought to have grown upward from the floor of the intrusion while the

96 Upper Border and Marginal Border Series crystallized from the roof and the walls of the

97 intrusion inward, respectively. The Layered Series is subdivided into the Lower (LZ), Middle

98 (MZ), and Upper Zone (UZ) according to the presence of index minerals.

99 For the MZ, Wager et al. (1957) describe a Cu-rich sulfide assemblage consisting of

100 chalcopyrite and bornite. The upper ~ 100 m of the MZ is host to the Triple Group, a sequence of

101 layered gabbros (Nielsen et al 2015) that contain zones of anomalously high Au and PGE

102 concentrations (Andersen et al. 1998). Individual layers within the Triple Group have been

103 correlated across the Intrusion (e.g., Andersen et al. 1998; Holwell and Keays 2013; Nielsen et

104 al. 2015) and discrete Pd-, Au-, and Cu-rich subzones, or “offsets”, have been identified (e.g.,

105 Bird et al. 1991; Andersen et al. 1998; Nielsen et al. 2005; Holwell and Keays, 2014).

106 Collectively, these metalliferous subzones are known as the Platinova Reef. Several petrogenetic

107 models for the Platinova Reef have been proposed and broadly align with the orthomagmatic or

108 hydromagmatic end-member models. Previous workers have suggested that the Platinova Reef is

109 the result of Rayleigh-fractionation processes acting to concentrate PGEs in sulfide (e.g.,

110 Holwell and Keays 2014). In this model, large 퐷푠푢푙/푠푖푙 values act to concentrate PGE in early

111 formed layers (Pendergast 2000; Holwell and Keays 2014). To explain high PGE tenors in

112 sulfide and the observed Pd, Au, Cu offsets of the Platinova Reef, Holwell and Keays (2014)

113 proposed a multi-stage sulfide saturation model whereby early formed sulfide droplets

114 concentrate chalcophile elements as they settle through the magma chamber. In their model, re-

115 dissolution of this early sulfide as it moves into hotter magma along the floor of the chamber acts

116 to concentrate chalcophile elements in the magma. This is followed by a second local (or in-situ)

117 sulfide saturation event that generates low-volume high-tenor sulfides. The observed Pd, Au, and

118 Cu offsets are thought by Howell and Keays (2014) to reflect variations in 퐷푠푢푙/푠푖푙. Indeed,

119 Godel et al. (2014) found textural evidence supportive of in-situ sulfide nucleation in the

120 Platinova Reef. Andersen et al. (1998) suggested that compaction of cumulates forced sulfides to

121 migrate vertically generating the observed metal offsets while potentially oxidizing existing

122 sulfides. Analogous to the model of Andersen et al. (1998), Nielsen et al. (2015) proposed that

123 the precious metal distribution observed in the Platinova Reef can be explained by the upward

124 migration of Fe-rich melts rich in volatiles and precious metals. Andersen (2006) proposed that

125 late hydrothermal fluids transported precious metals vertically along grain boundaries and fluid

126 pathways. In this model, redox barriers limit the PGE carrying capacity of the hydrothermal fluid

127 acting to separate the stratigraphic occurrence of Pd from Au.

128 Anomalously high Cu/S ratios observed throughout the Skaergaard stratigraphy (Cu/S = 1 –

129 7, Keays and Tegner 2016) have been attributed to early shallow-level degassing and S loss (Li

130 and Boudreau 2017), late S loss to hydrothermal fluids (Andersen 2006), magmatic oxidation

131 (Wohlgemuth-Ueberwasser et al. 2013), or anomalous Cu/S ratios in the Skaergaard parental

132 magma source region (Keays and Tegner 2016). Andersen (2006) concluded that sulfide texture

133 and mineralogy in the LZ is well explained by an exsolved hydrothermal fluid oxidizing and

134 replacing preexisting sulfides with low-Ti magnetite and redistributing Cu, S, and precious

135 metals to different stratigraphic levels. Li and Boudreau (2017) arrived at a similar conclusion

136 after conducting a modal analysis of sulfides found in the Lower Zone and Marginal Border

137 Series. Conversely, Godel et al. (2014) note that LZ sulfides are generally not accompanied by

138 low-Ti magnetite, ultimately suggesting that late hydrothermal fluids did not oxidize magmatic

139 sulfide assemblages.

140 There are a number of detailed studies of precious metal distribution through the

141 Platinova Reef (e.g., Bird et al. 1991; Godel et al. 2014; Nielsen et al. 2015; Holwell et al. 2016;

142 Keays and Tegner 2016). These studies focus on characterizing the volumetrically significant

143 Cu, Au, Pd, and Pt occurrences within the Reef. In the Platinova Reef, researchers have observed

144 the occurrence of Ag in native form (Nielsen et al. 2015), electrum (Holwell et al. 2015), Au-Cu-

145 Ag alloys (Bird et al. 1991 and Cabri et al. 2005), as well as Au-PGE alloys (Andersen et al.

146 1998). Additionally, Holwell et al. (2015) note Ag concentrations of ~ 250 ppm in sulfides from

147 the Triple Group. Broadly, Reef occurrences of Au include native Au (Bird et al., 1991), tetra-

148 auricupride (AuCu, Cabri et al. 2005; Nielsen et al. 2015), and Au-Cu-Pd alloys (Nielsen et al.

149 2015). Pd and Pt form the precious metal minerals (PMMs) skaergaardite and nielsenite while

150 Cu is a constituent of PMMs or metal alloys. Away from the Plainova reef, Andersen et al.

151 (2017) note the occurrence of Ag in Au-Cu-Ag alloys found in marginal basement schists that

152 formed as the Skaergaard magma interacted with its Archean host rock. In addition, they also

153 found electrum (Au-Ag) and various Bi-bearing PMM.

154

155 3. Sampling and Methods

156 A. Boudreau collected samples in 1990. Three samples, SK90-5, SK90-9, and SK90-13,

157 were published in a previous study (Li and Boudreau 2017) and were re-examined and included

158 in this study for completeness. In total, 11 polished thin-sections were systematically scanned

159 using the Cameca CAMEBAX microprobe at Duke University. Sulfide assemblages, and

160 accessory phases were characterized using energy dispersive spectrometry (EDS, accelerating

161 voltage = 20 KeV, beam current = 15 nA) and X-ray composite maps. The same methods were

162 used to identify and characterize the occurrences of Ag, Au, Pt, Cu and metal alloys. Sulfides,

163 native metals, metal alloys, and accessory minerals were imaged and imported into the photo

164 processing software IMAGE-J (Schneider et al. 2012) for area and perimeter measurements.

165 Where reported, semi-quantitative plagioclase data were collected using EDS. Uncertainties are

166 An# ± 4. For false color composite map figures, readers are directed to Fig. 2 for a list of relevant

167 elements and their corresponding colors.

168

169 4. Results

170 4.1 Summary of Stratigraphic Trends

171 As summarized in Fig. 3 and discussed in detail below, stratigraphic trends show a

172 general decrease in chalcopyrite (and a corresponding increase in bornite) from LZ through MZ

173 before chalcopyrite and digenite join the sulfide assemblage in the UZ. Similarly, the relative

174 modal abundance of low-Ti magnetite as a fraction of the sulfide assemblage is highest in LZ

175 samples (Fig. 3). The abundance of late hydrous minerals (e.g., biotite and amphibole), are

176 typically higher in and below the MZ and broadly parallels the bulk rock loss on ignition (LOI)

177 stratigraphic trends reported by Keays and Tegner (2016). The sulfide-magnetite assemblages

178 occur at both grain boundaries and within primocryst silicate minerals. Further, the relative

179 proportion of accessory minerals (calcite, biotite, apatite and compositionally distinctive

180 orthopyroxene and clinopyroxene with no exsolution lamellae and little iron content, see Fig. 4a)

181 as a function of a given sulfide assemblage is greatest in those samples from the LZ.

182 Occurrences of native Ag, Au, Pt, and Cu are generally restricted to MZ and LZ samples

183 and native Ag is far more common than Pt or Au in terms to total grains observed (12 grains of

184 Ag vs 3 for Au and Pt, respectively). Similarly, halite and sylvite were found only in those

185 samples from and below the MZ. Native Ag is commonly associated with trace quantities of Cl.

186 Although uncertain owing to the low total number of grains observed, there is a suggestion that

187 the size and total number of individual Ag grains per thin section broadly increases up

188 stratigraphy. A brief description of the individual sulfide (n = 165) and metal and metal-alloy (n

189 = 18) occurrences are presented in tables 1 and 2, respectively. Comprehensive descriptions of

190 individual sulfide assemblages can be found in Supplementary Table 1 (sulfides occurring in late

191 crosscutting albite + titanite veins are not included in this dataset).

192

193 4.2 General Description of Sulfide Assemblages and Associated Trace Minerals

194 LZa: Sulfides, mainly chalcopyrite and bornite, are typically found interstitial to olivine

195 and plagioclase. Pentlandite is observed in limited quantities and generally comprises a small

196 proportion of individual assemblages (average of ~ 4 vol. %). Sulfides are commonly non-

197 uniform in morphology and are commonly replaced, in part, by low-Ti magnetite (Fig. 4a, b).

198 Native Cu is observed where sulfides are extensively replaced by low-Ti magnetite (Fig. 4a).

199 Hydrous minerals, such as biotite and amphibole, are observed truncating sulfides or mantling

200 sulfide assemblages (Fig. 4b, c). Minor calcite is found mantling sulfides with no evidence of

201 other low or high temperature silicate alteration. When observed, calcite is rounded and

202 commonly no greater than 20 µm in size. Anhedral clino- and orthopyroxene also locally mantle

203 sulfides. When adjacent, plagioclase is increasingly calcic near sulfide assemblages (An# ~ 66,

204 Fig. 4a) but less calcic away from sulfide assemblages (An# ~ 50).

205 LZb: Li and Boudreau (2017) reported general descriptions for LZb sulfides. Briefly,

206 chalcopyrite and bornite make up the entire sulfide assemblage with near equal proportions of

207 low-Ti magnetite replacing the sulfides.

208 LZc: Observed sulfide proportions by area average 9% chalcopyrite and 91% bornite.

209 Sulfides occur both within silicates (cores and margins) and at silicate ground boundaries. Low-

210 titanium magnetite is less abundant relative to LZa. Biotite truncates silicate grain boundaries or

211 crosscuts sulfides. Calcite was observed in a single sulfide assemblage. As for LZa, plagioclase

212 becomes increasingly calcic towards sulfide assemblages where An# is similar to that observed in

213 LZa.

214 MZ: The sulfide assemblage below the Platinova Reef consists entirely of bornite with

215 variable amounts of low-titanium magnetite. It is important to note that this differs from the MZ

216 bornite and chalcopyrite assemblage reported by Wager et al. (1957). Sulfides are interstitial to,

217 or found entrained within, silicates. Sulfide morphology is typically irregular and is positively

218 correlated with the abundance of secondary silicates or low-Ti magnetite. Biotite and amphibole

219 are observed in limited quantities mantling sulfides, and calcite is only associated with native Ag

220 or Ag-halides. At pyroxene grain boundaries, olivine is locally observed enclosing sulfide

221 assemblages and, in some cases, low-Ti magnetite (Fig. 4d). When present, an increase in

222 plagioclase anorthite content in adjacent plagioclase is observed.

223 UZa: The sulfide assemblage is composed entirely of bornite (82%), chalcopyrite (13%),

224 and minor digenite (5%). Low-Ti magnetite is observed replacing sulfides in quantities similar to

225 that observed in LZc. Apatite is regularly associated with sulfides, typically surrounding the

226 periphery of the mineral. Sulfides are notably non-uniform in composition with Cu-rich, S-poor

227 areas (Fig. 4e). Anhydrite is observed in limited abundance where sulfides are in direct contact

228 with plagioclase. Additionally, plagioclase becomes increasingly calcic towards sulfide

229 assemblages. Near sulfide An# are as high as 72 (Fig. 4f)

230 UZb: Bornite and minor digenite comprise the entire sulfide assemblage. Sulfides are

231 commonly rounded and found at silicate-silicate or silicate-oxide grain boundaries and less

232 commonly within high-titanium magnetite. Additionally, sulfides are found interstitial to

233 modally abundant apatite. Low-Ti magnetite is seen in minor abundance and biotite is observed

234 in only one assemblage.

235

236 4.3 Occurrence of Precious and Transition Metals

237 Below the MZ, native Ag occurs as discrete rounded grains near the cores of major

238 minerals such as plagioclase, interstitial to late phases such as clinopyroxene and apatite (Fig.

239 5a), or at primocryst grain boundaries (Fig. 5b, c). In MZ samples, Ag occurs both as well-

240 rounded grains and irregular grains in silicate minerals (Fig. 5d) or at silicate grain boundaries.

241 Those occurrences of irregular morphology (6 out of 12) are accompanied by trace amounts of

242 Cl ± S ± calcite (Fig. 5d, e) and Ag-halides (Fig. 5d). Less common are Ag grains occurring with

243 Au intergrowths and trace Cu with notable trace quantities of Cl (Fig. 5e). Ag was not observed

244 above UZa. One occurrence of note (MZ), which is interpreted as a ruptured fluid inclusion, is

245 native Ag found in clinopyroxene with quartz and trace quantities of Na, Si, Cl, Ca, K, and S

246 (Fig. 6).

247 Grains of native Au or Pt are much less common than native Ag. Native Au occurs as

248 intergrowths (Fig. 5e) with Ag in the MZ and as singular occurrences with halite and sylvite in

249 plagioclase in LZa (Fig. 7). Native Pt was found within high-Ti magnetite from LZa. Further, Pt-

250 sulfides and Pt-Au alloys are observed at primocryst grain boundaries with trace Fe-Mg

251 amphibole and apatite in MZ samples (Fig. 8a). Of note is one occurrence of Pt in plagioclase in

252 UZb (Fig. 8b).

253 Native Cu is seen where significant replacement, or loss, of S from sulfide minerals

254 appears to have occurred (LZa, Fig. 5a). In some samples, however, Cu “colloids” are observed

255 at sulfide grain boundaries without evidence of sulfur loss or silicate replacement in LZ and UZ

256 samples (Fig. 5e). In MZ samples, Cu is observed in high-Ti magnetite and pyroxene with

257 associated trace Ag. Finally, trace quantities of Cu are observed in pyroxenes from LZb and

258 LZc and amphiboles from the MZ.

259 Approximately 70 % of the occurrences of precious and transition metals outside of the

260 mineralized Platinova Reef are accompanied by trace Cl or are spatially associated with hydrous

261 minerals such as biotite and amphibole. Furthermore, about 40 % of the occurrences are either at,

262 or close to, grain boundaries while the remaining 60 % occur in primocryst minerals. Finally,

263 except for one instance of Ag near sulfide (Fig. 4d), precious metals are not observed near

264 sulfide minerals, perhaps reflecting different genetic processes.

265

266

267 4.4 Fluid Inclusions

268 Fluid inclusions were not observed below the MZ; however the former presence of a high

269 temperature fluid is inferred from the presence of sylvite inclusions in clinopyroxene and

270 plagioclase primocrysts from LZa and LZc and halite crystals at silicate grain boundaries in LZa.

271 Within the MZ, fluid inclusions were observed as singular occurrences in pyroxene or as

272 inclusion trails in olivine (Fig. 9).

273 The fluid inclusions observed in pyroxene and inadvertently opened during thin section

274 preparation, are no greater than 50 µm in diameter. They contain Ag, Ag-Zn alloys, Cu and trace

275 quantities of Cl, Na, Ca, and K. Fluid inclusions observed in olivine can be categorized into

276 three principle groups that form inclusion trails, consisting of elongate or oblate shaped fluid

277 inclusions ~ 40 µm in diameter. One population of inclusions parallels olivine fracture surfaces,

278 while the other two are not associated with any identifiable fracture surface and differ only in

279 size and density of the inclusions (Fig. 9d). Broadly, inclusions in olivine number in the dozens

280 and are categorized as consisting of ~ 80 % vapor fill with a small proportion of fluid (Fig 9b, d).

281 Where observed, solids make up ~ 10 % of fluid inclusion volume (Fig. 9b). All olivine in

282 sample SK90-50 (MZ) host fluid inclusions.

283

284 5. Discussion

285 5.1 Sulfides and Associated Minerals

286 As noted above, the Ti-free magnetite has been previously interpreted to be replacements

287 of the original sulfide assemblage as the result of S-loss (e.g., Andersen 1998; Andersen 2006; Li

288 and Boudreau 2017); our observations are consistent with these interpretations. Specifically,

289 sulfides in and below UZa are commonly replaced by low-Ti magnetite as the result of S loss via

290 ƒO2-neutral reactions:

291 CuFeS2 + 1.15H2O → 0.27Fe3O4 + 0.19Cu5FeS4 + 1.15H2S + 0.04SO2 (1) 292 cpy vapor mt bn ------vapor ------293

294 and

295 CuFeS2 + 1.11H2O + 2HCl + SO2 → 0.33Fe3O4 + CuCl2 + 2.11H2S (2) 296 cpy ------vapor ------mt ---- vapor ---- 297 298 (Li and Boudreau 2017). Evidence of sulfur loss by these reactions is ubiquitous from LZa

299 through UZa.

300 Additionally, the abrupt decrease in secondary, or late hydrous minerals above the MZ is

301 consistent with loss on ignition (LOI) data reported by Keays and Tegner (2016) where values

302 approach 1 wt. % below UZa but are closer to 0.2 wt. % in, and above, UZa (Fig. 10).

303 Similarly, the enrichment in the An content in plagioclase suggests silica and Na loss to a

304 silica-under saturated fluid via:

+ 2+ + 305 2.0CaNaAlSi2.5O8 + 4.0H2O + 4.0H → CaAl2Si2O8 + 3.0H4SiO4 + Ca + 2.0Na (3) 306 plagioclase --- vapor --- anorthite ------vapor ------307

308 Support for this is the observation that calcium content, or An#, increase toward sulfide

309 assemblages in LZa through UZa (Fig. 4a, b, c, and f). Further, silica loss can produce olivine at

310 pyroxene grain boundaries via the reaction:

311 2.0(Mg0.5Fe0.5)SiO3 + 2.0H2O → (MgFe)2SiO4 + H4SiO4 (4) 312 orthopyroxene vapor olivine vapor 313

314 Similar observations have been made at the 2.7 Ga Stillwater Complex (Meurer et al. 1997)

315 where discordant troctolites were explained by silica loss to reactive fluids. Further work is

316 needed to confirm that olivine observed in MZ samples are metasomatic in nature though fluid

317 inclusions in olivine (Fig. 9) are consistent with this hypothesis.

318

319 5.2 Evidence for Volatiles in the Skaergaard Intrusion

320 The reactants involved in equations 1 through 4 require a sulfide-undersaturated volatile

321 phase to produce the observed assemblages. Direct evidence for a fluid in this study includes (i)

322 the observed fluid inclusions with associated Ag and quartz that retain trace quantities of Ca, Na,

323 Cl, and S (products seen in reactions in Section 5.1) and (ii) MZ clinopyroxene host Ag-Zn

324 alloys with Cu and trace Cl. Moreover, the association of Ag and Cl in MZ and LZ samples

325 suggests their concentration mechanism is the same. Further, native Au with proximal halite and

326 sylvite (LZa, Fig. 8) is supportive of a coexisting metal + brine assemblage. Finally, the

327 occurrence of Pt-Au at clinopyroxene-plagioclase grain boundaries with associated Fe-Mg

328 amphiboles suggests that PGE interacted with a hydrous fluid.

329 More broadly, Sonnenthal (1992) described Cl bearing fluid inclusions in late

330 anorthositic pegmatites of the UZ while Larsen et al. (1992) describe brine inclusions containing

331 up to ~ 22 wt. % NaCl in gabbroic pegmatites in and below the MZ. Sonnenthal (1992)

332 determined that Cl bearing fluid inclusions were trapped between 500 and 800°C while Larsen et

333 al. (1992) estimate a much narrower range of 655-770°C. These saline conditions and

334 temperatures might reflect those appropriate for the transport of Ag by an aqueous NaAgCl2

335 complex, as predicted by Yin and Zajacz (2018). Further, Larsen et al. (1992) suggest that fluid

336 inclusions present within interstitial phases suggests a crystal + (evolved) silicate liquid +

337 exsolved fluid assemblage, consistent with the presence of apatite, biotite, and salts. This agrees

338 with our observation of fluid inclusions in interstitial olivine (MZ) and plagioclase with apparent

339 silicate and iron-rich melt inclusions (MZ) as described by Jakobsen et al. (2011). Na

340 enrichment will occur during reaction (3) where plagioclase is converted to anorthite by a silica

341 under-saturated fluid. It is noted that, in some instances, plagioclase near sulfide assemblages

342 preserve An# ~ 62 near oxidized sulfides while An# of ~ 52 is preserved away from sulfides

343 (Fig. 4a, b, c). UZa plagioclase preserve larger variations, with one instance of near sulfide An#

344 ~ 63 while An# away from the sulfide is ~ 33 (Fig. 4d), suggesting this process becomes more

345 efficient, and total Na+ increases, up stratigraphy.

346 Finally, evidence for volatiles in other intrusions include saline fluid and halide melt

347 inclusions (Hanley et al. 2008) in the Archean Stillwater Complex, apatite halogen geochemistry

348 of samples from the Archean Stillwater and Bushveld Complexes (Boudreau et al. 1986), and

349 H2O ± NaCl ± CO2 ± CH4 inclusions in quartz reported for samples from the Bushveld Complex

350 (Ballhaus and Stumpfl 1985). Of note is the similarity in fluid inclusion compositions from the

351 Bushveld Complex described by Ballhaus and Stumpfl (1985) to those from the Skaergaard

352 Complex described by Larsen et al. (1992). The solutions described by Larsen et al. (1992) and

353 Sonnenthal (1992) are saline in nature, so the observation of halite and sylvite at silicate grain

354 boundaries reflect the former presence of a high temperature fluid rich in Cl. These fluids are of

355 interest in LMIs because of their ability to transport precious and transition metals through an

356 igneous body.

357

358

359 5.3 Precious Metals in an Exsolved Volatile Phase

360 Magmatic volatile phases are an important control on the distribution of precious metals

361 in epithermal systems (e.g., Heinrich et al. 2004) and Cu-Au porphyry systems (e.g., Candela

362 1989; Heinrich et al. 2004; Simon et al. 2008a). In addition, volatiles have also been argued to be

363 important for metal distribution in layered mafic intrusions (e.g., Boudreau and McCallum 1992;

364 Meurer et al. 1999, Hanley et al. 2008). Previous studies of the Skaergaard intrusion have

365 invoked magmatic volatile phases to explain dendritic anorthosites and pegmatites found in UZa

366 (Sonnenthal 1992), oxidized sulfides found throughout the stratigraphy (Andersen 2006), and

367 unusually high Cu/S ratios (Anderson 2006). Further, Godel et al. (2014) assert that in the

368 lowermost zone (Pd5) of the Platinova Reef, Au occurs at silicate ground boundaries and is never

369 enclosed by oxide minerals further implying that Au mineralization is not of primary magmatic

370 origin.

371 Several researchers (e.g., Simon et al. 2008a; Frank et al. 2011; Zajacz et al. 2013; Yin

372 and Zajacz 2018) have studied the behavior of silver in sulfur poor granitic and rhyolitic systems.

푓푙푢푖푑/푙푖푞푢푖푑 373 Zajacz et al. (2008) show that the Ag fluid-silicate liquid partition coefficient (퐷퐴푔 )

푓푙푢푖푑/푙푖푞푢푖푑 374 increases linearly with the molarity of Cl, mCl, in the exsolved fluid phase such that 퐷퐴푔

푓푙푢푖푑/푠푖푙푖푐푎푡푒 375  4 x mCl. Indeed, Cl has an affinity for the aqueous phase with 퐷퐶푙 values between

376 20 and 40 (Holland, 1972) or between .9 and 6 in molten basalts (Webster et al. 1999). A

377 pertinent question is whether an exsolved Cl-bearing fluid phase can transport significant

378 quantities of Ag by a neutral chloride complex:

379 AgO0.5 + HCl → AgCl + 0.5H2O (5) 380 melt vapor --- vapor --- 381 382

383 (Simon et al. 2008b) and by inference, other Cl soluble metals such as Au, Cu, and Zn. Yin and

384 Zajacz (2018) suggest that at higher temperatures (~ 900°C), the neutral NaAgCl2 is the favored

385 Ag species in Cl-bearing exsolved fluids via the modified H2-neutral reaction:

386 AgO + NaCl + HCl + 0.5SO2 → NaAgCl2 + 0.5H2SO4 (6) 387 melt ------vapor ------vapor ------388 389 390 Na cations can come from reaction (3) suggesting that the conversion of plagioclase to anorthite

391 by a silica-undersaturated fluid is intimately related to the transport of Ag as illustrated by

392 reaction (6).

393 Less well understood is the behavior of Ag in basaltic systems. Early studies investigated

394 the concentration of Ag and other precious metals in MORBs to understand their potential as

395 protoliths to greenstone Au deposits (Keays and Scott 1976). Recent experimental results have

푠푢푙/푠푖푙 396 allowed for the determination of 퐷퐴푔 of about 1000 (Li and Audetet 2012). Indeed, the

푠푢푙/푠푖푙 397 investigation of sulfide globules in MORB glasses lead to the determination of a similar 퐷퐴푔

398 of 1138 ± 245 (Patten et al. 2013). Workers have observed that MORB Cu/Ag values remain

399 constant at ~ 3500 with decreasing Mg#, suggesting the two elements behave similarly

400 throughout MORB fractionation (Jenner 2017). Additionally, similar Cu/Ag values are observed

401 in arc and back-arc environments prior to sulfide saturation (Jenner et al. 2010). Further support

402 for the associated nature of Cu and Ag is found in mantle peridotites where Cu/Ag values are

403 nearly identical to those observed in MORB (Wang and Becker 2015). Zajacz et al. (2013)

404 determined Ag solubilities for basaltic liquids to be ~ 11 and ~ 5.5 ppm for sulfide-bearing and

405 sulfide-free compositions, respectively. Because of the late S-saturation in the S-poor Skaergaard

406 system, it is appropriate to further investigate the partitioning of Ag into coevolving vapor in

407 basaltic liquids. Doing so will allow for a better understanding of Ag behavior in S-poor

408 intrusions.

409

410 5.4 Temperature and Timing

411 Temperature constraints might allow for better understanding of Ag (and other precious

412 and transition metals) mobilization during the Skaergaard cooling history. Physical relations such

413 as the interstitial nature of Ag in LZ samples and proximity to grain boundaries suggests that Ag

414 precipitation occurred late in the intrusion’s cooling history but early enough to be locally

415 entrained during overgrowth of primocrysts by crystallization of interstitial liquid. Further, ~ 40

416 % of the precious metal occurrences reported here occur at primocryst grain boundaries (see

417 Table 2) and in some instances, these occurrences are in close proximity to hydrous minerals

418 (Fig. 8a). Utilizing clinopyroxene solvus thermometry, Manning and Bird (1986) determined

419 minimum temperature of formation for late clinopyroxene rich veins to be between ~ 500 and

420 750°C which is within the range of temperatures reported for fluid inclusions by Sonnenthal

421 (1992) and Larsen et al. (1992). It is suggested that this is a viable minimum temperature range

422 for the precious metals reported here.

423 What remains unclear is why Ag is observed exclusively in native form in lithologies

424 below the MZ while Ag occurs as both alloys and native Ag and in sulfides in the MZ in general

425 and in the Platinova Reef in particular (Bird et al. 1991; Andersen et al. 1998; Cabri et al. 2005;

푠푢푙/푠푖푙 426 Holwell et al. 2015). High 퐷퐴푔 of ~ 1000 (Li and Audetet, 2012; Patten et al., 2013) suggest

427 that Ag should be retained in sulfide, if any is present. One possible explanation is that the LZ

428 magma was simply under-saturated in sulfide. Another scenario could be that LZ S was lost to an

429 exsolving volatile phase (e.g., Li and Boudreau 2017) leaving behind small amounts of Ag with

430 little or no S. However, in that case, one would expect to see alloys with Cu and in approximate

431 proportion to the metal concentration of the original sulfide. That this is not generally seen

432 suggests that metals have been effectively separated during volatile separation and transport.

433

434 5.5 Modelling Silver Solubility

435 To illustrate Ag enrichment during crystallization and degassing, a simple partitioning

436 model is calculated using Rayleigh fractionation conditions (Shaw 1970):

퐷̅−1 437 퐶푙푖푞 = 퐶표 ∗ 퐹 (1)

438 where 퐶푙푖푞 and 퐶표 represent the concentration of Ag in the silicate liquid and the initial

439 concentration, respectively. 퐹 corresponds to the weight fraction of silicate liquid remaining; 퐷̅ is

440 the bulk partition coefficient of Ag in the system given by:

̅ 푣푎푝/푙푖푞 푥푡푎푙/푙푖푞 441 퐷 = 퐷 ∗ 푋푣푎푝 + ∑(퐷푖 ∗ 푋푖) (2)

푣푎푝/푙푖푞 푥푡푎푙/푙푖푞 442 퐷 is the Ag vapor/liquid partition coefficient, 퐷푖 corresponds to the Ag

443 crystal/liquid partition coefficient for each mineral phase 푖. 푋푖 and 푋푣푎푝 are the weight fraction

444 of crystal 푖 and vapor separating from a silicate liquid, respectively. Because Ag does not readily

푥푡푎푙/푙푖푞 445 partition into common silicate minerals (Adam and Green 2006; Greaney et al. 2017), 퐷퐴푔

446 is equal to 0. Ultimately, the concentration of Ag in the vapor as crystallization proceeds is given

447 by:

퐷̅−1 푣푎푝/푙푖푞 448 퐶푣푎푝표푟 = 퐶표 ∗ 퐹 ∗ 퐷 (3)

449

푣푎푝/푙푖푞 푏푟푖푛푒/푙푖푞 450 Simon et al. (2008b) determine that 퐷퐴푔 in rhyolitic systems is 32 and 퐷퐴푔 is ~ 1000.

푣푎푝/푙푖푞 451 For the purposes of this calculation, a conservative 퐷퐴푔 of 10 is used. Because the initial Ag

452 concentration of the Skaergaard silicate liquid is unknown, end-member calculations are carried

453 out for a mid-ocean ridge basalt (MORB) (퐶퐴푔 = 0.0718 ppm, sample: T3-72D254-16-4, Keays

454 and Scott 1976) and a Hawaiian basalt from the Kilauea Iki lava lake (퐶퐴푔= 0.09 ppm, sample:

455 67-3-6.8, Greaney et al. 2017) under sulfide-free conditions. In this example, Kilauea Iki is

456 thought to be analogous to a sulfide-under saturated degassing cumulate pile.

457 As illustrated in Fig. 11, Ag increases in both the volatile and liquid phase as the weight

458 fraction of silicate liquid decreases. Also plotted is a dashed line corresponding to Ag solubility

459 in an S-free basaltic liquid (~ 5.5 ppm, Zajacz et al. 2013). For MORB liquid, extreme

460 fractionation (퐹 = 0.002) is required before MORB liquid reaches Ag saturation values of ~ 5.5

푣푎푝/푙푖푞 461 ppm. Because 퐷퐴푔 is 10, coevolving MORB vapor saturates in Ag much sooner during the

462 evolution of the liquid (퐹 = 0.02). The evolution of the Hawaiian basalt is identical in behavior

463 except that saturation of Ag in the liquid and vapor occurs earlier in the evolution of the liquid (퐹

464 = 0.015 and 0.15, respectively). In the absence of sulfide, Ag (and other metals with a similar

465 affinity for Cl) will preferentially concentrate in the coevolving vapor phase. This observation is

466 confirmed by a number of studies of gas condensates from volcanic fumeroles (Yudovskaya et

467 al. 2006; Yudovskaya et al. 2008; Chaplygin et al 2015).

468 In the following, the partitioning of Ag into a volatile phase in both sulfide-free (Case 1)

469 and sulfide-bearing (Case 2) systems is examined. We follow the methods of Boudreau and

470 McCallum (1992) where fractional vaporization from an interstitial liquid is assumed like

471 fractional melting and is described by equation (4);

472 1 퐶푟 (퐷−1) 473 = (1 – 퐹푣푎푝) (4) 퐶표 474

475 Where, 퐶푟 = the residual concentration of the element of interest, 퐶표 = the initial concentration of

476 the element of interest, 퐹푣푎푝 = the weight fraction of vapor exsolving from the silicate liquid and

477 퐷 = bulk partition coefficient given by equation (5);

푋푙푖푞 퐷푠푢푙푓 478 퐷 = ( ) + 푋푠푢푙푓 ∗ ( ) (5) 퐷푙푖푞 퐷푓

479 Where, 푋푙푖푞 = the weight fraction of silicate liquid remaining in the system, 푋푠푢푙푓 = the weight

480 fraction of sulfide in the system, 퐷푓 = the fluid/silicate liquid partition coefficient while 퐷푠푢푙푓 =

481 the sulfide/silicate liquid partition coefficient for the element of interest. Finally, the

482 concentration of the element of interest in the separated fluid at any interval is given by:

퐶 1 퐶 483 푓 = ∗ 푟 (6) 퐶표 퐷 퐶표

484 where, 퐶푓 = the concentration in the exsolved fluid.

485 Case 1. The partitioning of Ag into an exsolved fluid in a sulfide-free system (Fig. 12a,

486 b). Here, it is assumed that 75% silicate liquid has crystallized to form silicate minerals while

487 25% interstitial silicate liquid remains. The liquid is undersaturated in volatile at 25% interstitial

488 silicate liquid. The liquid is considered to saturate in Cl-bearing vapor at 5% such that for every

489 1% silicate liquid crystallized, 5% of that is Cl-bearing vapor. Initial Cl concentrations of 1, 5,

490 10, and 20 mol are assumed corresponding to modest values of 12.5, 62.5 125, and 250 ppm (ρ =

3 푓푙푢푖푑/푠푖푙푖푐푎푡푒 491 2.8 g/cm ) or 퐷퐴푔 values of 4, 20, 40, and 80, respectively (Zajacz et al., 2008). Cl is

492 perfectly incompatible in the solids prior to vapor saturation so that concentrations increase

493 exponentially as crystallization takes place. As shown in Fig. 12a and b, residual concentrations

494 are most affected by larger Ag fluid/silicate liquid partition coefficients where residual Ag

495 concentrations are half the original value after ~ 20% silicate liquid has crystallized and the bulk

496 weight fraction H2O has evolved to ~ 0.012 (Fig. 10a, b). Similarly, Ag in the exsolved fluid is

497 highest in the early stages of fluid evolution before values diminish due to dilution (Fig. 10b).

498 Case 2. Fig.12c and d show a similar calculation for a sulfide-bearing system composed

499 of 74.9% crystals, 25% silicate liquid, and .1% sulfide liquid. Cl molarities remain unchanged

500 from Case 1. It is assumed that the liquid is sulfide saturated at 400 ppm S. Each fraction of

501 water evolved contains 5% S which is subtracted from the starting S content before each

502 subsequent calculation. The Ag sulfide/silicate liquid partition coefficient is taken to be 1000

503 (Jenner 2017) while Ag fluid/silicate liquid partition coefficient values are as in the S-free

504 calculation. Calculations suggest that minimal change is observed in residual concentration prior

505 to sulfide being completely exhausted (vertical blue line, Fig. 12c, d). This is in large due to the

506 high Ag sulfide/silicate liquid partition coefficient values suggesting the partitioning of silver

507 into a fluid in a sulfide-bearing system is controlled by the amount of sulfide present. Indeed, for

푓푙푢푖푑/푠푖푙푖푐푎푡푒 508 each 퐷퐴푔 value, concentration in the exsolved fluid is maximized after sulfide is

509 entirely exhausted (Fig. 10d).

510 In either scenario, calculations suggest that a volatile phase efficiently removes Ag from

511 the residual silicate liquid. In the sulfide-free system, the earliest exsolved fluid (푋푤푎푡푒푟 < .002)

512 experiences the highest enrichment in Ag that decreases as Ag is removed from the system. The

513 residual silicate liquid experiences a gradual reduction in Ag content as vapor is lost. In the

514 sulfide-bearing system, Ag in the residual silicate liquid is most affected after sulfide is totally

515 resorbed during S loss to the vapor (at 푋푤푎푡푒푟 ~ .0094). This suggests that, if sulfide is present,

516 Ag will be preferentially partitioned into sulfide unless large fractions of volatiles evolve from

푓푙푢푖푑/푠푖푙푖푐푎푡푒 517 the interstitial silicate liquid or 퐷퐴푔 values are sufficiently high.

518

519 5.6 Evaluating Existing Genetic Models for the Skaergaard Intrusion Metal Budget

520 Although this contribution is based on a limited number of samples, reconciling

521 observations presented here with the late-stage evolution of the Complex and existing genetic

522 models for the Platinova Reef is worthwhile. Godel et al. (2014) and Nielsen et al. (2015)

523 suggest that discordant Au mineralization in the Platinova Reef is explained by the late-stage

524 evolution of hydrothermal fluids. Textural observations presented in this study confirm the

525 existence of metal + brine assemblages and support the ideas of Godel et al. (2014) and Nielsen

526 et al. (2015). The limited number of PGE outside of the Reef with the observed absence of Pd

527 might reflect differences in 퐷푠푢푙/푠푖푙 where 퐷푠푢푙/푠푖푙 Pd > Pt >> Au > Ag ≈ Cu (Mungall and

528 Brenan 2014). It is possible that existing sulfide would scavenge PGE preferentially based on

529 퐷푠푢푙/푠푖푙 values. Limited sulfide availability would prevent the partitioning of lower

530 퐷푠푢푙/푠푖푙 elements (e.g., Au, Ag, Cu) into sulfide allowing for concentration in the evolving

531 silicate liquid where exsolved hydrothermal fluids might ultimately transport these metals along

532 grain boundaries. One obvious complication to this hypothesis is the observation that a distinct

533 Cu-sulfide zone exists in the Platinova Reef and this occurrence is commonly attributed to

534 orthomagmatic processes (e.g., Holwell and Keays 2014; Keays and Tegner 2016). Further,

535 native Cu reported in this study is spatially related to sulfide minerals and is likely the result of

536 oxidation processes.

537 The limited number of PGE observed outside of the Reef might be the result the limited

538 number of samples examined in this study. While little can be said about the petrogenesis of the

539 Platinova Reef, it is important to emphasize that our results indicate that late magmatic

540 hydrothermal fluids are, in some instances, directly linked to precious metals, particularly Ag

541 and Au. For both Ag and Au, evidence suggestive of a coexisting metal + Cl-bearing phase are

542 observed. If fluids contemporaneously oxidized magmatic sulfides, the S lost from below the

543 Platinova Reef and its transport upward could give rise to sulfide saturation in the upper parts of

544 the crystal pile. In this case, the exsolved fluid would consist of metals complexed as Cl and S

545 species. Variable redox sensitivity (Cl+ vs HS-) would act to separate metals transported as Cl

546 complexes (Ag) from those transported as S complexes (Au) (Li and Boudreau, in prep). In this

547 scenario, the net result would not be unlike the Pd, Au, Cu offsets described by others (Bird et al.

548 1991; Andersen et al. 1998; Nielsen et al. 2005; Holwell and Keays, 2014).

549

550 5.7 Genetic Model

551 A model is proposed where Ag (with other Cl soluble elements having low

552 퐷푚푛푟푙/푚푒푙푡 partition coefficients, Au in particular) becomes increasingly concentrated in the

553 residual liquid during crystallization. At low liquid/silicate mineral mass fraction values (퐹), Ag

554 is concentrated to quantities where saturation becomes favorable. At this point, or shortly

555 thereafter, enough silicate liquid volume has crystallized to allow for the separation of H2O, CO2,

556 Cl and other volatile components with low 퐷푚푛푟푙/푚푒푙푡partition coefficients. Owing to its

557 relatively high 퐷푓푙푢푖푑/푠푖푙푖푐푎푡푒 partition coefficient, Ag partitions into the exsolved fluid as Ag-

558 chloride or perhaps Na-Ag-chloride. The exsolved fluid transporting Ag will travel vertically

559 along grain boundary networks, and fluid pathways, interacting with other interstitial phases.

560 Heating of the fluid as it moves into hotter rocks can lead to S and SiO2 loss from the host

561 assemblage while leaving behind a small portion of the Ag-bearing volatile phase. Eventual

562 cooling of the fluid can cause Ag precipitation or, when a fluid under-saturated interstitial liquid

563 is encountered, re-dissolution of the fluid initiates the precipitation of Ag and perhaps an

564 increased volatile content for that section of the intrusion. Minor amounts of metal + volatile are

565 trapped in the interstitial liquid and are presumed to have generated textures seen in Fig. 5a, and

566 Fig. 8. The cumulative effect of this process most affected the MZ where the silicate liquid

567 became saturated in fluid + vapor early enough to preserve fluid inclusions in olivine,

568 clinopyroxene, and plagioclase.

569

570 6. Conclusions

571 Evidence for S loss is present throughout the Skaergaard stratigraphy, with the relative

572 proportion of low-Ti magnetite replacing sulfide being most extensive in the LZ. Native Ag is

573 regularly associated with trace quantities of Cl in and below the MZ, consistent with a coexisting

574 metal + brine assemblage. Ag, and other Cl soluble metals can be readily mobilized by an

575 exsolved volatile Cl-bearing phase in both sulfide-free and sulfide-bearing liquids. This

576 contribution has implications for the origins of the Platinova Reef. In particular, our work

577 supports the assertions of Godel et al. (2014) and Nielsen et al. (2015) who attribute Au

578 mineralization to late hydrothermal fluids. Ultimately, it is suggested that a late fluid rich in

579 halides transported silver and perhaps other precious metals vertically through the Skaergaard

580 cumulate pile. Numerical modelling suggests that the Kilaeua Iki lava lake may be analogous to

581 a S-free degassing cumulate pile. Indeed, it is possible that Ag and Au are readily mobile

582 provided the magma is S-poor or experienced an early sulfide saturation event. Finally, this

583 contribution presents new evidence supportive of late hydrothermal processes in LMIs and

584 suggests that similar work may clarify late stage evolution of other intrusions, particularly those

585 intrusions where evidence of hydrothermal activity has been reported (e.g., Stillwater Complex,

586 Bushveld Complex).

587

588 Acknowledgements

589 Discussions with J. Cann, E. Klein, and T.L. Nielsen were helpful and improved the

590 quality of this manuscript. Alex Hammerstrom is acknowledged for his work compiling and

591 digitizing sample location information. Thorough and constructive reviews by David Holwell

592 and Katie McFall greatly improved the quality and presentation of this manuscript. Finally,

593 editorial handling by Wolfgang Maier is acknowledged and much appreciated.

594

595

596

597

598

599

600

601 References

602 Adam, J and Green, T (2006) Trace element partitioning between mica- and amphibole-bearing 603 garnet lherzolite and hydrous basanitic melts: 1. Experimental results and the 604 investigation of controls on partitioning behavior. Contrib Mineral Petr 152; 1-17

605 Aird, HM and Boudreau, AE (2013) High-temperature carbonate minerals in the Stillwater 606 Complex, Montana, USA. Contrib Mineral Petr 116; 1143-1160

607 Andersen, JCØ, Rasmussen, H, Nielsen, TFD, and Ronsbo, JG (1998) The Triple Group and the 608 Platinova gold and palladium reefs in the Skaergaard Intrusion; stratigraphic and 609 petrographic relations. Econ Geol 93 (4); 488-509

610 Andersen, JCØ, Rollinson, GK, McDonald, I, Tegner, C, and Lesher, CE (2017) Platinum-group 611 mineralization at the margin of the Skaergaard intrusion, East Greenland. Miner Depos 612 52 (6); 929-942

613 Anderson, JCØ (2006) Postmagmatic Sulphur loss in the Skaergaard Intrusion: Implications for 614 the formation of the Platinova Reef. Lithos 92 (1); 198-221

615 Ballhaus, CG, and Stumpfl, EF (1985) Occurrence and petrological significance of graphite in 616 the Upper Critical Zone, western Bushveld Complex, South Africa. Earth Planet Sci Lett 617 74 (1); 58-68

618 Bird, DK, Brooks, CK, Gannicott, RA, Turner, PA (1991) A Gold-Bearing Horizon In the 619 Skaergaard Intrusion, East Greenland. Econ Geol 86; 1083-1092

620 Bird, DK, Rogers, RD, and Manning, CE (1986) Mineralized fracture systems of the Skaergaard 621 intrusion, East Greenland. Meddelelser om Grønland, Geoscience 16, pp 68

622 Boudreau, AE, and McCallum IS (1989) Investigations of the Stillwater Complex: Part V. 623 Apatites as indicators of evolving fluid composition. Contrib Mineral Petr 102 (2); 138- 624 153

625 Boudreau, AE, and McCallum IS (1992) Concentration of platinum-group elements by magmatic 626 fluids in layered intrusions. Econ Geol 87 (7); 1830-1848

627 Boudreau, AE, Mathez, EA, and McCallum, IS (1986) Halogen Geochemistry of the Stillwater 628 and Bushveld Complexes: Evidence for Transport of the Platinum-Group Elements by 629 Cl-Rich Fluids. J Petrol 27 (4); 967-986

630 Boudreau, AE, Stewart, MA, and Spivack, AJ (1997) Stable Cl isotopes and origin of high-Cl 631 magmas of the Stillwater Complex, Montana. Geology 25 (9); 791-794

632 Brooks, CK, and Gleadow, AJW (1977) A fission-track age for the Skaergaard intrusion and the 633 age of the East Greenland basalts. Geology 5 (9); 539-540

634 Cabri, LJ, Beattie, M, Rudashevsky, NS, and Rudashevsky, VN (2005) Process mineralogy of 635 Au, Pd and Pt ores from the Skaergaard intrusion, Greenland, using new technology. 636 Miner. Eng. 18 (8); 887-897

637 Campbell, IH, and Naldrett, AJ (1979) The Influence of Silicate: Sulfide Ratios on the 638 Geochemistry of Magmatic Sulfides. Econ Geol 74 (6); 1503-1506

639 Candela, PA, and Holland, HD (1984) The partitioning of copper and molybdenum between 640 silicate melts and aqueous fluids. Geochim Cosmochim Acta 48; 373-380

641 Chaplygin, I, Yudovskaya, M, Vergasova, L, and Mokhov, A (2015) Native gold from volcanic 642 gases at Tolbachik 1975-76 and 2012-13 Fissure Eruptions, Kamchatka. J Volcanol 643 Geotherm Res 307; 200-209

644 Fleet, ME, Chryssoulis, SL, Stone, WE, and Weisener, CG (1993) Partitioning of platinum- 645 group elements and Au in the Fe-Ni-Cu-S system: experiments on the fractional 646 crystallization of sulfide melt. Contrib Mineral Petr 115; 36-44

647 Frank, MR, Simon, AC, Pettke, T, Candela, PA, and Piccoli, PM (2011) Gold and copper 648 partitioning in magmatic-hydrothermal systems at 800°C and 100 MPa. Geochim 649 Cosmochim Acta 75 (9); 2470-2482

650 Godel, B, Rudashevsky, NS, Nielsen, TFD, Barnes, SJ, and Rudashevsky, VN (2014) New 651 Constraints on the origin of the Skaergaard intrusion Cu-Pd-Au mineralization: Insights 652 from high-resolution X-ray computed tomography. Lithos 190-191; 27-36

653 Greaney, A, Rudnick, RL, Helz, RT, Gaschnig, RM, Piccoli, PM, and Ash, RD (2017) The 654 behavior of chalcophile elements during magmatic differentiation as observed in Kilauea 655 Iki lava lake, Hawaii. Geochim Cosmochim Acta 210; 71-96

656 Hanley, JJ, Mungall, JE, Pettke, T, Spooner, ETC, and Bray, CJ (2008) Fluid and Halide Melt 657 Inclusions of Magmatic Origin in the Ultramafic and Lower Banded Series, Stillwater 658 Complex, Montana, USA. J Petrol 49 (6); 1133-1160

659 Heinrich, CA, Driesner, T, Stefansson, A, and Seward, TM (2004) Magmatic vapor contraction 660 and the transport of gold from the porphyry environment to epithermal ore deposits. 661 Geology 32 (9); 761-764

662 Hirschmann, MM, Renne, PR, and McBirney, AR (1997) 40Ar/39Ar dating of the Skaergaard 663 intrusion. Earth Planet Sci Lett 146; 645–658 664 Holland, HD (1972) Granites, Solutions, and Base Metal Deposits. Econ Geol 67; 281-301

665 Holness, MB, Nielsen, TFD, and Tegner, C (2007) Textural maturity of cumulates: A record of 666 chamber filling, liquidus assemblage, cooling rate and large scale convection in mafic 667 layered intrusions. J Petrol (48); 141-157

668 Holwell, DA and Keays, RR (2014) The Formation of Low-Volume, High-Tenor Magmatic 669 PGE-Au Sulfide Mineralization in Closed Systems: Evidence from Precious and Base 670 Metal Geochemistry of the Platinova Reef, Skaergaard Intrusion, East Greenland. Econ 671 Geol 109 (2); 387-406

672 Holwell, DA, Keays, RR, McDonald, I, and Williams, MR (2015) Extreme enrichment of Se, Te, 673 PGE, and Au in Cu sulfide microdroplets: evidence from LA-ICP-MS analysis of sulfides 674 in the Skaergaard Intrusion, east Greenland. Contrib Mineral Petr 170:53

675 Huber, C, Bachmann, O, Vigneresse, J-L, Dufek, J, and Parmigiani, A (2012) A physical model 676 for metal extraction and transport in shallow magmatic systems. Geochem Geophys 13 677 (8); 2887-2905

678 Jakobsen, JK, Veksler, IV, Tegner, C, and Brooks, CK (2011) Crystallization of the Skaergaard 679 Intrusion from an Emulsion of Immiscible Iron- and Silica-rich Liquids: Evidence from 680 Melt Inclusions in Plagioclase. J Petrol 52 (2); 345-373

681 Jenner, FE (2017) Cumulate causes for the low contents of sulfide-loving elements in the 682 continental crust. Nat Geosci 10; 524-530

683 Jenner, FE, O’Neill, HSC., Arculus, RJ, and Mavrogenes, JA (2010) The magnetite crisis in the 684 evolution of arc-related magmas and the initial concentration of Au, Ag, and Cu. J Petrol 685 51; 2445-2464

686 Keays, RR, and Scott, RB (1976) Precious metal in ocean-ridge basalts; implications for basalts 687 as source rocks for gold mineralization. Econ Geol 71 (4); 705-720

688 Keays, RR, and Tegner, C (2016) Magma Chamber Processes in the Formation of the Low- 689 sulfide Magmatic Au-PGE Mineralization of the Platinova Reef in the Skaergaard 690 Intrustion, East Greenland. J Petrol 56 (12); 2319-2340

691 Larsen, RB, Brooks, CK, and Bird, DK (1992) Methane-bearing, aqueous, saline solutions in the 692 Skaergaard intrusion, east Greenland. Contrib Mineral Petr; 428-437

693 Li, C, and Boudreau, AE (2017) The origin of high-Cu/S sulfides by shallow level degassing in 694 the Skaergaard intrusion, East Greenland. Geology 45 (12); 1075-1078.

695 Li, P, and Boudreau, AE (2019) Vapor Transport of Silver and Gold in Basaltic Lava Flows, in 696 prep

697 Li, Y, and Audetat, A (2012) Partitioning of V, Mn, Co, Ni, Cu, Zn, As, Mo, Ag, Sn, Sb, W, Au, 698 Pb, and Bi between sulfide phases and hydrous basanite melt at upper mantle conditions. 699 Earth Planet Sci Lett 355; 327-340

700 Manning, CE, and Bird, DK (1986) Hydrothermal clinopyroxenes of the Skaergaard intrusion. 701 Contrib Mineral Petr 92; 437-447

702 McBirney, AR (1975) Differentiation of the Skaergaard Intrusion. Nature 253; 691-694

703 McBirney, AR, (1989) The Skaergaard Layered Series: I. Structure and Average Compositions. J 704 Petrol 30 (2); 363-397

705 Meurer, WP, Klaber, S, and Boudreau, AE (1997) Discordant bodies from olivine-bearing zones 706 III and IV of the Stillwater Complex, Montana – evidence for postcumulus fluid 707 migration and reaction in layered intrusions. Cotrib Min Pet 130 (1); 81-92

708 Meurer, WP, Willmore, CC, and Boudreau, AE (1999) Metal redistribution during fluid 709 exsolution and migration in the Middle Banded series of the Stillwater Complex, 710 Montana. Lithos 47; 143-156

711 Mungall, JE, and Brenan, JM (2014) Partitioning of platinum-group elements and Au between 712 sulfide liquid and basalt and the origins of mantle-crust fractionation of the chalcophile 713 elements. Geochim Cosmochim Acta 125; 265-289

714 Naslund, HR (1984) Petrology of the Upper Border Series of the Skaergaard Intrusion. Journ 715 Petrol 25 (1); 185-212

716 Nielsen, TFD (1975) Possible mechanism of continental breakup in the North Atlantic. Nature 717 253; 182-184

718 Nielsen, TFD (2004) The Shape and Volume of the Skaergaard Intrusion, Greenland: 719 Implications for Mass Balance and Bulk Composition. J Petrol 45 (3); 507-530

720 Nielsen, TFD, Andersen, JCO, Holness, MB, Keiding, JK, Rudashevsky, NS, Rudashevsky, VN, 721 Salmonsen, LP, Tegner, C, and Veksler, IV (2015) The Skaergaard PGE and Gold 722 Deposit: the Result of in situ Fractionation, Sulphide Saturation, and Magma Chamber- 723 scale Precious Metal Redistribution by Immiscible Fe-rich Melt. Journ Petrol 56 (8); 724 1643-1676

725 Patten, C, Barnes, S-J, Mathez, EA, and Jenner, FE (2013) Partition Coefficients of chalcophile 726 elements between sulfide and silicate melts and the early crystallization history of sulfide 727 liquid: LA-ICP-MS analysis of MORB sulfide droplets. Chem Geol 358; 170-188

728 Pendergast, MD (2000) Layering and Precious Metals Mineralization in the Rincon del Tigre 729 Complex, Eastern Bolivia. Econ Geol 95 (1); 113-130

730 Schneider, CA, Rasnad, WS, and Eliceiri, KW (2012) NIH Image to ImageJ: 25 years of image 731 analysis. Nat Methods 9 (7); 671

732 Shaw, DM (1970) Trace element fractionation during anataxis. Geochim Cosmochim Acta 34; 733 237-242

734 Shinohara, H (1994) Exsolution of immiscible vapor and liquid from a crystallizing silicate melt: 735 Imlications for chlorine and metal transport. Geochim CosmochimActa 58 (23); 5215- 736 5221

737 Simon, AC, Candela, PA, Piccoli, PM, Mengason, M, and Englander, L (2008a) The effect of 738 crystal-melt partitioning on the budgets of Cu, Au, and Ag. Amer Miner 93 (8-9); 1437- 739 1448

740 Simon, AC, Pettke, T, Candela, PA, and Piccoli, PM (2008b) The partitioning behavior of silver 741 in a vapor-brine-rhyolite melt assemblage. Geochim Cosmochim Acta 72 (6); 1638-1659

742 Sonnenthal, EL (1992) Geochemistry of dendritic anorthosites and associated pegmatites in the 743 Skaergaard Intrusion, East Greenland: Evidence for metasomatism by a chlorine-rich 744 fluid. J Volcanol Geotherm Res (1-3); 209-230

745 Wager, LR, and Deer, WA (1939) Geological investigations in East Greenland, Part III. The 746 petrology of the Skaergaard Intrusion, Kangerdlugssuaq, East Greenland, Meddelelser 747 om Grønland, 105, pp 1-352 748 Wager, LR, Vincent, EA, and Smales, AA (1957) Sulphides in the Skaergaard intrusion, east 749 Greenland. Econ Geol 52; 855–903 750 Wang, Z, and Becker, H (2015) Abundances of Ag and Cu in mantle peridotites and the 751 implications for the behavior of chalcophile elements in the mantle. Geochim 752 Cosmochim Acta 160 (1); 209-226

753 Webster, JD, Kinzler, RJ, and Mathez, EA (1999) Chloride and water solubility in basalt and 754 andesite melts and implications for magmatic degassing. Geochim Cosmochim Acta 65; 755 729-738

756 Wohlgemuth-Ueberwasser, CC, Fonseca, ROC, Ballhaus, C, and Berndt, J (2013) Sulfide 757 oxidation as process for the formation of copper-rich magmatic sulfides. Miner Deposita 758 48 (1); 115-127

759 Yin, Y, and Zajacz, Z (2018) The solubility of silver in magmatic fluids: Implications for silver 760 transfer to the magmatic-hydrothermal ore-forming environment. Geochem Cosmochem 761 Acta 238; 235-251

762 Yudovskaya, MA, Distler, VV, Chaplygin, IV, Mokhov, AV, Trubkin, NV, Gorbacheva, SA 763 (2006) Gaseous transport and deposition of gold in magmatic fluid: evidence from the 764 active Kudryavy volcano, Kurile Islands. Miner Deposita 40; 828-848

765 Yudovskaya, MA, Tessalina, S, Distler, VV, Chaplygin, IV, Chugaev, AV, and Dikov, YP 766 (2008) Behavior of highly-siderophile elements during magma degassing: A case study at 767 the Kudryavy volcano. Chem Geol 248; 318-341

768 Zajacz, Z, Candela, PA, Piccoli, PM, Sanchez-Valle, C, and Walle, M (2013) Solubility and 769 partitioning behavior of Au, Cu, Ag and reduced S in magmas. Geochim Cosmochim 770 Acta 112; 288-304

771 Zajacz, Z, Halter, WE, Pettke, T, and Guillong, M (2008) Determination of fluid/melt partition 772 coefficients by LA-ICPMS analysis of co-existing fluid and silicate melt inclusions: 773 Controls on element partitioning. Geochem Cosmochem Acta 72 (8); 2169-2197

774