Agangi Et Al 2015

Earth-Science Reviews 140 (2015) 2753
Contents lists available at ScienceDirect
Earth-Science Reviews
journal homepage: www.elsevier.com/locate/earscirev
Gold accumulation in the Archaean Witwatersrand Basin,

South Africa Evidence from concentrically laminated pyrite
A. Agangi a,, A. Hofmann a, C. Rollion-Bard b,1, J. Marin-Carbonne c,2, B. Cavalazzi d, R. Large e, S. Meffre e
a
Department of Geology, University of Johannesburg, Auckland Park 2006, Johannesburg, South Africa
CRPG-CNRS, Universit de Lorraine, 15 rue Notre Dame des Pauvres, 54500 Vandoeuvre-ls-Nancy, France
Department of Earth and Space Sciences, UCLA, 595 Charles Young Drive East, Los Angeles, CA 90095, USA
d
Dipartimento di Scienze Biologiche, Geologiche e Ambientali, Universit di Bologna, Via Zamboni 67, 40126 Bologna, Italy
e
CODES ARC Centre of Excellence in Ore Deposits, University of Tasmania, Hobart, Tasmania 7005, Australia
b
c
a r t i c l e
i n f o
Article history:
Received 17 January 2014
Accepted 30 October 2014
Available online 7 November 2014
Keywords:
Archaean
Gold
Pyrite
S isotopes
Fe isotopes
Witwatersrand
a b s t r a c t
Concentrically laminated pyrite is a relatively common, although volumetrically minor, component of auriferous
conglomerates in the Archaean (ca. 3.02.7 Ga) Witwatersrand Basin of South Africa. This type of pyrite contains
high amounts (several tens of ppm) of Au, but the origin of the pyrite is debated, and the timing of Au deposition
in these grains is not known. In order to constrain the formation of pyrite, we have studied concentrically laminated pyrite and other coexisting types of pyrite (inclusion-rich, massive pyrite) by analysing the contents and
distribution of Au and other trace elements by laser ablation ICP-MS, the S and Fe isotope composition by
SIMS, and the mineral inclusions by scanning electron microscope and laser Raman spectroscopy. Trace element
maps indicate that concentrically laminated pyrite is enriched in Sb, Mn, Au, Ag, Tl, Cu, Mo, Mn, and contains two
types of gold: nely dispersed Au (invisible gold, with Au/Ag ~0.1 and likely of primary origin) and Au inclusions with Au/Ag ~10 of secondary origin. The study of mineral inclusions revealed the presence of muscovite,
chlorite, ne-grained carbonaceous matter, monazite, Ti-oxides, and quartz. Iron and multiple S isotopes suggest
that concentrically laminated pyrite and inclusion-rich pyrite were formed from two separate pools of S and Fe
with different isotope characteristics. Sulfur was derived from atmospheric S that had undergone
with negative 33S that constituted concentrically lamimass-independent isotope fractionation to form SO2
4
nated pyrite, and elemental S with positive 33S that formed inclusion-rich pyrite. Iron pools were derived
from partial oxidation of Fe2+, so that concentrically laminated pyrite formed from a low-56Fe residual Fe2+
(average +0.2) and inclusion-rich pyrite formed from a high-56Fe Fe3+ pool (average +2.7). Biological ac34
tivity may have been involved in the reduction of SO2
4 , causing a wide spread of S values (~25, S reducing
microorganisms), as well as in the partial oxidation of Fe2+ (anaerobic photosynthetic Fe reducers or photosynthetic O2 producers), and in the formation of pyrite from Fe3+ (dissimilatory Fe reducers). We propose that concurrent biogenically-mediated pyrite formation and Au trapping suggest that microbial activity was responsible
for the accumulation of Au and other trace elements (e.g. Sb, Mn, Ag, Tl, Cu, Mo, Mn) which are commonly
enriched in organic matter-rich sediments.
2014 Elsevier B.V. All rights reserved.
Contents
1.
2.
3.
4.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Geological setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Previous investigations of CLP in the Witwatersrand Basin and other sedimentary successions
3.1.
CLP in the Witwatersrand Basin . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.
CLP in other successions . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Methods and analytical techniques . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.
Sample preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.
X-ray elemental maps and electron probe micro analysis (EPMA) . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Corresponding author at: Applied Geology Department, Curtin University, Kent St, Bentley 6102, WA, Australia Tel. + 61(0)892667074.
E-mail address: andrea.agangi@curtin.edu.au (A. Agangi).
1
Present address: Institut de Physique du Globe - Sorbonne Paris Cit, Universit Paris Diderot, CNRS, 1 rue Jussieu, 75238 Paris cedex 05, France.
2
present address: Laboratoire Magmas et Volcans Transferts Lithosphriques UMR CNRS 6524 Universit Jean Monnet 42023 Saint-Etienne, Cedex 2.
http://dx.doi.org/10.1016/j.earscirev.2014.10.009
0012-8252/ 2014 Elsevier B.V. All rights reserved.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
28
29
30
30
30
30
30
30
28
A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753
4.3.
4.4.
Laser Raman spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . .

Secondary ion mass spectrometry (SIMS) S and Fe isotope analysis . . . . . . . .
4.4.1.
Sulfur isotope measurements . . . . . . . . . . . . . . . . . . . . .
4.4.2.
Iron isotope measurements . . . . . . . . . . . . . . . . . . . . . .
4.5.
Laser ablation ICP-MS trace element maps . . . . . . . . . . . . . . . . . . .
5.
Sample description and pyrite textures . . . . . . . . . . . . . . . . . . . . . . . .
5.1.
Round massive pyrite . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.
Euhedral massive pyrite . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.
Round inclusion-rich pyrite . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.
CLP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.
Mineral inclusions (laser Raman and SEM-EDS study) . . . . . . . . . . . . . . . . .
7.
Elemental composition and intragranular textures of CLP (EPMA and LA-ICP-MS) . . . . .
7.1.
Electron microprobe X-ray maps . . . . . . . . . . . . . . . . . . . . . . .
7.2.
LA-ICP-MS maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.
Sulfur and Fe isotope composition of pyrite . . . . . . . . . . . . . . . . . . . . . .
8.1.
S isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.
Fe isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.1.
Textural constrains on the formation of CLP . . . . . . . . . . . . . . . . . . .
9.2.
Distinguishing primary and secondary gold in CLP . . . . . . . . . . . . . . . .
9.3.
Distinguishing primary vs. post-sedimentary isotope signatures . . . . . . . . .
9.4.
Two distinct S pools carrying MIF represented in VCR pyrite . . . . . . . . . . .
9.5.
Two pools of Fe: partial oxidation of Fe2 + via APIO or photosynthetic O2 production
9.6.
A model of CLP formation and Au accumulation for the VCR . . . . . . . . . . .
9.7.
Role of biological processes in Au mineralisation in the Witwatersrand Basin . . . .
10. Summary/conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Introduction
The origin of AuU deposits in the Archaean Witwatersrand Basin of
South Africa is the subject of longstanding discussion. Because of the
close spatial association of pyrite with Au and U in the mineralised
conglomerates, the origin of pyrite is a central aspect in all genetic
hypotheses of Witwatersrand mineralisation, in both primary placer
models (Robb and Meyer, 1995a,b; Frimmel, 2005), and postsedimentary, hydrothermal models (Barnicoat et al., 1997; Phillips
and Law, 2000). The two models have broad implications, given the
constraint that the stability of detrital suldes in a subaerial sedimentary environment would give on the composition of the Archaean atmosphere (Krupp et al., 1994; England et al., 2002a; Frimmel, 2005).
Further, high concentrations of Au and U are in some cases associated
with thin but laterally continuous layers of carbonaceous matter.
Although the carbonaceous matter is ultimately of biogenic origin, as indicated by carbon isotope ratios, the process of its formation is unclear,
either representing bitumen derived from migrated hydrocarbons
(Spangenberg and Frimmel, 2001; England et al., 2002b) or microbial
mat remains (Mossman et al., 2008). Other independent evidence for
the former presence of life exists in the Witwatersrand basin, including
microbially-induced sedimentary structures (Noffke et al., 2006), and
possible suldised microfossils (Schidlowski, 1965).
In this paper, new evidence for the association of biological activity
and the accumulation of gold in the Witwatersrand conglomerates is
provided by the volumetrically minor, but relatively widespread concentrically laminated pyrite (CLP). This type of pyrite, reported in the
Elsburg Reef, Basal Reef and, most commonly, in the Ventersdorp Contact Reef (Schweigart and von Rahden, 1965; Hallbauer, 1986;
England et al., 2002a) has been previously described as oolitic pyrite,
a term with a certain genetic implication. It locally occurs in association
with other types of round pyrite (buckshot pyrite), such as mudball
pyrite (e.g. Hallbauer, 1986). Mudball pyrite, characterised by roughly
globular shape and polygonal cracks on the surface, is common in Witwatersrand reefs and locally forms 210 mm thick-layers (Hallbauer,
1986; Barton and Hallbauer, 1996). Both types of pyrite are known to
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
31
31
31
31
31
31
32
32
32
34
34
34
34
34
37
37
39
40
40
41
42
43
45
46
48
49
51
51
contain appreciable amounts of gold (up to several tens of ppm of Au;

Hallbauer, 1986; Guy et al., 2010; Koglin et al., 2010; Agangi et al.,
2013). Preservation of delicate textures of CLP suggests short-distance
transport (Saager, 1970; Large et al., 2013). Thus, the study of this pyrite
type can potentially give information on the conditions at, or near, the
site of nal deposition of the conglomerates.
The analysis of multiple S isotope ratios in pyrite has proven to be a
powerful tool in the study of ancient rock successions, especially since it
was discovered that sedimentary rocks older than c. 2.3 Ga show massindependent S isotope fractionation, which is interpreted to be the result of photochemical reactions distinctively occurring in the early atmosphere (Thiemens, 1999; Farquhar et al., 2000; Farquhar and Wing,
2003; Whitehouse et al., 2005; Johnston, 2011). Further, recent advances in analytical techniques have demonstrated that Fe isotopes
can be applied to Archaean and Phanerozoic iron suphides to discriminate biological and abiotic processes for their origin (Beard et al.,
2003a; Yamaguchi et al., 2005; Whitehouse and Fedo, 2007). Finally,
the combined study of Fe and multiple S isotopes is emerging as a promising approach in addressing these issues (Archer and Vance, 2006;
Hofmann et al., 2009; Marin-Carbonne et al., 2014).
We have studied the chemical and Fe, S isotope compositions of CLP
grains and other types of pyrite from the Ventersdorp Contact Reef
(VCR), Witwatersrand Basin in order to gain information on their origin.
Abundant elemental and some isotopic compositional data are available
in the literature, which have shown that signicant differences in trace
element and isotopic composition exist between different conglomerates, and different generations of pyrite (e.g. Utter, 1978; Meyer et al.,
1990; Barton and Hallbauer, 1996; Zhao et al., 2006; Guy et al., 2012;
Large et al., 2013). However, most of the available chemical and isotopic
analyses were made on whole grains. Because Witwatersrand pyrite is
commonly zoned (Foya et al., 1999; Reimold et al., 2004; Agangi et al.,
2013), these whole-grain analyses must be considered as average compositions of different zones. Inclusions of sulde and silicate minerals
can also severely affect the results. Bearing this in mind, we have chosen
an in situ analytical approach, including ion and electron microprobes,
and laser ablation ICP-MS, to be able to analyse small-scale textures,
and to discriminate between different elemental and isotopic contributions. Our results have implications for the cycles of S and Fe in the Archaean, suggesting that the origin of sedimentary pyrite in the
Witwatersrand Basin was biologically mediated, and pointing to a link
with the concentration of gold.
2. Geological setting
The Meso- to Neoarchaean Witwatersrand Basin extends
for ~ 350 150 km in the central Kaapvaal Craton (Fig. 1), and
represents the remnant of a larger basin (Beukes and Cairncross, 1991).
The development of the Witwatersrand Basin accompanied and recorded
the complex tectonic evolution of the Kaapvaal Craton (Phillips and
Law, 2000). After the rst phase of lithosphere creation and accretion
(3.643.08 Ga), the Kaapvaal Craton experienced the deposition of the
Dominion Group, a ~ 2 km-thick volcano-sedimentary succession
with a minimum age of 3074 6 Ma (Armstrong et al., 1991).
Following a ~ 100-million-year hiatus, the Witwatersrand Supergroup was deposited.
The Witwatersrand Supergroup is composed of the West Rand
Group and the overlying Central Rand Group. The West Rand Group is
a 5 km-thick sedimentary succession resting on the Dominion
Group or granitoid-greenstone basement. The West Rand Group
(2985 14 Ma2914 8 Ma; Armstrong et al., 1990; Kositcin and
Krapez, 2004) is dominated by shallow marine to off-shore shelf sediments with minor uvial deposits, mostly sandstone, siltstone-shale,
and mudstone, deposited during a phase of thermal subsidence
(Eriksson et al., 1981; Coward et al., 1995; Frimmel et al., 2005). The
Central Rand Group (2902 13 Ma2872 6 Ma; Kositcin and
Krapez, 2004) lies unconformably on the West Rand Group, and includes a 2.8 km-thick, mostly subaerial (uvial) to shallow marine sequence of sandstone and quartz-pyrite pebble conglomerates with
minor shales deposited in a foreland or retroarc setting (Coward et al.,
1995; Els, 1998; Catuneanu, 2001; Kositcin and Krapez, 2004). The
Central Rand Group contains most of Au and U mineralisation associated with the conglomerates (or reefs) (Frimmel, 2005).
The 2.782.71 Ga Ventersdorp Supergroup unconformably overlies
the Witwatersrand Supergroup. The unconformity is marked by the
Venterspost Conglomerate Formation, also known as the Ventersdorp
Contact Reef (VCR), a uvial conglomerate that can locally reach high
Au grades (Krapez, 1985; Robb and Meyer, 1995a,b). The (VCR) is one
map location
29
of the major gold-bearing conglomerate horizons in the Witwatersrand

Basin, with a gold grade of 512 g/t (Frimmel, 2005). The VCR in the
Carlonville area is a thin, degradational sedimentary package generally
less than 2.5 m thick that unconformably overlies Central Rand Group
strata mantling an irregular palaeotopography and is overlain by subaerial lava ows of the Klipriviersberg Group. The latter includes mac
lava and silicic units that erupted during a short-lived phase of volcanic
activity at 2709 4 Ma (Armstrong et al., 1991). The VRC consists of
different unconformity-bounded stratigraphic packages of channelled
or sheet-like deposits of conglomerate and sandstone regarded to
have formed in a uvial braid-plain environment (Krapez, 1985;
Engelbrecht et al., 1986; McCarthy, 1994). Inter-reef lava ows
(Germs and Schweitzer, 1994) indicate syn-depositional volcanic activity. In this study, we used the term Witwatersrand Basin in the sense of a
long-lived intracratonic depository, in which Au- and U-bearing sediments of the Dominion Group, the Witwatersrand Supergroup and the
Ventersdorp Supergroup were deposited.
After sedimentation, the Witwatersrand Basin rocks have been
overprinted in various events that started during sedimentation with
meteoric water inltration, and continued during diagenesis and metamorphism and culminated in the Vredefort impact event (Phillips et al.,
1990; Robb and Meyer, 1995a,b). During diagenesis, maturation of
shale-hosted organic matter presumably liberated hydrocarbons that
migrated forming complex ions with Au and U (Drennan and Robb,
2006). High-grade Au and U mineralisation commonly occurs in association with fossil remnants of these hydrocarbons. Regional greenschistfacies metamorphism resulted in the formation of chlorite, muscovite,
pyrophyllite, and chloritoid-bearing mineral assemblages, with peak
conditions reaching 350 50 C and 3 kbar (Frimmel, 1997; Law
and Phillips, 2005). Peak metamorphic conditions have been
interpreted to be associated with the emplacement of the ~ 2050
2060 Ma Bushveld Complex (Stevens et al., 1997; Rasmussen et al.,
2007) or the 2023 4 Ma (Kamo et al., 1996) meteorite impact that
formed the Vredefort dome in the centre of the Witwatersrand Basin.
Higher temperatures (medium grade metamorphism) were reached locally around the Vredefort dome (Gibson and Reimold, 1999; Zartman
and Frimmel, 1999). Studies of post-sedimentary uid circulation in
the reefs indicate H2O-rich uids, with lesser amounts of CO2 and CH4,
generally low to moderate salinity ( 1018 wt.% NaCl equivalent),
and homogenization temperature T 370 C (Phillips et al., 1989,
1990; Drennan et al., 1999; Boer and Reimold, 2006; Jaguin et al., 2011).
Supergroup
Johannesburg
TauTona, Driefontein
mines
South
Africa
Bushveld
Complex
Transvaal
Carltonville
Sasolburg
Klerksdorp
Evander
Pretoria
Chuniespoort
Priel
Ventersdorp
27S
VD
Group
VCR
Platberg
Klipriviersberg
Central Rand foreland
basin
Edenville
Witwatersrand
West Rand
Bushveld Complex
28
N
0
km
Ventersdorp Supergroup
Central Rand Group
West Rand-Dominion Groups
<2.6 Ga cover
Welkom
50
27E
29E
Dominion
thermal
subsidence
rifting
Basement
Fig. 1. A) Surface and subsurface distribution of the Witwatersrand Supergroup, Ventersdorp Supergroup, and adjacent Bushveld Igneous Complex (modied from Catuneanu, 2001;
Frimmel, 2005). Sampling location (Tau Tona and Driefontein mines, Carltonville) indicated by a star. B) Simplied stratigraphic column of the Witwatersrand Basin and Transvaal Supergroup. VD: Vredefort Dome, VCR: Ventersdorp Contact Reef.
30
3. Previous investigations of CLP in the Witwatersrand Basin and

other sedimentary successions
3.1. CLP in the Witwatersrand Basin
CLP has been described from different conglomerates across the
Witwatersrand and Ventersdorp Supergroups (Ventersdorp Contact
Reef, Kimberly Reef, Basal Reef, Elsburg Reef), and in the younger succession of the Transvaal Supergroup (e.g. Black Reef; Schweigart and
von Rahden, 1965; Saager and Mihalik, 1967; Saager, 1970; Barton
and Hallbauer, 1996; England et al., 2002a; Reimold et al., 2004).
These and other studies include textural classications, description of
mineral inclusions, and elemental and isotopic analyses.
CLP grains have a diameter ranging approximately from 1 to 5 mm,
and are spherical to ellipsoidal. Angular fragments of broken spherical
grains are common, indicating reworking and/or transport. They have
concentric, colloform and radial textures, and in many cases, these textures can be found combined in single grains (e.g. Schweigart and von
Rahden, 1965; England et al., 2002a; Frimmel, 2005). Crystals forming
radial textures have been interpreted as former marcasite replaced by
pyrite during metamorphism (Barton and Hallbauer, 1996). Schweigart
and von Rahden (1965) noted a textural similarity between oolitic
pyrite of the VCR and iron ooids from the Pretoria Group, and ascribed
pyrite formation to biological activity. Hallbauer and Barton (1987)
considered a sedimentary origin for the CLP grains via winnowing and
accretion of pyrite mud and other ne-grained detritus in relatively
quiet waters.
England et al. (2002a) measured wide 34S variations between grains
and within single grains of CLP from the VCR, and interpreted this pyrite
as suldised carbonate or evaporite ooids. According to these authors,
the ooids would have been replaced in the source area, before introduction in the Witwatersrand conglomerates as detrital grains. An alternative interpretation involved post-depositional suldation of other
minerals, such as FeTi oxides (Fe pisoids; (Phillips and Law, 2000) during hydrothermal processes. This second interpretation related pyrite,
as well as Au and U, deposition in the Witwatersrand Basin to a hydrothermal/metamorphic crystallisation stage and challenged the detrital
origin of pyrite.
Quartz, chlorite, muscovite, rutile, orthobrannerite, galena, gold and
monazite were found as inclusions in CLPs (Saager, 1970;Ramdohr,
1958; Barton and Hallbauer, 1996). In addition, previous studies of porous and concretionary pyrite in different Witwatersrand conglomerates have reported uraninite (Schidlowski, 1965), chromite and zircon
(in mudball pyrite from the Black Reef and Kimberly Reef, Barton and
Hallbauer, 1986). Different studies have found that CLPs are enriched
in Au in comparison to massive pyrite in the same rocks (e.g. Barton
and Hallbauer, 1996; Koglin et al., 2010), and up to ~ 8 ppm Au have
been measured in laser ablation ICP-MS analyses (Koglin et al., 2010).
Proton microprobe (PIXE) major and trace element mapping of concentric pyrite from the Kimberley Reef and VCR has shown the heterogeneous distribution of Au, which is present both as invisible gold and as
visible inclusions (e.g. Foya et al., 1999; Reimold et al., 2004). Barton
and Hallbauer (1996) found that concretionary pyrite, consisting of
several concentric bands of radial pyrite after marcasite, has very radiogenic Pb compositions in comparison with massive detrital pyrite in the
same reef, which suggests addition of Pb derived from U decay.
3.2. CLP in other successions
CLP grains similar to those described from the Witwatersrand Basin
have been described from other sedimentary successions. Present in
Phanerozoic marine black shales and sandstones, such grains have
been interpreted as either replacements of iron ooids or primary diagenetic precipitates (Binda et al., 1985; Binda and Simpson, 1989;
Schieber and Riciputi, 2004, 2005). For example, Schieber and Riciputi
(2005) and Schieber (2007) described coated grains of alternating
cortices of pyrite and marcasite from the Ordovician shallow marine

sandstone of the Black Island Member of the Winnipeg Formation.
They suggested an origin of this type of grains by early diagenetic precipitation, and envisaged a process of intermittent reworking, abrasion,
reburial and further precipitation of concentric laminae. Formation
would have occurred at times of minimum sedimentation, when organic matter was enriched in surface sediments. Phanerozoic coated phosphate grains, also containing pyrite laminae, show textural similarity
with the grains described here. They formed authigenically just below
the sedimentwater interface and were affected by multiple episodes
of reworking during periods of stratigraphic condensation (Pufahl and
Grimm, 2003). Xenophontos et al. (1999, 2001) described concentric
pyrite grains with radial arrangement of crystals in carbonate horizons
in the Bulawayo-Bubi greenstone belt, Zimbabwe. The authors
interpreted these grains as due to replacement of carbonate ooids. Similar concentric textures were described in marcasite in Tertiary to recent
auriferous quartz-pyrite conglomerates from the Belle-Brook deposit,
New Zealand (Falconer et al., 2006). These grains are described as
formed by colloform-banded radial marcasite (Falconer et al., 2006),
and the resemblance with Witwatersrand concentric pyrite has also
been noted by England et al. (2002a). In modern environments, pyrite
concretions that show locally concentric and radial textures have been
shown to form in lagoonal environments (Frizzo et al., 1991).
4. Methods and analytical techniques
4.1. Sample preparation
Samples 01, 02, 03, 13 and 16 were collected from the Ventersdorp
Contact Reef (VCR) at the Tau Tona mine, Carltonville area (Fig. 1),
and derived from a large quartz pebble conglomerate at 66 level
(stope panel E1A). Samples VCR_E_Drie, VCRa, VCRb and VCRc were collected at the Driefontein gold mine, Carltonville gold eld. Samples
were prepared as 25 mm and 30 mm diameter polished rock chip
mounts, and described in reected optical microscopy and backscattered electron mode (SEMBSE) of scanning electron microscopy
using a Jeol JSM 5600 instrument at the University of Johannesburg.
Mineral identication was aided by energy dispersion spectrometry
(SEMEDS) on the same instrument.
4.2. X-ray elemental maps and electron probe micro analysis (EPMA)
X-ray compositional maps of pyrite were obtained using a 4
detector-equipped Cameca SX100 electron microprobe at the Spectrum
centre of the University of Johannesburg. Operating conditions were
20 kV acceleration, 40 to 100 nA beam current, and 250 to 500 ms
dwell time per spot. Arsenic (L), Co (K), Ni (K), and Pb (M)
were analysed by wavelength dispersion spectroscopy (WDS), whereas
S (K), Mg (K), K (K), Si (K), and Ca (K) were analysed by energy
dispersion spectroscopy (EDS). Maps are up to 256 256 pixels in size,
with 1.5 to 2.2 m step (distance between analysis spots, or spatial definition), and show element distributions as total counts. Some of the
largest maps (N500 m) show a shading effect towards the edges.
This is an analytical artefact due to the progressive misalignment of
the WDS spectrometers as the electron beam moved away from the centre of the mapped area; it does not reect compositional variations of
the sample. In comparison with laser ablation maps, electron microprobe X-ray maps have higher detection limit and a consequent lower
number of elements mappable. The main advantage of electron microprobe maps is the higher spatial denition, which allows ne textural
details down to ~1 m in size to be accurately mapped and described.
Electron microprobe spot analyses (a total of approximately 220
spots) were used in conjunction with X-ray compositional maps, to quantify the trace element contents of pyrite generations identied during Xray mapping. The following analytical setup was used: 20 kV acceleration
potential difference, 40 nA beam current, fully focused beam (b1 m). A
31
set of suldes and native elements, including FeS (for Fe and S), ZnS, PbS,
SbS, Cu, Co, As, Se, Bi, Ag, Cd, Sn, Te, and Mn, was used as reference materials. Analysed elements include Fe (K, analysed for 24 s both on-peak
and off-peak), Co (K, 21 s), Ni (K, 22 s), Cu (K, 24 s), Zn (K, 29 s)
(analysed on a LIF crystal), S (K, 30 s), Sb (L, 41 s), Te (L, 35 s), Mn
(K, 14 s) (PET crystal), As (L, 65 s), Se (L, 55 s) (TAP crystal), Pb
(M, 36 s), Bi (M, 40 s), Ag (L, 14 s), Cd (L, 20 s) and Sn (L, 10 s)
(LPET crystal). Detection limits, based on counting statistics, are between
100 and 350 ppm by weight for most minor/trace elements, except Cu
and Zn, around 600 ppm. Among minor and trace elements, only Co, Ni,
As and Pb revealed concentrations above detection limits of the instrument. Copper, Se, Bi, Ag, Cd, Sb, Te, Mn, and Zn are almost consistently
below detection limit. In CLP, Cd and Sb are locally present in detectable
amounts.
mode using two off-axis Faraday cups (C and H1). The gains of these
Faraday cups were determined at the beginning of the analytical session. The background of the detectors was measured during the presputtering (1 min) at the beginning of each analysis. The count rates
were typically 2.108 cps for 56Fe. The analysis consisted of 30 cycles
of 5 s acquisition time. No chromium was detected in the samples and
so no 54Cr correction on the 54Fe peak was necessary. The internal precision was typically less than 0.1 (2n), and the external reproducibility based on multiple measurements of pyrite in-house reference
material (Balmat, 56Fe = 0.399; Whitehouse and Fedo, 2007)
was better than 0.2 (2). Iron isotopes are expressed in the
standard notation relative to the IRMM-014 international standard.
4.3. Laser Raman spectroscopy
Pyrite trace element concentration maps were obtained at CODES,

University of Tasmania, on a New Wave 213-nm Nd-YAG solid-state
laser microprobe coupled to an Agilent 7500 quadrupole ICP-MS. The
ablation system is equipped with a small-volume cell, which assures
quick response (b1 s) and short wash-out time. Ablation was performed
in a He atmosphere, and the carrier gas was mixed with Ar immediately
after the ablation cell. The laser was operated at 10 Hz repetition and
3.5 J/cm2 uence. Spot size was between 10 and 22 m. Each image is
composed of multiple parallel lines ablated at a speed corresponding
to the spot size (e.g. 22 m/s with a 22 m spot size). After ablating
each line, a time of 15 s was allowed to purge the cell before the next
line. STDGL, a Li borate glass of fused sulde developed at CODES, was
used as the reference material, and was measured at the beginning
and the end of each map along with the background signal (blank acquisition) to correct for instrument drift. For a detailed description of
the method, including reference materials, see Large et al. (2009) and
Danyushevsky et al. (2011). The element intensities (counts per second) were then converted to concentrations (ppm by weight) using
in-house software. The element concentrations shown in each pixel of
the maps can be plotted as individual analyses, thus giving the opportunity to ascertain correlations between elements.
In situ, non-destructive Raman analyses on various minerals were

performed with a confocal laser Raman microscope (WITec alpha300R)
at the Department of Geology, University of Johannesburg. The spectral
range of the spectrometer is 1004400 cm1. Raman spectra were collected using 20 and 100 Nikon objectives and a frequency doubled
Nd:YAG (532 nm) Ar-ion 20-mW monochromatic laser source. Beam
centring and Raman spectra calibration were performed daily before
spectral acquisition using a Si standard (111). The optimum power for
analyses of different minerals was determined experimentally. Raman
spectra were compared with reference spectra from RRUFF Database
(Laetsch and Downs, 2006), and spectra from literature.
4.4. Secondary ion mass spectrometry (SIMS) S and Fe isotope analysis
4.4.1. Sulfur isotope measurements
Sulfur isotope compositions (34S and 33S) were measured by
secondary ion mass spectrometry (SIMS) using a Cameca ims 1280
HR2 at CRPG (Nancy, France). The analytical method is described in
details in Thomassot et al. (2009). Briey, a Cs+ primary beam of 5 nA intensity was focused to a spot of about 1520 m. 32S, 33S and 34S were simultaneously measured in three off-axis Faraday cups (L2, C and H1).
The relative gains of the Faraday cups were intercalibrated at the beginning of the analytical session. Typical 32S intensity was between 6 and
10 108 counts per second (cps) depending on the sulde mineral
analysed. Several sulde in-house reference minerals (Maine, Spain
and Balmat, Marin-Carbonne et al., 2014) were used to determine
(i) the instrumental mass fractionation and (ii) the reference mass discrimination line, from which 33S was calculated. A typical analysis
consisted of 2 min of pre-sputtering followed by 30 cycles of 3 s each.
The background of each detector was measured during the presputtering and was then corrected for each analysis. The internal precision achieved in these conditions was better than 0.05 for 34S and better than 0.10 for 33S (2n). The reproducibility, based on multiple
measurements of the reference materials, for 34S was 0.40 (2),
and the reproducibility for 33S was 0.06 (2n). Sulfur isotopes are
expressed as per mil variation relative to the Vienna Canyon Diablo Troilite (VCDT) international reference, as 3xS = 1000 [(3xS / 32S)sample
(3xS / 32S)VDVT / (3xS / 32S)VCDT]. Deviations from linear relations between
33S and 34S reect mass-independent fractionation (MIF), and can be
expressed as 33S = 33S 1000 [(1 34S / 1000)0.515 1], as per
mil variation.
4.4.2. Iron isotope measurements
Measurements of iron isotope compositions were performed by
SIMS using a Cameca ims 1270 at CRPG (Nancy, France). The basic settings are described in detail in Marin-Carbonne et al. (2011). Briey, a
16 2
O primary beam of about 10 nA intensity was focused to a spot of
about 15 m. The instrument was set to achieve a mass resolution of
about 7000. Secondary 54Fe and 56Fe were measured in multicollection
4.5. Laser ablation ICP-MS trace element maps
5. Sample description and pyrite textures

Samples were derived from clast-supported quartz-pyrite pebble
conglomerate. Pebbles are up to a few cm in size, and include various
types of chert and vein quartz, pyrite, and minor silicied mac rocks
(Fig. 2A). The matrix is sand-sized, mostly composed of quartz, pyrite
and minor amounts of monazite and apatite. Secondary quartz, chlorite
and muscovite occur in interstitial position and formed during
greenschist grade metamorphism (Phillips et al., 1990). The matrix of
sample VCR_E_Drie also contains a small lens or clast of mudrock.
Under the microscope, three main textural types of pyrite can be distinguished, mostly based on habit and the amount of inclusions or
pores. The term porous has been widely used in the literature to indicate inclusion-rich (mostly muscovite and chlorite) pyrites, although
these inclusions were likely trapped during the growth of pyrite and,
therefore, the pores were never cavities at any time. We thus refrain
from using this term.
The pyrite textural types are as follows: 1) anhedral (round to
angular) massive (inclusion-poor, also referred to as compact in the
previous literature) pyrite, mostly b 2 mm in size; 2) large (up
to ~ 5 mm or more) anhedral grains of inclusion-rich pyrite (Fig. 2B)
(cf. Saager, 1970; Utter, 1978; Hallbauer, 1986; Barton and Hallbauer,
1996; England et al., 2002a; Guy et al., 2010); and 3) euhedral massive
pyrite, up to a few mm in size, which can also form overgrowths on
round grains (Fig. 2C). Gold is present as anhedral grains, up to
100 m in size, and occurs in the matrix, as inclusions hosted in massive
pyrite along cracks (Fig. 2D) and as lling of secondary pores and cracks
in inclusion-rich pyrite. The rocks show limited evidence of compaction
and brittle-deformation and no evidence of plastic deformation, with
32
Fig. 2. Mineral textures of the Ventersdorp Contact Reef. A) Round pebbles of polycrystalline quartz (Qtz), mac rock and pyrite (Py) in sand-sized matrix (sample 13, scanned polished
mount). B) (and inset) Round inclusion-rich pyrite including angular fragments (arrowed). Holes from SIMS analysis (sample 01, reected polarised light). C) Banded (parallel laminated)
pyrite rimmed by massive sub-euhedral pyrite. Contact between two types of pyrite marked by dashed line (sample 03, reected polarised light). D) Fractured grain of massive round
pyrite; electrum (el) is present along fractures and at margin of the pyrite grain (sample VCRc, reected polarised light).
the exception of CLP grains, which show widespread evidence for

plastic deformation.
5.1. Round massive pyrite
This type of pyrite is generally up to 12 mm in size, with rounded or
anhedral margins. Etching of round massive pyrite from different reefs
and trace element mapping have shown euhedral growth textures,
invisible in optical microscopy, truncated along grain margins
(MacLean and Fleet, 1989; Frimmel, 2005; Agangi et al., 2013). Massive
round pyrite is interpreted to be detrital in origin (e.g. MacLean and
Fleet, 1989).
5.2. Euhedral massive pyrite
Euhedral massive pyrite is abundant in our samples and in the VCR
in general. Its shape and the textural relationships with other minerals
clearly indicate a secondary, post-sedimentary origin. Euhedral massive
pyrite forms whole grains and overgrowths around inclusion-rich and
laminated grains (Fig. 2C). Massive pyrite also occurs as veinlets, or as
inlling of larger cavities present in some inclusion-rich grains.

Although invisible in optical and electron microscopy, overgrowths
around massive detrital pyrite grains have also been shown through
etching and compositional maps (Frimmel, 2005; Agangi et al., 2013).
Round and euhedral massive pyrite types contain a number of secondary mineral inclusions: suldes (sphalerite, chalcopyrite, galena),
quartz, chlorite, and electrum (alloy of gold and silver). In contrast to
the inclusion-rich type, mineral inclusion typically form a minor
proportion (b1 vol.%) of the grains.
5.3. Round inclusion-rich pyrite
This type of pyrite is characterised by an abundance of mineral inclusions typically forming large proportions of the grains. Inclusions, mostly consisting of muscovite, chlorite and quartz, are typically distributed
homogeneously through the grain, conferring a spongy appearance to
the pyrite. Some grains have composite textures, and contain angular
massive domains, possibly clasts of detrital pyrite included during formation of the inclusion-rich grain (Fig. 2B). Round inclusion-rich pyrite
is interpreted to be detrital. Further, some grains (Fig. 2) are crossed by
Fig. 3. Pyrite textures in the Ventersdorp Contact Reef. A) Fragment of concentrically laminated pyrite plastically deformed at the contact with a more competent massive pyrite grain
(sample VCR02, reected light). Deformation in this grain may have taken place during burial. Inset: anhedral (detrital) monazite inclusion wrapped by the immediately external lamina
(BSE image). B) Fragment of concentrically laminated pyrite in a quartz (Qtz)-chlorite (Chl) matrix. Fine laminae alternate with more massive, botryoidal layers, which surround a nonlayered core (bottom right). Inset: apatite (Ap) and secondary monazite (Mnz) are present as inclusions (sample 02, BSE image). C (and insets)) Well-preserved concentrically laminated
pyrite in a quartz-chlorite-dominated matrix (sample VCR_E_Drie, reected light). Inset 1: colloform-textured core (arrowed); inset 2: plastically deformed, inward bent laminae overgrown by undeformed laminae (sample VCR_E_Drie, reected polarised light). Deformation must have occurred during grain formation. D) Concentric, nely laminated pyrite with a
round massive core; contact between the two types of pyrite marked by dashed line (sample 01, reected polarised light). Holes from SIMS analyses. E) Concentrically laminated pyrite
grain with partly exfoliated and truncated laminae. Electrum (el) particles are visible in the matrix (sample VCRb, reected polarised light). F) Concentrically laminated pyrite also containing radially arranged pyrite grains (arrowed). Portions of laminae are truncated (sample VCRb, reected polarised light). G) (and inset) Fragment of concentrically laminated pyrite
with secondary gold around the pyrite and along cracks (sample VCRc, reected polarised light).
33
34
veinlets of massive pyrite truncated at grain margins. These veinlets

predate transport and grain rounding.
5.4. CLP
Concentrically laminated grains are approximately spherical, up to a
few mm in diameter, and include nely (m-scale) laminated domains
alternating with thicker (tens of m) laminae, and non-laminated domains (Fig. 3AG). Several grains are broken into angular fragments,
and laminae in these grains are truncated by grain margins, indicating
reworking and transport. Non-laminated portions of grains have
variable microscopic textures, such as botryoidal or angular (possibly
clastic) (Fig. 3B). Radial arrangements of pyrite are also present
(Fig. 3F; e.g. Saager, 1970). In some of these grains, concentric laminae
surround non-concentric cores, such as complex cores including a botryoidal/colloform portion (Fig. 3C), and round massive pyrite cores
(Fig. 3D). Other grains, e.g. VCRb_04 (Fig. 3E) are concentrically laminated throughout, and do not show any core, although this may be an
artefact of sectioning. Some CLP contain chalcopyrite discontinuously
distributed between pyrite laminae.
Evidence of plastic and brittle-plastic deformation was observed in
several concentrically laminated grains. Local bending of laminae indicates plastic deformation at the contact with more competent massive
pyrite grains (Fig. 3A). In grain VCR_E_Drie_01, a set of laminae is truncated and bent inwards (Fig. 3C), and the indentation is inlled by the
immediately external laminae. This suggests plastic deformation before
formation of the grain was complete (syn-formational deformation),
followed by accretion of successive laminae. In grains VCRb_02 and
12_01, the external laminae appear to have been partly exfoliated and
truncated (Fig. 3E); partial exfoliation might may have also occurred
in grain VCRb_01, removing portions of laminae (Fig. 3F). In grain
VCRb_02, cracks cross-cut the grain and the fragments on both sides
of the cracks are partially displaced. The amount of displacement is uneven along the cracks, implying a relative compression of one of the
fragments. Evidence for brittle-plastic deformation, including indentations, attening and radial cracks, and hard minerals such as zircon
pressed into softer pyrite, has been previously described in similar
grains from different Witwatersrand conglomerates (Hallbauer and
Utter, 1977; Utter, 1978; Hallbauer, 1986).
The style of deformation of CLP is in sharp contrast with the purely
brittle deformation experienced by the rocks and other types of detrital
pyrite, and must be controlled by a difference in their rheological characteristics. It is also worth noting that pyrite is believed to behave rigidly
up to temperatures of N400 C (McClay and Ellis, 1983), and is not likely
to have behaved plastically during the metamorphic conditions estimated for these rocks (350 C). Plastic deformation of CLP must have occurred before lithication of the conglomerate, and suggests a softer
precursor to these grains.
6. Mineral inclusions (laser Raman and SEM-EDS study)
A combination of laser Raman spectroscopy and SEM-EDS was used
to identify pyrite-hosted mineral inclusions and minerals in the surrounding matrix. Inclusion-rich and laminated pyrite grains contain numerous mineral inclusions of quartz, chlorite (clinochlore-chamosite),
muscovite, apatite, monazite, Ti oxide (anatase and rutile, the latter
mostly found in inclusion-rich pyrite), and galena (mostly found in
inclusion-rich pyrite) (Fig. 4). Chalcopyrite and electrum (samples
VCRa, VCRb, VCRc) were found locally and occur between pyrite laminae and in cracks, apparently lling secondary pores and inter-lamina
spaces (Figs. 2D and 4G). Apatite and rutile mostly have an anhedral
habit and appear rounded, evidence of a detrital origin. Monazite inclusions either have a rounded shape, indicating a detrital origin (Fig. 3A),
or appear to inll secondary pores in pyrite, thus implying a secondary
origin (Fig. 3B). Carbonaceous matter has been detected (Fig. 4): 1) in
association with pyrite in concentrically laminated grains; 2) in
association with chlorite in the pores of some pyrite grains; and 3) in

ne grained portions of the conglomerate matrix. In the matrix,
clinochlore, quartz, muscovite, rutile, ilmenite-pyrophanite [(Fe,Mn)
TiO3] and pyrite were identied.
7. Elemental composition and intragranular textures of CLP
(EPMA and LA-ICP-MS)
7.1. Electron microprobe X-ray maps
X-ray element distribution maps show more complex textures than
observed by optical microscopy and BSE images. In CLP, the composition
of each lamina can be signicantly different from the adjacent laminae
(cf. EPMA traverse in Figs. 5 and 6). Massive-looking laminae in grain
02_02 are internally subdivided in m-scale sub-laminae (Fig. 5).
Electron microprobe spot analyses of CLP grains conrm these wide
compositional variations between laminae, particularly of As and Ni
(As 0.010.56, Ni 0.050.62 wt.%, Table 1). Composition traverses and
plots indicate good correlation between As and Ni (Fig. 5C). Cobalt
and Pb, which show minor or no variations in X-ray maps, are mostly
below 0.1 wt.%. Some laminae in grains VCR02_01 (Supplementary
Fig. 1) appear massive in BSE images, but show clastic texture in compositional maps, and contain distinct subhedral to anhedral (round or irregular) grains. These grains are mostly trace element-poor, and are
surrounded and cemented by trace element-rich pyrite. In X-ray
maps, K, Mg, and Si peaks are indicative of the distribution of negrained chlorite and muscovite, and Ca is associated with apatite situated between and within pyrite laminae.
7.2. LA-ICP-MS maps
CLP grains analysed by LA-ICP-MS (samples 02, VCR_E_Drie and
VCRb) contain signicant Au, Ag, As, Ni, Co, Sb, Mo, Cu, Tl, and Pb. Concentrations up to several tens ppm Au have been measured in grains without
visible gold. In grains containing visible gold, up to ~40,000 ppm Au were
occasionally measured, indicative of the presence of electrum inclusions.
Maximum As, Ni and Co concentrations are 1 wt.% or above, and all other
elements have concentrations up to several tens ppm (Tl) or thousands
ppm (Mo, Cu, Pb). A comparison between photomicrographs and elemental maps indicates that some elements tend to be included in pyrite
or locally in chalcopyrite (Co, Ni, Mo, Cu, As, Sb, Te, Tl, Pb), whereas others
are mostly associated with the silicates lling spaces between laminae
(Mg, Si, K, Mn, Ti, Zn, Sn, U). Gold and Ag occur both in pyrite and in
the interstices between laminae (Fig. 7). Co-occurrence of Ti, Ta, and locally U, in discrete peaks indicates the presence of inclusions of Ti oxide
(rutile, anatase), which can be U-bearing. Peaks of Gd, mostly present
in the matrix and between laminae, are probably due to monazite. The
enrichment of U, Zn, Pb, and locally Au and Ag, along cracks and less massive portions of grains demonstrates their post-sedimentary mobilisation.
Enrichments in Co, Ni, Mo, Cu, As, Sb, Te, Tl, and Pb have been previously
described in diagenetic pyrite from marine carbonaceous sediments
(black shale; Large et al., 2009).
In the Au vs. As plot (Fig. 8), the analyses of all grains form two
clusters/groups: 1) a large and dense group with Au up to ~1020 ppm,
and 2) a smaller group with higher Au concentrations, which suggests
the presence of gold inclusions. The upper limit of the main group of analyses (group 1) has a slight positive slope that lies at lower Au values than
reported by Reich et al. (2005) for sediment-hosted deposits at temperatures up to 250 C, possibly indicating lower Au solubility in the VCR pyrite. In the Au vs. Ag plot, the analyses are characterised by bimodal
distribution of Au/Ag, with values of ~0.1 and ~10, and the high-Au analyses mostly have high Au/Ag (Fig. 8). The co-occurrence of Au, Ag and Hg
indicates that high Au concentrations are due to electrum inclusions, locally (e.g. grain VCRb_02) visible under the microscope (Fig. 3E, G). Low
concentrations of Te and Se and the lack of correlation of these elements
with Au suggest that tellurides and selenides are not signicant carriers of
35
Fig. 4. Examples of ne-grained minerals identied by Raman spectroscopy. A (and inset)) Fine-grained portion of the conglomerate matrix. B) Contact between a concentrically laminated
grain and the surrounding matrix. C) Detail of the non-laminated core of a pyrite grain. CM indicate the rst- and the second-order Raman spectra of the carbonaceous matter marked with
D (disordered) and G (graphitic) bands. Crosses indicate location of spectra of the same number. Sample VCR_E_Drie.
Au in the analysed pyrite. The chlorite-muscovite-rich matrix around

grains is enriched in uranogenic Pb (206Pb, derived from the decay of
238
U), and has generally low 208Pb/206Pb ratios (208Pb/206Pb = 0.021
in comparison with bulk-Earth values; Stacey and Kramers, 1975), with

the possible exception of grain VCR_E_EDrie_01 map 2), indicating that
the 208Pb/206Pb was modied, likely by metamorphic uids enriched in
36
EPMA
traverse
BSE
grain 02_02
B
0.6
4
47
0.5
Mnz
17
46
13 14
45
11 12
10
15
18
16
20
21
19
element, wt.%
48
Co
Ni
As
Pb
0.3
0.2
0.1
44
43 42
VCR02_EPMA_02
0.4
36
35
200
400
mapped area
13
1000
1200
Co
Pb
Ni
As
600
800
distance, m
C
16
16
16
13
14
5
16
13
15
14
5
45
14
5
15
45
44
16
13
15
45
44
Si
14
45
44
16
13
15
S analyses
Fe analyses
44
Ca
Mg
Fig. 5. BSE image (A), electron microprobe spot analyses (B), and X-ray compositional maps (C) of concentrically laminated pyrite from the Ventersdorp Contact Reef. Nickel and arsenic
concentrations vary signicantly between successive laminae. Circles indicate position of SIMS analyses. Grain 02_02.
37
Fig. 6. BSE image (A), electron microprobe spot analyses (B) and X-ray compositional maps (C) of concentrically laminated pyrite from the Ventersdorp Contact Reef. Fine pyrite laminae
deposited around a massive core are marked by variations of As and Ni. Note veinlets of AsNi-rich pyrite truncating the grain. Circles indicate position of SIMS analyses. Grain 01_01.
U and radiogenic Pb. This result is consistent with previous studies of the
VCR (Zartman and Frimmel, 1999) and Black Reef (Barton and Hallbauer,
1996), and with the notion of U-rich uids in the Witwatersrand reefs
(Vollbrecht et al., 2002). In contrast, the pyrite grains have mostly low
uranogenic Pb concentrations (208Pb/206Pb 1), although the inlling
minerals in the most porous portions of pyrite and cracks also tend to
have high concentrations of Pb produced by decay of U, suggesting that
metamorphic uids locally inltrated the grains and partially reset their
isotopic composition. This alteration is more pronounced at grain margins, and proceeded inwards along cracks and spaces between laminae,
and is commonly associated with higher content of U, Zn, Pb, and locally
Au and Ag (Fig. 8).
8. Sulfur and Fe isotope composition of pyrite

8.1. S isotopes
CLP grains have wider variations of 34S ( 12.16 to +12.61) in
comparison with other types of pyrite (Table 2, Fig. 9,), a feature also
observed for CLP and inclusion-rich (sooty) pyrite from the Carbon
Leader Reef (Large et al., 2013). Signicant variations in 34S were
measured both between laminae of individual grains and between grains
(Figs. 10 and 11). Inclusion-rich pyrite grains have 34S varying
from 5.69 to +1.43, and distinct intergranular variations (Fig. 9).
Massive pyrite grains have 34S varying from 4.95 to +4.05.
38
Fig. 7. Reected light image (A), Raman images (B) and laser ablation ICP-MS compositional maps (C) of concentrically laminated pyrite (grain VCRb_01). 2D intensity Raman images were
generated using 1660 cm1 (G peak) and 463 cm1 characteristic spectral positions of carbonaceous matter and quartz, respectively. Note the presence of electrum inclusions (arrowed).
Comparison of 34S of different massive grains indicates that much of the

S isotopic variation is due to intergranular heterogeneity, whereas single
massive grains are relatively homogeneous, having ranges of b0.5
(Fig. 9). A few analyses of pyrite overgrowths have negative 34S
(4.95 to 3.04).
Mass-independent fractionation of S isotopes (MIF-S) is strongly dependent on pyrite textures, with values of 33S that vary between 1.35
and +0.64 (Fig. 12). CLP has negative 33S values (1.35 to 0.17),
whereas inclusion-rich pyrite has positive 33S values (up to +0.64),
except a few analyses affected by post-sedimentary recrystallisation or

veining. Massive pyrite has mostly near-zero or slightly positive 33S
(0.02 to 0.20, these values are within the range of mass-dependent
fractionation, Farquhar and Wing, 2003), with the exception of the
massive core of grain 01_01 (33S = 0.37 to 0.36; Fig. 9), which
has possibly been affected by post-sedimentary Ni-As-rich veinlets,
visible in X-ray maps. Post-depositional pyrite, occurring as overgrowths, recrystallised pyrite, and veinlets, have negative 33S values
(0.21 to 0.32; Fig. 13).
39
Fig. 8. BSE image, laser ablation ICP-MS compositional plots and maps of concentrically laminated pyrite. Spots are colour-coded based on Au content and the presence of gold (electrum)
inclusions, as indicated by the Au vs. As plot. Gold inclusions are concentrated along a crack (also affected by U inltration) and in the most inclusion-rich portions of the grain, and are of
likely secondary origin. Gold solubility in the VCR (dashed line in the Au vs. As plot) was probably lower than in Reich et al.'s (2005) model. A large proportion of electrum inclusions have
Au/Ag ~10 (grain 02_14). (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)
8.2. Fe isotopes
Values of 56Fe in all the pyrite analysed are between 1.5
and + 5.04 (Table 3, Fig. 14). These values partly overlap with, and
partly exceed, the values previously measured in late Archaean pyrite
from shales and in BIF (Johnson et al., 2003; Rouxel et al., 2005). Since
most available 56Fe data have been measured on whole-grains, the
wider scatter of data reported here can be partly explained by smallscale heterogeneity. This is also conrmed by recent analyses of Late
Archaean pyrite by in situ techniques (Yoshiya et al., 2012;
Marin-Carbonne et al., 2014), which indicate wide variations of 56Fe.
Comparison with previous whole-grain analyses of inclusion-rich and
massive VCR pyrite (Hofmann et al., 2009) indicates that in situ analyses
presented here have larger variations of 34S, 33S and 56Fe (Fig. 12),
suggesting that whole-grain analyses provide average values, and highresolution in situ analyses resolve small-scale heterogeneity.
CLP have wide 56Fe variations, spanning between 1.50
and + 3.32 (average + 0.20 0.98). Intragranular differences
between laminae can account for most of this variation (Figs. 10, 11
and 14). The 56Fe frequency histogram indicates an asymmetric
distribution of values, with a peak roughly between 0.50 and

0.00, and decreasing frequencies for higher 56Fe values (Fig. 14).
This distribution is complementary to the distribution of 56Fe for
inclusion-rich pyrite, which has a distribution biased towards higher
values. Comparison of different 56Fe traverses of CLP indicates contrasting Fe isotopic trends. These contrasting trends may be due to the
relatively large spot size (~ 15 m) in comparison with the thickness
of the laminae (a few m), or to varying isotopic composition along laminae. In concentrically laminated grain 02_02, average values increase
from internal laminae to the external ones, although standard deviation
is considerable (Supplementary Fig. 2).
Inclusion-rich pyrite has generally higher 56Fe values than other
types of pyrite, being largely strongly N0 and up to + 5.04,
average + 2.75 1.26, with one outlying analysis at 1.57. This
range of 56Fe values is due to both intragranular and intergranular variations. Locally, recrystallisation and veining have imparted lower 56Fe
values to inclusion-rich pyrite, as shown by partially recrystallised
inclusion-rich grain 01_04, which has moderate 56Fe values (+ 0.89
to + 1.18), and by a pyrite vein cross-cutting grain 01_02, which
has distinctively lower values (+ 0.40 to + 0.84, average + 0.56;
40
Table 1
EPMA of pyrite.
Type
Massive (n = 23)
Overgrowth (n = 6)
Concentrically laminated
(n = 35)
Inclusion-rich
(n = 30)
Avg
St dev
Min
Max
Avg
St dev
Min
Max
Avg
St dev
Min
Max
Avg
St dev
Min
Max
Weight%
S
Mn
Fe
Co
Ni
Cu
Zn
As
Se
Pb
Total
53.13
bdl
46.20
0.02
0.10
bdl
bdl
0.18
0.01
0.02
99.68
0.94
0.01
0.61
0.03
0.15
bdl
0.01
0.37
0.01
0.02
1.25
49.8
bdl
43.96
bdl
bdl
bdl
bdl
bdl
bdl
bdl
94.72
53.97
0.02
46.74
0.1
0.46
0.02
0.03
1.64
0.02
0.08
100.83
52.80
bdl
45.57
0.10
0.51
bdl
bdl
0.81
bdl
0.02
99.83
0.31
0.00
0.49
0.13
0.39
0.01
0.01
0.60
0.01
0.02
0.36
52.39
bdl
44.92
0.01
0.12
bdl
bdl
0.19
bdl
bdl
99.33
53.14
bdl
46.24
0.35
0.94
0.03
0.03
1.67
0.01
0.05
100.3
52.35
bdl
45.20
0.04
0.24
bdl
bdl
0.15
bdl
0.05
98.06
1.16
0.00
0.89
0.03
0.12
0.01
0.01
0.13
0.01
0.03
1.82
48.66
bdl
42.37
0.01
0.05
bdl
bdl
0.01
bdl
bdl
92.42
53.98
0.01
46.1
0.16
0.62
0.04
0.04
0.56
0.02
0.12
100.41
52.39
bdl
45.49
0.01
0.06
bdl
0.01
0.07
0.01
0.02
98.08
1.63
0.01
1.26
0.01
0.07
0.01
0.01
0.09
0.01
0.02
2.83
48.61
bdl
42.39
bdl
bdl
bdl
bdl
bdl
bdl
bdl
91.09
54.24
0.05
46.81
0.03
0.27
0.03
0.04
0.38
0.03
0.09
100.70
Atomic%
S
Mn
Fe
Co
Ni
Cu
Zn
As
Se
Pb
Co/Ni, weight
S/Fe, atomic
Co + Ni + Pb, atomic
66.58
bdl
33.24
0.01
0.07
bdl
bdl
0.10
bdl
0.00
0.28
2.00
0.15
0.30
bdl
0.21
0.02
0.10
bdl
0.01
0.20
0.00
0.01
0.34
0.02
0.17
65.8
bdl
32.9
bdl
bdl
bdl
bdl
bdl
bdl
bdl
0.00
1.96
0.01
67
0.01
33.78
0.08
0.32
0.01
0.02
0.89
0.01
0.02
1.00
2.04
0.56
66.30
bdl
32.85
0.07
0.35
bdl
bdl
0.44
bdl
0.00
0.55
2.02
0.62
0.30
0.00
0.34
0.09
0.27
0.01
0.01
0.33
0.00
0.01
1.16
0.02
0.37
65.91
bdl
32.41
bdl
0.08
bdl
bdl
0.1
bdl
bdl
0.03
2.00
0.21
66.63
bdl
33.27
0.24
0.65
0.02
0.02
0.9
0.01
0.01
2.92
2.04
1.08
66.66
bdl
33.05
0.03
0.17
bdl
bdl
0.09
bdl
0.01
0.16
2.02
0.32
0.24
0.00
0.18
0.02
0.09
0.01
0.01
0.08
0.00
0.01
0.10
0.02
0.16
66.07
bdl
32.78
0.01
0.03
bdl
bdl
0.01
bdl
bdl
0.03
1.98
0.09
67.11
0.01
33.45
0.12
0.46
0.02
0.03
0.31
0.01
0.02
0.50
2.04
0.85
66.67
bdl
33.24
0.01
0.04
bdl
0.01
0.04
0.00
0.00
0.17
2.01
0.10
0.19
0.01
0.20
0.01
0.05
0.01
0.01
0.05
0.00
0.01
0.23
0.02
0.08
66.29
bdl
32.77
bdl
bdl
bdl
bdl
bdl
bdl
bdl
0.00
1.98
0.00
67.07
0.03
33.59
0.03
0.19
0.02
0.03
0.20
0.01
0.02
1.00
2.05
0.31
bdl: below detection limit.
Fig. 14) than the rest of the grain (average +2.94). This vein is truncated by the grain margins, and therefore it pre-dates clastic transport
and nal deposition of the host conglomerate. Massive detrital pyrite
has 56Fe = 1.58 to + 3.74 (average = + 1.05 1.36). This
range is due to both large intergranular 56Fe variations and
intragranular heterogeneity (Fig. 14); for example, variation of N 3
was measured in grain 01_05.
Although no clear correlation can be seen between Fe and S isotopic
values, it is apparent (Fig. 12) that the highest 56Fe (3) have been
measured in inclusion-rich pyrite, which has 33S N0. No correlation
has been observed in pyrite nodules in 2.7 Ga old shales from the Bubi
Greenstone Belt, Zimbabwe (Marin-Carbonne et al., 2014). This is in
contrast with the results obtained by Archer and Vance (2006), who
have hypothesised a common microbial reduction of S and Fe. The
trend between 56Fe and 33S can be interpreted as reecting different
pyrite formation pathways, but may also be due to the fact that S and Fe
isotopes were measured in separate spots in proximity, but not coincident, and strong small-scale heterogeneity is clear from the analyses.
9. Discussion
9.1. Textural constrains on the formation of CLP
Textural features and the nature of mineral inclusions can help
constraining the environment of formation and the successive history
of CLP grains. The shape of laminae strongly suggests a growth of
these grains lamina by lamina, by successive accretion stages. This is indicated by the way laminae mantle detrital inclusions (Fig. 2F), and by
the succession of pyrite formation and syn-formational plastic deformation, followed by further deposition of laminae (Fig. 3C). The nature of
detrital inclusions gives important indications on the environment of
formation. Silt-sized detrital monazite, apatite and Ti oxides, together
with quartz (which recrystallised during metamorphism) are common
inclusions in concentrically laminated and inclusion-rich pyrite. In
addition, X-ray maps seem to indicate the presence of ne-grained

(b100 m) clastic pyrite in some grains (Supplementary Fig. 1). Muscovite and chlorite, which are abundant both as pyrite-hosted inclusions
and in the matrix, are interpreted to have formed as a result of metamorphism of clay minerals originally present between laminae. We
infer that all these minerals were present in the sediment at the time
of formation of CLP, and were scavenged by pyrite during growth. Siltsized detrital inclusions in CLP reect a depositional environment of
lower-energy than that of the conglomerates in which they are commonly found.
The occurrence of CLP as single-grain clasts in the conglomerate suggests that these grains were loose when included in the conglomerate
and not present in already lithied substrate. Derivation from the erosion of Central Rand Group strata is unlikely, as diagenetic pyrite is
rare in these rocks (Guy et al., 2010, 2012). Considering their mechanical instability and the angular shape of some CLP fragments, we infer
short transport with the gravel host (e.g. Saager, 1970). Further, the
studied grains bear evidence of brittle-plastic deformation that occurred
in situ (in the gravel), and during formation of the grains (Fig. 3A, C).
Thus, concentrically laminated grains were included in the gravel after
formation in a lower energy environment, and at a stage when they
were loose and still plastic. This in situ soft deformation is at odds
with the interpretation of these concentric grains via suldation of carbonate or evaporite ooids in the source region (England et al., 2002a),
because this process would have produced already rigid clasts and,
possibly, multi-grain clasts.
The concentric arrangement of pyrite laminae may have formed
while the grains were moving on the sediment surface, or during early
diagenetic growth under shallow sediment cover, as proposed for concentric sulde grains (Binda and Simpson, 1989; Pufahl and Grimm,
2003; Schieber and Riciputi, 2005). Successive reworking of early diagenetic grains and subsequent reburial is also possible. The absence of
replacement textures (e.g. Schieber and Riciputi, 2004) suggests a
primary origin of pyrite in these grains.
41
Table 2
Pyrite sulfur isotope analyses (SIMS).
Sample
Grain
Analysis ID
I (nA)
32S (cps)
34S,
33S,
33S,
Texture
01
01
01
01
01
01
01
01
01
01
01
01
01
02
02
01
01
02
02
02
02
02
02
02
02
02
01
01
01
01
01
01
01
01
02
02
02
02
02
02
02
02
02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_04
VCR01_04
VCR01_03
VCR01_03
VCR01_03
VCR01_03
VCR01_03
VCR01_03
VCR02_05
VCR02_05
VCR01_01
VCR01_01
VCR02_02
VCR02_06_a
VCR02_06_a
VCR02_06_a
VCR02_03
VCR02_03
VCR02_07
VCR02_07
VCR02_05
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
S_01_01
S_01_02
S_01_03
S_01_04
S_01_05
S_01_10
S_01_11
S_01_12
S_01_13
S_01_14
S_01_15
S_01_16
S_01_17
S_02_19
S_02_20
S_01_06
S_01_07
S_02_01
S_02_10
S_02_11
S_02_12
S_02_14
S_02_15
S_02_16
S_02_17
S_02_18
S_01_08
S_01_09
S_01_19
S_01_20
S_01_21
S_01_22
S_01_23
S_01_24
S_02_02
S_02_03
S_02_04
S_02_05
S_02_06
S_02_07
S_02_08
S_02_09
S_02_13
5.06
5.05
5.05
5.05
5.04
5.04
5.04
5.03
5.03
5.03
5.02
5.03
5.03
5.27
5.28
5.04
5.04
5.49
5.35
5.34
5.33
5.32
5.32
5.27
5.3
5.27
5.04
5.05
8.5
8.37
8.14
7.8
7.56
7.31
5.44
5.42
5.43
5.41
5.39
5.4
5.37
5.36
5.32
1,753,741,000
1,812,178,000
1,742,358,000
1,757,431,000
1,760,238,000
1,856,864,000
1,819,982,000
1,692,121,000
1,518,109,000
1,817,852,000
1,813,578,000
1,601,072,000
1,735,374,000
1,489,181,000
1,790,643,000
2,037,465,000
1,975,713,000
1,763,905,000
1,893,660,000
1,851,126,000
1,809,198,000
1,733,988,000
1,783,992,000
1,776,051,000
1,900,984,000
1,680,303,000
1,897,760,000
1,873,571,000
3,321,741,000
3,123,787,000
3,002,154,000
2,884,262,000
2,810,093,000
2,609,699,000
1,760,169,000
1,712,600,000
1,790,183,000
1,651,374,000
1,784,108,000
1,698,726,000
1,586,314,000
1,695,369,000
1,629,785,000
1.434
1.298
1.372
0.375
0.334
1.121
0.476
5.355
3.721
4.206
4.701
5.690
3.968
0.596
1.179
5.013
4.948
2.742
1.701
1.903
1.962
1.236
0.773
3.855
4.054
3.040
12.159
7.688
12.614
7.029
10.496
8.244
10.923
10.089
2.148
1.450
0.545
5.857
5.348
4.031
4.548
0.427
5.788
1.095
1.139
1.105
0.614
0.675
0.427
0.209
1.773
1.015
1.197
1.404
2.056
1.104
0.709
0.667
2.571
2.541
1.594
0.566
0.648
0.630
0.875
0.676
2.304
2.333
1.803
6.705
3.032
6.204
3.637
4.704
3.778
4.786
4.433
1.343
0.678
0.239
3.104
2.762
2.133
2.319
0.016
2.999
0.002
0.115
0.043
0.063
0.145
0.213
0.325
0.608
0.530
0.596
0.642
0.497
0.567
0.236
0.235
0.365
0.369
0.025
0.133
0.155
0.203
0.074
0.112
0.167
0.094
0.421
0.843
1.261
0.804
0.482
1.209
0.970
1.347
1.269
0.416
0.232
0.209
0.283
0.201
0.245
0.167
0.371
0.213
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich (veinlet?)
Inclusion-rich (veinlet?)
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich (mixed with veinlet)
Massive detrital, mixed with As-Ni veinlet
Masive detrital, mixed with AsNi veinlet
Masive detrital
Masive detrital
Masive detrital
Masive detrital
Masive detrital
Masive detrital
Masive detrital
Masive detrital
Overgrowth around inclusion-rich
All values referred to VCDT.
9.2. Distinguishing primary and secondary gold in CLP

High Au contents have been previously measured in whole-grain
analyses of concretionary pyrite, a category that also includes CLP
(Koppel and Saager, 1974; Utter, 1978; Barton and Hallbauer, 1996).
Gold in these grains is regarded as: 1) secondary and due to inltration
of Au-bearing metamorphic uids (Saager, 1970; Koppel and Saager,
1974); or 2) primary and due to detrital grains deriving from altered
granites) or meta-mac rocks (Utter, 1978; Hallbauer and Barton,
1987; Robb et al., 1990); or 3) primary and due to adsorption from
sea water during sedimentary pyrite growth in an anoxic environment
(Large et al., 2013).
In our samples, gold appears to be secondary in some of the grains
observed (in samples VCR_E_Drie, VCRa, VCRb, VCRc), where it occurs
as electrum grains lling cracks in pyrite (Fig. 3G). However, in other
samples (01, 02, 13, 16), Au is predominantly invisible, and concentrations up to several tens of ppm were measured by LA-ICP-MS, although
no electrum inclusions were observed under the microscope, or by
time-resolved laser ablation spectra (this study; Agangi et al., 2013).
These two occurrences of gold suggest two stages of gold incorporation:
a rst stage in which Au was incorporated during pyrite growth as ne
detrital particles or precipitated from solution, and a second stage in
which Au was remobilised under hydrothermal conditions. In trace element maps, texturally secondary gold is has high Au/Ag ratio (~ 10),
whereas invisible Au has with low Au/Ag (~0.1), and is broadly correlated with Sb, Tl, and Mo.
In some concentrically laminated grains, domains affected by U
inltration (cracks and most inclusion-rich domains) tend to have
high Au/Ag (~ 10) inclusions, whereas apparently less altered domains tend to have low Au/Ag (~ 0.1) (Fig. 8). This seems compatible
with studies that pointed out an increase of gold particle neness
(Au/(Au + Ag) 1000) with increasing metamorphism and alteration. Koppel and Saager (1974) and Saager (1970) indicated that
primary allogenic pyrite-hosted gold inclusions from the VCR are
whiter, more reective and Ag-rich in comparison with secondary
gold particles in the matrix of conglomerates, thus providing a criterion to distinguish metamorphic gold from primary inclusions.
Henckel and Schweitzer (1994) reported that, in the VCR, gold particles
from areas with strong alteration has lower Ag content (gold neness is
mostly 900) than that from areas less affected by alteration and base
metal sulde deposition (Au neness mostly 750900). Oberthr and
Saager (1986) measured gold neness largely N 900 in secondary gold
from the Carbon Leader Reef. However, a possible exception is represented by grain 02_01, which has lower Au/Ag in areas affected by
42
Concentrically
laminated
grain 02_02
grain 01_01
analyses
0
-12
-10
-8
-6
-4
-2
10
12
Inclusion-rich
grain 01_02
grain 01_03
grain 01_04
grain 02_05
analyses
-6
-4
-2
Massive
grain 02_05
grain 02_03
grain 02_06a
grain 02_02b
grain 01_01 (core)
grain 02_07
analyses
0
-6
-4
-2
34
S
Fig. 9. Frequency histograms of 34S for different textural types of pyrite in the Ventersdorp Contact Reef. Continuous line indicates probability density curve.
metamorphic uid circulation, and gold inclusions also Ag-rich are

associated with these areas. Frimmel and Gartz (1997) have found
wide compositional variations in gold grains from the VCR, despite the
rapid homogenisation expected under hydrothermal conditions. They
proposed that this variability reected detrital grain compositions,
and that intergranular compositional homogenisation was dependent
on local hydrothermal conditions, namely the uid-to-rock ratio.
Gold might either have been originally included in pyrite as sub-mscale particles transported in suspension, or have been precipitated
from solution. In the second scenario, the Au/Ag observed in pyrite
would be controlled by the solubility of the two metals in water, and
by the process of exsolution and deposition. Some authors (Reimer,
1984; Horscroft et al., 2011) proposed that gold was transported in solution or as colloid stabilised by S-complexes or humic acids, and was
precipitated by an increase in oxygen produced by microbial communities in localised environments.
9.3. Distinguishing primary vs. post-sedimentary isotope signatures

A uniquely hydrothermal S source of pyrite is ruled out by evidence
of MIF-S, a process that is largely thought to reect an atmospheric
source (e.g. Farquhar et al., 2000; Farquhar and Wing, 2003; Shen
et al., 2009; although see Watanabe et al., 2009), although hydrothermal remobilisation of sedimentary S could be a possibility. However,
large intergranular variations of S and Fe isotopic values in massive pyrite, inclusion-rich pyrite and CLP (Figs. 9 and 14) are consistent with a
separate origin and subsequent juxtaposition of pyrite grains, as expected during transport in a sedimentary environment. A detrital origin is
also supported by preservation of intragranular isotopic and textural
characteristics, despite burial diagenesis and metamorphism. For example, inclusion-rich grain 01_02 has 56Fe N 2, and is cross-cut by a
veinlet of trace element-poor massive pyrite (Table 3) with
56Fe b 1. This veinlet is truncated by the abraded margin of the
43
Fig. 10. Isotope traverses (56Fe and 34S) across a concentrically laminated pyrite grain with a massive core (sample 01). Uncertainty of 56Fe indicated as 2 error bars, uncertainty of 34S
less than symbol size.
grain, and must pre-date reworking. These characteristics are inconsistent with the hypothesis of in situ suldation of detrital Fe oxide grains
(e.g. Law and Phillips, 2005). In contrast, uniform 34S values have been
reported from concentrically laminated pyrite derived from replacement
of chamositic precursor ooids, indicative of constant composition of mineralizing uids during the replacement (Schieber and Riciputi, 2004).
9.4. Two distinct S pools carrying MIF represented in VCR pyrite
Concentrically laminated and inclusion-rich pyrite grains have a
comparable mineral inclusion assemblage and similar trace element
compositions, including high values of Sb, Tl, Mo, Mn, Cu, Au, and Ag
(this study; Agangi et al., 2013; Large et al., 2013). Secondly, the preservation of these non-massive grains, which were loose and partly plastic,
suggests that they did not undergo long transport in a high-energy
environment such as the quarzt and chert pebbles in the conglomerate
host. All this evidence implies that CLP and inclusion-rich pyrite formed in
a similar environment, and they formed near the site of nal deposition.
Despite these similarities, isotopic data of CLP and inclusion-rich pyrite presented in this study indicate that S from pyrite in the VCR was
sourced from two distinct reservoirs carrying MIF-S signal. One reservoir,
forming inclusion-rich pyrite, carried 33S N 0, the second, forming CLP,
had 33S b 0. MIF-S has been found in the Archaean and
Paleoproterozoic (Farquhar et al., 2000), and in sulfate from atmospheric
volcanic plumes (Baroni et al., 2008). Experiments have shown MIF-S
during UV irradiation of SO2 gas (Farquhar et al., 2001) through
photolysis, a chemical reaction that involves breaking of molecules via

photon bombardment. In one of these experiments, short-wavelength
(193 nm) irradiation resulted in elemental S (S0) with 33S N 0, and
with 33S b 0. In nature, such MIF can be produced by photolysis
SO2
4
of atmospheric SO2, or SO3 (Pavlov and Kasting, 2002) by UV radiation
under conditions of low atmospheric O2 and ozone concentrations. It is
largely believed that S in the rock record carrying 33S b 0 was produced
33
as SO2
4 and S with S N 0 was produced as elemental S via photolytic
process (Farquhar et al., 2001; Philippot et al., 2007; Ono et al., 2009). In
the following discussion, we also adopt this interpretation, although this
will need further testing.
The range of 34S measured in CLP ( 12 to + 12) is relatively
wide for the Archaean (Farquhar et al., 2011), although similar or
even wider ranges have been reported for rocks dated at around
2.7 Ga (Goodwin et al., 1976; Grassineau et al., 2001), and locally as
early as 3.5 Ga (Shen et al., 2001; Philippot et al., 2007; Ueno et al.,
2008). The distribution in the 34S vs. 33S plot (Fig. 12), with wide
34S variations for a small 33S range and broad scatter around the
Archaean array (Ono et al., 2003, 2009), cannot be explained by photolytic processes alone, and indicates strong mass-dependent S isotope
fractionation.
Two features are evident from the 34S vs. 33S plot: 1) analyses of
CLP are distributed approximately symmetrically around the Archaean
array with a 34S range of ~25; and 2) analyses of inclusion-rich pyrite
plot either near or at lower 34S values compared with this array, with a
total 34S range of ~ 7 (Fig. 12). Pyrite formation by sulfate and
44
4.0
3.0
4
34S
1.0
0.0
-1.0
-2.0
III
IV
500
1000
1500
2.0
-2
II
III
IV
500
massive
56Fe
2.0
1500
1000
56Fe
grain 02_02
1.0
d
a
c
f
i
0.0
V
II
-1.0
1.0
III
IV
200
400
600
e
j
800
56Fe
IV III II I
0.5
massive
pyrite
0.0
Fe isotope traverse
S isotope traverse
1 mm
-0.5
-1.0
-1.5
-2.0
0
2.0
II
200
III
400
IV
600
800
1000
1200
56Fe
1.5
1.0
0.5
0.0
I
II
III
IV
-0.5
0
200
400
600
800
1000
1200
distance, m
Fig. 11. Isotope traverses (56Fe and 34S) across a concentrically laminated pyrite grain (sample 02). The grain has been schematically subdivided into ve concentric rims (I to V). Uncertainty indicated as 2 error bars.
elemental S reduction would preferentially incorporate 32S, thus offering an explanation for the 34S values lower than the Archaean MIF-S
array (Ono et al., 2003, 2009). Large deviations from the photolytic
array have been explained by microbial sulfur disproportionation, and
by microbial sulfate reduction (Shen et al., 2001; Kamber and
Whitehouse, 2007; Philippot et al., 2007; Ueno et al., 2009). Under
low-sulfate conditions in sea water (b 200 M/l) expected for most of
the Archaean, the fractionation factor between sulfate and precipitating
pyrite would be strongly dependent on sulfate concentrations, and
would be limited to between a few and ~ 20 34S (Habicht et al.,
2002). These values could explain the lowest 34S values ( 12.16)
measured in our CLP samples, assuming an initial sulfate composition
along the Archaean array.
However, some analyses of CLP have higher 34S than the photolytic
array, and plot in the eld typically occupied by Archaean barite
(33S b 0 and 34S N 0; e.g. Ono et al., 2003; Ueno et al., 2008; Roerdink
et al., 2012). These compositions can be explained by a reservoir effect
in a closed system. In a closed system (e.g. a water body isolated from
the sea with limited S available), formation of pyrite with low 34S by sulfate reduction would progressively increase the 34S values of the residual sulfate. If precipitation proceeded until complete consumption of the
available S, the isotopic composition of pyrite would have evolved towards progressively increasing values of 34S. This process can be
modelled by Rayleigh fractionation according to the equation:
34
34
SPy Saq PySaq 1 ln f ;

where (Py-Saq) is the fractionation factor between aqueous S and
resulting sulde, and f expresses the degree of fractionation. This
model indicates that, assuming a starting composition along the
1.0
Archean
array
uncert.
0.5
33S
0.0
-0.5
-1.0
-1.5
-15
affected by
alteration/
recrystallisation
-10
-5
10
15
34S
6
5
4
56Fe
3
2
1
concentrically laminated
inclusion-rich
massive
VCR whole grains
uncert.
affected by
alteration/
recrystallisation
0
-1
-2
-1.5
-1.0
-0.5
0.0
0.5
1.0
S
6
5
4
56Fe
3
2
1
0
affected by
alteration/
recrystallisation
-1
-2
-15
-10
-5
10
15
34S
Fig. 12. Plots of Fe and S isotopic compositions (45Fe, 34S and 33S) of pyrite grains from
the VCR (samples 01 and 02). Uncertainty: 0.2 for 56Fe, 0.1 for 33S; lesser than symbol size for 33S (0.28) and 34S (0.4). Whole-grain analyses from Hofmann et al.
(2009). Grey areas: mass-dependent S isotopic fractionation (33S = 0 0.2) and
56Fe of pyrite from hydrothermal and igneous sources (0.5 to 0.1).
Archaean array (33S = 1 and 34S = 1), and fractionating pyrite with the lowest 34S of ~ 12 ((Py-Saq) = 11), pyrite with the
highest 34S measured (+12.6) can be produced by depositing 80
90% of S available in the uid (Fig. 16). An alternative origin for the
high 34S values measured is precipitation from sea water sulfate
(34S ~ 515, as constrained based on barite compositions; Ono et al.,
2003) by either biogenic or non-biogenic mechanisms.
9.5. Two pools of Fe: partial oxidation of Fe2+ via APIO or photosynthetic
O2 production
Negative 56Fe values measured in this study as low as 1.5 largely overlap with previous studies of pyrite in the Archaean (Johnson
45
et al., 2003; Dauphas et al., 2004; Rouxel et al., 2005; Yamaguchi et al.,
2005; Dauphas and Rouxel, 2006). Overall, the average 56Fe for our
analyses are in general agreement with previous analyses of whole
grains or grain fragments from the VCR and other reefs of the Witwatersrand Basin, which have 56Fe = 0.74 to + 1.21 (Hofmann
et al., 2009).
However, some grains, inclusion-rich in particular, have average
56Fe values exceeding the range measured for whole-grains (Fig. 12),
and some of the positive 56Fe values obtained from the VCR far exceed
any previous measurements for Archaean pyrite. Thus, the following
features require explanation: 1) inclusion-rich pyrite has higher average
56Fe (+ 2.75) in comparison with CLP (+ 0.2), thus together
forming a bimodal distribution of 56Fe; 2) each pyrite type has wide
Fe isotope variations; and 3) these two types of pyrite have an asymmetric and complementary distribution in a histogram of 56Fe values
(Fig. 14).
The bimodal Fe isotope compositions of CLP and inclusion-rich pyrite suggest precipitation from separate Fe pools. It is possible that
some of the isotopically heavier pyrite was formed by suldation of
Fe3 + species (e.g. 56Fe + 2.2 was measured in magnetite from
the 3.7 Ga BIF from the Isua Greenstone Belt; Whitehouse and Fedo,
2007 and 56Fe = + 0.9 to 5.2 was measured by SIMS in magnetite
from the 2.0 Ga old Gunint chert, Marin-Carbonne et al., 2011). Assuming ambient water contained Fe2+aq with 56Fe = 0 (although see
Beard et al., 2003b), partial oxidation of Fe would result in Fe3+ with
56Fe +3 (Welch et al., 2003) (Supplementary Fig. 3). This fractionation mechanism would have been able to create two separate Fe pools
(a Fe2 +aq pool with 56Fe ~ 0 or slightly negative, and a precipitated
Fe3 + pool with higher 56Fe) from which concentrically laminated
and inclusion-rich pyrite, respectively, may have formed. Modelling of
this oxidation as a steady-state process suggests that the creation of a
Fe pool with 56Fe = + 2.75 (the average measured for inclusionrich pyrite) would require mild oxidation (b 10%) of the Fe2+ present
in the water (Supplementary Fig. 3). Partial oxidation of Fe2 + may
imply the presence of free oxygen in surface water at 2.7 Ga. The idea
that some oxygen was present in the oceans and the atmosphere well
before the great oxidation event due to oxygenic photosynthesis
(Kasting, 1993), gains increasing acceptance (e.g. Anbar et al., 2007;
Nishizawa et al., 2010; Crowe et al., 2013; Planavsky et al., 2014). However, Fe oxidation may have occurred via anaerobic photosynthetic Fe
oxidation (APIO) by purple and green bacteria, a mechanism that does
not require free oxygen and produces similar Fe isotope fractionation
(Croal et al., 2004; Johnson et al., 2004). In both these cases, biological
activity would be required.
During sulde formation from each of these pools, further Fe isotope
fractionation would have occurred, causing the wide variations of 56Fe
values observed for both concentrically laminated and inclusion-rich
pyrite. The large variation of Fe isotope values suggests pyrite formation
in a closed system, in which case the nal average Fe isotope composition of pyrite would progressively tend towards the initial composition
of the initial Fe pool.
In the case of inclusion-rich pyrite, the lowest 56Fe measured
(+ 0.4, if one outlying analysis at 1.6 not considered at this
stage) requires formation of pyrite depleted by 2.3 in comparison
with the average value measured for this type of pyrite (56Fe =
2.7). Large fractionation of Fe isotopes with strongly negative values
(e.g. pyrite with 56Fe as low as 2.7; Archer and Vance, 2006) has
been interpreted as evidence for dissimilatory iron reduction (DIR), a
biologically-mediated process capable of producing Fe2+aq depleted in
56Fe by up to 2.0 in comparison with Fe3+aq. After Fe reduction, combination of Fe2 +aq with H2S would result in formation of FeS
(mackinawite, a common pyrite precursor) depleted in 56Fe by a further 0.3 to 0.9 (Butler et al., 2005; Guilbaud et al., 2011). This combined fractionation factor ( 2.3 to 2.9) would be able to explain
the lowest values measured in inclusion-rich pyrite. In a Fe-limited system, this would cause increase of heavy Fe isotopes in the residual Fe3+
46
BSE
20
19
compact
pyrite
33
33
S=
-0.24
33
S=
0.24
18
S=
-0.42
mapped area
grain 02_05
500 m
As
Ni
20
20
As-Ni veinlet
19
19
As-Ni overgrowth
18
18
Fig. 13. BSE image and X-ray compositional maps of round inclusion-rich pyrite rimmed by massive euhedral pyrite from the Ventersdorp Contact Reef. Circles indicate locations of in situ
isotope SIMS analyses. Analyses n. 18 and 20 sampled AsNi-rich secondary pyrite with 33S b 0. Sample 02.
reservoir, and consequently in the late forming pyrite. This process can
be modelled by Rayleigh fractionation, according to the equation:
56
56
FePy Feaq PyFeaq 1 ln f :

Starting from an initial composition of 56Fe = +2.7 and using a
fractionation factor (Py-Feaq) of 2.3, values of +45 for precipitating pyrite can be reached by fractionation of 8090% of Fe in the
reservoir.
In the case of CLP, precipitation of pyrite depleted by ~1.7 with respect to the average 56Fe values of this type of pyrite (+0.2) would
be necessary to explain the lowest measured compositions (56Fe =
1.5). Precipitation of pyrite without Fe3+ reduction is expected to
cause small Fe isotope fractionation (e.g. up to 0.9; Butler et al.,
2005). However, Guilbaud et al. (2011) have measured wide fractionations (1.73.0) between FeS mackinawite, which was reacted
(abiotically) with H2S, and the resulting pyrite. Fractionation modelled
by Rayleigh distillation for CLP using a starting composition of +0.2
and a fractionation factor of 1.7 would also explain the highest
values measured (22.5, if an outlying value at ca. 3 is not
considered) for 8090% fractionation. This type of fractionation is consistent with the increasing average 56Fe measured in core-to-rim traverses in sample 02 (Supplementary Fig. 2). Some of the outlying 56Fe
values measured (i.e. one analysis with 56Fe N 3 in CLP and one with
56Fe = 1.6 in inclusion-rich pyrite) may be explained by local

mixing of the two Fe pools, so that isotopically heavy Fe3+ may have locally contributed to CLP and Fe2+ to inclusion-rich pyrite.
In addition to the general S and Fe isotope trends discussed here,
intragranular variations have been observed, which result in smallscale heterogeneities, and locally apparently contrasting trends (e.g. Fe
isotope traverses Figs. 10 and 11). Small scale heterogeneities have
been interpreted as an indication of biogenic processes, given the significant variability in space and time of these metabolic processes and the
isotopic fractionation associated (e.g. Caneld, 2001; Wacey et al.,
2010). For example, experiments of Fe oxide reduction by microbial dissimilatory reduction of amorphous Fe3 +Si co-precipitate (PercakDennett et al., 2011) have produced ne-grained, poorly-crystalline
Fe2+ solid phases in intimate contact with residual Fe3+, leading the authors to interpret the small-scale heterogeneity of Fe isotope values as
distinctive of this biogenic process of Fe reduction. Thus, the smallscale wide scatter of S and Fe isotope data we observed is consistent
with, although not unambiguously related to, origin via metabolic processes.
9.6. A model of CLP formation and Au accumulation for the VCR
Sulfur and Fe isotope analyses indicate that CLP and inclusion-rich
pyrite in our samples originated from two separate and complementary
S and Fe pools. Sulfate with 33S b 0 and elemental S with 33S N 0,
47
Table 3
Pyrite iron isotope analyses (SIMS).
Sample
Grain
Analysis ID
56Fe,
Error
Texture
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_02
VCR01_04
VCR01_04
VCR01_04
VCR01_06
VCR01_06
VCR01_06
VCR01_06
VCR01_06
VCR01_06
VCR01_06
VCR01_06
VCR01_06
VCR01_06
VCR01_06
VCR01_06
VCR01_06
VCR01_06
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
Fe_01_01
Fe_01_02
Fe_01_03
Fe_01_04
Fe_01_05
Fe_01_06
Fe_01_07
Fe_01_10
Fe_01_08
Fe_01_09
Fe_01_11
Fe_01_12
Fe_01_13
Fe_01_14
Fe_01_15
Fe_01_16
Fe_01_57
Fe_01_58
Fe_01_59
Fe_01_10_b
Fe_01_14_b
Fe_01_17_b
Fe_01_18_b
Fe_01_19_b
Fe_01_20_b
Fe_01_40
Fe_01_41
Fe_01_42
Fe_01_43
Fe_01_44
Fe_01_45
Fe_01_46
Fe_01_47
Fe_01_48
Fe_01_49
Fe_01_50
Fe_01_51
Fe_01_52
Fe_01_53
Fe_01_54
Fe_01_55
Fe_01_56
Fe_01_20
Fe_01_21
Fe_01_22
Fe_01_23
Fe_01_24
Fe_01_25
Fe_01_26
Fe_01_27
Fe_01_28
Fe_01_29
Fe_01_30
Fe_01_31
Fe_01_32
Fe_01_33
Fe_01_03_b
Fe_01_04_b
Fe_01_06_b
Fe_02_07
Fe_02_08
Fe_02_09
Fe_02_10
Fe_02_11
Fe_02_12
Fe_02_13
Fe_02_14
Fe_02_15
Fe_02_16
Fe_02_17
Fe_02_18
Fe_02_19
Fe_02_20
Fe_02_21
0.646
0.461
0.400
0.528
0.471
0.843
0.696
0.459
2.998
2.786
2.824
2.906
3.103
3.749
3.565
4.113
3.955
3.894
4.074
0.93
1.88
1.57
3.96
3.51
3.28
0.896
1.182
0.886
3.776
2.558
5.044
2.842
2.389
2.133
2.653
2.018
3.863
2.892
2.000
2.739
3.458
2.197
0.585
0.523
0.352
0.418
0.989
0.727
0.803
0.373
0.369
0.018
0.215
0.234
0.529
0.165
0.39
0.8
0.89
3.32
2.13
1.57
1.91
1.35
1.08
0.76
0.34
0.79
1.08
1.22
0.59
0.06
1.68
0.03
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.19
0.19
0.19
0.19
0.19
0.19
0.16
0.16
0.16
0.16
0.16
0.17
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.17
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.19
0.19
0.19
0.23
0.22
0.23
0.23
0.24
0.23
0.23
0.23
0.22
0.23
0.24
0.22
0.23
0.23
0.22
Vein
Vein
Vein
Vein
Vein
Vein
Vein
Vein
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
Inclusion-rich
(continued on next page)
48
Table 3 (continued)
Sample
Grain
Analysis ID
56Fe,
Error
Texture
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
01
01
01
01
01
01
01
01
01
01
01
01
01
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR02_02
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_01
VCR01_05
VCR01_05
VCR01_05
VCR01_05
VCR02_03
VCR02_03
VCR02_03
VCR02_03
VCR02_03
VCR02_06_c
VCR02_06_b
VCR02_06_c
VCR02_06_d
VCR02_06_b
VCR02_06_b
VCR02_02
VCR02_04
VCR02_04
VCR02_04
VCR02_04
Fe_02_22
Fe_02_23
Fe_02_24
Fe_02_25
Fe_02_26
Fe_02_27
Fe_02_28
Fe_02_29
Fe_02_30
Fe_02_31
Fe_02_32
Fe_02_33
Fe_02_34
Fe_02_44
Fe_02_45
Fe_02_46
Fe_02_47
Fe_02_48
Fe_01_34
Fe_01_35
Fe_01_36
Fe_01_07_b
Fe_01_08_b
Fe_01_09_b
Fe_01_17
Fe_01_18
Fe_01_19
Fe_01_37
Fe_01_38
Fe_01_39
Fe_01_01_b
Fe_02_02
Fe_02_03
Fe_02_04
Fe_02_05
Fe_02_06
Fe_02_35
Fe_02_36
Fe_02_38
Fe_02_39
Fe_02_42
Fe_02_43
Fe_02_49
Fe_02_53
Fe_02_54
Fe_02_55
Fe_02_56
0.66
0.23
1.65
1.5
0.43
1.23
0.61
0.25
0.88
0.52
0.45
0.37
0.21
0.13
0.55
1.45
0.82
0.37
1.221
1.223
1.468
1.981
0.509
1.009
1.580
1.603
1.997
0.178
0.641
0.740
3.74
3.22
2.77
2.67
1.21
2.28
0.24
1.58
0.5
0.37
0.34
2.59
0.15
1.12
1.23
1.13
1.17
0.22
0.22
0.22
0.23
0.23
0.23
0.23
0.23
0.22
0.23
0.23
0.23
0.25
0.22
0.22
0.22
0.22
0.23
0.16
0.16
0.16
0.19
0.19
0.19
0.16
0.16
0.16
0.16
0.16
0.16
0.19
0.23
0.23
0.23
0.22
0.24
0.23
0.23
0.23
0.23
0.23
0.24
0.23
0.23
0.23
0.23
0.23
Massive detrital
Massive detrital
Massive detrital
Massive detrital
Massive detrital
Massive detrital
Massive detrital
Massive detrital
Massive detrital
Massive
Massive
Massive
Massive
Massive detrital
Massive detrital
Massive detrital
Massive detrital mixed with overgrowth
Massive detrital
Massive detrital?
Massive detrital?
Massive detrital?
Massive detrital?
Massive detrital
Massive detrital
Massive detrital?
Massive detrital
Massive detrital
Massive detrital
Massive detrital
All values referred to IRMM-014.
produced by photolysis of volcanic gases by UV radiation are the main

candidates to explain these S isotopic values. Iron isotopes suggest precipitation of pyrite from Fe2+ and Fe3+ pools, which may have formed
by partial oxidation of Fe2+aq. This process may have been directly or indirectly biologically-induced (through APIO or photosynthetic O2 production). Sulfate and Fe2+ are highly soluble in water, and would be
taken up by surface waters. Given low oxygen levels in the Archaean atmosphere and the consequent minor role of oxidative weathering, this
sulfate would have been predominant in surface sea water and continental fresh water (Habicht et al., 2002; Caneld, 2005), and would
have contributed signicantly to the S isotopic signal. Elemental S and
Fe3+ species (e.g. Fe(OH)3) are insoluble, and will tend to settle through
the water column and be preferentially incorporated in the sediment
(Fig. 15). Abundant S in volcanic gases would have been available for
UV photolysis, as a result of active mac volcanism during the deposition of the VCR (Armstrong et al., 1991). We envisage formation of
CLP by precipitation of pyrite from sulfate and Fe2+ from the water at,
or near, the watersediment interface and formation of inclusion-rich
pyrite from S0 and Fe3+ in the sediment.
The environment in which these pyrite grains formed was relatively
low-energy, and sedimentation was dominated by a clay- to silt-sized
terrigenous component, in the presence of an organic fraction. Fine-
grained sediment may have been deposited in ephemeral ponds concomitant with the early diagenetic growth of pyrite, having been eroded
during the increase in ambient energy responsible for the deposition of
the conglomerates, and might have left no sign of their existence in the
rock record, the only evidence being preserved as inclusions in pyrite
grains. Mudstone facies have only rarely been reported from the VCR
(Minter, 1982), consistent with the degradational nature of the VCR
forming a pediment mantling actively eroding palaeotopography
(McCarthy, 1994). Nevertheless, black shale clasts have been reported
to occur in VCR conglomerates (Krapez, 1985). These ponds, such as
abandoned channels or ood plains between river channels, developed
in a uvial braid-plain environment on a degradational surface (Krapez,
1985), and would have been the site where microbial mats developed.
Redox-sensitive elements (especially Ni, As, Zn, Mo, Mn) were concentrated during sedimentation, and then incorporated during growth of
diagenetic pyrite.
9.7. Role of biological processes in Au mineralisation in the
Witwatersrand Basin
The occurrence of nely-dispersed gold of likely primary origin in
biogenically-precipitated pyrite gives support to the hypothesis that a
Concentrically
laminated
49
grain 02_02
grain 01_01
analyses
7
6
5
4
3
2
1
0
-2
-1
9
8
grain 01_02
grain 01_03
grain 01_04 (recrystallised?)
grain 01_02 early vein
Inclusion-rich
analyses
6
5
4
hydrothermal and
magmatic pyrite
3
2
1
0
-2
-1
analyses
grain 02_03
grain 02_04
grain 01_01 (core)
grain 02_06b
grain 02_06c
grain 02_06d
grain 02_02b
grain 01_05
Massive
7
6
2.5-2.7 Ga black shales pyrite
5
4
3
2
BIF pyrite
Wits
1
0
-2
-1
56Fe
Fig. 14. Frequency histograms of 56Fe for different types of pyrite in the Ventersdorp Contact Reef. Data compared with 2.5 Ga Transvaal BIF pyrite (Johnson et al., 2003), 2.52.7 Ga black
shales pyrite (Rouxel et al., 2005) and Witwatersrand pyrite (indicated as Wits; Hofmann et al., 2009). Continuous line indicates probability density curve (early vein in grain 01_02 excluded from calculation). Grey eld denotes pyrite from hydrothermal and igneous sources (56Fe = 0.5 to 0.1; Rouxel et al., 2005).
link exists between biological activity and Au deposition in the Witwatersrand Basin (e.g. Horscroft et al., 2011). Given the evidence for hydrocarbon formation from the maturation of organic matter in the Wits
Basin and subsequent mobilisation (Drennan and Robb, 2006), the origin of carbonaceous matter detected in some of the pyrite grains cannot
be determined beyond doubt, as both primary and secondary origins
can be envisaged. However, its nely-dispersed nature, and the association with a suite of trace elements typical of carbonaceous sediments
seem permissive of a primary origin. If primary, carbonaceous matter
may represent the remnants of microbes, such as sulfate-reducing bacteria, or may have been present in the sediment and accidentally
trapped by crystallising pyrite. Both microbial communities and degraded carbonaceous matter present at the moment of pyrite formation
could have interacted with Au in different ways. It is known that some
bacteria can process Au as part of their life cycle. Experimental studies
have shown that microorganisms can mediate gold solubilisation by excretion of thiosulde, amino acids and cyanide, and are able to
precipitate gold intracellularly and extracellularly and as products of

their metabolism (Lengke and Southam, 2007; Reith et al., 2007). Further, degraded carbonaceous matter can produce compounds, such as
humic acids, that can form complexes with Au (Gray et al., 1998; Reith
et al., 2007), and adsorption onto organic matter, as well as clay minerals, is a known mechanism of Au accumulation (Greenwood et al.,
2013). It has also been pointed out that microbial mats, such as the
ones believed to have been present in the Witwatersrand Basin, may
have acted as physical traps for ne Au particles transported in the
suspended load (Noffke et al., 2006; Horscroft et al., 2011).
10. Summary/conclusions
1. Concentrically laminated pyrite (CLP) in the Ventersdorp Contact
Reef was formed by successive accretion of laminae, in some cases
around a pre-existing core (Fig. 3D), forming a soft precursor to the
grains observed in our samples. CLP formed in a depositional
50
UV solar radiation
SO
primary
Au deposits
SO
S
S<0 S>0
concentrically
laminated Py
SO
Fe +aq
+
Fe+
Fe
S
erosion surface
Fe+
inclusion-rich Py
Sedimentation area
fine grained terrigenous sediment
containing organic matter
Sediment source area

granitoids,
greenstone belts
Fig. 15. A possible model of CLP formation. Photodissociation of SO2 gas (of, e.g., volcanic origin) would produce SO2
with 33S b 0, and S0 with 33S N 0. Given their different solubility,
4
the two S species could have been segregated in water, thus forming separate S reservoirs. Sulfate would have been dissolved imparting negative MIF signal to the water. Concentrically
laminated pyrite formed from this reservoir at, or near, the watersediment interface and scavenging sediment containing carbonaceous matter and gold particles. Elemental sulfur would
have sunk and be included in the sediment. Inclusion-rich pyrite crystallised from this latter reservoir.
environment dominated by deposition of clay- to silt-sized clastic

sediment and, during growth, it scavenged some detritus (monazite,
apatite, quartz, clay minerals now metamorphosed to chlorite and
muscovite and possibly Ti oxides), and carbonaceous matter.
2. CLP and inclusion-rich pyrite are interpreted to have formed in the
same environment, near the site of nal deposition. This interpretation is based on: 1) poor preservation potential inferred from the
1.5
3.
4.
concentrically
laminated
inclusion-rich
MIF-S
1.0
MDF-SR
uncert.
33S
0.5
5.
0.0
-0.5
MDF-SR
-1.0
0.2
0.4
0.6
0.8
Archean array
-1.5
seawater sulfate
MIF-SO
-2.0
-20
-15
-10
-5
10
15
20
6.
34S
Fig. 16. Interpretation of S isotope fractionation mechanisms in inclusion-rich and concentrically laminated pyrite in the VCR (analyses affected by secondary pyrite are not included). Black dashed line indicates the Archaean array of photolysis-induced fractionation of
S isotopes (Ono et al., 2003). Black arrows indicate S isotope variations in pyrite by S reduction reactions. Blue line indicates S isotope variations in sulfate induced by Rayleigh
fractionation of pyrite with the lowest 34S (approximately 12) starting from a composition along the Archaean array; f: fraction of S precipitated from solution. Approximate
position of seawater sulfate composition eld (Ono et al., 2003) indicated by blue dashed
line. (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)
7.
delicate textures; 2) similar mineral inclusions; and 3) similar trace

element composition (high Sb, Tl, Mo, Cu, Au, Ag and Mn in comparison with other types of pyrite).
In CLP, the 208Pb/206Pb ratio and the distribution of U indicate areas of
the grains that were affected by circulation of U-rich (metamorphic)
uids. Mapping these areas allowed distinguishing altered portions
of the grains from areas more likely to have pristine composition.
CLP contains invisible gold and Au inclusions. Domains of grains containing invisible gold are characterised by low Au/Ag (around 0.1),
and may represent pristine portions of CLP grains, where Au was
trapped during pyrite growth. This gold was derived from ambient
uid, by precipitation from solution, or from ne particles in suspension. Electrum inclusions, mostly having high Au/Ag (around 10),
tend to occur along fractures and along grain margins and are
interpreted as being of secondary (hydrothermal) origin.
CLP and inclusion-rich pyrite were formed by separate pools of S and
Fe: 33S b 0 and relatively low 56Fe (average 0.2) for CLP, and
33S N 0 and high 56Fe (average 2.7) for inclusion-rich pyrite. Formation of these S pools is attributed to photolysis, and formation of
Fe pools is attributed to partial oxidation of Fe2 + aqueous to Fe3+
species. Separation of S and Fe pools may have occurred in water,
by segregation of soluble sulfate and Fe2 + from insoluble S0 and
Fe3+; the latter would have sunk and been included in the sediment.
Microbial activity may have caused sulfate reduction to form CLP (S
reducers), Fe2+ oxidation to form Fe3+ (O2-producing cyanobacteria
or Fe-oxidising green and purple bacteria), and dissimilatory Fe3+
reduction to form inclusion-rich pyrite. Traces of carbonaceous matter found in CLP may be the remnants of these microbes.
Concurrent biogenically-mediated pyrite formation and Au trapping
suggest that microbial activity was responsible for the accumulation
of Au and other trace elements (e.g. Sb, Mn, Ag, Tl, Cu, Mo, Mn) which
are commonly enriched in organic matter-rich sediments.
Supplementary data to this article can be found online at http://dx.

doi.org/10.1016/j.earscirev.2014.10.009.
Acknowledgments
This research was supported by the National Research Foundation of
South Africa via a Post-Doctoral Fellowship to AA., a research grant to
AH and by the DST-NRF Centre of Excellence for Integrated Mineral
and Energy Resource Analysis (CIMERA). We thank Katarien Deysel
(AngloGold Ashanti) and Bradley Guy for providing sample material.
Hartwig Frimmel is thanked for the constructive review of the
manuscript.
References
Agangi, A., Hofmann, A., Wohlgemuth-Ueberwasser, C.C., 2013. Pyrite zoning as a record
of mineralization in the Ventersdorp Contact Reef, Witwatersrand Basin, South Africa.
Econ. Geol. 108, 12431272.
Anbar, A.D., et al., 2007. A whiff of oxygen before the great oxidation event? Science 317,
19031906.
Archer, C., Vance, D., 2006. Coupled Fe and S isotope evidence for Archaean microbial
Fe(III) and sulfate reduction. Geology 34, 153156.
Armstrong, R.A., Compston, W., de Wit, M.J., Williams, I.S., 1990. The stratigraphy of the 3.
53.2 Ga Barberton Greenstone Belt revisited: a single zircon ion microprobe study.
Earth Planet. Sci. Lett. 101, 90106.
Armstrong, R.A., Compston, W., Retief, E.A., Williams, I.S., Welke, H.J., 1991. Zircon ion microprobe studies bearing on the age and evolution of the Witwatersrand triad.
Precambrian Res. 53, 243266.
Barnicoat, A.C., et al., 1997. Hydrothermal gold mineralization in the Witwatersrand
basin. Nature 386, 820824.
Baroni, M., Savarino, J., Cole-Dai, J., Rai, V.K., Thiemens, M.H., 2008. Anomalous sulfur
isotope compositions of volcanic sulfate over the last millennium in Antarctic ice
cores. J. Geophys. Res. 113, D20112.
Barton, E.S., Hallbauer, D.K., 1996. Trace-element and UPb isotope compositions of pyrite
types in the Proterozoic Black Reef, Transvaal Sequence, South Africa: implications on
genesis and age. Chem. Geol. 133, 173199.
Beard, B.L., et al., 2003a. Application of Fe isotopes to tracing the geochemical and biological cycling of Fe. Chem. Geol. 195, 87117.
Beard, B.L., Johnson, C.M., Von Damm, K.L., Poulson, R.L., 2003b. Iron isotope constraints on Fe cycling and mass balance in oxygenated Earth oceans. Geology
31, 629632.
Beukes, N.J., Cairncross, B., 1991. A lithostratigraphic-sedimentological reference prole for
the late Archaean Mozaan Group, Pongola Sequence; application to sequence
stratigraphy and correlation with the Witwatersrand Supergroup. S. Afr. J. Geol. 94,
4469.
Binda, P.L., Simpson, E.L., 1989. Petrography of sulphide-coated grains from the Ordovician Winnipeg Formation, Saskatchewan, Canada. Eur. J. Mineral. 1, 439453.
Binda, P.L., Koopman, H.T., Schwann, P.L., 1985. Sulphide ooids from the Proterozoic Siyeh
Formation of Alberta, Canada. Mineral. Deposita 20, 4349.
Boer, R.H., Reimold, W.U., 2006. Conditions of gold remobilization in the Ventersdorp
Contact Reef, Witwatersrand Basin, South Africa. Geol. Soc. Am. Spec. Pap. 405, 387402.
Butler, I.B., Archer, C., Vance, D., Oldroyd, A., Rickard, D., 2005. Fe isotope fractionation on
FeS formation in ambient aqueous solution. Earth Planet. Sci. Lett. 236, 430442.
Caneld, D.E., 2001. Biogeochemistry of sulfur isotopes. Rev. Mineral. Geochem. 43,
607636.
Caneld, D.E., 2005. The early history of atmospheric oxygen: homage to Robert M.
Garrels. Annu. Rev. Earth Planet. Sci. 33, 136.
Catuneanu, O., 2001. Flexural partitioning of the Late Archaean Witwatersrand foreland
system, South Africa. Sediment. Geol. 141142, 95112.
Coward, M.P., Spencer, R.M., Spencer, C.E., 1995. Development of the Witwatersrand
Basin, South Africa. Geol. Soc. Lond. Spec. Publ. 95, 243269.
Croal, L.R., Johnson, C.M., Beard, B.L., Newman, D.K., 2004. Iron isotope fractionation
by Fe(II)-oxidizing photoautotrophic bacteria. Geochim. Cosmochim. Acta 68,
12271242.
Crowe, S.A., et al., 2013. Atmospheric oxygenation three billion years ago. Nature 501,
535538.
Danyushevsky, L., et al., 2011. Routine quantitative multi-element analysis of sulphide
minerals by laser ablation ICP-MS: standard development and consideration of matrix effects. Geochem. Explor. Environ. Anal. 11, 5160.
Dauphas, N., Rouxel, O., 2006. Mass spectrometry and natural variations of iron isotopes.
Mass Spectrom. Rev. 25, 515550.
Dauphas, N., et al., 2004. Clues from Fe isotope variations on the origin of early Archaean
BIFs from Greenland. Science 306, 20772080.
Drennan, G.R., Robb, L.J., 2006. The nature of hydrocarbons and related uids in the
Witwatersrand Basin, South Africa: Their role in metal redistribution. Geol. Soc. Am.
Spec. Pap. 405, 353385.
Drennan, G.R., Boiron, M.C., Cathelineau, M., Robb, L.J., 1999. Characteristics of postdepositional uids in the Witwatersrand Basin. Mineral. Petrol. 66, 83109.
Els, B.G., 1998. The auriferous late Archaean sedimentation systems of South Africa:
unique palaeo-environmental conditions? Sediment. Geol. 120, 205224.
Engelbrecht, C.J., Baumbach, G.W.S., Matthysen, J.L., Fletcher, P., 1986. The West Wits Line.
In: Anhaeusser, C.R., Maske, S. (Eds.), Mineral Deposits of Southern Africa. Geological
Society of South Africa, pp. 599648.
England, G.L., Rasmussen, B., Krapez, B., Groves, D.I., 2002a. Archaean oil migration in the
Witwatersrand Basin of South Africa. J. Geol. Soc. 159, 189201.
51
England, G.L., Rasmussen, B., Krapez, B., Groves, D.I., 2002b. Palaeoenvironmental signicance
of rounded pyrite in siliciclastic sequences of the Late Archaean Witwatersrand Basin:
oxygen-decient atmosphere or hydrothermal alteration? Sedimentology 49,
11331156.
Eriksson, K.A., Turner, B.R., Vos, R.G., 1981. Evidence of tidal processes from the lower part
of the Witwatersrand Supergroup, South Africa. Sediment. Geol. 29, 309325.
Falconer, D.M., Craw, D., Youngson, J.H., Faure, K., 2006. Gold and sulphide minerals in
Tertiary quartz pebble conglomerate gold placers, Southland, New Zealand. Ore
Geol. Rev. 28, 525545.
Farquhar, J., Wing, B.A., 2003. Multiple sulfur isotopes and the evolution of the atmosphere. Earth Planet. Sci. Lett. 213, 113.
Farquhar, J., Bao, H., Thiemens, M., 2000. Atmospheric inuence of Earth's earliest sulfur
cycle. Science 289, 756758.
Farquhar, J., Savarino, J., Airieau, S., Thiemens, M., 2001. Observation of wavelengthsensitive mass-independent sulfur isotope ejects during SO2 photolysis: Implications
for the early atmosphere. J. Geophys. Res. Solid Earth 106, 3282932839.
Farquhar, J., Wu, N., Caneld, D.E., Oduro, H., 2011. Connections between sulfur cycle
evolution, sulfur isotopes, sediments, and base metal sulde Deposits. Econ. Geol.
105, 509533.
Foya, S.N., Reimold, W.U., Przybylowicz, W.J., Gibson, R.L., 1999. PIXE microanalysis of
gold-pyrite associations from the Kimberley Reefs, Witwatersrand basin, South Africa. Nucl. Instrum. Methods Phys. Res., Sect. B 158, 588592.
Frimmel, H.E., 1997. Chlorite thermometry in the Witwatersrand Basin: constraints on
the Paleoproterozoic geotherm in the Kaapvaal Craton, South Africa. J. Geol. 105,
601616.
Frimmel, H.E., 2005. Archaean atmospheric evolution: evidence from the Witwatersrand
gold elds, South Africa. Earth Sci. Rev. 70, 146.
Frimmel, H.E., Gartz, V.H., 1997. Witwatersrand gold particle chemistry matches model of
metamorphosed, hydrothermally altered placer deposits. Ore Geol. Rev. 32, 523530.
Frimmel, H.E., et al., 2005. The formation and preservation of the Witwatersrand
goldelds, the world's largest gold province. Econ. Geol. 100th Anniversary,
769798.
Frizzo, P., Rampazzo, G., Molinaroli, E., 1991. Authigenic iron sulphides in Recent sediments of the Venice Lagoon (northern Italy). Eur. J. Mineral. 3, 603612.
Germs, G.J.B., Schweitzer, J.K., 1994. A provisional model for the regional morphostratigraphy
of the Venterspost Conglomerate Formation in the West Rand and Carletonville Goldelds. S. Afr. J. Geol. 97, 279287.
Gibson, R.L., Reimold, W.U., 1999. The signicance of the Vredefort Dome for the thermal
and structural evolution of the Witwatersrand Basin, South Africa. Mineral. Petrol. 66,
523.
Goodwin, A.M., Monster, J., Thode, H.G., 1976. Carbon and sulfur isotope abundances
in Archaean iron-formations and early Precambrian life. Econ. Geol. 71, 870891.
Grassineau, N.V., et al., 2001. Antiquity of the biological sulphur cycle: evidence from sulphur and carbon isotopes in 2700 million-year-old rocks of the Belingwe Belt,
Zimbabwe. Proc. R. Soc. Lond. Ser. B Biol. Sci. 268, 113119.
Gray, D.J., Lintern, M.J., Longman, G.D., 1998. Chemistry of gold-humic interactions. Cooperative Research Centre for Landscape Evolution & Mineral Exploration Open File
Report (32 pp.).
Greenwood, P.F., et al., 2013. Organic geochemistry and mineralogy. I. Characterisation of organic matter associated with metal deposits. Ore Geol. Rev. 50, 127.
Guilbaud, R., Butler, I.B., Ellam, R.M., 2011. Abiotic pyrite formation produces a large Fe
isotope fractionation. Science 332, 15481551.
Guy, B.M., Beukes, N.J., Gutzmer, J., 2010. Paleaoenvironmental controls on the texture
and chemical composition of pyrite from non-conglomeratic sedimentary rocks of
the MesoArchaean Witwatersrand Supergroup, South Africa. S. Afr. J. Geol. 113,
195228.
Guy, B.M., Ono, S., Gutzmer, J., Kaufman, A.J., Lin, Y., Fogel, M.L., Beukes, N.J., 2012. A multiple sulfur and organic carbon isotope record from non-conglomeratic sedimentary
rocks of the MesoArchaean Witwatersrand Supergroup, South Africa. Precambrian
Res. 216219, 208231.
Habicht, K.S., Gade, M., Thamdrup, B., Berg, P., Caneld, D.E., 2002. Calibration of sulfate
levels in the Archaean ocean. Science 298, 23722374.
Hallbauer, D.K., 1986. The mineralogy and geochemistry of the Witwatersrand
pyrite, gold, uraninite and carbonaceous matter. In: Anhaeusser, C.R., Maske, S.
(Eds.), Mineral Deposits of Southern Africa. Geological Society of South Africa,
pp. 731752.
Hallbauer, D., Barton, J., 1987. The fossil gold placers of the Witwatersrand. Gold Bull. 20,
6879.
Hallbauer, D.K., Utter, T., 1977. Geochemical and morphological characteristics of gold
particles from recent river deposits and the fossil placers of the Witwatersrand. Mineral. Deposita 12, 293306.
Henckel, J., Schweitzer, J.K., 1994. Geochemical and mineralogical characteristics of a portion of the Ventersdorp Contact Reef at Elandsrand Gold Mine. S. Afr. J. Geol. 97,
332338.
Hofmann, A., Bekker, A., Rouxel, O., Rumble, D., Master, S., 2009. Multiple sulphur and iron
isotope composition of detrital pyrite in Archaean sedimentary rocks: a new tool for
provenance analysis. Earth Planet. Sci. Lett. 286, 436445.
Horscroft, et al., 2011. Witwatersrand metallogenesis: the case for (modied) syngenesis.
In: Noffke, Nora, Chafetz, Henry (Eds.), Microbial mats in siliciclastic depositional systems through time. SEPM Special Publication No. 101, pp. 7595.
Jaguin, J., Boulvais, P., Poujol, M., Boiron, M.C., Cathelineau, M., 2011. Stable isotope composition of quartz-calcite veins in the Witwatersrand basin, South Africa: implication
for basin-scale uid circulation. S. Afr. J. Geol. 113, 169182.
Johnson, C.M., Beard, B., Beukes, N., Klein, C., O'Leary, J., 2003. Ancient geochemical cycling
in the Earth as inferred from Fe isotope studies of banded iron formations from the
Transvaal Craton. Contrib. Mineral. Petrol. 144, 523547.
52
Johnson, C.M., Beard, B.L., Roden, E.E., Newman, D.K., Nealson, K.H., 2004. Isotopic constraints on biogeochemical cycling of Fe. Rev. Mineral. Geochem. 55, 359408.
Johnston, D.T., 2011. Multiple sulfur isotopes and the evolution of Earth's surface sulfur
cycle. Earth Sci. Rev. 106, 161183.
Kamber, B.S., Whitehouse, M.J., 2007. Micro-scale sulphur isotope evidence for sulphur
cycling in the late Archean shallow ocean. Geobiology 5, 517.
Kamo, S.L., Reimold, W.U., Krogh, T.E., Colliston, W.P., 1996. A 2.023 Ga age for the
Vredefort impact event and a rst report of shock metamorphosed zircons in
pseudotachylitic breccia and granophyre. Earth Planet. Sci. Lett. 144, 369387.
Kasting, J.F., 1993. Earth's early atmosphere. Science 259, 920926.
Koglin, N., Frimmel, H., Lawrie Minter, W., Brtz, H., 2010. Trace-element characteristics of
different pyrite types in Mesoarchaean to Palaeoproterozoic placer deposits. Mineral.
Deposita 45, 259280.
Koppel, V.H., Saager, R., 1974. Lead isotope evidence on the detrital origin of Witwatersrand pyrites and its bearing on the provenance of the Witwatersrand gold. Econ.
Geol. 69, 318331.
Kositcin, N., Krapez, B., 2004. Relationship between detrital zircon age-spectra and the
tectonic evolution of the Late Archaean Witwatersrand Basin, South Africa. Precambrian Res. 129, 141168.
Krapez, B., 1985. The Ventersdorp Contact placer: a gold-pyrite placer of stream and
debris-ow origins from the Archaean Witwatersrand Basin of South Africa. Sedimentology 32, 223234.
Krupp, R., Oberthr, T., Hirdes, W., 1994. The early Precambrian atmosphere and hydrosphere; thermodynamic constraints from mineral deposits. Econ. Geol. 89,
15811598.
Laetsch, T., Downs, R., 2006. Software for identication and renement of cell parameters
from powder diffraction data of minerals using the RRUFF project and American Mineralogist crystal structure databases. Abstracts from the 19th General Meeting of the
International Mineralogical Association, Kobe, Japan, 2328 July 2006.
Large, R.R., et al., 2009. Gold and trace element zonation in pyrite using a laser imaging
technique: implications for the timing of gold in orogenic and Carlin-style
sediment-hosted deposits. Econ. Geol. 104, 635668.
Large, R.R., Meffre, S., Burnett, R., Guy, B., Bull, S., Gilbert, S., Goemann, K., Danyushevsky, L.
, 2013. Evidence for an intrabasinal source and multiple concentration processes in
the formation of the Carbon Leader Reef, Witwatersrand Supergroup, South Africa.
Econ. Geol. 108, 12151241.
Law, J.D.M., Phillips, G.N., 2005. Hydrothermal replacement model for Witwatersrand
gold. Econ. Geol. 100th Anniversary, 799812.
Lengke, M.F., Southam, G., 2007. The deposition of elemental gold from gold(I)-thiosulfate
complexes mediated by sulfate-reducing bacterial conditions. Econ. Geol. 102,
109126.
MacLean, P.J., Fleet, M.E., 1989. Detrital pyrite in the Witwatersrand gold elds of South
Africa; evidence from truncated growth banding. Econ. Geol. 84, 20082011.
Marin-Carbonne, J., Rollion-Bard, C., Luais, B., 2011. In-situ measurements of iron isotopes
by SIMS: MC-ICP-MS intercalibration and application to a magnetite crystal from the
Gunint chert. Chem. Geol. 285, 5061.
Marin-Carbonne, J., Rollion-Bard, C., Bekker, A., Rouxel, O., Agangi, A., Cavalazzi, B.,
Wolgemuth-Ueberwasser, C.C., Hofmann, A., McKeegan, K.D., 2014. Coupled Fe and
S isotope variations in pyrite nodules from Archaean shales. Earth Planet. Sci. Lett.
392, 6779.
McCarthy, T.S., 1994. The tectono-sedimentary evolution of the Witwatersrand Basin with
special reference to its inuence on the occurrence and character of the Ventersdorp
Contact Reef A review. S. Afr. J. Geol. 97, 247259.
McClay, K.R., Ellis, P.G., 1983. Deformation and recrystallization of pyrite. Mineral. Mag.
47, 527538.
Meyer, F.M., Robb, L.J., Oberthr, T., Saager, R., Stupp, H.D., 1990. Cobalt, nickel, and gold in
pyrite from primary gold deposits and Witwatersrand reefs. S. Afr. J. Geol. 93, 7082.
Minter, W.E.L., 1982. The Golden Proterozoic (Chapter 4, pp. 115150). In: Tankard, A.J., et
al. (Eds.), Crustal Evolution of Southern Africa: 3.8 Billion Years of Earth History.
Springer-Verlag, New York, Heidelberg, Berlin, p. 523.
Mossman, D.J., et al., 2008. The indigenous origin of Witwatersrand carbon. Precambrian Res. 164, 173186.
Nishizawa, M., et al., 2010. Grain-scale iron isotopic distribution of pyrite from Precambrian shallow marine carbonate revealed by a femtosecond laser ablation multicollector
ICP-MS technique: possible proxy for the redox state of ancient seawater. Geochim.
Cosmochim. Acta 74, 27602778.
Noffke, N., Beukes, N., Gutzmer, J., Hazen, R., 2006. Spatial and temporal distribution of
microbially induced sedimentary structures: a case study from siliciclastic storm deposits of the 2.9 Ga Witwatersrand Supergroup, South Africa. Precambrian Res. 146,
3544.
Oberthr, T., Saager, R., 1986. Silver and mercury in gold particles from the Proterozoic
Witwatersrand placer deposits of South Africa; metallogenic and geochemical implications. Econ. Geol. 81, 2031.
Ono, S., et al., 2003. New insights into Archaean sulfur cycle from mass-independent sulfur isotope records from the Hamersley Basin, Australia. Earth Planet. Sci. Lett. 213,
1530.
Ono, S., Beukes, N.J., Rumble, D., 2009. Origin of two distinct multiple-sulfur isotope compositions of pyrite in the 2.5 Ga Klein Naute Formation, Griqualand West Basin, South
Africa. Precambrian Res. 169, 4857.
Pavlov, A.A., Kasting, J.F., 2002. Mass-independent fractionation of sulfur isotopes in Archean sediments: Strong evidence for an anoxic Archean atmosphere. Astrobiology
2, 2741.
Percak-Dennett, E.M., et al., 2011. Iron isotope fractionation during microbial dissimilatory iron oxide reduction in simulated Archaean seawater. Geobiology 9, 205220.
Philippot, P., et al., 2007. Early Archaean microorganisms preferred elemental sulfur, not
sulfate. Science 317, 15341537.
Phillips, G.N., Law, J.D.M., 2000. Witwatersrand gold elds; geology, genesis, and exploration. Rev. Econ. Geol. 13, 439500.
Phillips, G.N., et al., 1989. The Witwatersrand gold elds; Part I, Postdepositional history,
synsedimentary processes, and gold distribution. Econ. Geol. Monogr. 6, 585597.
Phillips, G.N., Law, J.D.M., Myers, R.E., 1990. The role of uids in the evolution of the Witwatersrand Basin. S. Afr. J. Geol. 93, 5469.
Planavsky, N., Asael, D., Hofmann, A., Reinhard, C.T., Lalonde, S.V., Knudsen, A., Wang, X.,
Ossa Ossa, F., Pecoits, E., Smith, B., Beukes, N., Bekker, A.O., Konhauser, K.O.,
Johnson, T.M., Lyons, T.W., Rouxel, O.J., 2014. Evidence for oxygenic photosynthesis
half a billion years before the Great Oxidation Event. Nat. Geosci. 7, 283286.
Pufahl, P.K., Grimm, K.A., 2003. Coated phosphate grains: proxy for physical, chemical,
and ecological changes in seawater. Geology 31, 801804.
Ramdohr, P., 1958. New observations on the ores of the Witwatersrand in South Africa
and their genetic signicance. Trans. Geol. Soc. S. Afr. 61, 150.
Rasmussen, B., Fletcher, I.R., Muhling, J.R., Mueller, A.G., Hall, G.C., 2007. Bushveld-aged
uid ow, peak metamorphism, and gold mobilization in the Witwatersrand basin,
South Africa: constraints from in situ SHRIMP U-Pb dating of monazite and xenotime.
Geology 35, 931934.
Reich, M., Kesler, S.E., Utsunomiya, S., Palenik, C.S., Chryssoulis, S.L., Ewing, R.C., 2005. Solubility of gold in arsenian pyrite. Geochim. Cosmochim. Acta 69, 27812796.
Reimer, T.O., 1984. Alternative model for the derivation of gold in the Witwatersrand Supergroup. J. Geol. Soc. 141, 263271.
Reimold, W.U., Przybylowicz, W.J., Gibson, R.L., 2004. Quantitative major and trace elemental mapping by PIXE of concretionary pyrite from the Witwatersrand Basin,
South Africa. X-Ray Spectrom. 33, 189203.
Reith, F., Lengke, M.F., Falconer, D., Craw, D., Southam, G., 2007. The geomicrobiology of
gold. ISME J. 1, 567584.
Robb, L.J., Meyer, F.M., 1995a. The nature of the Witwatersrand and hinterland: conjectures on the source area problem. Econ. Geol. 85, 511536.
Robb, L.J., Meyer, F.M., 1995b. The Witwatersrand Basin, South Africa: geological framework and mineralization processes. Ore Geol. Rev. 10, 6794.
Robb, L.J., Meyer, F.M., Ferraz, M.F., Drennan, G.K., 1990. The distribution of radioelements
in Archaean granites of the Kaapvaal Craton, with implications for the source of uranium in the Witwatersrand Basin. S. Afr. J. Geol. 93, 540.
Roerdink, D.L., Mason, P.R.D., Farquhar, J., Reimer, T., 2012. Multiple sulfur isotopes in
Paleoarchean barites identify an important role for microbial sulfate reduction in
the early marine environment. Earth Planet. Sci. Lett. 331332, 177186.
Rouxel, O.J., Bekker, A., Edwards, K.J., 2005. Iron isotope constraints on the Archaean and
Paleoproterozoic ocean redox state. Science 307, 10881091.
Saager, R., 1970. Structures in pyrite from the Basal Reef in the Orange Free State goldeld. Trans. Geol. Soc. S. Afr. 73, 2946.
Saager, R., Mihalik, P., 1967. Two varieties of pyrite from the Basal Reef of the Witwatersrand system. Econ. Geol. 62, 719731.
Schidlowski, M., 1965. Probable life-forms from the Precambrian of the Witwatersrand
System (South Africa). Nature 205, 895896.
Schieber, J., 2007. Oxidation of detrital pyrite as a cause for marcasite formation in marine
lag deposits from the Devonian of the eastern US. Deep-Sea Res. II Top. Stud.
Oceanogr. 54, 13121326.
Schieber, J., Riciputi, L., 2004. Pyrite ooids in Devonian black shales record intermittent
sea-level drop and shallow-water conditions. Geology 32, 305308.
Schieber, J., Riciputi, L., 2005. Pyrite and marcasite coated grains in the Ordovician Winnipeg Formation, Canada: an intertwined record of surface conditions, stratigraphic
condensation, geochemical reworking, and microbial activity. J. Sediment. Res. 75,
907920.
Schweigart, H., von Rahden, H.O., 1965. Oolithische Strukturen in Pyriten des
Ventersdorp-Contact-Reefs, Sdafrika. Geol. Rundsch. 54, 11431148.
Shen, Y., Buick, R., Caneld, D.E., 2001. Isotopic evidence for microbial sulphate reduction
in the early Archaean era. Nature 410, 7781.
Shen, Y., Farquhar, J., Masterson, A., Kaufman, A.J., Buick, R., 2009. Evaluating the role of
microbial sulfate reduction in the early Archaean using quadruple isotope systematics. Earth Planet. Sci. Lett. 279, 383391.
Spangenberg, J.E., Frimmel, H.E., 2001. Basin-internal derivation of hydrocarbons in the
Witwatersrand Basin, South Africa: evidence from bulk and molecular 13C data.
Chem. Geol. 173, 339355.
Stacey, J.S., Kramers, J.D., 1975. Approximation of terrestrial lead isotope evolution by a
two-stage model. Earth Planet. Sci. Lett. 26, 207221.
Stevens, G., Boer, R., Gibson, R., 1997. Metamorphism, uid-ow, and gold mobilization in
the Witwatersrand Basin; towards a unifying model. S. Afr. J. Geol. 100, 363375.
Thiemens, M.H., 1999. Mass-independent isotope effects in planetary atmospheres and
the early solar system. Science 283, 341345.
Thomassot, E., Cartigny, P., Harris, J.W., Lorand, J.P., Rollion-Bard, C., Chaussidon, M., 2009.
Metasomatic diamond growth: A multi-isotope study (13C, 15N, 33S, 34S) of sulphide inclusions and their host diamonds from Jwaneng (Botswana). Earth Planet.
Sci. Lett. 282, 7990.
Ueno, Y., Ono, S., Rumble, D., Maruyama, S., 2008. Quadruple sulfur isotope analysis of ca.
3.5 Ga Dresser Formation: new evidence for microbial sulfate reduction in the early
Archaean. Geochim. Cosmochim. Acta 72, 56755691.
Ueno, Y., et al., 2009. Geological sulfur isotopes indicate elevated OCS in the Archaean atmosphere, solving faint young sun paradox. Proc. Natl. Acad. Sci. 106, 1478414789.
Utter, T., 1978. Morphology and geochemistry of different pyrite types from the Upper
Witwatersrand System of the Klerksdorp Goldeld, South Africa. Geol. Rundsch. 67,
774804.
Vollbrecht, A., Oberthr, T., Ruedrich, J., Weber, K., 2002. Microfabric analyses applied to
the Witwatersrand gold- and uranium-bearing conglomerates: constraints on the
provenance and post-depositional modication of rock and ore components. Mineral.
Deposita 37, 433451.

Wacey, D., McLoughlin, N., Whitehouse, M.J., Kilburn, M.R., 2010. Two coexisting sulfur
metabolisms in a ca. 3400 Ma sandstone. Geology 38, 11151118.
Watanabe, Y., Farquhar, J., Ohmoto, H., 2009. Anomalous fractionations of sulfur isotopes
during thermochemical sulfate reduction. Science 324, 370373.
Welch, S.A., Beard, B.L., Johnson, C.M., Braterman, P.S., 2003. Kinetic and equilibrium Fe
isotope fractionation between aqueous Fe(II) and Fe(III). Geochim. Cosmochim.
Acta 67, 42314250.
Whitehouse, M.J., Fedo, C.M., 2007. Microscale heterogeneity of Fe isotopes in N3.71 Ga
banded iron formation from the Isua Greenstone Belt, southwest Greenland. Geology
35, 719722.
Whitehouse, M.J., Kamber, B.S., Fedo, C.M., Lepland, A., 2005. Integrated Pb- and S-isotope
investigation of sulphide minerals from the early Archaean of southwest Greenland.
Chem. Geol. 222, 112131.
Xenophontos, L., Stevens, G., Przybylowicz, W.J., 1999. Micro-PIXE elemental imaging of
pyrites from the Bulawayan-Bubi Greenstone Belt, Zimbabwe. Nucl. Instrum.
Methods Phys. Res. Sect. B 150, 496501.
Xenophontos, L., Stevens, G., Przybylowicz, W.J., 2001. The pyritegold association in
dolomites and carbonated metavolcanics from the Turk Mine, Zimbabwe: the link
between late shearing and gold introduction. S. Afr. J. Geol. 104, 217230.
53
Yamaguchi, K.E., Johnson, C.M., Beard, B.L., Ohmoto, H., 2005. Biogeochemical cycling of
iron in the Archaean-Paleoproterozoic Earth: constraints from iron isotope variations
in sedimentary rocks from the Kaapvaal and Pilbara Cratons. Chem. Geol. 218,
135169.
Yoshiya, K., Nishizawa, M., Sawaki, Y., Ueno, Y., Komiya, T., Yamada, K., Yoshida, N., Hirata,
T., Wada, H., Maruyama, S., 2012. In situ iron isotope analyses of pyrite and organic
carbon isotope ratios in the Fortescue Group: Metabolic variations of a Late Archaean
ecosystem. Precambrian Res. 212213, 169193.
Zartman, R.E., Frimmel, H.E., 1999. Rn-generated 206Pb in hydrothermal sulphide minerals
and bitumen from the Ventersdorp Contact Reef, South Africa. Mineral. Petrol. 66,
171191.
Zhao, B., Robb, L.J., Harris, C., Jordaan, L.J., 2006. Origin of hydrothermal uids and gold
mineralization associated with the Ventersdorp Contact Reef, Witwatersrand Basin,
South Africa: constraints from S, O, and H isotopes. Geol. Soc. Am. Spec. Pap. 405,
333352.
All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.

Agangi Et Al 2015

Caricato da

Informazioni sul documento

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Agangi Et Al 2015

Caricato da

Copyright:

Formati disponibili

Earth-Science Reviews 140 (2015) 2753

Contents lists available at ScienceDirect

Gold accumulation in the Archaean Witwatersrand Basin,

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

Laser Raman spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . .

contain appreciable amounts of gold (up to several tens of ppm of Au;

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

of the major gold-bearing conglomerate horizons in the Witwatersrand

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

3. Previous investigations of CLP in the Witwatersrand Basin and

cortices of pyrite and marcasite from the Ordovician shallow marine

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

4.3. Laser Raman spectroscopy

Pyrite trace element concentration maps were obtained at CODES,

In situ, non-destructive Raman analyses on various minerals were

4.5. Laser ablation ICP-MS trace element maps

5. Sample description and pyrite textures

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

the exception of CLP grains, which show widespread evidence for

inlling of larger cavities present in some inclusion-rich grains.

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

veinlets of massive pyrite truncated at grain margins. These veinlets

association with chlorite in the pores of some pyrite grains; and 3) in

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

Au in the analysed pyrite. The chlorite-muscovite-rich matrix around

in comparison with bulk-Earth values; Stacey and Kramers, 1975), with

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

8. Sulfur and Fe isotope composition of pyrite

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

Comparison of 34S of different massive grains indicates that much of the

except a few analyses affected by post-sedimentary recrystallisation or

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

distribution of values, with a peak roughly between 0.50 and

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

bdl: below detection limit.

addition, X-ray maps seem to indicate the presence of ne-grained

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

All values referred to VCDT.

9.2. Distinguishing primary and secondary gold in CLP

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

metamorphic uid circulation, and gold inclusions also Ag-rich are

9.3. Distinguishing primary vs. post-sedimentary isotope signatures

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

photolysis, a chemical reaction that involves breaking of molecules via

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

SPy Saq PySaq 1 ln f ;

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

FePy Feaq PyFeaq 1 ln f :

56Fe = 1.6 in inclusion-rich pyrite) may be explained by local

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

All values referred to IRMM-014.

produced by photolysis of volcanic gases by UV radiation are the main

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

2.5-2.7 Ga black shales pyrite

precipitate gold intracellularly and extracellularly and as products of

A. Agangi et al. / Earth-Science Reviews 140 (2015) 2753

Sediment source area

environment dominated by deposition of clay- to silt-sized clastic

delicate textures; 2) similar mineral inclusions; and 3) similar trace

Supplementary data to this article can be found online at http://dx.