Patten Et Al., 2017

Ore Geology Reviews 86 (2017) 487–508
Contents lists available at ScienceDirect
Ore Geology Reviews

journal homepage: www.elsevier.com/locate/oregeo
Hydrothermal mobilisation of Au and other metals in supra-subduction

oceanic crust: Insights from the Troodos ophiolite
C.G.C. Patten a,⇑, I.K. Pitcairn a, D.A.H. Teagle b
a
Department of Geological Sciences, Stockholm University, Sweden
b
Ocean and Earth Science, National Oceanography Centre Southampton, University of Southampton, United Kingdom
a r t i c l e i n f o a b s t r a c t
Article history: The Troodos ophiolite is an ideal location to investigate the relationships between the mobilisation of
Received 8 August 2016 metals by hydrothermal fluids from the lower oceanic crust and the formation of volcanogenic massive
Accepted 21 February 2017 sulphide (VMS) deposits. The ophiolite hosts the classic ‘‘Cyprus-type” Cu-rich VMS deposits as well as
Available online 22 February 2017
abundant zones of epidosite alteration in the lower sheeted dyke section that are significantly depleted
in base metals including Cu and Zn, and are considered to be the source of the metals enriched in the
Keywords: overlying deposits. Previous research indicates that the Troodos VMS deposits are irregularly enriched
VMS deposit
in Au and related elements As, Sb, and Se, but the behaviour of these elements during the hydrothermal
Hydrothermal alteration
Gold
alteration of the Troodos ophiolite hitherto has been poorly investigated. Low detection limit whole rock
Troodos ophiolite analyses of fresh glass samples reveal that the Troodos primitive crust has a similar metal content and
Magnetite crisis distribution to modern-day arc-related environments such as the Manus Basin. Compared to mid-
oceanic ridge basalt (MORB), the Troodos primitive crust is enriched in As, Sb and Pb most likely due
to addition from a subducting slab during crustal formation. During early stages of magmatic differenti-
ation (9–3.5 wt% MgO) Au, As, Sb, Se, Cu, Zn and Pb behave as incompatible elements due to the sulphide-
undersaturated nature of the melt. The onset of magnetite crystallisation, however, at 3.5 wt% MgO
leads to sulphide segregation and depletion of strongly chalcophile elements (Au, Cu and Se) during con-
tinued differentiation (<3.5 wt% MgO) whereas poorly chalcophile elements (As, Sb, Zn and Pb) remain
incompatible. These differences in metal behaviour can account for the Cu-rich, Zn-Pb-poor of the
Cyprus-type VMS deposits as the source area rocks show high Cu fertility compared to Zn and Pb.
Mobilisation of metals during hydrothermal alteration of the Troodos ophiolite is more extensive than
observed in hydrothermally altered MORB. Mass balance calculations show that the epidosite zones
are significantly depleted in Au (88 ± 16%), As (89 ± 23%), Sb (60 ± 12%), Se (91 ± 20%), Cu
(84 ± 18%), Zn (63 ± 9%) and Pb (60 ± 8%). Background altered diabase from outside epidosite zones
shows similar metal depletions which suggests that the source areas of VMS are not restricted to epi-
dosite zones but are extended to the lower sheeted dyke section. The masses of metals mobilised from
a source area of 10.9 km3, (composed of a 5 km3 epidosite zone and 5.9 km3 of background altered dia-
base) in the Solea graben are 47 t Au, 21 kt As, 1200 t Sb, 3100 t Se, 2.4 Mt Cu, 1.8 Mt Zn and 27 kt Pb.
Comparison of metal quantities mobilised from lower sheeted dike section in the Solea graben with those
hosted in VMS deposits shows trapping efficiencies ranging from 4 to 37% indicating that most of the
metals is lost by other processes.
Ó 2017 Elsevier B.V. All rights reserved.
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 488
2. Geological setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
2.1. The Troodos ophiolite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
2.2. Hydrothermal system of the Troodos ophiolite. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 490
⇑ Corresponding author.
E-mail address: [email protected] (C.G.C. Patten).
http://dx.doi.org/10.1016/j.oregeorev.2017.02.019
0169-1368/Ó 2017 Elsevier B.V. All rights reserved.
488 C.G.C. Patten et al. / Ore Geology Reviews 86 (2017) 487–508
2.3. The VMS deposits of the Troodos ophiolite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491

3. Sampling and analytical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491
4. Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
4.1. The fresh glasses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
4.2. The volcanic section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
4.3. The lower sheeted dyke complex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
4.4. Plutonic complex and ultramafics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494
5. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494
5.1. Composition of the primitive crust at Troodos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494
5.2. Behaviour of S in the primitive crust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494
5.3. Metal behaviour in the primitive oceanic crust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497
6. Mass variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 498
6.1. Mass variation in the volcanic section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
6.2. Mass variation in the lower sheeted dyke section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
6.2.1. Epidosite zones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
6.2.2. Background altered diabase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500
6.2.3. Formation of VMS deposits in the Solea graben . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500
6.3. The role of primitive crustal composition in the formation of VMS deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503
7. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505
7.1. The major outcomes of this study are . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 506
Appendix A. Supplementary data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 506
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 506
1. Introduction et al., 1990; Poulsen and Hannington, 1996; Hannington, 1999;

Barriga et al., 2007). Third, oceanic crustal rocks formed in arc-
Volcanogenic massive sulphide (VMS) deposits form by focused related settings have generally higher Au content than in MOR set-
discharge of hydrothermal fluids on or near the seafloor (e.g. tings, which increases the amount of Au available for mobilisation
Tornos et al., 2015). They are exploited primarily for the base met- by hydrothermal circulation (Huston, 2000; Moss et al., 2001;
als Cu, Zn and Pb (e.g. Galley et al., 2007; Mudd et al., 2013), and it Herrington et al., 2011; Pitcairn, 2011). In mid-oceanic ridge
is generally accepted that these metals are sourced in part from (MOR) settings significant masses of Au as well as As, Sb and Se
hydrothermal reaction zones in the lower sheeted dyke section (hereafter referred to as related elements) are mobilised during
of the oceanic crust (e.g. Alt, 1995). Depletions in base metals from alteration in the lower sheeted dykes, but the degree of mobilisa-
the lower sheeted dyke section has been observed in deep drill tion is similar to that of the base metals and alone is not sufficient
cores in the oceanic crust, such as Ocean Drilling Program (ODP) to generate the Au enrichment required to form Au-rich VMS
Hole 504B and Hole 1256D (e.g. Alt et al., 1989, 2010; Alt, 1995; deposits (Patten et al., 2015). Higher primary Au content in arc-
Bach et al., 2003; Teagle et al., 2006; Patten et al., 2015). Base metal related rocks has been previously suggested as the cause for Au
depletions have also long been recognised from epidosite zones in enrichment in VMS deposits (e.g. Moss et al., 2001; Herrington
ophiolites such as Troodos, Cyprus, (Richardson et al., 1987; et al., 2011). However, the behaviour of Au and related elements
Schiffman et al., 1987; Schiffman and Smith, 1988) where over during hydrothermal alteration of arc-related oceanic crust has
80% of the primary Cu and Zn has been removed (Jowitt et al., not been investigated.
2012). An important sub-class of VMS deposits are those enriched The Troodos ophiolite, Cyprus, which is one of world’s most
in Au, Ag, and a related suite of semi-metals As, Sb, Se, Te and Bi well studied ophiolites, shows a full exposure of the crustal
(Hannington, 1999; Huston, 2000; Dubé et al., 2007). In certain sequence from ultramafic rocks to seafloor sediments. It is supra-
deposits, Au contents are sufficiently high that they are classified subduction oceanic crust (e.g. Miyashiro, 1973; Rautenschlein
as Au-rich VMS deposits. Mercier-Langevin et al. (2011) charac- et al., 1985; Muenow et al., 1990) that probably formed during
terised Au-rich VMS deposits by a grade of >3.46 g/t Au and a ton- subduction initiation (Pearce and Robinson, 2010). The Troodos
nage of 31 t Au. Despite the economic significance, the causes of Au ophiolite has contributed greatly to our understanding of the for-
enrichments in these deposits are not well constrained. mation of VMS deposits as it hosts the eponymous Cyprus-type
Gold-rich VMS deposits occur mostly in arc-related tectonic VMS deposits; structurally controlled Cu-rich sulphide deposits
settings and are hosted in intermediate to felsic volcanic rocks hosted in mafic volcanic rocks that form at or near the seafloor
(e.g. Herzig et al., 1993; Barrie and Hannington, 1999). Arc settings (Constantinou and Govett, 1973; Adamides, 2010a,b). The ophiolite
display a number of characteristics that could potentially promote also exposes the hydrothermal reaction zones located at the base
Au enrichment in VMS deposits. First, arc-related magmas are gen- of the sheeted dyke complex that are considered to be the source
erally volatile-rich and undergo significant magmatic degassing areas for the base metals enriched in the deposits (Richardson
which may cause direct input of metals into hydrothermal fluids et al., 1987; Schiffman et al., 1987; Schiffman and Smith, 1988;
(e.g. Urabe et al., 1987; Stanton, 1990; Yang and Scott, 1996, Richards et al., 1989; Bettison-Varga et al., 1992; Jowitt et al.,
2002; Sillitoe et al., 1996; Kamenetsky et al., 2001; Moss et al., 2012). The Troodos ophiolite therefore provides an ideal locality
2001; Petersen et al., 2002; Sun et al., 2004; de Ronde et al., for investigation of the mobility of metals including Au during
2011). Second, oceanic crust produced in volcanic arc settings hydrothermal alteration of arc-related oceanic crust. The Au con-
can form in relatively shallow seawater (>2000–1500 mbsf, e.g. tent of the VMS deposits in Troodos is poorly documented
Monecke et al., 2014) which, due to the lower confining pressure, (Herzig et al., 1991) but there is evidence for significant but
promotes sub-seafloor boiling of hydrothermal fluids and metal irregular Au enrichment. The sulphide ore from the Limni deposit
fractionation that could cause Au-enrichment during VMS forma- for example, has 3.4 ppm Au grade (Table 1; Mosier et al., 1983).
tion (Urabe et al., 1987; Huston and Large, 1989; Butterfield Seafloor weathering of the VMS deposits has caused supergene
C.G.C. Patten et al. / Ore Geology Reviews 86 (2017) 487–508 489
Table 1
Gold and related elements content in Troodos VMS deposits, gossans, ochers and umbers.
References S (%) Cu (%) Zn (%) Pb (ppm) Au (ppm) As (ppm) Sb (ppm) Se (ppm)

Deposit Tonnage (Mt)
Limni massive body 1 0.42 1.4 3.4
Mavrovouni 1 15 49 3.8 0.5 0.3
Kavalasos/Mousoulos 1 6.9 35 1 0.5 100 1.7
Skouriotissa (Phoukasa) 1 5.4 48 2.3 0.06 0.5⁄
Skouriotissa (Phoenix) 1 15 0.5
Apliki 1 1.65 40 1.8 0.2
West Apliki 2 3.6 0.34
Limni 2 16 14 1 0.2
Three Hills 2 6.2 0.37
Memi 2 1.5 26 0.1
Mathiatis 2 2.8 33 0.2 1
2
Pyritic massive ore material (n)
Skouriotissa 1 8 2.78 0.02 0.5** 74** 3** 296**
Mathiati 1 45 48 0.17 0.26 97 0.8 160 11 20
Sha 1 1 36.7 1.9 1.2 94 1 657 20 26
Memi 1 2 49.8 0.02 2 0.02 23 35
Kambia 1 1 0.04 70 3
Kinousa-Uncle Charles 1 1 51.5 0.18 0.02 19 0.3 105 17 5
Agrokipia 1 2 46.9 0.07 0.03 30 0.14 201 3 144
Mavrovouni 1 1 3.69 0.8 1.05 170 330
VMS related Gossan/Ocher/Umber (n)
Ochers Skouriotissa 3 6 0.1 0.1 7.0 276 17.50 106.3
Umbers Skouriotissa 3 4 0.3 0.1 0.08 265 3.5
Meteoric gossan 3 1 0.1 0.0 0.23 18.0 1.0 4.0
Silicified umber Skouriotissa 4 13 2.35
Umber (n)
Perapedhi 4 13 0.08
Asgata 4 9 0.11
Pyrga 4 7 1.42
Kokkinovounaro 4 7 0.91
Typical Unsilicified Umber 4 10 0.01 303
Typical Silicified Umber 4 6 1.87 50
n = number of samples. ⁄Estimated Au grade.

**
One sample MH-16 (Hannington et al., 1998; Herzig et al., 1991). 1 = Hannington et al. (1998); 2 = Adamides (2010b); 3 = Herzig et al. (1991); 4 = Prichard and Maliotis
(1998).
Au enrichment in gossans and ochres (Herzig et al., 1991) and Au- et al., 2012). Regelous et al. (2014) alternatively propose formation
rich umbers, either primary or enriched by off-axis low tempera- in a fore-arc setting at a triple junction. The core of the ophiolite is
ture fluid circulation, are also common (Herzig et al., 1991; characterised by partially to completely serpentinised peridotites
Prichard and Maliotis, 1998). representing the mantle sequence (e.g. Gass, 1990; Batanova and
In this study we apply low detection limit analytical tech- Sobolev, 2000; Nuriel et al., 2009) surrounded by the plutonic
niques to investigate the behaviour of Au, As, Sb and Se, as well sequence composed of layered ultramafics, gabbros and pla-
Cu, Zn and Pb, during hydrothermal alteration of the Troodos giogranites corresponding to the lower part of the oceanic crust
ophiolite. The objectives are: (1) to determine the distribution of (e.g. Gass, 1990; Fig. 1). Higher in the sequence the extensively
Au and related elements in the primary crust of the Troodos ophi- exposed sheeted dyke complex represents the fossilised pathways
olite for comparison with modern day supra-subduction zone for magma migrating from the magmatic chambers to the volcanic
analogues; and (2) to quantify the mobility of Au and related ele- section (Baragar et al., 1987). The dykes were produced from differ-
ments from the different lithological units during hydrothermal ent magmatic chambers (Baragar et al., 1987) with intrusion verti-
fluid circulation, with specific focus on the epidosite zones. We cal, lateral and/or orthogonal to the magmatic chamber but always
then discuss the implications for formation of VMS deposits, parallel to the spreading-axis in a near-vertical plane (Baragar
particularly those that are Au-rich. et al., 1987; Staudigel et al., 1992, 2000). The upper oceanic crust
at Troodos, commonly referred to as the volcanic section, has been
2. Geological setting sub-divided, based on field evidence and analyses of fresh glass
samples (e.g. Miyashiro, 1973; Robinson et al., 1983; Cameron,
2.1. The Troodos ophiolite 1985; Rautenschlein et al., 1985; Thy et al., 1985; Bednarz and
Schmincke, 1994; Portnyagin et al., 1997), into the basal unit
The Troodos ophiolite was formed in the eastern Mediterranean (BU), the lower lava sequence (LLS) and the upper lava sequence
basin during the closure of the Tethys Ocean (Robertson, 1987; (ULS; Fig. 1). The basal unit is a combination of intrusive dykes
Garfunkel, 1998) and is a segment of supra-subduction oceanic and extrusive lavas that represents the transitional zone of the
crust (e.g. Miyashiro, 1973; Pearce, 1975). It is considered to have oceanic crust (Fig. 1). The lower and upper lava sequences have a
formed during subduction initiation (Pearce et al., 1984; composition that follows a basaltic andesite-andesite-dacite-rhyo
Rautenschlein et al., 1985; Pearce and Robinson, 2010) at 90– lite (BADR) trend but also host boninite and high magnesian ande-
92 Ma (Mukasa and Ludden, 1987; Osozawa et al., 2012). Boninitic site series (Bailey et al., 1991; Pearce and Robinson, 2010 and ref-
volcanism extended up to 75 Ma due to ridge subduction (Osozawa erences therein).
Fig. 1. Geological map of the Troodos ophiolite. The main epidosite zones, Au-bearing VMS deposits and Au-rich umbers are displayed. Modified from Cyprus Geological
Survey Mineral Resources Map of Cyprus (2007), Bickle and Teagle (1992), Bettison-Varga et al. (1992), Prichard and Maliotis (1998), Adamides (2010), Jowitt (2009) and
Osozawa et al. (2012).
2.2. Hydrothermal system of the Troodos ophiolite (>200 °C) alteration dominating in the sheeted dyke and plutonic
complexes. The upper section of the sheeted dyke complex was
The ancient hydrothermal system in the Troodos ophiolite altered mainly to sub-greenschist facies conditions (200–250 °C)
shares many similarities with those in modern-day oceanic crust whereas the lower section was altered at greenschist-
(e.g. Gillis and Robinson, 1988; Alt and Teagle, 2000). It can be sim- amphibolite facies conditions (>300 °C; Schiffman and Smith,
plified into two subdomains; the seawater-derived fluids domain 1988; Gillis and Robinson, 1990). Background altered dykes
that mainly affected the volcanic section, and the hydrothermal (referred as diabase) were partially to completely altered to amphi-
fluids domain that affected the sheeted dyke and plutonic com- bole, albite, and chlorite, with minor quartz, epidote, titanite and
plexes. The alteration in the seawater-derived fluids domain was secondary sulphides (e.g. Jowitt et al., 2012; Cann et al., 2015). A
caused by flow of low temperature fluids (50–200 °C) producing narrow range of elevated strontium isotope compositions
pervasive background alteration throughout the volcanic section (87Sr/86Sr 0.7053 ± 6) across the Troodos sheeted dykes indicates
with smectite, celadonite, zeolite and calcite being the most com- that the downwelling of seawater-derived hydrothermal recharge
mon secondary minerals (Gillis and Robinson, 1988; Bednarz and fluids was pervasive (Bickle and Teagle, 1992), rather than chan-
Schminke, 1994; Gillis and Robinson, 1990). An increase in nelled by deep faults.
alteration temperatures occurred from the lower lava sequence In contrast to modern-day oceanic crust where the discharge of
(50–100 °C) to the basal unit (100–200 °C; Gillis and Robinson, hydrothermal fluids appears to occur in the same pathways as fluid
1988). Late stage pervasive flow of low temperature (<50 °C) oxi- recharge (e.g., Harris et al., 2015), discrete zones of intensive
dised seawater at high water rock ratios led to seafloor weathering hydrothermal fluid alteration at high temperature (350–400 °C)
in the top of the lava pile in the upper volcanic section (Gillis and and high water rock ratio (Seyfried et al., 1988) are evident in
Robinson, 1988). Secondary alteration minerals in this zone the Troodos ophiolite through the localised development of epi-
include Fe-hydroxide, smectite, K-feldspar and calcite (Gillis and dosite zones (Richardson et al., 1987; Schiffman and Smith, 1988;
Robinson, 1988; Bednarz and Schminke, 1994; Gillis and Gillis and Robinson, 1990; Bickle and Teagle, 1992). Epidosites rep-
Robinson, 1990). Finally localised alteration halos associated with resent zones of extensive interaction between the focused upwel-
VMS deposits, stockworks and mineralised zones developed during ling high temperature hydrothermal fluids with newly intruded
interaction of hydrothermal fluids with seawater-derived fluids at dykes (Cann et al., 2015). The alteration can be a dyke-by-dyke
high temperature (300 °C; e.g. Gillis and Robinson, 1988; process that leads to rapid porosity generation and extensive dis-
Richards et al., 1989; Bednarz and Schminke, 1994; Gillis and solution of primary mineral assemblage (Cann et al., 2015). Six
Robinson, 1990), with smectite, illite, chlorite, quartz, pyrite and main epidosite zones have been recognised on Troodos: five within
chalcopyrite being the most abundant secondary minerals the sheeted dyke complex and one in plagiogranites (Richardson
(Richards et al., 1989; Gillis and Robinson, 1990). et al., 1987; Bettison-Varga et al., 1992; Jowitt et al., 2012; Cann
An important change in the hydrothermal system occurs et al., 2015). Epidosite zones are not restricted to the lower sheeted
between the basalt unit and the upper sheeted dyke section with dyke section and extend into the upper plutonics. They are several
the hydrothermal fluid domain characterised by high temperature hundred meters wide and up to a kilometre long relative to the
strike of the dykes (Richardson et al., 1987; Bettison-Varga et al., VMS deposits, additional Au enrichment occurred during acid sul-
1992; Cann et al., 2015). Within the epidosite zones the degree phate ground water leaching into the ‘‘Devil’s mud”, a siliceous
of alteration is highly variable both at outcrop and dyke scale giv- horizon underlying or overlying the orebody (Wilson et al., 1959;
ing the specific pistachio banded aspect (Jowitt et al., 2012; Cann Bear, 1963; Herzig et al., 1991). Gold was profitably mined from
et al., 2015). Diabase is the least altered and is characterised by Skouriotissa from the early 1930s to 1942 and a total of 2.2 t of
amphibole, plagioclase, chlorite, epidote, quartz, oxides and titan- Au have been extracted (Wilson et al., 1959). Gold was also mined
ite with conservation of igneous texture (Jowitt et al., 2012). It is from the Apliki deposit (80 kg; Wilson et al., 1959).
followed by transitional diabase-epidote which contains higher
chlorite (>30%), epidote (15–30%) and quartz (15–20%) abundances
but lacks igneous texture (Jowitt et al., 2012). Intermediate epi- 3. Sampling and analytical methods
dosite has high epidote and quartz content (>30%) but moderate
chlorite (15–30%), and finally end-member epidosite, which repre- A suite of 124 samples representative of the variation in lithol-
sents the most advance stage of hydrothermal alteration, is charac- ogy and alteration have been used in this study including 89 col-
terised by a granoblastic texture and high epidote-quartz content lected during fieldwork, and 25 sheeted dyke and gabbro
(>60%) and little chlorite (<10%, Jowitt et al., 2012). Replacement samples previously documented by Bickle and Teagle (1992). In
of Fe-Ti oxides by titanite increases with the degree of epidotisa- addition, ten fresh glass samples representative of the BADR series
tion (e.g. Jowitt et al., 2012). Mobilisation of Cu and Zn during epi- (Coogan et al., 2017; Appendix 1) have been analysed to determine
dotisation has been extensively documented (Richardson et al., the primary concentration of Au and related elements in the Troo-
1987; Schiffman et al., 1987; Schiffman and Smith, 1988; dos crust. From the basal unit and volcanic section, 26 samples
Humphris and Cann, 2000). More recently, Jowitt et al. (2012) pre- have been selected. Samples are mostly of pillow lava, lava flow,
cisely quantified base metal release with Cu (87% to 91%), Zn and hyaloclastite but also include samples of dykes, picritic lava
(63% to 65%), Ni (59% to 78%) and Mn (52% to 60%) being flows and alteration halos surrounding VMS mineralisation.
significantly depleted depending on the degree of epidotisation. Sixty-seven samples from the lower sheeted dyke section have
been selected mostly from the Spilia-Kannavia, Kykko-Pedhoulas,
2.3. The VMS deposits of the Troodos ophiolite Phterykoudi-Potamos and Lemithou epidosite zones (Fig. 1;
Bickle and Teagle, 1992; Richardson et al., 1987; Jowitt et al.,
Most of the Troodos VMS deposits occur at the lower lava 2012). Sheeted dyke samples are variably altered from diabase to
sequence and the upper lava sequence boundary within the vol- end-member epidosite (Richardson et al., 1987; Jowitt et al.,
canic section (Constantinou and Govett, 1973; Adamides, 2010b). 2012; Cann et al., 2015). From the plutonic sequence 11 gabbros
They are generally characterised, from top to bottom, by a sedi- and 6 plagiogranites have been selected, and from the mantle
mentary cover and an oxidised horizon, implying seafloor forma- sequence 2 dunites, one hazburgite and one serpentinised
tion, underlain by a massive sulphide lens and a stockwork zone pyroxenite.
(e.g. Constantinou and Govett, 1973; Adamides, 2010a,b; Keith Fresh glasses and altered samples have been analysed using the
et al., 2016). The Apliki VMS deposit, in the Solea graben (Fig. 1), same methods in order to have a coherent and homogeneous data-
is an exhalative structurally controlled deposit characterised by base. Because fresh glasses have not been investigated by microan-
moderate tonnage and Cu grade (Table 1) and is a good example alytical techniques, their analyses most likely include minor
of Cyprus-type VMS deposit. Some deposits, such as Mavrovouni microphenocrysts, and partially devitrified zones. The elements
(Table 1, Adamides, 2010a), lack evidence for formation on the sea- investigated in this study, however, are moderately to strongly
floor and show extensive alteration halos suggesting formation by chalcophile (e.g. Peach et al., 1990; Li and Audétat, 2015; Patten
sub-seafloor replacement (Adamides, 2010a; Tornos et al., 2015). et al., 2013) and are likely to be only weakly affected by the pres-
The Troodos VMS deposits are classified as mafic type (Barrie and ence of silicate microphenocrysts. Additionally, most of these ele-
Hannington, 1999) and range in size from 50 000 t up to 16 million ments show very limited mobility during low temperature
t with average grade of 1.3 ± 1.1% Cu and 0.8 ± 0.4% Zn (Hannington seawater alteration (e.g. Alt et al., 2010; Patten et al., 2015) and lit-
et al., 1998). The Au grade is rarely reported (Table 1; Herzig et al., tle mobilisation associated with devitrification is expected. Mag-
1991; Hannington et al., 1998) and for most deposits it is relatively matic sulphide droplets, that are potentially present within fresh
low (Table 1). A notable exception is the Limni deposit that had glass samples, are important hosts for the investigated elements
3.4 ppm Au grade (Table 1), almost classifying it as an auriferous (e.g. Peach et al., 1990; Patten et al., 2016) and therefore whole
deposit (Au > 3.46 ppm; Mercier-Langevin et al., 2011). The Skouri- rock analyses are preferable to in-situ analyses of thin sections that
otissa deposit, within the Solea graben (Fig. 1), is the largest and may miss these rare phases. Thus, the glass composition for the
most recently exploited VMS deposit in Troodos (Table 1). It is sub- elements investigated in this study are interpreted as representa-
divided into the Phoukasa and Phoenix deposits (Table 1, tive of magmatic composition. Major element concentrations were
Adamides, 2010a). The Phoukasa deposit is a typical exhalative carried out by X-ray Fluorescence (XRF), using a Rigaku ZSX Primus
Cyprus-type deposit whereas the Phoenix deposit is a result of par- II Spectrometer at Stockholm University. Accuracy and precision
tial supergene alteration of a sulphide lens that occurs on the sea- was controlled through analyses of the reference material AGV-2,
floor and probably also during later subaerial environment an andesite powder provided by USGS (Appendix 2). Trace element
(Adamides, 2010a). The Au grade of Skouriotissa deposit is not analyses were carried out by laser ablation-inductively coupled
clearly defined because of the extensive supergene alteration plasma-mass spectrometer (LA-ICP-MS) of whole rock glass beads
(Table 1; Hannington et al., 1998). During seafloor weathering of at Stockholm University using a ESI NWR-193 nm excimer laser
the massive sulphides, Au has been locally enriched in gossans coupled to a Thermo Fisher XSeries 2 quadrupole mass spectrom-
and ochres which average 7 ppm Au (0.6–28.4 ppm; Herzig et al., eter. A spot size of 150 lm was used with a laser pulse frequency of
1991). Umbers, which are sedimented from hydrothermal vent 10 Hz and a laser energy density of 7 J cm2. Calibration was car-
plumes, show low initial Au content (e.g. Skouriotissa ried out using NIST SRM 612 and Si as the internal standard. Accu-
Au = 0.08 ± 0.07 ppm; Herzig et al., 1991) but late Au enrichment racy and precision were verified by analyses of external reference
associated with silicification of umber during low temperature materials BCR-2G (Appendix 2). All trace elements are reproduced
off-axis fluid circulation occurs in places (up to 2.35 ppm Au, to better than 10% accuracy except for Sc, Zn, Ga, Nb, Cs, Ta, Pb and
Table 1; Prichard and Maliotis, 1998). At Skouriotissa and other U which are reproduced to better than 15% (Appendix 2). Data
reduction and calculation of trace element concentrations were Appendix 1). Significant trace metal variation is observed between
performed off-line through IOLITE software (Hellstrom et al., and within each rock type either due to primary magmatic vari-
2008; Paton et al., 2011). ability or remobilisation during alteration. We report the average
Gold whole rock analyses have been carried out following the concentrations and standard deviation for each lithological unit;
ultra-low detection limit technique developed by Pitcairn et al. the range is reported when the standard deviation is higher than
(2006a) at Stockholm University. To minimise possible nugget the average.
effects, 3 g grams of sample powder are digested with HNO3-HF-
aqua regia. Gold is extracted from the sample solution by chro-
4.1. The fresh glasses
matography using an inert resin saturated with di-isobutyl ketone
(DIBK) and is subsequently analysed using a Thermo Fisher XSeries
The 10 fresh glass samples analysed in this study have compo-
2 ICP-MS. The 3r method detection limit calculated from blank
sition ranging between 2.95 wt% and 7.91 wt% MgO and are repre-
digests is 0.027 ppb. Analytical accuracy and precision were con-
sentative of the BADR series. The bulk of the samples have a
trolled through analyses of CANMET reference material TDB-1
basaltic-andesitic composition (n = 8; 52–57 wt% SiO2, Appendix
and USGS reference materials WMS-1 and CH-4 (Appendix 3).
1), with one basalt (52 wt% SiO2), and one andesite (60 wt% SiO2;
Arsenic, Sb and Se analyses were carried out by Hydride Genera-
Appendix 1). The concentrations of Zn, As and Sb increase system-
tion Atomic Fluorescence Spectrometry (HG-AFS) at Stockholm
atically from basaltic to andesitic composition ranging between
University following the method described in Pitcairn et al.
55–100 ppm, 359–1520 ppb and 25–150 ppb respectively
(2006b). Using the same acid digests from Au analyses hydride
(Appendix 1). The basaltic-andesite samples have the highest Cu,
generation is achieved by mixing sample solutions with a reduc-
Se and Au concentrations (96 ± 36 ppm, 120 ± 61 ppb and
tant solution (0.4% NaOH and 0.7% NaBH4) and is analysed using
2.1 ± 1.0 ppb respectively) whereas the andesitic sample shows
a PSA 10.055 Millennium Excalibur instrument. The 3r method
the lowest concentrations (29 ppm, 69 ppb and 0.77 ppb respec-
detection limits are 0.059 ppb, 0.081 ppb and 0.039 ppb for As,
tively; Appendix 1). One fresh glass sample shows anomalously
Sb and Se respectively. Analytical accuracy and precision were also
high Au concentration (24 ppb) although a more realistic value of
controlled through analyses of CANMET reference material TDB-1
2.2 ppb can be calculated from the Au/Se correlation (R2 = 0.62).
and USGS reference materials WMS-1 and CH-4 (Appendix 3). Gold
This latter value, however, is not used in calculations.
and As are reproduced to better than 10% accuracy for TDB-1 and
better than 15% for WMS-1 and CH-4 whilst Sb is reproduced to
better than 10% accuracy for WMS-1 and CH-4 and better than 4.2. The volcanic section
15% for TDB-1 (Appendix 3). Selenium is reproduced to better than
15% for CH-4 but shows poorer accuracy for TDB-1 and WMS-1 Of the 26 whole rock samples from the volcanic section, 23
(>15%, Appendix 3). samples follow the BADR series and 3 are picritic. Significant vari-
ation occurs in the volcanic section samples due to the different
styles of alteration (Table 2). Samples altered by low temperature
4. Results
fluid circulation show slightly lower average concentrations of
Au, As, Sb, Se, Cu and Pb and higher Zn concentration than fresh
Based on immobile element concentrations, most volcanic,
glass samples (Table 2). Samples from alteration halos in the
sheeted dyke and plutonic sample compositions follow the BADR
vicinity of mineralised zones, which contain variable amount of
series although several samples are picritic (MgO > 30 wt%, Fig. 2,
secondary sulphides, have the highest Se, Cu, Zn and Pb concentra-
tions (170 ± 130 ppb; 162 ppm, 4.8–1200 ppm; 120 ± 110 ppm and
5.0 ppm, 0.28–23 ppm respectively) and also high As, Sb (Table 2).
Although a few samples show relative high Au content (>2 ppb,
n = 3), samples from alteration halos in the vicinity of mineralised
zones are not systematically enriched in Au relative to the other
alteration types (Table 2). Only one sample affected by seafloor
weathering alteration has been analysed in this study (TVS-033-
KA) and shows the highest CaO, K2O, As and Sb concentrations in
the volcanic section and high LOI (23.5 wt%, 5.4 wt%, 6.5 ppm,
230 ppb and 15.5% respectively, Table 2 and Appendix 1). The other
metals have concentrations within the range of the low tempera-
ture altered samples. Picritic samples have similar compositions
to the BADR background altered samples for all metals except
Au, which has significantly higher concentrations in the picrites
(2.0 ± 0.2 ppb). This value is similar to Au analyses of fresh
boninites reported by Hamelyn et al. (1984; 1.9 ± 1.9 ppb). Picrite
Se concentrations (97 ± 72 ppb) are in the same order of
magnitude than that reported by Hamlyn et al. (1985) for boninites
(53 ± 55 ppb).
4.3. The lower sheeted dyke complex
The major element compositions of the sheeted dyke samples

are controlled by the secondary mineral assemblages (Richardson
Fig. 2. Immobile element diagram of the samples investigated in this study. Most of et al., 1987; Jowitt et al., 2012). Increase of modal epidote relative
the samples plot in the island arc tholeiite field except for the picritic samples and
the extensively silicified samples from the basal unit. Diagram from Dilek et al.
to chlorite and amphibole leads to higher SiO2 and CaO, and lower
(2007) and Pearce (2003). IAT = island arc tholeiites, WPB = withinplate basalts, MgO, Na2O and K2O (Jowitt et al., 2012; Table 2). The
Bon = boninites. average concentrations of Au (0.28 ppb, 0.07–3.7 ppb), As
Table 2
Whole rock data of the different lithological units and alteration fascies in the Troodos ophiolite.
SiO2 TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O P2O5 K2O Au As Sb Se Cu Zn Pb Y
% % % % % % % % % % ppb ppb ppb ppb ppm ppm ppm ppm
Volcanic section Fresh glass Average (n = 10) 54.8 0.99 15.1 12.3 0.20 5.33 8.68 2.15 0.04 0.47 1.9 850 78 120 86 79 1.9 26
r 2.1 0.23 0.6 1.7 0.03 1.53 1.88 0.46 0.02 0.20 1.0 370 34 54 37 11 0.56 7.6
Median 54.0 1.03 15.0 12.8 0.20 5.45 8.41 2.02 0.03 0.41 1.5 910 64 120 82 81 2.1 25
Seafloor weathering n=1 43.8 0.46 13.5 8.5 0.12 3.46 23.5 1.21 0.04 5.41 1.2 6500 240 b.d.l. 47 58 1.9 12
Low temperature zone Average (n = 10) 53.8 0.87 15.5 10.6 0.22 6.78 8.70 2.70 0.04 0.86 1.0 450 71 50 46 100 1.2 20
r 5.9 0.27 1.3 1.9 0.08 2.68 6.28 1.76 0.04 0.79 0.62 450 84 53 23 39 0.35 8.8
Median 55.1 0.86 15.8 10.2 0.22 6.86 6.46 2.37 0.04 0.76 0.95 260 35 20 41 89 1.3 17
Mineralisation zone halo Average (n = 12) 62.5 1.13 14.1 10.7 0.26 4.05 3.31 3.15 0.06 0.66 1.0 2300 140 170 160 120 5.0 15
C.G.C. Patten et al. / Ore Geology Reviews 86 (2017) 487–508

r 8.7 0.28 2.3 3.9 0.16 2.64 1.00 1.26 0.01 0.59 1.10 2400 150 130 340 110 6.4 2.5
Median 60.8 1.16 14.7 12.5 0.20 4.03 3.51 3.00 0.07 0.45 0.36 1000 87 170 22 110 2.0 15
Picrites Average (n = 3) 47.2 0.20 6.3 9.3 0.15 32.2 4.16 0.32 b.d.l 0.07 2.0 320 33 97 56 59 1.1 5.0
r 0.5 0.02 0.6 0.1 0.00 1.74 0.46 0.10 0.00 0.17 81 3.3 72 5.4 3.0 0.78 0.51
Median 47.5 0.21 6.6 9.4 0.15 31.2 4.37 0.35 0.07 2.10 310 31 55 56 59 0.63 5.3
Sheeted dyke complex Diabase Average (n = 25) 58.0 0.94 15.4 10.4 0.14 5.14 6.55 3.16 0.05 0.30 0.24 110 37 54 29 45 1.1 25
r 5.2 0.28 0.9 1.8 0.07 1.88 2.10 1.10 0.05 0.21 0.11 89 25 110 49 36 1.80 9.2
Median 55.7 0.96 15.6 10.6 0.15 5.44 6.36 3.15 0.05 0.30 0.21 86 30 10 8.4 30 0.43 24
Transitional diabase-epidosite Average (n = 18) 56.8 1.04 15.2 12.0 0.17 5.01 8.15 1.53 0.05 0.09 0.43 86 35 10 24 52 0.63 24
r 4.4 0.29 1.5 2.3 0.06 1.76 2.46 1.65 0.03 0.14 0.83 35 21 9.1 28 32 0.38 6.9
Median 56.3 0.98 15.6 11.4 0.18 5.30 7.79 0.23 0.04 0.03 0.19 91 27 6.5 12 58 0.58 23
Intermediate epidosite Average (n = 18) 61.3 0.96 14.7 9.0 0.12 2.06 10.48 1.22 0.07 0.08 0.23 86 38 22 23 34 1.10 31
r 6.6 0.35 1.7 3.2 0.08 1.40 4.31 1.49 0.03 0.13 0.12 47 17 25 22 29 1.50 9.5
Median 59.5 0.97 14.6 9.9 0.15 1.90 11.03 0.27 0.07 0.02 0.21 72 36 12 14 27 0.57 30
End-member epidosite Average (n = 6) 55.3 1.08 15.6 11.1 0.09 2.53 14.26 0.01 0.04 0.02 0.18 110 27 7.8 10 19 0.89 24
r 1.6 0.23 0.6 1.6 0.04 1.27 2.71 0.01 0.02 0.01 0.02 24 16 10 12 10 0.60 5.3
Median 55.1 1.08 15.7 10.5 0.09 2.29 15.98 0.00 0.04 0.02 0.17 120 20 2.5 5.2 14 0.63 24
All dykes Average (n = 67) 58.3 0.98 15.2 10.5 0.14 4.04 8.73 1.92 0.06 0.16 0.28 97 36 27 24 42 0.97 26
r 5.6 0.31 1.3 2.6 0.07 2.19 3.79 1.69 0.04 0.20 0.44 64 21 68 35 33 1.4 9.0
Median 57.0 0.97 15.4 10.9 0.15 3.90 7.87 2.29 0.04 0.04 0.19 88 31 8.7 11 27 0.57 24
Plutonic complex Plagiogranites Average (n = 6) 69.0 0.58 14.0 5.0 0.04 1.30 6.60 3.03 0.07 0.41 0.19 110 40 47 4.8 26 0.40 41
r 3.9 0.15 1.4 2.9 0.04 0.52 4.05 1.38 0.04 0.24 0.06 51 25 35 2.5 29 0.24 9.5
Median 69.8 0.63 14.7 4.2 0.02 1.52 6.30 3.33 0.08 0.37 0.21 120 34 47 5.6 15 0.37 40
Gabbros Average (n = 11) 50.4 0.25 16.3 8.1 0.16 10.1 13.65 1.01 b.d.l 0.07 1.30 130 23 94 69 40 0.54 6.7
r 0.8 0.07 1.7 0.6 0.02 0.98 0.95 0.22 0.04 1.30 44 11 74 37 7.9 0.30 2.5
Median 50.3 0.26 16.4 8.1 0.16 10.3 13.94 1.06 0.05 0.88 140 23 62 72 40 0.47 6.3
Mantle sequence Average (n = 4) 43.8 0.08 3.4 10.3 0.15 39.5 2.53 0.14 b.d.l 0.04 0.33 100 22 33 14 47 0.39 1.7
r 1.9 0.10 4.4 1.8 0.04 11.5 3.80 0.25 0.03 0.26 55 5.0 31 16 6.0 0.07 2.3
median 44.6 0.03 1.1 9.9 0.14 44.4 0.41 0.00 0.03 0.23 110 22 12 5.5 45 0.36 0.5
All data are in Appendix 1. b.d.l. = below detection limit.
493
(100 ± 64 ppb), Sb (36 ± 12 ppb), Se (27 ppb, 0.8–430 ppb), the Troodos ophiolite (Pearce and Robinson, 2010). Fresh glass
Cu (24 ppm, 0.60–226 ppm), Zn (42 ± 33 ppm) and Pb (0.97 ppm, compositions from different arc-related environments normalised
0.10–6.9 ppm) in the sheeted dyke samples are significantly lower to MORB values show that the Troodos ophiolite and the Manus
than in the volcanic section and fresh glass samples (Table 2). basin have similar metal abundances, whereas the Lau basin shows
Strongly chalcophile elements (Au, Se and Cu; e.g. Peach et al., metal contents closer to MORB composition (Fig. 3). The Troodos
1990; Li et al., 2012; Patten et al., 2013) and Pb show large varia- ophiolite and the Manus basin are significantly enriched in As, Sb
tions (Table 2) due to a small number of samples with higher con- and Pb relative to MORB (Fig. 3), most likely due to contamination
centrations. These anomalies are likely due to the presence of of the mantle wedge by direct input from the subducting slab,
specific secondary minerals such as pyrite. Concentrations of Au, either from the sediments or the altered crust (Pearce and Peate,
related elements and base metals are generally similar in all types 1995; Noll et al., 1996; Hattori et al., 2005; Richards, 2011;
of alteration but subtle differences may still be observed. Timm et al., 2012; Hattori and Guillot, 2003; Regelous et al.,
End-member epidosite samples show the lowest concentrations 2014). Gold concentrations in the Troodos ophiolite (0.77–4.20 ppb;
of Au (0.18 ± 0.02 ppb), Sb (27 ± 16 ppb), Se (7.8 ppb, 1.6–24 ppb), Appendix 1) are within the range of MORB compositions
Cu (10 ppm, 2.7–36 ppm) and Zn (19 ± 10 ppm; Table 2). Diabase (0.2–5.8 ppb; Jenner et al., 2010) indicating that Au may not be
samples from outside epidosite zones show similar metal contents strongly mobilised by slab contamination (Fig. 3). Copper and Zinc
to those within epidosite zones. concentrations in the different arc settings are similar to those of
MORB suggesting that these elements are not released from the
slab during subduction (Fig. 3). Finally Se shows lower concentra-
4.4. Plutonic complex and ultramafics
tions in the Troodos ophiolite (119 ± 53 ppb) relative to MORB
range (243–723 ppb; Jenner et al., 2010) suggesting a depleted
Differences in the major element compositions between the
mantle source (Fig. 3).
gabbros and the plagiogranites are shown in Table 2. Metal
After the onset of mantle melting, the composition of the
concentrations in the plutonic samples are overall higher than
magma generated is controlled by differentiation caused by frac-
those in the lower sheeted dyke section but lower than in the
tional crystallisation in the lower crust. The fresh glass samples
volcanic section. Concentrations of Au, As, Se, Cu and Zn are
in the Troodos ophiolite that represent the melts which migrated
higher in the gabbros (1.3 ± 1.3 ppb; 130 ± 45 ppb; 94 ± 73 ppb;
upwards after differentiation, range from around 9 wt% MgO for
70 ± 37 ppm; 40 ± 8.0 ppm respectively) than in the plagiogranites
the most primitive compositions to < 1 wt% MgO for the most frac-
(0.19 ± 0.06 ppb; 110 ± 51 ppb; 48 ± 35 ppb; 4.8 ± 2.5 ppm;
tionated (Fig. 4). The supra-subduction affinity of the Troodos
26 ppm, 0.8–57 ppm respectively). Conversely, Sb and Pb show
ophiolite is highlighted by the covariation of TiO2 and MgO, which
higher concentrations in the plagiogranites (40 ± 8.0 ppb and
is characteristic of the BADR series (Fig. 4; Pearce and Robinson,
0.54 ± 0.30 ppm respectively) than the gabbros (23 ± 11 ppb and
2010). During magmatic differentiation the TiO2 concentration
0.54 ± 0.30 ppm respectively, Table 2). Gabbros and plagiogranites
gradually increases until around 3.5 wt% MgO, where magnetite
show variable degrees of alteration from background altered to
crystallisation occurs leading to a subsequent plateau and rapid
partially epidotised with varying proportions of chlorite, epidote
decrease in TiO2 as differentiation continues (Fig. 4; Pearce and
and quartz, which most likely caused metal mobility but insuffi-
Robinson, 2010; Regelous et al., 2014). The magmatic evolution
cient samples have been analysed to determine it. Agiorgitis and
of the Troodos BADR series can be divided into a first pre-
Becker (1979) reported Au concentrations of 0.77 ± 0.05 ppb and
magnetite crystallisation stage (9–3.5 wt% MgO) and a second
1.4 ± 0.1 ppb for a Troodos olivine gabbro and a bronzite gabbro
post-magnetite crystallisation stage (3.5–<1 wt% MgO). Similar
respectively, which are in the same range than the average gabbro
trends can be observed in the Manus Basin data that plots within
value from this study (1.3 ± 1.3 ppb). Ultramafic samples
the BADR field, and to some degree in the Lau Basin data although
from the mantle sequence have lower Au (0.33 ± 0.26 ppb), As
due to the higher TiO2 contents these samples plot in the MORB
(100 ± 55 ppb), Sb (22 ± 5.0 ppb), Se (33 ± 31 ppb), Cu (13 ppm,
field (Fig. 4). The crystallisation of magnetite has a strong effect
2.1–41 ppm) and Pb (0.39 ± 0.06 ppm) concentrations than the
on the behaviour of S and the other metals.
gabbros (Table 2). Only Zn shows higher concentrations
(47 ± 6 ppb; Table 2). The average Au concentration in ultramafic
5.2. Behaviour of S in the primitive crust
rocks (0.19 ± 0.08 ppb) is of the same order of magnitude as a sin-
gle data reported by Agiorgitis and Becker (1979; 0.52 ± 0.1 ppb).
Most of the base metals, Au and related elements are chal-
cophile elements and are affected by sulphide segregation during
5. Discussion magmatic differentiation (e.g. Peach et al., 1990; Barnes and
Lightfoot, 2005; Jowitt, 2009; Jenner et al., 2010; Patten et al.,
5.1. Composition of the primitive crust at Troodos 2013). It is thus necessary to understand the behaviour of S during
magmatic differentiation of the primitive crust at Troodos and the
Since Miyashiro (1973) highlighted the arc-related nature of the possible implications on metal distribution. Within the BADR ser-
Troodos ophiolite, the primitive crustal composition has been ies, S shows a similar trend to TiO2 during differentiation with
extensively investigated using fresh glass samples (Rautenschlein incompatible behaviour during the first stage (9–3.5 wt% MgO)
et al., 1985; Hamlyn et al., 1985; Flower and Levine, 1987; and compatible behaviour during the second stage (3.5–<1 wt%
Mehegan, 1988; Muenow et al., 1990; Bednarz et al., 1994; MgO, Fig. 4; Jowitt, 2009). The average S concentration in Troodos
Portnyagin et al., 1997; Jowitt et al., 2012; Appendix 4). It is now BADR series fresh glasses (Table 3; 459 ± 311 ppm; Muenow et al.,
accepted that Troodos ophiolite formed in a supra-subduction 1990; Bednarz et al., 1994; Jowitt, 2009) is within the range
setting probably during subduction initiation (e.g. Pearce and estimated by Alt (1994) for primary magmatic values (200–
Robinson, 2010). Subduction and release of aqueous fluids has a 930 ppm). The incompatible behaviour of S during the first stage
profound impact on geochemistry and melt generation in the of magmatic differentiation suggests that most of S is preserved
mantle wedge (e.g. Pearce 1982; Pearce and Peate, 1995), and gen- in the melt. The sudden drop of S at MgO values of around 3.5 wt
erates significantly different magma compositions compared to % indicates that at this stage of differentiation S starts to be effi-
MORB. The Manus and Lau basins have been suggested to repre- ciently removed from the melt. The two most likely processes to
sent modern-day supra-subduction oceanic crust analogues to remove S from the melt are 1) sulphide saturation and 2) magmatic
Fig. 3. a) Median concentration of Troodos ophiolite, Manus Basin and Lau Basin fresh glass data normalised to median MORB values from Jenner et al. (2012). The solid black
line corresponds to the average MORB values determined by Arevalo and MacDonough (2010) and the grey shaded area corresponds to the MORB range determined by the
upper and lower quartile of Jenner et al. (2012) database. b) Discriminant diagrams for As, Sb and Pb subduction contamination of volcanic arc mantle wedge. Modified from
Pearce and Peate (1995). Troodos data are from this study and Regelous et al. (2014); MORB data are from mid-oceanic ridges and seamounts samples from Jenner et al.
(2012); Manus Basin data are from Jenner et al. (2012), Sun et al. (2004, 2007); and Lau Basin data are from Jenner et al. (2012).
degassing (Kamenetsky et al., 2001; Yang and Scott, 2002; Sun calculated ranges between 437 ppm and 1071 ppm S, which is
et al., 2004; Jenner et al., 2010, 2012, 2015; Park et al., 2015; Sun below the range estimated for arc-related magmas by Jugo
et al., 2015). (2009) but higher than most of the S concentrations in the fresh
Arc-related magmas have high oxygen fugacity inherited from glasses (Fig. 4) and within the range estimated by Alt (1994);
their mantle source, which has been oxidised by fluids from the (200–1000 ppm S). During the first stage of magmatic differentia-
subducting slab, with S present as sulphate rather than sulphide tion the S concentration is lower than the SCSS, but as S is
(e.g. Mungall, 2002; Wallace and Edmonds, 2011; Richards, 2011, concentrated in the melt it eventually crosses the SCSS at MgO val-
2015). In Troodos ophiolite the oxygen fugacity averages FMQ + 1.5 ues of 4 to 3.5 wt% MgO (Fig. 4) concomitant with magnetite
compared to MORB (FMQ = 0; Coogan et al., 2003). Jugo (2009) crystallisation. This would cause sulphide segregation which
calculated that the sulphur content at sulphide saturation (SCSS) would remove S from the melt during the second stage of mag-
for arc-related magmas with FMQ + 0.5 < fO2 < FMQ + 1.7 should matic differentiation (Fig. 4). Despite the generally sulphide
have a SCSS between 1500 and 4500 ppm S. Using Bednarz et al. under-saturated nature of Troodos magma, primary magmatic sul-
(1994) fresh glass data and the Liu et al. (2007) method, the SCSS phides are reported (Alt, 1994) in the sheeted dykes and gabbros.
is calculated for the Troodos ophiolite (Fig. 4). The Liu et al. Only one fresh gabbro sample (10.8 wt% MgO) from this study
(2007) method is appropriate for Troodos SCSS calculation as it hosts unequivocal primary magmatic sulphides characterised by
predicts the SCSS in melts with a large composition range, high partially leached pyrrhotite with pentlandite exsolution and
water content (up to 9 wt% H2O) and high fO2 (up to FMQ + 2). chalcopyrite (e.g. Peach et al., 1990; Patten et al., 2012; Fig. 5). This
The liquidus temperatures used for the calculation are estimated suggests that in general the Troodos magma was sulphide under-
using the PELE 7.3 algorithm (Boudreau, 1999) and a pressure of saturated (Jowitt, 2009), but that locally sulphide saturation was
887 bar (assuming 3.5 km water depth; Keith et al., 2016; and a reached. This is also supported by the compositions of a few glass
magmatic chamber at 2 km extrapolated from CY1, CY1A and samples which have S contents higher than the SCSS during the
CY4 drill cores; e.g. Vine and Smith, 1990; Alt, 1994). The SCSS first stage of magmatic differentiation (Fig. 4).
Fig. 4. Fresh glass TiO2, S, Cu, Zn and Pb vs MgO plots showing the magmatic differentiation trends in the Troodos ophiolite, Manus Basin and Lau Basin. The sulphur content
at sulphide saturation (SCSS) is calculated for each sample from Bednarz et al. (1994) using the Liu et al. (2007) method. Dashed red curves are the SCSS ± the root mean
square deviation (RMSD). Sulphur in the Manus Basin (*S) is recalculated using Se as a proxy (Jenner et al., 2010). See text for details. Troodos data are from this study,
Rautenschlein et al. (1985), Flower and Levine (1987), Mehegan (1988), Muenow et al. (1990), Bednarz et al. (1994), Portnyagin et al. (1997), Jowitt (2009), Jowitt et al. (2012)
and Regelous et al. (2014). Manus Basin data are from Moss et al. (2001), Sun et al. (2004), Park et al. (2010), Jenner et al. (2010, 2012). Lau Basin data are from Cawood and
Fryer (1994) and Jenner et al. (2012). See Appendix 4 for details about the database.
References: 1-this study; 2-Rautenschlein et al. (1985); 3-Mehegan (1988); 4-Jowitt et al. (2012); 5-Muenow et al. (1990); 6-Bednarz and Schmincke (1994); 7-Portnyagin et al. (1997); 8-Flower and Levine (1987); 9-Jenner et al.
In the Manus Basin, primary magmatic S (calculated using Se as
References
a proxy; Jenner et al., 2010) shows similar behaviour to that
observed in Troodos, with incompatible behaviour until 3 wt%
9–13
9,14
1–8
MgO, after which S concentrations significantly decrease (Fig. 4).
15
9
In the Manus Basin this has been interpreted to be caused by mag-
netite crystallisation, which causes a significant decrease in the fO2
(ppm)
1.93
2.15
5.98
4.38
0.56
5.10
0.64
0.43
0.55
0.45
0.47
0.57
0.50
of the magma, triggering sulphur reduction. Sulphur reduction
Pb
leads either to sulphide saturation and segregation of a sulphide

phase (Jenner et al., 2010, 2012, 2015; Park et al., 2015) or the
(ppm)
exsolution of hydrosulphide complexes in cogenetic magmatic flu-

101
104
100
Zn
82
83
16
95
34
94
90
80
20
80
ids (e.g. Sun et al., 2004; Sun et al., 2015). Similar processes are
observed in the Lau Basin, albeit at higher MgO values of around
(ppm)
(2012); 10-Moss et al. (2001); 11-Park et al. (2010); 12-Sun et al. (2004); 13-Sun et al. (2007); 14-Cawood and Fryer (1994); 15-Arevalo and McDonough (2010). See Appendix 1 and 4.
5 wt% (Jenner et al., 2012).
134
136
Cu
62
51
57
72
88
32
43
88
26
79
70
Several lines of evidence suggest that magmatic fluid exsolution
occurred in the Troodos ophiolite. High-temperature brines in fluid
(ppb)
inclusions from late-stage gabbros and plagiogranites in Troodos

120
120
230
220
390
370
360
350
210
54
65
69
67
Se
suggest that magmatic fluids have exsolved during magmatic dif-

ferentiation although they undergo limited migration within the
(ppb)
plutonic rocks (Kelley et al., 1992). Additionally, negative d34S val-

78
34
64
87
35
17
16
13
14
14
80
20
10
Sb
ues of pyrite from the Skouriotissa VMS deposit have been inter-
preted as a possible signature of magmatic origin for some of the
(ppb)
2500
2300
S trapped within the deposit (Keith et al., 2016). Volatile contents

850
370
910
890
290
140
240
230
290
170
110
As
of fresh glass samples show that degassing of CO2 occurs during

differentiation at 4 wt% MgO, whereas H2O and Cl are preserved
(ppb)
0.97
in the magma (Fig. 6). The behaviour of S is suggested to be more

1.9
1.5
3.1
2.2
3.4
4.3
3.5
3.1
1.5
1.2
1.1
1.2
Au
similar to that of H2O rather than CO2 during degassing in a high

temperature magma (Wallace and Edmonds, 2011), suggesting
(ppm)
1110
1120
1100
that S may be preserved in the magma despite CO2 loss. This is dis-
498
445
194
666
685
826
346
194
303
690
S
tinct from the Manus basin where both degassing of CO2 and H2O is
observed (Kamenetsky et al., 2001). It is likely that some S loss
P2O5
0.31
0.22
0.14
0.15
0.12
0.11
0.15
0.14
0.13
0.30
0.20
0.07
0.10
occurred during magmatic degassing in the Troodos ophiolite but

(%)
it is not interpreted to be the main cause for the S drop observed

in fresh glass samples. Sulphide segregation is most likely to be
1.31
1.34
0.59
0.16
0.13
0.12
0.18
0.26
0.12
0.18
0.09
0.06
0.08
K2O
(%)
the main mechanism for S removal from the magma at 3 wt%

MgO in the Troodos ophiolite. The sulphide under-saturated nature
of Troodos magma during the first stage and subsequent sulphide
Na2O
2.26
2.28
3.86
4.35
2.55
2.57
2.53
2.51
2.71
0.82
1.04
0.55
0.45
(%)
segregation during the second stage of magmatic differentiation

both have important effects on the distribution of metals in the
oceanic crust before initiation of hydrothermal alteration.
8.92
2.47
8.56
5.53
3.25
11.6
1.57
11.7
11.6
11.6
11.5
4.60
0.87
CaO
(%)
5.3. Metal behaviour in the primitive oceanic crust

MgO
2.35
3.95
2.84
2.82
1.77
7.53
1.56
7.96
7.52
7.59
7.84
4.70
0.97
(%)
Despite the limited number of fresh glass samples analysed in

this study (n = 10) the composition of the primitive crust at Troo-
MnO
0.17
0.17
0.15
0.14
0.18
0.18
0.19
0.19
0.17
0.03
0.10
0.04
0.02
dos for the elements of interest can be investigated using previ-

(%)
ously published data and modern-day analogues. Like S, the

strongly chalcophile elements Cu, Au and Se show incompatible
9.21
1.73
6.56
2.18
6.21
1.46
9.07
10.4
10.1
10.2
1.40
FeO
9.9
9.4
(%)
behaviour during the first stage of magmatic differentiation in

Average composition of Troodos ophiolite and MORB fresh glass.
the Troodos samples, followed by compatible behaviour during

Al2O3
the second stage (Figs. 4 and 7). In the case of Cu, the change in
14.8
14.7
14.5
14.5
15.6
15.7
15.3
15.2
15.5
1.1
1.3
1.1
1.0
(%)
behaviour occurs at 4 wt% MgO rather than 3.5 wt% MgO, but this
is assumed to be within error of magnetite crystallisation and asso-
TiO2
ciated to sulphide segregation. Of the samples analysed for Au and

1.31
1.21
1.46
1.43
1.42
0.38
0.83
0.68
0.65
0.39
0.44
0.90
0.20
(%)
Se, only one sample has an MgO value below 3.5 wt% but low con-
centrations for both Au and Se in this sample point to a decrease in
SiO2
55.5
62.8
65.1
49.8
55.0
50.2
50.5
50.6
50.5
concentrations of these elements after sulphide segregation.

3.5
6.9
2.4
0.9
(%)
Strongly chalcophile elements (Au, Se and Cu) show strikingly sim-

ilar behaviour in the Manus Basin where the drop in Cu, Au and Se
Average
Average
Average
Average
Median
Median
Median
Median
concentrations match those of S and TiO2 at 3 wt% MgO (Fig. 7;

Jenner et al., 2010). The decrease in concentrations of these ele-
r
ments has been attributed both to volatile loss (Kamenetsky

et al., 2001; Yang and Scott, 2002; Sun et al., 2004; Sun et al.,
Manus Basin
2015) and to sulphide segregation (Jenner et al., 2010, 2012,

Lau Basin
Troodos
MORB
MORB
2015). The poorly chalcophile elements Zn, Pb, As and Sb are

Table 3
incompatible throughout magmatic differentiation in Troodos sug-

gesting that they are not significantly affected by magnetite and
Fig. 5. Primary magmatic sulphides in gabbroic sample TM-14-LE. Only one sample from this study hosts indubitably magmatic sulphide. Po = pyrrhotite, Pn = pentlandite,
Cpy = chalcopyrite, Py = pyrite. The spherical shape and the pentlandite exsolutions within the pyrrhotite are characteristic features of magmatic sulphides (e.g. Patten et al.,
2012). Primary magmatic sulphides are partially leached and initial limits are represented by white dashed lines. Chalcopyrite and pentlandite are more resistant to leaching
than pyrrhotite. Pyrite is secondary.
sulphide segregation. Their behaviour in the Manus Basin and the directly comparable to the fresh glass samples that represent the
Lau Basin is similar except for Zn in the Manus Basin which shows extracted melt. To observe the primary magmatic composition of
weak compatibility during the second stage of magmatic differen- the samples, Y is used rather than MgO and TiO2, as Y has been
tiation (Fig. 7). shown to be immobile during hydrothermal alteration and unaf-
Because of the sulphide under-saturated nature of the magma fected by magnetite crystallisation during magmatic differentia-
in the Troodos ophiolite, Cu, Au and Se behave as incompatible ele- tion (Fig. 7; Jowitt et al., 2012; Patten et al., 2015). A series of
ments during the first stages of magmatic differentiation. The high- differentiation curves have been determined for each element
est concentrations of these metals occurs immediately prior to during magmatic differentiation (Jowitt et al., 2012):
magnetite and sulphide segregation at 3.5–5 wt% MgO (basalt-
andesite). This is clearly distinct from MORB, where the highest
Ee ¼ AY B ð1Þ
concentrations of these metals occur in more mafic magma (e.g. where Ee is the estimated element concentration for a given Y con-
Jenner et al., 2010, 2012). During the second stage of magmatic centration and A and B are regression coefficients (Appendix 5). For
differentiation (andesite-dacite), Troodos magmas have lower Cu elements showing a change in behaviour during magmatic differen-
concentrations, and also most likely lower Au and Se concentra- tiation (Ti, S, Cu, Zn, Au and Se) two differentiation curves are deter-
tions, but higher Zn, As, Sb and Pb concentrations (Fig. 4 and 7). mined (Appendix 5). Due to lack of fresh glass samples below 3.5 wt
The metal behaviour during magmatic differentiation described % MgO, Au and Se differentiation trends below this value are extrap-
above shows that the quantities of different metals available for olated from the Manus Basin samples (Appendix 5). The error
hydrothermal fluid mobilisation is strongly controlled by the associated between the fresh glass data and the differentiation
primary magmatic composition. curves is determined using the root mean square deviation (RMSD,
Appendix 5; Jowitt et al., 2012):
6. Mass variation pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
RMSD ¼ RðEc EmÞ=n ð2Þ
The systematic changes in metal concentrations in the different with Ec the calculated element concentration, Em the measured ele-
rock types and alteration zones of the Troodos ophiolite indicates ment concentration and n the number of values used to calculated
that significant metal mobilisation has occurred. To quantify the differentiation curves. Mass variations are estimated from the
mobilisation, whole rock values are compared to the primitive deviation of a sample from the differentiation curves (Jowitt et al.,
crustal compositions determined from fresh glass samples. Mass 2012):
variations between samples from the volcanic section and the
% Variation ¼ ðEs Ee Þ=Ee 100 ð3Þ
sheeted dyke complex are calculated. The plutonic complex cannot
be included in the mass balance as these samples are the crys- where Esis the measured value and Ee the estimated primary crust
tallised residue of magmatic differentiation and are thus not composition from Eq. (1). Mass variations assume minor absolute
in the mass change value for Al2O3 is caused mainly by a small

number of extensively silicified samples (n = 3; Appendix 1) in the
basal unit of the volcanic section.
6.1. Mass variation in the volcanic section
The volcanic section has been affected by multiple styles and

generations of alteration (e.g. Gillis and Robinson, 1990). The
mobility of metals relative to each type of alteration can be esti-
mated, but accuracy is limited by low sample numbers in some
alteration types. The only sample altered by seafloor weathering
has the highest As, K2O and Rb concentrations of the volcanic sec-
tion (6.47 ppm, 5.4 wt% and 35.6 ppm respectively) and shows
high Sb and Cs concentrations (236 ppb and 385 ppb respectively).
These values are significantly higher than those of fresh glasses
(Table 2). This is comparable to the modern day oceanic crust
where K2O, Rb and Cs (Staudigel, 2003) as well as As and Sb
(Patten et al., 2015) are known to be enriched in the volcanic sec-
tion during seafloor weathering. In the low temperature altered
samples, concentrations of Au and related elements are of the same
order of magnitude as that of fresh glass (Table 2), suggesting little
bulk mobilisation although the larger range in concentrations indi-
cates local remobilisation has occurred. Within altered samples
proximal to mineralisation, significant enrichment of metals
occurs. Variations in the metal contents in these samples are most
likely due to variations in the extent and conditions of alteration
within this group of samples. In summary, the mobilisation of
metals in the bulk of the volcanic section is relatively limited.
Exceptions include enrichments in As and Sb during seafloor
weathering and metal enrichment in alteration halos proximal to
mineralised zones.
6.2. Mass variation in the lower sheeted dyke section
6.2.1. Epidosite zones

Mobility of metals from the lower sheeted dyke section is sig-
nificant in both epidosite zones and background altered diabase
(Fig. 8). From the Spilia epidosite zone, Jowitt et al. (2012) esti-
mated relative proportions of each sub-class of epidosite alteration
with 24% diabase, 24% transitional diabase-epidosite, 39% interme-
diate epidosite and 13% end-member epidosite. Assuming the
same proportions for a hypothetical epidosite zone, the metal
depletions are 88 ± 16% Au, 89 ± 23% As, 60 ± 12% Sb,
91 ± 20% Se, 84 ± 18% Cu, 63 ± 9% Zn and 60 ± 8% Pb (Table 4).
These degrees of depletion are significantly higher than those from
modern-day fast-spreading ridge oceanic crust where epidosite
zones are not encountered and depletions do not exceed 50%
Fig. 6. Volatile element concentrations in Troodos fresh glass. a) CO2 vs MgO, (ODP Hole 1256D, Patten et al., 2015). Although mass variation cal-
degassing of CO2 at 4% MgO. b) and c) H2O and Cl vs MgO, no apparent degassing culation for altered plutonic rocks cannot be properly constrained,
of H2O and Cl seems to occur during differentiation. Blues crosses are from Bednarz three epidotised plagiogranites show significant depletions of over
et al. (1994) and blue squares are from Regelous et al. (2014).
60% for all the elements of interest. The metals show differences in
degrees of mobility between the different alteration sub-groups
within the epidosite zones. Strongly chalcophile elements (Au, Cu
mass change during the different styles of hydrothermal alteration and Se) are overall more strongly depleted (>85%) than poorly chal-
that occurred in the Troodos ophiolite and that the primary Y con- cophile elements (Sb, Zn and Pb; e.g. Peach et al., 1990; Li et al.,
centrations have remained constant during alteration. This assump- 2012; Patten et al., 2013) that show less than 70% depletion. These
tion can be evaluated using the method described in Jowitt et al. differences are a function of the different host minerals and the
(2012), where Al2O3, which is immobile during alteration but com- effects of secondary mineralogy. Because of the sulphide under-
patible during magmatic differentiation, is used as a proxy for bulk saturated nature of the magma during the first stage of magmatic
mass change. A discrepancy between the calculated Al2O3 concen- differentiation, the strongly chalcophile elements in fresh dykes
trations and the measured ones would indicate bulk mass change. are likely distributed within rare sulphide blebs and in the silicate
Average mass variation of Al2O3 is 0.02 ± 1.55 wt% for samples matrix (see below). The highly reactive response of these phases to
from the volcanic section and the sheeted dyke complex which high temperature alteration, leads to efficient metal mobilisation
suggests that little absolute mass variation occurred in these units and depletions of over 85%. Rare secondary sulphides are observed
during the different alteration stages and it has minimal effect on in sheeted dykes within epidosite zones indicating sulphide
the metal budget (Appendix 5; Jowitt et al., 2012). The variation precipitation during the waning stage of hydrothermal fluid
Fig. 7. Behaviour of Au and related elements during magmatic differentiation in the Troodos ophiolite, the Manus Basin and the Lau Basin. A drop in Au and Se is
observed for the Troodos ophiolite and the Manus Basin after magnetite crystallisation. Arsenic shows the lowest R2 value (<2) which could be due to partial
devitrification as it is a highly mobile element during low temperature seawater alteration (Patten et al., 2015). This could increase the normalised root mean square of
deviation (NRMSD = 21.2%, Appendix 5). Manus Basin data are from Jenner et al. (2012) and Sun et al. (2007). Each value of Sun et al. (2007) data is the median value of
three replicates instead of the average (Appendix 4). Lau Basin data from Jenner et al. (2012). Differentiation curves for Au and Se in the Troodos ophiolite below 3.5 wt%
MgO are extrapolated from the Manus Basin data. Regression coefficients and NRMSD are given in Appendix 5. Gold value marked by ⁄ is recalculated based on Au/Se
relationship, see text for details.
circulation possibly explaining why some samples are less 2015). Jowitt (2009) argues that unlike the epidosite zones, metal
depleted in metals. The poorly chalcophile elements, Sb, Zn and depletion in background altered diabase is not systematic as loca-
Pb are most likely hosted by silicate or oxide minerals (e.g. Doe, lised metal redistribution occurs within the sheeted dyke complex.
1994; Jowitt et al., 2012; Patten et al., sub), which are less reactive Metals are locally enriched in the upper sheeted dykes and transi-
to alteration than the rare sulphide blebs and silicate matrix, pos- tional zone in Troodos (Alt, 1994), as is observed in the modern-
sibly explaining the lower depletions of these elements. The degree day oceanic crust (e.g. Alt et al., 1989, 2010; Teagle et al., 2006;
of metal depletion is also partly controlled by the extent to which Patten et al., 2015), but the extent of this mineralisation is minor
metals can be incorporated into the secondary minerals. For exam- in comparison to the degree of depletions from the lower sheeted
ple Zn, Ni and Mn that are less strongly mobilised have been shown dyke complex. We suggest that due to the large volume of these
to be hosted in chlorite and amphibole (Jowitt et al., 2012). Where rocks and the degree to which they are depleted in metals, the
these two minerals are more abundant, such as in the diabase and background altered diabase represents an equally important
transitional diabase-epidosites, Zn and also Sb depletions are less source of metals as the epidosite zones.
strong (Table 4; Jowitt et al., 2012). Arsenic is also poorly chal-
cophile but it is very strongly depleted possibly due to its highly 6.2.3. Formation of VMS deposits in the Solea graben
mobile nature during rock alteration (see previously; Noll et al., The Solea graben is an extensively studied locality that provides
1996; Hattori et al., 2005; Hattori and Guillot, 2003; Patten et al., a unique overview of the link between source areas and associated
2015) and limited incorporation into secondary minerals. ore deposits (Varga and Moores, 1985; Richardson et al., 1987;
Schiffman et al., 1987; Schiffman and Smith, 1988; Bettison-
6.2.2. Background altered diabase Varga et al., 1992). It is characterised by a north-northwest trend-
Mass variations of Au and related elements in background ing graben axis and is bounded by sedimentary cover to the north
altered diabase from outside the epidosite zones is similar to those and by the plutonic section to the south (Fig. 1; e.g. Granot et al.,
in the diabase within epidosite zones (Table 4). This indicates that 2006; Adamides, 2010a). The Solea graben hosts some of the lar-
mobilisation of Au and related elements is not restricted to gest VMS deposits of Troodos ophiolite such as the Skourioatissa,
epidosite zones but occurs throughout the lower sheeted dyke sec- Apliki and Mavrovouni deposits (Table 1; e.g. Adamides, 2010a).
tion. Significant loss of S (Alt, 1994) and Cu (Baragar et al., 1989) Within the sheeted dyke complex, three epidosite zones are recog-
from the lower sheeted dyke complex of Hole CY-4, which recovers nised with the largest in the lower to middle part of the sheeted
background altered diabases, are in agreement with the mobilisa- dyke complex covering an area of 25 km2 (Schiffman et al., 1987;
tion of metals throughout the lower sheeted dyke section. Signifi- Richardson et al., 1987; Schiffman and Smith, 1988; Bettison-
cant metal mobility has also been reported from the lower sheeted Varga et al., 1992). Within this central epidosite zone two areas
dyke section in the modern-day oceanic crust where significant of 8 km2 are pervasively altered to epidosite (Bettison-Varga
epidosite alteration has not been observed (e.g. ODP Hole 504B, et al., 1992). The two others epidosite zones are smaller, occurring
Alt et al., 1989; ODP Hole 1256D, Alt et al., 2010; Patten et al., at the base of the sheeted dyke complex in contact with the
Fig. 8. Whole rock concentrations of Au, As, Sb, Se, Cu, Zn, and Pb vs. Y. Solid black lines are calculated differentiation curves from fresh glass data shown in Figs. 3 and 6
represented by the blues shaded areas; regression coefficients are given in Appendix 5. Magnetite crystallisation occurs at 35 ppm Y. Dashed differentiation curves for Au
and Se above 35 ppm Y are extrapolated from the Manus Basin dataset. Dashed differentiation curves for Cu and Zn are from Jowitt et al. (2012). Mass variation corresponds to
the difference between sample whole rock values and the calculated differentiation curve. See text for details.
plutonic complex and at the top of the sheeted dyke complex
0.83
0.86
0.71
0.63
0.81
0.85
0.46
0.16
0.47
0.41
0.45
0.70
(Bettison-Varga et al., 1992). Both the ore deposits and the
1.0
ppm
0.12
1.37
1.90
0.52
0.73
1.6
1.4
Pb
epidosite zones occur to the west of the Solea graben axis

(Schiffman and Smith, 1988). Ore deposits within the Solea graben
ppm
46
57
42
52
34
29
59
68
67
70
46
56
55
occur within a surface area of 17.5 km2 (5 3.5 km2; Adamides,
7.7
Zn
38
45
36
36
14
39
2010a). Assuming a section of the oceanic crust of similar surface
ppm
57
74
54
74
55
71
52
72
75
83
55
73
74
area and 2 km depth for the volcanic section and the sheeted dyke
Cu
42
55
37
36
21
43
16
complex (extrapolated from CY-1, CY-1A and CY-4; Malpas and
103
Williams, 1991), a simplified hydrothermal cell system within
81
95
51
92
95
92
86
99
97
79
94
96
ppb
116
68
19
29
18
21
70
Se
the Solea graben can be modelled and the metal mobility quanti-
fied (Fig. 9). The source area of the Solea graben VMS deposits is
37
36
32
29
51
47
42
37
39
40
40
40
40
ppb
7.5
34
28
31
32
26
35
Sb
interpreted as both the epidosite zone and the background altered

diabase within the lower sheeted dyke section. Assuming a thick-
628
663
614
635
581
795
804
607
679
622
679
680
606
ppb
225
163
141
218
178
210
210
ness of 625 m for the lower sheeted dyke section (half of the
As
sheeted dyke complex estimated at 1250 m thick; Alt, 1994;

1.5
1.4
1.3
1.6
1.7
1.4
1.5
1.3
1.4
1.2
1.3
1.6
1.3
0.27
0.48
0.68
0.37
0.71
0.70
Vine and Smith, 1990) the total volume of the source area in the
ppb
1.0
Au
Solea graben is 10.9 km3. Using a surface area of 8 km2 for the epi-
dosite zones (Bettison-Varga et al., 1992) and a thickness of 625 m,
33
59
21
58
59
24
64
26
55
48
58
60
50
106
81
34
98
53
31
Pb
the volume of the epidosite zone is 5 km3. The volume of the back-
%
ground altered diabase is calculated to be the total volume of the

71
45
68
38
32
62
72
77
84
42
64
63
50
Zn
41
47
39
37
13
35
%
lower sheeted dyke section minus the volume of the epidosite

9
zone, which is 5.9 km3 (Fig. 9). Quantities of metals mobilised from
66
87
62
90
63
85
64
77
88
93
57
86
84
Cu
46
64
42
36
13
44
18
these volumes are calculated using the metal depletions shown in

%
Table 4. The mass balance shows that 47 ± 9 t Au, 21 ± 5 kt As,

75
92
47
85
95
79
89
92
98
90
91
90
90
104
61
22
11
20
Se
1200 ± 200 t Sb, 3100 ± 700 t Se, 2.3 ± 0.5 Mt Cu, 1.8 ± 0.25 Mt Zn
%
and 27 ± 4 kt Pb has been removed from the 10.9 km3 altered

45
59
49
57
41
50
54
65
57
71
66
60
137
54
26
23
14
12
50
30
Sb
source area in the Solea graben (Table 5). It should be emphasised

%
that these calculations are made for a simplified hydrothermal cell

87
88
85
89
88
87
91
83
82
83
86
89
90
14
23
system and they are an estimate of the actual mass of metal

As
%
mobilised.
88
84
86
71
88
79
87
90
76
87
88
80
90
Au
Comparing available data for tonnages of Au and related ele-

31
56
19
18
16
%
ments in the Solea graben VMS deposits (Table 1), with the quan-
Average (n = 63)
Average (n = 17)
Average (n = 18)
Average (n = 20)
tity of metal mobilised from the lower sheeted dyke complex

Average (n = 6)
allows estimation of metal trapping efficiencies (Table 5). How-

ever, the accuracy of these trapping efficiencies, is poor due to
Average
Median
Median
Median
Median
Median
Median
Median
Error
the limited constraints on the Au and related element concentra-

r
tions in the VMS deposits (Table 1). Gold, As, Se and Cu show mod-
erate trapping efficiencies within VMS deposits (10 to 35%)
Transtional diabase-epidosite (n = 1)
whereas <10% of Zn and Sb are trapped. These trapping efficiencies

are higher than those determined for fast spreading mid-oceanic
Transitional diabase-epidosite
ridge VMS deposits for which <1% of the metals mobilised from
Mass variation of samples from the lower sheeted dyke section of the Troodos ophiolite.
the lower oceanic crust are trapped as VMS deposits (Patten

End-member epidosite
Intermediate epidosite
et al., 2015). These differences in trapping efficiency are either

due to better trapping mechanisms or to an underestimation of
Calculated depletion of epidosite zone
Diabase (n = 5)
the source area size. Ohmoto (1996) estimated a source area for
Kuroko-type VMS deposits of 40 km3 which is significantly larger
Diabase
than the one estimated in this study (10.9 km3). It is important

to emphasise that the majority of the mobilised metal is lost from
the system. Several mechanisms can lead to poor metal trapping in
Greenschist diabase
VMS deposits, such as early metal precipitation as fluids rise, unfo-

cused fluid discharge, off-axis metal mobilisation, poor trapping
Epidosite zone
efficiency during mineralisation and metal mobilisation during

Errors are propagated as detailed in Appendix 6.
VMS seafloor weathering. The bulk of metal that is not trapped

All dykes
in VMS deposits is likely vented in the hydrothermal plume either

then precipitating as fine grained sediments such as umber, or
being released into seawater. Metal bearing umbers are common
Lower Sheeted Dyke Complex
in the Troodos ophiolite (Herzig et al., 1991; Prichard and

Maliotis, 1998), some of have been enriched in Au during late-
stage off axis silicification (Prichard and Maliotis, 1998). Zone
refining and seafloor weathering of VMS deposits has an important
impact on metal redistribution within VMS deposits (e.g. Phoenix
deposits; Herzig et al., 1991; Adamides, 2010a; Keith et al.,
2016). The Cu grade of the Phoenix deposit (0.5 wt% Cu), which
formed by seafloor weathering, is lower than that of the parent
Table 4
Skouriotissa deposit (Adamides, 2010a), suggesting that Cu might

have been lost during zone refining and/or seafloor weathering
Fig. 9. Hydrothermal system of the Solea graben. a) The source areas of the Solea graben VMS deposits are the epidosite zone and the background altered diabase. They are
constrained in the lower sheeted dyke complex. Extension of the epidosite zone and of the end–member epidosites is an assumption. ⁄For more clarity the graphical
expression of depth is exaggerated by a factor 2. Upward arrows represent discharging fluids whereas downward pointing arrows represent recharge fluids. Their distribution
is schematic. b) Quantity of mobilised metals from the lower sheeted dykes and percentages of metals trapped within VMS deposits or lost during mobilisation. Errors on the
quantity of metal mobilised are from the mass balance calculations in Appendix 6. Errors on the percentages of metal trapped or lost are minimum estimates as the actual
metal tonnage of the VMS deposits is not precisely constrained. LSDC = lower sheeted dyke complex; USDC = upper sheeted dyke complex. See text for details.
Table 5
Mass variation from the Solea graben source areas and comparison with VMS deposits.
Volume Density Mass Au (t) As (kt) Sb (t) Se (t)* Cu (kt) Zn (kt) Pb (kt)
(km3) (kg m3) (Mt)
Solea graben source area Epidosite zone 5 3095 15,475 D 23 11 610 1500 1100 850 13
r 4.2 2.8 120 330 240 120 1.8
Greenschists diabase 5.9 2847 16,904 D 24 11 620 1600 1200 950 14
r 4.5 2.8 120 350 260 130 1.9
Total source area 10.9 2961 32,379 D 47 21 1200 3100 2400 1800 27
r 8.7 5.5 240 680 510 250 3.7
Solea Graben VMS deposit Phoukasa 5 0.59 0.40 16.2 221.4 124.2 3.24
Phoenix 15 1.65 75
Mavrovrouni 15 4.5 2.55 615 570 75
Apliki 2 0.1 29.7
Total VMS 37 6.8 2.95 16.2 836.4 798.9 78.2 n.d.
Percentage metal trapped in VMS 15 14 1.4 27 33 4.3
Percentage metal lost 85 86 99 73 67 96
*
Se values of the Phoukasa and Mavrovrouni deposits in Table 1 appear excessively high relative to other Troodos VMS deposits and an average of 41 ppm is used for
calculation from the average of other VMS deposits (Table 1). D = mass variation; r = propagated error (Appendix 6); n.d. = not determined.
(Table 1). Finally, loss of metals during transport from the source 6.3. The role of primitive crustal composition in the formation of VMS
area to the surface can occur by formation of mineralised zones deposits
as observed mostly in the volcanic section but also in the upper
sheeted dyke section (Fig. 8, e.g. Gillis and Robinson, 1988). The The quantity of metal available for mobilisation during
main implication from the mass balance of metal mobilisation in hydrothermal alteration of a supra-subduction oceanic crust is
the Solea graben is that large masses of metal are mobilised in strongly controlled by its primary composition. Strongly chal-
the lower sheeted section from both epidosite zones and back- cophile metal concentrations (Au, Cu and Se) in the primary crust
ground altered diabase, and although a significant portion of the are at their maximum at around 3–5 wt% MgO just before mag-
metals mobilised at depth is trapped as VMS deposits, most is lost netite crystallisation and S reduction, and therefore crust of this
by several other processes. composition has high fertility for these metals (Fig. 10). Crust with
Fig. 10. Cumulative frequency plots of the Troodos fresh glass, Manus Basin fresh glass and Troodos source area databases. Hypothetical cumulative frequency of the Manus
Basin source area is calculated assuming as similar variation between fresh glass samples and source area samples than in Troodos ophiolite. Median values are shown and
light shaded boxes represent the upper and lower quartile. Cumulative frequency plots are compared to primitive crust metal composition, the Troodos area has high fertility
for Cu, Se and Au whereas the Manus Basin hypothetical source area has high metal fertility for Zn, As, Sb and Pb. Data are referenced in Appendix 4. Troodos source area data
are from this study and Jowitt (2009) and primary MgO is recalculated from Y concentration using the equation MgO = 4.54 * Y + 49.39. Mag cryst. = magnetite
crystallisation.
compositions below 3–3.5 wt% MgO, however, has much lower fer- of formation. In Troodos, the VMS deposits are Cu-rich but Zn-Pb
tility for these metals but high fertility for poorly chalcophile met- poor (Table 1, Hannington et al., 1998) which can be explained
als (As, Sb, Zn an Pb; Fig. 10). In the Troodos ophiolite, fresh glass by the composition of the source area which has higher fertility
compositions (median MgO = 4.3 wt%) indicate good potential for strongly chalcophile elements Cu, Se and Au compared to
metal fertility, especially for strongly chalcophile metals (Fig. 10). poorly chalcophile elements As, Sb, Zn and Pb. The source area
The source areas in the altered lower sheeted dyke section shows compositional control is highlighted through comparison with
slightly more mafic composition (median = 4.9 wt% MgO) but sim- the Manus basin where the VMS deposits are Zn-Pb-As rich but
ilar metal fertility (Fig. 10). The masses of metals mobilised from Cu poor (Herzig et al., 1993; Herzig and Hannington, 1995;
the lower sheeted dykes in Troodos are significantly higher than Beaudoin and Scott, 2009), and the hypothetical lower sheeted
the extent of metal mobilisation in the modern day oceanic crust dyke reaction zones likely have MgO concentrations below
such as at ODP Hole 1256D (e.g. Patten et al., 2015). This is likely 3–3.5 wt% indicating poor fertility for strongly chalcophile metals
a combination of high metal fertility in the source area, the sulphur but high fertility for poorly chalcophile metals (Fig. 10). Deep
under-saturated nature of the Troodos crust as the bulk of chal- water depth is has also been suggested to promote development
cophile metals most likely sit in rare sulphides and the glassy of Cu-rich VMS deposits (Monecke et al., 2014). Based on sediment
matrix which can be efficiently recrystallised, and also the larger records the water depth during formation of the Troodos deposits
time-integrated hydrothermal fluid fluxes (Bickle and Teagle, is suggested to be 3500 m (Keith et al., 2016), which would
1992; Alt and Teagle, 2000). enhance the development of Cu-rich affinity.
The composition of the Troodos VMS deposits is a function of Gold enrichment during formation of VMS deposits has been
the composition of the source area and the water depth at the time suggested to occur through three main processes; 1) addition of
magmatic volatiles (Urabe et al., 1987; Stanton, 1990; Yang and product of addition of magmatic volatiles to the hydrothermal sys-
Scott, 1996, 2002; Sillitoe et al., 1996; Kamenetsky et al., 2001; tem (Moss et al., 2001; Kamenetsky et al., 2001; Yang and Scott,
Moss et al., 2001), 2) vapour separation by sub-seafloor boiling 2002; Sun et al., 2004; Sun et al., 2015). Mungall et al. (2015) sug-
(Urabe et al., 1987; Huston and Large, 1989; Butterfield et al., gested that gas bubbles can act as carriers for sulphide droplets in
1990; Poulsen and Hannington, 1996; Hannington, 1999), and 3) degassing arc magmas. Such process would enable efficient metal
higher Au budget in source area rocks in arc related settings mobilisation from the magma to the hydrothermal system during
(Huston, 2000; Moss et al., 2001; Herrington et al., 2011; degassing despite possible sulphide segregation in the Manus
Pitcairn, 2011). This study shows that the Troodos crust has high Basin (Jenner et al., 2010, 2012, 2015; Park et al., 2015).
Au budget as the Au content of the source area rocks and the mass
of Au mobilised during alteration is significantly higher than those 7. Conclusions
of MORB settings (Patten et al., 2015). However, the mass of base
metals mobilised is also higher and therefore the Au: base metal 7.1. The major outcomes of this study are
ratio of mobilised metals in Troodos is similar to that observed in
MORB. Additional processes such as addition of magmatic volatiles The Troodos ophiolite shows clear subduction affinity with
and or sub-seafloor boiling would be required to enrich Au by an strong enrichments of As, Sb and Pb relative to MORB due to
order of magnitude relative to the base metals to reach the Au slab contamination. Gold contents in the Troodos ophiolitic
grade of Au-rich VMS deposits (Fig. 11). The relatively limited Au fresh glasses (1.9 ± 1.0 ppb Au) are slightly higher but within
enrichment in the Troodos VMS deposits suggests that these addi- the range of MORB values (1.2 ppb, Arevalo and McDonough,
tional processes were irregularly active if not absent. Sub-seafloor 2010; 1.52 ± 1.21, Jenner et al., 2012).
boiling and phase separation of hydrothermal fluids have occurred The Troodos primitive crust is sulphide-undersaturated during
in Troodos ophiolite (Cowan and Cann, 1988; Jowitt et al., 2005) the first stage of magmatic differentiation (9–3.5 wt% MgO)
but the water depth (3.5 km; Keith et al., 2016) most likely until magnetite crystallisation occurs. In Troodos, magnetite
limited the extent and therefore the impact on VMS deposits for- crystallisation is interpreted to trigger sulphide saturation
mation. Comparison with the Manus Basin also indicates that the rather than exsolution of hydrosulphide complexes in mag-
composition of the source rock is not the major control on develop- matic fluids. Strongly chalcophile elements, such as Cu, Se and
ment of Au-rich VMS deposits. Gold-rich deposits are common in Au, therefore behave as incompatible elements during the first
the Manus basin (PACMANUS hydrothermal field; e.g. Herzig stages of magmatic differentiation and have their maximum
et al., 1993; Herzig and Hannington, 1995; Moss et al., 2001; Ihle concentrations just prior to magnetite crystallisation. Their con-
et al., 2005) despite the assumed poor Au fertility of the hypothet- centrations in the melt decrease drastically after sulphide seg-
ical lower sheeted dyke section (Fig. 9). The Au: base metal ratio in regation. Poorly chalcophile elements, such as Zn, As, Sb and
the hypothetic source area of the Manus Basin is in fact even lower Pb are not affected by magnetite crystallisation and sulphide
than in the Troodos ophiolite (Fig. 10), yet Au-rich VMS mineralisa- segregation, and behave as incompatible elements throughout
tion is far more common. Although there is evidence for boiling of whole magmatic differentiation.
hydrothermal fluids at vent sites in the Manus Basin (PACMANUS; Mobilisation of metals during hydrothermal alteration of the
Barriga et al., 2007) the Au enrichment in this area is most likely a Troodos ophiolite is extremely extensive. Percentage depletions
of metals from epidosite zones range from 60% for Pb up to 90%
for Se. These depletions are considerably higher that depletion
observed in hydrothermally altered MORB (Patten et al.,
2015), albeit with the proviso that the mid-ocean ridge equiva-
lents of epidosites have not yet been sampled.
Element mobility in Troodos is not limited to epidosite zones
and strong metal depletions also occur in background altered
diabase (from 58% for Pb to 90% for Se) outside epidosite
zones in the lower sheeted dyke section. This implies that
source areas of Troodos VMS deposits are most likely not solely
restricted to the epidosite zones but could be enlarged to the
lower sheeted dyke section.
The masses of metals mobilised from a source area of 10.9 km3,
(composed of a 5 km3 epidosite zone and 5.9 km3 of background
altered diabase) in the Solea graben are 47 t Au, 21 kt As, 1220 t
Sb, 3080 t Se, 2.4 Mt Cu, 1.8 Mt Zn and 27 kt Pb. The proportions
of these metals trapped in the VMS deposits in the Solea Graben
are estimated to range from 4% for Zn up to 37% for Cu, indicat-
ing that most of the metals mobilised by hydrothermal circula-
tion are not trapped in the large VMS deposits but are lost by
several other processes.
The composition of the source area exerts strong control on the
composition of the VMS deposits. In Troodos, the high source
area fertility for Cu is most likely a major control on the Cu-
rich affinity of these deposits. However, the source area compo-
Fig. 11. Gold to base metal ratio of fresh glass samples from the Troodos ophiolite sition does not appear to control the development of Au-rich
and the Manus Basin; Au in ppm and base metals in weight percent. Troodos data VMS deposits as a significant increase in the Au to base metal
are from this study and fresh glass values of Manus Basin are from Jenner et al.
(2012). Manus basin hypothetic source area has a lower Au to base metal ratio than
ratio would be required to produce an Au-rich VMS deposit.
Troodos ophiolite. Both in Troodos ophiolite and Manus Basin the Au: base metal The most likely causes of Au enrichment are sub-seafloor boil-
ratio is 1 and is not high enough for Au-rich VMS grade.. ing or addition of Au enriched magmatic volatiles.
Acknowledgements Bednarz, U., Schmincke, H.-U., 1994. Petrological and chemical evolution of the
northeastern Troodos Extrusive Series, Cyprus. J. Petrol. 35, 489–523.
Bettison-Varga, L., Varga, R.J., Schiffman, P., 1992. Relation between ore-forming
This work was funded by Stockholm University and by the hydrothermal systems and extensional deformation in the Solea graben
Swedish Research Council (PRG 621-2007-4539). The authors spreading center, Troodos ophiolite, Cyprus. Geology 20, 987–990.
Bickle, M.J., Teagle, D.A., 1992. Strontium alteration in the Troodos ophiolite:
would like to thank Laurence Coogan for providing the fresh glass
implications for fluid fluxes and geochemical transport in mid-ocean ridge
samples used in this study. The authors would also like to thank hydrothermal systems. Earth Planet. Sci. Lett. 113, 219–237.
the reviewers for their comments which greatly improved the Boudreau, A., 1999. PELE—a version of the MELTS software program for the PC
platform. Comput. Geosci. 25, 201–203.
quality of the manuscript. Northern Lion Gold Corp. and Bob John-
Butterfield, D.A., Massoth, G.J., McDuff, R.E., Lupton, J.E., Lilley, M.D., 1990.
ston are thanked for the samples and the help provided during field Geochemistry of hydrothermal fluids from Axial Seamount hydrothermal
investigations. The Geological Survey Department, Ministry of emissions study vent field, Juan de Fuca Ridge: subseafloor boiling and
Agriculture, Natural Resources and Environment and Hellenic Cop- subsequent fluid-rock interaction. J. Geophys. Res. Solid Earth 95, 12895–
12921.
per Mines Ltd are also thanked for logistical support during field Cameron, W., 1985. Petrology and origin of primitive lavas from the Troodos
investigations. Cora Wohlgemuth-Ueberwasser and Runa Jacobson ophiolite, Cyprus. Contrib. Miner. Petrol. 89, 239–255.
are thanked for assistance with LA-ICP-MS and XRF analyses. Cann, J., McCaig, A., Yardley, B., 2015. Rapid generation of reaction permeability in
the roots of black smoker systems, Troodos ophiolite, Cyprus. Geofluids 15,
179–192.
Appendix A. Supplementary data Cawood, P., Fryer, B., 1994. Noble metal abundances in backarc basin basalts (Lau
Basin, southwest Pacific). In: Proceedings of the Ocean Drilling Program.
Scientific results. Ocean Drilling Program, pp. 595–602.
Supplementary data associated with this article can be found, in Constantinou, G., Govett, G., 1973. Geology, geochemistry, and genesis of Cyprus
the online version, at http://dx.doi.org/10.1016/j.oregeorev.2017. sulfide deposits. Econ. Geol. 68, 843–858.
Coogan, L., Banks, G., Gillis, K., MacLeod, C., Pearce, J., 2003. Hidden melting
02.019.
signatures recorded in the Troodos ophiolite plutonic suite: evidence for
widespread generation of depleted melts and intra-crustal melt aggregation.
References Contrib. Miner. Petrol. 144, 484–506.
Coogan, L., Gillis, K., Pope, M., Spence, J., 2017. The role of low-temperature
alteration of the oceanic crust in the global Li-cycle: insights from the Troodos
Adamides, N., 2010a. Mafic-dominated volcanogenic sulphide deposits in the
ophiolite. Geochim. Cosmochim. Acta 203, 201–215.
Troodos ophiolite, Cyprus Part 1–The deposits of the Solea graben. Appl. Earth
Cowan, J., Cann, J., 1988. Supercritical two-phase separation of hydrothermal fluids
Sci. Trans. Inst. Min. Metall. Sect. B 119, 65–77.
in the Troodos ophiolite. Nature 333, 259–261.
Adamides, N., 2010b. Mafic-dominated volcanogenic sulphide deposits in the
Dilek, Y., Furnes, H., Shallo, M., 2007. Suprasubduction zone ophiolite formation
Troodos ophiolite, Cyprus Part 2–A review of genetic models and guides for
along the periphery of Mesozoic Gondwana. Gondwana Res. 11, 453–475.
exploration. Appl. Earth Sci. Trans. Inst. Min. Metall. Sect. B 119, 193–204.
Doe, B.R., 1994. Zinc, copper, and lead in mid-ocean ridge basalts and the source
Agiorgitis, G., Becker, R., 1979. The geochemical distribution of gold in some rocks
rock control on Zn/Pb in ocean-ridge hydrothermal deposits. Geochim.
and minerals of the Troodos Complex, Cyprus. Neues Jahrb. Mineral. Monatsh..
Cosmochim. Acta 58, 2215–2223.
7, 316–320.
Dubé, B., Gosselin, P., Mercier-Langevin, P., Hannington, M., Galley, A., 2007. Gold-
Alt, J.C., 1994. A sulfur isotopic profile through the Troodos ophiolite, Cyprus:
rich volcanogenic massive sulphide deposits. In: Mineral Deposits of Canada: A
primary composition and the effects of seawater hydrothermal alteration.
Synthesis of Major Deposit-Types, District Metallogeny, the Evolution of
Geochim. Cosmochim. Acta 58, 1825–1840.
Geological Provinces, and Exploration Methods. Geological Association of
Alt, J.C., 1995. Subseafloor processes in mid-ocean ridge hydrothermal systems. In:
Canada, Mineral Deposits Division, pp. 75–94. Special Publication 5.
Seafloor Hydrothermal Systems: Physical, Chemical, Biological, and Geological
Flower, M.F., Levine, H.M., 1987. Petrogenesis of a tholeiite-boninite sequence from
Interactions.
Ayios Mamas, Troodos ophiolite: evidence for splitting of a volcanic arc?
Alt, J.C., Anderson, T.F., Bonnell, L., 1989. The geochemistry of sulfur in a 1.3 km
Contrib. Miner. Petrol. 97, 509–524.
section of hydrothermally altered oceanic crust, DSDP Hole 504B. Geochim.
Gillis, K.M., Robinson, P.T., 1988. Distribution of alteration zones in the upper
Cosmochim. Acta 53, 1011–1023.
oceanic crust. Geology 16, 262–266.
Alt, J.C., Teagle, D.A., 2000. Hydrothermal alteration and fluid fluxes in ophiolites
Gillis, K.M., Robinson, P.T., 1990. Patterns and processes of alteration in the lavas
and oceanic crust. Special Papers-Geological Society of America, 273–282.
and dykes of the Troodos Ophiolite, Cyprus. J. Geophys. Res. Solid Earth 95,
Alt, J.C., Laverne, C., Coggon, R.M., Teagle, D.A.H., Banerjee, N.R., Morgan, S., Smith-
21523–21548.
Duque, C.E., Harris, M., Galli, L., 2010. Subsurface structure of a submarine
Galley, A.G., Hannington, M., Jonasson, I., 2007. Volcanogenic massive sulphide
hydrothermal system in ocean crust formed at the East Pacific Rise, ODP/IODP
deposits. In: Mineral Deposits of Canada: A Synthesis of Major Deposit-Types,
Site 1256. Geochem. Geophys. Geosyst. 11.
District Metallogeny, the Evolution of Geological Provinces, and Exploration
Arevalo, R., McDonough, W.F., 2010. Chemical variations and regional diversity
Methods. Geological Association of Canada, Mineral Deposits Division, pp. 141–
observed in MORB. Chem. Geol. 271, 70–85.
161. Special Publication 5.
Bach, W., Peucker-Ehrenbrink, B., Hart, S.R., Blusztajn, J.S., 2003. Geochemistry of
Garfunkel, Z., 1998. Constrains on the origin and history of the Eastern
hydrothermally altered oceanic crust: Dsdp/ODP hole 504B–Implications for
Mediterranean basin. Tectonophysics 298, 5–35.
seawater-crust exchange budgets and Sr-and Pb-isotopic evolution of the
Gass, I.G., 1990. Ophiolites and oceanic lithosphere. In: Malpas, J., Moores, E.M.,
mantle. Geochem. Geophys. Geosyst. 4.
Panayiotou, A., Xenophontos, C. (Eds.), Ophiolites, Oceanic Crustal Analogues,
Bailey, D., Langdon, G., Malpas, J., Robinson, P., 1991. Ultramafic and related lavas
Proceeding Symposium ‘Troodos 1987’. Cyprus Geological Survey Department,
from the Margi area, Troodos ophiolite. Cyprus Crystal Study Project Initial
Nicosia, pp. 1–10.
Report. Holes CY-1 and Geological Survey Canada Paper, pp. 90–20.
Granot, R., Abelson, M., Ron, H., Agnon, A., 2006. The oceanic crust in 3D:
Baragar, W., Lambert, M., Baglow, N., Gibson, I., 1989. Sheeted dykes from CY-4 and
paleomagnetic reconstruction in the Troodos ophiolite gabbro. Earth Planet. Sci.
surface sections: Troodos ophiolite. Cyprus Crustal Study Project: Initial Report
Lett. 251, 280–292.
CY-4, Geological Survey of Canada, Paper, pp. 88–89.
Hamlyn, P.R., Keays, R.R., Cameron, W.E., Crawford, A.J., Waldron, H.M., 1985.
Batanova, V.G., Sobolev, A.V., 2000. Compositional heterogeneity in subduction-
Precious metals in magnesian low-Ti lavas: implications for metallogenesis and
related mantle peridotites, Troodos massif, Cyprus. Geology 28, 55–58.
sulfur saturation in primary magmas. Geochim. Cosmochim. Acta 49, 1797–
Baragar, W., Lambert, M., Baglow, N., Gibson, I., 1987. The sheeted dyke zone in the
1811.
Troodos ophiolite. ophiolites-Oceanic Crustal Analogues. In: Malpas, J., Moores,
Hannington, M.D., Galley, A.G., Herzig, P.M., Petersen, S., 1998, 28. Comparison of
E.M., Panayiotou, A., Xenophontos, C. (Eds.), Ophiolites, Oceanic Crustal
the TAG mound and stockwork complex with cyprus-type massive sulfide
Analogues, Proceeding Symposium ‘Troodos 1987’. Cyprus Geological Survey
deposits. In: Proceedings of the Ocean Drilling Program: Scientific Results, Leg
Department, Nicosia, pp. 37–51.
158, pp. 349–415.
Barnes, S.J., Lightfoot, P.C., 2005. Formation of magmatic nickel-sulfide ore deposits
Hannington, M., 1999. Volcanogenic gold in the massive sulfide environment
and processes affecting their copper and platinum-group element contents.
Volcanic-associated massive sulfide deposits: processes and examples in
Econ. Geol. 100th Anniversary Volume, 179–213.
modern and ancient settings. Rev. Econ. Geol. 8, 325–356.
Barrie, C., Hannington, M., 1999. Classification of volcanic-associated massive
Harris, M., Coggon, R.M., Smith-Duque, C.E., Cooper, M.J., Milton, J.A., Teagle, D.A.H.,
sulfide deposits based on host-rock composition. Rev. Econ. Geol. 8, 1–11.
2015. Channelling of hydrothermal fluids during the accretion and evolution of
Barriga, F., Binns, R., Miller, D., Herzig, P., 2007. 1. Leg 193 Synthesis: Anatomy of an
the upper oceanic crust: Sr isotope evidence from ODP Hole 1256D. Earth
active felsic-hosted hydrothermal system, eastern Manus Basin, Papua New
Planet. Sci. Lett. 416, 56–66.
Guinea.
Hattori, K.H., Guillot, S., 2003. Volcanic fronts form as a consequence of serpentinite
Bear, L.M., 1963. The mineral resources and mining industry of Cyprus, Cyprus.
dehydration in the forearc mantle wedge. Geology 31, 525–528.
Geol. Surv. Bull. 1, 1–208.
Hattori, K., Takahashi, Y., Guillot, S., Johanson, B., 2005. Occurrence of arsenic (V) in
Beaudoin, Y., Scott, S.D., 2009. Pb in the PACMANUS sea-floor hydrothermal system,
forearc mantle serpentinites based on X-ray absorption spectroscopy study.
eastern Manus Basin: Numerical modeling of a magmatic versus leached origin.
Geochim. Cosmochim. Acta 69, 5585–5596.
Econ. Geol. 104, 749–758.
Hellstrom, J., Paton, C., Woodhead, J., Hergt, J., 2008. Iolite: software for spatially Moss, R., Scott, S.D., Binns, R.A., 2001. Gold content of eastern manus basin volcanic
resolved LA-(quad and MC) ICPMS analysis. Mineralogical Association of Canada rocks: implications for enrichment in associated hydrothermal precipitates.
short course series 40, 343–348. Econ. Geol. 96, 91–107.
Herrington, R.J., Scotney, P.M., Roberts, S., Boyce, A.J., Harrison, D., 2011. Temporal Mudd, G.M., Weng, Z., Jowitt, S.M., 2013. A detailed assessment of global Cu
association of arc–continent collision, progressive magma contamination in arc resource trends and endowments. Econ. Geol. 108, 1163–1183.
volcanism and formation of gold-rich massive sulphide deposits on Wetar Muenow, D., Garciat, M., Aggrey, K., Bednarz, U., Schmincke, H., 1990. Volatiles in
Island (Banda arc). Gondwana Res. 19, 583–593. submarine glasses as a discriminant of tectonic origin: application to the
Herzig, P.M., Hannington, M.D., 1995. Polymetallic massive sulfides at the modern Troodos ophiolite. Nature 343, 159–161.
seafloor a review. Ore Geol. Rev. 10, 95–115. Mukasa, S.B., Ludden, J.N., 1987. Uranium-lead isotopic ages of plagiogranites from
Herzig, P.M., Hannington, M.D., Fouquet, Y., von Stackelberg, U., Petersen, S., 1993. the Troodos ophiolite, Cyprus, and their tectonic significance. Geology 15, 825–
Gold-rich polymetallic sulfides from the Lau back arc and implications for the 828.
geochemistry of gold in sea-floor hydrothermal systems of the Southwest Mungall, J.E., 2002. Roasting the mantle: slab melting and the genesis of major Au
Pacific. Econ. Geol. 88, 2182–2209. and Au-rich Cu deposits. Geology 30, 915–918.
Herzig, P.M., Hannington, M.D., Scott, S.D., Maliotis, G., Rona, P.A., Thompson, G., Mungall, J.E., Brenan, J.M., Godel, B., Barnes, S., Gaillard, F., 2015. Transport of metals
1991. Gold-rich sea-floor gossans in the Troodos ophiolite and on the mid- and sulphur in magmas by flotation of sulphide melt on vapour bubbles. Nat.
atlantic ridge. Econ. Geol. Bull. Soc. Econ. Geol. 86, 1747–1755. Geosci. 8, 216–219.
Humphris, S.E., Cann, J.R., 2000. Constraints on the energy and chemical balances of Noll, P., Newsom, H., Leeman, W., Ryan, J.G., 1996. The role of hydrothermal fluids in
the modern TAG and ancient Cyprus seafloor sulfide deposits. J. Geophys. Res. the production of subduction zone magmas: evidence from siderophile and
Solid Earth 105, 28477–28488. chalcophile trace elements and boron. Geochim. Cosmochim. Acta 60, 587–611.
Huston, D.L., Large, R.R., 1989. A chemical model for the concentration of gold in Nuriel, P., Katzir, Y., Abelson, M., Valley, J.W., Matthews, A., Spicuzza, M.J., Ayalon,
volcanogenic massive sulphide deposits. Ore Geol. Rev. 4, 171–200. A., 2009. Fault-related oceanic serpentinization in the Troodos ophiolite,
Huston, D., 2000. Gold in volcanic-hosted massive sulfide deposits: distribution, Cyprus: implications for a fossil oceanic core complex. Earth Planet. Sci. Lett.
genesis, and exploration. Rev. Econ. Geol. 13, 401–426. 282, 34–46.
Ihle, T., Petersen, S., Herzig, P., Hannington, M.D., 2005. Siting of gold and Ohmoto, H., 1996. Formation of volcanogenic massive sulfide deposits: the Kuroko
characteristics of gold-bearing massive sulfides from the interior of the felsic- perspective. Ore Geol. Rev. 10, 135–177.
hosted PACMANUS massive sulfide deposit, eastern Manus basin (PNG). Mineral Osozawa, S., Shinjo, R., Lo, C.-H., Jahn, B.-M., Hoang, N., Sasaki, M., Ishikawa, K.i.,
Deposit Research: Meeting the Global Challenge, 623–626. Kano, H., Hoshi, H., Xenophontos, C., 2012. Geochemistry and geochronology of
Jenner, F.E., O’Neill, H.S.C., Arculus, R.J., Mavrogenes, J.A., 2010. The Magnetite Crisis the Troodos ophiolite: an SSZ ophiolite generated by subduction initiation and
in the Evolution of Arc-related Magmas and the Initial Concentration of Au, Ag an extended episode of ridge subduction? Lithosphere 4, 497–510.
and Cu. J. Petrol. 51, 2445–2464. Park, J.-W., Campbell, I.H., Kim, J., Moon, J.-W., 2015. The role of late sulfide
Jenner, F., Arculus, R., Mavrogenes, J., Dyriw, N., Nebel, O., Hauri, E., 2012. saturation in the formation of a Cu-and Au-rich magma: insights from the
Chalcophile element systematics in volcanic glasses from the north western platinum group element geochemistry of Niuatahi-Motutahi lavas, Tonga rear
Lau basin. Geochem. Geophys. Geosyst. 13. arc. J. Petrol. 56, 59–81.
Jenner, F.E., Hauri, E.H., Bullock, E.S., König, S., Arculus, R.J., Mavrogenes, J.A., Park, S.-H., Lee, S.-M., Kamenov, G.D., Kwon, S.-T., Lee, K.-Y., 2010. Tracing the origin
Mikkelson, N., Goddard, C., 2015. The competing effects of sulfide saturation of subduction components beneath the South East rift in the Manus Basin,
versus degassing on the behavior of the chalcophile elements during the Papua New Guinea. Chem. Geol. 269, 339–349.
differentiation of hydrous melts. Geochem. Geophys. Geosyst. 16, 1490–1507. Paton, C., Hellstrom, J., Paul, B., Woodhead, J., Hergt, J., 2011. Iolite: freeware for the
Jowitt, S.M., 2009. Field, petrological and geochemical constraints on the release of visualisation and processing of mass spectrometric data. J. Anal. At. Spectrom.
base metals into hydrothermal fluids in Cyprus-type Volcanogenic Massive 26, 2508–2518.
Sulfide (VMS) systems: an investigation of the Spilia-Kannavia epidosite zone, Patten, C., Barnes, S.-J., Mathez, E.A., 2012. Textural variations in MORB sulfide
Troodos ophiolite, Cyprus (Unpublished Ph.D. thesis). University of Leicester. 1- droplets due to differences in crystallization history. Can. Min. 50, 675–692.
355. Patten, C., Barnes, S.-J., Mathez, E.A., Jenner, F.E., 2013. Partition coefficients of
Jowitt, S., Osborn, R., Thomas, R., Naden, J., Gunn, A., Herrington, R., Nicolaides, S., chalcophile elements between sulfide and silicate melts and the early
2005. ‘T’-type mineralisation—a pseudo-epithermal style of VHMS associated crystallization history of sulfide liquid: LA-ICP-MS analysis of MORB sulfide
gold mineralisation, Cyprus. Mineral Deposit Research: Meeting the Global droplets. Chem. Geol. 358, 170–188.
Challenge, pp. 635–637. Patten, C.G., Pitcairn, I.K., Teagle, D.A.H., Harris, M., 2015. Mobility of Au and related
Jowitt, S.M., Jenkin, G.R.T., Coogan, L.A., Naden, J., 2012. Quantifying the release of elements during the hydrothermal alteration of the oceanic crust: implications
base metals from source rocks for volcanogenic massive sulfide deposits: effects for the sources of metals in VMS deposits. Miner. Deposita, 1–22.
of protolith composition and alteration mineralogy. J. Geochem. Explor. 118, Patten, C.G.C., Pitcairn, I.K., Teagle, D.A.H., Harris, M., 2016. Sulphide mineral
47–59. evolution and metal mobility during alteration of the oceanic crust: insights
Jugo, P.J., 2009. Sulfur content at sulfide saturation in oxidized magmas. Geology 37, from ODP Hole 1256D. Geochim. Cosmochim. Acta 193, 132–159.
415–418. Peach, C.L., Mathez, E.A., Keays, R.R., 1990. Sulfide melt silicate melt distribution
Kamenetsky, V., Binns, R., Gemmell, J., Crawford, A., Mernagh, T., Maas, R., Steele, D., coefficients for noble-metals and other chalcophile elements as deduced from
2001. Parental basaltic melts and fluids in eastern Manus backarc Basin: MORB – implications for partial melting. Geochim. Cosmochim. Acta 54, 3379–
implications for hydrothermal mineralisation. Earth Planet. Sci. Lett. 184, 685– 3389.
702. Pearce, J.A., 1975. Basalt geochemistry used to investigate past tectonic
Keith, M., Haase, K.M., Klemd, R., Krumm, S., Strauss, H., 2016. Systematic variations environments on Cyprus. Tectonophysics 25, 41–67.
of trace element and sulfur isotope compositions in pyrite with stratigraphic Pearce, J.A., 1982. Trace element characteristics of lavas from destructive plate
depth in the Skouriotissa volcanic-hosted massive sulfide deposit, Troodos boundaries. Andesites 8, 525–548.
ophiolite, Cyprus. Chem. Geol. 423, 7–18. Pearce, J.A., Lippard, S., Roberts, S., 1984. Characteristics and tectonic significance of
Kelley, D.S., Robinson, P.T., Malpas, J.G., 1992. Processes of brine generation and supra-subduction zone ophiolites. Geol. Soc. London, Special Publications 16,
circulation in the oceanic crust: fluid inclusion evidence from the Troodos 77–94.
ophiolite, Cyprus. J. Geophys. Res. Solid Earth 97, 9307–9322. Pearce, J.A., Peate, D.W., 1995. Tectonic implications of the composition of volcanic
Li, Y., Audétat, A., 2015. Effects of temperature, silicate melt composition, and arc magmas. Annu. Rev. Earth Planet. Sci. 23, 251–286.
oxygen fugacity on the partitioning of V, Mn Co, Ni, Cu, Zn, As, Mo, Ag, Sn, Sb, W, Pearce, J.A., 2003. Supra-subduction zone ophiolites: the search for modern
Au, Pb, and Bi between sulfide phases and silicate melt. Geochim. Cosmochim. analogues. Special Papers–Geological Society of America, 269–294.
Acta 162, 25–45. Pearce, J.A., Robinson, P., 2010. The Troodos ophiolitic complex probably formed in a
Liu, Y., Samaha, N.-T., Baker, D.R., 2007. Sulfur concentration at sulfide saturation subduction initiation, slab edge setting. Gondwana Res. 18, 60–81.
(SCSS) in magmatic silicate melts. Geochim. Cosmochim. Acta 71, 1783–1799. Petersen, S., Herzig, P.M., Hannington, M.D., Jonasson, I.R., Arribas, A., 2002.
Malpas, J., Williams, D., 1991. Geology of the area surrounding the CY-1 and CY-1A Submarine gold mineralization near Lihir Island, New Ireland fore-arc, Papua
boreholes. In: Malpas, J., Robinson, P.T., Xenophontos, C. (Eds.), Cyprus Crustal New Guinea. Econ. Geol. 97, 1795–1813.
Study Project. Geological Survey of Canada, Ottawa, Ontario, pp. 29–40. Pitcairn, I.K., Warwick, P.E., Milton, J.A., Teagle, D.A.H., 2006a. Method for ultra-low-
Mehegan, J.M., 1988. Temporal, spatial and chemical evolution of the Troodos level analysis of gold in rocks. Anal. Chem. 78, 1290–1295.
ophiolite, Cyprus: Supra-subduction zone volcanism in the Tethyan Sea (PhD Pitcairn, I.K., Teagle, D.A.H., Craw, D., Olivo, G.R., Kerrich, R., Brewer, T.S., 2006b.
thesis). Dalhousie University, Halifax. Sources of metals and fluids in orogenic gold deposits: insights from the Otago
Mercier-Langevin, P., Hannington, M.D., Dube, B., Becu, V., 2011. The gold content of and Alpine Schists, New Zealand. Econ. Geol. 101, 1525–1546.
volcanogenic massive sulfide deposits. Miner. Deposita 46, 509–539. Pitcairn, I.K., 2011. Background concentrations of gold in different rock types. Appl.
Miyashiro, A., 1973. The Troodos ophiolitic complex was probably formed in an Earth Sci. 120, 31–38.
island arc. Earth Planet. Sci. Lett. 19, 218–224. Prichard, H.M., Maliotis, G., 1998. Gold mineralization associated with low-
Monecke, T., Petersen, S., Hannington, M.D., 2014. Constraints on water depth of temperature, off-axis, fluid activity in the Troodos ophiolite, Cyprus. J. Geol.
massive sulfide formation: evidence from modern seafloor hydrothermal Soc. 155, 223–231.
systems in arc-related settings. Econ. Geol. 109, 2079–2101. Portnyagin, M., Danyushevsky, L., Kamenetsky, V., 1997. Coexistence of two distinct
Mosier, D.L., Singer, D., Salem, B., 1983. Geologic and grade-tonnage information on mantle sources during formation of ophiolites: a case study of primitive pillow-
volcanic-hosted copper-zinc-lead massive sulfide deposits. U.S. Geological lavas from the lowest part of the volcanic section of the Troodos Ophiolite,
Survey Open-File Report 83–89, 1–78. Cyprus. Contrib. Miner. Petrol. 128, 230–287.
Poulsen, K., Hannington, M., 1996. Volcanic-associated massive sulphide gold. Staudigel, H., Tauxe, L., Gee, J., Bogaard, P., Haspels, J., Kale, G., Leenders, A., Meijer,
Geology of Canadian Mineral Deposit Types, Geological Survey of Canada, P., Swaak, B., Tuin, M., 2000. Geochemistry and intrusive directions in sheeted
Geology of Canada 8, 183–196. dikes in the troodos ophiolite: implications for mid-ocean ridge spreading
Rautenschlein, M., Jenner, G.A., Hertogen, J., Hofmann, A.W., Kerrich, R., Schmincke, centers. Geochem. Geophys. Geosyst. 1.
H.U., White, W.M., 1985. Isotopic and trace element composition of volcanic Staudigel, H., 2003. Hydrothermal alteration processes in the oceanic crust. Treatise
glasses from the Akaki Canyon, Cyprus: implications for the origin of the Geochem. 3, 511–535.
Troodos ophiolite. Earth Planet. Sci. Lett. 75, 369–383. Sun, W., Arculus, R.J., Kamenetsky, V.S., Binns, R.A., 2004. Release of gold-bearing
Regelous, M., Haase, K., Freund, S., Keith, M., Weinzierl, C., Beier, C., Brandl, P., fluids in convergent margin magmas prompted by magnetite crystallization.
Endres, T., Schmidt, H., 2014. Formation of the Troodos Ophiolite at a triple Nature 431, 975–978.
junction: evidence from trace elements in volcanic glass. Chem. Geol. 386, Sun, W., Binns, R., Fan, A., Kamenetsky, V.S., Wysoczanski, R., Wei, G., Hu, Y., Arculus,
66–79. R., 2007. Chlorine in submarine volcanic glasses from the Eastern Manus basin.
Richards, H., Cann, J., Jensenius, J., 1989. Mineralogical zonation and metasomatism Geochim. Cosmochim. Acta 71, 1542–1552.
of the alteration pipes of Cyprus sulfide deposits. Econ. Geol. 84, 91–115. Sun, W., Huang, R.-F., Li, H., Hu, Y.-B., Zhang, C.-C., Sun, S.-J., Zhang, L.-P., Ding, X., Li,
Richards, J.P., 2011. Magmatic to hydrothermal metal fluxes in convergent and C.-Y., Zartman, R.E., 2015. Porphyry deposits and oxidized magmas. Ore Geol.
collided margins. Ore Geol. Rev. 40, 1–26. Rev. 65, 97–131.
Richards, J.P., 2015. The oxidation state, and sulfur and Cu contents of arc magmas: Teagle, D., Alt, J., Umino, S., Miyashita, S., Banerjee, N., Wilson, D., 2006. the
implications for metallogeny. Lithos 233, 27–45. Expedition 309, 312 Scientists, Superfast spreading rate crust 2 and 3.
Richardson, C.J., Cann, J.R., Richards, H.G., Cowan, J.G., 1987. Metal-depleted root Proceedings of the Ocean Drilling Program 309 (312), 50.
zones of the Troodos ore-forming hydrothermal systems, Cyprus. Earth Planet. Timm, C., de Ronde, C.E., Leybourne, M.I., Layton-Matthews, D., Graham, I.J., 2012.
Sci. Lett. 84, 243–253. Sources of chalcophile and siderophile elements in Kermadec Arc Lavas⁄. Econ.
Robertson, A., 1987. Tectonic evolution of Cyprus. In: Malpas, J., Moores, E.M., Geol. 107, 1527–1538.
Panayiotou, A., Xenophontos, C. (Eds.), Ophiolites, Oceanic Crustal Analogues, Tornos, F., Peter, J.M., Allen, R., Conde, C., 2015. Controls on the siting and style of
Proceeding Symposium ‘Troodos 1987’. Cyprus Geological Survey Department, volcanogenic massive sulphide deposits. Ore Geol. Rev. 68, 142–163.
Nicosia. Thy, P., Brooks, C., Walsh, J., 1985. Tectonic and petrogenetic implications of major
Robinson, P.T., Melson, W.G., O’Hearn, T., Schmincke, H.-U., 1983. Volcanic glass and rare earth element chemistry of Troodos glasses, Cyprus. Lithos 18, 165–
compositions of the Troodos ophiolite, Cyprus. Geology 11, 400–404. 178.
de Ronde, C.E., Massoth, G.J., Butterfield, D.A., Christenson, B.W., Ishibashi, J., Urabe, T., Yuasa, M., Nakao, S., 1987. Hydrothermal sulfides from a submarine
Ditchburn, R.G., Hannington, M.D., Brathwaite, R.L., Lupton, J.E., Kamenetsky, V. caldera in the Shichito-Iwojima Ridge, northwestern Pacific. Mar. Geol. 74, 295–
S., 2011. Submarine hydrothermal activity and gold-rich mineralization at 299.
Brothers Volcano, Kermadec Arc, New Zealand. Miner. Deposita 46, 541–584. Varga, R.J., Moores, E.M., 1985. Spreading structure of the Troodos ophiolite, Cyprus.
Schiffman, P., Bettison, L., Smith, B., 1987. Mineralogy and geochemistry of Geology 13, 846–850.
epidosites from the Solea graben, Troodos ophiolite, Cyprus. In: Malpas, J., Vine, F., Smith, G., 1990. Structure and physical properties of the Troodos crustal
Moores, E.M., Panayiotou, A., Xenophontos, C. (Eds.), Ophiolites, Oceanic Crustal section at ICRDG drill holes CY1, 1a and 4. Ophiolites–oceanic crustal analogues.
Analogues, Proceeding Symposium ‘Troodos 1987’. Cyprus Geological Survey Geological Survey Department, Ministry of Agriculture and Natural Resources,
Department, Nicosia, pp. 673–684. Cyprus, pp. 113–124.
Schiffman, P., Smith, B.M., 1988. Petrology and oxygen isotope geochemistry of a Wallace, P.J., Edmonds, M., 2011. The sulfur budget in magmas: evidence from melt
fossil seawater hydrothermal system within the Solea Graben, northern inclusions, submarine glasses, and volcanic gas emissions. Rev. Mineral.
Troodos ophiolite, Cyprus. J. Geophys. Res. Solid Earth 93, 4612–4624. Geochem. 73, 215–246.
Seyfried, W.E., Berndt, M.E., Seewald, J.S., 1988. Hydrothermal alteration processes Wilson, R.A.M., Ingham, F., Geōrgias, Y., kai Perivallontos, P.P., Dēmokratia, K.,
at mid-ocean ridges; constraints from diabase alteration experiments, hot Episkopēsēs, T.G., 1959. The Geology of the Xeros-Troodos Area: With an
spring fluids and composition of the oceanic crust. Can. Mineral. 26, 787–804. Account of the Mineral Resources. authority of the Government of Cyprus.
Sillitoe, R.H., Hannington, M.D., Thompson, J.F., 1996. High sulfidation deposits in Yang, K., Scott, S.D., 1996. Possible contribution of a metal-rich magmatic fluid to a
the volcanogenic massive sulfide environment. Econ. Geol. 91, 204–212. sea-floor hydrothermal system. Nature 383, 420–423.
Stanton, R., 1990. Magmatic evolution and the ore type-lava type affiliations of Yang, K., Scott, S.D., 2002. Magmatic degassing of volatiles and ore metals into a
volcanic exhalative ores. Australasian Intitule of Mining & Metallurgy hydrothermal system on the modern sea floor of the eastern Manus back-arc
Proceeding 15, 101–107. basin, western Pacific. Econ. Geol. 97, 1079–1100.
Staudigel, H., Gee, J., Tauxe, L., Varga, R.J., 1992. Shallow intrusive directions of
sheeted dikes in the Troodos ophiolite: anisotropy of magnetic susceptibility
and structural data. Geology 20, 841–844.

Patten Et Al., 2017

Uploaded by

Copyright:

Available Formats

Patten Et Al., 2017

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Patten Et Al., 2017

Uploaded by

Copyright:

Available Formats

Ore Geology Reviews 86 (2017) 487–508

Contents lists available at ScienceDirect

Ore Geology Reviews

Hydrothermal mobilisation of Au and other metals in supra-subduction

2.3. The VMS deposits of the Troodos ophiolite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491

1. Introduction et al., 1990; Poulsen and Hannington, 1996; Hannington, 1999;

References S (%) Cu (%) Zn (%) Pb (ppm) Au (ppm) As (ppm) Sb (ppm) Se (ppm)

n = number of samples. ⁄Estimated Au grade.

4.3. The lower sheeted dyke complex

The major element compositions of the sheeted dyke samples

C.G.C. Patten et al. / Ore Geology Reviews 86 (2017) 487–508

All data are in Appendix 1. b.d.l. = below detection limit.

leads either to sulphide saturation and segregation of a sulphide

exsolution of hydrosulphide complexes in cogenetic magmatic flu-

inclusions from late-stage gabbros and plagiogranites in Troodos

suggest that magmatic fluids have exsolved during magmatic dif-

plutonic rocks (Kelley et al., 1992). Additionally, negative d34S val-

S trapped within the deposit (Keith et al., 2016). Volatile contents

of fresh glass samples show that degassing of CO2 occurs during

in the magma (Fig. 6). The behaviour of S is suggested to be more

similar to that of H2O rather than CO2 during degassing in a high

occurred during magmatic degassing in the Troodos ophiolite but

it is not interpreted to be the main cause for the S drop observed

the main mechanism for S removal from the magma at 3 wt%

segregation during the second stage of magmatic differentiation

5.3. Metal behaviour in the primitive oceanic crust

Despite the limited number of fresh glass samples analysed in

dos for the elements of interest can be investigated using previ-

ously published data and modern-day analogues. Like S, the

behaviour during the first stage of magmatic differentiation in

the Troodos samples, followed by compatible behaviour during

ciated to sulphide segregation. Of the samples analysed for Au and

concentrations of these elements after sulphide segregation.

Strongly chalcophile elements (Au, Se and Cu) show strikingly sim-

concentrations match those of S and TiO2 at 3 wt% MgO (Fig. 7;

ments has been attributed both to volatile loss (Kamenetsky

2015) and to sulphide segregation (Jenner et al., 2010, 2012,

2015). The poorly chalcophile elements Zn, Pb, As and Sb are

incompatible throughout magmatic differentiation in Troodos sug-

in the mass change value for Al2O3 is caused mainly by a small

6.1. Mass variation in the volcanic section

The volcanic section has been affected by multiple styles and

6.2. Mass variation in the lower sheeted dyke section

6.2.1. Epidosite zones

plutonic complex and at the top of the sheeted dyke complex

epidosite zones occur to the west of the Solea graben axis

interpreted as both the epidosite zone and the background altered

sheeted dyke complex estimated at 1250 m thick; Alt, 1994;

ground altered diabase is calculated to be the total volume of the

lower sheeted dyke section minus the volume of the epidosite

these volumes are calculated using the metal depletions shown in

Table 4. The mass balance shows that 47 ± 9 t Au, 21 ± 5 kt As,

and 27 ± 4 kt Pb has been removed from the 10.9 km3 altered

source area in the Solea graben (Table 5). It should be emphasised

that these calculations are made for a simplified hydrothermal cell

system and they are an estimate of the actual mass of metal

Comparing available data for tonnages of Au and related ele-

tity of metal mobilised from the lower sheeted dyke complex

allows estimation of metal trapping efficiencies (Table 5). How-