Magmatic Sulfide Ore Deposits
Stephen J. Barnes1, David A. Holwell2, and Margaux Le Vaillant1
1811-5209/17/0013-0089$2.50  DOI: 10.2113/gselements.13.2.89
M
agmatic sulfide ore deposits are products of natural smelting:
concentration of immiscible sulfide liquid (‘matte’), enriched in
chalcophile elements, derived from silicate magmas (‘slags’). Sulfide
ore deposits occupy a spectrum from accumulated pools of matte within small
igneous intrusions or lava flows, mined primarily for Ni and Cu, to stratiform
layers of weakly disseminated sulfides within large mafic–ultramafic intru-
sions, mined for platinum-group elements. One of the world’s most valuable
deposits, the Platreef in the Bushveld Complex (South Africa) has aspects of
both of these end members. Natural matte compositions vary widely between
and within deposits, and these compositions are controlled largely by the
relative volumes of matte and slag that interact with one another.
Keywords : nickel, sulfide, platinum, layered intrusions, large igneous provinces,
magma, igneous petrology
INTRODUCTION
Magmatic sulfide deposits are nature’s smelters. By the
same process that has been used since prehistoric times
to extract metals from ores, magmatic sulfide ores form
by the interaction between immiscible sulfide–oxide
liquids (mattes) with silicate magmas (slags). Scavenging
of chalcophile elements – Ni, Cu, Au and the platinum-
group elements (PGEs) – and the accumulation of the matte
component (Fig. 1) has produced some of the world’s most
valuable economic metal concentrations (Naldrett 2004).
These currently account for ~56% of the world’s Ni produc-
tion and over 96% of Pt, Pd and the other PGE production
(Mudd and Jowitt 2014).
Magmatic Sulfide Deposit Settings
On a deposit scale, magmatic sulfide accumulations are
found in a variety of host igneous rock bodies. Broadly,
they fall into two major categories: sulfide-rich, exploited
primarily for Ni and Cu; and sulfide-poor (typically less
than 5% sulfide), dominated by the platinum-group
elements (PGEs) and Au.
Sulfide-rich deposits dominated by Ni–Cu can be catego-
rized into three types:
1. Sulfide-rich accumulations in small mafic or mafic–
ultramafic intrusions, (Figs. 1A, B), usually identifiable
as magma conduits (Lightfoot and Evans-Lamswood
2015) and exploited mostly for Ni, Cu and Co.
1 CSIRO Mineral Resources
26 Dick Perry Avenue
Kensington WA 6151, Australia
E-mail Steve.Barnes@csiro.au Margaux.Levaillant@csiro.au
2 Department of Geology
University of Leicester
Leicester, LE1 7RH, United Kingdom
E-mail: dah29@le.ac.uk
E lements , V ol . 13, pp. 89–95 89
Important examples include
Voisey’s Bay (Canada), Jinchuan
(China) and the Noril’sk-Talnakh
deposits of Siberia (Russia). The
PGEs are minor by-products,
with the notable exception of the
Noril’sk-Talnakh ores where PGEs
are unusually abundant.
2. Accumulations of sulfide in
komatiites, such as the Kambalda
and Perseverance deposits in
Australia (Barnes 2006), or in
ferropicrites, such as the Pechenga
deposits in Russia (Hanski et al.
2011). Deposits are hosted in
lava flows (F ig. 2A) or shallow
subvolcanic intrusions. Exploited
dominantly for Ni only.
3. Sulfide accumulation beneath an impact-generated
crustal melt sheet: the unique example of the Sudbury
Ni–Cu–PGE ores (Keays and Lightfoot 2004).
Sulfide-poor deposits dominated by PGEs can be catego-
rized into two types:
1. Stratiform accumulations with a few percent dissemi-
nated sulfide in cumulates within large layered mafic–
ultramafic intrusions, including PGE-enriched “reefs”
(Naldrett et al. 2008). Such deposits are typically
exploited for PGEs with by-product Ni, Cu and Co.
They generally occur as remarkably thin and persistent
layers. The best-known example is the Merensky Reef
of the Bushveld Complex (Fig. 1C). This exceptional
layer is commonly only a few tens of centimetres thick
but extends continuously for over 400 km.
2. Stratabound sulfide disseminations, commonly
PGE-rich, in the marginal rocks of large layered
intrusions. An example is the Platreef of the Bushveld
Complex (McDonald and Holwell 2011).
The Form of Ore-Hosting Magma Bodies
Stratiform reef-style PGE deposits are exclusively hosted
within large sill-like or boat-shaped layered mafic–ultra-
mafic intrusions, usually several kilometres thick. But there
is a much greater diversity of form in the magma bodies that
host Ni–Cu-dominant, sulfide-rich orebodies (Fig. 2). These
latter types all represent the products of magma flowing
through restricted conduits or channels, leaving behind an
accumulated residue of sulfide liquid and cumulus silicate
minerals. These conduits can be feeder tubes or channels
within extensive komatiite lava-flow fields (Lesher 1989;
Barnes 2006) (Fig. 2A), or feeders to large igneous province
volcanism in the form of sill–dike combinations (Fig. 2C) or
tube-like conduits (Figs. 2D, E) (Barnes et al. 2016; Lightfoot
A pr il 2017
 AA. “Kambalda Type” -
A
A komatiite lava flows
Flow top
M$ Olivine-rich channel fill
Komatiite
Sediment
Sulfide orebody -
matrix/disseminated
Sulfide orebody -
massive ore
Xen Basalt

Pyritic layer Shale 0 100 m

5 cm
BB. “Mirabela Type”
- funnel/boat
B
B Hnf
Gabbro, gabbronorite
Mineralization (disseminated)
Pyroxenite
Dunite
0 ~1 km

CC. “Eagle/Kalatongke Type” -

Troc Sul elongated funnel,
flared dyke
DD. “Nebo-Babel (Limoeiro) Type” -
Plg
tubular chonolith Gabbro(norite), Troctolite
Mineralization

0 ~1 km

Hnf

0 ~200 m
2 cm
Disseminated ore

C
C Matrix ore
(Semi-) Massive ore
MelN
Olivine gabbro, troctolite, peridotite
Gabbro, dolerite
Chr

EE. “Noril’sk (Nkomati) Type” -

chonolith/elongate sill
D
MelN
Olivine gabbro, troctolite, peridotite
~ 100-200 m
Gabbro, dolerite

Massive ore
Disseminated ore
Cu-PGE breccia ore ~200-400 m
10 cm Chr
Chr 0
Anth SJBdbt008

Field characteristics of magmatic Ni–Cu–platinum Characteristic geometries of magmatic bodies that

Figure 1
group element (PGE) sulfide ores. (A) Massive sulfide
(M$) from the Eagle deposit (Michigan, USA) showing assimilation
of sulfidic black shale. Note narrow zone of partially molten shale
xenoliths (Xen) just above contact, and presence of a band of
sedimentary pyrite within the black shale providing a source of
sulfur for the orebody. (B) Sulfide-matrix ore breccia from the
Voisey’s Bay deposit (Labrador, Canada). Irregular-shaped xenoliths
of country rock hornfels (Hnf) and troctolite (Troc) fragments of
host intrusion within sulfide (Sul) also containing laths of plagio-
clase (Plg). (C) Underground exposure of the Merensky Reef,
Bushveld Igneous Complex (South Africa). Layer of coarse
melanorite (MelN) with disseminated sulfides between two thin
chromite seams (Chr), with anorthosite (Anth) below and
melanorite above. Photo : R ais L atypov. Inset (D), closeup of cut slab
showing disseminated interstitial sulfides (arrowed).
Figure 2
host Ni–Cu-platinum group element (PGE) mineraliza-
tion, showing relationship of ore distribution to shape of intrusion
or lava flow, as appropriate. (A) Komatiite flow-hosted style, named
after Kambalda (Western Australia). (B) Stratiform disseminated
sulfide layer within funnel- or boat-shaped layered intrusion, as at
Fazenda Mirabela (Brazil) —disseminated layer up to 100 m thick,
mined for Ni and Cu—and Munni Munni (Australia) —5m layer,
sub-ore grade PGE. (C) Sill–dike transition or elongate funnel style,
named after the Eagle deposit (Michigan, USA) and the Kalatongke
mine (Inner Mongolia, China). (D) Subhorizontal tube chonolith,
named after Nebo-Babel (Canada) and Limoeiro (Brazil).
(E) Subhorizontal ‘fat ski’ chonolith, named after Noril’sk (Russia)
and Nkomati (South Africa). Modified from Barnes et al. (2015).

and Evans-Lamswood 2015). Almost all examples show
much larger proportions of sulfide and cumulus silicate
minerals (typically olivine) within the flow or intrusion
than could have been dissolved in a volume of magma
equal to that of the host body. This suggests that magma
flowed through the conduit leaving crystals and sulfide
liquid behind. Commonly, there is evidence of thermal
or thermomechanical erosion, in the form of transgressive
footwall troughs beneath komatiite flows (Fig. 2A), or in
tube-like or ‘ski-like’ intrusions (chonoliths) that truncate
layering within the country rock and often contain
partially digested wall-rock fragments (Figs. 2D, E). Such
host bodies are usually very small compared to the total
volume of magmatism in the province: in the case of the
ore-hosting intrusions of the Noril’sk-Talnakh camp, it is
about 1 millionth the total volume of the Siberian Trap
lavas.
Tectonic Setting
Magmatic sulfide ore deposits generally occur in intracra-
tonic settings, commonly associated with mantle plume
activity. Sulfide-rich Ni– Cu-dominant deposits are
commonly located close to the margins of ancient Archean

E lements 90 A pr il 2017
 cratonic blocks. This is thought to be the result of plume
impingement at the base of the craton, with consequent
channeling of the magmas into major crustal fault systems
around the margins (Begg et al. 2010). A number of Ni–Cu
deposits, including those of the Central Asian Orogenic Belt
in China, appear to be associated with convergent tectonics
and subduction processes rather than with plumes (Li et
al. 2012). PGE-dominant deposits in layered intrusions are
more commonly located in the interiors of stable cratons.
THE NATURE OF MAGMATIC SULFIDE ORES
Mineralogy
Magmatic sulfide ores range in sulfide contents from less
than a tenth of a percent in some stratiform PGE ores to
100% sulfide in some Ni–Cu deposits (Fig. 1). Almost all
unaltered magmatic sulfide ores, regardless of sulfide mode,
have a characteristic assemblage of pyrrhotite–pentlandite–
chalcopyrite–platinum-group minerals (PGM). This assem-
blage formed from the cooling and crystallization of a
magma-derived sulfide matte. Natural mattes, consisting
predominantly of Fe, Ni, Cu and S, fractionate to form a
sequence of phases on cooling. Below ~1,100 °C, (Ni,Fe)S
monosulfide solid solution crystallizes to leave a Cu-rich
sulfide liquid enriched in Pt, Pd and semi-metals (e.g. Te,
Bi, As). At ~900 °C, the Cu-rich liquid crystallizes to inter-
mediate solid solution (approximately CuFeS2 ) to leave a
residual melt progressively enriched in Pt, Pd and semi-
metals (Li et al. 1996). This fractionation process takes
place on scales from that of individual centimetre-size
globules (Fig. 3) to entire orebodies, e.g. the supergiant
Cu–PGE rich Oktyabrysky orebody at Talnakh in Siberia
(Naldrett 2004). On further cooling to below ~700 °C,
monosulfide solid solution breaks down to pyrrhotite and
pentlandite (Fig. 3, Fig. 4E), intermediate solid solution
to chalcopyrite, and the low-T residual liquid crystallizes
Pt- and Pd tellurides, bismuthotellurides and arsenides.
The common co-occurrence of magnetite arises from the
ability of mattes to also dissolve substantial amounts of
FeO (Naldrett 2004).
1.5 mm
Pyrrhotite Chalcopyrite Pentlandite Olivine
A differentiated sulfide liquid globule from the Bear
Figure 3
Creek open pit of the Noril’sk-1 intrusion (Russia). The
Cu-rich upper zone (orange) is the crystallization product of a
residual liquid; the lower part of the globule is composed of
multiple stages of pentlandite (blue) exsolution from pyrrhotite
(green) that was originally formed as a monosulfide solid solution
(Fe,Ni)S cumulus phase.
E lements
Sulfide Ore Textures and Evidence for
Magmatic Origins
Textural relationships between sulfides and their host
silicates are key evidence for their origin (Fig. 4). One of
the critical textures, from an historical point of view, is
that interpreted by Hawley (1962) as a frozen emulsion of
immiscible silicate and sulfide liquids (Fig. 4A). Hawley’s
was one of the first papers to argue persuasively for the
primary magmatic origin of the Sudbury ores. Other
diagnostic magmatic features are (1) net- or matrix
textures (Fig. 4B), where sulfides form a continuous 3-D
matrix enclosing cumulus silicates; (2) interspinifex ores
in komatiites (Fig. 4C), where sulfide occupies the original
spaces between dendritic olivine plates (Barnes et al. 2017);
(3) sub-spherical globular ores, sometime associated with
infilled vesicles (Le Vaillant et al. 2017) (Fig. 4D); (4) breccia
textures, where sulfide liquid has percolated through the
pore space between wall rock clasts in an intrusion breccia
(Fig. 1B).
The other strong line of evidence for magmatic, as opposed
to hydrothermal, origins is in the ore chemistry, discussed
further below. The suite of elements concentrated in these
ores (Ni, Cu, Fe, Se, PGEs) are exactly those that are known
from experimental evidence to partition strongly into
immiscible sulfide liquids. Other ‘chalcophile’ elements,
such as Zn, Pb and Sn, are present only in trace propor-
tions in magmatic sulfide ores, consistent again with their
known low partition coefficients (Naldrett 2004).
ORE-FORMING PROCESSES
The great majority of magmatic sulfide deposits form from
much the same sequence of three processes: 1) Generation
of a sulfide–silicate liquid emulsion; 2) Physical separation
of a mixture of sulfide liquid droplets and cumulus silicate
minerals from this emulsion; 3) Deposition and coales-
cence of sulfide liquid in specific sites. In some cases, the
final disposition of the ores is influenced by post-deposi-
tion migration of coalesced sulfide liquid pools, driven
by the balance between surface tension and gravitational
forces (Barnes et al. 2017). Tectonic deformation can lead
to further modifications due to the differential rheology
of solid sulfides and silicates during strain.
Generation of Sulfide Liquids
This fundamental process can happen by a variety of
mechanisms.
Partial Melting of Sulfide-Bearing Mantle
Sulfide liquid can be generated at source where the degree
of partial melting is low enough to produce melting of the
sulfide component of the source but the sulfide liquid does
not completely dissolve in the silicate partial melt. The S
content of silicate magmas in equilibrium with matte (S
content at sulfide liquid saturation, or SCSS) increases with
decreasing pressure such that sulfide-saturated magmas
generated in the mantle are likely to be undersaturated
on arrival in the upper crust (Edmonds and Mather 2017
this issue).
Fractional Crystallization of Silicate Magma
Sulfur behaves as an incompatible element under sulfide-
undersaturated conditions. Thus, fractional crystallization
causes S content to increase. Sulfide liquation (i.e. nucle-
ation and growth of immiscible sulfide liquid droplets)
occurs once the S content of the magma exceeds the
SCSS, which itself decreases with decreasing tempera-
ture and Fe content. Within cumulate sequences, the first
91 A pr il 2017
 A S Fe Ca B C
Sil
Sil
Ol
Po Pn Po+Pn
2 mm
2 cm
Opx
D E
Ol
Ol+Plg+Cpx
Cpy
Po+Pn
1 cm
Five diagnostic textures of magmatic sulfide–silicate
Figure 4
intergrowths, (A) Emulsion textures showing mixtures
of sulfide and silicate (Sil) quenched liquids. From the Frood
deposit, Sudbury (Canada). (B) “Net-textured” or “matrix” ore,
showing chains and clusters of olivine grains in a continuous matrix
of pyrrhotite (Po) and pentlandite (Pn). From the Kattiniq deposit
(Canada). (C) Interspinifex ore–sulfide liquid (Sul) penetrating
between original spinifex olivine plates (blue/green) at the top of
a partially eroded komatiite flow. From Kambalda (Australia).
(D) Globular ores showing flattened, differentiated sulfide globules
onset of sulfide liquid saturation can generate extremely
PGE-enriched mattes, as in the Platinova Reef of the
Skaergaard intrusion (Greenland) (Nielsen et al. 2015).
Mixing of Two Magmas Both of which are
at or close to Sulfide-Liquid Saturation
This mechanism of mixing two magmas both of which
are at or close to sulfide-liquid saturation can give rise to
a hybrid magma with transient sulfide supersaturation.
This process has been invoked to explain the origin of
PGE reefs associated with major magma influxes in large
chambers, such as the Merensky Reef of the Bushveld
Complex (Campbell et al. 1983).
Incorporation of External Crustal S,
Giving Rise to Sulfide ‘Xenomelts’
Addition of external S is regarded as the dominant process
in the formation of all komatiite-hosted ores (Lesher 1989)
and in the great majority of intrusion-hosted deposits
(Ripley and Li 2013). Crustal rocks can have S isotope and
S/Se ratio signatures that are very distinct from mantle S.
Therefore, any sulfide in an orebody that has a contribution
from crustal sources can be traced using their S isotope
and S/Se compositions. Examples do exist of deposits
with mantle-like S isotope signatures: most notably, the
E lements
Po+Pn Po+Pn
Ol+Plg+Cpx
Sil
2 cm S Fe Si
Pn
Cpy
Po
Xen
Pn
1 cm
Ni Cu S
in a matrix of fine grained olivine (Ol) orthocumulate with intersti-
tial plagioclase (Plg) and clinopyroxene (Cpx). Globules are capped
by former vapour bubbles (arrowed) now infilled by fine-grained
silicate phases (Le Vaillant et al., 2017). From the Noril’sk-1 intru-
sion (Russia). (E) Coarse grained “loop texture” formed by grain
boundary exsolution of pentlandite (Pn) from pyrrhotite (Po), in
semi-massive ore with silicate xenoliths (Xen) and minor chalcopy-
rite (Cpy). From the Nova deposit, Western Australia. Note that (A),
(C) and (E) are multi-element microbeam XRF maps.
giant Jinchuan deposit (China) (Ripley et al. 2005). But
it is likely that, in such cases, extensive equilibration of
sulfide with much larger volumes of silicate melt (see
below) has caused the final signature to be dominated by
the mantle S component. A variety of mechanisms exist for
incorporating external S, but direct melting of physically
incorporated sulfidic country-rock fragments (xenoliths) to
form sulfide ‘xenomelts’ is the fastest and most effective
(Robertson et al. 2015).
COMPOSITION OF MAGMATIC
SULFIDE ORES
Variability in Ni and Cu Content
The first-order variability in Ni and Cu tenors (tenor =
concentration of the metal in 100% sulfide) is related to the
composition of the host rocks (Fig. 5A). There is a decrease
in the Ni:Cu ratio from values around 20:1 in komatiites,
4:1 in ores associated with komatiitic basalts, and between
0.5:1 to 5:1 in most deposits (both Ni–Cu sulfide-rich
type, and reef-style low sulfide PGE type) associated with
mafic magmas. Lower Ni:Cu ratios are found in ores where
there has been extensive sulfide liquid fractionation, as at
Noril’sk-Talnakh, the Sudbury footwall veins (Naldrett et al.
1997) and in disseminated ores associated with advanced
fractionation of tholeiitic mafic magmas.
92 A pr il 2017
 PN 1 Controls on Metal Tenors
u=
A
A Ni/
C
Magma Composition
R=100,000
.25
0 The variability in Ni/Cu shown in Figure 5A arises largely
10 C u=
Ni/ PR 5 from parent magma compositions. Values of Ni/Cu in
Cu in 100% sulfide wt %

ores correspond reasonably well with those in the parent

R=500 s magmas themselves (Barnes and Lightfoot 2005), which
s alt
Ba for mafic host magmas range from high-Mg basalts through
R=100,000
u ry .
db om 20 to fractionated tholeiites. Nickel becomes depleted in more
Su K
evolved magmas due to olivine crystallization, whereas
R=100
1 Cu becomes enriched due to incompatibility in the major
fic R=1000 crystallizing silicate phases. Hence, primitive high-T
Ma es
it komatiites have very high Ni and low Cu compared with
ati more fractionated mafic magmas that attain progressively
K om
R=100 lower Ni/Cu. This contrast is reflected in sulfide composi-
tions (Fig. 5A). The most extreme Cu-rich example is the
R=50 Cu vs Ni Platinova Reef (PN in Fig. 5) of the Skaergaard Intrusion
0.1 (Nielsen et al. 2015). Here, sulfide saturation occurred
1 10 very late in the crystallization history, such that the more
Ni in 100% sulfide wt %
Layered intrusion compatible Ni had already been depleted from the melt,
Mafic R factor model Kom R factor model
Reef sulfides whereas the more incompatible Cu was strongly enriched.
10,000 PN In this case, the magma attained saturation in Cu–Fe sulfide
B
B liquid rather than the typical Fe–Ni-dominated sulfide.

1000 R=100,000 The PGE contents of silicate melts are controlled through
R=100,000 the variable degree of igneous compatibility of the different
100 PGEs: moderately compatible in the case of Ir, Os and Ru,
)=
Pd in 100% sulfide ppm

100 u(
P d/C
10 pm
Pd/
Cu(
1.0
s Su
te
atii
Ko
m Mass Balance
0.1
.01
0.1
Ore compositions and bulk metal tenors (concentra-
Figure 5
tion of Cu, Ni and Pd in 100% sulfide) from the
complete spectrum of magmatic sulfide ore types. (A) Cu vs Ni;
(B) Pd vs. Cu. Coloured fields contain 95% of data for average bulk
compositions of individual ore deposits associated with komatiites,
mafic-parented intrusions, komatiitic basalts and the Sudbury
(Canada) impact melt sheet. Yellow field shows estimates of sulfide
tenors from various sections of the Platreef (PR). Red symbols are
individual deposit data points from PGE-rich reefs in layered mafic
intrusions including the Bushveld Complex; PN indicates approxi-
mate composition of sulfides from the Pd zone of the Platinova
Reef, Skaergaard intrusion. Curves show results of simple partition
coefficient/mass ratio (D/R) calculations (Eq. 1 in text) for model
komatiite and mafic magma starting compositions, assuming D
values of 150 and 400 for Ni in komatiite and basalt respectively;
1,000 for Cu, and 100,000 for Pd. Pd/Cu(pmn) is Pd/Cu normalised
to ratio in primitive upper mantle.
Variability in PGE Content
Platinum group element (PGE) tenors (represented in Fig.
5B by Pd) show a much wider range than those of Ni and
Cu. Platinum group element tenors range over nearly 6
orders of magnitude between the most depleted Ni–Cu
deposits to the most enriched, reef-style ores. Within each
deposit/setting type there is a strong positive correlation
between Pt and Pd, and a weaker correlation between Pd
and Cu (Fig. 5B) and between Pd and Ni contents, particu-
larly if the reef examples are considered along with the
mafic-hosted grouping (Campbell and Naldrett 1979).
pmn
R=1000 and strongly incompatible for Pt and Pd. Superimposed on
R=10000
1
this effect is the very strong tendency of all the PGEs to
PR )=
0 C u(p
mn become strongly depleted during fractional extraction of
n) =
1
alts Pd/
s sulfide liquid, owing to their extreme chalcophile character
Ba
m. (Mungall and Brenan 2014). A further consequence of this
100 Ko R=100
y
attribute of the PGEs is their strong susceptibility to mass
r
dbu 0.1 balance effects.
)=
u(p
mn
Pd/
C
fic A major control on sulfide liquid compositions is the mass
Ma
ratio, R, of silicate to sulfide liquid that react with one
Pd vs Cu another (Campbell and Naldrett 1979). This relationship
is expressed as follows:
1 10
[Xisul · Disul · (1 + R)] (1)
Cu in 100% sulfide wt % Yisul = –––––––––––––––––––––––
(R + Disul)
where Yisul is the final concentration of element i in the
sulfide liquid; Xisil is the initial concentration in the silicate
liquid; Disul is the partition coefficient between sulfide and
silicate melt. Formation of magmatic sulfides is treated here
as a batch equilibrium process: a batch of sulfide liquid
forms and equilibrates with a given volume of silicate melt.
The effect of variations in R is shown on the model curves
in Figure 5. Where R for element i is very low compared
with Disul, tenors are relatively low; there is relatively little
impact of changing D on the content of i in either melt,
and tenor depends almost linearly on R. If R is large relative
to Disul, then the opposite applies: both silicate and sulfide
melt have high metal contents that increase almost linearly
with D. The effect operates in natural systems through the
wide range in the partition coefficients for the different
chalcophile metals: typically, around the low hundreds
for Ni, ~1,000 for Cu, and of the order of hundreds of
thousands for the PGEs (Kiseeva et al. 2017 this issue).
Hence, the extremely chalcophile PGEs are much more
susceptible to R factor effects than Ni and Cu, as can be
seen in the model curves in Figure 5. Nickel and Cu tenors
approach maximum values where R is greater than about
1,000, while the PGE tenors continue to increase almost
indefinitely with increasing R owing to their extreme D
values. Extremely high silicate:sulfide ratios are necessary
to produce the high PGE tenors of reef-style deposits.

E lements 93 A pr il 2017
Sulfide liquid differentiation can produce significant S, lowering the PGE tenor of the sulfides and modifying
variability in the proportion of Cu to Ni and of Pt and Pd the isotopic and PGM compositions to different degrees
to Ir, Ru and Os, and is progressively more important in depending on the country-rock lithology (McDonald and
orebodies containing initially higher Cu contents where Holwell 2011).
the melting range of the sulfide component extends to
Recent exploration in the northern limb has identified a
much lower temperatures. This effect produces dispersion
much more regularly layered sequence, with traceable strat-
at metre- to decimetre scale within orebodies and at the
iform mineralization down dip from the surface outcrop
centimetre scale in individual droplets, as seen in Figure 3.
of the Platreef, with the addition of thick underlying
But in some very large systems, differentiation is accompa-
ultramafic cumulates. This ‘Flatreef’ sequence has many
nied by physical migration of residual Cu-rich liquid into
similarities to the Merensky Reef in the rest of the complex
veins and fractures. At Sudbury, this process generates very
and represents a transition from stratabound ‘contact-type’
high grade Cu–PGE orebodies hundreds of metres below
or ‘marginal’ styles, to stratiform, reef-style mineralization.
the base of the host magma body (Naldrett et al. 1997).

THE ELEPHANT IN THE SHIELD:

CASE STUDY: THE PLATREEF
SUDBURY (CANADA)
OF SOUTH AFRICA
No review of magmatic sulfide ores could ignore the world’s
The giant layered ultramafic–mafic Bushveld Igneous
largest known accumulation of magmatic ores: the extraor-
Complex (South Africa) hosts over three-quarters of the
dinary Sudbury impact structure in the southern part of
world’s PGE resources in three main deposits: the UG2
the Canadian Shield, Ontario. In this case, ore formation
chromitite, the Merensky Reef and the Platreef. The former
followed wholesale melting of almost the entire thickness
two are archetypal stratiform reef deposits, with a few
of the crust following a giant bolide impact (Mungall et
percent of PGE-rich sulfides associated with chromite and
al. 2004). Sulfide liquid segregated from the resulting
having thicknesses of a few centimetres to a few metres.
melt sheet on subsequent cooling. While many aspects
The Platreef is in the northern part of the complex and is
of Sudbury ore genesis are unique, one aspect is highly
a much thicker orebody (~10–400 m) emplaced as a series
significant to understanding processes in other deposits:
of sills with stratabound sulfides present in a package
the presence of extensively mobilized veins and dikes of
of mostly pyroxenites that rest directly on Archaean–
sulfide rich rocks that extend for distances of kilometres
Proterozoic country rocks (Fig. 6). It has been interpreted
below the original base of the melt sheet. These features
as the lateral equivalent of the Merensky Reef and as the
attest to the extreme physical mobility of sulfide magmas,
propagating marginal facies of the Bushveld Complex. The
or possibly sulfide-rich melt emulsions charged with rock
Platreef was probably formed as magma was squeezed out
fragments, driven by gravity into fracture systems. This
at the edges of the expanding magma chamber (Naldrett
is an important clue to the origin of late-stage injections
et al. 2008). The Platreef is likely to be the main future
of sulfide-rich breccia-textured ores, common in many
source of PGEs in the coming decades, making it one of
intrusion-hosted deposits (Barnes et al. 2016).
the most economically significant of all known magmatic
sulfide ore deposits.
RESEARCH QUESTIONS
The Platreef displays much complexity due to its multi-stage
Research on magmatic sulfides has made substantial
origin (McDonald and Holwell 2011). Mass independent S
advances over the five decades from the pioneering work
isotopes and S/Se ratios give evidence that sulfide satura-
of Naldrett and others, particularly in understanding
tion was initially triggered by a bulk assimilation event,
geochemical processes. But a number of fundamental
most likely of pyritic shales at depth (Penniston-Dorland et
questions remain about physical processes of ore formation.
al. 2008). Contaminated magma with a cargo of dispersed
sulfide liquid droplets was then emplaced higher in the 1. Is the addition of external S through assimilation of
crust, at which point interaction with the diverse range crustal rocks essential, and by what mechanisms does
of country rocks resulted in further addition of external this addition take place?

Block model showing a

Figure 6
schematic long section
through the Platreef and underlying
footwall lithologies of the Bushveld
Igneous Complex (South Africa), and
sections through normal westerly
dipping Platreef at the surface, with
transition to shallow-dipping ‘Flatreef’
at depth.

E lements 94 A pr il 2017
2. How is sulfide liquid transported in magmas, in what
proportions and in what physical form? How far can
sulfide liquids be transported from the original site of
liquation to the point of deposition?
3. What is the physical process of deposition, and to what
extent is it governed by mechanical sedimentation
versus in-situ chemical deposition of sulfide liquid at
the point of nucleation?
Magmatic sulfides continue to be fascinating topics of
research, leading to advances in exploration models as
well as new insights into magmatic, and even climatic,
processes. Applications have been made to meteoritics, in
studies of the origin of planetary cores, and into processes
of metal and S transport with implications for the origin
of porphyry deposits, climate-impacting giant eruptions,
and mass extinctions (Mungall et al. 2015; Le Vaillant et
al. 2017).
ACKNOWLEDGMENTS
We thank Kate Kiseeva and the Elements editorial team
for the kind invitation to contribute to this volume. Anais
Pages and Angus MacFarlane reviewed a preliminary draft.
Reviews by Tony Naldrett, Ed Ripley and Vera Lorenz
greatly improved the clarity of the manuscript. SJB and
MLV acknowledge support from the CSIRO Science Leader
program. We acknowledge the many colleagues who have
contributed data and insights drawn on here.
The image in Figure 3 was collected on the X-ray Fluorescence
Mapping beamline of the Australian Synchrotron, Clayton,
Victoria, Australia. We acknowledge financial support for
this facility from the Science and Industry Endowment
Fund (SIEF).

REFERENCES
Barnes S-J, Lightfoot PC (2005) Formation
of magmatic nickel sulfide ore deposits
and processes affecting their copper
and platinum group element contents.
Economic Geology 100: 179-214
Barnes SJ (2006) Komatiite-hosted
nickel sulfide deposits: geology,
geochemistry, and genesis. In: Barnes
SJ (ed) Nickel Deposits of the Yilgarn
Craton: Geology, Geochemistry, and
Geophysics Applied to Exploration.
Society of Economic Geologists, Special
Publication 13, pp 51-118
Barnes SJ, Cruden AR, Arndt NT, Saumur
BM (2016) The mineral system approach
applied to magmatic Ni–Cu–PGE
sulphide deposits. Ore Geology Reviews
76: 296-316
Barnes SJ and 8 coauthors (2017) Sulfide-
silicate textures in magmatic Ni-Cu-PGE
sulfide ore deposits: disseminated
and net-textured ores. American
Mineralogist 102: 473-506
Begg GC and 5 coauthors (2010)
Lithospheric, cratonic, and geodynamic
setting of Ni-Cu-PGE sulfide deposits.
Economic Geology 105: 1057-1070
Campbell IH, Naldrett AJ (1979) The
influence of silicate:sulfide ratios on
the geochemistry of magmatic sulfides.
Economic Geology 74: 1503-1506
Campbell IH, Naldrett AJ, Barnes SJ
(1983) A model for the origin of the
platinum-rich sulfide horizons in the
Bushveld and Stillwater Complexes.
Journal of Petrology 24: 133-165
Edmonds M, Mather TA (2017) Volcanic
sulfides and outgassing. Elements 13:
107-112
Hanski EJ, Luo Z-Y, Oduro H, Walker
RJ (2011) The Pechenga Ni-Cu sulfide
deposits, Northwestern Russia: A
review with new constraints from the
feeder dikes. In: Li C, Ripley EM (eds)
Magmatic Ni-Cu and PGE Deposits:
Geology, Geochemistry, and Genesis.
Reviews in Economic Geology 17, pp
145-162
Hawley J (1962) The Sudbury ores,
their mineralogy and origin; Part
1, the geological setting. Canadian
Mineralogist 7: 1-29
Keays RR, Lightfoot PC (2004) Formation
of Ni-Cu-platinum group element
sulfide mineralization in the Sudbury
impact melt sheet. Mineralogy and
Petrology 82: 217-258
Kiseeva ES, Fonseca ROC, Smythe DJ
(2017) Sulfides in the upper mantle:
controls over the chalcophile element
budget in magmatic systems. Elements
13: 113-118
Le Vaillant M, Barnes SJ, Mungall JE,
Mungall E (2017) Role of de-gassing
of the Noril’sk nickel deposits in the
Permo-Triassic mass extinction event.
Proceedings of the National Academy
of Sciences, in press and online,
doi/10.1073/pnas.1611086114
Lesher CM (1989) Komatiite-associated
nickel sulfide deposits. In: Whitney
JA, Naldrett AJ (eds) Ore Deposition
Associated with Magmas. Economic
Geology Publishing Company, El Paso,
pp 44-101
Li C, Barnes S-J, Makovicky E,
Rose-Hansen J, Makovicky M (1996)
Partitioning of nickel, copper, iridium,
rhenium, platinum, and palladium
between monosulfide solid solution
and sulfide liquid: effects of composi-
tion and temperature. Geochimica et
Cosmochimica Acta 60: 1231-1238
Li CS and 5 coauthors (2012) The
Kalatongke magmatic Ni–Cu deposits
in the Central Asian Orogenic Belt, NW
China: product of slab window magma-
tism? Mineralium Deposita 47: 51-67
Lightfoot PC, Evans-Lamswood D (2015)
Structural controls on the primary
distribution of mafic–ultramafic intru-
sions containing Ni–Cu–Co– (PGE)
sulfide mineralization in the roots of
large igneous provinces. Ore Geology
Reviews 64: 354-386
McDonald I, Holwell D (2011) Geology
of the northern Bushveld Complex and
the setting and genesis of the Platreef
Ni–Cu–PGE deposit. In: Li C, Ripley
EM (eds) Magmatic Ni–Cu and PGE
Deposits: Geology, Geochemistry, and
Genesis. Reviews in Economic Geology
17, pp 297-327
Mudd GM, Jowitt SM (2014) A detailed
assessment of global nickel resource
trends and endowments. Economic
Geology 109: 1813-1841
Mungall JE, Brenan JM (2014)
Partitioning of platinum-group
elements and Au between sulfide liquid
and basalt and the origins of mantle-
crust fractionation of the chalcophile
elements. Geochimica et Cosmochimica
Acta 125: 265-289
Mungall JE, Ames DE, Hanley JJ (2004)
Geochemical evidence from the
Sudbury structure for crustal redistribu-
tion by large bolide impacts. Nature
429:546-548
Mungall JE, Brenan JM, Godel B, Barnes
SJ, Gailard F (2015) Transport of metals
and sulphur in magmas by flotation
of sulphide melt on vapour bubbles.
Nature Geoscience 8: 216-219
Naldrett AJ (2004) Magmatic Sulphide
Deposits: Geology Geochemistry and
Exploration. Oxford University Press,
New York, 728 pp
Naldrett AJ, Ebel DS, Asif M, Morrison
G, Moore CM (1997) Fractional crystal-
lization of sulfide melts as illustrated at
Noril’sk and Sudbury. European Journal
of Mineralogy 9: 365-377
Naldrett T, Kinnaird JA, Wilson AH,
Chunnett G (2008) Concentration of
PGE in the Earth’s crust with special
reference to the Bushveld Complex.
Earth Science Frontiers 15: 264-297
Nielsen TFD and 8 coauthors (2015) The
Skaergaard PGE and gold deposit: the
result of in situ fractionation, sulphide
saturation, and magma chamber-scale
precious metal redistribution by immis-
cible Fe-rich melt. Journal of Petrology
56: 1643–1676
Penniston-Dorland SC and 6 coauthors
(2008) Multiple sulfur isotopes reveal
a magmatic origin for the Platreef
platinum group element deposit,
Bushveld Complex, South Africa.
Geology 36: 979-982
Ripley EM, Brophy JG, Li C (2002)
Copper solubility in a basaltic melt and
sulfide liquid/silicate melt partition
coefficients of Cu and Fe. Geochimica
et Cosmochimica Acta 66: 2791-2800
Ripley EM, Li C (2013) Sulfide saturation
in mafic magmas: Is external sulfur
required for magmatic Ni-Cu-(PGE) ore
genesis? Economic Geology 108: 45-58
Robertson JC, Ripley EM, Barnes SJ,
Li C (2015) Sulfur Liberation from
Country Rocks and Incorporation in
Mafic Magmas. Economic Geology 110:
1111-1123

E lements 95 A pr il 2017
Next Generation Excellence in
Mass Spectrometry

Find Out Why So Many Scientists Choose Isotopx For

Their TIMS And Noble Gas MS Solutions. Now Also
Providing Electrronics And Software Upgrades For VG,
GV, Micromass and MAP Noble Gas MS Systems
www.isotopx.com