Economic Geology; January 2005; v. 100; no. 1;
p. 43-61; DOI: 10.2113/100.1.0043
© 2005 Society of Economic Geologists
Geology, Petrology, and Controls on PGE Mineralization of the Southern Roby and Twilight Zones, Lac des Iles Mine, Canada
J. G. Hinchey
and
K. H. Hattori
Department of Earth Sciences, University of Ottawa, Ottawa, Ontario, Canada K1N 6N5
M. J. Lavigne*
North American Palladium Ltd., Thunder Bay, Ontario, Canada P7B 6T9
e-mail, john.hinchey{at}science.uottawa.ca
 |
Abstract
|
|---|
The Lac des Iles Pd mine, with reserves of 88 million metric tons (Mt)
containing 1.51 g/t Pd, is hosted by the 2.69 Ga Lac des Iles intrusive complex
in the southern Wabigoon subprovince of the Superior province of Canada. The
known economic concentration of Pd occurs in the Mine Block intrusion, the
central body of the intrusive complex, where gabbroic rocks range from
leucogabbro to pyroxenite and show complicated textures, such as breccias, magma
mingling, and pods and veins of pegmatite. The ore is characterized by low
concentrations of sulfide (typically less than 3 vol %) and exceptionally high
Pd contents (Pd/Pt ~10, Pd/Au ~13). Detailed mapping of the southern Roby and
Twilight zones shows that early leucocratic rocks are barren and that the bulk
of Pd was introduced by late melanocratic magmas. The average concentration of
Pd in the melanocratic rocks, excluding the High-Grade zone, is estimated to be
~4 ppm. The concentrations of sulfur correlate positively with those of base
metals, platinum group elements (PGE), and Au. Furthermore, sulfide grains
commonly show exsolution textures. The evidence suggests a magmatic origin of
the PGE mineralization where the PGE were concentrated in an immiscible sulfide
melt in the parental magma.
Bulk chemical compositions suggest that all mafic igneous rocks in the
mineralized zones, except for late clinopyroxenite, are cogenetic. The
hypothetical parental magmas have high MgO and low (~15x chondrite),
unfractionated rare earth elements (REE) with (Ce/Yb)chondrite <3,
suggesting that the magmas formed through high degrees of partial melting in a
moderately depleted mantle. Fractional crystallization of olivine, chromite, and
high-temperature platinum group minerals (PGM) resulted in high concentrations
of Pt, Pd, and Rh relative to Ir, Os, and Ru in the parental magmas. Extreme
enrichment of Pd in the late melanocratic magmas is interpreted to have been
attained through the incorporation of earlier formed sulfide melt. This
interpretation is supported by high Cu/Pd in early barren leucocratic rocks and
low Cu/Pd in fertile melanocratic rocks. Rocks in the volumetrically minor but
economically important High-Grade zone (>35% of Pd in the deposit) on the
eastern margin of the Roby zone have much higher concentrations of Pd than any
other rocks and do not show correlations between sulfur and precious and base
metals. Furthermore, the rocks are intensely and pervasively altered to
actinolite, talc, anthophyllite, hornblende, chlorite, sericite, calcite, and
quartz. These observations suggest subsolidus enrichment of Pd and mobility of
S. The lack of apparent fluid pathways within the High-Grade zone and the
distribution of the zone are consistent with the enrichment of Pd at high
temperatures by fluids that originated from the mafic magmas.
The textures of the Lac des Iles deposit are similar to those of contact-type
PGE deposits, but there are fundamental differences between the two. The Lac des
Iles deposit is not localized near the contact between the host intrusion and
the country rocks and evidence of the assimilation of the host rocks is lacking.
Instead, the mineralization at Lac des Iles has many features in common with
layered intrusion-hosted deposits, in which pulses of primitive magma introduced
the PGE. Unlike the quiescent magma chambers of most layered deposits, the
magmas at Lac des Iles were intruded energetically, forming breccias and
magma-mingling textures.
|
Introduction
|
|---|
THE LAC DES ILES MINE is the only primary Pd producer in Canada. The deposit
is hosted by the Neoarchean Lac des Iles mafic-ultramafic intrusive complex in
the southern Wabigoon subprovince of the Superior province (Fig. 1). In contrast
to many platinum group element (PGE) deposits, which are hosted by Ni sulfides
at the base of large mafic intrusions (e.g., Sudbury and Noril’sk) or as
stratiform layers within large layered intrusions (e.g., Bushveld Intrusion),
the Lac des Iles deposit is characterized by low contents of sulfides, common
alteration of silicate minerals, and complex igneous textures, such as breccias
and pegmatites (Lavigne and Michaud, 2001). The mineralization is extremely
enriched in Pd compared to other PGE, with Pd/Pt ratios of ~10 and Pd/Ir ratios
of ~10,000. These features are not easily explained by genetic models for other
PGE deposits. Previously proposed origins for the deposit include hydrothermal
and/or deuteric mineralization based on the occurrence of Bi-bearing platinum
group minerals (PGM; Watkinson and Dunning, 1979) and the close association of
secondary hydrous silicate minerals with PGM (Talkington and Watkinson, 1984), and a constitutional zone-refining process (Brügmann et al., 1989) based on
high Pd contents and coexisting melanocratic and leucocratic rocks in the
mineralized zones. The latter process, as originally proposed by McBirney
(1987), involves partial melting of gabbro cumulates.
View larger version (81K):
[in this window]
[in a new window]
|
FIG. 1. Regional geology of the western Superior province, illustrating the
boundaries of subprovinces, the locations of the Nipigon plate and the Lac des
Iles mine (modified from Ontario Geological Survey, 1991). The subprovince names
are shown in italics. The Nipigon plate is the northern extension of the igneous
province related to the Midcontinental rift. The inset displays a simplified map
of the Superior province. The square outlines the area shown in the figure.
|
|
We conducted detailed mapping of the southern Roby and Twilight zones at
scales of 1:60 and 1:120, respectively, in the summers of 2002 and 2003. Field
observations by Hinchey et al. (2003) show a clear relationship between
mineralization and rock types. This paper documents new data on the geochemistry
of the different rock types and discusses the origin of the mineralization.
Previous work and exploration history
Geologic investigations in the area began with reconnaissance mapping by
Jolliffe (1934), followed by more detailed mapping of the area by Pye (1968).
Economic interest in the area was sparked by the discovery of aeromagnetic
anomalies in the late 1950s. Significant Pd mineralization was first discovered
in the Roby zone in 1963 by prospectors and was subsequently investigated by
Gunnex Ltd. and Anaconda Ltd. The Ontario Geological Survey conducted several
mapping projects in the area (Sutcliffe and Sweeny, 1986; Macdonald, 1988;
Sutcliffe et al., 1989), and M.Sc. thesis projects on the deposit were completed
by Dunning (1979), Sweeny (1989), and Michaud (1998).
Madeleine Mines Ltd. commenced mining in 1990 but this lasted only several
months. Lac des Iles Mines Ltd. began production in 1993 at a rate of 2,000 t/d.
Today the Lac des Iles mine is an open-pit operation conducted by North American
Palladium Ltd. Current proven and probable reserves consist of 88 million metric
tons (Mt) grading 1.51 g/t Pd, 0.17 g/t Pt, 0.12 g/t Au, 0.06 percent Cu, and
0.05 percent Ni, with additional measured and indicated resources of 65 Mt
grading 1.58 g/t Pd, 0.17 g/t Pt, 0.11 g/t Au, 0.05 percent Cu, and 0.05 percent
Ni (North American Palladium Ltd. Annual Report, 2002).
|
Geology
|
|---|
Regional geologic setting
The Lac des Iles intrusive complex is located in the granite-greenstone
terrane of the Wabigoon subprovince of the Archean Superior province of the
Canadian Shield (Blackburn et al., 1992; Fig. 1). The intrusive complex is one
of a suite of Late Archean mafic-ultramafic plutons in the southern Wabigoon
subprovince close to the Quetico subprovince (Pye, 1968; Sutcliffe, 1986). It is
part of a 30-km-diameter circular array of mineralized mafic and ultramafic
intrusions, including Legris Lake, Tib Lake, and Buck Lake, of which the Lac des
Iles intrusive complex is the largest (Gupta and Sutcliffe, 1990; Fig. 2).
View larger version (65K):
[in this window]
[in a new window]
|
FIG. 2. Simplified regional geologic map of the study area, showing the
occurrence of mafic and/or ultramafic intrusions with PGE mineralization
(modified after Sutcliffe and Smith, 1988). Note the circular distribution of
mafic-ultramafic intrusions. The diabase sills belong to the igneous rocks of
the Nipigon plate.
|
|
Regional metamorphic grades in the area are low, up to greenschist facies (Blackburn
et al., 1992; Pettigrew and Hattori, 2002), which contrasts with granulites and
migmatites in the adjacent Quetico subprovince (Williams, 1992; Fig. 2). The low
grade of regional metamorphism in the study area is supported by the presence of
pristine clinopyroxene and orthopyroxene in the northern part of the complex and
in the gabbronorite of the Mine Block intrusion.
Geology of the Lac des Iles complex
The Lac des Iles intrusive complex (2692+4–2 Ma; Blackburn et
al., 1992) is subdivided into three main intrusive bodies (Lavigne and Michaud,
2001; Fig. 3): (1) the North Lac des Iles ultramafic intrusion, centered on the
lake, (2) the Mine Block intrusion, consisting of lithologically and texturally
complex gabbroic rocks, and (3) the Camp Lake intrusion, which is a homogeneous
hornblende gabbro southwest of Camp Lake. The three intrusions are separated by
tonalitic country rocks (Lavigne and Michaud, 2001). All rocks in the area have
been intruded by diabase dikes and sills, which range from ~2120 to ~1140 Ma (Buchan
and Ernst, 2004).
The North Lac des Iles intrusion (Fig. 3) ranges from clinopyroxenite and/or
websterite in the northern part to minor gabbro and/or gabbronorite in the
southern part. Igneous layering is common and pristine igneous minerals are well
preserved. The ultramafic rocks are in direct contact with the tonalitic country
rocks but are also locally separated by thin gabbroic layers (Lavigne and
Michaud, 2001). The gabbro and/or gabbronoritic rocks of Mine Block intrusion
(Figs. 3–4) range from anorthosite to clinopyroxenite. They have complicated
textures, including breccias and magma-mingling structures, which vary at scales
of less than 5 m.
View larger version (71K):
[in this window]
[in a new window]
|
FIG. 4. Simplified geologic map of the Mine Block intrusion of the Lac des
Iles intrusive complex (modified after Sutcliffe and Sweeney, 1986), showing the
distribution of East Gabbro, the locations of the Roby, Twilight, and High-Grade
zones, and the outline of the phase 3 open pit.
|
|
|
Mineralized Zones
|
|---|
Anomalous concentrations of PGE occur locally throughout the Lac des Iles
intrusive complex, but known economic mineralization occurs only in the Mine
Block intrusion (Fig. 3). The mineralized zones include the Roby, Twilight, and
Baker zones (Fig. 4). The Roby zone has dimensions of 950 m north-south, 700 m
east-west, and greater than 1,000 m vertically and contains the narrow
High-Grade zone on its eastern margin (Fig. 4). The High-Grade zone is 15 to 25
m wide and 400 m long and is bounded by the barren East Gabbro to the east. This
ore zone consists of pervasively altered melanocratic rocks and dips nearly
vertically to a depth of 250 m, at which depth it shallows out to the east. The
High-Grade zone comprises only 7.6 vol percent of the current reserves (Lavigne
and Michaud, 2001) but contains approximately 35 percent of the Pd in the Roby
zone.
Most of the Roby zone, including the High-Grade zone, is being mined in the
phase 3 open pit, but the southwest part of the Roby zone was still exposed on
the surface in 2003 and was mapped in detail in this study (approx 65 to 70 m
north-south and 35 to 40 m east-west: Fig. 5). The Twilight zone, with a surface
exposure of 175 to 200 m north-south and 200 to 225 m east-west, lies southeast
of the Roby zone (Fig. 4), with 50 to 70 m of East Gabbro between the two zones.
The Baker zone is located approximately 1 km northeast from the Roby and
Twilight zones (Fig. 4) and contains rocks and textures similar to those in the
Roby and Twilight zones.
View larger version (73K):
[in this window]
[in a new window]
|
FIG. 5. Illustration of the complicated distribution of various rock types in
the southern Roby zone in the Mine Block intrusion after 1:60 scale mapping by
Hinchey et al. (2003). Rocks are divided into early leucocratic, mineralized
melanocratic, and late barren melanocratic rocks. The mineralized melanocratic
rocks are commonly accompanied by pegmatitic veins and pods of various sizes
from several centimeters to 50 cm. Only large pegmatites are shown in the map.
Felsic and diabase dikes cut all rock types of the Lac des Iles intrusive
complex. Numbers on the sides correspond to the mine grid.
|
|
|
Petrology of the Southern Roby Zone
|
|---|
The detailed mapping by Hinchey et al. (2003) classified rocks into three
groups; (1) early leucocratic rocks, (2) sulfide-bearing melanocratic gabbroic
rocks, and (3) late, sulfide-free melanocratic gabbroic rocks. The area also
contains young felsic dikes and diabase dikes that cut all rocks of the Lac des
Iles intrusive complex (Fig. 5).
Most rocks of the southern Roby zone show cumulate textures, consisting of
subhedral to euhedral cumulus crystals of clinopyroxene, plagioclase, and minor
orthopyroxene with intercumulus material consisting of the same assemblage plus
minor biotite, magnetite, ilmenite, and sulfides. Breccias are common and are
named on the basis of matrix composition. For example, a breccia with
melanogabbro matrix is referred to as a melanogabbro breccia (Fig. 6A). Gabbroic
rocks with varying grain size from fine to coarse and local pods and veins of
pegmatite are termed "varitextured" gabbro.
View larger version (105K):
[in this window]
[in a new window]
|
FIG. 6. Representative photographs of the ore zone and the concentrations of
Pd in different rock types. Solid squares show the locations of samples used for
Pd analysis after the photographs were taken. The Pd value of the melanonorite
in the center of (C) is for a sample outside the field of view. The numbers in
parentheses correspond to the rock types described in the text. Rock hammer for
scale. A. Medium-grained melanogabbro breccia (2a) containing fragments of
earlier leucocratic rocks (1b-c) in the northern section of the southern Roby
zone. Note low concentrations of Pd in leucocratic rocks and a dike of late
barren clinopyroxenite (3a) cutting the melanogabbro breccia (2a). B. Magma
mingling between early leucogabbro (1d) and late melanogabbro (2c) from the
southern portion of the southern Roby zone. C. Magmatic brecciation of late
melanonorite (6) and earlier norite (5) in the Twilight zone.
|
|
Early leucocratic rocks (group 1)
The early leucocratic rocks consist of (1a) medium-grained anorthosite and/or
leucogabbro; (1b) medium-grained gabbro, which is commonly varitextured close to
pegmatite; (1c) medium-grained, light-gray gabbro with local layering that
strikes N 64-68° E, defined by alternating layers of plagioclase and
clinopyroxene; and (1d) medium-grained, locally foliated gabbro. The
relationships between these rocks are not certain due to the lack of exposed
contacts. All leucocratic rocks have similar mineral abundances, with 45 to 65
vol percent plagioclase, 35 to 55 vol percent clinopyroxene, and less than 10
vol percent of combined orthopyroxene, biotite, and epidote. They contain less
than 1 vol percent of finely disseminated pyrite, chalcopyrite, and oxides. The
plagioclase in the rocks commonly displays cumulus texture, with crystals 1 to 7
mm in length. Intercumulus material is dominated by clinopyroxene (grain sizes
of 1–3 mm), with lesser amounts of orthopyroxene, biotite, sulfides, and
oxides. The rocks are unaltered to moderately altered, with variably sericitized
plagioclase. Clinopyroxene is commonly replaced by actinolite, hornblende,
epidote, and chlorite.
Mineralized melanocratic rocks (group 2)
The mineralized melanocratic rocks range from gabbro to clinopyroxenite,
contain sulfide, and have irregularly shaped pods and veins of pegmatite.
Brecciation and magma-mingling structures with the earlier leucocratic rocks of
group 1 are common. From north to south, the melanocratic rocks are (2a)
medium-grained meso- to melanogabbro breccia, (2b) medium- to relatively
coarse-grained clinopyroxenite, and (2c) medium-grained, dark gabbro (Fig. 5).
Pegmatite (2d) is common within and in contact with these melanocratic rocks.
Most rocks are altered with very rare relict pyroxenes and plagioclase.
Clinopyroxene is commonly replaced by a mixture of hornblende, actinolite, and
chlorite, whereas interstitial clinopyroxene is replaced by chlorite.
Plagioclase is mostly sericitized.
Medium-grained mesogabbro to melanogabbro breccia (2a): This breccia
contains fragments of leucocratic rocks of group 1 (Fig. 6A). The matrix of the
breccia is dominated by subhedral, equigranular clinopyroxene (1–6 mm, 50–80
vol %) that commonly forms aggregates, with lesser amounts of plagioclase (1–4
mm, 30–50 vol %) and minor (<5 vol %) orthopyroxene and biotite.
Interstitial minerals are plagioclase (1–4 mm with minor aggregates of small
~0.5-mm crystals), minor clinopyroxene and biotite, and sulfide and oxide
minerals. A mixture of sulfide and oxide minerals occur as blebs (1–6 mm diam)
and fine-grained disseminations (1–3 vol % of the matrix). The sulfide
minerals are mostly pyrrhotite with exsolved chalcopyrite and pentlandite. Minor
pyrite and magnetite are present, and minor amounts of calcite occur in the
blebs of sulfide minerals. Chalcopyrite also commonly occurs along cleavage
planes of actinolite.
Medium- to coarse-grained clinopyroxenite (2b): These rocks range from
massive, medium-grained clinopyroxene-rich melanogabbro to clinopyroxenite and
consist of equigranular clinopyroxene (1–7 mm, 85–100 vol %) with
interstitial clinopyroxene, plagioclase, orthopyroxene, biotite, sulfides and
oxides, and rare gahnite. The rocks have an orthocumulate texture, with a minor
adcumulate texture. Sulfides (1–5 vol %) are disseminated, blebby (0.5–2
mm), and net textured, consisting of pyrrhotite with intergrown chalcopyrite and
pentlandite. The texture suggests that the sulfides were once monosulfide solid
solution. The clinopyroxenite also contains minor chalcopyrite along cleavage
planes of secondary actinolite (Fig. 7A).
View larger version (113K):
[in this window]
[in a new window]
|
FIG. 7. Photomicrographs showing the textures of sulfide minerals. A.
Chalcopyrite (Ccp) along cleavage planes of actinolite (Act). The sample is from
the medium-grained clinopyroxenite (2b) of the southern Roby zone. B. Primary
magmatic sulfide bleb with apparent exsolution of chalcopyrite (Ccp) and
pentlandite (Pn) in late fractures within pyrrhotite (Po). The sample is from
the medium-grained melanocratic gabbro (2c) of the southern Roby zone. Note the
thin lamellae of chalcopyrite along cleavage planes of actinolite (lower left).
|
|
Medium-grained dark gabbro (2c): This rock type occurs in the southern
portion of the mapped area (Figs. 5–6B) and due to its dark color, including
dark plagioclase in hand specimens, has been described in the field as
melanocratic. However, it contains less than 60 vol percent of ferromagnesian
minerals (40–60 vol % clinopyroxene, <5 vol % orthopyroxene). Disseminated,
blebby sulfides (1–5 vol %), 1 to 5 mm in size, consist of intergrown
pyrrhotite, chalcopyrite, pentlandite, and pyrite (Fig. 7B). Chalcopyrite is
commonly observed along cleavage planes of secondary actinolite. The gabbro has
an equigranular to orthocumulate texture with local adcumulus clinopyroxene.
Cumulus phases are dominated by subhedral clinopyroxene (1–5 mm) and minor
plagioclase, and the intercumulus material consists of plagioclase with minor
clinopyroxene and orthopyroxene.
Pegmatite (2d): Pegmatite occurs as dikes and pods and is composed of
coarse-grained plagioclase and clinopyroxene. Isolated pods of pegmatite on the
surface are commonly connected to large pods of pegmatite through veinlets. The
cores of the pods are commonly quartz rich and contain minor biotite, magnetite,
ilmenite, pyrite, and traces of chalcopyrite. The sulfide aggregates commonly
enclose fine-grained (<0.5-mm) calcite.
Late-stage intrusions (group 3)
Postmineralization intrusions include (3a) medium-grained, sulfide-free
clinopyroxenite and (3b) medium-grained, salt-and-pepper-textured gabbro. The
clinopyroxenite cuts earlier leucocratic rocks (group 1) and commonly displays
magma mingling with the earlier clinopyroxenite (2b). The clinopyroxenite (3a)
and gabbro (3b) display a cumulate texture and are altered with uralization of
clinopyroxene and sericitization of plagioclase.
Intrusions unrelated to the Lac des Iles complex
Late intrusions that cut rocks of the southern Roby zone are felsic and
diabase dikes. The felsic dikes are several centimeters to 1 m in width with
random orientations and range in composition from tonalite to granodiorite. The
dikes are commonly accompanied by narrow (<5-cm) symmetrical alteration halos
with reddish K-feldspar. Very fine grained diabase dikes, ranging from a few
centimeters to 1.5 m in width, cut all rock types.
Contact relationships between different rock types
The contacts between the early leucocratic rocks of group 1 and the
melanocratic rocks of group 2 are well exposed on the outcrop. The latter
contains fragments of leucocratic gabbro of group 1 (Fig. 6A) and also intruded
the partially solidified leucocratic rocks, developing magma-mingling structures
(Fig. 6B). This suggests that mafic magmas intruded prior to the solidification
of the earlier magmas, which produced the leucocratic rocks. Varitextured and
pegmatitic phases (rock type 2d) are commonly developed near the contact between
groups 1 and 2, suggesting the release of aqueous fluids from the magmas that
produced the melanocratic rocks.
Hydrothermal alteration
All rocks show some degree of alteration, especially melanocratic rocks which
are pervasively altered, but rocks adjacent to the Roby zone, including those of
the Twilight zone and East Gabbro, are not altered. Although minor faults occur
in the southern Roby zone, as observed by Michaud (1998), alteration is not
controlled by these faults. The alteration most likely took place at high
temperatures and was possibly related to aqueous fluids released from the magmas
for the melanocratic rocks. This interpretation is supported by abundant pods
and veins of pegmatite associated with the melanocratic rocks.
The High-Grade zone within the Roby zone is composed of intensely altered
medium-grained melanogabbro and pyroxenite, similar to the melanocratic rocks of
group 2 in the southern Roby zone. Primary minerals are entirely replaced by
secondary amphiboles, talc, anthophyllite, chlorite, sericite, and calcite. The
occurrence of blue quartz is also common. Sulfide minerals are dominated by
pyrite (up to 10 vol %), with lesser amounts of pyrrhotite, chalcopyrite, and
pentlandite.
|
Petrology of the Twilight Zone
|
|---|
The Twilight zone (Fig. 8) is dominated by noritic and/or gabbronoritic
intrusions, which are divided into three types, following from the rock types
recognized in the southern Roby zone: (4) leuconorite and/or gabbronorite, (5)
medium-brown norite and/or gabbronorite, and (6) sulfide-bearing melanonorite
and/or gabbronorite. Minor phases include (7) anorthosite, (8) gabbro, (9)
melanogabbro and/or clinopyroxenite, and (10) fine-grained gabbro. Late felsic
and diabase dikes, similar to those in the Roby zone, are also present in the
Twilight zone. They cut all other rock types and are not related to the Lac des
Iles intrusive complex. This zone also has complex structures but alteration is
not as intense and pervasive as in the Roby zone. Orthopyroxene has been
variably replaced by talc and magnetite, clinopyroxene by actinolite, and
plagioclase by sericite.
View larger version (40K):
[in this window]
[in a new window]
|
FIG. 8. Illustration of the complicated distribution of different rock types
in the Twilight zone based on the 1:120 scale mapping by Hinchey et al. (2003).
Dashed lines show the mine grid. The numbers in parentheses after rock names
correspond to the rock types in the text. The mineralized melanonorite (6) is
the latest rock type of the Lac des Iles intrusive complex in the outcrop. It
shows the intrusive contacts with fine-grained gabbro (10), leuconorite (4),
norite (5), and medium-grained gabbro (8).
|
|
All noritic and/or gabbronoritic intrusions (rock types 4–6) are medium
grained with granular, orthocumulate textures and are relatively unaltered. The
early leuconorite (4) and the medium-brown norite (5) are very similar in
texture and mineralogy. The leuconorite and/or gabbronorite (4) consists of
subhedral to euhedral cumulus orthopyroxene (2–4 mm, 20–30 vol %),
equigranular plagioclase (60–70 vol %), and intercumulus clinopyroxene (<10
vol %). The medium-brown norite and/or gabbronorite (5) contain equal amounts of
cumulus orthopyroxene and intercumulus plagioclase. Other minor intercumulus
phases (<10 vol %) are clinopyroxene and biotite. Disseminated aggregates of
sulfide and oxide (<1–2 vol %) occur in both rock types where they are in
contact with melanonorite and/or gabbronorite (6), including intergrown
pyrrhotite, chalcopyrite, and pentlandite, and rounded magnetite. Orthopyroxene
crystals in both the early leuconorite (4) and the medium-brown norite (5) are
locally poikilitic, containing smaller crystals of pyroxene and plagioclase.
Some of the orthopyroxene crystals contain exsolution lamellae of clinopyroxene
(inverted pigeonite).
The mineralized melanocratic norite and/or gabbronorite (6) have a texture
similar to other norite and/or gabbronorite although it is rich in sulfides.
This rock consists of subhedral to euhedral cumulus orthopyroxene (3–10 mm,
70–95 vol %), intercumulus plagioclase (5–30 vol %), intercumulus
clinopyroxene (<10 vol %), minor biotite, and interstitial blebby sulfides
and oxides (2–8 vol %). The sulfides consist of intergrown pyrrhotite,
chalcopyrite, and pentlandite and oxides are magnetite and ilmenite. As in the
other noritic and/or gabbronoritic rocks, clinopyroxene occurs as an exsolved
phase in orthopyroxene (inverted pigeonite). Minor amounts of chalcopyrite and
magnetite occur along the cleavage of secondary amphiboles.
Contact relationships of different rocks
The three noritic and/or gabbronoritic rocks display breccia and
magma-mingling structures (Fig. 6C). Matrix and clast relationships indicate
that the leucocratic norite and/or gabbronorite (4) is the earliest and that the
melanocratic norite and/or gabbronorite orthopyroxene (6) is the youngest. The
magma-mingling structures suggest that successive intrusions occurred while
earlier intrusions were not completely solidified.
Anorthosite (7), gabbro (8), and melanogabbro and/or clinopyroxenite (9) were
contemporaneous with the norite and/or gabbronorite (5). These rocks commonly
occur as fragments in a matrix of norite and/or gabbronorite. Fragments of
gabbro and anorthosite are especially common in the northern part of the
Twilight zone. These rocks are medium grained, moderately to intensely altered,
and free of sulfide.
The south-central portion of the Twilight zone is dominated by late,
medium-grained melanogabbro and/or clinopyroxenite (9), which produces breccias
containing fragments of norite, gabbronorite, and gabbro. The breccias and
intrusions contain irregularly shaped pods and veins of pegmatite, which consist
of plagioclase, pyroxene, quartz, magnetite, and sulfide dominated by pyrite.
The rocks are more intensely altered than the volumetrically dominant norite
and/or gabbronorite. Alteration has produced actinolite after clinopyroxene, and
chlorite and sericite after intercumulus clinopyroxene and plagioclase. These
rocks commonly contain blebs (up to 5 vol %) that consist of pyrrhotite,
chalcopyrite, pentlandite, and minor magnetite. The textures and alteration are
very similar to those of the melanogabbro and clinopyroxenite of the southern
Roby zone.
|
Geochemistry of Mineralized Zones
|
|---|
Sampling and analytical methods
Samples representing various rock types were collected from the southern Roby
and Twilight zones following detailed mapping. Samples of the High-Grade zone
were collected from the pit. Major and minor elements were determined on fused
disks using a Philips PW 2400 X-ray fluorescence spectrometer at the University
of Ottawa. For samples with high Cu content, LiBr was added to prevent the fused
glass from sticking to the Pt crucible. Precision based on 11 replicate runs was
0.35 percent for Al2O3, 0.48 percent for MgO, 3.7 percent
for Sr, 1.3 percent for Cr, and 9.2 percent for Ni. The accuracy, which was
monitored using international references MRG-1 and Sy-2, was within 0.039
percent for Al2O3, 0.28 percent for MgO, 3.5 percent for
Sr, 3.4 percent for Cr, 4.0 percent for Ni, and better than 1 and 10 percent for
other major and minor elements, respectively. Loss of ignition was determined
after heating samples at 1,050°C for over 1 h. Sulfur contents were determined
using an elemental analyzer (Carlo Erba 1110) at the University of Ottawa.
Precision based on nine runs of one sample was 5.4 percent, and the accuracy
based on a set of references was 1.4 percent. Blank samples yielded less than
the detection limit of sulfur (0.001 wt %).
The concentrations of Ir, Os, and Ru were determined at the University of
Ottawa by an isotope dilution technique using a solution enriched in 191Ir,
190Os, and 99Ru. The analytical procedure was essentially
identical to that described in Guillot et al. (2000). Precious metals in 3-g
samples mixed with 6 g of Na2B4O7 were
concentrated into a Ni sulfide bead at 1,050°C. The bead was dissolved in HCl,
and the insoluble residue was dissolved in HNO3 before isotope ratio
measurements using an inductively coupled plasma-mass spectrometry (ICP-MS;
model HP-4500). Blank contributions of Ir, Ru, and Os, mostly from the Ni
powder, were 0.005, 0.011, and 0.008 ng/g of flux, respectively, and less than 1
percent of the concentrations of the metals in the samples. Gold, Pt, Pd, and Rh
were determined by a Pb collection fire assay followed by an ICP-MS analysis on
30-g sample splits at Acme Analytical Laboratories Ltd. in Vancouver, with
detection limits of 1, 0.1, 0.5, and 0.05 ppb, respectively. Precision and
accuracy of the analysis based on nine replicate analyses of a reference were
4.50 and 9.2 percent for Au, 0.61 and 2.03 percent for Pt, 0.49 and 4.68 percent
for Pd, and 16.4 percent and an undetermined accuracy for Rh. The concentrations
of REE and other trace elements were determined at Acme Analytical Laboratories
Ltd., using ICP-MS after digesting samples with HNO3-HClO4-HF-HCl.
Acid digestion technique rather than fusion technique was selected because of
lower detection limits for many elements and because of the lack of refractory
minerals in the samples. The precision of the REE analyses based on four
replicates was mostly better than 10 percent, but the analyses of samples with
concentrations close to the detection limits had precisions of only 28 percent.
The precision and accuracy of REE analyses of reference materials with high
concentrations of REE were better than 10 percent. Concentrations of Cu and
other chalcophile elements were determined at Acme Analytical Laboratories Ltd.
after aqua regia digestion followed by an ICP-MS analysis. The precision and
accuracy of the Cu determinations based on replicate analyses were 0.85 and 0.87
percent, respectively.
Compositional variation
In all of the studied rocks, FeO(total) (total Fe expressed as
FeO) and MnO concentrations increase with MgO, whereas Al2O3,
Na2O, K2O, and Sr decrease (Fig. 9). The higher
concentrations of Al2O3, Na2O, and Sr
correspond to greater abundances of plagioclase in the rocks, whereas higher FeO(total)
and MgO correspond to greater abundances of pyroxene and oxides in the rocks.
View larger version (24K):
[in this window]
[in a new window]
|
FIG. 9. MgO (wt %) vs. Al2O3 (wt %) and Sr (ppm) for
whole rocks from the southern Roby, Twilight, and High-Grade zones. Note the
well displayed correlations suggesting a common parental origin of all rock
types. Numbers in parentheses are rock types in the text.
|
|
A positive correlation between Zr and Hf confirms a cogenetic origin for most
rocks (Fig. 10; Pearce and Norry, 1979; Lesher et al., 1991; Jenner, 1996). The
late clinopyroxenite (3a) has distinctly higher concentrations of incompatible
elements than other rocks (not plotted in Figs. 10–20 for clarity), suggesting
that this rock type is not related to other igneous rocks in the study area.
View larger version (11K):
[in this window]
[in a new window]
|
FIG. 10. Hf vs. Zr concentrations for the rocks from the southern Roby,
Twilight, and High-Grade zones. Note the well displayed correlation suggesting a
cogenetic origin of all rock types. See legend in Figure 9. Correlation
coefficient (r) is shown in the lower right and error bars in 2
are given in the upper left of the diagram.
|
|
View larger version (29K):
[in this window]
[in a new window]
|
FIG. 20. Temporal evolution of palladium mineralization at the southern Roby
and Twilight zones, Lac des Iles mine. The concentrations of MgO (wt %) and Pd
(ppb) are shown by box-whisker plots where a box is defined by the 1st
and 3rd quartile values. The total ranges of values are shown as
lines and the median values of rock types are connected by the thick line. Note
the increased MgO and Pd concentrations associated with the sulfide-bearing
melanocratic rocks. The numbers on the left side of the diagram correspond to
the rock types: 1a = leucogabbro and/or anorthosite, 1b = varitextured gabbro,
1c = medium-grained light gray gabbro, 1d = medium-grained gabbro, 2a = matrix
of melanogabbro breccia, 2b = clinopyroxenite, 2c = dark gabbro, 3a = late
clinopyroxenite, and 3b = late gabbro.
|
|
Almost all lithologic rock types, except the late clinopyroxenite (3a), show
relatively unfractionated REE (Fig. 11). The average values of (Ce/Yb)chondrite
for the melanocratic and leucocratic rocks from the southern Roby zone are 1.3
and 1.9, respectively, and the values for the melanonorite and the norite and/or
leuconorite of the Twilight zone are 1.0 and 1.5, respectively. The late
clinopyroxenite (3a) shows a fractionated REE pattern with overall higher
concentrations of REE compared to the other rocks.
View larger version (31K):
[in this window]
[in a new window]
|
FIG. 11. Chondrite-normalized rare earth element (REE) plots for a variety of
rocks: (A) the melanocratic rocks and (B) the leucocratic rocks from the
southern Roby zone, (C) the melanonorite and (D) norite and/or leuconorite rocks
from the Twilight zone, and (E) normal and enriched midoceanic ridge basalts (N-
and E-MORB, respectively) compared to the hypothetical parental melt for the
rocks in the southern Roby and Twilight zones. The composition of the
hypothetical melt was calculated following the method of Bédard (1994). Note
the low degrees of REE fractionation for both the bulk rocks and calculated
melt. The calculation of the composition of the hypothetical melt used CIPW
normative minerals of plagioclase, clinopyroxene, orthopyroxene, and olivine.
The remaining is attributed to a trapped melt fraction. Using the REE
concentrations of bulk rocks, the modal abundance of minerals, and partition
coefficients of the minerals and melt, the concentrations of REE in different
phases were calculated using mass balance. We used the concentration of REE in
clinopyroxene and the partition coefficients between clinopyroxene and melt
listed in Bédard (1994) to calculate the REE concentrations of the parental
melt. Chondrite and MORB values are from McDonough and Sun (1995), and Sun and
McDonough (1989), respectively.
|
|
The concentrations of Pd, Pt, and Au vary widely within each rock type, but
the melanocratic rocks (2a-c, with >14 wt % MgO) generally have higher metal
concentrations than the leucocratic rocks (1a-d, with <10 wt % MgO; Fig. 12),
suggesting that the bulk of PGE was introduced by the more Mg rich magmas.
Within the melanocratic rocks, dark gabbro (2c) has the highest average
concentrations of Pd, Pt, and Au, and the clinopyroxenite (2b) has higher
concentrations than the melanocratic breccia matrix (2a; Table 1).
View larger version (19K):
[in this window]
[in a new window]
|
FIG. 12. Bivariate plots of MgO (wt %) vs. Au, Pd, and Pt (ppm). Note the
high concentrations of precious metals in the melanocratic rocks in the southern
Roby and Twilight zones. The dashed line broadly separates the melanocratic from
the leucocratic rocks.
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 1 TABLE 1. Representative Whole-Rock Major and Trace Element Compositions of Various Rock Types from the Southern Roby, Twilight, and High-Grade Zones
|
|
View larger version (14K):
[in this window]
[in a new window]
|
FIG. 13. Bivariate plots of sulfur (wt %) vs. Ni, Pt, and Pd (ppm). Note the
positive correlations, suggesting a sulfide control on mineralization. Elevated
Ni concentrations in samples without sulfides suggest the presence of minor Ni
in silicate minerals. Preliminary electron microprobe data confirm that
clinopyroxene contains significant Ni. Correlation coefficients (r) are
calculated for samples, excluding the High-Grade zone.
|
|
View larger version (23K):
[in this window]
[in a new window]
|
FIG. 14. Primitive mantle-normalized plot for Ni, Cu, and PGE of the southern
Roby, Twilight, and High-Grade zones. The data have been recalculated to 100
percent sulfide following Naldrett (1981). Note similar patterns for all rocks
with low concentrations of Ni, Ir, Os, and Ru and high concentrations of Cu, Rh,
Pt, and Pd. Primitive mantle values for Ni and Cu are from McDonough and Sun
(1995) and PGE values from Guillot et al. (2000).
|
|
View larger version (29K):
[in this window]
[in a new window]
|
FIG. 15. A. Plot of Pd (ppm) vs. Ir/Pd for samples from the southern Roby,
Twilight, and High-Grade zones compared to other PGE deposits. Note that the Lac
des Iles data plots along the general trend of other orthomagmatic deposits. B.
Plot of Pd (ppm) vs. Pt/Pd for rocks from the Lac des Iles deposit. Note that
the majority of the melanocratic rocks show comparable Pt/Pd ratios, but samples
from the High-Grade zone have higher Pd and therefore low Pt/Pd. The scatter is
most likely due to a nugget effect of coarse-grained platinum group minerals.
Data sources: Komatite-hosted deposits (Picard et al., 1995); Marathon deposit
in the Coldwell Complex, Ontario (Good and Crocket, 1994); Duluth Complex
(Theriault et al., 1997); J-M reef (J-M: Naldrett, 1981); Merensky reef,
Bushveld Complex (M: Naldrett, 1981).
|
|
View larger version (12K):
[in this window]
[in a new window]
|
FIG. 16. A. Plot of total REE vs. Zr for all rock types. Zirconium is an
incompatible and immobile element. The positive correlation suggests that the
REE were also immobile. B. Plot of Rb vs. Zr, illustrating that Rb acted as a
mobile element during alteration. C. Plot of H2O (wt %) vs. Pd (ppm)
for variably altered melanocratic rocks of the southern Roby zone. The lack of
any positive correlation suggests that aqueous fluid likely was not important in
the concentration of Pd.
|
|
All rocks, except in the High-Grade zone, show positive correlations between
sulfur and Pd (correlation coefficient r = 0.891), Pt (r = 0.907),
and Ni (r = 0.901; Fig. 13). Although the rocks from the Roby zone have
higher concentrations of metals than those of the Twilight zone, the two sets of
data show the same positive correlations (Fig. 13), suggesting that the bulk of
the PGE in these zones are associated with sulfide. This is further supported by
a positive correlation between precious metals and Cu (Pd vs. Cu, r =
0.89; not plotted). The minor scatter between sulfur and Ni is most likely
related to the Ni present in silicate minerals, such as pyroxene, as shown by
the elevated Ni concentrations in sulfur-free rocks (Fig.13). The scatter of Pt
and Pd in Figure 13 may be explained by the subsolidus mobility of PGE or by
nugget effects of PGM. Although we do not discount the mobility of PGE by
hydrothermal fluids, the variation of PGE concentrations in Figure 13 is most
likely due to the presence of discrete PGM. Mineralogic studies carried out by
Watkinson and Dunning (1979) and Edgar and Sweeny (1991) showed that the bulk of
the PGE in the ore form discrete PGM.
Primitive mantle-normalized plots of Ni, Cu, and PGE in mineralized rocks
show low Ni, Os, Ir, and Ru compared to Cu, Rh, Pt, Pd, and Au (Fig. 14). The
High-Grade zone samples have slightly higher ratios of platinum-group PGE
(Pt-PGE) to iridium-group PGE (Ir-PGE) than other samples (Fig. 14). The Pt-PGE
include Pt, Pd, and Rh and the Ir-PGE include refractory Ir, Os, and Ru. This is
also illustrated by the low Ir/Pd, with an average value of 1.5 x 10–4
(Fig. 15A).
Element mobility during alteration and metamorphism
The total concentrations of REE and Zr show a positive correlation (Fig.
16A), suggesting that REE acted as immobile elements during alteration. In
contrast, considerable scatter is evident in the plot of Rb versus Zr (Fig.
16B), as expected for a mobile alkali element. The immobile behavior of the REE
is further supported by the similar, relatively flat patterns of
chondrite-normalized REE for all rocks, regardless of the intensity of
alteration (Fig. 11). As light REE are more soluble in fluids than heavy REE,
this pattern would have been modified during alteration.
|
Discussion
|
|---|
Parental magma composition
Most rocks of the study area are cumulates and the bulk chemical compositions
do not represent those of their parental magmas. Therefore, we calculated the
compositions of parental melt following the method described by Bédard (1994).
This method requires the primary mineralogy of rocks. Due to the variable degree
of alteration in the rocks, we used CIPW normative minerals of plagioclase,
clinopyroxene, orthopyroxene, and olivine in the calculation. The remaining
fraction is assigned as a trapped melt, which ranged up to 12 wt percent, with
an average of 9.5 wt percent.
The REE in the calculated parental melt for each rock type showed relatively
flat normalized patterns, about 10 to 15x chondrite values (Fig. 11). The
primary melt would have contained less than the calculated values because the
magmas were enriched in REE through fractional crystallization of chromite and
olivine that do not incorporate REE. The average (Ce/Yb)chondrite for
the parent magmas of the melanocratic and leucocratic rocks in the southern Roby
zone are 2.4 and 2.5, respectively, whereas the average values for the
melanonorite and the norite and/or leuconorite of the Twilight zone are 1.7 and
2.6, respectively.
Light REE are preferentially incorporated into a melt during partial melting.
The relatively unfractionated REE patterns of our samples suggest that the
source was a moderately refractory mantle where previous partial melting
resulted in lower light REE than heavy REE. This interpretation is further
supported by the similarity between the calculated concentrations and patterns
of REE in the parental melts and enriched midoceanic ridge basalts. The melt for
typical oceanic ridge basalts forms through relatively high degrees of partial
melting in the moderately depleted mantle (e.g., Saunders, 1984).
The MgO content of the parental melt of the melanocratic rocks is calculated
to be 8.9 wt percent, using the partition coefficient for MgO between
clinopyroxene and melt of Hart and Dunn (1993) and an MgO content of 15.7 wt
percent for clinopyroxene in group 2 rocks. The primary melt probably had much
higher concentrations of MgO due to fractional crystallization of olivine. For
example, 20 percent fractional crystallization of olivine (Fo92) from
the melt yields 17.3 wt percent MgO in the original parental melt. Fo92
is the composition of olivine in equilibrium with mantle peridotites (e.g.,
Arai, 1992), but the choice of olivine composition does not significantly affect
the result. Therefore, the MgO content of the primary melt was similar to that
of picritic or komatiitic basalts which form through high degrees of partial
melting in the mantle. We suggest that the primary magmas for the Lac des Iles
intrusive complex were most likely products of relatively high degrees of
partial melting of a moderately depleted, refractory mantle source. This
interpretation is further supported by Nd isotope compositions of the
mineralized rocks. Recalculation of 
Nd values at
2690 Ma from data obtained by Brügmann et al. (1997) yields values ranging from
1.0 to 1.5 for most rocks in the Mine Block intrusion. The values are similar to
the Nd value of 2 for Late Archean igneous rocks
derived from a depleted mantle (e.g., Stern et al., 1989; Hattori et al., 1996).
Sulfur saturation in the parental magma
The ratio of Cu/Pd changes in a magma during sulfide separation because Pd
has a higher partition coefficient between sulfide and silicate melt than Cu at
a given temperature and fO2 (i.e., ~34,000 vs. 1,400; Peach
and Mathez, 1996, Crocket, 2002). Therefore, the ratio of Cu/Pd reflects the
timing of sulfur saturation in the silicate magmas (e.g., Barnes et al., 1993).
A lower Cu/Pd ratio than that of the primitive mantle implies that there was no
early removal of sulfide from the magma. Higher ratios of Cu/Pd imply early
removal of sulfide from the magma or the retention of sulfide at the source.
Samples from the Lac des Iles complex have a large range of Cu/Pd ratios (Fig.
17). The majority of the melanocratic rocks have lower Cu/Pd than the primitive
mantle (Fig. 17), suggesting that there was no early separation of sulfide melt.
In contrast, the leucocratic phases have high Cu/Pd, suggesting early removal of
sulfide from the magma or that residual sulfide remained in the mantle source.
View larger version (29K):
[in this window]
[in a new window]
|
FIG. 17. Pd vs. Cu/Pd plot for samples from the southern Roby and Twilight
zones compared to other PGE deposits. Note that the rocks from Lac des Iles show
a wide variation in Cu/Pd ratios. The mineralized melanocratic rocks show low
Cu/Pd ratios, similar to those for other primary PGE deposits, whereas the
barren leucocratic rocks display high Cu/Pd ratios. Abbreviations: JM = JM reef,
Stillwater, United States (Barnes and Naldrett, 1985); MR = Merensky reef,
Bushveld (Maier et al., 1996) and UG = UG2 reef, Bushveld (Maier and Bowen,
1996), South Africa; N = Noril’sk (Smirnov, 1966), FP = Federov Pansky
(Schissel et al., 2002), Russia; Sud = Sudbury, Canada (Naldrett, 1981).
|
|
High Pt-PGE/Ir-PGE in the ore
Mineralized rocks from the southern Roby and Twilight zones have similar
primitive mantle-normalized PGE patterns (Fig. 14), suggesting a common origin
for the mineralization in these zones. Palladium has a high solubility in
aqueous fluids compared to other PGE (e.g., Wood, 2002), but the pattern and
ratios of Pt-PGE/Ir-PGE are similar, regardless of the H2O contents
of the rocks, which vary from 0.1 to 5.5 wt percent. There is no correlation
between Pd and H2O (Fig. 16C; high H2O values not shown),
suggesting that aqueous fluids had a limited role in concentrating Pd. In
contrast, the samples from the High-Grade zone show extreme enrichment of Pd, as
indicated by low Ir/Pd and Pt/Pd ratios (Fig. 15). In addition, there is no
correlation between Pd and S. This is consistent with the hydrothermal
enrichment of Pd, as suggested by previous workers (Talkington and Watkinson,
1984).
There are two possible processes for mafic magmas to acquire fractionated PGE
with high Pt-PGE: low degrees of partial melting and fractional crystallization.
Ir-PGE are preferentially incorporated into olivine, chromite, and
high-temperature PGM, such as laurite and iridosmine (Puchtel and Humayun, 2001;
Andrews and Brenan, 2002; Sattari et al., 2002; Righter et al., 2004). Retention
of these minerals in the mantle during low degrees of partial melting results in
low Ir-PGE in the melt. This is consistent with low Ir/Pd in gabbro-hosted
deposits compared to komatiite-hosted deposits (Fig. 15). Evolved magmas also
have fractionated PGE because olivine, chromite, and high-temperature PGE are
removed during the early stages of fractional crystallization. We discount low
degrees of partial melting of the parental magmas as the cause of high Pt-PGE
because this is not consistent with the flat normalized REE patterns and the
high MgO contents of the calculated parental magmas. Therefore, high Pt-PGE in
our samples is best explained by the removal of Ir-PGE during early
crystallization of the magmas.
High palladium in the southern Roby and Twilight zones
The average concentration of Pd in the melanocratic rocks, excluding the
High-Grade zone, is estimated to be ~4 ppm, on the basis of average grade,
tonnage, and abundance (~20 vol %) of mineralized melanocratic rocks in the Roby
zone. This is very high compared to average Pd concentrations of basalts (0.46
ppb) and komatiites (~11 ppb; Crocket, 2002). Various processes were suggested
for such enrichment of Pd. Watkinson and Dunning (1979) proposed that sulfur,
Ni, Cu, and PGE were enriched during evolution of the parental magmas, followed
by immiscible separation of Pd-bearing aqueous fluid from the magmas. Talkington
and Watkinson (1984) argued for an important role of hydrothermal activity in
the mineralization, based on the occurrence of PGM containing Te, As, and Bi and
the spatial association of PGM with secondary hydrous minerals and pyrite.
Macdonald (1988) also suggested hydrothermal enrichment of PGE by fluids that
originated from the parental magmas, also based on the abundant hydrothermal
minerals in the mineralized zones. Brügmann et al. (1989) proposed
"constitutional zone refining" for the enrichment of Pd and other
Pt-PGE relative to Ir-PGE. Constitutional zone refining (McBirney, 1987)
involves the formation of volatile-rich magmas through fractional
crystallization and partial melting of earlier formed cumulates. Brügmann et
al. (1989) suggested that a volatile-rich silicate magma remelted gabbro
cumulates and selectively incorporated the Pt-PGE from sulfides in the
cumulates, and that the melanocratic and leucocratic rocks in the Roby zone
represent the residue and partial melt, respectively. In another model, Lavigne
and Michaud (2001) suggested that the mineralization involved the forceful
intrusion of a PGE-, Ni- and Cu-rich immiscible sulfide liquid into the
partially crystallized overlying magma chamber. Exsolution of aqueous fluids
from magmas resulted in the redistribution of precious metals.
Our data, showing positive correlations between sulfur and base and precious
metals, are not consistent with hydrothermal concentration of Pd in the southern
Roby and Twilight zones. Furthermore, our detailed mapping shows that the bulk
of PGE are in late melanocratic rocks. Hydrothermal processes cannot explain the
preferential enrichment of Pd in the mafic rocks. Therefore, we discount any
significant contributions of hydrothermal fluids to the Pd mineralization in the
southern Roby and Twilight zones. However, aqueous fluids may have been
responsible for mineralization in the High-Grade ore samples as indicated by the
scatter in the plots of sulfur and base and precious metals (see below).
The zone-refining process is also not consistent with our data. First, the
zone-refining process produces leucocratic melt and melanocratic and ultramafic
rocks as the residue (Brügman et al., 1989). Our detailed mapping shows that
melanocratic rocks carry most of the Pd, and leucocratic rocks are essentially
barren (Fig. 12, Table 2). Second, the zone-refining process involves partial
melting of a cumulate (Brügmann et al., 1989) containing clinopyroxene,
orthopyroxene, and a solidified, trapped melt. Partial melting would dissolve
the solidified melt first, then clinopyroxene. The solidified melt contains high
concentrations of incompatible elements, such as REE, compared to any other
phases. Clinopyroxene also contains high concentrations of incompatible elements
compared to olivine and orthopyroxene. Therefore, any partial melt should
contain elevated incompatible elements compared to earlier rocks. The late
clinopyroxenites (3a) are the only rock type with high REE, but these rocks are
free of sulfides and barren of Pd. Similar concentrations of REE in all rocks in
the southern Roby and Twilight zones (Fig. 11) argue against the zone-refining
process. Dissolution of significant amounts of clinopyroxene would have
increased the Sc content of the melt, as Sc is preferentially included in
clinopyroxene, but this is inconsistent with the similar Sc/Y ratios of all
rocks (Fig. 18).
View this table:
[in this window]
[in a new window]
|
TABLE 2 TABLE 2. Palladium, Platinum, and Gold Concentrations in Various Rock Types in the Southern Roby Zone and Twilight Zone (reported as median, first quartile, and third quartile values in ppb)
|
|
View larger version (8K):
[in this window]
[in a new window]
|
FIG. 18. Sc vs. Sc/Y for samples from the southern Roby zone. Note that the
Sc/Y ratios of the melanocratic rocks are similar to or lower than those of the
leucocratic rocks.
|
|
We suggest that Pd in the late melanocratic magmas was enriched by
incorporating preexisting sulfide melt formed by earlier magmas. This is
consistent with all data, including low Cu/Pd ratios in the ore and high Cu/Pd
in earlier leucocratic rocks (Fig. 17). Earlier magmas lost sulfide melt with
low Cu/Pd at depth, and later melanocratic magmas incorporated the sulfide melt.
Enrichment of palladium in the High-Grade zone of the Roby zone
Although the High-Grade zone was not the focus of this investigation, it is
important because it contains high Pd, ~8 ppm in most samples, and hosts
approximately 35 percent of the Pd in the mine. The samples from the High-Grade
zone do not plot on the correlation trends of sulfur versus base metals and
sulfur versus precious metals (Fig. 13). The data suggest that Pd enrichment in
this zone may have been caused by aqueous fluids. This interpretation is
supported by abundant quartz aggregates, the lack of exsolution textures in
sulfides, and the common occurrence of pyrite.
The High-Grade zone is located on the eastern margin of the Roby zone,
adjacent to East Gabbro that is older than the rest of the Mine Block
intrusions. The High-Grade zone is adjacent to high-grade breccia ore in the
Roby zone (Lavigne and Michaud, 2001), suggesting a genetic link between the
two. We propose that aqueous fluid exsolved from late fertile melanocratic
magmas migrated into the area of the High-Grade zone. The barren East Gabbro
would have acted as a physical barrier for such an aqueous fluid, resulting in
precipitation of Pd along the boundary with the East Gabbro. The mineralized
zone is subvertical at present, but this steepening probably occurred during the
docking of the Quetico accretionary prism and Wawa arc to the Wabigoon
subprovince to the north (Percival, 1989).
A model of igneous activity and PGE mineralization at Lac des Iles
This section summarizes the evolution of the southern Roby and Twilight
zones, which is illustrated schematically in Figure 19. First, relatively high
degrees of partial melting in a moderately depleted mantle formed the parental
magmas of early gabbroic rocks. These magmas became enriched in Cu and Pt-PGE
during the fractional crystallization of olivine, chromite, and high-temperature
PGM. The magmas eventually reached sulfur saturation, forming an immiscible
sulfide melt with low Cu/Pd in the conduit, which resulted in high Cu/Pd ratios
in the evolving magma. The magmas reached the site of the deposit and partly
solidified as leucocratic gabbros. A new batch of magmas, also formed by
relatively high degrees of partial melting of a similar source, passed through
the same conduit as the earlier, leucocratic magmas and incorporated the
preexisting sulfide melt formed by the earlier magmas. The melanocratic magmas
reached sulfur saturation, forming a second sulfide melt and inheriting the low
Cu/Pd ratios of the earlier sulfide melt. In this model the late mafic magmas
became rich in water and other volatiles, which caused the brecciation and
pegmatite formation observed in the surface outcrops. The hydrothermal fluids
released from melanocratic magmas migrated upward along the boundary between the
East Gabbro and mineralized gabbroic rocks, resulting in intense hydrothermal
alteration at the margin of the Roby zone and forming the High-Grade zone.
View larger version (91K):
[in this window]
[in a new window]
|
FIG. 19. Schematic model showing the history of mineralization at the
southern Roby zone. Magmas formed by a high degree of partial melting in a
depleted mantle source (A1) became enriched in Cu, Pt, and Pd through
fractional crystallization of olivine, chromite, and high-temperature PGM (A2),
segregated sulfide melt that had low Cu/Pd ratios along the conduit and the base
of the magma chamber (A3), and solidified as the early leucocratic
gabbros. A second episode of partial melting in the mantle source produced
another batch of fertile magma. As with the early magma, this magma was enriched
in Cu, Pt, and Pd through fractional crystallization (A2). This magma
incorporated the earlier sulfide melt and intruded forcefully into the partially
crystallized leucocratic rocks (B1), causing brecciation and magma
mingling, and solidified as fertile melanocratic gabbro. Aqueous fluids that
separated from the melanocratic magma percolated through the cumulates,
partially dissolving Pd and concentrating it in the High-Grade ore zone adjacent
to barren East Gabbro (B2).
|
|
Comparison with other types of PGE mineralization
Major PGE deposits occur as stratiform layers in large layered intrusions or
are hosted by Ni sulfides at the base of large mafic igneous intrusions. The
former type includes the Merensky reef in the Bushveld (Von Gruenewaldt et al.,
1985; Schoenberg et al., 1999), the Main sulfide zone in the Great Dyke (Wilson
and Prendergast, 2001), and the J-M reef of the Stillwater Complex, Montana
(Todd et al., 1982). PGE ore in layered intrusions is interpreted to result from
additions of significant volumes of magma or pulses of relatively primitive
magmas into quiescent magma chambers (Todd et al., 1982; Von Gruenewaldt et al.,
1985; Maier et al., 1998; Schoenberg et al., 1999). The Ni sulfide-hosted
deposits include the Noril’sk and Sudbury ore deposits (Lightfoot and
Naldrett, 1994). Contact-type deposits are represented by the River Valley-East
Bull Lake suite of intrusions in Ontario (Peck et al., 2001), the Coldwell
Complex, Ontario (Barrie et al., 2002), the Federov Pansky intrusion in Russia
(Schissel et al., 2002), and the Portimo intrusion, Finland (Alapieti et al.,
1989). Typical contact-type deposits occur near the basal inclusion-rich zones
of the intrusions that contain abundant xenoliths of country rocks and felsic
intrusive rocks. The partial digestion of these fragments is thought to have
resulted in increased H2O and SiO2 contents in the magmas,
forming breccias and pegmatites, and leading to sulfide saturation (Barrie et
al., 2002).
Comparison with contact-type PGE mineralization
The southern Roby and Twilight zones of the Lac des Iles intrusive complex
share many characteristics with contact-type deposits. They include low sulfide
abundance in the mineralized rocks, high Pd compared to other PGE, and the
common occurrence of breccias and pegmatite. However, the PGE mineralization at
the southern Roby and Twilight zones is not localized near the contacts with
country rocks. Instead, the mineralization is in the center of the Mine Block
intrusion. In addition, there is no evidence that enrichment of SiO2 resulted
in sulfur saturation in the ore deposit. In con
Top
Abstract
Introduction
