Petrogenesis of Archean PGM-bearing chromitites and associated ultramafic-mafic-anorthositic rocks from the Guelb el Azib layered complex (West African craton, Mauritania)

Abstract

The Archean Guelb el Azib layered complex (GAC) in the West African craton of Mauritania is composed of an association of serpentinites, chromitites, amphibolites and anorthosites with few fine-grained amphibolite dykes. The complex forms tectonic slices in 2.9-3.5 Ga TTG gneiss terrains in close association with supracrustal rocks (BIFs, impure marbles, amphibolites). It was affected by a main granulite-facies grade metamorphism (up to 900 degrees C at 5-6 kbar) with subsequent retrogression in amphibolite and greenschist facies conditions. The preserved igneous macrostructures, the mineral compositions and the nature of relic magmatic assemblages have been used to constrain the composition of the parental melts and the conditions of crystallization. According to petrological observations and to comparison with experimental data, the formation of the complex can be explained by fractionation of a slightly hydrous high-alumina basaltic melt at low pressure. The early fractionation of olivine and the absence of massive clinopyroxene fractionation before plagioclase saturation led to crystallization of highly calcic plagioclase with Fe-, Al-rich but Cr-poor chromite from a hydrous tholeiitic parental magma, similar to worldwide Archean tholeiites. The complex shares many similarities with Archean anorthosite layered complexes, possibly formed in a supra-subduction zone environment according to results obtained on similar 2.9-3.0 Ga complexes from Greenland and India (namely Fiskenaesset and Sittampundi). Three phases of PGE mineralization affected the GAC chromitites: (i) igneous crystallization of laurite; (ii) formation of late magmatic IPGE sulpho-arsenides (irarsite-hollingworthite) and (iii) hydrothermal Pt-Pd mineralization represented by sperrylite and rustenburgite. (c) 2012 Elsevier B.V. All rights reserved.

Full text

(This is a sample cover image for this issue. The actual cover is not yet available at this time.)
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright

Author's personal copy
Precambrian
Research
224 (2013) 612–
628
Contents
lists
available
at
SciVerse
ScienceDirect
Precambrian
Research
journa
l
h
o
me
pa
g
e:
www.elsevier.com/locate/precamres
Petrogenesis
of
Archean
PGM-bearing
chromitites
and
associated
ultramafic–mafic–anorthositic
rocks
from
the
Guelb
el
Azib
layered
complex
(West
African
craton,
Mauritania)
Julien
Bergera,∗,
Hervé
Diotb,
Khalidou
Loc,
Daniel
Ohnenstetterd,
Olivier
Féméniase,
Marjorie
Pivina, Daniel
Demaiffea,
Alain
Bernarda,
Bernard
Charlierf
aDépartement
des
Sciences
de
la
Terre
et
de
l’Environnement
&
FRS-FNRS,
Université
Libre
de
Bruxelles,
Belgium
bUMR
CNRS
6112
“Laboratoire
de
Planétologie
et
de
Géodynamique
de
Nantes”,
Université
de
Nantes,
France
cUniversité
de
Nouakchott,
Mauritania
dCentre
de
Recherches
Pétrographiques
et
Géochimiques,
CNRS,
F-54501
Vandoeuvre
lès
Nancy,
France
eIAMGOLD
Corporation
–
West
Africa,
Mali
fDepartment
of
Earth,
Atmospheric
and
Planetary
Sciences,
Massachusetts
Institute
of
Technology,
USA
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
7
March
2012
Received
in
revised
form
4
October
2012
Accepted
8
October
2012
Available online xxx
Keywords:
Anorthosite
Layered
complex
Archean
tholeiite
Granulites
Mafic–ultramafic
rocks
a
b
s
t
r
a
c
t
The
Archean
Guelb
el
Azib
layered
complex
(GAC)
in
the
West
African
craton
of
Mauritania
is
composed
of
an
association
of
serpentinites,
chromitites,
amphibolites
and
anorthosites
with
few
fine-grained
amphi-
bolite
dykes.
The
complex
forms
tectonic
slices
in
2.9–3.5
Ga
TTG
gneiss
terrains
in
close
association
with
supracrustal
rocks
(BIFs,
impure
marbles,
amphibolites).
It
was
affected
by
a
main
granulite-facies
grade
metamorphism
(up
to
900 ◦C
at
5–6
kbar)
with
subsequent
retrogression
in
amphibolite
and
greenschist
facies
conditions.
The
preserved
igneous
macrostructures,
the
mineral
compositions
and
the
nature
of
relic
magmatic
assemblages
have
been
used
to
constrain
the
composition
of
the
parental
melts
and
the
conditions
of
crystallization.
According
to
petrological
observations
and
to
comparison
with
experimental
data,
the
formation
of
the
complex
can
be
explained
by
fractionation
of
a
slightly
hydrous
high-alumina
basaltic
melt
at
low
pressure.
The
early
fractionation
of
olivine
and
the
absence
of
massive
clinopyroxene
frac-
tionation
before
plagioclase
saturation
led
to
crystallization
of
highly
calcic
plagioclase
with
Fe-,
Al-rich
but
Cr-poor
chromite
from
a
hydrous
tholeiitic
parental
magma,
similar
to
worldwide
Archean
tholeiites.
The
complex
shares
many
similarities
with
Archean
anorthosite
layered
complexes,
possibly
formed
in
a
supra-subduction
zone
environment
according
to
results
obtained
on
similar
2.9–3.0
Ga
complexes
from
Greenland
and
India
(namely
Fiskenaesset
and
Sittampundi).
Three
phases
of
PGE
mineralization
affected
the
GAC
chromitites:
(i)
igneous
crystallization
of
laurite;
(ii)
formation
of
late
magmatic
IPGE
sulpho-arsenides
(irarsite–hollingworthite)
and
(iii)
hydrothermal
Pt–Pd
mineralization
represented
by
sperrylite
and
rustenburgite.
© 2012 Elsevier B.V. All rights reserved.
1.
Introduction
The
association
of
highly
calcic
anorthosite
and
Fe-rich
chromi-
tite
in
ultramafic–mafic
layered
intrusions
is
almost
exclusively
restricted
to
Archean
terrains
(Windley
et
al.,
1981;
Ashwal,
1993;
Rollinson
et
al.,
2002,
2010;
Dutta
et
al.,
2011).
These
ultramafic–mafic–anorthosite
(hereafter
UMA)
layered
complexes
are
systematically
closely
associated
with
supracrustal
rocks
in
strongly
metamorphosed
and
deformed
TTG
terrains.
They
have
∗Corresponding
author.
Present
address:
ETH
Zurich,
NO
E
59,
Sonneggstrasse,
5,
8092
Zurich,
Switzerland.
Tel.:
+41
44
632
81
67;
fax:
+41
44
632
10
30.
E-mail
address:
julien.berger@erdw.ethz.ch (J.
Berger).
been
recognized
and
well
studied
in
the
North
Atlantic
craton
(the
2.97
Ga
Fiskenaesset
complex
and
2.98
Naajat
Kuuat
com-
plex
in
Greenland;
Myers,
1976;
Windley
and
Garde,
2009;
Polat
et
al.,
2010;
Rollinson
et
al.,
2010;
Hoffmann
et
al.,
2012),
in
the
Indian
Darhwar
craton
(the
2.9
Ga
Sittampundi
and
related
com-
plexes;
Dutta
et
al.,
2011;
Dharma
Rao
et
al.,
in
press
and
references
therein),
in
the
Limpopo
belt
linking
the
Zimbabwe
and
Kaapvaal
cratons
(the
3.3
Ga
Messina
layered
intrusion;
Hor
et
al.,
1975;
Barton,
1996;
Mouri
et
al.,
2009).
Archean
anorthosite–chromitite
complexes
in
the
Australian
Pilbara
craton
have
also
probably
similar
origins
(Hoatson
and
Sun,
2002).
According
to
petrolog-
ical,
geochemical
and
isotopic
studies
of
the
Fiskenaesset
and
Sittampundi
complexes,
UMA
are
thought
to
represent
the
plu-
tonic
section
of
oceanic
arc
crust
formed
above
subducting
slabs
0301-9268/$
–
see
front
matter ©
2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.precamres.2012.10.005

Author's personal copy
J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628 613
Fig.
1.
Geological
maps
of
the
West
African
craton
(a)
and
the
Amsaga
Area
(b).
The
map
of
the
Amsaga
and
the
geochronological
informations
are
from
Potrel
et
al.
(1998).
GAC:
Guelb
el
Azib
Complex.
(Polat
et
al.,
2009,
2010;
Windley
and
Garde,
2009;
Rollinson
et
al.,
2010;
Dutta
et
al.,
2011;
Hoffmann
et
al.,
2012).
Their
close
associ-
ation
with
TTG
(tectonic
or
intrusive
relationships),
interpreted
as
slab
melts
(Martin
et
al.,
2005)
or
as
partial
melts
from
the
base
of
thickened
arc
crust
(Hoffmann
et
al.,
2011),
also
argues
in
favour
of
a
supra-subduction
zone
origin.
Hence,
these
UMA
associations
are
markers
to
understand
Archean
arc
magmatism
and
crustal
recycling
processes
at
sub-
duction
zones,
and
to
characterize
fluxes
between
mantle
and
island
arc
crust
in
the
light
of
early
oceanic
and
continental
crustal
growth.
Disagreements
persist
about
the
geochemical
nature
of
the
primitive
melt
and
its
evolution
with
differentiation
(tholeiitic
alu-
minous
basalt,
Mg-
and
Al-rich
ultrabasic
magma;
Weaver
et
al.,
1981;
Ashwal,
1993;
Rollinson
et
al.,
2010;
Polat
et
al.,
2011),
the
origin
of
highly
calcic
plagioclase
and
the
occurrence
of
metamor-
phic
versus
igneous
plagioclase–chromite–amphibole
associations.
In
this
study,
we
present
new
data
on
the
Archean
Guelb
el
Azib
complex,
a
new
occurrence
of
layered
anorthosite
complex
located
in
the
West
African
craton
of
Mauritania.
It
is
composed
of
a
suite
of
ultramafic
cumulates,
chromitites,
layered
gabbros,
leucogabbros
and
anorthosites
metamorphosed
under
amphibolite
and
granulite
grades.
The
primitive
melt
composition
and
its
liquid
line
of
descent
are
estimated
in
order
to
explain
the
relation
between
the
different
rock-types,
the
origin
of
high-Ca
anorthosites
and
their
association
with
Fe-rich
chromitites.
The
degree
and
effect
of
high-grade
meta-
morphism
on
the
original
igneous
mineralogy
of
the
complex
and
the
link
between
PGE
mineralization
and
chromitites
are
also
dis-
cussed.
Finally,
we
emphasize
the
importance
of
such
complexes
for
a
better
understanding
of
Archean
intra-oceanic
subduction
zone
magmatism.
2.
Regional
geology
The
West
African
Craton
(WAC)
comprises
two
large
shields
where
Archean
and
Paleoproterozoic
terrains
crop
out
at
the
southern
and
northern
borders
of
the
Neoproterozoic
and
Meso-
zoic
Taoudeni
basin:
the
Leo-Man
rise
in
the
south
(Ivory
Coast,
Guinea.
.
.)
and
the
Reguibat
rise
in
the
northwestern
Africa,
within
the
Sahara
desert
(Fig.
1a).
The
Amsaga
area
(Fig.
1b)
(Barrère,
1967),
located
within
the
Reguibat
rise,
belongs
to
the
Choum-Rag
el
Abiod
terrane
(Key
et
al.,
2008)
which
is
characterized
by
major
magmatic
events
at
3.5,
2.99,
2.8
and
2.7
Ga
with
a
main
tectono-metamorphic
event
bracketed
between
2.95
and
2.73
Ga
(Auvray
et
al.,
1992;
Potrel,
1994;
Potrel
et
al.,
1996,
1998;
Key
et
al.,
2008).
The
northern
part
of
the
Amsaga
area
(Fig.
1b)
is
composed
of
supracrustal
metasediments
(metaquartzites,
metagreywackes
and
metapelites)
intruded
by
2.99
Ga
metamorphosed
charnockitic
plutons
(Potrel,
1994;
Potrel
et
al.,
1998).
The
southern
domain
(Fig.
1b)
is
dominated
by
migmatitic,
locally
garnet-bearing,
quartzo-feldspathic
gneisses
of
TTG
affinity
with
layers,
sheets
or
lenses
of
amphibolite
(former
mafic
dykes
or
lavas),
banded
iron
formation
(BIF),
cipolin
and
chromitite/anorthosite-
bearing
mafic–ultramafic
layered
complexes
dissected
by
Neoarchean
shear
zones.
Post-metamorphic/late-kinematic
formations
(Fig.
1b)
include
the
2.73
Ga
Touijenjert–Modreïgue
granite
and
the
2.7
Ga
Iguilid
mafic
intrusion
(Potrel
et
al.,
1998).
Except
for
volumetrically
minor
Paleoproterozoic
to
Jurassic
mafic
dykes,
the
Amsaga
area
underwent
no
major
magmatic
and/or
metamorphic
events
since
2.7
Ga.
The
formation
and

Author's personal copy
614 J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628
Fig.
2.
Detailed
geological
map
of
the
Guelb
el
Azib
complex
with
the
location
of
chromitite
outcrops
in
the
area.
The
rectangle
refers
to
bloc
diagram
of
Fig.
3.
reworking
of
the
Mauritanides,
a
mixed
Pan-African
and
Variscan
orogenic
belt,
has
led
to
thrusting
of
tectonic
sheets
onto
the
Archean
rocks
in
the
southwestern
part
of
the
Reguibat
rise,
but
has
not
affected
the
studied
area.
3.
Chromitites
in
the
Amsaga
area
and
the
Guelb
el
Azib
complex
(GAC)
Our
fieldwork
together
with
the
observation
of
Barrère
(1967)
has
revealed
seven
main
occurrences
of
chromitites
aligned
along
a
NE–SW
strike.
Chromitites
are
associated
with
serpentinites,
ultramafic
meta-cumulates,
mafic
rocks
(coarse-grained
and
fine-
grained
amphibolites)
and
anorthosites.
The
whole
association
forms
complexes
dissected
by
late
Archean
N35–45◦dextral
shear
zones
(Fig.
1b).
The
largest
complex
(300–1500
m
wide
and
10
km
long)
located
about
5
km
south
of
Guelb
el
Azib
(Figs.
1b
and
2)
is
surrounded
by
porphyroclastic
granodioritic
and
tonalitic
orthogneisses,
amphi-
bolites
and
impure
marbles
(Fig.
2).
BIFs
lodes
have
only
been
observed
on
the
eastern
side
of
the
complex,
whereas
few
leucoc-
ratic
garnet-sillimanite
mylonitic
orthogneisses
occur
exclusively
on
its
western
side
(Figs.
2
and
3).
Most
foliations
are
upright
with
down-dip
stretching
lineations;
some
foliation
planes
are
however
only
gently
dipping
away
from
the
major
shear
zone.
Late
dex-
tral
strike-slip
movements
are
evidenced
by
horizontal
lineations
and
asymmetric
mantled
porphyroclasts
in
shear
zones
of
several
tens
of
kilometres
length.
This
last
ductile
event
is
also
attested
by
upright
folds
with
vertical
axis
that
are
well
expressed
in
the
surrounding
BIFs
(Fig.
3).
The
GAC
is
an
allochthonous,
highly
tectonized,
layered
body
affected
by
high-
and
low-grade
metamorphism.
Its
original
igneous
stratigraphy
can
therefore
not
be
reconstructed.
The
geometries
of
boundaries
between
the
different
lithological
units
presented
on
the
geological
map
(Fig.
2)
are
simplified;
they
more
closely
correspond
to
those
drawn
on
the
block
diagram
of
Fig.
3.
Chromitites
crop
out
as
lenses
of
several
metres
width
and
tens
of
metre
length
(Fig.
3).
They
often
occur
within
serpentinites
and
metawebsterites.
Chromitites
are
either
massive,
brecciated
(Fig.
4b)
or
disseminated
within
the
host
cumulates.
Some
lay-
ered
chromite
pods
have
also
been
observed
within
anorthosite
and
leuco-amphibolite
layers
(Fig.
4c
and
sample
MA
435),
a
com-
mon
feature
of
Archean
ultramafic–mafic–anorthosite
complexes
(Rollinson
et
al.,
2010).
Serpentinites
are
deeply
silicified,
carbonated
and
oxidized
(Fig.
4d).
Consequently,
neither
high
temperature
anhydrous
min-
erals
(olivine,
pyroxenes.
.
.)
nor
primary
magmatic
textures
or
structures
are
preserved.
Spinel
veins
(Fig.
4d)
and
rodingitic
dykes
(hydrogrossular–prehnite–chlorite
rocks)
cut
across
the
ser-
pentinites.
Talc
is
often
present
either
as
small
patches
or
as
Fig.
3.
Interpretative
bloc
diagram
of
the
Guelb
el
Azib
complex.
The
letters
refer
to
field
photographs
in
Fig.
4.
The
location
of
the
cross
section
is
shown
in
Fig.
2.

Author's personal copy
J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628 615
Fig.
4.
Field
photographs
of
rocks
forming
the
Guelb
el
Azib
complex:
(A)
layered
chromitite;
(B)
brecciated
and
retrogressed
chlorite
chromitite;
(C)
chromitite
pod
in
a
leuco-amphibolite/anorthosite
block;
(D)
spinel
vein
in
carbonated
and
oxidized
serpentinite;
(E)
layered
hornblendite/mela-amphibolite;
(F)
layered
leuco-amphibolite.
individualized
spinel-talc
rocks.
Elliptical
bodies
of
metaweb-
sterite,
hornblendite,
amphibolite,
anorthosite
and
chromitite
are
scattered
within
the
main
mass
of
serpentinite
(Fig.
3).
The
ultramafic
cumulates
consist
of
partly
serpen-
tinized
metawebsterites,
metatroctolites,
hornblendite
and
mela-amphibolite,
with
some
lenses
of
chromitites.
Elliptical
bodies
of
mafic
material
and
serpentinite
are
also
commonly
found
(Fig.
3).
Despite
strong
deformation
and
pervasive
metamorphism,
igneous
layering
is
preserved
in
both
mela-amphibolites
(Fig.
4e),
ultramafic
cumulates
and
a
few
chromitites.

Author's personal copy
616 J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628
Table
1
Modal
and
textural
properties
of
chromitite
samples.
Sample
X
Y
Texture
%
Chromite
Matrix
mineral
MA
17
643,562
2,242,634
Massive
95
Chlorite
MA
44 653,300 2,254,645
Disseminated
42
Chlorite–calcite
MA
226 657,930
2,274,050
Massive
78
Chlorite
MA
238
654,701
2,253,581
Massive
61
Chlorite
MA
239
654,656
2,253,739
Brecciated
87
Chlorite
MA
273
653,400
2,255,000
Massive
61
Amphibole
MA
400
654,028
2,256,615
Massive
87
Chlorite
MA
420 653,144 2,253,504 Massive 59 Amphibole
MA
422 652,474 2,253,808
Disseminated
21
Srp-Talc
MA
425 653,050
2,254,000
Layered
80–40
Amphibole
MA
436
653,165
2,254,250
Brecciated
65
Chlorite
MA
440
653,372
2,254,820
Massive
69
Amphibole
X–Y:
UTM
coordinates
(m),
zone
28,
datum
WGS
84.
Anorthosites,
leuco-amphibolites
and
meso-amphibolites
form
a
distinct
unit
despite
the
presence
of
a
few
bodies
of
serpentinite.
Igneous
layering
can
still
be
found
(Fig.
4f)
but
both
metamorphism
and
ductile
deformation
affected
the
amphibolite.
Scarce
chromi-
tite
pods
are
interlayered
within
anorthosites/leuco-amphibolite.
Two
generations
of
basic
dykes
are
found
within
the
GAC
(Fig.
3):
the
oldest
one
of
Archean
age
is
affected
by
HT
metamorphism
and
deformation,
while
the
youngest
one
cuts
across
Archean
structures
and
is
oblique
to
foliation.
4.
Petrographic
description
4.1.
Chromitites
Chromitites
have
been
subdivided
according
to
the
modal
pro-
portion
of
spinels
and
the
nature
of
the
matrix
silicate
phase
(Table
1).
Massive
chromitites
contain
more
than
50
vol.%
spinels
(Fig.
5a–c).
The
main
host
silicate
is
pale
greenish
chlorite
(MA
17,
226,
238,
400)
or
green
hornblende
(MA
273,
420,
440).
Chromite
grains
are
angular
and
coarse
(up
to
5
mm)
in
the
chlorite-bearing
samples
(Fig.
5c),
while
they
are
small
(<1
mm)
and
rounded
in
the
amphibolite
(Fig.
5a
and
b).
Ferritchromit
generally
forms
a
thin
rim
around
primary
spinel
or
around
chlorite
and
amphibole
inclusions
(Fig.
5b).
Other
silicate
and
oxide
inclusions
are
scarce;
they
consist
of
nearly
pure
anorthite,
magnesian
olivine
and
rutile.
Small
shuiskite,
Cr-pumpellyite
and
Cr-grossular
have
been
found,
respectively,
in
the
matrix
and
as
inclusions
in
chromite
from
sample
MA
440.
Minute
(<7
␮m)
sulphide
and
arsenide
inclusions
are
common,
the
dominant
species
being
euhedral
millerite
(NiS),
which
is
frequently
found
in
close
association
with
anorthite
and
rutile
inclusions
(Fig.
4d).
The
primary
sulphides
frequently
display
exsolution
or
replacement
bands
of
pyrite
and
pentlandite.
Cov-
elite
(CuS),
chalcopyrite
(CuFeS2)
and
a
single
isolated
gersdorffite
(NiAsS)
were
also
found
within
chromite
grains.
Brecciated
chromitites
owe
their
aspect
to
the
development
of
a
network
of
chlorite-filled
fractures.
The
host
matrix
mineral
is
chlorite
and
spinel
with
a
spongy
texture
characterized
by
numer-
ous
inclusions
of
chlorite
(Fig.
5c).
The
ferritchromit
rim
is
largely
developed;
it
can
entirely
invade
some
chromite,
surround
chlo-
rite
inclusion
and
outline
fractures.
Anhydrous
silicate,
oxide
and
sulphide
inclusions
have
not
been
observed.
The
only
sample
of
layered
chromitite
(MA
425)
selected
for
detailed
investigation
is
in
close
association
with
anorthosite
and
leucogabbro.
Despite
its
layered
structure,
this
sample
has
exactly
the
same
petrographical
features
as
amphibole-bearing
massive
chromitites,
except
for
a
larger
mean
grain
size
(up
to
5
mm
wide)
in
the
silicate-rich
band.
Disseminated
chromitites
contain
less
than
50
vol.%
spinels;
their
matrix
is
either
composed
of
serpentine
with
talc
(MA
422)
or
chlorite
with
calcite
(MA
44).
Chromite
grains
are
characterized
by
a
thick
ferritchromit
rim,
a
frequent
spongy
texture
due
to
numer-
ous
mineral
inclusions
(same
nature
as
in
the
matrix),
and
the
lack
of
anhydrous
silicate,
oxide
and
sulphide
inclusions
(Fig.
5e).
4.2.
Ultramafic
metacumulates
Four
subgroups
have
been
distinguished
on
the
basis
of
the
lithological
nature
of
the
sample:
olivine-amphibole
rocks,
spinel-
amphibolite,
mela-amphibolite
and
hornblendite
(Table
2).
Olivine-amphibole
rocks
(Fig.
6a)
are
composed
of
olivine,
spinel
and
amphibole.
They
are
deeply
serpentinized
(more
than
60
vol.%
of
olivine
were
serpentinized),
but
the
presence
of
rounded
olivine
relics
surrounded
by
amphibole
crystals,
probably
pseudomorphs
after
former
pyroxenes,
allows
us
to
identify
a
former
poikilitic
tex-
ture.
Green
spinel
is
surrounded
or
totally
replaced
by
magnetite.
Spinel-amphibolites
(Fig.
6b)
show
the
asso-
ciation
spinel–amphibole–anorthite
(MA
264)
or
spinel–amphibole–olivine–orthopyroxene
(MA
25).
These
samples
are
affected
by
a
strong
deformation
outlined
by
large
elongated
amphibole
prisms
(up
to
1
cm).
Granular
polyhedral
amphibole
neoblasts
crystallized
at
the
borders
of
larger
porphyroclasts
indicate
a
high-temperature
recrystallization
process.
Plagioclase
grains
are
polyhedral
and
do
not
show
evidence
of
internal
strain.
They
are,
however,
partly
altered
into
albite–epidote
intergrowths.
Green
spinel
and
olivine
(<2
mm)
are
interstitial
to
amphibole
porphyroclasts.
4.3.
Amphibolites
and
anorthosites
Mela-amphibolites
(Fig.
6c)
consist
of
brown
amphi-
bole
+
plagioclase
±
ilmenite
±
clinopyroxene
±
orthopyroxene.
Two
types
of
textures
have
been
observed:
(i)
undeformed
poly-
hedral
plagioclase
grains
with
large
prismatic
amphibole
(MA
262
and
401),
(ii)
undulatory
plagioclase
with
deformation
twins
and
small
granular
amphibole
representing
the
recrystallization
prod-
uct
of
larger
prophyroclasts
(MA
28
and
34).
Former
anhydrous
granulitic
assemblages
and
granular
textures
are
still
observed
in
a
few
samples
(ex:
MA
401).
Brown
amphibole
grows
at
the
expense
of
clinopyroxene
grains
(Fig.
6c)
indicating
that
amphibole
is
not
a
primary
igneous
phase.
Plagioclase
is
partly
altered
into
association
of
epidote,
albite
and
sulphides.
One
sample
of
hornblendite
(MA
258)
has
been
selected;
it
is
exclusively
composed
of
large
(up
to
5
cm
long)
oriented
prismatic
brown
amphibole
(Fig.
4e).
The
anorthosite
MA
424
(Fig.
6d)
is
exclusively
composed
of
plagioclase
(with
a
mean
size
of
3
mm)
which
displays
lobate
boundaries,
internal
strain
evidenced
by
undulose
extinction
and
deformation
twins.
The
rock
has
been
affected
by
both
high-temperature
grain-boundary-migration
(GBM)
and
subgrain

Author's personal copy
J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628 617
Fig.
5.
Microphotographs
of
chromitites
from
the
GAC.
(a)
Rounded
chromite
in
amphibole-rich
matrix
(MA
273).
Note
the
laurite
inclusion
in
the
inner
rim
of
a
chromite
grain.
(b)
Rounded
chromite
grains
with
ferritchromit
rim
in
a
calcite–talc–serpentine
matrix
from
a
chlorite
chromitite
(MA
238).
Sperrylite
is
exclusively
present
in
the
ferritchromit
rim.
(c)
Spongy
chromite
from
a
brecciated
chlorite
chromite
(MA
239).
The
ferritchromit
rim
develops
around
chlorite
inclusions.
(d)
Sulphides
(millerite
replaced
by
pyrite
and
pentlandite),
rutile
and
anorthite
inclusions
within
a
chromite
grain
from
an
amphibole
chromitite
(MA
273).
(e)
Chromite
grain
with
a
thick
ferritchromit
rim
in
a
dessiminated
chromitite
(MA
422).
chr:
chromite,
fct:
ferritchromit,
hbl:
hornblende,
chl:
chlorite,
cc:
calcite,
tlc:
talc,
srp:
serpentine;
rt:
rutile;
pl:
plagioclase,
mi:
millerite,
py:
pyrite;
pe:
pentlandite.

Author's personal copy
618 J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628
Table
2
Modal
composition
of
silicate
rocks
from
the
GAC.
Ol
Cpx
Opx
cAmp
Antophyllite
Pl
Qz
Zrc
Ap
Spl
Ilm
Sulfides
%Alt
Olivine-amphibole
rocks
Ma
426 62 25 13 35
(srp)
Ma
43 63
29
8
60
(srp)
Ma
259
75
17
8
75
(srp)
Spinel-amphibolite
Ma
25
7
67
20
6
+
Ma
264
23
1
63
13
15
(Srp
+
chl)
Hornblendite
Ma
258 100 + + +
Mela-amphibolites
Ma
28
46
25
24
+
5
25
(ant)
Ma
34
62
38
+
+
2
(ep
+
ab)
Ma
262
+
63
32
+
+
4
8
(ep,
ab,
msc)
Ma
401
15
26
28
31
+
Leuco-amphibolites
Ma
29
23
77
+
Ma
27
14
86
+
+
Ma
37 1 13 33 53 2
(srp,
chl)
Anorthosites
Ma
424
1
99
2
(ep,
ab)
Ma
423
10
90
5
(ep,
ab)
Fine-grained
amphibolite
Ma
256 10 66 24 +
7
(ep,
ab)
Ma
267
3
16
45
36
+
8
(ep,
ab)
Ma
408 4
15
47
34
+
11
(ep,
ab)
Ma
412
3
14
47
36
+
12
(ep,
ab)
Ma
435
27
41
32
1
+
+
+
%Alt:
modal
proportion
of
secondary
low
temperature
phases
(the
nature
of
the
phases
is
indicated
between
brackets).
+:
accessory;
srp:
serpentine;
ant:
antophyllite;
chl:
chlorite,
ep:
epidote,
ab:
albite,
msc:
muscovite.
rotation
(SGR)
recrystallization,
the
latter
being
shown
by
the
presence
of
small
polygonal
grains
bordering
the
larger
porphyro-
clasts.
Sample
MA
423
(Table
2)
is
an
anorthosite
with
∼8
vol.%
of
small
prismatic
green
amphibole
patches
replacing
a
former
larger
clinopyroxene.
Plagioclase
is
polyhedral
with
sharp
straight
grain
boundaries.
It
shows
evidence
for
subgrain
rotation
recrystalliza-
tion
(SGR),
leading
to
the
presence
of
both
large
deformed
grains
(up
to
1
cm
long)
and
small
unstrained
neoblasts.
Leuco-amphibolites
(Fig.
6e)
are
characterized
by
the
associ-
ation
of
green–brown
amphibole
(14–33
vol.%)
and
plagioclase
(53–86
vol.%),
sample
MA
37
showing
additional
granular
orthopy-
roxene.
Plagioclase
is
largely
dominant
over
amphibole;
no
oxide
has
been
observed.
Textural
evidence
suggests
that
green
amphi-
bole
progressively
consumed
brown
amphibole
during
retrograde
phase.
Two
samples
(MA
27
and
29)
have
a
bimodal
size
distribution
of
plagioclase
with
slightly
lobate
strained
polyhedral
porphyro-
clasts
surrounded
by
neoblasts
formed
by
SGR.
MA
37
sample
displays
a
strong
shape
preferred
orientation
of
both
plagioclase
aggregates
and
amphibole,
plagioclase
being
unstrained
and
poly-
hedral.
Epidote
and
albite
are
common
low-temperature
alteration
phases
of
plagioclase
in
this
sample.
4.4.
Fine-grained
amphibolites
(amphibolites)
Samples
MA
267,
408
and
412
show
the
metamorphic
association
plagioclase
+
brown
amphibole
+
clinopyroxene
+
orthopyroxene
+
ilmenite
(Fig.
6f,
Table
2).
Former
porphyro-
clasts
of
clinopyroxene
(up
to
1
mm
wide)
are
surrounded
by
a
fine-grained
granulitic
matrix
of
polygonal
plagioclase,
granular
orthopyroxene
and
brown
amphibole
blasts,
the
latter
clearly
consuming
large
clinopyroxene.
Samples
MA
238
and
435
show
a
more
equilibrated
texture
characterized
by
polygonal
plagio-
clase
and
brown
amphibole.
Few
relics
of
clinopyroxene
are
still
observed
in
the
core
of
amphibole.
Quartz,
apatite
and
pyrite
are
common
accessory
phases
in
these
samples.
5.
Mineral
chemistry
For
the
following
sections,
analytical
methods
are
provided
as
Supplementary
material
A
and
tables
of
microprobe
analyses
as
Supplementary
file
B.
Supplementary
data
associated
with
this
article
can
be
found,
in
the
online
version,
at
http://dx.doi.org/10.1016/
j.precamres.2012.10.005.
5.1.
Chromite
and
other
spinels
Chromites
from
GAC
chromitites
define
a
large
compositional
field,
the
Fe2+#
ranges
from
53
to
98
and
the
100
×
Cr/(Cr
+
Al)
ratio
from
40
to
75
(Fig.
7a)
with
low
Ti
contents
(max
0.54
wt%
TiO2).
Compared
to
worldwide
Archean
and
Paleoproterozoic
chromites,
those
from
the
GAC
are
characterized
by
high
Fe#
and
low
Cr/(Cr
+
Al)
(Fig.
7a)
comparable
to
chromites
from
Fiskenaesset
(Rollinson
et
al.,
2010).
Disseminated
and
brecciated
chromitites
show
large
compositional
variations
(ex:
Fe#:
62–96
in
sample
MA
422)
with
grains
surrounded
by
a
ferritchromit
rim.
Three
compositional
trends
can
be
observed
for
brecciated
and
dissem-
inated
chromites
(Fig.
7a):
(i)
a
strong
increase
in
Cr#
coupled
to
a
slight
increase
of
Fe#
towards
ferritchromit
compositions
(Fe#:
96,
Cr#:
99).
(ii)
A
decrease
of
both
Fe#
and
Cr#
towards
green
spinels
from
spinel-amphibolites
(Fe#:
30,
Cr/(Cr
+
Al):
0).
(iii)
Large
variations
in
Fe#
(61–85)
at
constant
Cr/(Cr
+
Al)
ratio
(∼61–64)
with
an
endmember
composition
represented
by
the
massive
chlorite–chromitite
MA
226
(Fe#:
53–57;
Cr/(Cr
+
Al):
61–62)
also
characterized
by
the
development
of
ferritchromit
rims.
Massive
and
layered
chromitites
(Fig.
7b),
either
amphibole
or
chlorite-bearing,
do
not
show
compositional
variation
from

Author's personal copy
J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628 619
Fig.
6.
Microphotographs
of
some
silicate
rocks
from
the
GAC:
(A)
olivine–amphibole
rock
MA
426,
note
the
former
poikilitic
texture;
(B)
spinel
amphibolite
MA
25;
(C)
mela-amphibolite
MA
262
showing
the
progressive
replacement
of
clinopyroxene
by
hornblende;
(D)
anorthosite
MA
424;
(E)
leuco-amphibolite
MA
29;
(F)
fine-grained
opx-bearing
amphibolite
MA
435.
grain
to
grain
in
a
given
sample
(ex:
MA
273,
Fe#:
59.2–60.8;
Cr/(Cr
+
Al):
51.8–52.8)
and
no
chemical
zoning
within
individual
grains.
Chromite
in
these
samples
are
characterized
by
slight
vari-
ation
in
Cr/(Cr
+
Al)
ratio
(46–53)
with
Fe#
ranging
from
59
to
78;
with
the
massive
chlorite
chromitite
MA
238
showing
maxi-
mal
values
for
both
Fe#
and
Cr/(Cr
+
Al)
ratio
(85–88
and
54–56,
respectively).
Chromite
in
massive
and
layered
chromitites
is
the
most
similar
to
those
from
Archean
layered
anorthiste
complexes
(Fig.
7b).
On
the
Fe3+–Cr–Al
plot
(Fig.
8),
brecciated
and
disseminated
chromites
are
located
between
massive
chromitites
and
either
fer-
ritchromit
(Al#:
0–10,
Fe3+#:
35–40)
or
Mg–Al
green
spinels
(Al#:
90–100).
As
a
whole,
massive
and
layered
chromitites
are
compa-
rable
to
Fiskenaesset
chromites
though
having
low
Fe3+#
(4–10).

Author's personal copy
620 J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628
0
10
20
30
40
50
60
70
80
90
100
0 102030405060708090
100
2+ 2+
Fe /(Fe +Mg)
Cr/(Cr+Al)
Chromitites
Massive
50-70%
70-90%
90-100%
Brecciated
Layered
Disseminated
Ultram. metacumulates
Spl-amphibolites
Chl Amp Srp
Metamorphic spinel
from meta-troctolites
Ferritchromit
Fiskenaesset
Komatiites
Greenstone
Cont. Intrusions
10 20 30 40 50 60 70 80 90 100
MA 400
Sittampundi
AB
Fig.
7.
Fe#
vs
Cr#
plot
for
the
GAC
chromites
and
spinels.
(A)
Plot
of
all
spinels
and
chromites
from
the
GAC
with
the
three
metamorphic
trends
(thick
black
arrows)
described
in
the
text.
(B)
Plot
of
GAC
chromites
with
preserved
igneous
composition
and
trend
(thick
grey
arrow)
compared
to
chromites
from
those
from
Archean/Paleoproterozoic
terrains.
Comparison
fields
are
from
Barnes
and
Roeder
(2001)
except
for
Fiskenaesset
(Rollinson
et
al.,
2010)
and
Sittampundi
chromites
(Dutta
et
al.,
2011).
The
nature
of
the
MA
400
chromites
(metamorphic
vs
igneous)
is
discussed
in
the
text.
Chl,
Amp
and
Srp
refer
to
chlorite-,
amphibole-
and
serpentine-chromitites,
respectively.
On
the
Mg#
vs
Al
plot
from
chromites,
the
three
trends
observed
for
brecciated
and
disseminated
chromites
are
well
defined.
The
massive
chromitites
define
a
trend
of
constant
Al
for
decreasing
Mg#
and
the
sample
with
lowest
Mg#
is
characterized
by
lowest
Al
content
(Fig.
9).
5.2.
Plagioclase
and
amphibole
(Figs.
10
and
11)
Plagioclase
compositions
range
from
An50 to
An98 (Fig.
10)
in
the
whole
Guelb
el
Azib
complex
and
are
unzoned
(less
than
2
mol.%
An
difference
between
core
and
rim,
except
in
two
retrogressed
fine-grained
and
one
mela-amphibolite)
with
no
com-
positional
difference
between
large
strained
grains
and
small
Al
0 102030405060708090100
3+
Fe
0
10
20
30
40
50
60
70
80
90
100
Cr
0
10
20
30
40
50
60
70
80
90
100
Metamorphic spinel
from meta-troctolites
Chromitites
Massive
50-70%
70-90%
90-100%
Brecciated
Layered
Disseminated
Ultram. metacumulates
Spl-amphibolites
Chl Amp Srp B
Ferritchromit
Fiskenaesset
Komatiites
Greenstone
Cont. Intrusions
Fig.
8.
Fe3+–Al–Cr
plot
from
the
GAC
spinels
and
chromites
compared
to
chromites
from
Archean/Paleoproterozoic
settings.
Same
references
as
in
Fig.
7.
polygonal
new
grains.
The
most
An-rich
plagioclase
is
found
in
the
spinel-amphibolite
MA
25
(An97–98);
it
is
associated
with
alu-
minous
spinel
and
pargasitic
amphibole
(Mg#:
90–93;
up
to
0.6
(Na
+
K)Aa.p.f.u.).
The
plagioclase
from
mela-amphibolites
varies
from
An56 to
An92,
the
most
calcic
being
found
in
the
sample
MA
401
which
also
has
the
most
Na
and
Al-rich
amphibole
(up
to
1.9
IVAl
a.p.f.u.,
Fig.
10).
Amphibole
from
this
group
is
highly
variable
in
composition
from
one
sample
to
another
(Mg#:
43–78, IVAl:
0.15–1.9
a.p.f.u.;
Fig.
11).
Again,
plagioclase
crystals
are
unzoned
Mg#
0 1020304050
0.4
0.6
0.8
1.0
1.2
Al (p.f.u.)
MA 400
Chromitites
Massive
50-70%
70-90%
90-100%
Brecciated
Layered
Disseminated
Chl Amp Srp
Compositional field
of massive igneous
GAC chromite
Toward high-Al
metamorphic spinel
Toward
ferritchromit
Fe-Mg exchange trend
Fig.
9.
Mg#
vs
Al
plot
for
chromites
from
the
GAC.
The
grey
field
represents
the
composition
of
preserved
igneous
chromites.

Author's personal copy
J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628 621
50 60 70 80 90 100
An (mol.%)
µ-amphibolite
Amp-anorthosite
Hornblendite
Leuco-amphibolite
Mela-amphibolite
Spl-amphibolite
Anorthosite
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
Al in amphibole (p.f.u.)
Ma34
8
2
a
M
Fig.
10.
Relationship
between
composition
of
amphibole
and
plagioclase
in
the
GAC
silicate
rocks.
The
lines
are
joining
plagioclase–amphibole
pairs
from
a
same
sample.
and
limited
chemical
variations
(up
to
4
mol.%
An)
from
grain
to
grain
are
observed
within
a
single
rock.
Leuco-amphibolites
are
characterized
by
plagioclase
with
compositional
range
An75–93,
the
most
calcic
plagioclase
(MA
37)
being
associated
with
Na-
and
Al-rich
tschermakitic/pargasitic
amphibole
(Mg#:
85–87;
(Na
+
K)A
up
to
0.5).
The
mono-mineralic
anorthosite
(MA
424)
has
calcic
plagioclase
(An85–87),
but
the
amphibole
anorthosite
(MA
423)
has
plagioclase
with
distinctly
lower
Ca
contents
(An50–52)
com-
pared
to
classical
Archean
anorthosites
(Phinney
et
al.,
1988),
and
is
associated
with
a
Al-
and
Na-poor
Mg-hornblende
(0.1–0.2
(Na
+
K)A;IVAl:
∼0.8
p.f.u.).
The
different
fine-grained
amphibolites
show
calcic
plagioclase
(An64–70)
and
Al-rich
Mg-hornblende
(Mg#:
54–68)
except
for
sample
MA
435
which
is
characterized
by
Al-
and
Na-poor
amphibole
(Mg#:
78–80).
Amphiboles
composition
from
chromitites
largely
overlaps
the
one
from
spinel–amphibolite
and
olivine–amphibole
rocks,
they
are
magnesiohornblende
and
5.56.06.57.07.58.0
40
50
60
70
80
90
100
(Na+K)A
0.0
0.2
0.4
0.6
0.8
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
µ-amphibolite
Amp-anorthosite
Hornblendite
Leuco-amphibolite
Mela-amphibolite
Spl-amphibolite
Meta-ultramafic
actinolite
tremolite
magnesiohornblende
ferrohornblende
ferroactinolite
tschermakite (Na+K) <0.5
or
pargasite (Na+K) >0.5
tschermakite sub.
.
b
u
s
e
t
i
s
a
g
r
a
p
edenite sub.
ferropargasite
Si (p.f.u.)
IVAl (p.f.u.) Mg#
Chromitite
Fig.
11.
Composition
of
amphiboles
from
the
GAC.
(A)
Classification
diagram
from
Leake
et
al.
(1997).
(B)
(Na
+
K)Avs IVAl
plot
showing
the
trends
produced
by
various
substitutions.
tschermakite
(Mg#:
84–96; IVAl:
0.9–1.7
p.f.u.).
Amphibole
from
the
hornblendite
MA
238
has
a
Fe-rich
composition
(Mg#:
60)
compared
to
other
ultramafic
samples
(Mg#
>
85;
Fig.
11).
A
few
amphiboles
from
mela-amphibolites
and
chromitites
in
chlorite-rich
microdomains
are
actinolite
or
tremolite
formed
dur-
ing
low-temperature
late
stage
alteration.
5.3.
Olivine
Olivines
have
a
restricted
range
of
composition
from
Fo83 to
Fo84
both
in
spinel-amphibolite
MA
264
and
metawebsterites
43
and
426.
The
olivine
within
chromitites
is
Mg-rich
compared
to
silicate
rocks:
Fo91 in
sample
MA
17
and
Fo94 in
sample
MA
400.
The
Cr
content
is
high
in
olivine
from
chromitites
(up
to
1.04
wt%
Cr2O3)
compared
to
olivine
from
silicate
rocks
(<0.04
wt%
Cr2O3).
5.4.
Pyroxenes
Clinopyroxenes
from
mela-amphibolites
and
fine-grained
amphibolites
are
augites
with
Mg#
ranging
from
0.70
to
0.80;
they
have
low
contents
of
non-quadrilateral
elements
(Al:
0.03–0.06
a.p.f.u.
and
Ti:
0.002–0.004
a.p.f.u.).
Orthopyroxene
from
the
spinel
amphibolite
MA
264
is
the
most
magnesian
(Mg#:
0.84).
Within
the
group
of
mafic
rocks,
the
olivine-bearing
leuco-
amphibolite
has
the
most
Mg-rich
orthopyroxene
(Mg#:
0.76–0.78)
compared
to
mela-amphibolite
(0.59–0.62)
and
fine-grained
amphibolites
(0.52–0.66).
The
Al
content
increases
with
decreasing
Mg#
in
Mg-rich
samples
(from
0.050–0.067
to
0.074–0.098
p.f.u.);
it
is
significantly
lower
in
mela-
and
fine-grained
amphibolites
(0.011–0.028
p.f.u.).
6.
P–T
calculations
and
thermodynamic
modelling
Despite
preserved
igneous
macrostructures,
such
as
layer-
ing,
scarce
igneous
textures
and
minerals,
the
rocks
from
the
Guelb
el
Azib
complex,
are
strongly
metamorphosed
under
gran-
ulitic
grade
with
subsequent
retrogression
under
amphibolite
and
greenschist
facies
grade.
The
P–T
calculations
will
thus
only
constrain
the
metamorphic
conditions
registered
by
the
GAC
com-
plex.
According
to
petrographical
observations,
samples
from
the
GAC
were
subject
to
anhydrous
granulitic
grade
metamorphism
that
is
evidenced
by
granular
opx–cpx–plag
assemblages
in
mafic
rocks
and
olivine–spinel–pargasite
in
ultramafic
rocks.
It
was
then
affected
by
pervasive
hydrous
recrystallization
in
the
low
T
gran-
ulite
to
amphibolite-facies
conditions
(growth
of
brown
and
green
amphibole)
and
then
subject
to
local
recrystallization
in
the
green-
schist
facies
(rodingites,
chlorite,
epidote).
Amphibole–plagioclase
thermometry
of
Holland
and
Blundy
(1994)
was
used
to
compute
equilibration
temperature
in
GAC
rocks.
As
recommended
by
the
authors,
plagioclase–amphibole
pairs
characterized
by
XAn largely
above
0.9
(MA
25:
0.97–0.98)
were
excluded
of
the
calculations.
The
thermometer
was
applied
to
amphibole–anorthosite,
fine-grained
amphibolites,
leuco-amphibolites
and
mela-amphibolites.
Pressure
has
been
fixed
at
5
kbar
(see
below)
but
this
calibration
is
only
slightly
pressure-dependent
with
a
temperature
increase
of
7◦C/kbar.
The
calculated
temperatures
vary
from
650
to
960 ◦C
(Fig.
12),
the
mela-
amphibolites
showing
the
highest
values
and
most
temperatures
fall
within
the
range
750–850 ◦C
(granulitic
conditions).
The
pla-
gioclase
and
amphibole
compositions
from
mela-amphibolite
MA
401
have
mineral
compositions
close
to
the
limits
of
applicability
of
this
thermometer
(XAn:
0.89–0.93,
with
most
values
at
0.91; IVAl
in
amphibole
between
1.66
and
1.84
a.p.f.u.).
Computed
tempera-
tures
range
from
960
to
1000 ◦C
in
this
sample
but
the
uncertainty
is

Author's personal copy
622 J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628
0
1
2
3
4
5
6
7
8
9
10
650
700
750
800 850 900 950 1000
Amp-anorthosite
µamphibolite
Leuco-amphibolite
Mela-amphibolite
Number of results
Fig.
12.
Bar
plot
of
temperatures
calculated
with
hornblende–plagioclase
thermom-
etry
(Holland
and
Blundy,
1994)
for
the
various
rock
types
of
the
Guelb
el
Azib
complex.
Pressure
is
fixed
at
5
kbar.
probably
larger
than
the
one
recommended
by
Holland
and
Blundy
(1994)
(±40 ◦C).
Pressure
is
difficult
to
estimate
for
the
metamorphic
assem-
blages
lacking
baro-dependent
minerals.
One
olivine
leuco-
amphibolite
sample
(MA
37)
has
been
selected
to
draw
a
fixed
composition
phase
diagram
because
it
has
a
low
variance
assemblage
(pl
+
amph
+
ol
+
opx)
compared
to
other
samples.
The
pseudosection
(Fig.
13)
has
been
built
with
Perple
X
(Connolly,
2005)
using
the
thermodynamic
dataset
of
Holland
and
Powell
(1998,
updated
in
2003)
and
the
solution
models
of
Diener
et
al.
(2007)
for
amphibole,
of
Holland
and
Powell
(1996)
for
ortho-
and
clino-pyroxene
and
of
Holland
and
Powell
(2003)
for
plagioclase.
The
stability
of
the
pl
+
amph
+
opx
+
ol
assemblage
occupies
a
small
portion
of
the
grid:
below
6.6
kbar
at
800–910 ◦C.
Mineral
isopleths
have
been
drawn
for
the
Mg#
of
orthopyroxene;
the
isopleth
that
4.0
5.2
6.4
7.6
8.8
10.0
P(kbar)
opx amph pl
sp cpx
opx amph
pl sp
opx amph
pl ol sp
opx pl ol
cpx H2O
opx amph pl
sp cpx H2O
opx pl sp
cpx H2O
opx amph
pl ol H2O
opx amph pl
ol sp H2O
opx amph pl
ol sp cpx
opx amph pl
ol cpx H2O
opx pl ol
sp cpx H2O
opx amph pl
ol sp cpx H2O
8
7
n
E
7
7
n
E
mph pl
px H2O
opx pl
cpx H
7
m
px
pl
H
7
7
MA 37: olivine leuco-amphibolite (ol pl amph opx)
SiO Al
OFeO
MgO CaO Na
O
H
O
45.57
25.59 5.57 8.42
13.22 1.13 0.50 (wt%)
Range of T° calculated
using hbl-pl thermometry
760
820
880
940
1000
T(°C)
700
Fig.
13.
Pseudosection
built
with
Perple X
for
sample
MA
37.
The
grey
shaded
area
represents
the
temperature
range
measured
with
hornblende–plagioclase
ther-
mometry
for
the
same
sample.
The
stability
field
for
the
assemblage
observed
in
sample
MA
37
is
outlined
by
a
thick
line.
matches
the
composition
measured
by
microprobe
(En77)
crosses
the
pl
+
amph
+
opx
+
ol
field
in
the
highest
pressure
part
of
the
sta-
bility
field.
When
combining
the
results
of
hornblende–plagioclase
thermometry
(shaded
grey
area
in
Fig.
13)
and
those
from
phase
diagram
calculation,
the
best
fit
for
pressure
is
between
5.2
and
6.4
kbar.
This
pressure
estimation
is
in
good
agreement
with
pre-
vious
studies
on
metapelitic
rocks
of
the
Amsaga
area
(5
±
1
kbar;
Potrel
et
al.,
1998)
but
the
maximal
temperature
estimations
from
this
author
are
lower
(800
±
50 ◦C)
compared
to
ours
(880–910 ◦C).
The
diagram
also
support
that
the
growth
of
brown
amphibole
at
the
expense
of
pyroxene
is
a
retrograde
reaction
linked
to
both
tem-
perature
decrease
and
hydration
within
the
granulite–facies
space
(Fig.
13),
as
deduced
from
petrographic
observations.
7.
Concentration
and
speciation
of
platinum-group
elements
Platinum-group
minerals
(PGM)
have
only
been
observed
in
massive
and
layered
chromitites.
The
most
abundant
PGM
is
lau-
rite
(RuS2)
which
occurs
as
small
(<5
␮m,
mostly
2–3
␮m
wide)
and
euhedral
inclusions
(Fig.
14a
and
b)
in
the
core
and
the
rim
of
the
chromite.
Quantitative
electron
microprobe
analyses
have
not
been
undertaken
due
to
the
small
size
of
the
inclusions
that
approaches
the
diameter
of
microprobe
beam.
Semi-quantitative
analyses
with
energy
dispersive
spectrometer
(EDS)
show
that
irid-
ium
and
osmium
are
the
most
common
elements
substituted
to
Ru
into
laurite:
the
range
of
measured
composition
is
(Ru0.80–0.87Os0.09–0.12 Ir0.04–0.08)S2.
Sulfoarsenides
of
the
irarsite
(Ir,
Rh)AsS–hollingworthite
(RhAsS)
solid
solution
are
the
second
most
frequent
PGM
in
the
Amsaga
chromitites.
They
are
found
as
small
overgrowths
on
laurite
(Fig.
14a
and
b)
or
as
isolated
anhedral
grains
(<2
␮m
wide)
sometimes
associated
with
rutile
and
anorthite
(Fig.
14c).
The
composition
of
irarsites
measured
by
EDS
in
samples
MA
420
and
226
corresponds
to
the
formula
Ir0.60–0.69Rh0.31–0.40 AsS.
Sperrylite
(PtAs2)
is
the
coarser
(up
to
15
␮m
long)
PGM
found
(Fig.
14d)
and
is
located
in
the
ferritchromit
rim
of
spinel
from
samples
MA
273
and
238.
Minute
(<1
␮m)
rustenburgite
(Pt3Sn)
has
also
been
observed
forming
small
grains
overgrowing
laurite
in
sample
MA
226.
Six
bulk
chromitites
samples
have
been
analysed
for
bulk
PGE
content.
Three
subgroups
can
be
distinguished
on
the
basis
of
the
chondrite-normalized
PGE
pattern
(Fig.
15):
-
MA
226,
400
and
422
show
a
pronounced
Ru
peak
with
low
PPGE
(Rh,
Pt,
Pd)
compared
to
IPGE
(Ir,
Ru):
i.e.
low
(Pt/Rh)Nratios
with
(Rh/Ir)Nclose
to
unity
(0.5–1.9)
and
high
(Pt/Pd)Nratio
(3.2–3.4).
-
The
two
samples
containing
sperrylite
(MA
273
and
MA
238)
are
characterized
by
high
PGE
contents
compared
to
the
previous
group
and
high
PPGE/IPGE
ratios
((Rh/Ir)N:
3.5–4.9).
-
The
layered
chromitite
MA
425
(found
within
an
anorthositic
unit)
has
the
lowest
PGE
contents
with
a
more
or
less
flat
profile
and
a
slight
positive
peak
for
Ru.
8.
Discussion
8.1.
Impact
of
metamorphism
on
chromite
compositions
and
mineral
assemblages
Four
different
trends
can
be
observed
in
Fig.
7a
and
b
for
chromite
compositions.
The
increase
in
both
Fe#
and
Cr/(Cr
+
Al)
ratio
towards
ferritchromit
composition
can
be
ascribed
to
reequi-
libration
during
low-temperature
hydrous
metamorphism.
Indeed,
many
chromites
from
the
samples
following
this
trend
show
spongy
textures
characterized
by
(i)
numerous
inclusions
of

Author's personal copy
J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628 623
Fig.
14.
Backscattered
electron
images
of
PGM
enclosed
in
chromite
grains:
(a)
euhedral
laurite
with
irarsite
overgrowth
(MA
420);
(b)
euhedral
laurite
with
hollingworthite
bud
(MA
226);
(c)
laurite
(white)
associated
with
rutile
(grey)
and
plagioclase
(black)
in
a
chromite
(dark
grey)
(MA
420);
(d)
sperrylite
grain
surrounded
py
Pd
oxides
within
the
ferritchromit
rim
(sample
MA
273).
chlorite,
(ii)
thick
rim
of
ferritchromit
composition,
(iii)
strong
compositional
variation
from
grain
to
grain.
Ferritchromit
rims
or
grains
are
often
interpreted
as
low
grade
reequilibration
of
chromite
under
an
oxidizing
environment
(see
Mukherjee
et
al.,
2010).
Chromite
lying
along
this
trend
cannot
be
used
to
infer
the
composition
of
their
parental
magma.
Some
ultramafic
metacumulates
have
preserved
olivine–amphibole
assemblages
with
an
aluminous
spinel.
The
composition
of
these
spinels
and
the
association
of
high-Al
amphibole
and
green
aluminous
spinel
with
olivine
are
typical
of
high
temperature
(granulite
to
amphibolites
facies
conditions)
meta-troctolites
(Tenthorey
et
al.,
1996;
Berger
et
al.,
2010;
see
Figs.
7a
and
8)
and
sakenites
(Giuliani
et
al.,
2006;
Raith
et
al.,
2008).
Green
spinels
from
spinel–amphibolite
and
olivine–amphibolite
are
strongly
aluminous
and
depleted
in
Cr.
Their
composition
suggests
that
high-grade
metamorphism
has
led
to
enrichment
in
the
spinel
end-member
with
leaching
of
Cr,
a
shift
comparable
to
the
replacement
trend
described
by
Rollinson
et
al.
(2002)
in
metamorphosed
Archean
chromitite–anorthosite
associations
from
Greenland.
At
variance
with
our
interpretation,
these
authors
interpret
the
replacement
trend
as
a
result
of
late
magmatic
interaction
between
high-Cr
spinels
and
evolved
interstitial
melt.
Recent
studies
on
chromitites
from
the
Archean
Sittampundi
complex,
in
India
(Dutta
et
al.,
2011)
also
propose
a
metamorphic
origin
for
the
aluminous
green–blue
spinel.
Chromite
composition
trending
towards
these
Mg–Al
spinels
in
Fig.
7a
(i.e.
brecciated
chromitite
MA
436)
are
interpreted
to
have
been
partially
equili-
brated
during
high
grade
metamorphism
and
they
will
not
be
used
to
calculate
parental
magma
compositions.
The
strong
variations
in
Fe#
for
a
constant
Cr/(Cr
+
Al)
ratio
observed
for
disseminated
chromitites
(Fig.
7a)
also
character-
ized
by
the
development
of
ferritchromit
rims
are
more
difficult
to
interpret.
Because
these
samples
show
strong
compositional
variation
from
grain
to
grain
with
the
frequent
development
of
a
ferritchromit
rim,
one
can
infer
that
their
composition
has
been
modified
by
metamorphism.
One
of
these
sample
(MA
422)
is
a
serpentine-bearing
disseminated
chromitite
probably
represent-
ing
a
former
olivine–chromitite.
Fe–Mg
exchange
between
spinel
and
olivine
is
a
quick
reaction
that
led
to
partial
reequilibration
of
these
two
phases
at
PT
conditions
(Ozawa,
1983).
As
olivine

Author's personal copy
624 J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628
Ir Ru Rh Pt Pd
0.001
0.01
0.1
1
10
MA 226
MA 238
MA 273
MA 400
MA 422
MA 425
Chondrite normalised concentrations
Fig.
15.
PGE
pattern
of
selected
chromitites
from
the
GAC.
Normalization
values
are
from
Naldrett
and
Duke
(1980).
contains
only
minor
amounts
of
Al
and
Cr,
reequilibration
with
spinel
will
not
modify
the
Cr/(Cr
+
Al)
ratio
of
the
latter
but
will
strongly
affect
the
partition
of
Fe
and
Mg.
The
disseminated
and
the
massive
chromitite
(MA
226)
lying
on
the
Fe–Mg
exchange
trend
with
constant
Cr/(Cr
+
Al)
ratios
represent
chromite
that
has
been
reequilibrated
with
a
low
Al
and
Cr
but
high
Fe–Mg
phase,
probably
olvine
and/or
clinopyroxene.
The
massive
chromitites
have
chromite
as
main
phase
(up
to
95
vol.%).
Reequilibration
with
a
minor
host
silicate
phase
will
thus
have
minor
effects
on
the
composition
of
these
chromites.
More-
over,
massive
chromitites
(except
MA
226)
show
homogeneous
chromite
compositions
in
a
given
sample,
they
do
not
show
fer-
ritchromit
rims,
they
are
not
affected
by
low-temperature
growth
of
chlorite
as
in
brecciated
chromitites
and
they
are
very
compa-
rable
to
Fiskenaesset
and
Sittampundi
chromites
with
preserved
igneous
compositions
(Fig.
7b;
Rollinson
et
al.,
2010;
Dutta
et
al.,
2011).
The
last
trend
observed
for
massive
chromite
can
thus
be
interpreted
as
a
preserved
igneous
feature.
Silicate
rocks
have
also
been
affected
by
HT
and
LT
meta-
morphism.
The
composition
of
their
minerals
cannot
be
used
to
estimate
compositions
of
their
parental
melts
The
granu-
lar
opx–cpx–plag
assemblages
in
fine-grained
amphibolites
and
mela-amphibolites
argue
for
a
peak
HT–MP
granulite
grade
meta-
morphism:
up
to
910 ◦C
at
5–6
kbar
in
anhydrous
environment
(Fig.
13).
Pyroxenes
were
nearly
totally
consumed
to
amphi-
bole
during
a
retrogressive
hydrous
event
in
the
amphibolite
and
lower
granulite
facies
conditions
(900–650 ◦C).
A
few
samples
have
preserved
evidences
for
direct
replacement
of
clinopyrox-
ene
by
amphibole
(Fig.
6c),
which
is
concordant
with
a
retrograde
evolution
in
the
pseudosection
built
for
MA
37
bulk
composition
(Fig.
13).
There
is
a
relationship
between
the
An
content
of
plagioclase
and
the
Na
and
Al
contents
of
co-existing
amphibole
from
fine-grained,
leuco-
and
mela-amphibolites,
the
most
calcic
plagioclase
being
at
equilibrium
with
a
Na-
and
Al-rich
amphibole
(Fig.
10).
This
cer-
tainly
indicates
a
metamorphic
origin
for
the
amphibole
in
those
rocks,
the
Na
enclosed
in
former
igneous
plagioclase
has
been
trans-
ferred
into
amphibole
thanks
to
a
pargasite
substitution
during
clinopyroxene
breakdown
(the
latter
cannot
deliver
large
quanti-
ties
of
Na).
The
An
content
of
the
plagioclase
in
amphibole-bearing
rocks
thus
no
more
reflects
the
original
igneous
composition,
as
attested
by
nearly
pure
anorthite
composition
(0.97–0.98
An
mol.%;
Fig.
10)
of
plagioclase
in
spinel-amphibolite
MA
25.
Exceptions
are
observed
in
Fig.
10
for
two
mela-amphibolites
(MA
28
and
MA
34).
This
anomaly
cannot
be
explained
by
different
equilibration
temperatures
in
these
samples
(800–850 ◦C)
compared
to
other
amphibolites
(720–930 ◦C)
but
could
reflect,
however,
the
preser-
vation
of
igneous
amphibole
that
has
been
reequilibrated
with
plagioclase
during
granulite-facies
metamorphism.
As
discussed
above,
olivine–amphibole
rocks
and
spinel
amphibolites
share
many
characteristic
with
meta-troctolites
metamorphosed
under
HT
amphibolite
to
granulite
grade.
The
origin
of
chromite–amphibole
association
in
some
chromi-
tites
is
more
ambiguous.
These
samples
are
found
in
close
spatial
association
with
anorthosites.
In
the
Fiskenaesset
complex,
chromi-
tites
in
anorthosites
have
amphibole
as
matrix
phase
(see
Rollinson
et
al.,
2010).
On
the
basis
of
comparison
with
worldwide
unmeta-
morphosed
amphibole–chromitite
occurrences,
Rollinson
et
al.
(2010)
proposed
that
amphibole
from
chromitites
is
of
igneous
origin.
Dutta
et
al.
(2011)
however
observed
textural
evidence
for
clinopyroxene
replacement
by
amphibole
in
chromitites
from
the
Sittampundi
complex.
In
our
samples,
there
is
no
textural
evi-
dence
preserved
but
the
amphibole
composition
matches
the
one
of
spinel–amphibole
metamorphic
rocks
(see
Fig.
11).
We
therefore
propose
that
amphibole
in
chromitite
is
of
metamorphic
origin
in
the
Guelb
el
Azib
complex.
As
a
conclusion,
only
the
mono-mineralic
rocks
(or
nearly
so)
are
expected
to
have
preserved
their
original
igneous
mineral
compo-
sitions.
This
is
most
probably
the
case
for
the
anorthosites
and
the
chromites
from
massive
and
layered
chromitites.
8.2.
Determination
of
the
chromitite
parental
melts
Kamenetsky
et
al.
(2001)
have
shown
that
the
Al
and
Ti
contents
of
igneous
chromites
are
linearly
correlated
to
that
of
the
melt.
Maurel
and
Maurel
(1983)
proposed
an
equation
to
compute
the
FeO/MgO
ratio
in
the
parental
melt
of
igneous
chromites.
The
value
of
this
ratio
in
the
spinel
is
however
strongly
sensitive
to
variations
oxygen
fugacity
and,
as
a
consequence,
large
errors
are
attached
to
the
determination
of
the
FeO/MgO
ratio
in
the
melt.
Only
primary
chromites
following
the
proposed
igneous
trend
(Fig.
7b)
and
showing
no
petrographical
nor
chemical
evidence
for
hydrothermal
alteration,
oxidation
or
high
grade
metamor-
phism
were
used
for
the
calculations.
Few
rutile
exsolutions
were
observed
in
chromites,
the
TiO2content
of
the
equilibrium
melt
is
most
probably
underestimated
(see
Rollinson
et
al.,
2002)
and
will
therefore
not
be
used
as
a
discriminating
factor.
Furthermore,
evolved
chromites
with
high
Fe#
show
a
strong
decrease
of
their
Al
content.
They
have
probably
co-crystallized
with
plagioclase
and
their
Al
content
has
been
buffered
by
the
feldspar.
Consequently,
only
the
results
for
the
chromites
crystallized
during
and
before
the
Al
peak
(Fig.
9)
are
plotted
in
the
FeO/MgO
vs
Al2O3diagram
(Fig.
16).
The
melt
in
equilibrium
with
the
most
primitive
chromite
(highest
Mg#,
MA
273)
has
15
wt%
Al2O3and
plots
both
in
the
field
of
Archean
tholeiites
(Fig.
16).
There
is
almost
no
increase
of
Al2O3
contents
with
increasing
FeO/MgO
ratio
in
melts
that
crystallized
chromitites,
up
to
15–16
wt%
for
the
most
evolved
chromite
(low-
est
Mg#,
MA
238)
parental
melts,
confirming
that
there
was
no
massive
fractionation
of
plagioclase
during
this
sequence
of
differ-
entiation.
All
calculated
melts
fall
in
the
field
of
Archean
tholeiite
in
Fig.
16.
Chromites
from
MA
400
are
not
in
equilibrium
with
common
Archean
basic–ultrabasic
melts
(Fig.
16).
They
also
plot
outside
the
igneous
trend
formed
by
other
massive
chromitites
(Fig.
9).
Although
there
is
no
petrographic
evidence
for
metamorphic
reequilibration,
the
chromite
from
this
sample
probably
underwent
metamorphic
reequilibration
with
host
phase.
It
is
indeed
the
most
Fe3+-rich
spinel
of
all
massive
chromitites
(up
to
0.17
a.p.f.u.).

Author's personal copy
J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628 625
FeO/MgO
0
1
2
3
4
5
6
8
10
12
14
16
18
20
Basaltic komatiites
Archean tholeiites
Izu-Bonin
Al O
23
Cascades
MA 400
Chromitites
Massive
50-70%
70-90%
90-100%
Layered
Chl Amp
Fig.
16.
Composition
of
the
melt
in
equilibrium
with
GAC
chromites
with
preserved
igneous
composition
(Fig.
7b)
compared
to
basaltic
komatiites,
Archean
tholeiites,
Izu-Bonin
and
Cascades
arc
magmas
(data
from
the
online
GEOROC
database).
The
most
evolved
chromite
from
sample
MA
226
has
not
been
plotted
in
this
diagram
because
it
has
probably
been
crystallizing
along
with
plagioclase.
8.3.
Inferences
on
the
primary
magma
and
the
conditions
of
crystallization
The
igneous
chromite
with
highest
Mg#
crystallized
from
a
melt
having
a
Mg#
around
47.
This
is
too
low
to
be
in
equilibrium
with
a
classical
mantle
peridotite
with
pyrolitic
or
depleted
harzburgitic
composition.
The
melt
that
has
crystallized
the
igneous
chromite
was
either
evolved
or
it
has
been
formed
by
a
pertubated
Fe-rich
mantle.
The
GAC
also
present
large
volumes
of
ultramafic
rocks
in
the
form
of
serpentinite
(former
spinel-bearing
olivine-rich
rocks)
and
ultramafic
metacumulates
(former
olivine–plagioclase–spinel
rocks).
It
is
not
possible
to
estimate
the
composition
of
their
parental
melts
because
of
the
strong
imprint
of
both
HT
granulitic
metamorphism
and
late
greenschist
facies
alteration.
However,
their
presence
indicates
at
least
that
the
primitive
magma
was
more
Mg-rich
compared
to
the
one
that
crystallized
chromites.
Using
information
on
rocks
with
preserved
igneous
mineralogy
and/or
mineral
composition
and
on
the
nature
of
the
igneous
pre-
cursors
of
metamorphic
rocks,
the
primary
magma
and
is
liquid
line
of
descent
must
fit
to
the
following
observations:
(i)
The
melt
has
crystallized
primitive
ultramafic
olivine-rich
cumulates
(olivine–pyroxenes–spinel
±
plagioclase
rocks),
chromites,
mela-
and
leuco-gabbros
(plagioclase
and
pyrox-
ene,
±amphibole
rocks)
and
highly
calcic
anorthosites
(Fig.
4).
The
hornblendites
(amphibole
cumulates
or
meta-pyroxenite)
form
only
small
elliptical
bodies
within
ultramafic
units
and
are
consequently
not
a
major
constituent
of
the
GAC.
(ii)
The
anorthosite
and
some
leucogabbros
have
a
highly
calcic
plagioclase.
The
formation
of
such
rocks
requires
an
aluminous
parental
melt
with
high
Ca/Na
ratio.
Massive
fractionation
of
clinopyroxene
did
certainly
not
occur
before
plagioclase
sat-
uration
because
it
would
have
strongly
decreased
the
Ca/Na
ratio
of
the
melt
parental
to
the
anorthosite.
(iii)
The
compositional
trends
observed
for
chromitites
(decrease
in
Al
content
for
the
most
Fe-rich
igneous
chromites),
together
with
field
observation,
suggest
that
plagioclase
saturation
in
the
melt
was
reached
during
chromite
crystallization
(Fig.
9).
(iv)
Compared
to
Archean
and
Paleoproterozoic
chromites
from
komatiites,
greenstone
belts
and
layered
intrusions,
those
from
Guelb
el
Azib
have
higher
Fe#
and
lower
Cr#
mean
values
(Figs.
7
and
8).
The
parental
melt
to
chromitites
has
thus
high
Fe/Mg
ratio
and/or
large
amount
of
a
phase
with
low
Fe/Mg
ratio
(olivine)
has
been
fractionated
from
a
high-Mg
melt
before
chromite
saturation
(as
suggested
by
the
presence
of
ultramafic
olivine-rich
cumulates
in
GAC).
Considering
these
lines
of
evidence,
the
primitive
melt
should
have
been
an
Mg-rich,
high
alumina
basaltic
melt.
Such
a
compo-
sition
explains
the
presence
of
former
olivine-rich
meta-igneous
rocks
co-existing
with
large
volume
of
plagioclase-rich
metagab-
bros
and
anorthosites.
In
addition,
experiments
have
shown
that
increasing
water
contents
in
the
melt
promotes
the
crystalliza-
tion
of
Ca-rich
plagioclase
(see
Grove
and
Baker,
1984;
Sisson
and
Grove,
1993;
Takagi
et
al.,
2005;
Feig
et
al.,
2006).
This
is
com-
patible
with
the
probable
presence
of
few
igneous
amphiboles
in
mela-amphibolites
MA
28
and
34
and
with
the
occurrences
of
hornblendite
that
are
commonly
formed
from
hydrous
mag-
mas
in
similar
context
(see
Polat
et
al.,
2012).
Increasing
water
content
leads
to
the
enlargement
of
the
temperature
interval
of
pyroxene
crystallization
and
reduces
the
window
where
plagio-
clase
and
chromite
co-precipitate
(Berndt
et
al.,
2005;
Feig
et
al.,
2006;
Hamada
and
Fujii,
2008).
The
primitive
melt
was
thus
not
saturated
with
water,
it
was
slightly
hydrous.
Experimental
studies
of
subalkaline
melt
crystallization
demon-
strate
that
increasing
pressure
tends
to
stabilize
pyroxenites
over
olivine–plagioclase
cumulates
and
leads
to
the
formation
of
igneous
garnet
(Muntener
et
al.,
2001;
Villiger
et
al.,
2004,
2007;
Alonso-Perez
et
al.,
2009).
Pyroxenites
are
absent
or
minor
in
the
GAC,
garnet
was
not
observed
neither
as
igneous
nor
metamorphic
phase
while
metamorphosed
olivine–plagioclase
rocks
are
present.
We
can
thus
infer
a
low
pressure
of
crystallization,
at
depth
corre-
sponding
to
upper
or
middle
crust.
Formation
of
the
GAC
through
low
pressure
crystallization
of
a
hydrous
magma
is
in
agreement
with
previous
studies
on
similar
complexes
(Weaver
et
al.,
1981;
Rollinson
et
al.,
2010;
Polat
et
al.,
2011;
Dutta
et
al.,
2011).
The
melt
became
subsequently
enriched
in
Al
through
fractional
crystallization
of
olivine
and
chromite
and
reached
a
composition
matching
that
of
Archean
tholeiites
(Fig.
16,
see
below)
just
before
plagioclase
saturation.
An
Archean
tholei-
ite
composition
for
the
Archean
anorthosite
parental
melt
agrees
with
results
obtained
by
Henderson
et
al.
(1976),
Ashwal
(1993)
and
Rollinson
et
al.
(2010)
on
the
basis
of
inverted
REE
contents
of
plagioclase.
8.4.
PGM
crystallization
and
PGE
fractionation
PGM
co-crystallized
along
with
chromites
and
base
metal
sulp-
hides.
Detailed
observations
with
the
electron
microscope
and
bulk
PGE
patterns
have
shown
that
three
phases
of
PGE
mineralization
can
be
distinguished
in
chromitites.
(i)
The
earliest
one
is
evidenced
by
the
formation
of
euhedral
lau-
rite
and
millerite
(Fig.
14a
and
b).
The
shape
of
laurite
together
with
experimental
results
(Brenan
and
Andrews,
2001)
show
that
laurite
is
a
high-temperature
igneous
PGM
in
the
GAC
as
in
ophiolitic
chromitites
where
it
is
entrapped
during
mag-
matic
growth
of
chromite
(Augé,
1985).
This
mineralization
phase
has
characteristic
PGE
patterns
with
high
IPGE
contents
compared
to
PPGE
(samples
MA
226,
400
and
422;
Fig.
15)
which
fits
relatively
well
with
that
of
chromitites
from
ophio-
lites,
komatiites
and
layered
intrusions
(Cabri,
2002;
Naldrett
et
al.,
2012;
Pagé
et
al.,
2012).
Chromite
crystallized
in
komati-
itic
and
tholeiitic
melts
concentrates
IPGE.
These
elements
may
be
present
in
chromite
as
solid
solution
(Pagé
et
al.,
2012;
Brenan
et
al.,
2012)
or
co-crystallized
with
IPGE
min-
erals
due
to
their
low
solubilities
in
basic
magmas
(Pagé
et
al.,

Author's personal copy
626 J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628
Table
3
Major
and
trace
element
composition
of
bulk
chromitite
samples.
MA
226
MA
238
MA
240
MA
241
MA
273
MA
400
MA
422
MA
425
Major
elements
(wt%)
SiO23.02
8.94
3.33
12.85
12.14
1.65
19.71
10.23
TiO20.31
0.36
0.29
0.24
0.54
0.38
0.17
0.44
Al2O316.23
17.11
23.54
22.17
20.56
11.77
7.29
22.12
Fe2O327.54
25.13
30.59
24.39
19.58
24.59
26.66
23.57
MnO
0.51
0.69
0.61
0.69
0.32
0.44
0.50
0.37
MgO
10.03
10.30
4.92
11.53
11.68
8.39
19.86
8.02
CaO
0.10
0.09
0.12
0.18
2.79
0.13
1.21
2.54
Na2O
0.14
0.21
0.16
0.22
0.37
0.16
0.09
0.32
K2O
0.02
0.03
0.02
0.14
0.08
0.02
0.02
0.07
P2O50.00
0.00
0.00
0.00
0.01
0.01
0.02
0.01
Cr2O340.45
29.58
35.13
21.04
32.61
50.35
18.32
32.34
LOI
−0.3
6.02
0.3
3.9
−0.9
−0.25
6.05
−1.11
98.06
98.46
99.01
97.34
99.77
97.65
99.90
98.93
Trace-elements
(ppm)
Co 169
132
214
176
103
120
128
136
Cu
29
27
37
49
23
36
26
21
Ni
1597
1295
847
1188
1334
833
2237
969
Zn 1372 3207 2780 2919
556
2615
1115
1969
V
439
772
648
555
987
342
391
940
PGE
(ppb)
Ir
42
241
75
63
28
10
Ru
329
583
242
709
355
45
Rh
8
442
97
37
20
6
Pt 6 783 1700 42
22
21
Pd
1
463
418
6
10
2012).
The
positive
correlation
between
Cu
and
Ru
contents
in
chromitites
and
the
absence
of
correlation
between
PGE
and
Cr
or
Si
(Table
3)
contents
suggest
that
the
PGE
(and
espe-
cially
IPGE)
have
been
entrapped
by
sulphide
phases
that
now
form
inclusions
within
chromite
grains.
The
igneous
PGE
min-
eralization
most
probably
originated
through
separation
of
a
sulphide
fraction
from
the
silicate
melt,
in
agreement
with
the
more
compatible
behaviour
of
IPGE
relative
to
PPGE
in
sul-
phide
melts
compared
to
silicate
magmas
(Sattari
et
al.,
2002).
The
sulphide
melt/silicate
magma
immiscibility
event
could
have
been
triggered
either
by
mixing
between
evolved
and
primitive
magmas
(Irvine,
1977)
or
by
assimilation
of
enclos-
ing
gabbros
by
the
fractionating
melt
(Bédard
and
Hébert,
1998;
Gervilla
et
al.,
2005).
Considering
the
strong
meta-
morphic
imprint
on
the
GAC,
more
detailed
geochemical
and
isotopic
analyses
are
needed
to
choose
between
these
two
processes.The
layered
chromitite
(MA
425)
associated
with
anorthosite–leuco-amphibolite
samples
also
shows
a
PGE
pat-
tern
with
a
peak
in
Ru
but
with
lower
bulk
PGE
contents
and
lower
Ru/Rh
ratio
compared
to
MA
226–400–422.
This
deple-
tion
in
bulk
PGE
and
in
Ru
relative
to
other
PGE
could
be
explained
by
strong
depletion
of
Ru
and,
to
a
lesser
extent,
bulk
PGE
contents
in
the
parental
melt.
Indeed,
on
the
basis
of
sim-
ulations
with
MELTS
and
according
to
mineral
compositions,
it
is
stressed
that
the
chromitite
in
anorthosite
has
crystallized
from
evolved
aluminous
melts
in
comparison
to
more
primi-
tive
Mg-rich
and
Al-poor
chromitites.
Hence,
the
parental
melt
to
evolved
chromites
was
already
depleted
in
PGE,
especially
IPGE
over
PPGE,
due
to
the
igneous
segregation
of
laurite
with
more
primitive
chromites.
(ii)
The
second
phase
of
mineralization
is
evidenced
by
the
over-
growth
of
anhedral
irarsite-hollingworthite
grains
on
laurite
or
as
isolated
grains
within
chromite
(Fig.
14).
This
event
of
sulpho-arsenide
crystallization
is
not
marked
in
bulk
PGE
pat-
terns
as
it
remobilizes
the
PGE
in
a
closed
system
but
it
is
linked
to
an
increase
in
arsenic
activity.
A
few
irarsite
and
hollingwor-
thite
were
observed
along
with
rutile
and
anorthite
(Fig.
14c),
an
igneous
or
late-magmatic
origin
is
more
plausible.
(iii) The
third
mineralizing
phase
consists
of
the
crystallization
of
sperrylite
in
the
ferritchromit
rim
of
chromite
grains
and
scarce
rustenburgite
overgrowth
at
the
rim
of
laurite.
A
preferen-
tial
incorporation
of
PPGE
is
observed
from
both
mineralogy
and
bulk
PGE
patterns
of
samples
MA
238
and
273
(Fig.
15)
which
are
characterized
by
higher
Rh,
Pt
and
Pd
contents
(up
to
1700
ppb
Pt).
The
evolution
from
IPGE
to
nearly
flat
PGM
pattern
due
to
PPGE
enrichment
(Fig.
15)
is
similar
to
that
observed
in
Paleoproterozoic
ferropicrites
from
Pechenga
(Brügmann
et
al.,
2000).
Since
ferritchromit
preferentially
develops
around
inclusions
of
chlorite
and
along
the
border
of
chromite
grains,
this
phase
of
Pt–Pd
mineralizations
is
inter-
preted
as
a
result
of
hydrothermal
activity
under
high
arsenic
activities.
Similar
conclusions
have
been
drawn
for
Pt–Pd
min-
eralization
in
hydrothermally
altered
ophiolitic
chromitites
(Leblanc,
1991;
Prichard
et
al.,
2008)
and
for
the
evolution
of
the
PGM
from
sulphides,
to
sulphoarsenides
and
arsenides
followed
by
hydrothermal
intermetallic
alloys
that
are
also
observed
in
the
Two
Duck
lake
intrusion
from
the
Coldwell
complex
(Watkinson
and
Ohnenstetter,
1992).
8.5.
The
Guelb
el
Azib
complex:
the
metamorphosed
equivalent
of
Archean
anorthosite
layered
bodies
The
GAC
shares
many
similarities
with
Archean
anorthosite
complexes.
It
is
occurring
within
TTG
terrains
in
close
spatial
asso-
ciation
with
suprecrustals
(impure
marbles,
amphibolites,
BIF),
it
is
characterized
by
highly
calcic
anorthosite
and
Fe-rich
chromite
and
it
is
structured
as
layered
sequences
of
olivine-rich
cumulate
rocks,
former
gabbros
and
anorthosites
with
minor
pyroxenites
and
few
hornblendites
bodies
(see
Windley
and
Garde,
2009;
Rollinson
et
al.,
2010;
Polat
et
al.,
2011,
2012).
It
is
difficult
to
ascribe
a
tectonic
setting
to
the
GAC
due
to
strong
recrystallization
under
granulite
to
greenschist
facies
conditions.
But
several
lines
of
evidences
point
to
a
subduction
zone
origin:
(i)
The
highly
calcic
composition
of
plagioclase
in
anorthosites
and
the
hydrous
nature
of
the
primitive
melts
are
characteris-
tic
of
modern
hydrous
arc
magmas
and
xenoliths
(Arculus
and

Author's personal copy
J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628 627
Wills,
1980;
Takagi
et
al.,
2005).
(ii)
As
pointed
out
by
Rollinson
et
al.
(2010),
the
association
amphibole–calcic
plagioclase–ferrian
chromites
is
also
found
in
arc
xenoliths
(Arculus
and
Wills,
1980),
even
if
the
primary
or
secondary
origin
of
amphibole
is
debated
in
layered
anorthosite
complexes
(Owens
and
Dymek,
1997;
Rollinson
et
al.,
2010;
Dutta
et
al.,
2011;
this
study).
(iii)
The
inter-
pretation
of
spinel
compositions
in
terms
of
tectonic
setting
is
not
straightforward.
Indeed,
compositional
fields
of
chromites
from
various
modern
Phanerozoic
tectonic
settings
are
largely
overlap-
ping
(Barnes
and
Roeder,
2001)
and
most
GAC
chromites
plot
in
the
fields
of
modern
flood
basalts
in
the
Fe#–Cr#
plot
but
falls
in
the
ophiolite
field
in
the
trivalent
ion
plot.
Rollinson
et
al.
(2010)
moreover
noticed
that
some
spinels
in
modern
arc
tholeiites
have
composition
close
to
those
analysed
in
UMA
complexes.
The
FeO/MgO
ratio
and
Al2O3content
of
the
melts
in
equilibrium
with
igneous
chromites
are
very
comparable
to
modern
continental
(Cascades)
and
oceanic
(Izu-Bonin)
arc
basalts
(Fig.
16,
comparison
data
from
the
online
GEOROC
database).
The
GAC
chromites
with
preserved
igneous
compositions
are
comparable
to
those
analysed
in
the
Archean
Fiskenaesset
and
Sit-
tampundi
UMA
complexes
(Figs.
7b
and
8).
A
supra-subduction
zone
origin
is
proposed
by
Polat
et
al.
(2009,
2010,
2011,
2012)
for
the
Fiskenaesset
complex
on
the
basis
of
trace-element
and
iso-
topic
data
(negative
Nb
anomalies
in
most
samples,
positive
initial
␧
Nd).
Although
the
few
petrological
evidences
converge
to
a
supra-
subduction
zone
setting
for
the
GAC,
more
detailed
geochemical
analysis
are
needed
to
confirm
this
hypothesis.
9.
Conclusions
The
Archean
Guelb
el
Azib
complex
in
the
West
African
cra-
ton
is
a
metamorphosed
equivalent
of
famous
Archean
anorthosite
bodies.
Despite
granulite
(up
to
900 ◦C,
5
kbar)
to
greenschist
meta-
morphic
events
that
have
affected
the
complex,
few
preserved
igneous
mineral
compositions
and
the
lithological
nature
of
the
igneous
precursors
before
metamorphism
lead
to
the
following
conclusions:
•The
primitive
melt
was
a
slightly
hydrous
high
alumina
basaltic
melt.
It
evolved
towards
Archean
tholeiite-like
composition
through
the
massive
fractionation
of
olivine
and
chromite
gener-
ating
a
sequence
of
ultramafic
cumulates
now
transformed
into
serpentinites
and
olivine–amphibole–spinel
rocks
(metatrocto-
lites).
•Fe-rich,
Cr-poor
chromites,
An-rich
anorthosite
and
metamor-
phosed
gabbros
formed
from
an
Archean
tholeiite
parental
melt.
The
wide
crystallization
window
of
chromite
and
the
calcic
nature
of
the
plagioclase
are
promoted
by
the
lack
of
massive
fractionation
of
clinopyroxene.
•IPGE
minerals
such
as
laurite
were
precipitated
during
igneous
crystallization
of
chromitites
and
IPGE
were
remobilized
dur-
ing
late-magmatic
crystallization
of
hollingworthite
and
irarsite.
A
Pt–Pd
phase
of
mineralization,
represented
by
sperrylite
and
rustenburgite
is
linked
with
late
low-temperature
hydrothermal
metamorphism.
•The
main
petrological
characteristics
of
the
GAC
are
compatible
with
low-pressure
of
crystallization
at
depth
corresponding
to
upper
or
middle
crust.
Geochemical
analyses
are
needed
to
pre-
cise
the
tectonic
setting
of
the
GAC
but,
by
comparison
with
other
Archean
anorthosite
complexes,
it
could
have
been
formed
in
a
supra-subduction
zone
setting.
Acknowledgements
This
study
was
funded
by
a
FRS-FNRS
grant
to
JB.
We
would
like
to
thank
Mohamed
Dahmada,
Ousmane
N’Diaye
(deceased
in
November
2010),
Maloum
Baba,
Med
Salem
and
the
OMRG
for
their
support
and
for
the
fantastic
fieldtrip
in
the
Amsaga
during
November
2008.
Reviews
made
by
Hugh
Rollinson
and
an
anony-
mous
referee
together
with
the
editorial
handling
of
Guochun
Zhao
were
greatly
appreciated.
References
Arculus,
R.J.,
Wills,
K.J.A.,
1980.
The
petrology
of
plutonic
blocks
and
inclusions
from
the
lesser
antilles
island
arc.
Journal
of
Petrology
21,
743–799.
Alonso-Perez,
R.,
Müntener,
O.,
Ulmer,
P.,
2009.
Igneous
garnet
and
amphibole
fractionation
in
the
roots
of
island
arcs:
experimental
constraints
on
andesitic
liquids.
Contributions
to
Mineralogy
and
Petrology
157,
541–558.
Ashwal,
L.D.,
1993.
Anorthosites.
Springer-Verlag,
Berlin,
422pp.
Augé,
T.,
1985.
Platinum
group
mineral
inclusions
in
ophiolitic
chromites
from
the
Vourinos
complex,
Greece.
The
Canadian
Mineralogist
23,
163–171.
Auvray,
B.,
Peucat,
J.J.,
Potrel,
A.,
Burg,
J.P.,
Caruba,
C.,
Lo,
K.,
1992.
Données
géochronologiques
nouvelles
sur
1‘Archéen
de
1‘Amsaga
(Dorsale
Réguibat,
Mauritanie).
Comptes
Rendues
de
1‘Académie
dés
Sciences
de
Paris
315,
63–70.
Barnes,
S.J.,
Roeder,
P.L.,
2001.
The
range
of
spinel
compositions
in
terrestrial
mafic
and
ultramafic
rocks.
Journal
of
Petrology
42,
2279–2302.
Barrère,
J.,
1967.
Le
groupe
précambrien
de
l‘Amsaga
entre
Afar
et
Akjoujt
(Mau-
ritanie).
Etude
d‘un
métamorphisme
profond
et
de
ses
relations
avec
la
migmatisation.
Unpublished
PhD
Thesis,
Clermont-Ferrand,
France,
275pp.
Barton,
J.M.,
1996.
The
messina
layered
intrusion,
Limpopo
Belt,
South
Africa:
an
example
of
in-situ
contamination
of
an
Archean
anorthosite
complex
by
conti-
nental
crust.
Precambrian
Research
78,
139–150.
Bédard,
J.H.,
Hébert,
R.,
1998.
Formation
of
chromitites
by
assimilation
of
crustal
pyroxenites
and
gabbros
into
peridotitic
intrusions:
North
Arm
Mountain
massif,
Bay
of
Islands
ophiolite,
Newfoundland,
Canada.
Journal
of
Geophysical
Research
103,
5165–5184.
Berger,
J.,
Féménias,
O.,
Ohnenstetter,
D.,
Plissart,
G.,
Mercier,
J.C.C.,
2010.
Origin
and
tectonic
significance
of
corundum–kyanite–sapphirine
amphibolites
from
the
Variscan
French
Massif
Central.
Journal
of
Metamorphic
Geology
28,
341–360.
Berndt,
J.,
Koepke,
J.,
Holtz,
F.,
2005.
An
experimental
investigation
of
the
influence
of
water
and
oxygen
fugacity
on
differentiation
of
MORB
at
200
MPa.
Journal
of
Petrology
46,
135–167.
Brenan,
J.M.,
Andrews,
D.,
2001.
High-temperature
stability
of
laurite
and
Ru–Os–Ir
alloy
and
their
role
in
PGE
fractionation
in
mafic
magmas.
The
Canadian
Miner-
alogist
39,
341–360.
Brenan,
J.M.,
Finnigan,
C.F.,
McDonough,
W.F.,
Homolova,
V.,
2012.
Experimental
constraints
on
the
partitioning
of
Ru,
Rh,
Ir,
Pt
and
Pd
between
chromite
and
silicate
melt:
the
importance
of
ferric
iron.
Chemical
Geology
302–303,
16–32.
Brügmann,
G.E.,
Hanski,
E.J.,
Naldrett,
A.J.,
Smolkin,
V.F.,
2000.
Sulphide
segregation
in
ferropicrites
from
the
Pechenga
Complex,
Kola
Peninsula,
Russia.
Journal
of
Petrology
41,
1721–1742.
Cabri,
L.J.,
2002.
The
geology,
geochemistry,
mineralogy
and
mineral
beneficiation
of
platinum-group
elements.
CIM
Special
Volume
54,
852.
Connolly,
J.A.D.,
2005.
Computation
of
phase
equilibria
by
linear
programming:
a
tool
for
geodynamic
modeling
and
its
application
to
subduction
zone
decarbon-
ation.
Earth
and
Planetary
Science
Letters
236,
524–541.
Dharma
Rao,
C.V.,
Santosh,
M.,
Sajeev,
K.,
Windley,
B.F.
Chromite–silicate
chemistry
of
the
Neoarchean
Sittampundi
Complex,
southern
India:
impli-
cations
for
subduction-related
arc
magmatism.
Precambrian
Research,
http://dx.doi.org/10.1016/j.precamres.2011.11.012,
in
press.
Diener,
J.F.A.,
Powell,
R.,
White,
R.W.,
Holland,
T.J.B.,
2007.
A
new
ther-
modynamic
model
for
clino-
and
orthoamphiboles
in
the
system
Na2O–CaO–FeO–MgO–Al2O3–SiO2–H2O–O.
Journal
of
Metamorphic
Geology
25,
631–656.
Dutta,
U.,
Bhui,
U.K.,
Sengupta,
P.,
Sanyal,
S.,
Mukhopadhyay,
D.,
2011.
Magmatic
and
meta-morphic
imprints
in
2.9
Ga
chromitites
from
the
Sittampundi
layered
complex,
Tamil
Nadu,
India.
Ore
Geology
Reviews
40,
90–107.
Feig,
S.,
Koepke,
J.,
Snow,
J.,
2006.
Effect
of
water
on
tholeiitic
basalt
phase
equilibria:
an
experimental
study
under
oxidizing
conditions.
Contributions
to
Mineralogy
and
Petrology
152,
611–638.
GEOROC
online
database.
http://georoc.mpch-mainz.gwdg.de/georoc/
Gervilla,
F.,
Proenza,
J.A.,
Frei,
R.,
González-Jiménez,
J.M.,
Garrido,
C.J.,
Melgarejo,
J.C.,
Meibom,
A.,
Díaz-Martínez,
R.,
Lavaut,
W.,
2005.
Distribution
of
platinum-group
elements
and
Os
isotopes
in
chromite
ores
from
Mayarí-Baracoa
Ophiolitic
Belt
(eastern
Cuba).
Contributions
to
Mineralogy
and
Petrology
150,
589–607.
Giuliani,
G.,
Fallick,
A.,
Rakotondrazafy,
M.,
Ohnenstetter,
D.,
Andriamamonjy,
A.,
Ralantoarison,
T.,
Rakotosamizanany,
S.,
Razanatseheno,
M.,
Offant,
Y.,
Garnier,
V.,
Dunaigre,
C.,
Schwarz,
D.,
Mercier,
A.,
Ratrimo,
V.,
Ralison,
B.,
2007.
Oxygen
isotope
systematics
of
gem
corundum
deposits
in
Madagascar:
relevance
for
their
geological
origin.
Mineralium
Deposita
42,
251–270.
Grove,
T.L.,
Baker,
M.B.,
1984.
Phase
equilibrium
controls
on
the
tholeiitic
versus
cal-
calkaline
differentiation
trends.
Journal
of
Geophysical
Research
89,
3253–3274.
Hamada,
M.,
Fujii,
T.,
2008.
Experimental
constraints
on
the
effects
of
pressure
and
H2O
on
the
fractional
crystallization
of
high-Mg
island
arc
basalt.
Contributions
to
Mineralogy
and
Petrology
155,
767–790.
Henderson,
P.,
Fishlock,
S.J.,
Laul,
J.C.,
Cooper,
T.D.,
Conard,
R.L.,
Boynton,
W.V.,
Schmitt,
R.A.,
1976.
Rare
earth
element
abundances
in
rocks
and
minerals
from
the
Fiskenaesset
Complex,
West
Greenland.
Earth
and
Planetary
Science
Letters
30,
37–49.

Author's personal copy
628 J.
Berger
et
al.
/
Precambrian
Research
224 (2013) 612–
628
Hoatson,
D.M.,
Sun,
S.-S.,
2002.
Archean
layered
mafic–ultramafic
intrusions
in
the
West
Pilbara
Craton,
Western
Australia:
a
synthesis
of
some
of
the
old-
est
orthomagmatic
mineralizing
systems
in
the
world.
Economic
Geology
97,
847–872.
Hoffmann,
J.E.,
Münker,
C.,
Næraa,
T.,
Rosing,
M.T.,
Herwartz,
D.,
Garbe-Schönberg,
D.,
Svahnberg,
H.,
2011.
Mechanisms
of
Archean
crust
formation
by
high
precision
HFSE
systematics
on
TTGs.
Geochimica
et
Cosmochimica
Acta
75,
4157–4178.
Hoffmann,
J.E.,
Svahnberg,
H.,
Piazolo,
S.,
Scherstén,
A.,
Münker,
C.,
2012.
The
geodynamic
evolution
of
Mesoarchean
anorthosite
complexes
inferred
from
the
Naajat
Kuuat
Complex,
southern
West
Greenland.
Precambrian
Research
196–197,
149–170.
Holland,
T.,
Blundy,
J.,
1994.
Nonideal
Interactions
in
calcic
amphiboles
and
their
bearing
on
amphibole–plagioclase
thermometry.
Contributions
to
Mineralogy
and
Petrology
116,
433–447.
Holland,
T.J.B.,
Powell,
R.,
1996.
Thermodynamics
of
order-disorder
in
minerals
2:
symmetric
formalism
applied
to
solid
solutions.
American
Mineralogist
81,
1425–1437.
Holland,
T.J.B.,
Powell,
R.,
1998.
An
internally
consistent
thermodynamic
data
set
for
phases
of
petrological
interest.
Journal
of
Metamorphic
Geology
16,
309–343.
Holland,
T.J.B.,
Powell,
R.,
2003.
Activity–composition
relations
for
phases
in
petro-
logical
calculations:
an
asymmetric
multicomponent
formulation.
Contributions
to
Mineralogy
and
Petrology
145,
492–501.
Hor,
A.K.,
Hutt,
D.K.,
Smith,
J.V.,
Wakefield,
J.,
Windley,
B.F.,
1975.
Petrochemistry
and
mineralogy
of
early
Precambrian
anorthositic
rocks
of
the
Limpopo
belt,
southern
Africa.
Lithos
8,
297–310.
Irvine,
T.N.,
1977.
Origin
of
chromite
layers
in
the
Muskox
intrusion
and
other
strat-
iform
intrusions:
a
new
interpretation.
Geology
5,
273–277.
Kamenetsky,
V.S.,
Crawford,
A.J.,
Meffre,
S.,
2001.
Factors
controlling
chemistry
of
magmatic
spinel:
an
empirical
study
of
associated
olivine,
Cr-spinel
and
melt
inclusions
from
primitive
rocks.
Journal
of
Petrology
42,
655–671.
Key,
R.M.,
Loughlin,
S.C.,
Gillespie,
M.,
Del
Rio,
M.,
Horstwood,
M.S.A.,
Crowley,
Q.G.,
Darbyshire,
D.P.F.,
Pitfield,
P.E.J.,
Henney,
P.J.,
2008.
Two
Mesoarchaean
terranes
in
the
Reguibat
shield
of
NW
Mauritania.
Geological
Society
of
London,
Special
Publications
297,
33–52.
Leake,
B.E.,
Woolley,
A.R.,
Arps,
C.E.S.,
Birch,
W.D.,
Gilbert,
M.C.,
Grice,
J.D.,
Hawthorne,
F.C.,
Kato,
A.,
Kisch,
H.J.,
Krivovichev,
V.G.,
Linthout,
K.,
Laird,
J.,
Man-
darino,
J.,
1997.
Nomenclature
of
amphiboles:
Report
of
the
Subcommittee
on
Amphiboles
of
the
International
Mineralogical
Association
Commission
on
New
Minerals
and
Mineral
Names.
Mineralogical
Magazine
61,
295–321.
Leblanc,
M.,
1991.
Platinum-group
elements
and
gold
in
ophiolitic
complexes:
dis-
tribution
and
fractionation
from
mantle
to
oceanic
floor.
In:
Tj
Peters,
et
al.
(Eds.),
Ophiolite
Genesis
and
Evolution
of
the
Oceanic
Lithosphere,
Oman.
Kluwer,
Dordrecht,
pp.
231–260.
Martin,
H.,
Smithies,
R.H.,
Rapp,
R.,
Moyen,
J.F.,
Champion,
D.,
2005.
An
overview
of
adakite,
tonalite–trondhjemite–granodiorite
(TTG),
and
sanukitoid:
relation-
ships
and
some
implications
for
crustal
evolution.
Lithos
79,
1–24.
Maurel,
C.,
Maurel,
P.,
1983.
Étude
expérimentale
de
l.équilibre
Fe2+–Fe3+ dans
les
spinelles
chromifères
et
les
liquides
silicatés
basiques
coexistants,
à
1
atm.
Comptes
Rendus
de
l’Académie
des
Sciences
295,
209–212.
Mouri,
H.,
Whitehouse,
M.J.,
Brandl,
G.,
Rajesh,
H.M.,
2009.
A
magmatic
age
and
four
successive
metamorphic
events
recorded
in
zircons
from
a
single
meta-
anorthosite
sample
in
the
Central
Zone
of
the
Limpopo
Belt,
South
Africa.
Journal
of
the
Geological
Society
166,
827–830.
Mukherjee,
R.,
Mondal,
S.,
Rosing,
M.,
Frei,
R.,
2010.
Compositional
variations
in
the
Mesoarchean
chromites
of
the
Nuggihalli
schist
belt,
Western
Dharwar
Craton
(India):
potential
parental
melts
and
implications
for
tectonic
setting.
Contrib-
utions
to
Mineralogy
and
Petrology
160,
865–885.
Muntener,
O.,
Kelemen,
P.B.,
Grove,
T.L.,
2001.
The
role
of
H2O
during
crystallization
of
primitive
arc
magmas
under
uppermost
mantle
conditions
and
genesis
of
igneous
pyroxenites:
an
experimental
study.
Contributions
to
Mineralogy
and
Petrology
141,
643–658.
Myers,
J.S.,
1976.
Channel
deposits
of
peridotite,
gabbro
and
chromitite
from
tur-
bidity
currents
in
the
stratiform
Fiskenæsset
anorthosite
complex,
southwest
Greenland.
Lithos
9,
281–291.
Naldrett,
A.J.,
Duke,
J.M.,
1980.
Platinum
metals
in
magmatic
sulfide
ores.
Science
208,
1417–1428.
Naldrett,
A.,
Wilson,
A.,
Kinnaird,
J.,
Yudovskaya,
M.,
Chunnett,
G.,
2012.
The
origin
of
chromitites
and
related
PGE
mineralization
in
the
Bushveld
Complex:
new
mineralogical
and
petrological
constraints.
Mineralium
Deposita
47,
209–232.
Owens,
B.E.,
Dymek,
R.F.,
1997.
Comparative
petrology
of
Archaean
anorthosites
in
amphibolite
and
granulite
facies
terranes,
SW
Greenland.
Contributions
to
Mineralogy
and
Petrology
128,
371–384.
Ozawa,
K.,
1983.
Evaluation
of
olivine–spinel
geothermometry
as
an
indicator
of
thermal
history
for
peridotites.
Contributions
to
Mineralogy
and
Petrology
82,
52–65.
Pagé,
P.,
Barnes,
S.-J.,
Bédard,
J.H.,
Zientek,
M.L.,
2012.
In
situ
determination
of
Os,
Ir,
and
Ru
in
chromites
formed
from
komatiite,
tholeiite
and
boninite
magmas:
implications
for
chromite
control
of
Os,
Ir
and
Ru
during
partial
melting
and
crystal
fractionation.
Chemical
Geology
302–303,
3–15.
Phinney,
W.C.,
Morrison,
D.A.,
Maczuga,
D.E.,
1988.
Anorthosites
and
related
megacrystic
units
in
the
evolution
of
Archean
crust.
Journal
of
Petrology
29,
1283–1323.
Polat,
A.,
Appel,
P.W.U.,
Fryer,
B.,
Windley,
B.,
Frei,
R.,
Samson,
I.M.,
Huang,
H.,
2009.
Trace
element
systematics
of
the
Neoarchean
Fiskenæsset
anorthosite
complex
and
associated
meta-volcanic
rocks,
SW
Greenland:
evidence
for
a
magmatic
arc
origin.
Precambrian
Research
175,
87–115.
Polat,
A.,
Frei,
R.,
Scherstén,
A.,
Appel,
P.W.U.,
2010.
New
age
(2970
Ma),
mantle
sources
and
geodynamic
constraints
on
the
Archean
Fiskenæsset
anorthosite
complex.
Chemical
Geology
277,
1–20.
Polat,
A.,
Fryer,
B.,
Appel,
P.W.U.,
Kalvig,
P.,
Kerrich,
R.,
Dilek,
Y.,
Yang,
Z.,
2011.
Geochemistry
of
anorthositic
differentiated
sills
in
the
Archean
(∼2970
Ma)
Fiskenæsset
Complex,
SW
Greenland:
implications
for
parental
magma
com-
positions,
geodynamic
setting,
and
secular
heat
flow
in
arcs.
Lithos
123,
50–72.
Polat,
A.,
Fryer,
B.J.,
Samson,
I.M.,
Weisener,
C.,
Appel,
P.W.U.,
Frei,
R.,
Windley,
B.F.,
2012.
Geochemistry
of
ultramafic
rocks
and
hornblendite
veins
in
the
Fiskenæs-
set
layered
anorthosite
complex,
SW
Greenland:
evidence
for
hydrous
upper
mantle
in
the
Archean.
Precambrian
Research
214–215,
124–153.
Potrel,
A.,
1994.
Evolution
tectono-métamorphique
d‘un
segment
de
croüte
con-
tinentale
archéenne.
Exemple
de
1‘Amsaga
(R.I.
Mauritanie),
Dorsale
Réguibat
(craton
Quest
Africain).
Unpublished
PhD
Thesis,
Rennes,
France,
400pp.
Potrel,
A.,
Peucat,
J.J.,
Fanning,
C.M.,
Auvray,
B.,
Burg,
J.P.,
Caruba,
C.,
1996.
3.5
Ga
old
terranes
in
the
West
African
Craton,
Mauritania.
Journal
of
the
Geological
Society
153,
507–510.
Potrel,
A.,
Peucat,
J.J.,
Fanning,
C.M.,
1998.
Archean
crustal
evolution
of
the
west
African
craton:
example
of
the
Amsaga
area
(Reguibat
rise).
U–Pb
and
Sm–Nd
evidence
for
crustal
growth
and
recycling.
Precambrian
Research
90,
107–117.
Prichard,
H.M.,
Neary,
C.R.,
Fisher,
P.C.,
O’Hara,
M.J.,
2008.
PGE-rich
podiform
chromi-
tites
in
the
Al’Ays
Ophiolite
Complex,
Saudi
Arabia:
an
example
of
critical
mantle
melting
to
extract
and
concentrate
PGE.
Economic
Geology
103,
1507–1529.
Raith,
M.M.,
Rakotondrazafy,
R.,
Sengupta,
P.,
2008.
Petrology
of
corundum–spinel–sapphirine–anorthite
rocks
(sakenites)
from
the
type
locality
in
southern
Madagascar.
Journal
of
Metamorphic
Geology
26,
647–667.
Rollinson,
H.R.,
Appel,
P.W.U.,
Frei,
R.,
2002.
A
metamorphosed,
early
Archaean
chromitite
from
west
Greenland:
implications
for
the
genesis
of
Archaean
anorthositic
chromitites.
Journal
of
Petrology
43,
2143–2170.
Rollinson,
H.,
Reid,
C.,
Windley,
B.,
2010.
Chromitites
from
the
Fiskenæsset
anorthositic
complex,
West
Greenland:
clues
to
late
Archaean
mantle
processes.
Geological
Society
of
London,
Special
Publications
338,
197–212.
Sattari,
P.,
Brenan,
J.M.,
Horn,
I.,
McDonough,
W.F.,
2002.
Experimental
constraints
on
the
sulfide-
and
chromite-silicate
melt
partitioning
behavior
of
rhenium
and
platinum-group
elements.
Economic
Geology
97,
385–398.
Sisson,
T.W.,
Grove,
T.L.,
1993.
Experimental
investigations
of
the
role
of
H2O
in
calcalkaline
differentiation
and
subduction
zone
magmatism.
Contributions
to
Mineralogy
and
Petrology
113,
143–166.
Takagi,
D.,
Sato,
H.,
Nakagawa,
M.,
2005.
Experimental
study
of
a
low-alkali
tholeiite
at
1–5
kbar:
optimal
condition
for
the
crystallization
of
high-An
plagioclase
in
hydrous
arc
tholeiite.
Contributions
to
Mineralogy
and
Petrology
149,
527–540.
Tenthorey,
E.A.,
Ryan,
J.G.,
Snow,
E.A.,
1996.
Petrogenesis
of
sapphirine-bearing
metatroctolites
from
the
Buck
Creek
ultramafic
body,
southern
Appalachians.
Journal
of
Metamorphic
Geology
14,
103–114.
Villiger,
S.,
Ulmer,
P.,
Muntener,
O.,
Thompson,
A.B.,
2004.
The
liquid
line
of
descent
of
anhydrous,
mantle-derived,
tholeiitic
liquids
by
fractional
and
equilibrium
crystallization
–
an
experimental
study
at
1
center
dot
0
GPa.
Journal
of
Petrology
45,
2369–2388.
Villiger,
S.,
Ulmer,
P.,
Muntener,
O.,
2007.
Equilibrium
and
fractional
crystalliza-
tion
experiments
at
0.7
GPa;
the
effect
of
pressure
on
phase
relations
and
liquid
compositions
of
tholeiitic
magmas.
Journal
of
Petrology
48,
159–184.
Watkinson,
D.H.,
Ohnenstetter,
D.,
1992.
Hydrothermal
origin
of
platinum-group
mineralization
in
the
Two
Duck
Lake
intrusion,
Coldwell
Complex,
northwestern
Ontario.
Canadian
Mineralogist
30,
121–136.
Weaver,
B.L.,
Tarney,
J.,
Windley,
B.,
1981.
Geochemistry
and
petrogenesis
of
the
Fiskenaesset
anorthosite
complex,
southern
West
Greenland:
nature
of
the
par-
ent
magma.
Geochimica
Et
Cosmochimica
Acta
45,
711–725.
Windley,
B.F.,
Bishop,
F.C.,
Smith,
J.S.,
1981.
Metamorphosed
layered
igneous
com-
plexes
in
Archean
Granulite-Gneiss
Belts.
Annual
Reviews
in
Earth
Planetary
Science
9,
175–198.
Windley,
B.F.,
Garde,
A.A.,
2009.
Arc-generated
blocks
with
crustal
sections
in
the
North
Atlantic
craton
of
West
Greenland:
crustal
growth
in
the
Archean
with
modern
analogues.
Earth
Science
Reviews
93,
1–30.