Origin Of Scapolite-Hosted Sapphire (Corundum) Near Kimmirut, Baffin Island, Nunavut, Canada

Abstract

Gem-quality corundum (sapphire) occurs in scapolite-rich calc-silicate rock hosted in marble of the Lake Harbour Group near Kimmirut, southern Baffin Island. A deposit of blue and colorless gem corundum (Beluga occurrence) is compared to a similar calc-silicate pod generally lacking corundum but containing nepheline (Bowhead occurrence) and located 170 m to the SSW. Corundum formation was made possible by three equally important sequential metamorphic reactions: (1) formation of nepheline, diopside, and K-feldspar (inferred) at granulite facies peak metamorphic conditions; (2) partial retrograde replacement of the peak assemblage by phlogopite, oligoclase, calcite, and scapolite (Me-50-Me-67) as a result of CO2-, H2O-, Cl-, F-bearing fluid influx at 1782.5 +/- 3.7 Ma (P-T < 720 degrees C, 6.2 kbar); and (3) retrograde breakdown of scapolite + nepheline (with CO2-and H2O-bearing fluid) to form albite, muscovite, corundum, and calcite. Late, low-temperature zeolite mineralization is common in corundum-bearing zones. Based on thermodynamic models, the corundum-forming reaction only occurs in a <100 degrees C window with an upper limit determined by scapolite-nepheline stability, and a lower limit determined by the formation of Al-silicate rather than corundum. The protolith is inferred to be dolomitic argillaceous marl with no evidence to suggest the initial presence of evaporites. The enrichment of trace metals V and Cr, and the depletion of Co, Ni, and Mn, suggest reducing diagenetic conditions in the initial sediment. Beluga calc-silicate rock is strongly depleted in REE (Total REE similar to 18 ppm). Oxy-dravite delta B-11 (+3.9 +/- 0.7%) is consistent with a marine boron source. The oxygen isotope composition of corundum (delta O-18(VSMOW) = 16.4 +/- 0.1%) is comparable to that of corundum in marble or desilicated pegmatite associated with marble. Phlogopite and muscovite Ar-40/(39) Ar ages and calculated closure temperatures (considered estimates) are ca. 1640 Ma (T-c = 455 to 515 degrees C) and 1510 Ma (Tc = 410 to 425 degrees C), respectively. In the Lake Harbour Group, the most prospective areas for gem corundum exploration are expected to be contiguous to the thrust fault separating the Lake Harbour Group and Narsajuaq terranes, where the retrograde, amphibolite facies overprint of the granulite peak assemblages was most pervasive.

Full text

The Canadian Mineralogist
Vol. 55, pp. 669-699 (2017)
DOI: 10.3749/canmin.1700018
ORIGIN OF SCAPOLITE-HOSTED SAPPHIRE (CORUNDUM) NEAR KIMMIRUT, BAFFIN
ISLAND, NUNAVUT, CANADA
PHILIPPE M. BELLEY
§
,TASHIA J. DZIKOWSKI
*
,ANDREW FAGAN
†
,JAN CEMP´
IREK
‡
,LEE A. GROAT, AND
JAMES K. MORTENSEN
Department of Earth, Ocean and Atmospheric Sciences, University of British Columbia, Vancouver, British Columbia V6T 1Z4,
Canada
MOSTAFA FAYEK
Department of Geological Sciences, University of Manitoba, Winnipeg, Manitoba R3T 2N2, Canada
GASTON GIULIANI
Universit´
e de Lorraine, IRD and CRPG UMR 7358 CNRS-UL, BP 20, 15 rue Notre-Dame-des-Pauvres,
54501 Vandœuvre-l`
es-Nancy, France
ANTHONY E. FALLICK
Isotope Geosciences Unit, S.U.E.R.C., Rankine Avenue, East Kilbride, Glasgow G75 0QF, Scotland
PAUL GERTZBEIN
¶
Indigenous and Northern Affairs Canada, Iqaluit, Nunavut X0A 0H0, Canada
ABSTRACT
Gem-quality corundum (sapphire) occurs in scapolite-rich calc-silicate rock hosted in marble of the Lake Harbour Group
near Kimmirut, southern Baffin Island. A deposit of blue and colorless gem corundum (Beluga occurrence) is compared to a
similar calc-silicate pod generally lacking corundum but containing nepheline (Bowhead occurrence) and located 170 m to
the SSW. Corundum formation was made possible by three equally important sequential metamorphic reactions: (1)
formation of nepheline, diopside, and K-feldspar (inferred) at granulite facies peak metamorphic conditions; (2) partial
retrograde replacement of the peak assemblage by phlogopite, oligoclase, calcite, and scapolite (Me
50
–Me
67
) as a result of
CO
2
-, H
2
O-, Cl-, F-bearing fluid influx at 1782.5 63.7 Ma (P-T ,720 8C, 6.2 kbar); and (3) retrograde breakdown of
scapolite þnepheline (with CO
2
- and H
2
O-bearing fluid) to form albite, muscovite, corundum, and calcite. Late, low-
temperature zeolite mineralization is common in corundum-bearing zones. Based on thermodynamic models, the corundum-
forming reaction only occurs in a ,100 8C window with an upper limit determined by scapolite-nepheline stability, and a
lower limit determined by the formation of Al-silicate rather than corundum. The protolith is inferred to be dolomitic
argillaceous marl with no evidence to suggest the initial presence of evaporites. The enrichment of trace metals V and Cr, and
the depletion of Co, Ni, and Mn, suggest reducing diagenetic conditions in the initial sediment. Beluga calc-silicate rock is
strongly depleted in REE (Total REE ~18 ppm). Oxy-dravite d
11
B(þ3.9 60.7%) is consistent with a marine boron source.
The oxygen isotope composition of corundum (d
18
O
VSMOW
¼16.4 60.1%) is comparable to that of corundum in marble or
desilicated pegmatite associated with marble. Phlogopite and muscovite
40
Ar/
39
Ar ages and calculated closure temperatures
(considered estimates) are ca.1640Ma(T
c
¼455 to 515 8C) and 1510 Ma (T
c
¼410to4258C), respectively. In the Lake
Harbour Group, the most prospective areas for gem corundum exploration are expected to be contiguous to the thrust fault
§
Corresponding author e-mail address: phil.belley@gmail.com
*
Current address: Swiss Gemmological Institute SSEF, Aeschengraben 26, 4051 Basel, Switzerland
†
Current address: True North Gems Inc., P.O. Box 11108, Vancouver, British Columbia, V6E 3P3, Canada
‡
Current address: Dept. of Geological Sciences, Masaryk University, Kotla
´ˇ
rska
´2, Brno, Czech Republic
¶
Deceased
669

separating the Lake Harbour Group and Narsajuaq terranes, where the retrograde, amphibolite facies overprint of the
granulite peak assemblages was most pervasive.
Keywords: sapphire, corundum, marble, calc-silicate rock, scapolite, gem deposits, metamarl, nepheline,
proterozoic.
INTRODUCTION
Gem-quality blue corundum (sapphire) accounts for
a significant portion of the gemstone market, and prices
continue to rise in response to demand (Shor & Weldon
2009, Genis 2016). For this reason, there is considerable
interest in understanding the genesis of gem-quality
corundum and constraining the types of environments in
which it forms. Improved exploration methodologies
based on sapphire genetic models will aid in the
development of a Canadian colored gemstone industry
that will be competitive on the world market and could
be applied in exploration efforts worldwide.
Gem corundum deposits have been found in syenite,
monzonite, kimberlite, lamprophyre, basalt (xeno-
crysts), gneiss, amphibolite, marble, skarn, and various
contact metasomatic rocks (Simonet et al. 2008,
Dzikowski et al. 2014, Giuliani et al. 2014). Gem-
quality blue and colorless corundum was discovered in
2002 near Kimmirut, southern Baffin Island, Nunavut,
Canada, and subsequent exploration led to the discovery
of blue, colorless, yellow, and pink gem corundum
showings (True North Gems 2007). These occurrences
are the first reported examples of gem corundum hosted
in scapolite-rich calc-silicate pods in marble. The
present study examines two calc-silicate pods, the
Beluga deposit, which is blue-corundum-bearing, and
the Bowhead occurrence, a nearby calc-silicate pod in
which corundum is rare, with the objective of
determining the cause and timing of mineralization
and the nature of the protolith.
REGIONAL GEOLOGY
The Beluga and Bowhead calc-silicate pods are
hosted by a sequence of marble and calc-silicate rock
structurally overlying garnet psammite, meta-semi-
pelite, metapelite, and rare orthoquartzite. The marble,
calc-silicate rock, and metaclastic rocks form the Lake
Harbour Group (LHG) supracrustal suite (Fig. 1; Scott
1997, Scott et al. 2002). The LHG is interpreted to be
the sedimentary cover sequence of the Meta Incognita
microcontinent, which was accreted to the southern
margin of the Rae Craton by the Trans-Hudson
orogeny (THO; St-Onge et al. 2009). Detrital zircon
ages indicate a Paleoproterozoic sediment source
(rarely Archean) for LHG sediments and suggest the
depositional age of the LHG to be later than ca. 1.93
Ga, the youngest detrital zircon age (Scott et al. 2002).
The supracrustal rocks of the THO, including the
LHG, were subject to multiple deformational and
metamorphic events (see St-Onge et al. 2007). In
deformation event D
0
, which post-dates the deposition
of the LHG and pre-dates the Cumberland batholith
emplacement, basement rocks were imbricated with
cover rocks. The Cumberland batholith, an Andean-type
granitic batholith, was emplaced from 1865 þ4/–2 Ma to
1848 62 Ma. The Narsajuaq oceanic arc was accreted
to Meta Incognita (deformation event D
1A
), and the
suture closed between 1845 62 Ma and 1842 þ5/–3
Ma. Prograde granulite facies metamorphism (M
1A
)
occurred ca. 1849–1835 Ma as a result of crustal
thickening (accretion of the Narsajuaq arc) and heat
advection (emplacement of the Cumberland batholith).
A high-temperature thermal perturbation (M
1B
), ca.
1833–1829 Ma, represents either continued mineral
growth in the late stages of M
1A
or a distinct thermal
event related to felsic intrusions. Retrograde upper-
amphibolite facies metamorphism (M
2
) occurred from
1820 61 Ma to 1813 62Mawithreactivationand
further shortening of the Soper River suture (D
2
), which
separates Meta Incognita from the Narsajuaq terrane.
Recrystallization to S
2
assemblages was noted to be
most significant in samples with strong foliation and
occurring contiguous to D
2
thrusts. This localized
recrystallization is interpreted by St-Onge et al. (2007)
as being controlled by deformation-enhanced fluid
circulation. Lastly, post-D
2
thermal and fluid activity
occurred ca. 1797–1785 Ma, possibly related to late
felsic intrusive rocks of similar age and speculated to be
a potential cause of minor greenschist facies retrogres-
sion (secondary chlorite, epidote, sericitization).
Pressure and temperature determinations by St-
Onge et al. (2007) are as follows: S
1A
mineral
assemblages provided P-T values of ca. 810 8C and
8.0 kbar; thermobarometry of S
2
mineral assemblages
indicate metamorphic conditions of upper amphibolite
facies, ca. 720 8C and 6.2 kbar.
LAKE HARBOUR GROUP MARBLES AND
CALC-SILICATE ROCKS
Marbles and calc-silicate schists showing centime-
ter- to meter-scale compositional layering comprise a
670 THE CANADIAN MINERALOGIST

FIG. 1. Location of the Beluga and Bowhead calc-silicate pods, True North Gems property, near Kimmirut, Nunavut. Modified
after St-Onge et al. (2001).
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 671

significant portion of the Lake Harbour Group (Scott
1997). LHG metacarbonate and calc-silicate rocks
commonly contain calcite, diopside, phlogopite, for-
sterite, amphibole, spinel, graphite, or scapolite with
the occasional occurrence of apatite, titanite, wollas-
tonite, humite, or plagioclase (Scott 1997, Herd et al.
2000, St-Onge et al. 2000). Nepheline and K-feldspar
occur in calc-silicate bands and marble on Aliguq
island, 165 km NW of Kimmirut (Butler 2007).
Layering in the marbles is generally convoluted,
probably due to ductile flow (Scott 1997), and
silicate-rich lithologies occur as both stratiform
continuous layers and boudins or pods (Butler 2007).
Marble in the vicinity of the Beluga and Bowhead
calc-silicate pods is medium to very coarse-grained
(3–30 mm) and commonly contains phlogopite, green
or purplish diopside, graphite, scapolite, apatite, and
plagioclase (Gertzbein 2005).
PREVIOUS STUDIES
LeCheminant et al. (2005) suggested that formation
of the corundum-bearing calc-silicate rocks at the
Beluga and Bowhead occurrences involved a contam-
inated syenitic magma emplaced in marble late in the
D
2
deformation event. The nepheline- and diopside-
rich assemblage was proposed to have been formed by
the reaction of syenitic magma with LHG marble and
possibly evaporites, followed by retrogression forming
phlogopite-oligoclase-scapolite at the expense of
diopside and nepheline, and a second retrograde event
resulting in the fracture-controlled alteration of
nepheline to form corundum. In a study of nephe-
line-bearing LHG metacarbonates on Aliguq Island,
Butler (2007) determined that nepheline-K-feldspar
formed late based on textural relations with forsterite,
diopside, and phlogopite. Butler (2007) proposed that
this assemblage formed at peak metamorphic condi-
tions (ca. 800 8C and 8 kbar) at low X
CO2
(0.15) by
the reaction 3 albite þ3 calcite þphlogopite K-
feldspar þ3 nepheline þ3 diopside þH
2
Oþ3CO
2
.
Dzikowski (2013) suggested that the protolith for the
Beluga and Bowhead calc-silicate pods was evaporite-
black shale deposited in the LHG carbonate shelf and
proposed the following paragenetic sequence: (1)
prograde diopside þnepheline (810 8C, 8.3 kbar);
(2) alteration of the peak assemblage by NaCl-bearing
fluids (,710 8C, 6 kbar) forming phlogopite-oligo-
clase ‘symplectites’ with late scapolite rims; (3)
hydrous fluids, post-D
2
, causing the breakdown of
nepheline and either scapolite or anorthite to form
albite, muscovite, and corundum or alternatively, the
introduction of Na-bearing hydrous fluids breaking
down anorthite to albite and corundum.
Lepage & Davison (2007) reported the use of
ultraviolet LED technology to explore for fluorescent
scapolite, which they assumed would coexist with gem
corundum in the Kimmirut area. Turner et al. (2014)
investigated the possible application of hyperspectral
imaging in exploration for Kimmirut-type gem
corundum deposits.
In this study, we re-evaluate the data presented by
Dzikowski (2013) in light of additional isotopic and
petrographic data, and apply the results to refine
interpretations on the origin and petrogenesis of gem
corundum in Kimmirut-type deposits.
RECENT EXPLORATION
The area around the Beluga and Bowhead calc-
silicate pods was extensively explored by True North
Gems, Inc. (TNG) from 2002 to 2009. Ultraviolet light
prospecting for fluorescent scapolite assisted in the
discovery of 8031 scapolite showings in and around
the Beluga Project property, in addition to 45 named
spinel outcrops, and 44 named corundum localities
(some of which consist of multiple contiguous
showings). The distribution of corundum, scapolite,
and spinel showings is plotted on a geological map of
the TNG property in Figure 2. Several corundum
occurrences produced notable gems as part of the
exploration and deposit assessment work: the Beluga
South occurrence produced 34 yellow sapphire
gemstones totaling 6.98 carats, notably including two
yellow stones weighing 1.47 ct and 1.09 ct (Fig. 3A),
in addition to colorless and pale blue stones. The
Aqpik occurrence produced two virtually flawless
colorless stones, 2.50 ct and 2.59 ct in weight, and a
light blue, lightly included 7.81 ct gemstone (Fig. 3B).
Some of the Aqpik rough sapphire turned blue as a
result of heat treatment. One heat-treated, rich blue
cushion-cut stone weighs 2.43 ct (Fig. 3A). The
Beluga occurrence is the most important gem
corundum occurrence found in Nunavut to date,
containing grades of gem sapphire rough between 33
g/t (4.29 t bulk sample) and 19 g/t (22.5 t bulk sample).
Colorless, light blue, and more commonly saturated to
dark blue corundum occurs as crystals up to ca.7cm
long. One of the more notable stones in the bulk
sample is a 1.17 ct, deep blue, extra fine sapphire
gemstone (Fig. 3A and 3C) showing even color when
viewed through the table facet, but which in fact is
primarily colorless with a central, dark blue patch
(Wilson 2014). Lastly, pinkish corundum was found at
one location.
In total, 2607 polished corundum gemstones,
totaling 169.95 ct, were cut from gem rough recovered
by regional and bulk sampling. The great majority of
these gemstones originated from the Beluga deposit.
672 THE CANADIAN MINERALOGIST

FIG. 2. Bedrock geology map of the True North Gems property with markers for scapolite, spinel, and corundum occurrences
found during the 2006 mapping season. The most important mineralized areas are named. Small plutons consist of granite or
ultramafic plugs. Beluga is located 2.5 km SSW of Kimmirut airport. UTM zone 19 V (NAD83). Map courtesy of True
North Gems Inc.
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 673

ANALYTICAL METHODS
Chemical analysis
Chemical analyses for all minerals, with the
exception of tourmaline, were performed with a
CAMECA SX-50 electron microprobe (at the Univer-
sity of British Columbia) in wavelength-dispersion
(WD) mode. The operating voltage was 15 kV, with 20
nA beam current and 5 lm beam diameter. Counts
were collected for 20 s for each element with the
exception of F and Cl (50 s) and La, Ce, Pr, Nd, Sm,
Gd, Th, and U in zircon (60 s). The following
standards were used (all Kalines): synthetic phlogo-
pite (F), albite (Na), synthetic phlogopite (Mg and Si
in phlogopite, K in micas), diopside (Mg, minerals
other than phlogopite), kyanite (Al in diopside and
FIG. 3. (A) Corundum (sapphire) gemstones from the Kimmirut occurrences. Left: Colorless sapphire, Aqpik occurrence, 2.50
and 2.59 ct. Center: Deep blue, extra fine sapphire, 1.17 ct, from the Beluga occurrence, and a heat-treated, rich blue 2.43 ct
sapphire from Aqpik. Right: Yellow sapphire, Beluga South occurrence, 1.09 ct and 1.47 ct. Photograph courtesy of True
North Gems Inc. (B) Lightly included, light blue sapphire (7.81 ct) from the Aqpik occurrence. Photograph by Brad Wilson.
(C) Dark blue corundum crystal, 36 34 mm, in calc-silicate rock, and a 1.17 ct sapphire gemstone from the Beluga
occurrence. Photograph by Brad Wilson, courtesy of True North Gems Inc.
674 THE CANADIAN MINERALOGIST

micas), corundum (Al in corundum), anorthite (Al in
other minerals), zircon (Si in zircon), baryte (S),
scapolite (Cl), orthoclase (K in micas, Si in muscovite,
nepheline, muscovite, and scapolite), rutile (Ti),
synthetic V (V), synthetic magnesiochromite (Cr),
synthetic rhodonite (Mn), and synthetic fayalite (Fe).
The following additional standards were used for
zircon (all Lalines except UMa): Y
3
Al
5
O
12
(Y),
zircon (Zr), Ca-Al-Si-glass (Drake & Weill 1972; La,
Pr, Nd, Sm, Gd), CeO
2
(Ce), ThO
2
glass (Th), and
UO
2
glass (U). Matrix correction calculations were
done using the ‘PAP’ /(qZ) method (Pouchou &
Pichoir 1985).
Chemical analysis of tourmaline was performed
with a CAMECA SX100 instrument at Masaryk
University, Brno, Czech Republic. The operating
voltage was 15 kV with 10 nA beam current and 5
lm beam diameter. Counts were collected for 10 s
with the following exceptions: 15 s (V), 20 s (Mg, Cr,
Ca, Zn), 30 s (Cl), 40 s (K), and 60 s (F). The
following standards were used (all Kalines): topaz (F),
albite (Na), pyrope (Mg), sanidine (Al, Si, K),
fluorapatite (P), vanadinite (Cl), wollastonite (Ca),
titanite (Ti), scandium vanadate (V), chromite (Cr),
spessartine (Mn), almandine (Fe), and gahnite (Zn).
B
2
O
3
concentrations of 10.63 wt.% were estimated for
matrix correction of tourmaline data. Matrix correction
was done using the X-PHI method (Merlet 1994).
Boron isotopes
The boron isotope composition of tourmaline was
determined at the University of Manitoba with a
Cameca IMS 7f ion microprobe using secondary ion
mass spectroscopy (SIMS), a primary O-beam (5 nA
accelerated at 12.5 kV), with a 15 lm beam diameter,
a sample accelerating voltage þ10 kV, electrostatic
analyzer þ10 kV, and ETP 133H electron multiplier
coupled with an ion counting system having an overall
deadtime of 21 ns. The entrance slit was set at 36 lm
with a mass resolving power of 1450. Counts for
11
B,
10
B, and
30
Si were collected in succession for 50
cycles with 1 s measurements of each isotope per
cycle, 30 s pre-sputter, and 0 V offset. The analytical
procedure was similar to that used by Chaussidon &
Albar`
ede (1992). Instrumental mass fractionation
(IMF) and analytical quality were assessed by replicate
analyses of an elbaite tourmaline reference material
(No. 98114, see below); repeatability of the reference
material was 0.4%. Precision (1r) for the unknown
sample during the sessions was 60.3%. The boron
isotope composition is expressed in delta notation as a
per mil deviation from boric acid standard NIST RM
951 (
11
B/
10
B¼4.0437 60.0033, Catanzaro et al.
1970): d
11
B¼([
11
B/
10
B]
sample
/[
11
B/
10
B]
SRM951
–
1)*1000.
For the analysis of Beluga oxy-dravite, elbaite
standard No. 98114 (
11
B/
10
B¼4.0014 60.0007,
Leeman & Tonarini 2001) was used. Ludwig et al.
(2011) reported no significant matrix effects for B
isotope analysis of tourmaline (using dravite, elbaite,
and schorl standards) with SIMS. Cabral et al. (2012)
found a 1%to 2%IMF offset with elbaite,
comparable to overall uncertainty and therefore not
significant. A þ1.6%discrepancy in Leeman &
Tonarini (2001) standard elbaite No. 98114 relative
to dravite No. 108796 was deemed significant by
MacGregor et al. (2013).
Oxygen isotopes of corundum
The oxygen isotope composition (
18
O/
16
O) of
corundum was measured at the Isotope Geosciences
Unit, Scottish Universities Environmental Research
Centre, Glasgow, Scotland using the laser fluorination
method of Giuliani et al. (2005). Precision (1r)of
analyses on a quartz standard is 60.1%. Data is
reported in delta notation as a per mil deviation from
the
18
O/
16
O value of Vienna Standard Mean Ocean
Water (VSMOW) standard NIST RM 8535.
Whole-rock geochemistry
Whole rock major and trace elements were
analyzed by ALS Chemex in Vancouver, Canada
using a combination of inductively coupled plasma
mass spectrometry (ICP-MS) and atomic emission
spectroscopy (ICP-AES). Carbon was determined by
combustion furnace, and the rock sample was subject
to a lithium borate fusion for resistive elements, a four
acid digestion, and aqua regia digestion.
Radiometric dating
Zircon crystals were analyzed using conventional
ID-TIMS (isotope dilution thermal ionization mass
spectrometry) at the Pacific Center for Isotopic and
Geochemical Research (PCIGR), University of British
Columbia, with the methods described by Mortensen
et al. (1995) and Beranek & Mortensen (2011). Errors
attached to individual analyses were calculated using
the numerical error propagation method of Roddick
(1987), and decay constants are those recommended
by Steiger & J¨
ager (1975). Compositions for initial
common Pb were taken from the model of Stacey &
Kramer (1975). The zircon grains were strongly air
abraded prior to dissolution in order to try to minimize
the effects of post-crystallization Pb loss.
Muscovite and phlogopite were analyzed for
40
Ar/
39
Ar dating in the Noble Gas Laboratory of the
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 675

PCIGR. Mineral separates were hand-picked, washed
in acetone, dried, wrapped in aluminum foil, and
stacked in an irradiation capsule with a neutron flux
monitor (Fish Canyon Tuff sanidine, 28.02 Ma, Renne
et al. 1998) and 1071 Ma hornblende HB3Gr as an age
check (which yielded a flat J-curve and an age of 1069
62 Ma). The samples were irradiated at the McMaster
Nuclear Reactor in Hamilton, Ontario, for 90 MWH,
with a neutron flux of approximately 6 310
13
neutrons/cm
2
/s. Analyses (n¼45) of 15 neutron flux
monitor positions produced errors of ,0.5% in the J
value. At PCIGR, the mineral separates were step-
heated at incrementally higher powers in the defocused
beam of a 10 W CO
2
laser (New Wave Research
MIR10) until fused. The gas evolved from each step
was analyzed with a VG5400 mass spectrometer
equipped with an ion-counting electron multiplier.
All measurements were corrected for total system
blank, mass spectrometer sensitivity, mass discrimi-
nation, radioactive decay during and subsequent to
irradiation, atmospheric Ar contamination, and the
effect of irradiation on Ca, Cl, and K. The plateau and
correlation ages were calculated using Isoplot v3.09
(Ludwig 2003). Errors are quoted at the 2-sigma (95%
confidence) level and are propagated from all sources
except mass spectrometer sensitivity and age of the
flux monitor. The best statistically justified plateau and
plateau age were picked based on the following
criteria: (1) three or more contiguous steps comprising
more than 50% of the
39
Ar; (2) probability of fit of the
weighted mean age greater than 5%; (3) slope of the
error-weighted line through the plateau ages equals
zero at 5% confidence; (4) ages of the two outermost
steps on a plateau are not significantly different from
the weighted-mean plateau age (at 1.8r, six or more
steps only); and (5), the outermost two steps on either
side of a plateau must not have nonzero slopes with the
same sign (at 1.8r, nine or more steps only).
RESULTS
Outcrop descriptions
The Beluga (N 62.8287348, W 69.8940728; Fig. 4)
and Bowhead (170 m to the SSW) occurrences are
marble-hosted calc-silicate pods with surface expo-
sures of 4.2 33.7 m and 2 31.5 m, respectively.
Contacts with the host marble are sharp, with the
exception of parts of the Beluga pod, where euhedral
crystals of dark brownish-purple diopside (1–3 cm) or
yellowish-grey scapolite (1–5 cm) occur in very coarse
pale-orange calcite between the calc-silicate rock and
marble. The rock in the calc-silicate pods is beige with
abundant 1 to 6 cm spots of brown mottling. Beige
areas are primarily composed of scapolite, albite,
muscovite, calcite, and corundum at Beluga and
scapolite-nepheline-calcite at Bowhead. The brown
mottling is phlogopite-rich with oligoclase and
diopside. The relative abundance of the leucocratic
and melanocratic assemblages is variable: for exam-
ple, at Beluga, the melanocratic component ranges
between 10 and 90% on a decimeter scale but overall
represents roughly half of the total rock volume based
on visual estimation.
Petrography and mineral compositions
Beluga occurrence. The calc-silicate rock at
Beluga is coarse to very coarse-grained (5 mm to .5
FIG. 4. Contact between marble and sapphire-bearing calc-silicate rock at the Beluga occurrence. Note the variation in the
distribution of light-colored and dark mineral assemblages within the calc-silicate pod.
676 THE CANADIAN MINERALOGIST

cm) and commonly consists of randomly oriented
centimeter-scale crystals, occasionally with maximum
dimensions ca. 4–7 cm. The calc-silicate rock can be
divided into three distinct mineral assemblages (Fig.
5): (1) dark areas composed of phlogopite, oligoclase,
diopside, and subordinate calcite; (2) light-colored,
scapolite-rich zones, where scapolite is frequently
altered to a mixture of fine-grained silicate minerals,
generally separating the first assemblage from the
corundum-bearing assemblage; and (3) pods or zones
of light-colored corundum-albite-muscovite-calcite
assemblage containing small amounts of graphite and
pyrrhotite.
Dark brown, Fe-, F-, and Ti-bearing phlogopite
(Table 1), pale grey oligoclase (Ab
80
; Table 2), and
subordinate calcite occur as graphic oriented inter-
growths ranging in size from several millimeters to 2
cm across. Phlogopite-oligoclase symplectite up to 3
mm across is uncommon. The oriented intergrowths
form coronae around irregularly shaped grains of
purplish-brown Al-bearing, Si-poor diopside (Fig. 6)
from several millimeters to 3 cm in size (0.31 Al pfu,
0.12 Na pfu; Table 3) and are infrequently pseudo-
morphic after diopside. The replacement of diopside
by phlogopite-oligoclase is substantial (ca. 30–80%)
with considerable local variation, and small relict
diopside grains occur sparsely in the phlogopite-
oligoclase. Minor pale-orange calcite is present in
the coronae and is infrequently visible in hand sample.
Titanite (,0.5 mm), zircon (Table 4), and apatite (0.1
mm) are uncommon. The phlogopite-oligoclase inter-
growths are locally associated with 1–3 cm masses of
equigranular, medium-grained oligoclase with subor-
dinate phlogopite and calcite.
Anhedral, coarse grained to very coarse grained
crystals of light yellowish-grey to grey scapolite occur
FIG. 5. Corundum-bearing calc-silicate rock in situ at the Beluga pit. Phlogopite-oligoclase (Phl þPl) intergrowths occur near
partly albitized scapolite (Ab þScp) and albite, muscovite, and blue corundum (Ab þMs þCrn). The corundum- and
phlogopite-bearing assemblages contain calcite. Minor graphite (Cg) is present.
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 677

around the phlogopite-rich coronae. The scapolite
averages 57 mol.% meionite component with consid-
erable variation (Me
50
–Me
62
; Table 5) and minor
silvialite component (average 0.014 SO
4
pfu). Much of
the scapolite is pervasively replaced by milky white,
fine-grained mixtures (Fig. 7) of the following mineral
phases, identified via X-ray powder diffraction: albite
with variable quantities of prehnite, analcime, and
thomsonite, and small amounts of calcite and musco-
vite. The altered scapolite retains its cleavage in hand
sample. Uncommon, microscopic fractures cutting
scapolite contain albite, calcite, and probable Mg-
chlorite; other such fractures contain analcime.
Unaltered scapolite fluoresces bright yellow when
exposed to long- and short-wave ultraviolet radiation.
Pristine scapolite, easily seen when exposed to UV, is
generally restricted to the periphery of phlogopite-
oligoclase mineralization and as isolated ‘pods’ within
the leucocratic zones, where intensely altered scapolite
is always contiguous to corundum-bearing zones (Figs.
7 and 8). Phlogopite crystals associated with scapolite,
occurring on the periphery of phlogopite-oligoclase
intergrowths, have well-developed basal faces and no
embayed grain boundaries with scapolite, which
suggests that these minerals form a stable assemblage.
Small pods and zones, ranging in size from 0.5 31
cm to more than 4 310 cm, are composed of fine- to
coarse-grained grey endmember albite (Ab
98
; Table 2),
silvery-grey muscovite (Table 1), idiomorphic blue
corundum, pale-yellow calcite (Fig. 7), subordinate
graphite, and uncommon grains of anhedral pyrrhotite
(,6 mm). Oxy-dravite (Table 6) occurs very uncom-
monly at Beluga, and was only a significant constit-
uent of a corundum-bearing pod in one sample, where
it occurs as a 40 332 mm friable medium-brown mass
near two dark brown, short-prismatic, euhedral crystals
of the same mineral (up to 32 mm in length). The
corundum-bearing zones locally constitute up to 70%
of the rock volume, estimated from visual inspection
of outcrop. Parts of the calc-silicate rock with high
abundance of phlogopite-oligoclase-diopside (~90%
of rock volume) tend to have little or no corundum
mineralization. The corundum-bearing zone contains
open cavities where the rock surface, including
euhedral corundum, is entirely covered by a 1–2 mm
thick coating of prismatic thomsonite with local
analcime and grey, scalenohedral calcite. Some
cavities contain thin thomsonite prisms up to 8 mm
in length with analcime and calcite crystals no larger
than 2.5 mm. LeCheminant et al. (2005) noted
thomsonite seemingly penetrating a corundum crystal.
The surface of those corundum crystals enclosed in
albite-muscovite-calcite and those enclosed in thom-
sonite are identical in morphology and show no
corrosion or alteration.
The corundum crystals are euhedral and tapered
along the caxis; their faces are striated perpendicular
to c. Most crystals have dimensions in the range of 8–
14 mm long and 1–4 mm wide, but larger crystals are
not infrequent (2–4 cm long) and length:diameter
aspect ratios range from approximately 3:1 to 9:1. The
largest crystal recovered intact at Beluga measured 7.7
32.1 cm (LeCheminant et al. 2005). The corundum is
royal blue, dark blue, or dark greyish blue in hand
sample. When observed in cut cross-sections or
gemstones, the corundum varies from colorless and
light blue to bright blue, dark blue, or dark blue with a
grey tinge in well-defined oscillatory zoning patterns
and irregular sector zoning. The corundum has good
TABLE 1. AVERAGE COMPOSITION OF
PHLOGOPITE AND MUSCOVITE
wt.%
phlogopite
Beluga
n¼41
phlogopite
Bowhead
n¼18
muscovite
Beluga
n¼12
muscovite
Bowhead
n¼2
SiO
2
39.79 40.99 44.96 45.00
TiO
2
2.53 1.84 0.04 0.04
Al
2
O
3
15.58 15.01 37.39 38.47
Cr
2
O
3
0.03 0.01 0.01 0.00
MgO 23.32 25.12 0.02 0.08
CaO 0.02 0.01 0.01 0.14
BaO 0.04 0.06 0.01 0.09
MnO 0.03 0.02 0.02 0.00
FeO 2.92 2.14 0.04 0.03
Na
2
O 0.27 0.31 0.81 1.76
K
2
O 10.56 10.63 10.68 8.96
F 1.39 1.96 0.00 0.00
Cl 0.02 0.02 0.02 0.04
H
2
O* 3.57 3.36 4.46 4.51
O¼F,Cl –0.59 –0.83 0 –0.01
Total 99.48 100.65 98.46 99.11
Normalized to 12 anions with OH* þFþCl ¼2
Si (apfu) 2.820 2.860 3.020 2.985
Ti 0.135 0.097 0.002 0.002
Al (IV) 1.180 1.140 0.980 1.015
Al (VI) 0.121 0.094 1.980 1.993
Cr 0.002 0.001 0.001 –
Mg 2.463 2.613 0.002 0.008
Ca 0.002 0.001 0.001 0.010
Ba 0.001 0.002 0.000 0.002
Mn 0.002 0.001 0.001 –
Fe 0.173 0.125 0.002 0.002
Na 0.037 0.042 0.105 0.226
K 0.955 0.946 0.915 0.758
F 0.312 0.432 – –
Cl 0.002 0.002 0.002 0.004
OH* 1.686 1.565 1.998 1.996
Rcations 7.891 7.922 7.009 7.001
* Calculated, OH ¼2–F–Cl
678 THE CANADIAN MINERALOGIST

transparency and is sparsely included by calcite and
apatite. Gem quality is principally controlled by the
degree of fracturing. The corundum at Beluga contains
an average of 0.08 wt.% TiO
2
(range 0.00–0.30 wt.%)
and 0.07 wt.% FeO (range 0.02–0.13 wt.%), and there
is a weak positive correlation between Fe and Ti
content (correlation coefficient ¼0.26; see Dzikowski
2013 for spot analyses data).
Rare rutile, sanbornite, thorianite, monazite, and
uraninite are noted from the Beluga calc-silicate pod
(Dzikowski 2013, LeCheminant et al. 2005).
Bowhead occurrence. Mineralization in the Bow-
head calc-silicate pod is relatively similar to that at
Beluga with the notable exception that it is largely
devoid of corundum mineralization, whereas nepheline
is common. LeCheminant et al. (2005) noted trace
corundum in the surface exposure. The rock at
Bowhead is composed of coarse to very coarse grained
phlogopite-oligoclase intergrowths (partly replacing
diopside), diopside, scapolite, calcite, and nepheline.
The diopside and phlogopite at Bowhead are slightly
poorer in Ti and Fe relative to Beluga (Tables 1 and 3).
The oligoclase is slightly more calcic than that at
Beluga (Ab
76
; Table 2), and the same is true of
scapolite (average Me
60
, range Me
50
–Me
67
; Table 5).
Both nepheline and scapolite have irregular curved
boundaries with calcite. Some scapolite grains appear
to embay nepheline. Scapolite rims of variable
thickness, some of which are very thin and difficult
to observe in low magnification, occur around
nepheline grains contiguous to calcite (Fig. 9).
Multiple nepheline grains are proximal to phlogo-
pite-oligoclase intergrowths bordering diopside. The
extent of diopside replacement by phlogopite-oligo-
clase in the nepheline-bearing Bowhead samples
TABLE 2. AVERAGE COMPOSITION OF PLAGIOCLASE AND NEPHELINE
wt.%
oligoclase
Beluga
n¼42
albite
Beluga
n¼9
oligoclase
Bowhead
n¼8
albite
Bowhead
n¼5
nepheline
Bowhead
n¼20
SiO
2
63.27 66.31 62.65 68.16 43.49
Al
2
O
3
23.27 20.61 23.74 19.96 34.45
MgO 0.13 0.07 0.00 0.00 0.00
CaO 3.93 0.41 4.96 0.38 2.37
MnO 0.01 0.00 0.01 0.02 0.00
FeO 0.05 0.02 0.03 0.01 0.00
Na
2
O 9.23 11.61 8.82 11.33 15.69
K
2
O 0.19 0.11 0.07 0.10 3.92
Total 100.08 99.14 100.28 99.96 99.92
Norm. 8 O 8 O 8 O 8 O 4 O
Si (apfu) 2.793 2.932 2.766 2.978 1.035
Al 1.211 1.074 1.235 1.028 0.966
Mg 0.009 0.005 – – –
Ca 0.186 0.019 0.235 0.018 0.061
Mn 0.000 – 0.000 0.001 –
Fe 0.002 0.001 0.001 0.000 –
Na 0.790 0.995 0.755 0.960 0.724
K 0.011 0.006 0.004 0.006 0.119
Or 0.01 0.01 0.00 0.01
Ab 0.80 0.98 0.76 0.98
An 0.19 0.02 0.24 0.02
FIG. 6. Diopside (Di) surrounded by phlogopite-oligoclase
symplectite and coarser oriented intergrowths of phlog-
opite (Phl), oligoclase (Pl), and calcite (Cal). Beluga
occurrence. Cross-polarized light.
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 679

(~15–30%) is inferior to the average level of
replacement at Beluga and calcite is a major
component of the nepheline-bearing rock, whereas it
is less abundant at Beluga. A single, 50 lm-wide
nodule of muscovite (Table 1) and endmember albite
(Ab
98
; Table 2) was found in scapolite.
Whole rock composition – Beluga deposit
The Beluga calc-silicate pod is primarily composed
of SiO
2
(Table 7; ca. 46 wt.%), CaO (17 wt.%), Al
2
O
3
(14 wt.%), and MgO (10 wt.%). The rock is richer in
Na than K (average Na
2
O¼2.71 wt.%, K
2
O¼2.02
wt.%, Na/K ¼2.04) and contains 2.02 wt.% total Fe
and 0.98 wt.% Ti. The atomic ratio of Al/Si is 0.36.
Loss on ignition is on the order of 5%, where CO
2
is
the principal volatile. The rock is relatively enriched in
V (Table 8; 237 ppm) and, to a lesser degree, Cr (108
ppm). The Beluga occurrence is very poor in Co (2.5
ppm), Ni (5 ppm), and Cu (5 ppm). Fluorine is
relatively abundant (1260 ppm). The concentration of
Bis~80 ppm, and Cl is ~545 ppm. The average total
rare earth element (REE) concentration is relatively
low, 18.1 ppm, and the rock is slightly enriched in
LREE with a smooth, flat HREE chondrite-normalized
signature and variable negative Eu anomaly (see
discussion).
B and O isotope compositions
The oxygen isotopic composition (d
18
O
VSMOW
)of
corundum from the Beluga occurrence is 16.4 6
0.1%. Oxy-dravite has an average boron isotope
composition (d
11
B) of þ3.9 60.7%(Table 9).
Zircon U-Pb geochronology
Zircon was recovered from the phlogopite-oligo-
clase portion of rock from the Beluga occurrence. The
zircon in the sample displays a range of morphologies,
including coarse, irregular grains with rounded edges
and sharp-faceted, stubby prismatic to equant grains.
All the grains are clear to translucent but range from
medium to dark brown in color. Five single grains
(mostly ca. 130 lm in maximum dimension) were
selected for analysis, including one coarse (.180 lm
diameter), dark brown, translucent grain fragment with
all rounded or broken edges (fraction A); two single,
clear, stubby euhedral, prismatic brown grains (frac-
tions B and C); and two single, clear, slightly elongate
brown prismatic grains (fractions D and E). The five
zircon fractions contain relatively high U concentra-
tions (608–1258 ppm) and have low Th/U ratios (0.06–
0.12; Table 10). The analyses are slightly to strongly
discordant (2.4–13.3%) but define a linear array,
interpreted to reflect mainly recent Pb loss from a
single age of metamorphic zircon, with calculated
upper and lower intercepts (calculated using Model 1
of Ludwig 2003) of 1782.5 63.7 Ma and –6 6140
Ma, respectively, with calculated MSWD ¼0.23 (Fig.
10).
TABLE 3. AVERAGE COMPOSITION OF DIOPSIDE
wt.%
Beluga Bowhead Beluga Bowhead
n¼18 n¼9 Normalized to 6 anions
SiO
2
51.24 51.99 Si (apfu) 1.856 1.888
TiO
2
1.29 0.78 Ti 0.035 0.021
Al
2
O
3
7.25 5.76 Al 0.310 0.247
Cr
2
O
3
0.02 0.01 Cr 0.001 0.000
MgO 14.24 15.01 Mg 0.769 0.813
CaO 22.52 23.28 Ca 0.874 0.906
MnO 0.04 0.03 Mn 0.001 0.001
FeO 1.59 1.20 Fe 0.048 0.036
Na
2
O 1.69 1.56 Na 0.119 0.110
Total 99.88 99.62 Rcations 4.013 4.022
TABLE 4. COMPOSITION OF ZIRCON FROM THE
PHLOGOPITE-OLIGOCLASE PORTION OF BELUGA
CALC-SILICATE ROCK
wt.%n¼2
SiO
2
33.54
ZrO
2
65.34
HfO
2
1.36
La
2
O
3
0.01
Ce
2
O
3
0.01
Pr
2
O
3
0.07
Nd
2
O
3
0.02
Sm
2
O
3
0.13
Gd
2
O
3
0.03
ThO
2
0.05
UO
2
0.00
Total 100.56
680 THE CANADIAN MINERALOGIST

Ar-Ar ages of mica and estimate of T
c
Two
40
Ar-
39
Ar data collections from Beluga
phlogopite (Table A1 in the appendix, available from
the Depository of Unpublished Data on the MAC
website, document Kimmirut sapphire CM55-
4_10.3749/canmin.1700018) produced: (1) a flat
spectrum yielding a plateau age of 1635.9 68.4 Ma
(Fig. A1.A) where data corresponding to 91.4% of the
39
Ar volume was used, and (2) a non-ideal saddle-
shaped spectrum with a flat minimum yielding a
plateau age of 1646.8 68.6 Ma (Fig. A1.B) where
63.7% of the data was used. Two analyses of
muscovite (Table A1) yielded the following plateau
ages: (1) 1511.7 68.4 Ma using 34% of the
39
Ar
volume (Fig. A1.C) and (2) 1510.4 68.3 Ma using
78% of the
39
Ar volume (Fig. A1.D). Closure
temperatures were calculated using CLOSURE v1.2
(Brandon et al. 1998) and different sets of activation
energies and frequency factors for phlogopite and
muscovite (Table 11). Using the activation energies
and frequency factors of Villa (2010) and Harrison et
al. (2009), we estimate the cooling rate to be on the
order of ca. 0.35–0.75 8C/Ma, and the closure
temperatures of phlogopite and muscovite to be ca.
455–515 8C and 410–425 8C, respectively. The
accuracy of these results is limited by the quality of
the models, the validity of underlying assumptions
(size of diffusion domain), and the use of only two
minerals as geothermometers.
DISCUSSION
Paragenetic sequence and metamorphic history
Butler (2007) suggested the following possible
nepheline-forming reactions for rocks on Aliguq
island:
TABLE 5. AVERAGE COMPOSITION OF SCAPOLITE WITH STANDARD DEVIATION
AND MINIMUM/MAXIMUM Me%COMPOSITIONS
wt.%
Beluga
n¼67 rMin.** Max.**
Bowhead
n¼15 rMin.** Max.**
SiO
2
49.04 1.78 49.90 46.84 47.95 1.35 49.54 45.44
Al
2
O
3
26.30 1.00 25.52 27.09 26.21 0.75 25.31 27.58
MgO 0.06 0.22 0.00 0.00 0.00 0.01 0.00 0.00
CaO 14.34 1.77 13.03 15.13 15.10 1.17 14.02 17.38
MnO 0.01 0.02 0.00 0.00 0.01 0.02 0.00 0.01
FeO 0.02 0.03 0.10 0.00 0.02 0.02 0.00 0.08
Na
2
O 5.52 0.93 6.12 4.97 5.31 0.66 5.87 4.16
K
2
O 0.45 0.18 0.49 0.39 0.33 0.08 0.39 0.24
SO
3
0.12 0.05 0.08 0.09 0.12 0.08 0.20 0.06
Cl 2.59 0.92 3.65 2.30 2.65 0.57 3.24 1.66
CO
2
* 1.82 1.02 0.31 1.90 1.71 0.70 0.71 2.66
O¼Cl –0.58 0.21 –0.82 –0.52 –0.6 0.13 –0.73 –0.37
Total 99.69 0.81 98.37 98.19 98.81 0.49 98.55 98.9
Normalized to Al þSi ¼12
Si (apfu) 7.353 7.487 7.136 7.189 7.490 6.996
Al 4.647 4.513 4.864 4.631 4.510 5.004
Mg 0.013 – – 0.000 – –
Ca 2.304 2.095 2.470 2.426 2.271 2.867
Mn 0.001 – – 0.001 – 0.001
Fe 0.003 0.013 – 0.003 – 0.010
Na 1.605 1.780 1.468 1.544 1.721 1.242
K 0.104 0.094 0.076 0.076 0.075 0.047
SO
4
0.014 0.009 0.010 0.014 0.023 0.007
Cl 0.658 0.928 0.594 0.673 0.830 0.433
CO
3
* 0.328 0.063 0.396 0.313 0.147 0.560
Me(%)57 5062 60 5067
See Dzikowski (2013) for full data set
* Calculated assuming CO
3
þCl þSO
4
¼1
** Examples of minimum and maximum Me(%) analyses
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 681

dolomite þalbite !diopside þnepheline þ2CO
2
ð1Þ
phlogopite þ3 calcite þ3 albite
!3 diopside þ3 nepheline þK-feldspar
þ3CO
2þH2Oð2Þ
Based on the inclusion of nepheline by diopside,
phlogopite, and forsterite, Butler (2007) ruled out
reaction (1), since it occurs at lower metamorphic
grade than a forsterite-bearing assemblage. The
implied late formation of nepheline and its intergrowth
with K-feldspar led Butler (2007) to infer that
nepheline formed from reaction (2). A thermodynamic
model of the reaction presented by Butler (2007)
suggests that at 8 kbar pressure, reaction 2 only occurs
at X
CO2
,0.15. This seems unlikely because the
reaction produces CO
2
at a ratio of 3:1 relative to H
2
O,
and therefore the reaction would be self-limiting under
these P-T conditions unless X
CO2
was buffered by a
large supply of H
2
O-dominant metamorphic fluid. In
the Beluga and Bowhead calc-silicate pods, the
breakdown of diopside to a mixture of phlogopite,
oligoclase, and calcite clearly implies the retrograde
reversal of the peak metamorphic assemblage in
reaction 2. Thermodynamic modeling of reaction 2
in TWQ v1 (Fig. 11; Berman 1988, 1991) places the
phase boundary at ~950 8C for 8.0 kbar pressure, far
higher in temperature than the determined value for
peak metamorphic grades, approximately 810 8C (St-
Onge et al. 2007). While the phase boundary of
reaction 1 is most consistent with the P-T history of the
LHG, reaction 2 is significantly more consistent with
the whole rock composition and mineral parageneses.
The expected whole rock composition based on the
mineral proportions in reaction 1 differs significantly
from the measured composition of Beluga calc-silicate
rock: the abundance of K relative to Na (whole rock
K/Na ¼0.49) cannot be accounted for by the relatively
small K concentrations found in nepheline (Bowhead
nepheline K/Na ¼0.16). Potassium feldspar was not
observed in samples from Beluga or Bowhead, but it
has been noted in other corundum-bearing calc-silicate
pods in the area (Hansen 2008). Therefore, the
presence of K-feldspar in the peak metamorphic
assemblage is inferred from its suggested role in
metamorphic reactions, its coeval relationship with
nepheline in calc-silicate rocks elsewhere in the LHG
(Butler 2007), and its ability to account for whole rock
K concentrations higher than expected for the peak
metamorphic paragenesis partly preserved in Beluga
and Bowhead calc-silicate rock.
Zircon recovered from phlogopite-oligoclase coro-
nae have a concordia intercept age of 1782.5 63.7
Ma, interpreted as the age of zircon neocrystallization
and breakdown of the peak metamorphic assemblage
during the post-D
2
fluid incursion suggested by St-
Onge et al. (2007) in P-T conditions slightly inferior to
M
2
metamorphism (720 8C, 6.2 kbar). The timing of
this retrograde mineralization event is consistent with
slow cooling rates estimated from phlogopite-musco-
vite ages (0.35–0.75 8C/Ma). The extent to which the
retrograde mineral assemblage replaces the peak
assemblage [i.e., reversed reaction (2)] is dependent
on the availability of CO
2
and H
2
O in fluids introduced
during this episode of retrograde metamorphism.
FIG. 7. Contact between coarse-grained scapolite (Scp),
phlogopite-oligoclase-calcite (Phl-Pl-Cal), and the corun-
dum-bearing zone. The latter zone contains idiomorphic
corundum (Crn) with albite (Ab), calcite (Cal), and
muscovite (Ms) of variable grain size. Dark zones of fine-
grained alteration (Alt) consist of mixtures of the
following minerals in variable abundance: albite, calcite,
muscovite, analcime, prehnite, and thomsonite. Beluga
occurrence. Plane polarized light.
682 THE CANADIAN MINERALOGIST

Scapolite appears stable with phlogopite since the
phlogopite crystals possess well-developed basal
faces, and neither mineral embays the other: thus we
infer that the scapolite is part of the phlogopite-
oligoclase-calcite assemblage, although since it is not
present within the phlogopite coronae, it may have
formed slightly later from destabilization of nepheline
not consumed in the phlogopite-forming reaction. The
higher X
Ca
of scapolite (Me
50–67
) relative to oligoclase
(An
19–24
) is consistent with compositional data for
coexisting scapolite-plagioclase pairs in metamorphic
rocks, but F and Cl contents of phlogopite (X
Mg
’
0.93, F ’1.39–1.96 wt.%, Cl ¼0.02 wt.%) are
significantly different from values in biotite associated
with scapolite in these metasediments (X
Mg
’0.6, F
’0.6 wt.%, Cl ’0.2 wt.%; Mora & Valley 1989). At
the Bowhead occurrence, rims of scapolite separate
nepheline and calcite; this suggests that during the first
episode of retrograde metamorphism, scapolite formed
from the reaction of nepheline with calcite. The
availability of NaCl in fluid stabilizes scapolite
relative to plagioclase in the presence of calcite
(Ellis 1978), but the effect of fluid Cl content on the
relative stability of nepheline-scapolite-calcite is
unknown. In corundum-bearing calc-silicate rock at
Pitcairn, New York, rims of NaCl-bearing scapolite
separating nepheline from calcite or apatite imply the
formation of scapolite from the reaction of nepheline
with calcite or apatite and CO
2
- and Cl-bearing fluid
(P.M. Belley, unpublished data). We suggest that
scapolite formation at Beluga and Bowhead may have
been limited by Cl availability, whereas fluid Na
content may not be a controlling factor, since the
nepheline is Na-rich. Therefore, the extent of scapolite
formation at the expense of nepheline and calcite may
be controlled by the availability of Cl introduced to the
system concomitant with CO
2
- and H
2
O-influx during
retrograde metamorphism.
Pods and zones of albite, muscovite, calcite, and
corundum are always surrounded by pervasively
albitized (Ab
98
) scapolite. A similar phenomenon is
observed on a centimeter scale around small corun-
dum-bearing zones in scapolite at Pitcairn, New York,
but muscovite is not present at this locality (P.M.
Belley, unpublished data). This consistent spatial
relationship, together with the presence of calcite in
the fine-grained alteration, suggests that the carbonate-
bearing scapolite is a reactant in corundum-forming
reaction (3), below. Reaction (3) excludes graphite and
pyrrhotite, two minerals occurring in the corundum-
FIG. 8. Beluga calc-silicate rock under shortwave ultraviolet light showing the fluorescent scapolite (yellow; Scp), variable
diopside-phlogopite-oligoclase assemblages (Di-Phl-Pl), albitized scapolite (Ab), and albite-corundum-muscovite
assemblages (Ab-Crn-Ms). Minor calcite is present. Purple coloration is an artefact of the UV light source.
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 683

bearing assemblage at Beluga which probably formed
as a result of carbonate and sulfate (released from
scapolite breakdown) reduction, respectively.
Scapolite Me57
ðÞþ0:84 nepheline þ0:60H2O
þ2:02CO
2!0:76 corundum þ2:35 calcite
þ2:14 albite þ0:60muscovite
þ0:66 Cl ð3Þ
Reaction (4), an approximation of reaction (3), was
modeled in TWQ:
meionite þ3 nepheline þ3CO
2
!3 corundum þ4 calcite þ3 albite ð4Þ
The phase boundary for reaction (4) is in the 650–700
8C range for 2.5–4 kbar pressure (Fig. 11). The phase
boundary temperature for the breakdown reaction of
scapolite and nepheline with CO
2
is probably overes-
timated by the TWQ thermodynamic model, since
Kimmirut scapolite is richer in Na (Me
50–67
) and
endmember meionite becomes stable at higher tem-
perature relative to Na-bearing scapolite (Goldsmith &
Newton 1977). Moreover, the TWQ modeled reaction
does not take into account the formation of muscovite,
an important constituent of the albite-calcite-corun-
dum assemblage at Beluga, although this may not be
an important control on mineralization, since musco-
vite is absent in a similar retrograde corundum-calcite-
albite assemblage at Pitcairn, New York.
Zeolites, whose presence is characteristic of low
grade metamorphism, formed in corundum-bearing
zones at Beluga relatively late (,1500 Ma) and are
accompanied by the breakdown of calcite, scapolite, or
possibly relict nepheline, resulting in the creation of
open space.
Uncommon oxy-dravite crystals occur near the
border of phlogopite-oligoclase and the corundum-
bearing zone. Due to the wide P-T stability range of
tourmaline (van Hinsberg et al. 2011) and the lack of
mineral textures that could provide information on its
paragenesis, the position of oxy-dravite in the
paragenetic sequence could not be determined.
Controls on corundum genesis
At Kimmirut, corundum was formed by the
following sequence of events: (1) formation of a
nepheline-diopside-rich peak metamorphic assemblage;
(2) high-temperature partial retrogression of this
assemblage where some nepheline is preserved due to
low availability of fluids (CO
2
,H
2
O), and where
scapolite forms at its expense but is limited by the
availability of calcite and Cl; and (3) introduction of
CO
2
-rich fluid at slightly lower temperature, causing
nepheline and scapolite to react, forming albite, calcite,
corundum, and muscovite. Muscovite formation is
controlled by the availability of K (from nepheline)
and H
2
O (fluid) and is probably not essential to the
corundum-forming reaction. Corundum mineralization
is extensive at the Beluga occurrence, whereas it is rare
at Bowhead, where the nepheline is largely unaltered.
Since the occurrences are separated by a distance of
only 170 m, it is evident that local variations play an
important role in controlling the reaction.
TABLE 6. AVERAGE COMPOSITION OF OXY-
DRAVITE AT THE BELUGA OCCURRENCE
wt.%
Beluga
n¼7 Site
Normalized to
15 Y þZþT cations
SiO
2
36.18 T Si (apfu) 5.844
TiO
2
0.36 T Al 0.156
B
2
O
3a
10.76 B
a
3.000
Al
2
O
3
37.10 Z Al (Z) 6.000
V
2
O
3
0.03 Y Ti 0.044
Cr
2
O
3
0.01 Y Al 0.906
FeO 1.14 Y V 0.004
MnO 0.00 Y Cr 0.001
ZnO 0.02 Y Fe
2þ
0.154
MgO 7.84 Y Mn –
CaO 0.46 Y Zn 0.002
Na
2
O 2.82 Y Mg 1.888
K
2
O 0.04 X Ca 0.080
F 0.30 X Na 0.883
Cl 0.00 X K 0.008
O¼F,Cl –0.13 X Vacancy 0.029
Total 99.68 V OH
b
2.950
VO
b
0.050
W F 0.153
W Cl 0.000
WO
b
0.847
a
Boron calculated assuming ideal 3 B apfu.
b
OH–O calculated by cation charge balance.
FIG. 9. Thin scapolite rim between a grain of nepheline and
calcite. Bowhead occurrence. Cross-polarized light.
684 THE CANADIAN MINERALOGIST

Corundum forms from the breakdown of nepheline
þscapolite in a narrow temperature range with upper
and lower limits controlled by nepheline þscapolite
stability and corundum versus Al-silicate stability,
respectively. Al-silicate forms by reaction (5):
2 meionite þ3 nepheline þ6CO
2
!8 calcite þ6Al
2SiO5þ3 albite ð5Þ
A TWQ thermodynamic model (Fig. 11) indicates a
70–90 8C window in which nepheline and meionite
would break down to form corundum rather than Al-
silicate for expected pressures in a barrovian retro-
grade P-T path. The actual position of this window in
P-T space could only be determined with improved
thermodynamic models for Na-bearing scapolite.
A possible magmatic origin?
LeCheminant et al. (2005) proposed that the
Beluga and Bowhead calc-silicate pods consist of
TABLE 7. WHOLE ROCK MAJOR ELEMENT COMPOSITION OF THE BELUGA
OCCURRENCE
wt.%BA-1 BA-2 BA-3 BA-4 BA-5 Average
P
2
O
5
0.03 0.02 0.02 0.02 0.02 0.02
SiO
2
45.19 45.86 46.07 45.37 46.54 45.81
TiO
2
0.92 1.07 1.03 0.93 0.95 0.98
Al
2
O
3
15.26 11.47 13.3 15.32 14.92 14.05
Cr
2
O
3
0.01 0.01 0.01 0.01 0.01 0.01
Fe
2
O
3
1.47 1.83 1.53 1.43 1.47 1.55
FeO 1.16 1.42 1.16 1.16 1.09 1.2
MnO 0.02 0.03 0.03 0.02 0.02 0.02
MgO 9.28 11.46 10.53 9.59 9.87 10.15
CaO 15.82 18.86 17.62 16.07 16.92 17.06
SrO 0.02 ,0.01 0.01 0.03 0.02 0.02
BaO 0.02 0.01 0.02 0.02 0.02 0.02
Na
2
O 2.96 2.43 2.68 2.94 2.55 2.71
K
2
O 2.45 1.55 1.86 2.37 1.88 2.02
LOI 5.58 4.56 4.37 5.1 3.95 4.71
Total 100.19 100.58 100.24 100.38 100.23 100.33
H
2
O- 0.04 0.01 0.06 0.04 0.02 0.03
H
2
Oþ0.89 0.66 0.66 1.12 1.22 0.91
C 1.15 0.95 0.93 1.05 0.66 0.95
CO
2
4.2 3.5 3.4 3.8 2.4 3.46
Mol.%
P 0.009 0.006 0.006 0.006 0.006 0.006
Si 15.338 15.805 15.832 15.293 15.900 15.634
Ti 0.235 0.277 0.266 0.236 0.244 0.252
Al 6.104 4.659 5.387 6.086 6.008 5.651
Cr 0.003 0.003 0.003 0.003 0.003 0.003
Fe
3þ
0.375 0.475 0.396 0.363 0.378 0.398
Fe
2þ
0.329 0.409 0.333 0.327 0.311 0.343
Mn 0.006 0.009 0.009 0.006 0.006 0.006
Mg 4.696 5.888 5.395 4.819 5.027 5.164
Ca 5.753 6.964 6.488 5.804 6.194 6.238
Sr 0.004 ,0.002 0.002 0.006 0.004 0.004
Ba 0.003 0.001 0.003 0.003 0.003 0.003
Na 1.948 1.624 1.786 1.921 1.689 1.793
K 1.061 0.681 0.815 1.019 0.819 0.879
H– 0.091 0.023 0.138 0.090 0.046 0.068
Hþ2.015 1.517 1.513 2.518 2.780 2.072
C 1.953 1.638 1.599 1.770 1.128 1.622
C 1.946 1.647 1.595 1.749 1.119 1.612
O 58.132 58.373 58.435 57.983 58.336 58.252
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 685

TABLE 8. WHOLE ROCK TRACE ELEMENT COMPOSITION OF THE BELUGA
OCCURRENCE
ppm BA-1 BA-2 BA-3 BA-4 BA-5 Average
Rb 80.2 47.1 56 71.2 55.9 62.08
Cs 3.5 1.9 2.5 3.2 2.1 2.64
Sr 305 131.5 230 368 423 291.5
Ba 64.9 38.5 49.7 51 69.5 54.72
Sc 11.8 13.8 12.3 9.9 11.2 11.8
V 218 263 236 221 247 237
Cr 110 130 100 100 100 108
Zr 196.0 248.0 212.0 198.5 220.0 215.0
Hf68777 7
Nb43433 3.4
Ta 0.5 0.5 0.5 ,0.5 0.5 0.5
Mo ,2,2,2,2,2,2
W11123~2
Co 2.4 3.2 2.3 2.5 2.3 2.5
Ni56556 5.4
Cu5765,5~5
Zn 32 35 31 29 34 32
Ag ,1,1,1,1,1,1
Ga 27 27 26 27 28 27
Tl ,0.5 ,0.5 ,0.5 ,0.5 ,0.5 ,0.5
Sn33333 3
Pb ,5,5,5,5,5,0.5
Re 0.002 0.003 ,0.002 0.003 ,0.002 ~0.002
Se39443 4.6
Te ,0.05 0.27 0.07 0.06 ,0.05 ~0.1
Y 3.5 3.9 3.7 3.3 3.5 3.6
La 2.7 2.7 2.7 2.9 2.6 2.7
Ce 6.4 6.4 6.2 5.5 5.7 6.0
Pr 0.8 0.9 0.9 0.9 0.8 0.9
Nd 3.9 4.3 4.3 3.5 3.7 3.9
Sm 1.2 1.4 1.3 1.1 1.2 1.2
Eu 0.2 0.2 0.2 0.2 0.3 0.2
Gd 0.9 1 1 0.9 0.9 0.9
Tb 0.1 0.2 0.2 0.1 0.1 0.1
Dy 0.8 0.9 0.8 0.7 0.8 0.8
Ho 0.1 0.2 0.2 0.1 0.2 0.2
Er 0.4 0.5 0.4 0.4 0.4 0.4
Tm ,0.1 0.1 0.1 ,0.1 ,0.1 ~,0.1
Yb 0.4 0.5 0.4 0.4 0.4 0.4
Lu 0.1 0.1 0.1 ,0.1 0.1 0.1
Th11111 1
U 1.7 1.8 2.1 1.7 1.7 1.8
Li 17.6 14.2 13.9 14.5 16.2 15.3
Be 1.84 2.01 1.93 1.78 1.84 1.88
B 60 60 90 100 80 78
F 1580 1050 1240 1390 1030 1260
Cl* 570 230 690 800 460 550
Cl** 490 270 620 810 510 540
* Neutron activation analysis
** Specific ion electrode analysis.
686 THE CANADIAN MINERALOGIST

TABLE 9. BORON ISOTOPE COMPOSITION OF OXY-
DRAVITE FROM THE BELUGA OCCURRENCE
d
11
B(%)r(%)
4.6 0.3
4.3 0.3
4.4 0.3
3.7 0.3
4.2 0.3
2.8 0.3
2.7 0.3
3.3 0.3
3.9 0.3
4.4 0.3
4.4 0.3
Average
3.9 0.7
TABLE 10. U-Pb ANALYTICAL DATA FOR ZIRCON FROM THE BELUGA OCCURRENCE
Sample
description
1
A: N2, þ180,t B: N2, þ134,s,c C: N2, þ134,s,c D: N2, þ134,e,c E: N2, þ134,e,c
Wt. (mg) 0.090 0.033 0.011 0.032 0.023
U (ppm) 1233 608 632 776 1258
Pb
2
(ppm) 355 169 170 229 385
206
Pb/
204
Pb
(meas.)
3
3977 1294 1569 15970 1344
Tot. common Pb
(pg)
513 277 76 30 418
%
208
Pb
2
1.5 2.0 2.1 1.6 3.0
206
Pb/
238
U
4
(6%1r)
0.29734(0.19) 0.28611(0.14) 0.27582(0.17) 0.30505(0.07) 0.31143(0.11)
207
Pb/
235
U
4
(6%1r)
4.4700(0.26) 4.3091(0.42) 4.1416(0.36) 4.5840(0.09) 4.6856(0.39)
207
Pb/
206
Pb
4
(6%1r)
0.10903(0.14) 0.10923(0.35) 0.10890(0.28 0.10899(0.03) 0.10912(0.33)
206
Pb/
238
U age
(Ma; 6%2r)
1678.1(5.5) 1622.1(3.9) 1570.3(4.8) 1716.3(2.1) 1747.7(3.2)
207
Pb/
206
Pb age
(Ma; 6%2r)
1783.3(5.0) 1786.6(12.8) 1781.1(10.1) 1782.6(1.2) 1784.8(12.0)
Discordance %6.7 10.4 13.3 4.2 2.4
Th/U (calc.) 0.057 0.078 0.083 0.063 0.116
1
N2 ¼non-magnetic at 28side slope on Frantz magnetic separator; grain size given in microns; t ¼translucent; c ¼
clear; s ¼stubby prismatic grains; e ¼slightly elongate prismatic grains.
2
Radiogenic Pb; corrected for blank, initial common Pb, and spike.
3
Corrected for spike and fractionation.
4
Corrected for blank Pb and U, and common Pb.
FIG. 10. U-Pb concordia diagram for zircon recovered from
the phlogopite-rich assemblage at the Beluga occurrence
calculated using Model 1 of Ludwig (2003). Error ellipses
represent 2r. See Table 10 and text for U-Pb data and
sample descriptions.
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 687

TABLE 11. ESTIMATED COOLING RATE AND CLOSURE TEMPERATURES FOR PHLOGOPITE AND MUSCOVITE USING DIFFERENT DETERMINATIONS OF
ACTIVATION ENERGY (E
A
) AND FREQUENCY FACTOR (D
0
) FOR PHLOGOPITE (PHL) AND MUSCOVITE (MS) AND CALCULATED IN CLOSURE V1.2
Ms data Phl data
Ms Ms Phl Phl Phl - Ms Phl - Ms Estimated
cooling rate
(8C/Ma)
Phl Ms
E
a
(kJ/mol) D
0
(cm
2
/s) E
a
(kJ/mol) D
0
(cm
2
/s) ~DT
C
(8C)* ~Dt (Ma)**
T
C
(8C)
Approx.
T
C
(8C)
Approx.
Robbins (1972),
Hames & Bowring (1994)
Giletti (1974) 180 4 310
–4
242 242 67 124 0.54 400 330
Robbins (1972),
Hames & Bowring (1994)
Villa (2010)
B
180 4 310
–4
299.4 205 131 124 1.06 470 340
Robbins (1972),
Hames & Bowring (1994)
Villa (2010)
B
180 4 310
–4
359.2 93611 180 124 1.45 525 345
Harrison et al. (2009)
A
Giletti (1974) 268 20 242 242 –20 124 n/a
Harrison et al. (2009)
A
Villa (2010)
B
268 20 299.4 205 44 124 0.35 455 410
Harrison et al. (2009)
A
Villa (2010)
B
268 20 359.2 93611 93 124 0.75 515 425
A750lm spherical radius (approximating a 500 lm cylinder radius) was used as the effective dimension of the diffusion domain.
A
Estimate of D
0
for 5 kbar.
B
Two possible sets of E
a
and D
0
from a re-evaluation of Giletti’s (1974) experiments.
* 0.01–10 8C/Ma cooling rate range (value is a 9 data point average).
** (PHL-A) – (MS-B)
688 THE CANADIAN MINERALOGIST

syenitic magma that intruded, and were contaminated
by, marble during D
2
deformation. Partly contaminat-
ed syenitic intrusions have not been noted in the area,
but they would be expected given the widespread
occurrence of scapolite-bearing and corundum-bearing
calc-silicate pods in LHG marble near Kimmirut.
Some scapolite-rich pods are small (ca.1min
maximum dimension) and isolated within the regional
marble (Hansen 2008). Moreover, the contacts be-
tween marble and the calc-silicate rocks are generally
sharp with no apparent zoning. One author (PMB) has
observed multiple granite pegmatites in marble
throughout the Central Metasedimentary Belt, Gren-
ville Province, where metamorphic grades were in
upper amphibolite to granulite facies, and these dikes
typically have no to minor contamination in the form
of Ca-Mg-bearing silicate minerals. They typically
have thin metasomatic aureoles at their contact with
marble (e.g., tremolite around zircon-rich pegmatite
near Bryson, Qu´
ebec), or are contiguous with coarsely
crystalline vein- or pod-like bodies of clinopyroxene
and feldspar/scapolite with calcite-rich cores (e.g.,
rocks in the Lac Tortue area, ZEC Bras-Coup´
eD
´
esert,
Qu´
ebec). Relative to these examples, the Beluga and
Bowhead calc-silicate pods are remarkably uniform,
despite local variations in mineral abundance. Th/U
ratios in Beluga zircons (Th/U
avg
¼0.08) are within
the range considered to be characteristic for metamor-
phic zircon (Th/U ,0.1) by Rubatto et al. (1999),
although M¨
oller et al. (2003) cautioned against
attributing the origin of zircon using Th/U. Lastly,
the paragenetic sequence of the Beluga and Bowhead
calc-silicate pods appears most consistent with a
FIG. 11. Mineral reactions modeled in TWQ v1 (Berman 1988, 1991). Based on the paragenetic sequence of Bowhead and
Beluga calc-silicate rock (present study) and metasediments on Aliguq Island (Butler 2007), the diopside (Di), nepheline
(Ne), and K-feldspar (Kfs) assemblage is suggested to be possible at M1A metamorphic conditions. Alternatively, the
formation of nepheline and diopside from albite (Ab) and dolomite (Do) is more consistent with regional P-T conditions but
differs significantly from the paragenetic sequence at Beluga and Bowhead. The peak assemblage is partly replaced by
phlogopite (Phl), calcite (Cal), and albite at conditions slightly inferior to M2. The meionite (Me) – nepheline breakdown
reactions are probably overestimates since the measured scapolite compositions contain significant Na. The latter minerals
break down to form corundum (Crn), calcite, and albite at higher temperature than a similar reaction forming Al-silicate,
calcite, and albite. Regional P-T conditions after St-Onge et al. (2007). See text for a detailed discussion of mineral
reactions and paragenetic sequence in the context of regional metamorphism.
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 689

metamorphic origin where the peak metamorphic
assemblage was subject to two distinct, high-temper-
ature retrograde metamorphic events, and the inferred
peak P-T formation of nepheline was noted in
carbonate-bearing metasediments in another region
of the LHG (Butler 2007).
Nature of the protolith
Beluga calc-silicate rock plots between LHG
metapelites and psammites in Al/Si and Na/K ratios
(Fig. 12A), is slightly Fe-depleted relative to shale and
metapelites (Fig. 12B), and contains a TiO
2
concen-
FIG. 12. Comparison of Beluga calc-silicate rock major element composition with selected representative Lake Harbour Group
metasediments (Th´
eriault et al. 2001, Butler 2007), monzogranites (Butler 2007), lazurite-bearing metaevaporites (Hogarth
& Griffin 1978), averages for platform sediments (Carmichael 1989), and averages for syenite (Nockolds 1954).
690 THE CANADIAN MINERALOGIST

tration (~1 wt.%) consistent with a shale/pelitic
component (Fig. 12D). The rock is significantly richer
in CaO and MgO than LHG clastic metasediments and
intrusive rocks but with similar concentrations to LHG
lapis lazuli metaevaporite layers in marble (Fig. 12C).
These data suggest a protolith of mixed clastic and
carbonate composition, where the clastic component is
an intermediate between the pelitic and psammitic
sediments, i.e., a silty clay. While a small quantity of
halite could explain the greater abundance of Na
relative to K, the relative abundance of these elements
at Beluga is consistent with an intermediate between
LHG metapelite and psammite, and the Na/K value at
Beluga (~2) is significantly smaller than that at the
Soper River lapis lazuli metaevaporite (~5 to 19;
Hogarth & Griffin 1978). Metamorphic reactions
resulted in significant decarbonation of the rock, but
initial carbonate content can be estimated by subtract-
ing an estimated MgO and CaO siliciclastic contribu-
tion (an intermediate of LHG metapelite-psammites, 2
wt.% CaO, 3–6 wt.% MgO) from the total bulk CaO
and MgO and recalculating the difference as calcite
and dolomite with a second iteration correcting
assumed initial CaO and MgO values for the amount
of original allochthonous material (i.e., 1.6 wt.% CaO,
2.6–5.2 wt.% MgO). The resulting estimation is a
protolith with ca. 19 wt.% dolomite and 13 wt.%
calcite to 29 wt.% dolomite and 7 wt.% calcite.
Therefore, the most likely protolith based on major
element bulk composition is silty or sandy dolomitic
argillaceous marl, although a thinly interlayered dolo-
stone-shale sequence is equally possible given the
coarse size of metamorphic recrystallization. It should
be noted that the initial relative abundance of Na
versus K may have been affected by diagenetic
processes, notably albitization of detrital feldspars
(e.g., Baccar et al. 1993).
The trace-element composition of Beluga calc-
silicate rock is not as simple to interpret, and it is
important to consider that the complex metamorphic
history of the deposit may have obscured the original
trace element signature in the protolith. Vanadium
(237 ppm) is significantly enriched relative to Cr (108
ppm), while Co, Ni, and Cu (5 ppm) are extremely
depleted. The V concentration at Beluga is similar to
V-rich LHG metapelite and psammite, while Cr is
consistent with shale, argillite, or marl, and Co, Ni,
and Cu concentrations are closer to that of marble (Fig.
13). According to Tuttle et al. (2000), ‘‘aluminum and
titanium are mostly bound to phases that are relatively
unreactive in marine environments; therefore, both of
these elements provide a good estimate of the amount
of allochthonous detritus.’’ Using this assumption, V,
Cr, and Ni concentrations in Beluga rock, LHG
metapelite, and psammite are plotted against the
expected clastic contribution of trace elements in
average shale (Fig. 14). The comparison with shale
suggests significant V enrichment, typical detrital Cr
contribution, and significant Ni (and, by proxy, Co and
Cu) depletion. One LHG psammite (95-C084 of
Th´
eriault et al. 2001) shows similar V-enrichment,
expected Cr, and a weaker Ni-depletion (29 ppm).
A study of transition metal behavior in response to
different early diagenetic environments of modern
sediments (Shaw et al. 1990) demonstrated the
trapping of Ni and Co with manganese oxides, which
are enriched and preserved in the oxic zone of
sediments and released under reducing conditions. In
contrast, near-anoxic reducing conditions (e.g., in
organic-matter-bearing sediment) favor the enrichment
and preservation of Cr, V, and Mo, where accumula-
tion of Cu is moderately enriched in response to
reducing conditions and closely correlated with
biogenic material flux (i.e., the Cu binding capacity
of sediment decreases in slowly accumulating sedi-
ments where a smaller fraction of biogenic detritus
survives). In strongly reducing H
2
S-rich environments
which favor high V/(Ni þV) ratios, maximum molar
V/(Ni þV) are on the order of ~0.85 (black shale-
carbonate sequences of Wenger & Baker 1986) –
significantly lower than that at Beluga [V/(Ni þV) ’
0.98]. Vanadium is strongly enriched in the Beluga
calc-silicate rock, and Cr well-preserved. Molybde-
num concentration is below detection level (,2 ppm),
but considering the low concentration in average shale
(2 ppm; Carmichael 1989) and significant dilution by
the carbonate component in the protolith, Mo enrich-
ment may not be detectable. Low MnO (0.01 wt.%)
relative to a sample of LHG marble (0.17 wt.%; Butler
2007), average shale, and average marine carbonate
(0.08 wt.%, 0.07 wt.%; Carmichael 1989) is consistent
with the proposed mechanism of Co and Ni depletion,
but many LHG metapelites and psammites contain low
MnO with highly variable Ni (Fig. 13).
The extremely low REE concentrations relative to
other lithologies in the LHG and Meta Incognita
peninsula (Fig. 15) could partly be explained by early,
reducing diagenetic conditions since REE also con-
centrate on Mn oxides, although REE could be
preserved in phosphate in anoxic conditions
(Takahashi et al. 2015). The V-enriched, Ni-depleted
LHG psammite is far richer in total REE (~335 ppm)
relative to Beluga (~18 ppm), which does not support
this hypothesis. In LHG metapelites, concentrations of
V, Ni, and Cr show some degree of positive
correlation. Diagenetic processes could potentially
dissolve phosphate minerals and mobilize REEs.
Given the complex metamorphic history of the Beluga
calc-silicate rock, the attribution of early diagenetic
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 691

processes to observed trace element profiles carries
high uncertainty.
The oxygen isotope composition of corundum at
Beluga is slightly higher thanfor corundum in desilicated
pegmatite in marble, and at the low end of the marble
range (Fig. 16). A d
18
O value of þ16.4%is consistent
with the equilibration of oxygen isotopic composition
with marble, and it is higher than for corundum in skarn
developed in marble (Giuliani et al. 2014).
On evaporites
Dzikowski (2013) suggested that the Beluga and
Bowhead protoliths contained an evaporitic compo-
nent. Here, we compare the bulk composition of
Beluga calc-silicate rock with the geochemistry of
evaporite-bearing argillites, dolomitic marls, and
mixed sulfate rocks. These evaporite-bearing or
evaporite-related sediments, in comparison with com-
mon platform sediments, have high Mg, high Mg/Ca
ratios (except Ca-sulfate-rich rock), low Fe contents,
high K with low Na (except halite-bearing rocks), and
specifically in argillites, high Li, F, and B/Al (Moine et
al. 1981). The Al-Mg-Ca geochemical signature of
Beluga calc-silicate rock is within the range of
platform marl and clay-shale and just outside the
compositional domain for evaporites and metaevapor-
FIG. 13. Abundance of selected trace metals in Beluga calc-silicate rock compared to representative LHG metasediments
(Th´
eriault et al. 2001), sediments (Carmichael 1989, Moine et al. 1981 and references therein), and average granite
(Carmichael 1989).
692 THE CANADIAN MINERALOGIST

ites (Fig. 17). Hypothetically, the evaporite-associated
sediment best matching the bulk major element
composition of Beluga rock is argillaceous marl
(clay/silt with dolomite and calcite) containing halite
(Na .K), but Na-K contents are also consistent with a
mixed clastic component (i.e., intermediate between
LHG metapelites and psammites). The boron concen-
tration at Beluga (80 ppm) is similar to that of average
shale (100 ppm; Carmichael 1989) but far lower than
in argillaceous evaporitic rocks (200–400 ppm; Moine
et al. 1981). The whole-rock B measurements at
Beluga may be affected by localized oxy-dravite
mineralization (i.e., oxy-dravite is very uncommon,
but one sample contains a ~8 cm cluster). The
concentration of Li at Beluga (15 ppm) is far lower
than the range for most evaporitic rocks (45–300 ppm),
however Li concentrations are low in Mg-clay-poor,
dolomite-rich sedimentary rocks associated with
evaporites (Moine et al. 1981). Beluga F content
(1260 ppm), mostly held in phlogopite, is comparable
to evaporitic argillite (1000–2000 ppm; Moine et al.
1981) and approximately double that of shale (500–
740 ppm; Carmichael et al. 1989). Beluga calc-silicate
rock is also enriched in Cl (~500 ppm) relative to
shale (160–180 ppm; Carmichael et al. 1989).
Fluorine-rich phlogopite and Cl-rich scapolite are
retrograde assemblages: F and Cl are not highly
compatible constituents within the major mineral
phases in the peak assemblage (nepheline, diopside,
K-feldspar), therefore it can be surmised that these
elements were probably introduced with metamorphic
fluids during retrograde metamorphism.
FIG. 14. Abundance of V, Cr, and Ni relative to Al
2
O
3
. Line represents expected concentration of metals relative to alumina by
varying amount of allochthonous detritus based on metal-Al
2
O
3
ratios from a shale average (Carmichael 1989). Selected
examples of metapelite and psammite from the Lake Harbour Group (Th´
eriault et al. 2001) are included for comparison.
FIG. 15. Chondrite-normalized REE trace element profile of Beluga calc-silicate rock compared to rocks from the Lake Harbour
Group (*Butler 2007, **Th´
eriault et al. 2001). After Dzikowski (2013).
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 693

The boron isotope composition of oxy-dravite is
similar to tourmaline with B sourced from marine
boron. Due to the overlap of tourmaline d
11
Bin
marble, evaporite-associated metapelite, and B-rich
metaevaporite (Fig. 18), an evaporitic protolith cannot
be inferred from the data.
In summary, bulk rock major and trace element
data are consistent with non-evaporitic marine plat-
form sediments. There is no evidence to support an
evaporitic origin of the protolith, although the
possibility that the protolith contained a small
evaporite component cannot be ruled out.
FIG. 16. Oxygen isotope composition of Beluga corundum compared to values from different types of gem corundum deposits
(modified after Giuliani et al. 2014). Color in diamonds represent the color of gem-quality corundum from different deposits
(red ¼ruby, colored sapphire ¼others).
FIG. 17. Al–Mg–Ca diagram showing the distribution of the calc-silicate rock from the Beluga corundum deposit relative to
domains for platform marls and clay-shale, evaporites, and meta-evaporites (modified after Moine et al. 1981) and in
comparison to marbles and intercalated schists and gneisses from ruby-bearing marbles in central and southeast Asia (after
Garnier et al. 2008).
694 THE CANADIAN MINERALOGIST

Implications for gem corundum exploration
Since corundum mineralization is dependent on
two episodes of amphibolite facies retrograde meta-
morphism in Beluga- and Bowhead-like rocks, parts of
the LHG with pervasive amphibolite-facies retrograde
overprinting of peak metamorphic assemblages would
be more prospective than areas with well-preserved,
peak granulite facies assemblages. In the LHG,
domains with pervasive amphibolite facies retrograde
metamorphism are restricted to the periphery of the
thrust fault between the LHG and the Narsajuaq
terrane (see Fig. 5 of St-Onge et al. 2000).
Kimmirut-type gem corundum deposits are pro-
duced by a very specific P-T history and protolith
composition, and therefore are probably rare com-
pared to gem corundum deposits with simpler
paragenetic sequences and more common protoliths
(especially marble-hosted deposits, see Giuliani et al.
2014). A corundum occurrence similar to the
Kimmirut deposits is located in Pitcairn, New York,
USA, where purplish-red, opaque corundum occurs
with albite and calcite formed at the expense of
scapolite and nepheline (P.M. Belley, unpublished
data; Chamberlain et al. 2015).
OPPORTUNITIES FOR FUTURE WORK
Multiple questions arise from this study. First,
improved thermodynamic data for scapolite along the
meionite–marialite solution series and nepheline are
necessary to properly assess the proposed paragenetic
sequence at Beluga and Bowhead. A comparison of
corundum color with calc-silicate rock bulk composi-
tions for Kimmirut-area occurrences may yield
interesting results, and protolith and diagenetic
considerations may explain variations in chromophore
element abundances (Fe, Ti, Cr; Fritsch & Rossman
1987, 1988). Fluid inclusion studies in minerals from
the three successive parageneses may provide infor-
FIG. 18. Boron isotope composition of Beluga oxy-dravite compared to tourmaline from different environments. Sources of
data: (1) Swihart & Moore 1989, (2) Palmer 1991, (3) serendibite-rich calc-silicate rock interpreted to be metamorphosed
illite layer deposited in a hypersaline environment (Grew et al. 1991), (4) serendibite-rich calc-silicate rock similar to the
latter, but with both prograde and retrograde tourmaline (Belley et al. 2014).
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 695

mation on the composition of metamorphic fluids at
different stages and whether evaporites were present
during peak metamorphism. Phlogopite and muscovite
40
Ar/
39
Ar data suggest a slow rate of cooling; regional-
scale geologic interpretations relating to the retrograde
P-T path would be improved by additional geo/
thermochronological work.
CONCLUSIONS
In Lake Harbour Group calc-silicate rocks near
Kimmirut, Nunavut, corundum formation was made
possible by three equally important sequential meta-
morphic reactions: (1) formation of nepheline, diopside,
and K-feldspar (inferred) at peak metamorphic (gran-
ulite facies) conditions from a metamorphosed dolo-
mitic argillaceous marl precursor; (2) partial retrograde
replacement of the peak assemblage by phlogopite,
oligoclase, calcite, and scapolite as a result of CO
2
-,
H
2
O-, Cl-, F-bearing fluid influx at 1782.5 63.7 Ma,
~30 Ma younger than the end of M
2
metamorphism
(therefore P-T ,720 8C, 6.2 kbar); and (3) retrograde
breakdown of scapolite þnepheline (þCO
2
þH
2
O) to
form albite, muscovite, corundum, and calcite. As
evidenced by the abundance of corundum at Beluga and
its near absence at Bowhead (170 m away), conditions
favoring the alteration of scapolite-nepheline were
locally heterogeneous. The corundum-forming reaction
only occurs in a ,100 8C window with a lower limit
determined by the formation of Al-silicate rather than
corundum. Thermodynamic models of these reactions
are inconsistent with known metamorphic conditions in
the LHG, but do not take into account variations in the
composition of major phases. They are interpreted to be
overestimates of phase boundary P-T since they exceed
estimates based on the paragenetic sequence and the
regional metamorphic history. The inferred mineral
reactions are in strong agreement with petrologic
observations, mineral assemblages, and bulk composi-
tion.
Phlogopite and muscovite
40
Ar/
39
Ar ages and
calculated closure temperatures are ca. 1640 Ma (T
c
¼455 to 515 8C) and 1510 Ma (T
c
¼410–425 8C),
respectively. The late formation of thomsonite,
analcime, and prehnite at Beluga is expected to be
considerably younger than the muscovite closure age.
Local-scale and outcrop-scale field observations,
combined with bulk and trace element geochemistry,
suggest that the Beluga and Bowhead calc-silicate
pods are metasedimentary. The relative abundances of
Ca, Mg, Al, Si, Ti, and Fe are consistent with
dolomitic argillaceous marl, where the siliciclastic
component is an intermediate between that of LHG
metapelites and psammites (i.e., sandy or silty clay).
Significant V enrichment, high Cr, and very low Ni,
Co, and Cu may be related to early diagenetic
processes such as a slow sedimentation rate and
reducing conditions caused by the presence of organic
matter.
The Li and B concentrations do not suggest an
evaporite-related protolith, although concentrations of
these elements in evaporite-associated sediments can
be low. The Na/K ratios are intermediate between
LHG metapelites and psammites. High F and Cl
concentrations (ca. 1260 and 500 ppm, respectively)
appear to have been introduced by metamorphic fluid
during retrograde metamorphism. Tourmaline boron
isotopes are consistent with a marine boron source but
cannot be used to distinguish between normal marine
and hypersaline environments. Therefore, data do not
suggest that the Beluga protolith was evaporite-
related; i.e., Beluga whole rock geochemistry is within
the range expected in non-evaporitic marine shelf
sediments. However, the possibility that an initial,
minor evaporitic component was present in the
protolith cannot be rejected. We infer a similar origin
for the Bowhead calc-silicate rock due to its
mineralogical and geological similarity to the Beluga
occurrence.
In the Lake Harbour Group, the most prospective
areas for gem corundum exploration are expected to be
contiguous to the thrust fault separating the LHG and
Narsajuaq terranes, where the retrograde, amphibolite
facies overprint of the granulite peak assemblages is
most pervasive.
ACKNOWLEDGMENTS
We are very grateful to the staff and management of
True North Gems Inc. for granting access to the area,
providing data and maps, and for discussions about the
geology. Brad Wilson provided useful field observations
and helped with field work. Allison Brand provided
assistance in the field. Andrea Cade helped in sample
preparation and data collection for Ar-Ar geochronology.
Ryan Sharpe assisted MF with boron isotope analysis.
Corundum-bearing samples from Pitcairn, NY were
provided by Dr. George Robinson. Donald Lake and
David Turner provided useful suggestions. Access to
Inuit Owned Land to conduct research was granted by
the Qikiqtani Inuit Association and the community of
Kimmirut. Helpful comments from Mackenzie Parker,
Guest Editor Dr. Dan Marshall, and reviewers Drs.
Yannick Branquet and Eloise Gaillou improved the
quality of the manuscript. Financial support was provided
by the Northern Scientific Training Program in the form
of a grant to PMB, project GA17-17276S to JC, and by
the Natural Sciences and Engineering Research Council
of Canada in the form of a Discovery Grant to LAG.
696 THE CANADIAN MINERALOGIST

REFERENCES
BACCAR, M.B., FRITZ, B., & MADE
´, B. (1993) Diagenetic
albitization of K-feldspar and plagioclase in sandstone
reservoirs: Thermodynamic and kinetic modeling. Journal
of Sedimentary Petrology 63, 1100–1109.
BELLEY, P.M., GRICE, J.D., FAYEK, M., KOWALSKI, P.M., &
GREW, E.S. (2014) A new occurrence of the borosilicate
serendibite in tourmaline-bearing calc-silicate rocks,
Portage-du-Fort marble, Grenville Province, Qu´
ebec:
evolution of boron isotope and tourmaline compositions
in a metamorphic context. Canadian Mineralogist 52,
595–616.
BERANEK, L.P. & MORTENSEN, J.K. (2011) The timing and
provenance record of the Late Permian Klondike orogeny
in northwestern Canada and arc-continent collision along
western North America. Tectonics 30, TC5017, doi:10.
1029/2010TC002849.
BERMAN, R.G. (1988) Internally-consistent thermodynamic
data for minerals in the system Na
2
O–K
2
O–CaO–MgO–
FeO–Fe
2
O
3
–Al
2
O
3
–SiO
2
–TiO
2
–H
2
O–CO
2
.Journal of
Petrology 29, 445–522.
BERMAN, R.G. (1991) Thermobarometry using multiequili-
brium calculations: a new technique with petrologic
applications. Canadian Mineralogist 29, 833–855.
BRANDON, M.T., RODEN-TICE, M.K., & GARVER, J.I. (1998)
Late Cenozoic exhumation of the Cascadia accretionary
wedge in the Olympic Mountains, northwest Washington
State. Geological Society of America Bulletin 110, 985–
1009.
BUTLER, J.P. (2007) Petrogenesis of nepheline- and scapolite-
bearing metacarbonates from southwestern Baffin Island,
Nunavut, Canada. B.Sc. Thesis, Dalhousie University,
Halifax, Nova Scotia, Canada.
CABRAL, A.R., WIEDENBECK, M., RIOS, F.J., GOMES, A.A.S., JR,
FILHO, O.G.R., & JONES, R.D. (2012) Talc mineralisation
associated with soft hematite ore, Gongo Soco deposit,
Minas Gerais, Brazil: petrography, mineral chemistry and
boron-isotope composition of tourmaline. Mineralium
Deposita 47, 411–424.
CARMICHAEL, R.S. (1989) Physical properties of rocks and
minerals. CRC Press Inc., Boca Raton, Florida, United
States, 741 pp.
CATANZARO, E.J., CHAMPION, C.E., GARNER, E.L., MARINENKO,
G., SAPPENFIELD, K.M., & SHIELDS, W.R. (1970) Boric
acid: isotopic and assay standard reference materials.
National Bureau of Standards (US) Special Publications
260-17, 1–71.
CHAMBERLAIN, S.C., WALTER, M., BAILEY, D.G., & ROBINSON,
G.W. (2015) A new collecting site at the Rose Road
Locality, Pitcairn, St. Lawrence County, New York.
Rochester Mineralogical Symposium, Rochester, NY.
Abstract. Rocks & Minerals 90, 440–446.
CHAUSSIDON,M.&ALBARE
`DE, F. (1992) Secular boron
isotope variations in the continental crust: An ion
microprobe study. Earth and Planetary Science Letters
108, 229–241.
DRAKE, M.J. & WEILL, D.F. (1972) New rare earth element
standards for electron microprobe analysis. Chemical
Geology 10, 179–181.
DZIKOWSKI, T.J. (2013) A comparative study of the origin of
carbonate-hosted gem corundum deposits in Canada.
Doctoral Thesis, University of British Columbia, Van-
couver, British Columbia, Canada, 259 pp.
DZIKOWSKI, T.J., CEMP´
IREK, J., GROAT, L.A., DIPPLE,G.M.,&
GIULIANI, G. (2014) Origin of gem corundum in calcite
marble: The Revelstoke occurrence in the Canadian
Cordillera of British Columbia. Lithos 198–199,281–
297.
ELLIS, D.E. (1978) Stability and phase equilibria of chloride
and carbonate bearing scapolites at 750 8C and 4000 bar.
Geochimica et Cosmochimica Acta 42, 1271–1281.
FRITSCH,E.&ROSSMAN, G.R. (1987) An update on color in
gems. Part 1: Introduction and colors caused by dispersed
metal ions. Gems & Gemology 23, 126–139.
FRITSCH,E.&ROSSMAN, G.R. (1988) An update on color in
gems. Part 2: Colors involving multiple atoms and color
centers. Gems & Gemology 24, 3–15.
GARNIER, V., GIULIANI, G., OHNENSTETTER, D., FALLICK, A.E.,
DUBESSY, J., BANKS, D., HOA
`NG, Q.V., LHOMME,TH.,
MALUSKI, H., PE
ˆCHER, A., BAKHSH, K.A., PHAM, V.L.,
PHAN, T.T., & SCHWARZ, D. (2008) Marble-hosted ruby
deposits from Central and Southeast Asia: towards a new
genetic model. Ore Geology Reviews 34, 169–191.
GENIS, R. (2016) Retail gemstone price trends (1975–2015).
The Gemstone Forecaster 34, http://www.precious
gemstones.com/gfspring16.html [Retrieved December
20, 2016].
GERTZBEIN, P.J. (2005) Geology surrounding the Beluga
sapphire occurrence, Kimmirut, Nunavut: A preliminary
examination. Canadian Gemmologist 26, 50–57.
GILETTI, B.J. (1974) Studies in diffusion I: argon in phlogopite
mica. In Geochemical Transport and Kinetics (A.W.
Hoffman, B.J. Giletti, H.S. Yoder, & R.A. Yund, eds.).
Publication 634, Carnegie Institution of Washington,
107–115.
GIULIANI, G., FALLICK, A.E., GARNIER, V., FRANCE-LANORD, C.,
OHNENSTETTER, D., & SCHWARZ, D. (2005) Oxygen isotope
composition as a tracer for the origins of rubies and
sapphires. Geology 33, 249–252.
GIULIANI, G., OHNENSTETTER, D., FALLICK, A.E., GROAT, L.A.,
&F
AGAN, A.J. (2014) The geology and genesis of gem
corundum deposits. In Geology of Gem Deposits (L.A.
Groat, ed.). Mineralogical Association of Canada Short
Course 44, 29–112.
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 697

GOLDSMITH, J.R. & NEWTON, R.C. (1977) Scapolite-plagio-
clase stability relations at high pressures and temperatures
in the system NaAlSi
2
O
8
–CaAl
2
Si
2
O
8
–CaCO
3
–CaSO
4
.
American Mineralogist 62, 1063–1081.
GREW, E.S., YATES, M.G., SWIHART, G.H., MOORE, P.B., &
MARQUEZ, N. (1991) The paragenesis of serendibite at
Johnsburg, New York, USA: an example of boron
enrichment in the granulite facies. In Progress in
metamorphic and magmatic petrology; a memorial
volume in honor of D.S. Korzhinskiy (L.L. Perchuk,
ed.). Cambridge University Press, Cambridge, England
(247–285).
HAMES,W.E.&BOWRING , S.A. (1994) An empirical
evaluation of the argon diffusion geometry in muscovite.
Earth and Planetary Science Letters 124, 161–167.
HANSEN, M. (2008) A comparison of sapphire bearing and
non-sapphire bearing lenses from the Lake Harbour
Group, south-eastern Baffin Island. B.Sc. Thesis, Carleton
University, Ottawa, Ontario, Canada, 61 pp.
HARRISON, T.M., CE
´LE
´RIER, J., AIKMAN, A.B., HERMANN, J., &
HEIZLER, M.T. (2009) Diffusion of
40
Ar in muscovite.
Geochimica et Cosmochimica Acta 73, 1039–1051.
HERD, C.D.K., PETERSON, R.C., & ROSSMAN, G.R. (2000)
Violet-colored diopside from southern Baffin Island,
Nunavut, Canada. Canadian Mineralogist 38, 1193–1199.
HOGARTH, D.D. & GRIFFIN, W.L. (1978) Lapis lazuli from
Baffin Island – a Precambrian meta-evaporite. Lithos 11,
37–60.
LECHEMINANT, A.N., GROAT, L.A., MORTENSEN,J.K.,
GERTZBEIN, P., & ROHTERT, W. (2005) Sapphires from
Kimmirut, Baffin Island, Nunavut, Canada. Geochimica et
Cosmochimica Acta 69, supplement 280 (Abstracts of the
15
th
annual V.M. Goldschmidt conference).
LEEMAN, W.P. & TONARINI, S. (2001) Boron isotopic analysis
of proposed borosilicate mineral reference samples.
Geostandards Newsletter 25, 399–403.
LEPAGE,L.&DAVISON, J.G. (2007) Exploration advances with
ultraviolet LED technology; Beluga sapphire, Nunavut.
Abstracts with program, Geological Association of
Canada – Mineralogical Association of Canada joint
annual meeting, Yellowknife, Northwest Territories,
2007.
LUDWIG, K.R. (2003) Isoplot 3.09: A geochronological toolkit
for Microsoft Excel. Berkeley Geochronology Center
(bgc.org), Special Publication 4.
LUDWIG, T., MARSCHALL, H.R., POGGE VON STRANDMAN N,
P.A.E., SHABAGA, B.M., FAYEK, M., & HAWTHORNE, F.C.
(2011) A secondary ion mass spectrometry (SIMS) re-
evaluation of B and Li isotopic compositions of Cu-
bearing elbaite from three global localities. Mineralogical
Magazine 75, 2485–2494.
MACGREGOR, J.R., GREW, E.S., DEHOOG, J.C.M., HARLEY,
S.L., KOWALSKI, P.M., YATES, M.G., & CARSON, C.J.
(2013) Boron isotopic composition of tourmaline, pris-
matine, and grandidierite from granulite facies para-
gneisses in the Larsemann Hills, Prydz Bay, East
Antarctica: Evidence for a non-marine evaporite source.
Geochimica et Cosmochimica Acta, http://dx.doi.org/10.
1016/j.gca.2013.05.030
MERLET, C. (1994) An accurate computer correction program
for quantitative electron probe microanalysis. Micro-
chimica Acta 114, 363–376.
MOINE, B., SAUVAN, P., & JAROUSSE, J. (1981) Geochemistry of
evaporite-bearing series: a tentative guide for the
identification of metaevaporites. Contributions to Miner-
alogy and Petrology 76, 401–412.
MO
¨LLER, A., ´
OBRIEN, P.J., KENNEDY, A., & KRO
¨NER, A. (2003)
The use and abuse of Th-U ratios in the interpretation of
zircon. EGS-AGU-EUG joint assembly, Nice, France,
April 6–11, 2003, Abstracts with program, abstract
#12113.
MORA, C.I. & VALLEY, J.W. (1989) Halogen-rich scapolite and
biotite: Implications for metamorphic fluid-rock interac-
tion. American Mineralogist 74, 721–737.
MORTENSEN, J.K., GHOSH, D., & FERRI, F. (1995) U-Pb age
constraints of intrusive rocks associated with copper-gold
porphyry deposits in the Canadian Cordillera. In Porphyry
deposits of the northwestern Cordillera of North America
(T.G. Schroeter, ed.). Canadian Institute of Mining and
Metallurgy Special Volume 46, 142–158.
NOCKOLDS, S.R. (1954) Average chemical compositions of
some igneous rocks. Bulletin of the Geological Society of
America 65, 1007–1032.
PALMER, M.R. (1991) Boron isotope systematics of hydro-
thermal fluids and tourmalines: A synthesis. Chemical
Geology (Isotope Geoscience Section) 94, 111–121.
POUCHOU, J.L. & PICHOIR, F. (1985) PAP u(qZ) procedure for
improved quantitative microanalysis. Microbeam Analysis
Proceedings (J.T. Armstrong, ed.). San Francisco Press,
California, United States (104–106).
RENNE, P.R., SWISHER, C.C., DEINO, A.L., KARNER, D.B.,
OWENS, T.L., & DEPAOLO, D.J. (1998) Intercalibration of
standards, absolute ages and uncertainties in
40
Ar/
39
Ar
dating. Chemical Geology 145, 117–152.
ROBBINS, G.A. (1972) Radiogenic argon diffusion in musco-
vite under hydrothermal conditions. Master’s Thesis,
Brown University, Providence, Rhode Island.
RODDICK, J.C. (1987) Generalized numerical error analysis
with application to geochronology and thermodynamics.
Geochimica et Cosmochimica Acta 51, 2129–2135.
RUBATTO, D., WILLIAMS, I.S., & GU
¨NTHER, D. (1999) Trace-
element characterization of metamorphic zircons. 9
th
Annual V.M. Goldschmidt conference, Abstract 7111,
L.P.I. Contribution 971, Lunar and Planetary Institute,
Houston, Texas.
698 THE CANADIAN MINERALOGIST

SCOTT, D.J. (1997) Geology, U-Pb, and Pb-Pb geochronology
of the Lake Harbour area, southern Baffin Island:
implications for the Paleoproterozoic tectonic evolution
of northeastern Laurentia. Canadian Journal of Earth
Sciences 34, 140–155.
SCOTT, D.J., STERN, R.A., ST-ONGE, M.R., & MCMULLEN, S.M.
(2002) U-Pb geochronology of detrital zircons in
metasedimentary rocks from southern Baffin Island:
implications for the Paleoproterozoic tectonic evolution
of Northeastern Laurentia. Canadian Journal of Earth
Sciences 39, 611–623.
SHAW, T.J., GIESKES, J.M., & JAHNKE, R.A. (1990) Early
diagenesis in differing depositional environments: the
response of transition metals in pore water. Geochimica et
Cosmochimica Acta 54, 1233–1246.
SHOR,R.&WELDON , R. (2009) Ruby and sapphire production
and distribution: a quarter century of change. Gems &
Gemology 45, 236–259.
SIMONET, C., FRITSCH, E., & LASNIER, B. (2008) A classifica-
tion of gem corundum deposits aimed towards gem
exploration. Ore Geology Reviews 34, 127–133.
ST-ONGE,M.R.,WODICKA,N.,&LUCAS, S.B. (2000)
Granulite- and amphibolite-facies metamorphism in a
convergent-plate-margin setting: synthesis of the Quebec-
Baffin segment of the Trans-Hudson Orogen. Canadian
Mineralogist 38, 379–398.
ST-ONGE, M.R., SCOTT, D.J., & WODICKA, N. (2001) Geology,
Meta Incognita Peninsula, Baffin Island, Nunavut.
Geological Survey of Canada, Map 2009A.
ST-ONGE, M.R., WODICKA,N.,&IJEWLIW, O. (2007)
Polymetamorphic evolution of the Trans-Hudson Orogen,
Baffin Island, Canada: Integration of petrological, struc-
tural and geochronological data. Journal of Petrology 48,
271–302.
ST-ONGE, M.R., VAN GOOL, J.A.M., GARDE, A.A., & SCOTT,
D.J. (2009) Correlation of Archaean and Palaeoproter-
ozoic units between northeastern Canada and western
Greenland: constraining the pre-collisional upper plate
accretionary history of the Trans-Hudson orogeny.
Geological Society of London Special Publications 318,
193–235.
STACEY, J.S. & KRAMER, J.D. (1975) Approximation of
terrestrial lead isotope evolution by a two stage model.
Earth and Planetary Science Letters 26, 207–221.
STEIGER,R.H.&JA
¨GER, E. (1977) Subcommission on
geochronology: convention on the use of decay constants
in geo- and cosmochronology. Earth and Planetary
Science Letters 36, 359–362.
SWIHART, G.H. & MOORE, P.B. (1989) A reconnaissance of the
boron isotopic composition of tourmaline. Geochimica et
Cosmochimica Acta 53, 911–916.
TAKAHASHI, Y., HAYASAKA, Y., MORITA, K., KASHIWABARA, T.,
NAKADA, R., MARCUS, M.A., KATO, K., TANAKA, K., &
SHIMIZU, H. (2015) Transfer of rare earth elements (REE)
from manganese oxides to phosphates during early
diagenesis in pelagic sediments inferred from REE
patterns, X-ray absorption spectroscopy, and chemical
leaching method. Geochemical Journal 49, 653–674.
THE
´RIAULT, R.J., ST-ONGE, M.R., & SCOTT, D.J. (2001) Nd
isotopic and geochemical signature of the Paleoproter-
ozoic Trans-Hudson Orogen, southern Baffin Island,
Canada: implications for the evolution of eastern
Laurentia. Precambrian Research 108, 113–138.
TRUE NORTH GEMS INC. (2007) True North Gems recovers 139
carat sapphire from new Baffin Island discovery. Press
release, December 17, 2007: http://www.marketwired.
com/press-release/true-north-gems-recovers-139-carat-
sapphire-from-new-baffin-island-discovery-tsx-venture-
tgx-803810.htm [Retrieved December 20, 2016].
TURNER, D.J., GROAT, L.A., RIVARD, B., FENG, J., DZIKOWSKI,
T., & BELLEY, P.M. (2014) Hyperspectral imaging of
marble-hosted sapphire from the Beluga occurrence,
Baffin Island, Nunavut, Canada. Abstracts with program,
Geological Society of America annual meeting, Vancou-
ver, British Columbia, Canada, 2014.
TUTTLE, M.L., KLETT, T.R., RICHARDSON, M., & BREIT, G.N.
(2000) Geochemistry of two interbeds in the Pennsylva-
nian Paradox Formation, Utah and Colorado – A record of
deposition and diagenesis of repetitive cycles in a marine
basin. United States Geological Survey Bulletin 2000-N,
86 pp.
VAN HINSBERG, V.J., HENRY, D.J., & MARSCHALL, H.R. (2011)
Tourmaline: an ideal indicator of its host environment.
Canadian Mineralogist 49, 1–16.
VILLA, I.M. (2010) Disequilibrium textures versus equilib-
rium modelling: geochronology at the crossroads. In:
Advances in interpretation of geological processes:
refinement of multi-scale data and integration in
numerical modelling (M.I. Spalla, A.M. Marotta, & G.
Gosso, eds.). Ge ological Society of London Special
Publication 332, 1–15.
WENGER, L.M. & BAKER, D.R. (1986) Variations in organic
geochemistry of anoxic-oxic black shale-carbonate se-
quences in the Pennsylvanian of the Midcontinent, U.S.A.
Advances in Organic Geochemistry 10, 85–92.
WILSON, B.S. (2014) Colored gemstones from Canada. In
Geology of Gem Deposits (L. Groat, ed.). Mineralogical
Association of Canada Short Course 44, 375–405.
Received February 7, 2017. Revised manuscript accepted
June 7, 2017.
ORIGIN OF SAPPHIRE NEAR KIMMIRUT, NUNAVUT 699