ArticlePDF Available

Clinoferrogedrite in the contact-metamorphosed Biwabik Iron Formation, Northeastern Minnesota


Abstract and Figures

We describe a sample of Proterozoic banded iron-formation from the inner contact aureole of the Duluth complex near Babbitt, Minnesota, that contains a compositionally unique Fe-rich monoclinic ferromagnesian amphibole. Its content of TAl (1.3 atoms per 23 anhydrous oxygens) and ANa (0.55) places it in a compositional field (with the label clinoferrogedrite) outside those recognized in the IMA nomenclature for amphiboles. With declining Al, it grades into Al-rich grunerite, and finally to Al-poor grunerite at the crystal margins. While spot-to-spot compositional trends define basic gedrite substitutions, they differ systematically from the well-established anthophyllite-gedrite trends found in amphibolites. They resemble instead two instances of sodic gedrite described in the literature that are enriched in ANa apfu and depleted in octahedral Al. These unusual trends may be the result of very high temperatures of formation, or a high activity ratio of Fe + Mg to Si. Associated minerals in our sample include orthopyroxene Fs74-76, olivine Fa85-91, almandine, quartz, plagioclase, and accessories graphite, pyrrhotite, ilmenite, biotite, apatite, zircon, monazite, and löllingite. Conditions are estimated to have been ≥800 °C and ≈1.3 kbar, with fO2 close to FMQ-2 log10 units. Clinogedrite should be considered extremely rare, but not non-existent. In the monoclinic structure the symmetry of the double Si-O chains limits the uptake of gedritic components to amounts lower than found in the orthorhombic structure, where Al is distributed across three of the four distinct tetrahedral sites. We also present electron microprobe data for a highly aluminous cummingtonite grading compositionally into clinogedrite (ANa = 0.3, M2Al + Fe3+ + 2Ti = 1.0, TAl = 1.1) that was grown from fused and S-augmented Pinatubo dacite in the laboratory at 780 °C, 2.2 and 3.9 kbar. Growth of this MgFe-clinoamphibole was enabled by the capture of Ca into anhydrite away from potential hornblende.
Content may be subject to copyright.
§ Corresponding author e-mail address:
The Canadian Mineralogist
Vol. 52, pp. 533-554 (2014)
DOI : 10.3749/canmin.52.3.533
Brian r. JOY§
Department of Geological Sciences and Geological Engineering, Queen’s University, Kingston, Ontario, K7L 3N6, Canada
Bernard W. eVanS
Department of Earth and Space Sciences, Box 351310, University of Washington, Seattle 98195-1310, USA
We describe a sample of Proterozoic banded iron-formation from the inner contact aureole of the Duluth complex near
Babbitt, Minnesota, that contains a compositionally unique Fe-rich monoclinic ferromagnesian amphibole. Its content of TAl
(1.3 atoms per 23 anhydrous oxygens) and ANa (0.55) places it in a compositional eld (with the label clinoferrogedrite) outside
those recognized in the IMA nomenclature for amphiboles. With declining Al, it grades into Al-rich grunerite, and nally to
Al-poor grunerite at the crystal margins. While spot-to-spot compositional trends dene basic gedrite substitutions, they differ
systematically from the well-established anthophyllite-gedrite trends found in amphibolites. They resemble instead two instances
of sodic gedrite described in the literature that are enriched in ANa apfu and depleted in octahedral Al. These unusual trends may
be the result of very high temperatures of formation, or a high activity ratio of Fe + Mg to Si. Associated minerals in our sample
include orthopyroxene Fs74–76, olivine Fa85–91, almandine, quartz, plagioclase, and accessories graphite, pyrrhotite, ilmenite,
biotite, apatite, zircon, monazite, and löllingite. Conditions are estimated to have been ≥800 ºC and ≈1.3 kbar, with fO2 close
to FMQ-2 log10 units. Clinogedrite should be considered extremely rare, but not non-existent. In the monoclinic structure the
symmetry of the double Si–O chains limits the uptake of gedritic components to amounts lower than found in the orthorhombic
structure, where Al is distributed across three of the four distinct tetrahedral sites. We also present electron microprobe data for
a highly aluminous cummingtonite grading compositionally into clinogedrite (ANa = 0.3, M2Al + Fe3+ + 2Ti = 1.0, TAl = 1.1)
that was grown from fused and S-augmented Pinatubo dacite in the laboratory at 780 ºC, 2.2 and 3.9 kbar. Growth of this MgFe-
clinoamphibole was enabled by the capture of Ca into anhydrite away from potential hornblende.
Keywords: clinoferrogedrite, gedrite, clinogedrite, Al-cummingtonite, Al-grunerite, BIF, Duluth complex aureole
Natural orthorhombic ferromagnesian amphiboles
(space group Pnma, or very rarely protoanthophyllite
Pnmn, Konishi et al. 2002, 2003) extend in composition
from Al-free anthophyllite (Si = 8 apfu) through alumi-
nous anthophyllite to a gedrite endmember with Si ≤ 6
apfu (Rabbitt 1948, Robinson et al. 1971, Hawthorne
et al. 2008). The addition of Al and Na to endmember
anthophyllite is achieved via a combination of two
heterovalent substitutions at three crystallographic sites
(Robinson et al. 1971, Hawthorne et al. 2008):
A + TSi ANa + TAl (1)
M2Mg + TSi M2Al + TAl (2)
These two exchanges are commonly referred to in the
petrological literature as edenite (ed) and tschermakite
(ts), respectively (mineral abbreviations after Whitney
& Evans 2010); the two cations at the M2 site, and
three at M1 and M3, are collectively C-cations in the
formula unit. The natural monoclinic counterparts of
orthorhombic MgFe-amphibole, cummingtonite (C2/m
or P21/m) and grunerite (C2/m), appear on the other
hand to accommodate very little Al, as reected for
example in the amphibole classication of Leake et al.
(1997). Volcanic cummingtonite, which forms in silicic
and intermediate volcanic rocks between about 650
and 800 ˚C, contains at most 0.5 TAl and 0.2 MAl apfu.
Cummingtonite in amphibolites, metaperidotites, and
quartzofeldspathic schists and gneisses, and grunerite in
meta-ironstones, are mostly less aluminous than those in
534 the canadian mineralOgiSt
volcanic rocks (Deer et al. 1997), although two samples
from Sutherland, Scotland, with Al2O3 contents of
5.02 and 8.65 wt.% (TAl = 0.64 and 0.87, respectively,
Collins 1942) appear to be exceptional; these merit
further examination if at all possible. The compositions
of coexisting cummingtonite-grunerite and orthoamphi-
bole (Al-anthophyllite and gedrite) clearly demonstrate
the much greater capacity of orthorhombic amphiboles
to take up Al (Robinson et al. 1982).
Monoclinic MgFe-amphiboles encompass almost the
entire compositional range from 90% Mg-endmember
to 100% Fe-endmember, whereas, with rare exceptions
(Ferré 1989, Bozhilov & Evans 2001), anthophyllite
seems to be limited generally to no more than 50%
Fe-endmember. On the other hand, several occurrences
of ferrogedrite have been reported (Deer et al. 1997).
Given the capacity of orthoamphibole over MgFe-
clinoamphibole to accommodate Al, we believe that
nding cummingtonite and grunerite with substantial
ANa, and TAl approaching and exceeding 1.0 apfu,
as described in this work, deserves comment. The
latter compositions are logically called clinogedrite
and clinoferrogedrite if we follow the IMA amphibole
nomenclature and conventions of Leake et al. (1997,
2003). Here we seek answers to whether these hitherto
unheard-of compositions are the result of unusually high
temperature, unusual bulk compositions, or some other
phase-equilibrium control. In support of our proposed
extension of the compositional eld of the monoclinic
MgFe-amphiboles, we review here some recent experi-
mental work on fused, S-augmented samples of Pina-
tubo dacite. For reasons given below, we choose to use
the nomenclature of Leake et al. (1997) in preference
to that of Hawthorne et al. (2012).
While this paper is not intended to dispel the notion
of systematic crystal-chemical differences between
orthorhombic and monoclinic MgFe-amphiboles, it is
hoped that by documenting two exceptional cases we
may be able to throw further light on what controls their
occurrence and compositions.
BiWaBik irOn FOrmatiOn
Our clinoferrogedrite-bearing rock (sample 98D14-
150) is from the early Proterozoic Biwabik Banded
Iron Formation (BIF), which was intruded and ther-
mally metamorphosed ca. 1.1 Ga by the mac Duluth
Complex. The sample comes from sub-member Q,
which is also known as the “Intermediate Slate”
(Gundersen & Schwartz 1962, Bonnichsen 1968, 1975).
The sample was collected in situ in August, 1998, near
the oor of the now-ooded Dunka Pit, which was
once part of the extensive open-pit mining operations
in the Mesabi Range, and is located roughly 7 km east
of Babbitt, Minnesota. The exact distance of the collec-
tion site from the contact with the Duluth Complex
was difcult to determine in the eld, but it is certainly
no more than 100 m. In addition to clinoferrogedrite,
the sample contains orthopyroxene, fayalitic olivine,
almandine, quartz, plagioclase, graphite, pyrrhotite,
ilmenite, grunerite, biotite, apatite, zircon, monazite,
and löllingite. The mineralogy is basically that of the
pyroxene hornfels facies. Magnetite and greenalite also
occur in the rock as low-temperature alteration minerals.
The whole-rock composition of sample 98D14-150 has
been estimated by integration of electron-beam scans
(Table 1). Bonnichsen (1968) described a mineralogi-
cally similar rock (sample M-12088) from sub-member
Q in drill core taken roughly 900 m south of our
sample. Bonnichsen’s sample contains a pale brown
clinoamphibole with composition approaching that of
clinoferrogedrite (reported wt.% Al2O3 ranges up to
7.0%); it also contains garnet, but is lacking in quartz,
plagioclase, graphite, and biotite.
Within the aureole of the Duluth Complex, orthopy-
roxene-bearing rocks occur in the Biwabik Formation
Vol.% Density
Composition (wt.%)
SiO2Al2O3TiO2FeO MnO MgO CaO BaO Na2O K2O F Cl C S Total
Opx 44.4 3.8 48.25 1.18 0.14 41.62 1.35 8.00 0.36
Ol 23.3 4.3 30.82 64.18 1.68 4.77 0.03
Grt 11.5 4.2 37.27 21.05 0.02 35.09 3.44 2.28 1.94
Qz 9.8 2.65 100.00
Pl 7.3 2.7 53.69 29.84 0.22 11.65 4.96 0.09
Gr 3.1 2.23 100.00
Po 0.3 4.6 79.13 39.12
Ilm 0.3 4.7 51.55 46.39 1.21 0.12
Cfg/Gru 0.2 3.5
Bt 0.05 3.0 35.48 13.67 2.67 24.96 0.13 9.18 0.01 1.08 0.33 8.61 1.56 1.10
Whole rock: 44.35 4.80 0.26 40.74 1.50 5.15 1.03 0.00 0.26 0.01 0.00 0.00 1.84 0.14 100.00
to a map distance as great as ~2 km from the intrusive
contact (Gundersen & Schwartz 1962, French 1968,
Joy 2009). The Dunka Pit lies entirely within the
orthopyroxene zone. Although the mineralogy of the
Biwabik Formation in the orthopyroxene zone varies
between “cherty” and “slaty” sub-members (Gundersen
& Schwartz 1962, Bonnichsen 1968, 1975), it is gener-
ally dominated by quartz and magnetite, and in addition
contains varying proportions of olivine, orthopyroxene,
and/or calcic clinopyroxene. In some rocks from the
Dunka Pit, orthopyroxene contains coarse blebs of
clinopyroxene that are interpreted to have formed from
inverted pigeonite (Bonnichsen 1968, 1969, 1975).
Grunerite and hornblende (sensu lato) are abundant and
are texturally late. Where both amphiboles occur in the
same rock, the grunerite forms rims around the horn-
blende and thus postdates it; both minerals commonly
contain exsolution lamellae of the other. In comparison
to typical orthopyroxene-zone rocks in the aureole,
sample 98D14-150 is mineralogically unusual. Alman-
dine-rich garnet, biotite, ilmenite, zircon, and monazite
are generally absent from the Biwabik Formation, and
are limited exclusively to the Intermediate Slate (sub-
member Q). Compared to the average Biwabik BIF, the
Intermediate Slate contains less SiO2, Fe2O3, and CaO,
and more Al2O3, TiO2, FeO, MgO, Zr, and V. The high
whole-rock Al content and relative abundances of Ti,
Zr, V, and rare earth elements have been attributed to
a pyroclastic component (Morey 1992). The presence
of graphite and absence of prograde magnetite indicate
unusually reducing conditions. If it is assumed that
the molar O:H ratio of the pore uid was 1:2, as is
consistent with uid production due to dehydration
reactions (Connolly & Cesare 1993), then the oxygen
fugacity was between 1.5 and 2.3 log10 units below the
fayalite-magnetite-quartz (FMQ) buffer for tempera-
tures between 600 and 800 °C.
The pressure and peak temperature of metamor-
phism are difcult to quantify petrologically owing
to extensive retrograde mineral growth, mineral
compositional zoning, and resetting of mineral compo-
sitions during cooling. Application of the orthopy-
roxene-garnet-plagioclase-quartz thermobarometer of
Pattison et al. (2003) is problematic due to the fact
that plagioclase zoning is truncated, and plagioclase of
varying composition lies in contact with orthopyroxene
of essentially uniform composition (Joy 2009). The
presence of inverted pigeonite in some rocks in the
Dunka Pit implies a temperature in excess of ~820 °C,
at least locally (Davidson & Lindsley 1989), and it is
reasonable to assume that peak temperature experienced
by the sample under study was in the vicinity of 800
°C, if not somewhat greater. Textural relations between
clinoferrogedrite and the anhydrous ferromagnesian
silicates have been obscured by the growth of clearly
retrograde grunerite. Thus, it is not clear whether some
or all of the clinoferrogedrite was present at peak T or
if it grew solely during retrograde metamorphism, as
discussed below.
Insight into the pressure of metamorphism may be
gained by examination of the assemblage olivine +
orthopyroxene + calcic clinopyroxene + quartz, which
occurs sporadically within the Biwabik Formation. In
most cases, mineral compositions are described almost
completely with components SiO2, FeO, MgO, and
CaO; MnO is generally minor, and pyroxene charge
balance constraints indicate that Fe2O3 is minor as
well. Noting that the four-phase assemblage (Fig. 1a)
is isobarically univariant, Joy (2009) demonstrated that
the compositions of the ferromagnesian minerals in this
assemblage are generally not representative of a frozen
equilibrium. This can be seen when compositions of
orthopyroxene from inner aureole rocks are plotted
in a diagram (Fig. 1b) with superimposed contours
for temperature and pressure according to QUILF
(equilibria among Fe-Mg-Mn-Ti oxides, pyroxenes,
olivine, and quartz, Andersen et al. 1993), because their
compositions do not conform to an isobar. Core-to-rim
zoning in orthopyroxene is retrograde in origin and is
characteristic either of growth with falling tempera-
ture or of diffusion-induced cation exchange. Various
processes can account for deviation of orthopyroxene
from equilibrium composition with falling temperature,
and these processes are illustrated schematically in
Figure 1c. Orthopyroxene in the four-phase assemblage
is the most Fe- and Ca-rich orthopyroxene that can exist
in equilibrium at given pressure and temperature in the
SiO2–FeO–MgO–CaO system (Fig. 1c, heavy black
line). Since the analyzed grains are characterized by
increasing molar Fe/(Fe + Mg) progressing outward
from interiors, the only process that can produce
orthopyroxene more ferroan for a given molar Ca is
intragranular Fe-Mg interdiffusion. Considering this
fact, points lying on the true isobar must plot on the
Figure some distance to the left of the most Fe-rich
analyses for a given molar Ca.
Orthopyroxene grains most likely representative of
equilibrium in the four-phase assemblage are those that
grew relatively quickly with at compositional proles
and that do not lie adjacent to other ferromagnesian
minerals. Grains best meeting these criteria are present
in pegmatitic feldspar-bearing patches in one sample
from the Northshore Mine described by Joy (2009).
Application of the QUILF model suggests nucleation
(or homogenization) of one such grain (OPX203) at
a temperature of 780 °C and a pressure of 1300 bar
(Fig. 1b). The slope on the plot dened by composi-
tions outward from the grain interior roughly parallels
the isobar, although near-rim compositions deviate
leftward from it due to late growth or diffusion while
orthopyroxene was no longer at equilibrium in the four-
phase assemblage. Thus, 1300 bar is considered to be
the best estimate of metamorphic pressure; uncertainty
due to analytical precision is no more than 200 bar. This
536 the canadian mineralOgiSt
assessment of pressure is consistent with the sequence
of mineral assemblages observed in overlying pelites
of the Virginia Formation (Andrews & Ripley 1989),
which conform to facies series 1a of Pattison & Tracy
(1991). In their study of the correlative Rove Formation
~100 km to the northeast, Labotka et al. (1981) deter-
mined a similar pressure of metamorphism.
PetrOgraPhY and mineralOgY
OF SamPle 98d14-150
The sample displays a weak foliation dened by
variations in mineralogy and preferred orientation of
graphite grains. Bands of coarse orthopyroxene domi-
nate on a centimeter-scale, and the mineral occurs as
ellipsoidal poikiloblasts (Fig. 2a) measuring as much
as 8 mm across; inclusions within the orthopyroxene
comprise all the other minerals in the rock. Thinner
bands in the rock are dominated alternately by ner-
grained quartz ± plagioclase, olivine + quartz +
plagioclase, olivine + garnet + graphite + ilmenite
(without quartz), and pyrrhotite + chalcopyrite + garnet
+ ilmenite + graphite. Graphite is dispersed throughout
the rock (Fig. 2b) and locally dominates in thin, discon-
tinuous bands. Hornblende is not present in sample
98D14-150, although it is common (as ferro-edenite and
ferro-pargasite) in nearby samples of the Intermediate
Slate. Olivine typically displays polygonal outlines
and is separated from quartz by rims of orthopyroxene.
Equidimensional, xenoblastic garnet contains inclusions
of all the other minerals save orthopyroxene. Grain sizes
of olivine, garnet, quartz, and plagioclase rarely exceed
0.5 mm. Lenses of intergrown quartz and plagioclase
measure as much as 2.5 cm across and contain unusually
coarse grains as large as 3 mm. In some cases, coarse
pyrrhotite is interstitial to locally idiomorphic crystals
of quartz and plagioclase. These features suggest the
possible coexistence at some stage of silicate and sulde
melts. Late veinlets of greenalite + magnetite produce
an apparent foliation in the rock that is parallel to
Fig. 1. (a) Phase relations in the pyroxene quadrilateral
calculated using QUILF (Andersen et al. 1993) for P = 1500
bar. Isobars for each mineral in the olivine + orthopyroxene +
clinopyroxene + quartz assemblage are shown as heavy black
curves. XFe represents molar Fe/(Fe+Mg), and XCa represents
molar Ca/(Ca+Fe+Mg). (b) Selected Opx analyses from a
number of samples of the Biwabik Fm also containing Ol,
Cpx, and Qz. Each color represents a single grain or traverse.
Isotherms and isobars have been calculated with QUILF. (c)
Calculated paths of change in Opx composition with falling
temperature in the SiO2–FeO–MgO–CaO system. Black: a
possible isobaric path taken by Opx in equilibrium with Ol,
Cpx, and Qz. The pigeonite terminal reaction is shown for
reference as a polybaric, polythermal curve. Red: possible
isobaric paths taken by Opx in equilibrium with Qz and either
Ol or Cpx. Green: possible isobaric Opx paths due to exchange
of Fe and Mg with adjacent Ol. Yellow: possible isobaric
Opx paths due to due exsolution of Cpx. Blue: paths taken by
retrograde growth-zoned Opx interiors due to intragranular Fe-Mg interdiffusion. The last process is the only one that can
create Opx more Fe-rich for given XCa than that produced by growth in equilibrium with Qz, Ol, and Cpx.
Fig. 2. (a) and (b) Low and high magnication backscattered electron images of micro-textures in sample 98D14-150. Figure
2b shows the area covered by the X-ray maps in Figure 3, and the location (2-1) of the compositional prole in Figure 5a. (c)
Photomicrograph in PPL of amphibole 6, zoned from brown clinoferrogedrite cores to colorless grunerite rim; high relief:
olivine and orthopyroxene; black: graphite and ilmenite; colorless: plagioclase and quartz. Figure 2c is found on page 539.
538 the canadian mineralOgiSt
compositional banding; this feature was also observed
by Bonnichsen (1968) in samples of the Intermediate
Slate. Late hydrous alteration is minimal in comparison
to typical samples of the high-grade Biwabik Formation.
Electron microprobe compositions obtained from
sample 98D14-150 reveal the orthopyroxene to be
weakly zoned, with wt.% Al2O3 varying within interiors
between ~1.0 and ~1.3 and decreasing to as low as ~0.8
within tens of microns of grain rims. Molar Fe/(Fe +
Mg) varies more regularly and ranges between 0.740
in cores and 0.755 at rims. Adjacent to quartz, Fe/(Fe
+ Mg) of matrix olivine generally varies between 0.89
and 0.91, whereas in quartz-free bands it ranges as low
as 0.85. Garnet is weakly zoned, with rims of matrix
grains slightly depleted in pyrope and grossular compo-
nents and enriched in almandine and spessartine relative
to cores; garnets included in orthopyroxene interiors
show little or no zoning. Average mole fractions of
almandine, pyrope, spessartine, and grossular are 0.777,
0.090, 0.077, 0.055, respectively; upon recalculation for
Fe3+, andradite accounts for less than one mole percent.
Plagioclase shows relatively pronounced zoning; the
zoning is mostly irregular and/or abruptly truncated, but
generally rims are more sodic than interiors; mole frac-
tion of anorthite ranges from 0.685 in interiors to 0.491
at rims. Average mole fractions of anorthite, albite, and
orthoclase are 0.563, 0.433, and 0.005, respectively.
Clinoferrogedrite (ANa = 0.5 to 0.6, TAl > 1.0) is
dispersed sparsely within the sample in vaguely dened
patches as much as 2 cm across; grains vary in size
between 0.1 and 0.5 mm and range from idioblastic to
irregular in shape. Modally, the mineral represents only
about 0.1% of the rock, it being difcult to distinguish
from grunerite in backscattered electron images. The
clinoferrogedrite occurs adjacent to a variety of other
minerals and also occurs within the patches of coarse
quartz + plagioclase. No pattern in the distribution of
the mineral is readily discernible. In isolated cases,
the amphibole occurs as an inclusion in garnet, though
it is not clear whether it was included during growth
of the garnet or if it represents an alteration product.
Although the mineral is not strongly colored in thin
section (Fig. 2c), its pleochroism is perhaps its most
distinctive feature; observation via universal stage
yields the following: X = colorless, Y = pale brown, Z =
pale brown. The optic sign is negative and 2V = 85 ± 1˚.
Birefringence is roughly 0.023, and the mineral shows
dispersion with v > r. The extinction angle Z^c was
540 the canadian mineralOgiSt
not veried with the universal stage, but a longitudinal
grain oriented to give an off-center ash gure gave
Z^c = 15–16˚. The clinoferrogedrite is almost always
rimmed by colorless grunerite and grades continuously
into it optically, as conrmed by the marginal decline
in Al and Na X-ray intensities (Figs. 3a and 3d). The
grunerite rims typically display polysynthetic twins,
but the clinoferrogedrite cores are either untwinned
or simply twinned. Grunerite rims commonly contain
irregular inclusions of olivine and other minerals, but
the clinoferrogedrite cores generally are inclusion-free,
and thus the timing of its growth relative to the anhy-
drous ferromagnesian silicates is ambiguous. The clino-
ferrogedrite possesses exsolution lamellae of a more
Ca-rich amphibole in two distinct orientations. The
lamellae are only resolved at high magnication under
the petrographic microscope, feebly in CaKα X-ray
maps (Fig. 3b), and better in backscattered electron
images (Fig. 4). Individual lamellae typically measure
no more than ~100 nm in width and are oriented close to
(100) and (101). Where both sets of lamellae are visible
in appropriately oriented prismatic section (optic plane),
the angle of intersection is 103.5 ± 1º.
chemical cOmPOSitiOn
OF clinOamPhiBOle in SamPle 98d14-150
Analytical methods
Wavelength dispersive analyses of the clinoamphi-
bole were performed using the JEOL JXA-8230 electron
microprobe housed in the Department of Geological
Sciences and Geological Engineering, Queen’s Univer-
sity, Kingston, Ontario. The accelerating potential was
15 kV and the beam current 30 nA. The beam was defo-
cused to ~3 µm in order to sample both host and several
exsolution lamellae where the latter were present (as
the lamellae were impossible to avoid completely).
Standards were diopside (Si, Ca), synthetic fayalite (Fe),
synthetic forsterite (Mg), anorthite glass (Al), albite
(Na), rutile (Ti), rhodonite (Mn), V2O5 (V), chromite
(Cr), synthetic uorophlogopite (F), and scapolite (Cl);
between nine and 11 spots were analyzed on each stan-
dard. Peak and background count times were 20 or 30
s, except for Si (10 s) and Cr (50 s); total analysis time
was roughly 260 s. Raw data were processed using the
X-PHI matrix correction program of Merlet (1994). The
analytical data are available from the Mineralogical
Association of Canada Depository of Unpublished Data
(document clinoferrogedrite CM52_533).
Both VKα and CrKα were measured using a large-
area LiF crystal, and an overlap correction was applied
in each case to correct for interference from TiKb and
VKb, respectively. In the former case, the magnitude
of the correction was minimal. In the latter case, where
wt.% V2O3 ranged between 0.20 and 0.25 and wt.%
Cr2O3 averaged ~0.04, 10–15% of net counts at the
CrKα peak position were found to be due to VKb. A
sealed Xe counter was used to measure CrKα, and it is
worth noting that the XeL1 absorption edge occurs at a
wavelength 0.00160 nm below that of the CrKα peak
position (corresponding to a difference of 1.11 mm in
diffracting crystal position). However, examination of
wavelength scans collected at high beam current within
a grain interior revealed that any discontinuity in the
continuum was small relative to the peak height, and
so Cr background was measured both at high and low
offsets on opposing sides of the absorption edge. Detec-
tion limit (3 standard deviations above background) was
calculated according to the method of Williams (1987),
but with matrix corrections applied; respective values
for TiO2, Cr2O3, V2O3, K2O, F, and Cl were ~0.025,
~0.025, ~0.035, ~0.010, ~0.15, and ~0.010 wt.%. All
other measured elements were present well in excess
of the detection limit. For analyses with wt.% Al2O3
greater than about 6.0, uncertainty due to counting error
(one standard deviation) was less than 0.5% (relative)
for Si, Fe, and Al; ~0.5% for Mg and Ca; ~1.5% for Ti,
Na, and Cl; ~3% for Mn, ~10% for V, K, and F; and
generally 20–40% for Cr.
Analytical results and structural formulae
Representative analyses spanning the range of
compositions observed in the ferromagnesian amphi-
bole are presented in Table 2. Concentric zoning is
characteristic of almost all the individual grains of
amphibole. Compositional proles for two examples,
one a composite of two grains (6a and 6b, as illustrated
in Fig. 3), are shown in Figure 5. The compositional
zoning is continuous from cores of clinoferrogedrite (Si
atoms per formula unit < 7.0) through aluminous grune-
rite to rims of grunerite (Si apfu ~7.8). In every case, the
greatest values of wt.% Al2O3, Na2O, CaO, TiO2, V2O3,
Cr2O3, and Cl, and the lowest values of SiO2, FeO, and
MgO occur in grain interiors. Al2O3 and Na2O in grain
interiors attain values as large as 9.4 and 1.9 wt.%,
respectively. Typical values of wt.% CaO, TiO2, V2O3,
and Cr2O3 are 2.4, 0.90, 0.25, and 0.04 in Al-rich inte-
riors. Levels of wt.% FeO-total increase from 33.5–34.5
in Al-rich interiors to as high as 36.5 near rims, while
wt.% MgO increases concomitantly from 5.6 to ~9.0;
thus, molar Fe/(Fe + Mg) decreases from interior to
rim, and most values fall within the range 0.77–0.69.
Grains are enriched in MgO where they lie adjacent
to olivine and where included in garnet. Interestingly,
compositional variation in the amphibole occurs in
such a manner that the electron backscatter coefcient
remains nearly constant across zoned grains, such that
the chemical variation is largely invisible in backscat-
tered electron images (e.g., Fig. 3a). The ne (<100 nm)
exsolution lamellae of calcic amphibole visible in BSE
images collected at high gain (Fig. 4) implies that each
~3 µm spot analysis represents an area that probably
sampled several lamellae in many cases. The lamellae
do not appear to be present in grain 7, which occurs as
a ~20 µm inclusion in quartz and contains ~3.2 wt.%
CaO (Table 2, analysis 76/7); this grain also contains
greater wt.% K2O (albeit small, ~0.15) than any other
analyzed grain. The grain may have nucleated at high
temperature and did not experience exsolution, or it may
be the product of coalescence of lamellae that escaped
from a nearby, larger grain. Although it is clear that the
exsolution lamellae are rich in Ca relative to the host,
their small width precludes precise characterization
a b
c d
Fig. 3. X-ray maps of zoned amphibole 6 and amphibole 7 (20 µm grain in lower left) in sample 98D14-150. (a) AlKα (b)
CaKα, (c) MgKα, (d) NaKα. In Figure 3b, essentially vertical exsolution lamellae near {100} are visible in amphibole grain
6b (upper left) and roughly bisect the acute angle between the {110} prismatic cleavages. The cleavage traces running from
upper right to lower left give the appearance of lamellae due to their orientation relative to the X-ray detector.
542 the canadian mineralOgiSt
Fig. 4. Backscattered electron image showing exsolution of Ca-amphibole in clinoferrogedrite (amphibole grain 9, Table 2).
Circles with inner diameter corresponding to the spot size show analysis locations. The inset displays the compositional
prole along the line dened by the illustrated analysis spots.
of their composition by conventional electron probe
analysis (cf. Smelik & Veblen 1992).
Although it is implicit in application of the matrix
corrections that the analyzed material is homogeneous
on a submicroscopic scale, the nearly ubiquitous nature
of the exsolution lamellae renders them impossible to
avoid, and use of a defocused beam is the only recourse.
In grain 9, which was oriented to give an off-center
ash gure, lamellae near both {100} and {101} are
oriented roughly normal to the plane of section (Fig. 4).
The rst set of lamellae is parallel to the traces of the
{110} cleavage. Since the lamellae in this case are
clearly visible and are oriented at high angle to the
plane of section, one may test whether uneven sampling
Fig. 5. (a) and (b) Composition proles across zoned amphibole grains 6 and 13 (Table
2) in sample 98D14-150.
544 the canadian mineralOgiSt
of lamellae has inuenced measured mineral composi-
tions. Since the lamellae might be expected to consist of
a calcic, aluminous amphibole in which both the edenite
and tschermakite components are prominent, analyses
that include them preferentially could be enriched in
Ca, Al, and Na. In fact, one might even argue that the
most aluminous and sodic analysis results are simply
due to preferential sampling of these lamellae. However,
in grain 9, two of the three most Al-rich compositions
(e.g., Table 2, analysis 156/9) are located in a portion
of the grain (Fig. 4) in which no lamellae are visible
(though this does not rule out the presence of submi-
croscopic lamellae). In only one case (analysis 153/9)
in which exsolution lamellae are prominent does wt.%
CaO appear anomalous (2.9 wt.%); at this location,
wt.% Al2O3 and Na2O are somewhat below interior
plateau values (Fig. 4). Grain 6b, a prole of which
is displayed in Figure 5b, was oriented to produce a
nearly centered Bxo gure; in this orientation, the {100}
lamellae are vertical. If the lamellae were sampled
unevenly, then one should observe abrupt variation in
the compositional prole, particularly within the grain
interior. However, no such erratic variation is observed.
Further, in the set of X-ray maps displayed in Figure 3,
98D14-150 PIN69
Analysis/grain 1/6a 2/6a 15/6b 82/13 76/7 174/1 156/9 107/13 34/6b 24 32 39 102 108
SiO2 (wt.%) 50.20 48.17 46.28 44.39 42.96 43.09 42.24 42.58 42.51 50.46 50.07 49.62 49.86 48.46
Al2O31.42 3.39 5.22 7.08 8.37 8.82 9.08 9.29 9.40 7.68 8.27 8.52 9.58 9.94
TiO20.08 0.30 0.46 0.67 0.86 0.93 0.88 0.94 0.90 0.68 0.74 0.83 0.76 0.82
Cr2O30.00 0.00 0.01 0.01 0.03 0.02 0.04 0.04 0.04
V2O30.02 0.05 0.05 0.13 0.27 0.27 0.24 0.25 0.23
Fe2O3* 0.60 0.60 0.58 0.57 0.55 0.55 0.55 0.56 0.56 1.57 1.61 1.56 1.68 1.66
FeO 35.75 35.41 34.29 33.55 32.48 32.36 32.75 32.80 33.01 12.73 13.05 12.65 13.62 13.48
MnO 0.91 0.97 0.94 1.02 0.95 1.07 1.11 1.08 1.02 0.68 0.69 0.82 1.00 1.06
MgO 8.74 8.08 7.46 6.75 5.67 6.62 6.23 6.03 5.75 19.90 18.80 18.75 18.85 18.37
CaO 0.49 0.85 1.41 1.91 3.26 2.23 2.31 2.31 2.58 1.91 2.11 2.97 2.77 2.98
Na2O 0.35 0.76 1.05 1.44 1.48 1.75 1.80 1.77 1.77 0.82 1.02 1.15 1.10 1.07
K2O 0.01 0.01 0.03 0.04 0.15 0.06 0.06 0.05 0.06 0.03 0.07 0.00 0.07 0.06
F 0.43 0.35 0.58 0.58 0.52 0.55 0.54 0.49 0.57
Cl 0.20 0.21 0.29 0.40 0.44 0.46 0.47 0.49 0.51
H2O** 1.68 1.70 1.56 1.52 1.52 1.53 1.51 1.54 1.50 2.10 2.09 2.10 2.14 2.10
O = F, Cl 0.23 0.20 0.31 0.33 0.32 0.34 0.33 0.32 0.35
Total 100.64 100.66 99.91 99.71 99.20 99.97 99.49 99.90 100.06 98.56 98.52 98.97 101.42 99.99
apfu relative to 23 anhydrous oxygen atoms
Si (apfu) 7.79 7.50 7.27 7.01 6.84 6.78 6.71 6.73 6.72 7.21 7.17 7.10 6.99 6.91
TAl 0.21 0.50 0.73 0.99 1.16 1.22 1.29 1.27 1.28 0.79 0.83 0.90 1.01 1.09
∑T 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00
M2Al 0.05 0.13 0.24 0.33 0.41 0.42 0.42 0.46 0.47 0.50 0.57 0.53 0.57 0.57
Ti 0.01 0.03 0.05 0.08 0.10 0.11 0.11 0.11 0.11 0.07 0.08 0.09 0.08 0.09
Cr 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01
V 0.00 0.01 0.01 0.02 0.03 0.03 0.03 0.03 0.03
Fe3+ 0.07 0.07 0.07 0.07 0.07 0.06 0.07 0.07 0.07 0.17 0.17 0.17 0.18 0.18
Fe2+ 4.64 4.61 4.51 4.43 4.33 4.26 4.35 4.34 4.36 1.52 1.56 1.51 1.60 1.61
Mn 0.12 0.13 0.13 0.14 0.13 0.14 0.15 0.14 0.14 0.08 0.08 0.10 0.12 0.13
Mg 2.02 1.88 1.75 1.59 1.35 1.55 1.48 1.42 1.35 4.24 4.02 4.00 3.94 3.90
Ca 0.08 0.14 0.24 0.32 0.56 0.38 0.39 0.39 0.44 0.29 0.32 0.46 0.42 0.45
M4Na 0.01 0.00 0.01 0.02 0.03 0.04 0.00 0.03 0.03 0.13 0.19 0.15 0.10 0.07
∑M 7.00 7.00 7.00 7.00 7.00 7.00 7.00 7.00 7.00 7.00 7.00 7.00 7.00 7.00
Fe2+/(Fe2++Mg) 0.697 0.711 0.721 0.736 0.763 0.733 0.747 0.753 0.763 0.264 0.280 0.275 0.289 0.292
ANa 0.09 0.23 0.31 0.42 0.43 0.50 0.55 0.51 0.51 0.10 0.10 0.17 0.19 0.23
K 0.00 0.00 0.01 0.01 0.03 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.01 0.01
∑A 0.09 0.23 0.31 0.43 0.46 0.51 0.56 0.52 0.52 0.11 0.11 0.17 0.21 0.24
*Fe3+ = 1.5 mol.% of total Fe for 98D14-150 and 10 mol.% of total Fe for PIN69, **OH calculated as 2 – F – Cl apfu for for 98D14-150
exsolution lamellae in grain 6b are obvious in the CaKα
image, but are not even faintly visible in the AlKα
image. This fact suggests that the major compositional
difference between host and lamellae is their Ca content
and that the two do not differ appreciably in Al content.
Differences in Na content between host and lamellae are
difcult to assess due to lower X-ray intensity.
Estimation of the Fe3+ content is a prerequisite
for the correct assignment of cations to amphibole
crystallographic sites. Variation in the assumed molar
ratio Fe3+/Fe2+ affects the calculated distribution of Al
between tetrahedral and octahedral sites, as well as the
distribution of Na between M4 and A. Thus the Fe3+/
Fe2+ ratio has important consequences for assessing
substitution mechanisms, especially in the present case,
given the large amounts of total Fe that are present
(Table 2). In the general case there are three possible
approaches to the estimation of Fe3+. We will assume
for our ferromagnesian amphibole that Fe is the only
element that exists in more than one oxidation state and
that the formula unit contains 22 O2– and two monova-
lent anions such that the total anion charge is –46.
(1) Fe3+ may be determined from separates by wet-
chemical analysis (e.g., Deer et al. 1997), or the ratio
Fe3+/Fetotal (hereafter called the iron ratio) evaluated
from Mössbauer spectrometry (e.g., Schindler et al.
2008). Micro-scale methods are also available (electron
energy-loss spectroscopy EELS, the ank and peak-
shift methods, and Mössbauer spectrometry), but their
accuracy and precision are problematic.
(2) Absent these direct methods, a second approach
is to assess the range in acceptable Fe3+ content by
performing calculations that apply general restric-
tions based on site-occupancy and electroneutrality
(Robinson et al. 1982, Hawthorne 1983, Holland &
Blundy 1994, Schumacher 1997). For many compo-
sitions presented here, the assumption that all Fe is
divalent produces acceptable site occupancies; in such
cases, a small fraction of Na resides at M4. However,
for a subset of analyses, if it is assumed that all Fe is
divalent, then the normalized sum of cations excluding
Na and K exceeds 15, and therefore some Ca is forced
to be at the A site. In such cases, minimum structurally
acceptable Fe3+ is produced by normalizing the sum of
cations excluding Na and K to 15 (which restricts Na to
the A site only). Thus, it may be concluded that at least
some of the Fe present is trivalent. Maximum allowable
Fe3+ in most cases is produced by normalization of the
formula unit such that all Na resides at M4 (i.e., sum
cations excluding K equals 15). For some of our compo-
sitions with low wt.% Al2O3, this normalization results
in Si + Al < 8 apfu. In any case, this (maximum Fe3+)
assumption is patently inappropriate for the gedritic
amphiboles in question here, given the redox state as
discussed above. It is important to note that recalculated
values of Fe3+ on a spot by spot basis are very sensitive
to analytical errors, as uncertainties in all measured
oxides are propagated in the calculation. Thus, values
obtained by this method must be viewed with caution.
Formula amounts of TAl may not be strongly inuenced
by the assumed amount of Fe3+, but clearly the M2
cation population will be strongly affected.
(3) A third approach is to simply adopt for a given
population (such as within a rock sample) a xed value
of Fe3+/Fetotal that is consistent with site occupancy
constraints, likely substitution mechanisms, and esti-
mated values of intensive variables characterizing the
metamorphism (e.g., P, T, fO2). We applied this third
approach to our analyses of ferromagnesian amphibole
in sample 98D14-150 and estimated Fe3+/Fetotal = 0.015,
as justied below.
The edenite and tschermakite substitutions in antho-
phyllite-gedrite, expressions (1) and (2) above, refer to
simple-system compositions. In the ferroanthophyllite-
ferrogedrite series, M2Fe2+ will accompany M2Mg, and
in most natural ferromagnesian amphiboles we must
also consider the presence of Fe3+, Ti, Cr, and V at
the M2 site, and Ca and Mn at M4. We anticipate that
most, if not all, of the Na in high-temperature gedrite
and clinogedrite is likely to be present at the A site (e.g.,
Schindler et al. 2008). The probability of important
glaucophane substitution (M4Na) in the Duluth contact
aureole is small, given the P-T conditions (Table 2).
Our amphibole is compositionally in the ferromag-
nesian group and, although monoclinic, it would seem
very likely that component substitutions will mimic
those in the anthophyllite-gedrite series. If this is the
case, then we can say from expressions (1) and (2) that
analyzed TAl should ideally exceed M2Al (+ Fe3+ + 2Ti
+ Cr + V) apfu by an amount given by ANa apfu; that is:
TAl – (M2Al + Fe3+ + 2Ti + Cr + V) = ANa. (3)
These constituents are plotted in Figures 6a, 6b, and
6c. Equation (3) implies that, as the amount of gedrite
substitution tends towards zero, all values in Figures
6a, 6b, and 6c should converge simultaneously on zero.
We nd that this requirement is best achieved in sample
98D14-150, where we take the Fe ratio to be 1.5%.
Corresponding straight-line least-square ts possess
y-intercept values (x = 0) of 0.0275, 0.0014, and 0.0046
apfu in Figures 6a, 6b, and 6c, respectively. An arbitrary
choice of 5% instead for the Fe ratio changes these
y-intercepts by 0.12, –0.06, and –0.15 apfu respectively.
ANa in the clinogedrite becomes 0.05 apfu smaller and
M2Al + Fe3+ + Cr + V + 2Ti increases by 0.11 apfu.
Also, the Al-free endmember, grunerite, is predicted
to have negative ANa, which implies a cation total of
less than 15. The choice of Fe ratio has no effect on
the slope of the ANa versus TAl curve (Fig. 6b). The
optimized Fe ratio implies the presence of very small
amounts of M4Na.
Mössbauer spectrometry of 20 samples of antho-
phyllite and gedrite from amphibolites (Schindler et
al. 2008) found 3 to 8% of the Fe to be Fe3+, with an
average of 5%. We use these Fe ratios for the Schindler
546 the canadian mineralOgiSt
Fig. 6. (a), (b), (c) Formula-unit chemical variations among microprobe spot analyses
of ferroan clinoamphibole from Biwabik Formation sample 98D14-150 (lled black
circles), assuming an Fe ratio of 1.5%. For comparison are shown anthophyllite-gedrites
(grey-lled squares) from Schindler et al. (2008, Table 3), and sodic gedrites from
Labrador (grey-lled circles) after Berg (1985) and Berg & Wiebe (1985) and from
southern Sweden (grey-lled triangles) after Claeson & Meurer (2002), assuming in
both cases minimum Fe3+. The grey lines on Figure 6a delineate the region accessible
by combinations of edenite and tschermakite exchange. Above the upper grey line,
at least some Na must be present at the M4 site. (d) Ca versus TAl in sample 98D14-
150 compared with cummingtonite in natural volcanic rocks (grey-lled triangles),
synthetic Al-cummingtonite PIN 69 (grey-lled circles), and anthophyllite-gedrite
(open squares) from Schindler et al. (2008), and gedrite from Labrador (open circles)
and from southern Sweden (open triangles). Figures 6c and 6d are found on the next
et al. (2008) samples plotted in Figures 6a, 6b, and 6c.
Analyses of ferrogedrite in the literature show compa-
rable proportions of Fe3+ (Deer et al. 1997). While our
optimal estimate of 1.5% for the Fe ratio in the present
sample is lower than the amphibolite gedrites, it may
be deemed more appropriate given the low fO2 of our
The Figure 6a,b,c graphs each take two variables
from three variables belonging to the same dataset. The
best-t lines taken together reect the internal depen-
dency among these variables. Equation (3) provides a
test with respect to how well the plotted data reect
the basic gedrite substitutions (1) and (2) together. For
example, if we divide equation (3) throughout by TAl,
then the slope of the ANa versus TAl curve (Fig. 6b)
Fig. 6. See caption on preceding page.
548 the canadian mineralOgiSt
should be equal to 1.0 minus the slope of the M2Al +
Fe3+ + 2Ti + Cr + V versus TAl curve (Fig. 6a). This
comparison results in a disagreement of 2.4%. If,
instead, we divide equation (3) in turn by M2Al and ANa,
we get disagreements among the slopes (Fig. 6a cf. 6c,
and Fig. 6b cf. Fig. 6c) of 7.6% and 5.1% respectively.
We judge these disagreements as not seriously contrary
to our belief in the control of compositions largely by
the gedrite substitutions (1) and (2).
Calcium declines from 0.45 apfu in our most Si-poor
(or TAl-rich) compositions to 0.1 apfu in the most
Si-rich (Fig. 6d). This trend would be consistent with
marginal change accompanying falling temperature.
The Ca decline is counterbalanced by an increase in Mg
(Fig. 5). The monoclinic symmetry of the amphibole
is consistent with its high CaO (content of actinolite
component) which, except for amphibole rims, is
2–3 wt.%, distinctly higher than any orthorhombic
anthophyllite-gedrite solution (Fig. 6d).
Next, we compare our results with data for antho-
phyllite and gedrite from the literature.
Literature comparisons of composition trends
Hawthorne et al. (2008) showed that anthophyllite
and gedrite in amphibolites follow constrained trajec-
tories in composition space, and that this fact can be
attributed to anion bond-valence requirements in the
Pnma amphibole structure. The trajectories are essen-
tially the same as those shown in Robinson et al. (1971).
A feature of the composition trends for amphibolite
samples is an ANa to M2Al ratio of 4:10 (Hawthorne
et al. 2008).
Our monoclinic amphibole, on the other hand,
denes a trend for ANa relative to M2Al that is much
steeper (7:10) than the amphibolite trend (Fig. 6c). The
trend of our amphibole in ANa relative to TAl (4.7:10)
is also steeper (Fig. 6b) than the amphibolite trend
(2.7:10). The differences are smaller in the plot of M2
cations versus TAl (Fig. 6a). In short, our sample shows
signicantly greater advance along the ed exchange
vector (Eq. 1) than the amphibolites, and conversely
much smaller advance along the ts exchange vector
(Eq. 2).
Let us now test the anthophyllite-gedrite data in
Schindler et al. (2008) with respect to equation (3)
as we did above for sample 98D14-150. In this case,
disagreements are 4.8%, 1.9%, and 22%, respectively.
The third test result of the Schindler et al. (2008) data
is much poorer than all the others, perhaps because we
are dividing by ANa. In both populations one can draw
the conclusion that the datasets are largely consistent
with the combination of ed and ts substitutions (1 and
2). Built into these tests, of course, are uncertainties in
calculated formulae, e.g., the calculated partition of Na
between the M4 and A sites, that no Ca is present at A,
analytical errors, and the estimated Fe ratio.
Exceptions to the common amphibolite trend, like
our sample, were noted (Hawthorne et al. 2008) in the
sodic gedrites described by Berg (1985) and Berg &
Wiebe (1985) from a xenolith in granite gneiss, Nain
complex, Labrador (~800 °C rst metamorphic stage,
615 °C second stage, 2 kbar) and from a troctolite
cumulate described by Claeson & Meurer (2002) from
southern Sweden (~900 °C, 4–6 kbar). To plot these
samples (Fig. 6), we used a minimum value for the Fe
ratio determined via site occupancy constraints. Trends
dened by our clinoferrogedrite, although actually lower
in both absolute TAl and M2Al, are rather similar to these
sodic gedrites in the literature. Sodium is slightly higher
for given values of TAl and M2 substitution.
Differences in environmental intensive variables
could conceivably explain these three examples of
high-Na gedritic amphibole. Two of the three examples
are associated with temperatures in excess of 800 °C,
and perhaps the third (Labrador) as well. According
to Hawthorne et al. (2008), it is possible that bond-
valence requirements adjust to different bond lengths at
the higher temperatures. It is well known in petrology
that the ed and ts substitutions in calcic amphiboles are
promoted by higher temperatures, although very likely
to different degrees as a function of the mineral para-
genesis. We note, however, that the “high-Na” (high ed)
signature of our amphibole is maintained throughout the
traverse from core to low-Al grunerite rim, with corre-
lated declines in Na, TAl, and M2Al (Figs. 6b and 6c).
The rim can only have grown during cooling of the
aureole because grunerite is unstable (Ghiorso & Evans
2002) at the very high temperatures reached in the inner
aureole, so temperature may not be the driving factor
behind the high-ed signature.
Alternative controls on the gedrite trends can be
explored if we view expressions (1) and (2) as complex
exchange reactions governed by the balance of energies
among the phases present (the chemical potentials of the
components). For example, silica undersaturation can
promote increased TAl according to both exchanges. We
see this effect in the substitution of TAl in pyroxenes and
calcic amphiboles from silica-undersaturated igneous
(e.g., Giret et al. 1980) and metamorphic environments
(pargasite, hastingsite, and sadanagaite in ultramacs
and calc-silicate skarns with spinel, olivine, nepheline,
corundum, etc., Deer et al. 1997). Both the Labrador
(olivine Fa84, Opx Fs63, Al-spinel, etc.) and Swedish
(olivine Fa23, Opx Fs20, Pl An88) gedrite-bearing rocks
contain olivine, as does sample 98D14-150. Thus, an
alternative explanation for the “high-Na” compositions
might be that Si was undersaturated at the time. Our
sample contains quartz, however, which is in bands and
lenses accompanied by plagioclase. At >800 °C and
1.3 kbar, these lenses were possibly melt, that is, our
sample is perhaps a migmatite, so at the metamorphic
maximum it might well be the case that Si was, perhaps
locally, undersaturated (unlike normal BIF). Again, this
explanation must remain effective from core to cooler
rim of our amphibole when we would expect any melt
to have crystallized to plagioclase and quartz.
Unlike silica, the chemical potential of Mg (or total
femics) has a direct effect only on the ts exchange:
higher femics in the potential equation should lower the
amounts of total Al, while the ed exchange is unaffected.
This suggests a role for the presence of olivine. Other
workers have suggested the “assemblage” control in
the case of uptake of Al in calcic amphibole. Whatever
the cause, the presence of alternate gedrite trends in
both orthorhombic and monoclinic MgFe-amphibole
structures shows that the “amphibolite” trend is as good
as it is probably because of constrained values of inten-
sive petrologic variables, including chemical potentials
dened by the mineral assemblage.
Effect of symmetry
Despite its apparent high temperature of formation,
our clinoferrogedrite has less total Al than the Labrador
and Swedish sodic gedrites. This difference may be a
result of crystal symmetry. In monoclinic amphibole,
long-range disorder of Al is limited to two T sites,
whereas in orthorhombic amphibole long-range disorder
is possible among at least three T sites (Hawthorne et
al. 2008).
The clustering of data-points in Figures 6a and 6b
at TAl ~ 1.25 illustrates the high relative modal abun-
dance of that composition in our sample, as compared
to the spread of points in the Al-grunerite range. Our
clinoferrogedrite is thus robust in its presence. For this
reason, we believe that adjustment of the thermody-
namic properties of gedrite to preclude the formation
of clinogedrite under all conditions (Diener & Powell
2012) is not necessarily appropriate.
SYnthetic al-rich cummingtOnite
and clinOgedrite
We believe it is no accident that our finding of
MgFe-amphibole in the “unusual” clinoferrogedrite
range in the Duluth aureole is to a degree matched by
results from synthesis work. In laboratory experiments
using fused glass of 1991 Pinatubo dacite (Scaillet
& Evans 1999), some experiments were performed
with added S. Among other product minerals, we
identied in polished mounts of three experiments an
aluminous ferromagnesian amphibole that we called
gedrite (Scaillet & Evans 1999). More detailed elec-
tron microprobe work later showed relatively high
CaO contents, and a check by selected-area electron
diffraction, kindly performed by K. Bozhilov, showed
that the amphibole is in fact monoclinic. With Al2O3
contents ranging from 6 to 10 wt.%, these synthetic
cummingtonites are in a compositional range not found
in natural volcanic cummingtonite. Their TAl content
ranges from 0.6 to slightly greater than 1.1. Experiment
no. PIN 69, held for 14 days at 3890 bar, 780 ºC, ΔNNO
= 2.63, with 8 wt.% H2O and 1.35 wt.% S, provided
the best sample for detailed electron microprobe study,
and these compositions are plotted here (Figs. 6d, 7a,
7b, and 7c). Accompanying product minerals were
plagioclase, biotite, magnetite, ilmenite, and anhydrite.
Figure 7 shows that loci of analysis points for Na, TAl,
and MAl + 2Ti + Fe3+ differ from the Biwabik material
and the Al-anthophyllite-gedrite solid solution trend in
amphibolites. ANa is relatively low and M4Na is high;
two outlier points had anomalously low Fe and Mg and
were edited out. Molar Fe/(Fe+Mg) of the amphibole
in PIN 69 ranges between 0.29 and 0.31. An Fe ratio
of 10% was adopted for the calculation of formula
contents used in the plots, recognizing the fact that fO2
both in the experiments and in the natural arc-magma
samples also plotted in the Figures was higher than in
the Biwabik BIF sample. The volcanic samples are from
the St. Helens dacite, Mt. Pinatubo dacite, Martinique
dacite, Taupo rhyolite, Mt. Daisen (Japan) andesite,
Cacoyos (Guatemala) ash, and Catalina Island (Santo-
rini) rhyolite (Evans et al. 2001). Ilmenite-magnetite
geothermometry indicates a temperature range from 700
to 800 °C for cummingtonite in rhyolites and dacites
(Ghiorso & Evans 2008). Neither quartz nor olivine
was present among the products of experiment PIN
69. The experimental temperature of 780 ºC is scarcely
higher than that of natural volcanic cummingtonites. In
this case, we tentatively suggest that Si undersaturation
was responsible for the greater edenite and tschermakite
(gedrite) content of the experimental clinoamphibole
in comparison to the natural volcanic cummingtonites
(TAl < 0.5).
Volcanic cummingtonite is invariably accompanied
by hornblende. In S-spiked experiments of siliceous
magmas at relatively high fO2, crystallization of horn-
blende is diminished because Ca is taken away to form
anhydrite, the released Mg and Fe then entering biotite
(e.g., Costa et al. 2004) and cummingtonite. Horn-
blende did not form at all in experiment PIN 69. With
further addition of S, dacitic liquid evolves towards
peraluminous rhyolite in composition, and crystalliza-
tion products include cordierite, osumilite, and mullite
(Evans & Scaillet, unpublished data).
In this paper we adopted the conservative approach
of using the system of nomenclature for MgFeMn-
amphiboles that has been employed for many years
(Leake 1978, Deer et al. 1992, 1997, Leake et al. 1997,
2003), in which subdivisions are based on the Mg/Fe
ratio at 50:50 and the occupancy of the T site (Si7). We
use this nomenclature in preference to that suggested
by Hawthorne et al. (2012), where the occupancy of
the M2 site by 3+ and 4+ cations is used to separate
550 the canadian mineralOgiSt
anthophyllite from gedrite rather than the TSi content.
While their proposal in principle has certain merits, it
unfortunately depends on acquiring an accurate measure
of Fe3+, which is a challenge for many workers to do,
especially in situ and in the numerous cases where
compositions vary on a micron scale. The issue here
is not so much the mineral formulae but the mineral
names. Choice of the latter is not trivial in importance,
given their appearance in searchable titles, abstracts,
and citations. We also use the conventions governing
prexes such as clino, ferro and ferri, as in Leake et
al. (1997, 2003).
Fig. 7. (a), (b), (c) Formula unit chemical variations among spot analyses of synthetic
clinoamphibole from experiment PIN 69 at 781 °C and 2 kbar, NNO+2.5 (lled black
circles). For comparison are shown cummingtonites from volcanic rocks (black-lled
triangles) and the trend of anthophyllite-gedrite from Schindler et al. (2008). For the
PIN 69 experiment and the volcanic rocks, the Fe ratio is taken to be 10%. Figure 7c
is found on the following page.
We identify the gedritic character of MgFe-amphi-
boles as arising from advancement along both the ed and
ts vectors in composition space, as in the type gedrite
from Gèdres, France. Silica apfu is a measure of both
substitutions combined, whereas ANa apfu alone and
M2(Al + Fe3+ + 2Ti) apfu alone signify the ed and ts
substitutions, respectively. For this reason, we might
view Si apfu as having greater signicance in assigning
the name gedrite to an MgFe-amphibole. The amphibole
in sample 98D14-150 is clinoferrogedrite according
to the compositional boundaries in the 1978 to 2003
IMA nomenclature, but because M2(Al + Fe3+ + Cr +
V + 2Ti) does not exceed 1.0 apfu, it is sodic alumi-
nous grunerite according to Hawthorne et al. (2012).
Averaging 0.5 apfu, our clinoferrogedrite in fact has
more ANa than any of the gedrites in the Schindler et
al. (2008) study (Figs. 6b and 6c). Schumacher (2007)
showed that the diversity of metamorphic amphibole
compositions arising from the four principal heterova-
lent cation substitutions (ts, ed, gln, rct) can be depicted
satisfactorily, it would seem, in plots of vectors dened
by the occupancies of the A, M4, and T sites.
A possible role for Si activity was mentioned
above. Silica activity plays an important role (among
other inuences) as a controlling intensive parameter
in igneous and metamorphic petrology. The ed and
ts substitutions, which involve Al and Si at the T-site
(unlike gln and rct), occur not only in the MgFeMn-
amphiboles, but also in the Ca-amphiboles, the NaCa-
amphiboles, and the Na-amphiboles. In contrast to
Fe2O3, SiO2 can be analyzed in situ with high accuracy
and precision. It happens that a T-site fully occupied
by Si (or nearly so) is a feature of all the asbestiform
to acicular amphiboles; TAl serves to increase bond
strengths across the tetrahedral chains, leading to more
prismatic amphiboles. Similarly, the most important
compositional control on the compressional velocities
of amphiboles was found to be the Si-content (J.M.
Brown, pers. commun.). The proportional changes in Si
apfu here are small, but the effect on physical proper-
ties is major; we lose much by ignoring the Si content
of the T-site.
Very aluminous cummingtonite and grunerite, and
compositions that can be called clinogedrite and clino-
ferrogedrite, are extremely rare. In this paper, we have
described a eld example and a synthetic example.
Clearly, there are crystal-chemical reasons for their
rarity, in contrast to compositionally equivalent ortho-
rhombic examples, which occur quite commonly. We
have to inquire then, what are the special environmental
conditions that enable the monoclinic versions to occur
Among the requirements for clinoferrogedrite in the
place of Fe-rich orthopyroxene and orthoferrogedrite
are elevated H2O pressure and relatively high tempera-
ture (e.g., Diener & Powell 2012), respectively. In our
amphibole, the wide range in Ca, ed, and ts contents
Fig. 7. See preceding page for caption.
552 the canadian mineralOgiSt
(in the one rock) are indicative of crystallization over a
substantial range in temperature. Calcium apfu extends
in crystal cores to higher values than in virtually all
volcanic cummingtonites, where it is matched by the
PIN 69 crystals, which grew at 780 °C (Fig. 6d). We
have noted that the clinoferrogedrite occurs as inclu-
sions in garnet and that it is generally inclusion-free. We
cannot conclude, however, that it was present at peak
temperature (≈ 820 °C), although it must have appeared
shortly after temperature relaxation began. As in the
case of ordinary gedrites, the abundance of hornblende
in the rock must be diminished on chemical grounds
by appropriately low system values of the exchange
potential ΔµCa(Mg,Fe)–1. In the present example from
the Biwabik Formation, there is a possible role for
high values for the chemical potential of femics rela-
tive to that of Si. Aluminous anthophyllite and gedrite,
for reasons of crystal structure, are by contrast able to
accommodate gedritic substitutions at lower tempera-
tures, and so they are much more common in nature.
Krassimir Bozhilov and Bruno Scaillet are thanked
for laboratory input in parts of this paper. We also thank
John Schumacher and an anonymous reviewer for their
helpful comments.
anderSen, d.J., lindSleY, d.l., & daVidSOn, P.m. (1993)
QUILF: a Pascal program to assess equilibria among
Fe–Mg–Mn–Ti oxides, pyroxenes, olivine, and quartz.
Computers and Geoscience 19(9), 1337–1350.
andreWS, m.S. & riPleY, e.m. (1989) Mass transfer and
sulfur xation in the contact aureole of the Duluth Com-
plex, Dunka Road Cu–Ni deposit. Canadian Mineralogist
27, 293–310.
Berg, J.h. (1985) Chemical variation in sodium gedrite from
Labrador. American Mineralogist 70, 1205–1210.
Berg, J.h. & WieBe, r.a. (1985) Petrology of a xenolith of
ferroaluminous gneiss from the Nain complex. Contribu-
tions to Mineralogy and Petrology 90, 226–235.
BOnnichSen, B. (1968) General geology and petrology of the
metamorphosed Biwabik Iron Formation (Precambrian),
Dunka River area, Minnesota. Ph.D. thesis, University of
Minnesota, 269pp.
BOnnichSen, B. (1969) Metamorphic pyroxenes and amphi-
boles in the Biwabik Iron Formation, Dunka River area,
Minnesota. Mineralogical Society of America Special
Paper 2, 217–239.
BOnnichSen, B. (1975) Geology of the Biwabik Iron Forma-
tion, Dunka River area, Minnesota. Economic Geology
70, 319–340.
BOzhilOV, k. & eVanS, B.W. (2001) Ferroanthophyllite in
Rockport grunerite: A transmission electron microscopy
study. American Mineralogist 86, 1252–1260.
claeSOn, d.t. & meurer, W.P. (2002) An occurrence of igne-
ous orthorhombic amphibole, Eriksberg gabbro, southern
Sweden. American Mineralogist 87, 699–708.
cOllinS, r.S. (1942) Cummingtonite and gedrite from Suther-
land. Mineralogical Magazine 26, 254–259.
cOnnO llY, J.a.d. & ceSa re, B. (1993) C–O–H–S fluid
composition and oxygen fugacity in graphitic metapelites.
Journal of Metamorphic Geology 11, 379–388.
cOSta, F., Scaillet, B., & PichaVant, m. (2004) Petrological
and experimental constraints on the pre-eruption condi-
tions of Holocene dacite from Volcán San Pedro (36°S,
Chilean Andes) and the importance of sulphur in silicic
subduction-related magmas. Journal of Petrology 45(4),
daVidSOn, P.m. & lindSleY, d.l. (1989) Thermodynamic
analysis of pyroxene-olivine-quartz equilibria in the sys-
tem CaO–MgO–FeO–SiO2. American Mineralogist 74,
deer, W.a., hOWie, r.a., & zuSSman, J. (1992) An Intro-
duction to the Rock-Forming Minerals. Longman Group,
London, United Kingdom.
deer, W.a., hOWie, r.a., & zuSSma n, J. (1997) Double-
Chain Silicates, Volume 2B, Second Edition. The Geologi-
cal Society, London, United Kingdom.
diener, J.F.a. & POWell, r. (2012) Revised activity-compo-
sition models for clinopyroxene and amphibole. Journal of
Metamorphic Geology 30, 131–142.
eVanS, B.W., ghiOrSO, m.S., Yang, h., & medenBach, O.
(2001) Thermodynamics of the amphiboles: Anthophyllite-
ferroanthophyllite and the ortho-clino phase loop. Ameri-
can Mineralogist 86, 640–651.
Ferré, e. (1989) Les gneiss à cordiérite-grenat-orthoamphi-
bole de Topiti: témoin possible d’un soile métamorphique
du Protérozoique en Corse occidentale. Comptes Rendues
de l’Académie de Sciences, Paris, Série 2 309, 893–898.
French, B.m. (1968) Progressive contact metamorphism of
the Biwabik iron-formation, Mesabi Range, Minnesota.
Minnesota Geological Survey Bulletin 45, 103.
ghiOrSO, m.S. & eVanS, B.W. (2002) Thermodynamics of the
amphiboles: Ca-Mg-Fe2+ quadrilateral. American Miner-
alogist 87(1), 79–98.
ghiOrSO, m.S. & eVanS, B.W. (2008) Thermodynamics of
rhombohedral oxide solid solutions and a revision of the
Fe-Ti two-oxide geothermometer and oxygen barometer.
American Journal of Science 308, 957–1039.
giret , a., BOnin, B., & lé ger, J-m. (1980) Amphibole
composition trends in oversaturated and undersaturated
alkaline plutonic ring-complexes. Canadian Mineralogist
18, 481–495.
gunderSen, J.n. & SchWarz, g.m. (1962) The geology of the
metamorphosed Biwabik Iron Formation, eastern Mesabi
district, Minnesota. Minnesota Geological Survey Bulletin
43, 139.
haWthOrne, F. (1983) The crystal chemistry of the amphi-
boles. Canadian Mineralogist 21, 173–480.
haWthOrne, F.c., Schindler, m., aBdu, Y., SOkOlOVa, e.,
eVanS, B.W., & iShida, k. (2008) The crystal chemistry
of the gedrite-group amphiboles. II. Stereochemistry and
chemical relations. Mineralogical Magazine 72, 731–745.
haWthOrne, F.c., OBerti, r., harlOW, g.e., mareSch, W.V.,
martin, r.F., Schumacher, J.c., & Welch, m.d. (2012)
Nomenclature of the amphibole supergroup. American
Mineralogist 97, 2031–2048.
hOlland, t. & BlundY, J. (1994) Non-ideal interactions
in calcic amphiboles and their bearing on amphibole-
plagioclase thermometry. Contributions to Mineralogy and
Petrology 116, 433–447.
JOY, B.r. (2009) Chemical disequilibrium in the mineral
assemblage olivine + orthopyroxene + augite + quartz
from the contact metamorphosed Biwabik Iron-Formation,
northeastern Minnesota. M.Sc. thesis, University of Cali-
fornia, Davis.
kOniShi, h., dódOnY, l., & BuSeck, P.r. (2002) Protoantho-
phyllite from three metamorphosed serpentinites. Ameri-
can Mineralogist 87, 1096–1103.
kOniShi, h., grOY, t.l., dódOnY, l., miYaWaki, r., matSuB-
ara, S., & BuSeck, P.r. (2003) Crystal structure of pro-
toanthophyllite: A new mineral from the Takase ultramac
complex, Japan. American Mineralogist 88, 1718–1723.
laBOtka, t.c., PaPike, J.J., Vaniman, d.t., & mOreY, g.B.
(1981) Petrology of contact metamorphosed argillite from
the Rove Formation, Gunint Trail, Minnesota. American
Mineralogist 66, 70–86.
leake, B.e. (1978) Nomenclature of amphiboles. Canadian
Mineralogist 16, 501–520.
leake, B.e., WOOlleY, a.r., arPS, c.e.S., Birch, W.d.,
gilBert, m.c., grice, J.d., haWthOrne, F.c., katO, a.,
kiSch, h.J., kriVOVicheV, V.g., linthOut, k., laird, J.,
mandarinO, J.a., mareSch, W.V., nickel, e.h., rOck,
n.m.S., Schu macher, J.c., Smith, d.c., StePhenSOn,
n.c.n., unga ret ti, l., Wh ittak er, e.J.W., & guO,
Y.g. (1997) Nomenclature of amphiboles: Report of the
Subcommittee on Amphiboles of the International Miner-
alogical Association, Commission on New Minerals and
Mineral Names. Canadian Mineralogist 35, 219–246.
Mineralogical Magazine 61, 295–321.
leake, B.e., WOOlleY, a.r., Burch, W.d., Burke, e.a.J.,
FerrariS, g., grice, J.d., haWthOrne, F.c., kiSch, h.J.,
kriVOVic heV , V.g., Schumacher, J.c., SteP henSOn,
n.c.n., & Whittaker, e.J.W. (2003) Nomenclature of
amphiboles: additions and revisions to the International
Mineralogical Association’s amphibole nomenclature.
Canadian Mineralogist 41, 1355–1370.
merlet, c. (1994) An accurate computer correction program
for quantitative electron probe microanalysis. Mikrochi-
mica Acta 114/115, 363–376.
mOreY, g.B. (1992) Chemical composition of the eastern
Biwabik Iron Formation (early Proterozoic), Mesabi
Range, Minnesota. Economic Geology 87, 1649–1658.
PattiSOn, d.r.m. & tracY, r.J. (1991) Phase equilibria
and thermobarometry in metapelites. In Contact Meta-
morphism (D.M. Kerrick, ed.). Mineralogical Society of
America Reviews in Mineralogy (105–206).
PattiSOn, d.r.m., chackO, t., Farquar, J., & mcFarlane,
c.r.m. (2003) Temperatures of granulite-facies metamor-
phism: constraints from experimental phase equilibria
and thermobarometry corrected for retrograde exchange.
Journal of Petrology 44, 867–900.
raBBitt, J.c. (1948) A new study of the anthophyllite series.
American Mineralogist 33, 263–323.
rOBinSOn, P., JaFFe, h.W., rOSS, m., & klein, c. (1971)
Orientation of exsolution lamellae in clinopyroxenes and
clinoamphiboles: Consideration of optimal phase boundar-
ies. American Mineralogist 56, 909–939.
rOBinSO n, P., rOSS, m., & JaFFe, h.W. (1971) Composi-
tion of the anthophyllite-gedrite series, comparisons of
gedrite-hornblende, and the anthophyllite-gedrite solvus.
American Mineralogist 56, 1004–1041.
rOBin SOn, P., SPear, F.S., Schuma cher , J.c., laird, J.,
klein, c., eVanS, B.W., & dOOlan, B.l. (1982) Phase
relations of metamorphic amphiboles: natural occurrence
and theory. In Amphiboles: Petrology and Experimental
Phase Relations (D.R. Veblen & P.H. Ribbe, eds.). Reviews
in Mineralogy 9B, 1–228.
Scaillet , B. & eVanS, B.W. (1999) The June 15, 1991
eruption of Mount Pinatubo. I. Phase equilibria and pre-
eruption P–T–fO2–fH2O conditions of the dacite magma.
Journal of Petrology 40(3), 381–411.
Schindler, m., SOkOlOVa, e., aBdu, Y., haWthOrne, F.c.,
eVanS, B.W., & iShida, k. (2008) The crystal chemistry
of the gedrite-group amphiboles. I. Crystal structure and
site populations. Mineralogical Magazine 72(3), 703–730.
Schumacher, J.c. (1997) Appendix 2. The estimation of the
proportion of ferric iron in electron-microprobe analyses of
amphiboles. Canadian Mineralogist 35, 238–246.
Schumacher, J.c. (2007) Metamorphic amphiboles: Compo-
sition and Coexistence. Chapter 10. In Amphiboles: Crys-
tal Chemistry, Occurrence, and Health Issues (F.C. Haw-
thorne, R. Oberti, G. Della Ventura, & A. Mottana, eds.).
Reviews in Mineralogy and Geochemistry 67, 359–416.
554 the canadian mineralOgiSt
Smelik, e.a. & VeBlen, d.r. (1992) Exsolution of horn-
blende and the solubility limits of calcium in orthoamphi-
bole. Science 257(5077), 1669–1672.
WhitneY, d.l. & eVanS, B.W. (2010) Abbreviations for
names of rock-forming minerals. American Mineralogist
95, 185–187.
WilliamS, k.l. (1987) Introduction to X-ray Spectrometry.
Allen & Unwin, London, United Kingdom.
Received August 23, 2013, revised manuscript accepted
June 4, 2014.
... In the paper on which we are commenting (Joy & Evans 2014), the authors (hereafter referred to as J&E) introduced a new mineral belonging to the amphibole supergroup, gave it a name, and discussed its relations with coexisting amphiboles and amphiboles reported in previous literature. ...
... However, to have its formal recognition as a valid species, its mineral description (and the selected rootname) should be submitted to the IMA-CNMNC for approval. Chemical compositions for synthetic PIN69, also reported in Table 2 of Joy & Evans (2014), for which monoclinic symmetry had been proved, are referred to in the paper as "highly aluminous cummingtonite grading into clinogedrite". Actually, they all refer to the simplified formula A □ B Mg 2 C Mg 5 T Si 8 O 22 W (OH) 2 [because A (Na+K) ≤ 0.50 apfu, C M 3+ < 1 apfu, and B,C Mg > B,C Fe 2+ ], which corresponds to the synthetic analogue of cummingtonite. ...
Full-text available
In the paper on which we are commenting ([Joy & Evans 2014][1]), the authors (hereafter referred to as J&E) introduced a new mineral belonging to the amphibole supergroup, gave it a name, and discussed its relations with coexisting amphiboles and amphiboles reported in previous literature. First,
... However, the temperatures of arsenopyrite precipitation (465-560°C) and of the low-to mid-amphibolite metamorphic facies are above the registered in most classic orogenic gold deposits (Groves et al., 1998). Associations of biotite, amphiboles, Fe-rich garnet, and Fe-Ti-bearing oxides (e.g., titanomagnetite, ilmenite) like those described for São Sebastião have been interpreted as contact metassomatic alterations at the Proterozoic Biwabik BIF in northeastern Minnesota (Joy and Evans, 2014) and at the Archean BIF-hosted gold deposits of Nevoria (Yilgarn Craton, Australia - Mueller, 1997;Mueller et al., 1994) and Lupin (Slave Craton, Canada - Bullis et al.,1994;Meinert, 1998), considered by cited authors as skarn-like gold deposits, but also described as hypozonal deposits by Kolb et al. (2015). Archean gold deposits associated to amphibolite facies terrains develop over a 500-700°C range (Kolb et al., 2015) and are considered to form at the lower crustal section or "bottom" of greenstone belts (Hagemann et al., 2013). ...
São Sebastião is a recently discovered Archean gold deposit at the northwesternmost part of the Quadrilátero Ferrífero (QF), southern São Francisco craton, Brazil. The gold ore is strata-confined in two levels of banded iron formation (BIF) of the basal stratigraphic unit in the Pitangui greenstone belt (2.8 Ga). Pitangui is correlated to the widely known Rio das Velhas greenstone belt, which hosts several orogenic gold deposits such as the world-class Cuiabá. The gold mineralization at São Sebastião is associated to replacement-style sulfidation over magnetite, where the main developed phases are: pyrrhotite, chalcopyrite, arsenopyrite, pyrite and arsenian-pyrite. Native Au occurs as inclusions in pyrrhotite formed at dilatation domains (e.g., breccias, fold hinges) during compressional events (Rio das Velhas Orogeny at 2.8 – 2.75 Ga), but mainly as inclusions in late- to post-kinematic pyrite and arsenopyrite, as fracture infills in arsenopyrite and in contact with gangue minerals. Electron-microprobe analyses in arsenopyrite reveal a 465 – 560oC precipitation temperature interpreted from As atomic percentage and indicate that high-temperature fluids were active after the deformational event. The temperature increase caused the incorporation of LA-ICP-MS-detected lattice-bound Ni and Co in arsenian-pyrite and arsenopyrite, related to the mafic-ultramafic signature of boundary rocks. High temperature conditions are supported by pyrite and ISS (intermediate-solid solution in the Cu-Fe-S system) formation at 600oC and melting-derived textures resulting from the precipitation of Bi- and As-melts, typical of amphibolite-facies environments with high fS2. Whole-rock geochemical assays of drill core samples display two clear trends on Ag, Bi, Cu vs. Au binary plots and on principal component analysis results. The trends are a product of the heterogeneous spatial distribution of melts and high-temperature fluids. Melting features and superposition of mineralization styles are common in hypozonal gold deposits experiencing higher heat influx. At São Sebastião, the thermal effect is likely related to the 2.7 Ga granitic intrusions that surround the Pitangui belt. São Sebastião shows distinct characteristics than those described for the mesozonal gold deposits in the Rio das Velhas greenstone belt and may depict the deeper roots of the QF gold system.
... The cores of the monoclinic (Z ∧ c = 15-16°; (100) ∧ ( 101) = 103.5 ± 1°) Mg-Fe-Mn amphibole in our sample 98D14-150 (Joy & Evans 2014) are compositionally equivalent (except for Ca) to a roughly 40:60 solid solution of endmember grunerite and "sodicgedrite" (or ferro-rootname 2) from Labrador and Sweden (Berg & Wiebe 1985, Claeson & Meurer 2002. They are not a solid solution of grunerite and hornblende because A Na (0.50 to 0.55 apfu) and B Ca (0.38 to 0.39 apfu) are clearly not present in proportions that can remotely resemble an extrapolated "hornblende" endmember ( Table 2 in Joy & Evans 2014). ...
We report the first detailed petrological and pressure-temperature data from the pelitic schists and am-phibolites of the Keffi area, north-central Nigeria, which is located on the eastern flank of the western Nigerian based, Great Nigerian Schist Belt. In one fresh exposure of the schist, staurolite-bearing and staurolite-absent, garnet-rich assemblages occur. All pelitic samples contain garnet, quartz, biotite, plagioclase, chlorite and ilmenite, but the staurolite-bearing assemblage contains euhedral to subhedral staurolites and very subordinate retrograde chlorites in addition. Mineral compositions applied to calculate metamorphic PeT conditions using different approaches reveal a temperature range of 570e630 C for the garnet-biotite geothermometry. PeT pseudosection analyses calculated using THERMOCALC software for the suitable rock types, constrain garnet/staurolite equilibration within the range of 6.4e7.7 kbar and 610 e630 C. Empirical calculations and pseudo-section approaches indicate a clockwise P-T path for the rocks of the study area. The result of geothermobarometry (peak conditions) from this study is consistent with previous PeT estimations for the Pan-African episode on several areas within the Trans-Saharan Belt. All evidences point towards a magmatic arc tectonic setting for this area of study.
Full-text available
The amphiboles are the most complex group of rock-forming minerals, exhibiting wide chemical variation and a bewildering variety of parageneses. They are common constituents across the complete range of igneous rocks. In sedimentary rocks, amphiboles occur both as detrital and authigenic phases. In metamorphic rocks, amphiboles are important constituents from very low grade to high grade and over a wide variety of rock compositions. This study summarizes the existing body of knowledge on the crystallography and crystal chemistry of the amphiboles. It is shown that amphiboles belong to five principal structure-types, with space groups C2/m, P2/a, P2//1/m, Pnma, and Pnmn, but the C2/m and Pnma structures are by far the most common. The M(4) site is of major importance in amphibole crystal chemistry. Spectroscopic analysis data a presented and analyzed, and cation distributions in amphiboles, factors affecting cation ordering, oxidation-dehydroxylation characteristics, electrical, magnetic, and elastic properties, and deformation behavior in amphiboles are discussed in detail.
Full-text available
A new classification and nomenclature scheme for the amphibole-supergroup minerals is described, based on the general formula AB 2 C 5 T 8 O 22 W 2 , where A = o, Na, K, Ca, Pb, Li; B = Na, Ca, Mn 2+ , Fe 2+ , Mg, Li; C = Mg, Fe 2+ , Mn 2+ , Al, Fe 3+ , Mn 3+ , Ti 4+ , Li; T = Si, Al, Ti 4+ , Be; W = (OH), F, Cl, O 2-. Distinct arrangements of formal charges at the sites (or groups of sites) in the amphibole structure warrant distinct root names, and are, by implication, distinct species; for a specific root name, different homovalent cations (e.g., Mg vs. Fe 2+) or anions (e.g., OH vs. F) are indicated by prefixes (e.g., ferro-, fluoro-). The classification is based on the A, B, and C groups of cations and the W group of anions, as these groups show the maximum compositional variability in the amphibole structure. The amphibole supergroup is divided into two groups according to the dominant W species: W (OH,F,Cl)-dominant amphiboles and W O-dominant amphiboles (oxo-amphiboles). Amphiboles with (OH, F, Cl) dominant at W are divided into eight subgroups according to the dominant charge-arrangements and type of B-group cations: magnesium-iron-manganese amphiboles, calcium amphiboles, sodium-calcium amphiboles, sodium amphiboles, lithium amphiboles, sodium-(magnesium-iron-manganese) amphi-boles, lithium-(magnesium-iron-manganese) amphiboles and lithium-calcium amphiboles. Within each of these subgroups, the A-and C-group cations are used to assign specific names to specific compositional ranges and root compositions. Root names are assigned to distinct arrangements of formal charges at the sites, and prefixes are assigned to describe homovalent variation in the dominant ion of the root composition. For amphiboles with O dominant at W, distinct root-compositions are currently known for four (calcium and sodium) amphiboles, and homovalent variation in the dominant cation is handled as for the W (OH,F,Cl)-dominant amphiboles. With this classification, we attempt to recognize the concerns of each constituent community interested in amphiboles and incorporate these into this classification scheme. Where such concerns conflict, we have attempted to act in accord with the more important concerns of each community.
Full-text available
The introduction of a fifth amphibole group, the Na-Ca-Mg-Fe-Mn-Li group, defined by 0.50 < (B)(Mg,Fe2+,Mn2+,Li) < 1.50 and 0.50 less than or equal to (B)(Ca,Na) less than or equal to 1.50 apfu (atoms per formula unit), with members whittakerite and ottoliniite, has been required by recent discoveries of (B)(LiNa) amphiboles. This, and other new discoveries, such as sodicpedrizite (which is herein slightly, but significantly changed from the original idealized formula), necessitate amendments to the IMA 1997 definitions of the Mg-Fe-Mn-Li, calcic, sodic-calcic and sodic groups. The discovery of obertiite and the finding of an incompatibility in the IMA 1997 subdivision of the sodic group, requires further amendments within the sodic group. All these changes, which have IMA approval, are summarized.
Full-text available
For the first time in Western Corsica, near Cargese, Cordierite-garnet-orthoamphibole gneisses have been found. These metamorphic rocks are associated with the gneissic country rocks in which the high Mg/K granitic rocks had intrude. The age of these unusual formations is discussed. -English summary
A thermodynamic model has been developed and calibrated for monoclinic and orthorhombic amphiboles compositionally contained within the Ca-Mg-Fe2+ amphibole quadrilateral. The model incorporates the energetic consequences of cation ordering of Fe2+ and Mg over nonequivalent sites in the crystal structure, and accounts for the temperature, pressure, and compositional dependence of the orthorhombic-monoclinic phase transition. Calibration is based on previously published work on the thermodynamic properties of Fe2+-Mg amphiboles and tremolite, and experimental cation-ordering data, along with solvus width and tieline orientations of natural coexisting amphiboles in the quadrilateral. Among derived parameters is the enthalpy of formation of end-member ferroactinolite (-10534.966 kJ/mol). A calculated FeMg-1 isopotential solvus at mid-composition in the quadrangle agrees well with a revised calibration of experimental data from Cameron (1975). The solvus is not strongly asymmetric and narrows with increasing Fe/Mg ratio and temperature. Phase relations along the Mg and Fe sides of the quadrilateral are compared and contrasted. Cummingtonite-anthophyllite phase relations are shown in the T-XFe plane, with and without coexisting calcic amphibole. The breakdown reactions of quadrilateral amphiboles to assemblages of two pyroxenes, olivine, quartz, and H2O are depicted as functions of temperature, pressure, composition, H2O activity, and oxygen fugacity. Some highlights of their relevance to metamorphic and igneous systems are discussed.