Phlogopite and tetra-ferriphlogopite from Brazilian carbonatite complexes: petrogenetic constraints...
-
Upload
independent -
Category
Documents
-
view
1 -
download
0
Transcript of Phlogopite and tetra-ferriphlogopite from Brazilian carbonatite complexes: petrogenetic constraints...
Phlogopite and tetra-ferriphlogopite from Brazilian carbonatitecomplexes: petrogenetic constraints and implications for
mineral-chemistry systematicsq
J.A. Broda,*, J.C. Gaspara, D.P. de ArauÂjoa, S.A. Gibsonb, R.N. Thompsonc,T.C. Junqueira-Broda
aInstituto de GeocieÃncias, Universidade de BrasõÂlia, 70.910-970, Brasilia, DF, BrazilbDepartment of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EQ, UK
cDepartment of Geological Sciences, University of Durham, South Road, Durham DH1 3LE, UK
Accepted 2 August 2000
Abstract
The results of a mineral-chemistry study of phlogopite and tetra-ferriphlogopite in carbonatites and associated alkaline silicate rocks from
the Brazilian carbonatite complexes of Jacupiranga, Tapira and CatalaÄo are presented. A wide textural variety of mica is found, ranging from
primary magmatic crystals to late-stage metasomatic phases. Primary micas from the carbonatites and from the associated plutonic silicate
rocks follow distinct evolution paths with magma differentiation. In all three complexes, micas from the silicate rocks evolve from phlogopite
towards annite, although in Jacupiranga they are more Al-rich. Among the micas crystallising from carbonatite liquids, those from
Jacupiranga are typically Al- and Mg-rich, while in Tapira and CatalaÄo they are extremely Al-poor. Metasomatic micas bridge the gap
between primary micas from silicate-rocks and from carbonatites. The complete phlogopite±tetra-ferriphlogopite series is reported from the
CatalaÄo and Tapira complexes. This study shows that micas from carbonatite complexes may span a wide compositional range, largely
overlapping the ®elds of micas from several types of alkaline ultrapotassic rocks, especially regarding Ti and Al contents. The use of
phlogopite composition and evolution to discriminate between different types of alkaline rocks should be undertaken with caution. q 2001
Elsevier Science Ltd. All rights reserved.
Keywords: Mineral-chemistry systematics; Phlogopite and tetra-ferriphlogopite; Brazilian carbonatite complexes
1. Introduction
The chemistry of phlogopite from alkaline rocks has often
been used to discriminate between different types of alka-
line rocks and their respective tectonic environments (e.g.
Mitchell and Bergman, 1991; Mitchell, 1995b). In this
context, a remarkable motivation for the study of phlogopite
was its use in the mineral-chemistry systematics of ultrapo-
tassic rocks, especially with a view to identifying kimber-
lites and lamproites, which has obvious implications for the
diamond exploration industry. This paper deals with the
chemistry of phlogopites from Cretaceous Brazilian carbo-
natite complexes (Fig. 1), and investigates the possible
petrologic causes of its widely variable composition. The
implications for mineral chemistry systematics of world-
wide alkaline rocks are also discussed.
2. Chemical variation of trioctahedral mica fromalkaline rocks and carbonatites
The currently accepted (Rieder et al., 1998) ideal end-
member compositions of trioctahedral micas relevant to
this work are given in Table 1.
A number of cation substitutions is known to occur in
trioctahedral micas (Bailey, 1984). Tetrahedral cations are
primarily Si41 and Al31, although Fe31 can substitute for
Al31, a common feature in micas from alkaline igneous
rocks and carbonatites. Additionally, some authors have
argued for the presence of tetrahedral Ti41 (e.g. Farmer
and Boetcher, 1981). The most common cations in the octa-
hedral site are Mg21, Al31, Fe21, and Fe31. Less frequently,
Ti41, Mn21, Li1, and Cr31, among others, can also occupy
octahedral positions. The interlayer site is occupied mostly
Journal of Asian Earth Sciences 19 (2001) 265±296
1367-9120/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S1367-9120(00)00047-X
www.elsevier.nl/locate/jseaes
q This paper is part of the Special Issue: Alkaline and Carbonatitic
Magmatism and Associated Mineralization±Part II. Guest Editors: L.G.
Gwalani, J.L. Lytwyn.
* Corresponding author. Tel.: 155-61-307-2873; fax: 155-61-272-4286.
E-mail address: [email protected] (J.A. Brod).
by K1 and Na1, with Ca21 and Ba21 as possible common
substitutes.
2.1. Tetrahedral cations
Among the possible tetrahedral substitutions,
Fe31, Al31 is of paramount importance in alkaline rocks
and carbonatites. This substitution de®nes the phlogopite±
tetra-ferriphlogopite and annite±tetra-ferri-annite series,
and is commonly indicated by:
(a) strong negative correlation between Fe31 and Al31;
(b) de®ciency in the sum of the common tetrahedral
cations (i.e. Si 1 Al , 8);
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296266
Fig. 1. Geological map showing the main geological units and alkaline occurrences in southern Brazil (after Ulbrich and Gomes, 1981). Carbonatite complexes
discussed in this paper have their names underlined.
Table 1
Names and ideal end-member compositions of some trioctahedral micas
(after Rieder et al., 1998)
End-member name Ideal end-member composition
Annite KFe213 AlSi3O10(OH)2
Phlogopite KMg3AlSi3O10(OH)2
Siderophyllite KFe212 Al Al2Si2O10(OH)2
Eastonite KMg2Al Al2Si2O10(OH)2
Tetra-ferri-annite KFe213 Fe31Si3O10(OH)2
Tetra-ferriphlogopite KMg3Fe31Si3O10(OH)2
(c) excess of octahedral charges, caused by overestimated
Fe21 from electron microprobe analyses.
The presence of (IV)Fe31 is related to the reverse pleochro-
ism �a . b � g� which is typical of tetra-ferriphlogopite.
Farmer and Boetcher (1981) described reversely-pleochroic
mica with Fe2O3 as low as 0.66 wt.% (0.07 (IV)Fe31 p.f.u.),
whilst ArauÂjo (1996) detected a sharp change in the(IV)Fe31/(IV)Al ratio at 0.5 (IV)Fe31 p.f.u, coincident with
pleochroism reversal (see below). Additional evidence for
the presence of tetrahedral Fe31 in tetra-ferriphlogopite is
provided by MoÈssbauer spectroscopy studies (Dyar, 1987;
ArauÂjo, 1996; Lalonde et al., 1996).(IV)Al-de®ciency is typical of micas from lamproites and
orangeites and has been interpreted as a direct consequence
of the peralkalinity of the magma (Mitchell and Bergman,
1991; Mitchell, 1995b). Low Al concentration in the liquid
and/or high f O2 have also been recognised as major
inducing factors for the formation of tetra-ferriphlogopite
(e.g. Arima and Edgar, 1981; Heathcote and McCormick,
1989; Brigatti et al., 1996). Conditions such as these are
fairly common in carbonatites, and are consistent with the
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 267
Table 2
Example of Fe31/Fe21 recalculation (ArauÂjo, 1996) from microprobe analysis of phlogopite. Best overall results are highlighted in bold type. See text for
explanation
C61-7 N1 N2 N3 N4 N5 N6 N7
SiO2 37.870 37.870 37.870 37.870 37.870 37.870 37.870 37.870
TiO2 2.770 2.770 2.770 2.770 2.770 2.770 2.770 2.770
Al2O3 7.730 7.730 7.730 7.730 7.730 7.730 7.730 7.730
Cr2O3 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Fe2O3 4.941 5.647 5.973 6.252 6.535 6.829
FeO 23.930 23.930 19.493 19.362 19.645 20.003 20.387 20.789
MnO 0.310 0.310 0.310 0.310 0.310 0.310 0.310 0.310
MgO 12.630 12.630 12.630 12.630 12.630 12.630 12.630 12.630
Li2O 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
BaO 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
CaO 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Na2O 0.090 0.090 0.090 0.090 0.090 0.090 0.090 0.090
K2O 9.700 9.700 9.700 9.700 9.700 9.700 9.700 9.700
Rb2O 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
H2O 3.640 3.640 3.640 3.640 3.640 3.640 3.640 3.640
F 0.290 0.290 0.290 0.290 0.290 0.290 0.290 0.290
Cl 0.000 0.000 0.000 0.000 0.000 0.000
Total 98.960 98.960 99.464 100.039 100.648 101.285 101.951 102.647
22 O 24 (OH,O,F) 24 (OH,O,F) 24 (OH,O,F) 24 (OH,O,F) 24 (OH,O,F) 24 (OH,O,F)
Si 5.976 5.924 5.898 5.875 5.852 5.828 5.803(IV)Al 1.438 1.426 1.419 1.414 1.408 1.402 1.396(IV)Fe31 0.586 0.650 0.683 0.711 0.740 0.770 0.801
T site 8.000 8.000 8.000 8.000 8.000 8.000 8.000(VI)Al 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Ti 0.329 0.326 0.324 0.323 0.322 0.321 0.319
Cr 0.000 0.000 0.000 0.000 0.000 0.000 0.000(VI)Fe31 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Fe21 2.573 2.482 2.501 2.535 2.572 2.610 2.651
Mn 0.041 0.041 0.041 0.041 0.041 0.040 0.040
Mg 2.970 2.945 2.931 2.920 2.908 2.897 2.884
Li 0.000 0.000 0.000 0.000 0.000 0.000 0.000
O site 5.913 5.794 5.798 5.819 5.843 5.868 5.894
Ba 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Ca 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Na 0.028 0.027 0.027 0.027 0.027 0.027 0.027
K 1.953 1.936 1.927 1.920 1.912 1.904 1.896
Rb 0.000 0.000 0.000 0.000 0.000 0.000 0.000
A site 1.980 1.963 1.954 1.947 1.939 1.931 1.923
O 20.023 20.058 20.076 20.091 20.106 20.122 20.139
OH 3.832 3.799 3.782 3.767 3.752 3.737 3.721
F 0.145 0.143 0.143 0.142 0.142 0.141 0.141
Cl 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Charges (1) 44.441 44.126 44.097 44.105 44.120 44.136 44.152
Charges (2) 244.023 244.058 244.076 244.091 244.106 244.122 244.139
Balance 0.41740 0.06841 0.02105 0.01417 0.01330 0.01333 0.01349
frequent occurrence of tetra-ferriphlogopite in these rocks.
Nonetheless, Dyar (1987) pointed out that there is MoÈss-
bauer spectroscopy evidence for tetrahedral iron in both Al-
poor and -rich micas and Lalonde et al. (1996) reported
micas with (IV)Fe31 formed in an Al-rich environment.
An alternative hypothesis has been proposed to account
for tetrahedral de®ciency, involving the substitution of(IV)Ti41 instead of (IV)Fe31. Farmer and Boetcher (1981)
suggested that Ti41 precedes Fe31 in the order of preference
to occupy tetrahedral positions, with Fe31 entering this site
only if there is still some de®ciency left. Nevertheless, this
is not supported by the data obtained during this research
(see discussion later in the text).
Fe31/Fe21 calculations: The role of tetra-ferriphlogopite
and tetra-ferri-annite can only be properly assessed if Fe31/
Fe21 is known. In the case of microprobe analyses, however,
ferrous and ferric iron cannot be distinguished, and this ratio
has to be estimated. Procedures for iron recalculation invari-
ably rely on a series of assumptions regarding the type of
tetrahedral substitutions prevailing in each case. Dymek
(1983) and Droop (1987) proposed recalculation methods
for the estimation of Fe31/Fe21. However, Dymek's itera-
tive normalisation procedure was devised for the recalcula-
tion of (VI)Al-bearing biotites, which are substantially
different from some of the phlogopites relevant to this
study. Droop's method is not applicable to minerals with
cation vacancies and, as suggested by Mitchell (1995b),
may result in overestimated Fe31.
MoÈssbauer spectroscopy studies (ArauÂjo, 1996) have
shown that Fe31 is the main substituting cation in the tetrahe-
dral position in tetra-ferriphlogopites from the CatalaÄo
complex. Furthermore, crystal-chemistry studies of phlogo-
pites from the Tapira complex (Brigatti et al., 1996) suggested
that (IV)Fe31, (IV)Al is the main tetrahedral substitution in
these micas also. During this research, the microprobe
analyses of Al-de®cient phlogopites were recalculated using
the method suggested by ArauÂjo (1996), as described in the
paragraphs below and exempli®ed in Table 2.
Microprobe data were ®rst recalculated on the basis of 22
oxygen (column N1, Table 2), and an appropriate amount of
Fe was recast as (IV)Fe31 in order to complete the tetrahedral
site occupancy according to the equation:
�IV�Fe31 � 8 2 Si 2 �IV�Al
After adjusting FeO and Fe2O3 to the calculated ratio, the
analyses were recalculated on the basis of 24 oxygen
(column N2, Table 2), with H2O calculated by stoichiometry.
In some cases, the ®rst estimate of Fe2O3 by this procedure still
leaves a small tetrahedral de®ciency. The method may then be
successively repeated (columns N3±N7) until adequate
charge balance and/or analysis total were achieved (column
N4). Further recalculations (column N5) may slightly improve
the charge balance, but result in undesirable increase in the
analysis total. An excessive number of recalculations
(columns N6 and N7) progressively deteriorate both the
charge balance and the analysis total. The good overall results
in column N4 indicate that this procedure can provide an
adequate estimate of Fe31/Fe21 in the studied micas.
It should be noted that this recalculation procedure was
not applied to phlogopites from the Jacupiranga complex
(see below). This is because: (a) many of these micas
contain Al in excess of that necessary to compensate for
tetrahedral Si de®ciency; (b) Gaspar (1989) demonstrated
that when Si 1 Al , 8, Mg enters the tetrahedral site of
Jacupiranga phlogopites through the substitution scheme(VI)Mg 1 (IV)Si, (IV)Mg 1 (VI)Ti.
2.2. Octahedral cations
Substitution of Fe21 for Mg21 in the octahedral site
de®nes the phlogopite-annite series, and is probably the
most common substitution in trioctahedral micas from
silicate igneous rocks. Since Fe21 in phlogopite usually
increases with magma evolution, it can be used to assess
differentiation within an igneous suite. In their investigation
of phlogopite from ultrapotassic rocks, Edgar and Arima
(1983) demonstrated a sympathetic variation between
bulk-rock and phlogopite Fe/(Fe 1 Mg), although the latter
is systematically more magnesian than the coexisting liquid.
Increase in Fe/(Fe 1 Mg) ratio in phlogopites from the
silicate rocks associated with carbonatite complexes is also
a common feature (Gaspar, 1989; Brod, 1999, see discussion
of the Jacupiranga and Tapira carbonatite complexes below).
In carbonatite liquids, however, this relationship is not as
straightforward. McCormick and Le Bas (1996) suggested
that the Fe/Mg ratio of phlogopites crystallised from the
Busumbu and Sukulu carbonatites (Uganda) is controlled
by the co-precipitation of magnetite. They observed an
initial decrease in the Fe/Mg ratio in zoned phlogopite, as
Fe is consumed during the crystallisation of magnetite.
When magnetite ceases to crystallise, the Fe/Mg ratio of
phlogopite increases progressively, leading towards biotite.
They also found Al to decrease in mica, as a result of the
progressively lower availability of this element in the
carbonatite magma, except in the late-stage micas when
Al was added to the magma via country-rock assimilation.
Titanium occurs in high concentrations in alkaline
magmas and phlogopite from leucitites, melilitites and
lamproites can be distinctively Ti-rich (Mitchell and
Bergman, 1991; Mitchell, 1995a). Greenwood (1998)
reported up to 12.5 wt% TiO2 in phlogopites from a lampro-
phyric dyke in the Trindade Island, South Atlantic. The
solubility of Ti in phlogopite may be affected by paragenetic
and/or physicochemical constraints. In the absence of Ti-
bearing oxides phlogopite is the preferential site for TiO2 in
ultrapotassic rocks (Edgar and Arima, 1983). However, if
Ti-magnetite or perovskite form concomitantly with
phlogopite, titanium will partition preferentially to the
former two minerals, and the TiO2 content in phlogopite
may approximate that of the clinopyroxene. Arima and
Edgar (1981) suggested that the solubility of titanium in
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296268
phlogopite is most strongly dependent on the physical
conditions prevailing during crystallisation. According to
their review of experimental data, increasing temperature,
increasing f O2, and decreasing pressure all increase Ti
solubility in phlogopite.
Mn can also substitute for Fe21 in the octahedral site.
Lalonde et al. (1996) described Mn enrichment with Fe/
(Fe 1 Mg) in mica from the Mont Saint-Hilaire intrusion
(Canada), suggesting that the increase of Mn correlates
positively with the amount of tetrahedral Fe31.
2.3. Interlayer site cations
Amongst the common substitutions in the 12-fold co-
ordinated interlayer site, Ba enrichment has often been
described in phlogopites from carbonatites and alkaline
igneous rocks. Ba-rich eastonite (up to 5% Ba) occurring
as overgrowths on previously formed phlogopite in carbo-
natites from Arkansas has been interpreted as a product of
the latest stages of groundmass crystallisation (McCormick
and Heathcote, 1987). Phlogopites from a Trindade Island
lamprophyre contain up to 7.11 wt% BaO (Greenwood,
1998). Gaspar and Wyllie (1982, 1987) found up to 10.3%
BaO in phlogopites from the Jacupiranga carbonatites (see
discussion below). To our knowledge, the highest Ba
contents in phlogopite (13±16 wt% BaO) are those reported
by Seifert and Kampf (1994) for phlogopite in a nephelinite
from Bohemia, which are interpreted as a result of late-stage
Ba enrichment.
3. Phlogopites from Brazilian carbonatite complexes
The chemical and textural variations of phlogopite in
carbonatites and silicate rocks of three Brazilian carbona-
tite-bearing plutonic complexes (Tapira, CatalaÄo and
Jacupiranga) is illustrated below. The complexes were
chosen to give a wide range of mica composition and to
offer maximum petrographic and petrogenetic constraints.
The excellent rock exposures at Jacupiranga, especially at
the phosphate mine area, allow a very tight ®eld and petro-
graphic control, and the easy discrimination between rocks
of magmatic and metasomatic origin. The CatalaÄo complex
represents the opposite extreme, where a pervasive metaso-
matic alteration obliterates many of the primary magmatic
features but, in turn, allows the investigation of phlogopite
crystallised as the result of post-magmatic processes. The
Tapira complex contains well preserved examples of both
primary magmatic rocks and metasomatic alteration, avail-
able from drill cores.
3.1. The Jacupiranga complex
The Jacupiranga complex is located in the valley of the
Ribeira River, SaÄo Paulo State, southeast Brazil. Many
aspects of the Jacupiranga complex have been studied in
detail by several authors (Gaspar and Wyllie, 1982,
1983a,b, 1987; Gaspar, 1989, 1992; Huang et al., 1995;
Mitchell, 1978; Morikiyo et al., 1990; Roden et al., 1985;
Santos and Clayton, 1995). According to Gaspar (1989), the
complex is an elliptical intrusion (10.5 £ 6.7 km2),
composed of two main rock bodies, dunites in the northern
part and magnetite clinopyroxenites in the south. The
magnetite clinopyroxenite body is intruded by a crescent-
shaped body of ijolite and by an elongated carbonatite.
Melteigites, phlogopite clinopyroxenites and nepheline-
bearing clinopyroxenites occur in an elongated region
along the northwest margin of the magnetite clinopyroxe-
nite. Several plagioclase-bearing rock-types surround the
dunite and magnetite clinopyroxenite bodies and crosscut-
ting them near the margins, as dyke swarms and small intru-
sives. These rocks range from andesine-bearing phlogopite
clinopyroxenites and mela-gabbros to quartz-monzonites
and quartz syenites. Veins of medium- to very coarse-
grained nepheline-syenites are widespread in the margins
of the complex. Fenitisation occurs mainly near the margins
of the complex.
Five different intrusions of carbonatites were recognised
(C1±C5, from the oldest to the youngest, Gaspar and
Wyllie, 1983b). Compositionally, C1, C3 and C4 are sovites
(calcite±carbonatite), C2 is a dolomite±sovite and C5 is a
rauhaugite (dolomite±carbonatite). All ®ve intrusions
contain phlogopite, although this mineral is usually present
in minor amounts. A metasomatic reaction zone develops at
the contact between the carbonatites and the host pyroxenite
(jacupirangite). The rock produced in this zone is charac-
terised by an alternation of carbonate and silicate-rich
bands, millimetres to centimetres wide, where phlogopite
comprises 70±90% of the silicate-rich bands. This rock
was informally called ªreaction rockº by Gaspar and Wyllie
(1983b). In the remainder of this text it will be referred to as
ªmetasomatic phlogopititeº, in order to maintain the
consistency with the descriptions of CatalaÄo and Tapira
complexes.
The micas from Jacupiranga silicate rocks and carbonatites
show contrasting composition and evolution. Representative
analyses are given in Table 3.
3.1.1. Micas from silicate rocks
Micas in the Jacupiranga silicate rocks are mostly
biotites, with variable Fe/(Fe 1 Mg) and Ti contents
(Gaspar, 1989). The Fe/(Fe 1 Mg) ratio increases with the
differentiation of the silicate rock; phlogopite from a
pyroxenite yielded the lowest Fe/(Fe 1 Mg) ratio (0.154),
whilst the maximum value was observed in biotite from a
phonolite (0.671). This is the expected behaviour in an
evolution dominated by crystal fractionation processes,
and was also observed in the silicate-rock sequence of the
Tapira complex (see below). TiO2 varies in the opposite
direction. The highest Ti contents observed in Jacupiranga
are from a Mg-biotite in a mela-gabbro (11.2% TiO2), while
the lowest Ti content was observed in a biotite from an
alkali-feldspar syenite. A consistent phlogopite evolution
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 269
J.A.
Bro
det
al.
/Jo
urn
al
of
Asia
nE
arth
Scien
ces19
(2001)
265
±296
27
0
Table 3
Representative analyses of phlogopites from the Jacupiranga complex
Analysisa Carbonatites Silicate rocks
C1 C2 C3 C4 C5 Undersaturate
1 2 3 4 5 6 7 8 9 10 11 12
SiO2 39.70 38.20 31.70 38.10 35.40 38.70 37.30 41.80 40.30 42.90 36.90 36.60
TiO2 0.17 0.28 0.02 0.05 0.01 0.11 0.03 0.00 0.07 0.03 6.04 6.43
Al2O3 15.10 16.50 20.90 16.70 18.30 16.20 17.80 13.40 14.90 11.70 15.50 15.30
FeO 2.75 2.61 1.57 1.83 2.17 2.70 2.80 3.78 2.22 2.44 8.67 11.00
MnO 0.02 0.01 0.00 0.00 0.01 0.01 0.01 0.02 0.01 0.02
MgO 26.10 25.60 22.50 25.70 24.40 25.80 25.30 27.80 27.20 28.20 18.80 16.90
CaO 0.05 0.12 0.27 0.58 0.20 0.05 0.06 0.17 0.07 0.11 0.35 0.40
Na2O 0.21 2.00 0.47 0.12 0.13 0.44 0.41 0.36 1.20 0.51 0.65
K2O 10.20 7.23 6.63 9.59 8.99 9.56 8.94 10.00 8.94 10.30 8.50 9.31
BaO 1.57 3.11 10.30 4.15 5.03 2.54 3.56 0.10 0.89 0.26
Nb2O5
H2O 4.22 4.20 3.93 4.20 4.07 4.21 4.19 4.33 4.27 4.30 4.16 4.11
Total 100.09 99.86 98.29 101.02 98.71 100.32 100.40 101.76 100.07 100.77 99.57 100.05
Cations on the basis of 24 O (OH,F,Cl)
Si 5.636 5.448 4.835 5.438 5.212 5.510 5.337 5.789 5.656 5.977 5.341 5.344(IV)Al 2.364 2.552 3.165 2.562 2.788 2.490 2.663 2.185 2.344 1.920 2.644 2.633(IV)Mg 0.026 0.103 0.015 0.023
T site 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000(VI)Al 0.161 0.219 0.589 0.245 0.385 0.226 0.337 0.000 0.119 0.000
Ti 0.018 0.030 0.002 0.005 0.001 0.012 0.003 0.000 0.007 0.003 0.657 0.706
Fe21 0.327 0.311 0.200 0.218 0.267 0.322 0.335 0.438 0.261 0.284 1.050 1.343
Mn 0.002 0.001 0.000 0.000 0.001 0.001 0.001 0.002 0.001 0.002(VI)Mg 5.524 5.443 5.116 5.469 5.356 5.476 5.397 5.713 5.691 5.754 4.042 3.655
Nb
O site 6.032 6.004 5.907 5.937 6.010 6.037 6.073 6.153 6.079 6.043 5.749 5.704
Ca 0.008 0.018 0.044 0.089 0.032 0.008 0.009 0.025 0.011 0.016 0.054 0.063
Na 0.058 0.553 0.139 0.033 0.037 0.121 0.114 0.097 0.327 0.138 0.182
K 1.847 1.316 1.290 1.746 1.689 1.737 1.632 1.767 1.601 1.831 1.570 1.734
Ba 0.087 0.174 0.616 0.232 0.290 0.142 0.200 0.005 0.049 0.014
A site 2.000 2.061 2.089 2.100 2.048 2.008 1.955 1.894 1.988 1.999 1.806 1.797
Total 16.032 16.065 15.996 16.037 16.058 16.045 16.028 16.047 16.067 16.042 15.555 15.501
Charges (1) 43.992 43.988 43.991 43.988 43.989 43.992 43.990 43.993 43.995 43.995 43.998 44.001
Charges (2) 244.000 244.000 244.000 244.000 244.000 244.000 244.000 244.000 244.000 244.000 244.000 244.000
Balance 20.008 20.012 20.009 20.012 20.011 20.008 20.010 20.007 20.005 20.005 20.002 0.001
J.A.
Bro
det
al.
/Jo
urn
al
of
Asia
nE
arth
Scien
ces19
(2001)
265
±296
271
Table 3 (continued)
Analysisa Silicate rocks Metasomatic
Undersaturate Feldspar-bearing Phlogopitites
13 14 15 16 17 18 19 20 21 22 23 24
SiO2 35.90 35.60 35.70 36.90 34.80 35.80 36.00 37.10 36.30 37.10 41.60 42.50
TiO2 4.50 3.92 1.24 6.78 10.90 7.76 6.48 4.97 6.36 4.23 0.21 0.32
Al2O3 13.30 13.70 14.90 15.40 16.10 15.00 14.40 12.90 13.00 13.70 13.70 8.83
FeO 21.70 20.10 26.60 10.00 10.40 13.70 15.50 18.50 19.10 20.70 2.25 7.81
MnO 1.14 0.01 0.04
MgO 11.70 12.90 7.31 17.30 16.00 14.30 14.20 12.80 11.90 11.50 27.80 27.80
CaO 0.17 0.40 0.18 0.15 0.14 0.07 0.08
Na2O 0.37 0.63 1.40
K2O 9.25 9.39 9.43 9.02 8.30 9.28 9.46 9.10 9.30 9.33 9.67 7.73
BaO 0.73 0.07
Nb2O5
H2O 3.90 3.92 3.80 4.13 4.12 4.05 4.01 3.98 3.95 3.94 4.31 4.24
Total 100.42 99.93 100.12 99.90 100.80 100.04 100.19 99.35 99.91 100.50 100.98 100.82
Cations on the basis of 24 O (OH,F,Cl)
Si 5.494 5.447 5.621 5.359 5.021 5.299 5.369 5.635 5.515 5.622 5.785 6.013(IV)Al 2.399 2.471 2.379 2.636 2.738 2.617 2.531 2.309 2.328 2.378 2.215 1.471(IV)Mg 0.107 0.082 0.005 0.242 0.085 0.100 0.055 0.157 0.516
T site 8.000 8.000 8.000 8.000 8.001 8.001 8.000 7.999 8.000 8.000 8.000 8.000(VI)Al 0.386 0.068 0.028
Ti 0.518 0.451 0.147 0.740 1.183 0.864 0.727 0.568 0.727 0.482 0.022 0.034
Fe21 2.777 2.572 3.502 1.215 1.255 1.696 1.933 2.350 2.427 2.623 0.262 0.924
Mn 0.152 0.001 0.005(VI)Mg 2.562 2.861 1.716 3.740 3.199 3.070 3.057 2.843 2.539 2.598 5.763 5.347
Nb
O site 5.857 5.884 5.903 5.695 5.637 5.630 5.717 5.761 5.693 5.771 6.048 6.310
Ca 0.028 0.066 0.028 0.024 0.022 0.010 0.012
Na 0.104 0.170 0.384
K 1.806 1.833 1.894 1.671 1.528 1.752 1.800 1.763 1.803 1.804 1.715 1.395
Ba 0.040 0.004
A site 1.834 1.899 1.894 1.775 1.556 1.776 1.822 1.763 1.803 1.804 1.935 1.795
Total 15.691 15.783 15.797 15.470 15.194 15.407 15.539 15.523 15.496 15.575 16.011 16.105
Charges (1) 43.999 44.000 44.001 43.999 44.006 44.005 44.001 43.998 44.001 44.000 43.994 43.996
Charges (2) 244.000 244.000 244.000 244.000 244.000 244.000 244.000 244.000 244.000 244.000 244.000 244.000
Balance 20.001 0.000 0.001 20.001 0.006 0.005 0.001 20.002 0.001 0.000 20.006 20.004
a Rock type: 1±8, sovite; 9, rauhaugite; 10, beforsite; 11, magnetite±clinoyroxenite; 12, olivine±phlogopite±nepheline clinopyroxenite; 13, melteigite; 14, ijolite; 15, phonolite; 16, andesine-bearing
clinopyroxenite; 17, mela-gabbro; 18, mela-diorite; 19, diorite; 20, olivine monzonite; 21, monzonite; 22, quartz monzonite; 23, 24, reaction rock (metasomatic phlogopitite).
pattern can be deduced for both the nepheline-bearing
(pyroxenite through ijolite to phonolite) and the plagioclase-
bearing Jacupiranga silicate rocks (gabbro through monzonite
to syenite), despite the signi®cant overlap occurring
between and within these two rock groups. Gaspar (1989)
interpreted the increasing Fe/(Fe 1 Mg) and decreasing
TiO2 as a result of magma evolution (decreasing tempera-
ture of mica crystallisation), but pointed out that this is only
valid if the fractionation processes took place at nearly
constant pressure.
None or very little Fe31 resulted from charge-balance
calculations, indicating that the tetra-ferriphlogopite end
member is not relevant for the evolution of the micas in
Jacupiranga silicate rocks. This is con®rmed in Fig. 2,
where micas from both saturated and undersaturated silicate
rocks follow a trend of increasing Fe and decreasing Mg
with magma evolution, at a level of Al signi®cantly higher
than the phlogopite±annite join.
3.1.2. Micas from carbonatites
Phlogopites from Jacupiranga carbonatites (Gaspar and
Wyllie, 1982, 1987) contrast in composition and evolution
with those from the associated silicate rocks. In the carbo-
natites, the mica is characterised by very low Fe/(Fe 1 Mg)
(0.065±1.2), very low TiO2 (,0.44 wt%), high MgO (22.5±
28.2%), widely variable but occasionally very high BaO
(0.1±10.3 wt%), and Na2O as high as 2.77 wt%.
An important characteristic of micas from Jacupiranga
carbonatites is that they usually have Al in excess of that
necessary to ®ll the tetrahedral site in addition to Si. This
results in the presence of octahedral Al and, therefore, in
signi®cant participation of the eastonite end-member (Fig.
2). As will be seen later in this text, this feature is in sharp
contrast with the observations for magmatic phlogopites in
carbonatites from Tapira and CatalaÄo.
Contrary to the phlogopite from silicate rocks, the MgO
content of the mica increases with decreasing age of the
carbonatite intrusion, from C2 to C5 (Figs. 2 and 3). BaO
progressively decreases in this direction suggesting that in
Jacupiranga carbonatites, Ba enrichment in phlogopite is not
related to late-stage processes, contradicting the observa-
tions of McCormick and Heathcote (1987) for the Arkansas
carbonatites. On a local (grain) scale, Gaspar and Wyllie
(1987) report BaO zoning in both directions (increasing
and decreasing towards the rim), but point out that high-
Ba rims are more common than high-Ba cores. Mica from
the oldest intrusion C1 does not ®t the progressive variation
of the other Jacupiranga carbonatites, showing BaO and
MgO contents similar to the younger C4 and C5.
TiO2 content in micas from C2 to C5 is generally low
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296272
Al
Mg Fe(T)
Eastonite Siderophyllite
Phlogopite Annite
TFP Tetra-ferri-annite
JacupirangaPhlogopites
Carbonatites
SilicaterocksMetasomatic
rocks
Fig. 2. Composition of Jacupiranga micas in terms of Al, Fe and Mg (p.f.u.). The arrows depict the evolution trends for micas from (undersaturated) silicate
rocks, carbonatites and metasomatic micas. Fe(T)� total iron calculated as Fe21; TFP� tetra-ferriphlogopite.B
aO(w
t.%)
MgO(wt.% )
10
8
6
4
2
023 25 26 27 2824
C1
C2
C3
C4
C5 B5
Jacupirangacarbonatitephlogopites
Fig. 3. BaO variation with MgO for micas from Jacupiranga carbonatite.
Modi®ed from Gaspar and Wyllie (1987).
(less than 0.12 wt%), and does not vary systematically with
intrusion age. Again, C1 micas are distinguished by their
higher TiO2 (up to 0.38 wt%), compared to micas from other
Jacupiranga carbonatite intrusions.
The compositional trends of phlogopite from Jacupiranga
carbonatites were interpreted as product of differentiation,
either in the parent carbonatite magma or in an evolving
silicate magma from which successive batches of carbonatite
derived (Gaspar and Wyllie, 1987). It was also concluded that
the general chemical characteristics of these phlogopites are
representative of some of the least evolved phlogopites from
worldwide carbonatites.
3.1.3. Metasomatic micas
Metasomatic phlogopites from the reaction rock associated
with carbonatites C1, C3 and C5 compositionally resemble
those of the carbonatites, in terms of major elements, although
they do not necessarily coincide with the micas from the
respective associated carbonatite intrusion. In general, their
Al2O3 contents are lower and their TiO2 and MgO contents
>are higher than most phlogopites from Jacupiranga carbona-
tites. Some micas of this type may develop greenish colours
and have K2O contents as low as 7.7 wt%, suggesting an
incipient chloritisation.
3.2. The CatalaÄo-I and CatalaÄo-II complexes
The CatalaÄo I and II intrusions are the northernmost known
carbonatite occurrences in the Alto ParanaõÂba Igneous
Province. The two complexes are approximately 10 km
apart and are interpreted as co-genetic bodies comprising an
ultrama®c phase (dunite and clinopyroxenite) and several
carbonatite phases. The carbonatites interacted with the
primary ultrama®c rocks forming carbonate-, phlogopite-
and clinopyroxene-bearing metasomatic rocks. In many
cases, this resulted in metasomatic phlogopitites (Fig. 4),
formed by alternating parallel thin bands of phlogopite and
carbonate. The metasomatic phlogopitites are the equivalent
of the ªreaction rockº described by Gaspar and Wyllie (1983b)
from the Jacupiranga complex. Phoscorites also occur and are
intimately associated with the carbonatite (Gaspar et al.,
1998). In the CatalaÄo-I complex, a breccia with a phlogopite-
and olivine-rich matrix cuts the previously formed rocks.
CatalaÄo I, situated approximately 20 km to the NE of the
city of CatalaÄo, is the largest (27 km2) and best known of the
two complexes (e.g. Baecker, 1983; Danni et al., 1991;
Gaspar and ArauÂjo, 1995; Gaspar et al., 1994; ArauÂjo,
1996). This roughly circular-shaped, multi-stage intrusion
domed the Late-Proterozoic schists and quartzites of the
Araxa Group and was dated 85 ^ 6.9 Ma. (recalculation
by Sonoki and Garda, 1988 of a K/Ar age by Hasui and
Cordani, 1968). The rock-types occurring in the complex
are mainly dunites, clinopyroxenites (bebedourites), carbo-
natites and phoscorites. Carbonatite occurs as a central
massive sovite body, as well as widespread dykes and
veins, whilst the ultrama®c and metasomatic rocks dominate
the external portions of the complex. Gibson et al. (1995)
reported the occurrence of phlogopite±picrite dykes up to
5 m thick from drill cores of CatalaÄo I. Important deposits of
phosphate, niobium, rare-earth elements, titanium and
vermiculite are present in CatalaÄo I (Carvalho and Bressan,
1981; Gierth and Baecker, 1986), the complex is currently
mined for apatite and pyrochlore.
Brecciation and fenitisation of the country rock are
conspicuous, resulting in the formation of orthoclase,
aegirine and riebeckite in the fenitised quartzites. Danni et
al. (1991) described the occurrence of aegirine-bearing
nepheline syenites in the southern and western borders of
the complex, but pointed out that these rocks appear to grade
to the fenites. A metasomatic origin for these syenites was
also favoured by Carvalho (1974); Carvalho and Bressan
(1981). The dominance of phlogopitites over the other
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 273
Fig. 4. Drill core section of a banded metasomatic phlogopitite from CatalaÄo. This rock is typical of the contact between carbonatite and silicate ultrama®c
rocks, and occurs in the Jacupiranga and Tapira complexes also. In many cases the banded phlogopitite grades into both the carbonatite, to one side, and into
the unaltered ultrama®c rock, to the other side of the contact.
rock-types testi®es to the particularly intense metasomatic
processes that affected this complex.
The complex of CatalaÄo II is intrusive in the meta-
sedimentary rocks of the Araxa Group. The shape of
this 18 km2 complex is irregular, sinuous, elongated in
the NE±SW direction. Machado Junior (1992a)
described the rock-types present in the CatalaÄo II
complex, as follows: (a) pyroxenites composed of
augite, biotite, apatite, magnetite, zircon and accessory
amphibole, K-feldspar, sphene and calcite; (b) felsic
rock-types comprise quartz-and alkali feldspar±syenite,
locally grading to more ma®c varieties (up to 30% of
predominantly sodic pyroxene, amphibole and mica);
(c) carbonatites include petrographic varieties of
sovites, silicocarbonatites and beforsites; (d) lampro-
phyre occurs as thin dykes with olivine phenocrysts
set in a phlogopite±carbonate groundmass; (e) phoscor-
ites are a product of metasomatic alteration of the
ultrama®c rocks. The phoscorites were later reinter-
preted as magmatic by Melo (1999). Machado Junior
(1992b) obtained a Rb±Sr age of 83.4 ^ 0.9 Ma for
CatalaÄo II.
3.2.1. Primary mica from silicate rocks
The pervasive metasomatic processes acting on the
silicate plutonic rocks at CatalaÄo have a direct impact on
phlogopite compositions, making it dif®cult to assess the
magmatic trends of phlogopite evolution. Nevertheless,
possibly magmatic (and/or intercumulus?) phlogopite is
rarely preserved as relict nuclei in metasomatic tetra-
ferriphlogopite in some clinopyroxenites and phlogopitites.
These cores have normal pleochroism �a , b � g�; and a
sharp contact with the reversely pleochroic �a . b � g�rims. As a rule, the cores of these strongly zoned micas
are enriched in Fe21, Al, Ti and Ba, while the rims are
enriched in Mg, Fe31 and Si. These major chemical changes
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296274
Table 4
Representative analyses of CatalaÄo phlogopites
Analysis Phlogopitite Carbonatite
1 2 3 4 5 6 7 8 9 10 11 12
SiO2 39.83 38.39 39.78 39.64 40.03 40.87 40.71 39.92 41.08 39.83 40.56 40.85
TiO2 0.61 2.58 1.60 0.52 0.67 0.32 0.10 2.29 0.08 0.43 0.06 0.02
Al2O3 0.69 12.13 8.61 2.20 3.13 6.13 3.88 10.06 4.56 2.34 0.13 0.12
Cr2O3 0.00 0.00 0.00 0.00 0.00 0.02 0.01 0.00 0.00 0.00 0.00 0.00
Fe2O3 14.87 1.59 5.10 13.02 11.89 8.29 11.47 1.22 9.56 12.67 15.57 15.47
FeO 9.21 12.58 9.59 9.36 7.86 2.23 2.09 16.28 2.30 7.01 4.28 4.03
MnO 0.01 0.09 0.08 0.05 0.05 0.02 0.02 0.23 0.00 0.06 0.08 0.05
MgO 20.33 18.25 20.41 20.25 21.33 25.95 26.12 15.12 25.36 21.66 23.76 24.05
BaO 0.00 0.18 0.02 0.00 0.00 0.00 0.06 0.00 0.00 0.00 0.21 0.00
CaO 0.00 0.00 0.00 0.25 0.35 0.00 0.00 0.03 0.23 0.36 0.00 0.05
Na2O 0.18 0.06 0.09 0.18 0.17 0.06 0.02 0.08 0.19 0.07 0.07 0.14
K2O 9.86 10.13 10.27 10.05 9.94 10.53 10.28 9.65 10.45 10.05 10.24 10.11
H2O 3.67 3.84 3.73 3.63 3.58 3.78 3.78 3.51 3.61 3.63 3.65 3.63
F 0.46 0.42 0.62 0.56 0.77 0.60 0.54 0.93 0.87 0.57 0.58 0.65
Cl 0.00 0.01 0.01 0.01 0.02 0.02 0.02 0.00 0.01 0.01 0.03 0.01
OyF 0.194 0.176 0.263 0.234 0.324 0.251 0.228 0.393 0.368 0.240 0.244 0.272
OyCl 0.000 0.003 0.003 0.002 0.004 0.006 0.004 0.000 0.001 0.003 0.007 0.003
Total 99.915 100.410 100.190 99.962 100.107 99.075 99.347 99.690 98.667 98.947 99.457 99.453(IV)Si 6.146 5.700 5.919 6.095 6.080 6.018 6.041 6.061 6.125 6.114 6.188 6.209(IV)Al 0.126 2.122 1.510 0.398 0.561 1.064 0.678 1.800 0.802 0.423 0.024 0.021(IV)Fe31 1.727 0.178 0.571 1.507 1.359 0.919 1.281 0.139 1.073 1.464 1.788 1.769
T site 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000(VI)Al 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000(VI)Ti 0.071 0.287 0.179 0.060 0.077 0.035 0.011 0.262 0.009 0.049 0.007 0.002
Cr 0.000 0.000 0.000 0.000 0.000 0.002 0.001 0.000 0.000 0.000 0.000 0.000
Fe31 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Fe21 1.189 1.562 1.193 1.203 0.998 0.274 0.260 2.067 0.286 0.900 0.546 0.513
Mn 0.001 0.011 0.009 0.007 0.006 0.003 0.003 0.030 0.000 0.008 0.010 0.007
Mg 4.678 4.039 4.527 4.642 4.831 5.695 5.778 3.423 5.636 4.955 5.404 5.449
O site 5.939 5.899 5.908 5.912 5.911 6.009 6.053 5.782 5.932 5.913 5.966 5.971
Ba 0.000 0.010 0.001 0.000 0.000 0.000 0.003 0.000 0.000 0.000 0.013 0.000
Ca 0.000 0.001 0.000 0.040 0.056 0.000 0.000 0.004 0.037 0.059 0.000 0.007
Na 0.054 0.018 0.026 0.054 0.050 0.016 0.005 0.023 0.055 0.021 0.020 0.041
K 1.941 1.918 1.950 1.972 1.926 1.978 1.946 1.869 1.988 1.969 1.993 1.961
A site 1.995 1.947 1.977 2.067 2.032 1.994 1.954 1.897 2.079 2.048 2.025 2.010
O 20.000 20.000 20.000 20.000 20.000 20.000 20.000 20.000 20.000 20.000 20.000 20.000
OH 3.775 3.800 3.703 3.727 3.626 3.716 3.741 3.551 3.587 3.720 3.712 3.686
F 0.225 0.197 0.294 0.270 0.369 0.277 0.254 0.449 0.412 0.277 0.280 0.311
Cl 0.000 0.003 0.004 0.002 0.004 0.006 0.005 0.000 0.001 0.003 0.008 0.003
and the corresponding variation in optical properties are
always sharp and, as such, are not likely to have been
produced by ªnormalº magmatic evolution processes (e.g.
fractional crystallisation). A carbonatitic breccia from
CatalaÄo also contains mica with a normal pleochroic core
and a reversely pleochroic rim. The breccia shares some of
the features of the lamprophyric (correlated to the Tapira
phlogopite picrites, see below) magmatism in CatalaÄo
(ArauÂjo, 1996).
In some coarse-grained clinopyroxenites, phlogopite
occurs as euhedral crystals showing no signs of reaction,
which can be interpreted as of magmatic origin. The repla-
cement of pyroxene by amphibole or phlogopite, and the
replacement of amphibole by phlogopite are certainly reac-
tional features. However, in some cases it is not clear
whether phlogopite results from the reaction of pyroxene
with the intercumulus liquid or with newly introduced
(metasomatic) material. A good correlation between Fe
and Mg is observed in these micas, and is consistent with
the trends observed for the magmatic phlogopite in Tapira
silicate rocks.
3.2.2. Primary micas from carbonatites
Phlogopite is relatively rare in CatalaÄo carbonatites.
Two samples were analysed by ArauÂjo (1996). These
micas have low TiO2 (0.02±0.19 wt%) and low Al2O3
(0.11±4.8 wt%). The MgO content ranges from 23.7 to
25.2 wt%), the Fe21/(Fe21 1 Mg) ratio varies from 0.04
to 0.09 and the Fe31/(Fe31 1 Al) from 0.55 to 0.98.
Texturally, they appear to be in equilibrium with the
carbonatite, and have therefore been interpreted as of
primary magmatic origin (ArauÂjo, 1996). Phlogopite
from the phoscorites associated with the carbonatite
shows the same chemical characteristics as the carbona-
tite micas.
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 275
Table 4 (continued)
Analysis Pyroxenite Breccia
13 14 15 16 17 18 19 20 21 22 23 24
SiO2 38.71 37.89 39.64 39.38 37.64 37.66 41.08 41.11 39.52 40.31 38.52 40.45
TiO2 1.32 1.71 1.36 1.04 0.79 0.89 0.44 0.25 2.02 0.18 3.08 0.15
Al2O3 9.24 10.47 7.99 6.81 1.42 1.37 6.48 5.36 10.48 3.71 12.62 3.71
Cr2O3 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00
Fe2O3 4.26 3.44 5.09 6.89 13.64 14.28 7.99 9.36 4.40 11.75 2.25 12.20
FeO 12.32 13.22 11.58 11.91 16.81 15.70 2.61 2.23 4.49 3.02 5.67 2.68
MnO 0.21 0.22 0.31 0.40 0.43 0.40 0.00 0.01 0.03 0.07 0.07 0.09
MgO 18.69 18.00 18.90 18.79 14.74 15.60 25.75 25.97 24.14 25.40 22.60 25.88
BaO 0.00 0.00 0.04 0.18 0.00 0.04 0.00 0.19 0.27 0.12 0.43 0.10
CaO 0.00 0.08 0.00 0.09 0.00 0.02 0.00 0.00 0.05 0.08 0.00 0.01
Na2O 0.08 0.11 0.03 0.04 0.02 0.07 0.05 0.05 0.01 0.03 0.00 0.04
K2O 10.04 10.00 9.88 9.83 9.45 9.49 10.72 10.62 10.10 9.93 10.08 10.26
H2O 3.87 3.87 3.69 3.65 3.53 3.49 4.00 3.96 3.94 3.67 3.92 3.68
F 0.18 0.18 0.58 0.59 0.37 0.53 0.21 0.25 0.42 0.73 0.44 0.78
Cl 0.00 0.00 0.00 0.03 0.02 0.03 0.01 0.01 0.01 0.00 0.01 0.01
OyF 0.075 0.075 0.242 0.250 0.157 0.224 0.089 0.107 0.176 0.305 0.185 0.328
OyCl 0.000 0.000 0.001 0.006 0.005 0.007 0.003 0.002 0.001 0.000 0.002 0.002
Total 99.013 99.262 99.342 99.903 99.036 99.819 99.454 99.471 100.031 99.294 99.882 100.371
Si 5.864 5.739 5.996 5.990 6.074 6.023 6.006 6.037 5.729 6.024 5.594 5.992(IV)Al 1.650 1.869 1.424 1.221 0.271 0.259 1.116 0.928 1.790 0.654 2.159 0.648(IV)Fe 0.486 0.392 0.580 0.789 1.655 1.719 0.878 1.035 0.480 1.321 0.246 1.360
T site 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000(VI)Al 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Ti 0.151 0.195 0.154 0.118 0.095 0.107 0.049 0.028 0.220 0.021 0.336 0.017
Cr 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.000 0.000 0.000 0.000 0.000(VI)Fe31 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Fe21 1.561 1.674 1.465 1.514 2.268 2.100 0.319 0.273 0.545 0.377 0.689 0.332
Mn 0.026 0.029 0.040 0.052 0.059 0.054 0.000 0.001 0.004 0.009 0.008 0.011
Mg 4.220 4.063 4.263 4.261 3.545 3.720 5.613 5.687 5.217 5.659 4.894 5.716
O site 5.959 5.961 5.922 5.945 5.967 5.980 5.981 5.989 5.985 6.066 5.927 6.076
Ba 0.000 0.000 0.002 0.011 0.000 0.003 0.000 0.011 0.015 0.007 0.024 0.006
Ca 0.000 0.012 0.000 0.015 0.000 0.003 0.000 0.000 0.007 0.012 0.000 0.002
Na 0.024 0.032 0.010 0.012 0.008 0.021 0.015 0.014 0.003 0.008 0.000 0.010
K 1.940 1.932 1.907 1.907 1.946 1.937 2.000 1.990 1.868 1.893 1.868 1.939
A site 1.965 1.977 1.919 1.945 1.954 1.964 2.015 2.015 1.893 1.920 1.893 1.957
O 20.000 20.000 20.000 20.000 20.000 20.000 20.000 20.000 20.000 20.000 20.000 20.000
OH 3.915 3.915 3.724 3.707 3.804 3.723 3.899 3.880 3.807 3.657 3.796 3.633
F 0.085 0.085 0.275 0.286 0.190 0.269 0.098 0.118 0.192 0.343 0.202 0.365
Cl 0.000 0.000 0.001 0.007 0.006 0.008 0.003 0.002 0.001 0.000 0.002 0.002
3.2.3. Metasomatic micas
The intensity of the carbonatite metasomatism on the
ultrama®c plutonic rocks is much higher at CatalaÄo than at
Tapira and Jacupiranga, making this an ideal locality to
study the metasomatically driven changes in phlogopite
composition. It was from CatalaÄo that ArauÂjo (1996)
reported the ®rst known occurrence of the complete
phlogopite±tetra-ferriphlogopite series. Two categories of
replacement textures involving phlogopite are observed in
the rocks from CatalaÄo.
The ®rst consists of localised replacement of pyroxene by
phlogopite in the pyroxenites. Some of the phlogopites
formed in this way are Al-rich, show normal pleochroism
and plot along the phlogopite±annite series. Although the
contacts between phlogopite and clinopyroxene are clearly
reactional, they do not imply necessarily in the introduction
of extraneous (carbonatitic) material, since there is no
evidence for a pervasive alteration in the remainder of the
rock. Instead, the textural and chemical evidence suggests
that these micas are products of the crystallisation of the
intercumulus liquid.
The second category is much more common and
clearly related to carbonatite metasomatism. This type
of mica is Al-poor, Ti-poor, and Fe31-rich. Its metaso-
matic origin is con®rmed by: (a) abundant evidence of
textural disequilibrium involving minerals other than the
phlogopite; (b) presence of interstitial carbonate in the
rock; (c) association with stockworks of carbonatite
dykes and veins, with development of reaction margins;
(d) spatial association with carbonatite intrusions; (e)
presence of preserved relicts of high-Al cores and the
sharp changes in chemical and optical properties from
core to rim; (f) substitution of high-Al phlogopite, by
Al-poor, Fe31-rich phlogopite along cleavages and
fractures.
3.2.4. The complete phlogopite±tetra-ferriphlogopite series
and the relationship of composition to optical properties
Representative analyses of phlogopites from CatalaÄo are
given in Table 4, and their distribution in terms of Mg, Fe
and Al is shown in Fig. 5. When compared with Jacupiranga
phlogopites (Fig. 2), these micas show substantially lower
Al contents, which drives them towards the tetra-ferriphlo-
gopite and tetra-ferri-annite end members. Because of the
extensive metasomatic alteration observed in CatalaÄo rocks,
the distinction between magmatic and metasomatic trends in
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296276
Al
Mg Fe (T)
Eastonite Siderophyllite
Phlogopite Annite
TFP Tetra-ferri-annite
Catalão IPhlogopites
Fig. 5. Al, Mg and total Fe distribution of CatalaÄo phlogopites. Note that
most samples plot below the phlogopite-annite line, due to Al de®ciency.
Also noteworthy is the complete range of composition along the phlogopite
tetra-ferriphlogopite line, and various less well-de®ned trends extending
towards Al-poor mica at different Fe/Mg ratios.
5
5.5
6
6.5
0 0.5 1 1.5 2 2.5 3
Al(IV)
Si
(IV
)
Phlogopite- Tetra-ferriphlogopite series
Catalão I Complex
Normal Pleochroism
Reverse Pleochroism }}
Fig. 6. Si versus Al (p.f.u.) variation in the tetrahedral site of phlogopites from CatalaÄo.
phlogopite composition is not easily established on a
complex-wide scale. Therefore, we opted for not attempting
to distinguish such trends in Fig. 5. Nevertheless, the overall
distribution of CatalaÄo micas in the diagram is similar to that
observed in Tapira, and the relative roles of magmatic
evolution and metasomatism will become clear when we
discuss Tapira micas, later in this work.
The chemical composition of CatalaÄo phlogopites is
strongly related to their optical properties. They can be
divided into two groups, based on chemical and optical
criteria: (a) high-Al, -Ti and low-Si, -Mg and -Fe(t)
phlogopites, with normal pleochroism and (b) high-Fe(t)
and -Si, low-Al and -Ti tetra-ferriphlogopites with reverse
pleochroism. The ranges of compositional variation of Si,
Al and Fe throughout the complete phlogopite±tetra-
ferriphlogopite series, and their relation to the reversed
pleochroism are shown in Fig. 6.
Si plays an important role in cationic substitutions of the
normal phlogopites. Fig. 6 shows that in phlogopites with
normal pleochroism there is an important negative correla-
tion between Si and Al in the tetrahedral site. At approxi-
mately (IV)Al� 1.5 p.f.u. (coincident with the change in
pleochroism), this correlation disappears entirely, and a
strong correlation between tetrahedral Al and Fe prevails.
In both cases, there is a small dispersion of the data near the
zero values of (IV)Al.
The high (IV)Al content is accompanied by high (VI)Ti and
low (VI)Mg, suggesting that the most important coupled
substitution in high-Al members of the series could be�VI�Ti41 1 2�IV�Al31 , �VI�Mg21 1 2�IV�Si41
: On the other
hand, the reversely-pleochroic tetra-ferriphlogopites show a
very weak negative correlation between (IV)Si and (IV)Al, but(IV)Fe and (IV)Al are strongly antipathetic, a feature lacking in
normal Al-rich phlogopites (Fig. 7). Furthermore, the more
pronounced (VI)Fe21, (VI)Mg21 substitution in the tetra-
ferriphlogopites led ArauÂjo et al. (1998) to propose the coupled
substitution �IV�Fe31 1 �VI�Fe21 , �VI�Mg21 1 �IV�Al31 for
the high Fe31 members of the series.
3.2.5. MoÈssbauer evidence
Three phlogopite specimens from CatalaÄo were analysed
at the MoÈssbauer Spectrometry Facility of the University of
EspõÂrito Santo, Brazil. The samples were submitted to g-ray
absorption using a 57Co/Rh source and pattern transmission
geometry. The preliminary results and interpretation were
presented in ArauÂjo et al. (1996, 1998) and are discussed
below. The spectrum obtained for one of the samples is
presented in Fig. 8, and the corresponding MoÈssbauer para-
meters are given in Table 5.
The resulting MoÈssbauer spectrum is typically complex,
comprising two strong central lines formed by doublet
compositions. Five different Fe-bearing sites were identi®ed
(I±III, Ib, and Ic, Table 5). The sites I±III are octahedral (III
is signi®cantly distorted) and contain 61.6% of all Fe present,
in the form of Fe21. Site Ib is identi®ed as a tetrahedral position
occupied by Fe31, as expected for micas belonging to the
phlogopite±tetra-ferriphlogopite series.
The interpretation of the second Fe31-bearing site (Ic) is
not yet clear. It is unlikely that this site is tetrahedral, since
the combined average amounts of Si and Al in phlogopites
from this sample only allow for a limited amount of (IV)Fe31
(21.6% of total iron content). A second alternative would be
to interpret Ic as a site occupied by octahedral Fe31.
To gain further insight on this issue, the relative propor-
tions of Fe21 and Fe31 were recalculated considering: (a) the
percentage of ferric iron contained in the tetrahedral Ib site
only; and (b) the sum of ferric iron percentages in Ib and Ic.
The ratio of Fe31 to total iron for each case was then updated
as Fe2O3 in the original electron microprobe analysis.
Finally, the corresponding formulae were calculated on
the basis of 24 O (OH, F, Cl). The results of this procedure
are presented in Table 6. The formula obtained by direct
stoichiometric calculation (column 1) and that calculated
considering the Fe31 content in the Ib site only (column 2)
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 277
0
0.5
1
1.5
2
2.5
0 0.5 1 1.5 2 2.5
Al(IV)
Fe
(IV
)
Phlogopite- Tetra-ferriphlogopite series
Catalão I Complex
Normal Pleochroism
Reverse Pleochroism
}}Fig. 7. Al versus Fe31 (p.f.u.) variation in the tetrahedral site of phlogopites
from CatalaÄo.
Fig. 8. MoÈssbauer spectrum for a single phlogopite crystal from CatalaÄo I
complex (sample C61), after ArauÂjo et al. (1996)
are roughly similar, suggesting that the assumptions made for
stoichiometric calculations are consistent. On the other hand,
results in column 3 (considering ferric iron from sites Ib and Ic)
resulted in a signi®cant amount of octahedal Fe31 and showed
the highest octahedral vacancy.
A third possible interpretation is that Fe31 in the Ic site
represents superparamagnetic hematite microcrystallites
(average diameter ,100 AÊ ). This hypothesis would satis-
factorily explain the likelihood of results in columns 1 and
2 (Table 6), besides not requiring large amounts of ferric
iron to be allocated to octahedral positions. Neither would it
imply signi®cant increase in octahedral vacancies. If this
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296278
Table 5
MoÈssbauer parameters of subspectra of CatalaÄo phlogopite (sample C61) at
room temperature (after ArauÂjo et al. (1996))
Site d a DEqb G Areac Remarks Site type
I 1.11 2.66 0.30 35.8 Mica (A) Octahedral
II 1.03 2.39 0.31 17.4 Mica (B) Octahedral
III 0.58 0.55 0.30 8.4 Mica (C) Octahedral
Ib 0.01 0.14 0.30 20.8 Fe31 Tetrahedral
Ic 0.21 0.47 0.30 17.6 Fe31 ?
a Isomer shift, d in mm/s, relative to a-Fe.b Quadrupole splitting, DEq in mm/s.c Percent area.
Table 6
Calculated formulae (24 O, OH, F, Cl) of phlogopite from sample c61. Column 1 shows the results of direct recalculation from electron microprobe analyses,
after adjustment of Fe21/Fe31 to complete the tetrahedral site occupancy. Column 2 shows the calculated formula after updating the amount of Fe31 contained
in the tetrahedral Ib site (see interpretation of the MoÈssbauer spectrum, above) as Fe2O3 in the original analysis. Column 3 shows the calculated formula
considering all Fe31 detected by MoÈssbauer spectroscopy (i.e. the sum of Ib and Ic sites)
1 2 3
Stoichiometry recalculation MoÈssbauer (Ib) MoÈssbauer (Ib 1 Ic)
SiO2 37.87 37.87 37.87
TiO2 2.77 2.77 2.77
Al2O3 7.73 7.73 7.73
Fe2O3 0.00 6.73 10.21
FeO 23.93 17.87 14.74
MnO 0.31 0.31 0.31
MgO 12.63 12.63 12.63
Na2O 0.09 0.09 0.09
K2O 9.70 9.70 9.70
H2O 3.64 3.73 3.75
F 0.29 0.29 0.29
Cl 0.00 0.00 0.00
Total 98.89 99.60 99.97
OyF 0.12 0.12 0.12
OyCl 0.00 0.00 0.00
Si 5.875 5.889 5.835(IV)Al 1.414 1.417 1.404(IV)Fe31 0.711 0.695 0.762
T site 8.000 8.000 8.000
(VI)Al 0.000 0.000 0.000
Ti 0.323 0.324 0.321(VI)Fe31 0.000 0.093 0.422
Fe21 2.535 2.324 1.900
Mn 0.041 0.041 0.040
Mg 2.920 2.928 2.901
O site 5.819 5.710 5.584
Na 0.027 0.027 0.027
K 1.920 1.924 1.907
A site 1.947 1.951 1.933
Total 15.766 15.661 15.517
#O 20.091 20.000 20.000
#OH 3.767 3.857 3.859
#F 0.142 0.143 0.141
#Cl 0.000 0.000 0.000
Charge (1) 44.106 44.000 44.000
Charge (2) 244.091 244.000 244.000
Balance 0.015 0.000 0.000
%Fe31/Fe(T) 21.9 25.3 38.4
interpretation is correct, site Ic must be disregarded and the
ratio of Fe31 to total iron in the mica recalculated to 25.2%.
This value is not unrealistically higher than the one obtained
by stoichiometric calculations (21.9%).
The detection of Fe31 in the tetrahedral site of CatalaÄo
micas con®rms the existence of the tetra-ferriphlogopite
end-member in the studied rocks. The wide measured
range of Al in these micas characterises the occurrence of
a continuous solid solution along the phlogopite±tetra-
ferriphlogopite series, as reported by ArauÂjo (1996) and
ArauÂjo et al. (1998).
3.3. The Tapira complex
The Tapira complex is the southernmost of the Late-
Cretaceous, carbonatite-bearing alkaline plutonic
complexes which, together with kamafugites, lamproites
and kimberlites, form the Late-Cretaceous Alto ParanaõÂba
Igneous Province (APIP, Gibson et al., 1995). Economic
concentrations of Ti, P, Nb, REE and vermiculite are asso-
ciated with the development of a thick weathering cover.
The complex consists dominantly of bebedourite (a plutonic
rock formed mainly by diopsidic pyroxene and variable
amounts of phlogopite, perovskite, apatite, magnetite, Ti-
garnet and rare sphene), with subordinate carbonatite,
serpentinite (dunite), syenite and ultrama®c ultrapotassic
dykes. Bebedouritic pegmatites are common. Rare melilito-
lites (uncompahgrites) are found near the northeast margin
of the complex.
The coarse-grained silicate rocks were collectively
named Silicate Plutonic Series (SPS, Brod, 1999), further
subdivided into ultrama®c rocks (B1 unit in the centre and
B2 unit in the northern margin of the complex) and syenites.
Drill core relationships and the mineral chemistry of olivine,
clinopyroxene, perovskite, opaque minerals and phlogopite
indicate a general path of magmatic evolution in the sense
B1) B2) syenites for the SPS as a whole. Within the B1
and B2 units, there is a recurrent evolution sequence in the
sense (dunite)) wehrlite) bebedourite.
In the ultrama®c sequence, crystal accumulation
processes produced rock-types richer in olivine, perovskite,
magnetite or apatite. Of these, only olivine-rich facies
(dunites and wehrlites) are discussed individually in this
text. The remainder is considered as modal variations of
the bebedourite. This approach is justi®ed by the mineral
chemistry of key mineral phases (Brod, 1999) and by the
typical lack of olivine and chromite in bebedourites (and in
their modal variations), indicating that these rocks are
relatively evolved cumulates.
Syenites occur either as angular fragments in carbonatite±
syenite breccias or as independent intrusions. They are
essentially composed of K-feldspar plus phlogopite and/or
aegirinic pyroxene, with accessory zircon and sphene.
The carbonatites range from plugs through dykes to small
veinlets and were sub-divided into the C1±C5 units, accord-
ing to their location in the complex, petrographic and
compositional features. C1 is the largest carbonatite body,
occupying the centre of the complex. C2 is a smaller intru-
sion located to the northwest of C1. C3 and C4 are small
plugs occurring at the northern and southern margins,
respectively, and C5 comprises a set of dykes and veins
scattered throughout the complex. Carbonatite intrusion in
the SPS often leads to the transformation of ultrama®c rocks
into banded phlogopitites, and to the production of breccias
ranging in style from magmatic stoping to more explosive
diatreme-facies. Three compositional types of carbonatite
are recognised: (1) Sovites in C1, C3 and C4, (2) Dolomite
sovites in C1 and C2; and (3) Beforsites in C5. Tapira carbo-
natites were produced by an intricate combination of liquid
immiscibility and fractional crystallisation processes, which
is clearly recorded in the mineral chemistry and whole-rock
geochemical signatures of both carbonatites and their
silicate counterparts (Brod, 1999).
All types of Tapira plutonic rocks are crosscut by ®ne-
grained ultrama®c dykes. These are usually a few
centimetres or tens of centimetres thick, rarely exceeding
one metre and can be divided, on the basis of chemical and
mineralogical criteria, into phlogopite-picrite and bebedourite
dykes. Phlogopite-picrites are the most primitive rocks in
Tapira, consisting of olivine phenocrysts set in a carbonate-
phlogopite-rich groundmass. Bebedourite dykes are slightly
more evolved, containing rare phenocrysts of phlogopite,
clinopyroxene and/or apatite, set in a groundmass composed
of these phases plus carbonate and magnetite. The parental
character and the kamafugitic af®nity of the ultrama®c
dykes were established by Brod (1999) and Brod et al.
(2000), thus providing a strong link between the carbonatite-
bearing plutonic complexes and the voluminous kamafugitic
volcanism in APIP.
3.3.1. Tapira phlogopites
Two main phlogopite types occur. In the SPS the most
common micas belong to the phlogopite±annite series,
whilst in carbonatites and metasomatic phlogopitites they
have an important tetra-ferriphlogopite±tetra-ferri-annite
component. Representative microprobe analyses are
reported in Table 7.
Fig. 9 illustrates the compositional variation of mica by
rock group, in terms of the relevant end-members. Starting
from magnesian phlogopite, the following compositional
trends are depicted: (a) variation along the phlogopite±
annite join, marked by micas from the SPS and the cores
of phlogopites from some carbonatites; (b) a trend towards
tetra-ferri-annite, de®ned by the core-to-rim evolution of
phlogopite in some phlogopite-picrites and bebedourite
dykes; (c) a trend towards tetra-ferriphlogopite, de®ned by
micas from carbonatites, some phlogopite-picrites and the
rims of micas from metasomatic phlogopitites. Cores of the
latter are not depicted in the ®gure, but are roughly coin-
cident with the ®eld for wehrlites. The ®eld for carbonatite
micas includes both xenocrystic and primary phlogopite
(see discussion later in the text).
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 279
3.3.2. Silicate plutonic rocks series (SPS)
The SPS phlogopites show a trend of increasing Fe21 and
decreasing Mg21 with magmatic evolution (Fig. 9).
Although some of the groups slightly overlap, there is a
clear progression from magnesian phlogopite (Fe21/
(Fe21 1 Mg) as low as 0.07) in wehrlites towards biotite
(Fe21/(Fe21 1 Mg) up to 0.49) in the more evolved syenites.
There is little solid solution towards more Al-rich compo-
sitions, but it must be stressed that it does not represent a
true eastonite±siderophyllite component, because all
aluminium present is (IV)Al. The slight deviation from the
phlogopite-annite line towards the Al corner is probably the
combined result of Al, Si variation in the tetrahedral site
and normalisation of the variables to 100% in the triangular
diagram. The slight deviation of the trend from the ideal
phlogopite±annite line (Fig. 9) may be accounted for by
higher amounts of Ti (up to 3.69 wt% TiO2) and Mn (up
to 1.27 wt% MnO) in micas from more evolved rocks.
The tetrahedral cations Si, Al and Fe31 do not vary
substantially with Fe21/(Fe21 1 Mg). In fact, the differences
between rock types are always smaller than the magnitude
of the internal variation in each group. This suggests that the
phlogopite±tetra-ferriphlogopite substitution is of little or
no importance in the SPS. Indeed, as for phlogopites in
Jacupiranga silicate rocks, the recalculation of microprobe
analyses resulted in little or no Fe31.
Ti increases in phlogopite with the Fe21/(Fe21 1 Mg)
ratio, both between and within the rock groups, although
the individual trends overlap considerably (Fig. 10). The
lowest Ti was observed in a wehrlite (0.078 p.f.u.) and the
highest in a syenite (0.433 p.f.u.).
The TiO2 increase in phlogopite, from wehrlites to
syenites, contrasts with the observations from Jacupiranga.
This behaviour is dif®cult to reconcile with the tendency of
Ti solubility in phlogopite to decrease with decreasing
temperatures. Only on a local scale temperature appears to
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296280
Table 7
Representative analyses of Tapira phlogopites. Rock types: 1,2, wehrlites; 3±6, bebedourites; 7,8, syenites; 9,10, metasomatic phlogopitites; 11±16, carbo-
natites; 17±19, phlogopite picrites; 20±24, bebedouritic dykes (b.d.� below detection limit; n.a.� not analysed)
Silicate plutonic rock series Metasomatic phlogopitites Carbonatites (primary mica)
Unit B1 B2 Syenite C1
Analysis 1 2 3 4 5 6 7 8 9 rim 10 core 11 12
SiO2 38.77 39.95 37.66 38.65 36.84 38.71 36.22 35.88 39.53 38.50 43.11 39.76
TiO2 2.78 1.12 1.62 2.50 2.43 1.85 2.83 3.69 0.19 1.38 0.19 0.16
Al2O3 12.80 12.18 12.24 12.63 11.23 10.12 9.86 12.55 1.74 11.62 0.07 0.00
Cr2O3 0.03 0.03 b.d b.d 0.04 0.05 0.01 0.05 0.06 0.00 b.d 0.05
FeO 3.91 4.71 12.29 6.82 13.85 12.55 19.37 19.86 5.91 3.78 5.55 7.05
Fe2O3 2.17 1.57 1.60 1.65 3.26 3.62 4.14 1.10 14.58 2.54 14.47 16.23
MnO 0.10 0.14 0.21 0.12 0.38 0.32 1.27 0.60 0.06 0.12 0.06 0.09
MgO 23.69 24.65 18.69 21.81 16.61 18.00 11.64 11.56 22.82 24.32 23.74 21.60
BaO 0.40 0.45 0.61 0.50 0.71 0.13 0.20 0.06 n.a. n.a. b.d 0.03
CaO 0.05 0.06 0.08 0.06 b.d 0.08 0.03 0.07 0.02 b.d 0.05 0.08
Na2O 0.42 0.27 0.34 0.64 0.30 0.30 0.28 0.22 0.19 0.34 0.08 0.29
K2O 10.14 9.90 9.97 9.67 9.90 10.19 9.93 9.71 9.97 10.07 8.65 9.73
F n.a. 0.29 n.a. n.a. n.a. n.a. 0.20 0.17 n.a. n.a. 0.23 0.30
Cl 0.01 n.a. 0.01 b.d 0.04 0.02 0.02 n.a. n.a. n.a. b.d n.a.
H2O 4.16 4.02 3.99 4.11 3.93 4.00 3.72 3.78 3.94 4.06 3.94 3.75
Total 99.43 99.35 99.31 99.17 99.52 99.93 99.72 99.32 99.01 96.75 100.13 99.12
(IV)Si 5.590 5.761 5.654 5.645 5.610 5.803 5.685 5.572 6.018 5.693 6.378 6.121(IV)Al 2.173 2.069 2.164 2.172 2.013 1.786 1.823 2.296 0.312 2.024 0.011 0.000(IV)Fe31 0.235 0.170 0.181 0.181 0.374 0.409 0.488 0.129 1.668 0.283 1.609 1.878
T site 7.998 8.000 7.999 7.998 7.997 7.998 7.996 7.997 7.998 8.000 7.998 7.999
(VI)Al 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000(VI)Ti 0.301 0.122 0.183 0.275 0.279 0.208 0.334 0.431 0.021 0.154 0.021 0.019(VI)Fe21 0.471 0.568 1.543 0.833 1.764 1.573 2.542 2.579 0.752 0.468 0.687 0.907(VI)Cr 0.004 0.004 0.000 0.000 0.004 0.006 0.001 0.007 0.007 0.000 0.000 0.006(VI)Mn 0.012 0.017 0.027 0.015 0.049 0.041 0.169 0.079 0.008 0.015 0.008 0.011(VI)Mg 5.092 5.299 4.182 4.749 3.771 4.022 2.724 2.676 5.178 5.361 5.237 4.958
O site 5.880 6.010 5.935 5.872 5.867 5.850 5.770 5.772 5.966 5.998 5.953 5.901
Ba 0.023 0.025 0.036 0.029 0.042 0.008 0.012 0.004 0.000 0.000 0.000 0.002
Ca 0.008 0.009 0.013 0.010 0.000 0.012 0.005 0.011 0.004 0.000 0.008 0.013
Na 0.116 0.076 0.098 0.182 0.088 0.087 0.085 0.067 0.056 0.097 0.023 0.086
K 1.866 1.822 1.910 1.802 1.923 1.949 1.988 1.923 1.935 1.900 1.632 1.910
A site 2.013 1.932 2.057 2.023 2.053 2.056 2.090 2.005 1.995 1.997 1.663 2.011
F 0.000 0.132 0.000 0.000 0.000 0.000 0.100 0.084 0.000 0.000 0.107 0.148
Cl 0.003 0.000 0.002 0.000 0.011 0.005 0.006 0.000 0.000 0.000 0.000 0.000
OH 3.997 3.868 3.998 4.000 3.989 3.996 3.895 3.916 4.000 4.000 3.893 3.852
be the most important constraining factor, resulting in a
slight Ti decrease from core to rim in phlogopites from
some samples. In the SPS as a whole, the Ti content of
phlogopite seems to be controlled by the combination of
one or more of the following factors: (a) decreasing pres-
sure; (b) increasing f O2; (c) Ti availability in the liquid.
Oxygen fugacity is expected to increase with differentiation
and could lead to the observed Ti-enrichment in syenite
phlogopites. Furthermore, ®eld and petrographic evidence
indicates the emplacement of SPS at least partially as a crystal
mush, suggesting that different rocks may have been formed at
variable depths. The pace of Ti variation with Fe/(Fe 1 Mg)
decreases considerably between the wehrlites and the
bebedourites. (Fig. 10). This might be related to a change in
the availability of Ti in the liquid (e.g. extensive fractionation
of perovskite and/or Ti-magnetite or the onset of liquid immis-
cibility). Note that the trend for syenite mica is a progression of
the trend for B2 bebedourites.
Mn correlates positively with the Fe21/(Fe21 1 Mg) ratio in
the SPS, increasing steadily through the fractionation
sequence, from less than 0.01 p.f.u. in the wehrlites to 0.08
atoms p.f.u. in the syenites. Potassium increases from
1.75 p.f.u. in wehrlites to 1.98 p.f.u. in syenites. Na decreases
slightly with differentiation but the variability within groups
exceeds the differences between different rock-types. Ba (up to
0.05 p.f.u.) and Ca (up to 0.03 p.f.u.) are generally low and do
not vary systematically. Ba is remarkably low (less than
0.01 p.f.u.) in mica from some syenites, but this feature is
not part of a Ba-decreasing trend, and may indicate a prefer-
ential partition of Ba into the alkali feldspar.
Fig. 11 shows a series of microprobe pro®les across
individual phlogopite grains from various rock types.
The ®rst four columns show the evolution of primary
(magmatic) mica. Wehrlite and bebedourite micas show
similar behaviour, whereby the rims are slightly enriched
in MgO and SiO2, with a corresponding depletion in
TiO2 and Al2O3. This suggests that the substitution
mechanism (VI)Mg 1 2(IV)Si, (VI)Ti 1 2(IV)Al (Robert,
1976) is in place. Note that there is no Fe increase associated
with Al depletion. Syenite micas are not signi®cantly zoned
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 281
Table 7 (continued)
Carbonatites (xenocrystic mica) Phlogopite picrite dykes Bebedouritic dykes
Unit C1 C2
Sample 13 14 15 16 17 18 rim 19 core 20 21 core 22 rim 23 core 24 rim
SiO2 39.13 40.49 38.38 41.44 38.53 39.40 38.18 38.17 39.20 37.60 37.73 39.23
TiO2 3.03 0.67 0.13 0.17 3.48 0.76 2.80 4.77 3.92 5.04 4.85 3.08
Al2O3 10.93 5.61 3.17 8.12 10.95 7.20 12.05 12.61 9.26 4.64 12.64 10.57
Cr2O3 0.00 0.01 b.d b.d 0.05 0.07 b.d 0.46 0.03 b.d 0.24 b.d
FeO 10.23 9.88 14.24 2.09 5.98 9.35 5.25 5.88 5.10 18.10 6.61 7.08
Fe2O3 2.47 7.32 11.34 5.74 4.50 7.69 3.20 2.77 6.40 9.50 2.81 4.65
MnO 0.53 0.49 0.27 0.08 0.16 0.36 0.13 0.06 0.20 1.02 0.00 0.20
MgO 18.56 19.63 17.02 26.48 21.50 20.80 22.62 20.33 21.08 8.78 19.95 21.12
BaO b.d 0.01 0.30 0.19 0.79 0.27 0.81 0.21 1.02 0.38 0.27 0.72
CaO 0.04 0.12 0.20 0.16 0.12 0.31 0.11 0.05 0.16 0.37 0.09 0.09
Na2O 0.16 0.09 0.06 0.03 0.34 0.34 0.36 0.49 0.57 0.99 0.42 0.61
K2O 10.00 9.98 9.38 10.36 9.39 9.81 9.67 9.72 9.64 9.33 9.57 9.69
F 0.28 0.41 0.17 0.06 0.34 n.a. n.a. 0.42 0.64 0.28 0.27 0.49
Cl 0.01 0.00 0.02 0.02 n.a. b.d b.d n.a. n.a. n.a. n.a. n.a.
H2O 3.90 3.75 3.71 4.11 3.95 4.02 4.11 3.93 3.81 3.64 3.98 3.91
Total 99.27 98.47 98.39 99.04 100.07 100.37 99.31 99.87 101.03 99.67 99.43 101.45
(IV)Si 5.809 6.155 6.064 5.992 5.624 5.873 5.574 5.541 5.709 5.983 5.513 5.686(IV)Al 1.911 1.005 0.589 1.383 1.882 1.264 2.072 2.155 1.588 0.869 2.175 1.805(IV)Fe31 0.276 0.837 1.348 0.624 0.494 0.861 0.352 0.303 0.701 1.137 0.309 0.507
T site 7.996 7.997 8.001 7.999 8.000 7.998 7.998 7.999 7.998 7.989 7.997 7.998(VI)Al 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000(VI)Ti 0.338 0.077 0.015 0.019 0.382 0.086 0.308 0.521 0.429 0.603 0.533 0.336(VI)Fe21 1.270 1.256 1.881 0.253 0.730 1.165 0.641 0.714 0.621 2.409 0.808 0.859(VI)Cr 0.000 0.001 0.000 0.000 0.005 0.008 0.000 0.052 0.003 0.000 0.027 0.000(VI)Mn 0.067 0.063 0.036 0.009 0.019 0.045 0.015 0.008 0.025 0.137 0.000 0.024(VI)Mg 4.107 4.449 4.010 5.708 4.677 4.622 4.924 4.398 4.576 2.084 4.346 4.564
O site 5.782 5.846 5.942 5.989 5.813 5.926 5.888 5.693 5.654 5.233 5.714 5.783
Ba 0.000 0.001 0.019 0.010 0.045 0.016 0.046 0.012 0.058 0.023 0.015 0.041
Ca 0.007 0.020 0.034 0.024 0.018 0.050 0.018 0.007 0.025 0.063 0.015 0.015
Na 0.045 0.028 0.017 0.009 0.097 0.097 0.102 0.137 0.162 0.304 0.120 0.172
K 1.894 1.935 1.891 1.911 1.749 1.865 1.801 1.799 1.791 1.895 1.785 1.792
A site 1.946 1.984 1.961 1.954 1.909 2.028 1.967 1.955 2.036 2.285 1.935 2.020
F 0.133 0.195 0.085 0.030 0.158 0.000 0.000 0.194 0.297 0.140 0.125 0.223
Cl 0.003 0.001 0.007 0.004 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
OH 3.865 3.805 3.909 3.967 3.843 4.000 4.000 3.806 3.703 3.860 3.875 3.777
but have higher FeO and lower MgO than micas from ultra-
ma®c rocks. The zoning of phlogopite from carbonatite and
metasomatic mica from the reaction rock (Fig. 11) will be
discussed in the following sections.
3.3.3. Metasomatic phlogopites
At the contact between the carbonatite intrusions and
ultrama®c rocks, the interaction of carbonatitic liquid with
dunite, wehrlite or bebedourite resulted in a phlogopitite
formed by alternate bands of carbonate and phlogopite 1magnetite. The phlogopitite grades into the original ultra-
ma®c rock away from the contact. These features are similar
to those of the Jacupiranga metasomatic phlogopitites and
are common in other APIP carbonatite complexes (Issa
Filho et al., 1984; ArauÂjo, 1996; ArauÂjo and Gaspar, 1993).
The effects of carbonatite metasomatism on phlogopite
are recorded in the form of replacement or overgrowth rims
of reversely-pleochroic tetra-ferriphlogopite (Fig. 12),
whilst the original composition of the magmatic phlogopite
is often preserved in cores with normal pleochroism. The
chemical differences between core and rim are sharp and
coincide with the inversion of pleochroism. A chemical
pro®le of mica from the reaction rock is presented in the
right-hand-side column of Fig. 11.
The zoning pattern of metasomatic micas is easily
distinguished from the SPS magmatic ones. Of the oxides
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296282
Phlogopite
Al
Mg Fe(T)
Eastonite Siderophyllite
Annite
TFP Tetra-ferri-annite
Tapira micas
Biotite
SyenitesB2 bebedourites
B1 bebedouritesWehrlites
CarbonatitesMetasomatic Dykesphlogopitites
Fig. 9. Compositional variation of mica in all Tapira rock-types. Solid arrow indicates the compositional shift from the cores (wehrlite-like) to rims in the
reaction rock. Note the progression from phlogopite towards biotite with differentiation in the SPS, and the overlap of the carbonatite ®eld with the ®elds of
ultrama®c rocks, but not with the syenites (see text). TFP� tetra-ferriphlogopite.
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80.0
0.1
0.2
0.3
0.4
0.5 CA
B
Fe2+/(Fe2++Mg)
Syenites
B2 - bebedourites
B1 - bebedourites
B1 - wehrlites
Ti(
p.f.
u.)
Tapira micas
Fig. 10. Ti variation with the Fe21/(Fe21 1 Mg) ratio. Note the different slopes for the trends of (A) wehrlites; (B,C) two types of bebedourites; and (D)
syenites, suggesting that Ti substitution is more strongly correlated with Fe/Mg in more differentiated types.
J.A.
Bro
det
al.
/Jo
urn
al
of
Asia
nE
arth
Scien
ces19
(2001)
265
±296
283
Fig. 11. Microprobe pro®les across individual phlogopite grains from various Tapira rock types.
represented in the pro®le, only SiO2 shows a gradation from
core to rim. In contrast, MgO, Al2O3 and TiO2 decrease and
total FeO increases abruptly. The magnitude of the chemical
changes is remarkable, especially if compared with the
smoother and much less pronounced zoning of magmatic
micas. Another outstanding feature is the similar magnitude
of the coupled FeO(T) increase and Al2O3 depletion. This
strongly suggests Fe31 substitution for Al in the tetrahedral
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296284
Fig. 13. Euhedral zoned phlogopite, in equilibrium with carbonatite. Note the strong pleochroism from yellow-orange to deep red-brown. Plane-polarised
light. Field of view� 2.15 £ 1.3 mm2.
Fig. 12. Replacement of phlogopite by tetra-ferriphlogopite along the margins of a carbonate veinlet. Note the deep red colour and the inversion in the
direction of highest absorption in the tetra-ferriphlogopite. Plane-polarised light. Field of view� 1.3 £ 0.77 mm2.
site, a feature clearly lacking in the magmatic micas.
Finally, MgO is low in the rims of metasomatic mica,
whilst in micas from wehrlites and bebedourites it tends
to increase towards the rim. BaO is always below detec-
tion limit.
The chemical and petrographic evidence indicates that
cores and rims of these micas crystallised under dramatically
different chemical and/or physicochemical conditions.
Moreover, the cores are compositionally similar to the
magmatic phlogopite in the adjacent ultrama®c rocks
(compare, for instance, with wehrlite and syenite in Fig.
11). This is consistent with a metasomatic origin for the
rims, rather than magmatic evolution.
3.3.4. Carbonatites
Three textural types of phlogopite were recognised in
Tapira carbonatites. These are summarised below:
Type 1 Ð Large crystals of relatively Al-rich (up to
12.63% Al2O3) phlogopite in C1, C3 and C4, often
showing evidence of: (a) resorption by the carbonatite
magma; (b) replacement by rims and irregular patches
of reversely-pleochroic tetra-ferriphlogopite; (c) tetra-
ferriphlogopite overgrowth. The textural evidence
suggests that the phlogopite cores are xenocrystic, and
the tetra-ferriphlogopite is a product of reaction with the
carbonatite liquid. Texturally and compositionally, these
are the analogues of the metasomatic micas from the
phlogopitites.
Type 2 Ð Large, euhedral tetra-ferriphlogopite crystals
(Fig. 13), showing concentric (often oscillatory) zoning and
no evidence of disequilibrium. Inclusions of primary
carbonatite minerals, such as pyrochlore, apatite and carbo-
nate are common. A typical chemical pro®le is presented in
Fig. 11. The zoning involves variation of SiO2, FeO and
MgO and results in different shades of red in the overall
reversely-pleochroic grains. A possible substitution scheme
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 285
.1
.1
1
1
10
10
.1
.1
1
1
10
10
100
100
1000
1000
10000
10000
FeO(t)/MgO
FeO(T)/MgO
Carbonatites
Pri
mar
ym
ica
inca
rbon
atit
es
Pri
mar
ym
ica
inca
rbon
atit
es
Wehrlites to syenites
Primary mica in
silicate rocks
Primary mica in
silicate rocks
xenocrysts
in
carbonatite
xenocrysts
in
carbonatite
P
P
to Ann
to Ann
to TFP
to TFP
Tapira micas
Tapira micas
FeO
(T)/
AlO 2
3F
eO(T
)/A
lO 23
A
B
Fig. 14. Iron variation with aluminium and magnesium in Tapira phlogopites. (A) carbonatite micas. The analyses within a solid line are xenocryst cores (see
text). The evolution of carbonatite micas towards tetra-ferriphlogopite is marked by the sharp increase in the FeO(t)/Al2O3 ratio. (B) Phlogopites from silicate
plutonic rocks clearly plot along the phlogopite±annite trend and do not show the same FeO(t)/Al2O3 enrichment as those from carbonatites. P� phlogopite;
Ann� annite; TFP� tetra-ferriphlogopite. Note that the log scale is necessary to picture the wide range of Fe/Al variation.
could be
Si41 1 Mg21 , 2Fe31
Aluminium and titanium are extremely low and do not
vary. An interesting feature is the antipathetic variation of
SiO2 and FeO, suggesting mutual substitution. This variety
of phlogopite is interpreted as magmatic, crystallised
directly from the carbonatite liquid during a quiescent
period, when minor chemical or physicochemical changes
in the system became imprinted in the mica (e.g. Heathcote
and McCormick, 1989).
Type 3 Ð Minute, interstitial mica ¯akes. In C1 and C4
these are mainly tetra-ferriphlogopite, interpreted as crystal-
lised directly from the carbonatite (by compositional analogy
with Type 2). In C2, some ¯akes are tetra-ferriphlogopite and
some are intermediate members of the phlogopite±tetra-
ferriphlogopite series. The extremely ®ne grain-size and
paucity of mica ¯akes in C2 effectively preclude a decision
as to their magmatic or xenocrystic origin. Nevertheless, the
simultaneous occurrence of both types as scattered ¯akes in
the same sample suggests a mechanical mixture of ®ne-
grained phlogopite of distinct origins.
In summary, the only Tapira micas with clear textural
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296286
(IV)Al (IV) 3+Fe
Phlogopite TFP
Si
70 70
8080
Tapira carbonatitemicas
xenocryst core
xenocryst rim
C2 flakesmagmatic
metasomatic phlogopitite
end-members
Fig. 15. Compositional variation of carbonatite phlogopites, in terms of the three main cations in the tetrahedral site. Fields for cores and rims of micas in the
reaction rock are plotted for comparison (arrow indicates core to rim). TFP� tetra-ferriphlogopite.
(IV)Al( p.f.u.)
Phlogopite
TFP
Tapira carbonatitemicas
xenocryst core
xenocryst rim
C2 flakesmagmatic
metasomatic phlogopitite
end-members
2
1
0
0 1 2 3
(IV
)3+
Fe
(p.f
.u.)
..
.
Fig. 16. (IV)Al and (IV)Fe31 variation in micas from carbonatites. Note the negative correlation between Al and Fe de®ned by the rims of xenocrysts and in the
reaction rock, as well as some C2 mica ¯akes. TFP� tetra-ferriphlogopite.
evidence for magmatic crystallisation from a carbonatite
magma are the euhedral, zoned crystals of tetra-ferriphlo-
gopite in C1 (Type 2, Fig. 13).
Barium is low (up to 0.22% BaO) in all three types of
carbonatite micas and do not allow discrimination between
them. High-Ba micas, such as those from the Jacupiranga
carbonatites, were not found in Tapira.
The composition of carbonatite micas is compared with
those of the SPS in Fig. 14. Micas from carbonatites
and silicate rocks clearly follow different crystallisation
paths. A sudden increase in FeO(t)/Al2O3 ratio (at FeO(t)/
MgO < 0.75), signals the crystallisation of tetra-ferriphlo-
gopite in the carbonatites, whilst the plutonic silicate rocks
follow the pattern of the phlogopite±annite series. The high-
Al cores of xenocrystic micas from C1, C3 and C4 follow
the phlogopite±annite trend and have composition similar to
the micas of wehrlites and bebedourites in the lower range
of FeO(t)/MgO.
The tetra-ferriphlogopite and tetra-ferri-annite end-
members are signi®cant components of micas from Tapira
carbonatites. Consequently, tetrahedral site de®ciency and
substitutions must play a much more relevant role here
than in the SPS. The chemical behaviour of the major
tetrahedral cations in these micas is summarised in Figs.
15 and 16. The analyses cover the whole range of compo-
sitions in the phlogopite±tetra-ferriphlogopite series, but
cluster preferentially near either one of the ideal end-
members (Fig. 15).
Micas for which textural evidence of a magmatic origin
is unequivocal invariably plot within a very short distance
of the tetra-ferriphlogopite end-member. Many micas in
this group either have very low (IV)Al or lack aluminium
completely, in which case the tetrahedral site is ®lled
exclusively with (IV)Fe31 and Si. A strong negative correla-
tion between (IV)Fe31 and Si (not shown), and the antipa-
thetic behaviour of these elements in Fig. 11 both indicate
reciprocal substitution. However, the oscillatory nature of
the zoning in these micas suggests that this results from
subtle physical changes rather than a magmatic evolution-
ary trend. In turn, the high-Al cores of mica xenocrysts in
C1, C3 and C4 contain little (IV)Fe31 (less than 0.6 p.f.u.)
and most of the observed variation is between Si and (IV)Al
(Fig. 15). The analyses of this group cluster near the
phlogopite end-member, like the micas from wehrlites
and bebedourites.
The reversely-pleochroic rims and patches replacing
phlogopite xenocrysts show a distinctive behaviour with
respect to the tetrahedral cations. Only in this variety and
in some interstitial ¯akes from C2, a strong negative
correlation between (IV)Fe31 and (IV)Al is present (Fig.
16). In fact, this group of analyses ªbridges the gapº
between the high-Al cores of xenocrysts and the high-
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 287
FeO(T)/MgO
FeO
(T)/
AlO 2
3
.1.1
1
1 10
10
100
1000
1000 carbonatite trend
silicate plutonicrock trend
phlogopite picritesbebedouritic dykes
Tapira dyke-rockmicas
Fig. 17. Iron variation relatively to magnesium and aluminium in micas from phlogopite-picrites and bebedourite dykes. Fields for SPS and carbonatite micas
(including all groups) are shown for comparison. Most analyses follow the trend of the SPS, but some phlogopite rims deviate towards the trend of
carbonatites. Filled symbols� cores; open symbols� rims.
Fe31 primary carbonatitic mica, which otherwise seems
to have evolved independently of the (IV)Fe31, (IV)Al
substitution.
3.3.5. Ultrama®c dykes
Phlogopite is an essential constituent of phlogopite-picrites
and bebedourite dykes. Groundmass mica is present in all
dykes and some samples contain phlogopite phenocrysts and
microphenocrysts. In general, the phlogopite-picrites
contain less evolved (more magnesian) phlogopite than
the bebedourite dykes, but there is a signi®cant overlap.
Fig. 17 compares phlogopites from the dyke rocks with
the carbonatites and SPS. The dykes cover most of the span
of SPS micas and ®t well along that trend, although some
rim compositions tend to approximate the ®eld of carbona-
tite micas. A possible reason for this behaviour is reaction
between early-formed phlogopites and carbonate-rich
residual liquids, producing low-Al, high-Fe rims. This
is consistent with the abundant groundmass carbonate in
some of the dykes.
Phlogopite from the two types of dykes have the highest
Ba content amongst Tapira micas (up to 2.95% BaO in
phlogopite-picrites and up to 1.71% BaO in bebedourite
dykes), with Ba usually decreasing from core to rim.
Cr2O3 is generally low, but may exceptionally reach 1.2%.
Some micas in the bebedourite dykes seem to be enriched in
chromium (up to 1% Cr2O3). In any case, Cr2O3 decreases
from core to rim.
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296288
Al O (wt% )2 3
TiO
(wt%
)2
4 6 8 10 12
P
G
FeO/(FeO+MgO)
0.2 0.4 0.6 0.80
1
2
3
4
5
6
GP
Singlecrystal
Singlecrystal
phenocryst phenocryst
groundmass groundmass
Tapira dyke-rock micas
Fig. 18. Chemical zoning of phlogopite in the phlogopite-picrites (solid
lines) and bebedourite dykes (dashed lines). In the phlogopite-picrites,
the lines represent chemical pro®les across individual grains, from core
(solid circles) to rim (open circles). In the bebedouritic dyke the lines
connect pairs of core-rim analyses of phenocryst (P) and groundmass (G)
phlogopite.
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90
1
2
3
4
5
6
7
FeO/(Feo+MgO)
xenocrysts
magmatic
Carbonatitephlogopites
Wehrlites ( )Phlogopite-picrites ( )Low-Cr dykes ( )Bebedourites B1 ( )?
Bebedourites B2 ( )Syenites ( )Some low-Cr dykes( )
TiO
2
Fig. 19. Variation of TiO2 with FeO/(FeO 1 MgO) for micas of the SPS, phlogopite-picrites, and bebedourite dykes. Selected analyses of phlogopite rims in
both types of dyke are plotted, and linked to the respective cores by the dashed arrows. Compositional ®elds for xenocrystic and magmatic phlogopites from
carbonatite are also shown.
The variation of TiO2, Al2O3 and FeO/(FeO 1 MgO) in
zoned phlogopites from different dykes is illustrated in Fig.
18. In the phlogopite-picrites (solid lines in Fig. 18) titanium
increases initially with the FeO/(FeO 1 MgO) ratio, before
decreasing towards the crystal rims. The early TiO2 enrich-
ment may or may not be coupled with Al2O3, increase, and
the magnitude of Ti-depletion in the rims appears to be
attenuated with increasing TiO2.
Micas from the bebedouritic dykes (dashed lines in Fig.
18) have higher TiO2 than those from phlogopite-picrite
dykes. This is analogous to the evolution of the SPS, where
Ti increases in micas of the more differentiated rock types.
Phenocrysts and groundmass phlogopite show contrasting
behaviour, whereby TiO2 decreases in the former and
increases in the latter, from core to rim. The trend of the
phenocrysts in the bebedourite dykes is similar to the broad
core±rim variation in the phlogopite-picrites. In the bebe-
dourite dykes, the similarity between the phenocryst rims
and the groundmass phlogopite cores suggests a continued
crystallisation trend, where TiO2 decreases at ®rst, perhaps
during crystallisation of perovskite or Ti-magnetite, and
increases again towards the later stages of fractionation. In
all three examples, the phlogopite rims are always depleted in
Al2O3 relatively to the cores, which probably re¯ects
decreasing aluminium availability in the residual liquid.
It was seen in a previous section that micas from different
rock types in the SPS show different Ti behaviour with FeO/
(FeO 1 MgO). Fig. 19 compares those trends with the
composition of micas in dykes and carbonatites. The dashed
arrows in the diagram represent the core±rim variation of
phlogopites in the dykes, except for the two highest TiO2
pairs, where the arrows connect phenocrysts to groundmass
phlogopite. The evolution trends of cores and rims (or
groundmass) parallel, respectively, the trends of
wehrlites 1 bebedourites (B1) and bebedourites (B2) 1syenite. At least in the case of the bebedourite dykes, this
may suggest that micas started crystallising at a greater
depth, whereas the rims were formed at higher levels, during
or after emplacement of the dykes. In the latter case, the
lower pressure would favour Ti-solubility in the mica (e.g.
Arima and Edgar, 1981). This shallower crystallisation site
would, presumably, coincide with the level of emplacement
of B2 and the syenites, whose micas have FeO(T)/
(FeO(T) 1 MgO) and TiO2 similar to the rims of phlogopite
in the dykes. Also noteworthy in Fig. 19 is the compara-
tively high TiO2 content of the cores of mica xenocrysts in
carbonatites (dotted-limited ®eld). They cover a large span
of SPS mica compositions (except, maybe, the syenites),
suggesting that the carbonatite magmas assimilated mica
crystals from different ultrama®c rocks. The micas crystal-
lised from the carbonatite magma, on the other hand, show
extremely low TiO2 contents.
An alternative reason for the behaviour of Ti and Al in
magmatic micas from Tapira may be changes in the activity
of these elements in the magma. If the whole composition span
of Tapira rocks is considered, Al2O3 and TiO2 show substantial
variation. The rapid variations observed in some phlogopites
from dykes may indicate abrupt chemical changes in the
system, such as the onset of liquid immiscibility.
4. Oxygen fugacity
The Fe21±Fe31 relationship in micas can be used to gain
some insight on the oxygen fugacity conditions under which it
crystallised. The discussion in this section will focus on Tapira
micas only. As demonstrated earlier, phlogopites from Jacu-
piranga do not have signi®cant Fe31 content. Micas from Cata-
laÄo, on the other hand, show remarkable compositional
similarities with those from Tapira, and it is likely that the
same oxygen fugacity constraints apply to both complexes.
Nevertheless, CatalaÄo micas are excluded from this discussion
because of the dif®culty in distinguishing the magmatic and
metasomatic sources of the composition trends.
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 289
BebedouritesB1Fe2+ Wehrlites
HM
NNO
FMQ
BebedouritesB2
Syenites
CarbonatitesDykes
WM- WI
Mg
Fe3+
Fig. 20. Composition of primary Tapira micas plotted in the diagram of Wones and Eugster (1965). Oxidation state of the system indicated by the oxygen
buffers HM (hematite-magnetite), NNO (nickel±nickel oxide), FMQ (fayalite±magnetite±quartz), WM (wustite±magnetite), and WI (wustite±iron). Solid
circles indicate phlogopite ¯akes from carbonatite C2.
Fig. 20 provides an estimate of the oxidation state of the
liquids from which the Tapira phlogopites crystallised.
Micas from the SPS plot in a trend sub-parallel to the
nickel±nickel oxide buffer line, indicating that little varia-
tion in oxygen fugacity was involved in the differentiation
of the plutonic silicate rocks. Primary tetra-ferriphlogopite
from carbonatites, on the other hand, plots well above the
line of the hematite±magnetite buffer, indicating the highly
oxidised character of the Tapira carbonatite magmas. Micas
from phlogopite-picrites and bebedourite dykes plot along a
number of trends of increasing Fe31 at variable Fe21/Mg,
suggesting that variations in the oxygen fugacity may have
occurred during the crystallisation of the dykes. Most phlo-
gopites from dykes plot between the lines of hematite±
magnetite and nickel±nickel oxide buffers.
The phlogopite xenocrysts in the Tapira carbonatites (not
plotted in the diagram) span most of the SPS range, suggest-
ing that: (1) the original mica crystallised in less oxidising
conditions than those prevailing in the carbonatite magma
and/or (2) the carbonatites assimilated phlogopite from
distinct rock-types. Phlogopite ¯akes from the C2 carbona-
tite de®ne a trend of increasing Fe31 at high Mg. This trend,
also present in metasomatic micas from Tapira and CatalaÄo
(not shown in Fig. 20), is similar to the one proposed by
Brigatti et al. (1996, their Fig. 8) for the evolution of Tapira
rocks. As stated earlier in this work, Jacupiranga micas do
not contain Fe31.
5. Origin of tetra-ferriphlogopite
Tetra-ferriphlogopite has been interpreted as both a result
of post-magmatic processes (ArauÂjo, 1996; Zaitsev and
Polezhaeva, 1994; McCormick and Heathcote, 1987;
Mitchell, 1995a) and primary crystallisation (Heathcote
and McCormick, 1989; Brod, 1999).
In many cases, a secondary origin is supported by petro-
graphic evidence, such as mantling of pre-existing phlogo-
pite crystals by tetra-ferriphlogopite, sharp compositional
changes and various disequilibrium textures. Less
frequently, the opposite pattern is observed. Farmer and
Boetcher (1981) described phlogopite with reversely-pleo-
chroic cores and normally-pleochroic rims in kimberlites
and associated peridotite xenoliths. They point out,
however, that the cores are richer in Al2O3 and poorer in
FeO(T) than in other optically similar phlogopites, suggest-
ing that the reverse pleochroism is not necessarily related to
the Fe31, Al substitution. We demonstrated in this work
that the pleochroism reversal is intimately related to the
Fe31, Al substitution in phlogopites from both the Tapira
and CatalaÄo complexes.
Tetra-ferriphlogopite is a common variety of mica in
other APIP carbonatite complexes. Besides occurring in
Tapira and CatalaÄo (Brigatti et al., 1996; Brod, 1999; ArauÂjo
and Gaspar, 1993; ArauÂjo, 1996, and this work), it has also
been described from the intrusions of Araxa (Cruciani et al.,
1995) and Salitre (Lloyd and Bailey, 1991). It is, however,
apparently absent at Jacupiranga.
Lloyd and Bailey (1991) described tetra-ferriphlogopite
from the Salitre complex, associated with carbonate-silicate
cryptocrystalline material in bebedourites. They interpreted
the mica as a late-stage product, but consider the evidence
inconclusive as to whether it crystallised from the residual
liquid or resulted from introduction of material after crystal-
lisation. Morbidelli et al. (1995) reported the presence of
tetra-ferriphlogopite in carbonatites and fenites from Salitre.
Issa Filho et al. (1984) interpreted the glimmerites of the
Araxa complex as being of metasomatic origin, formed by
phlogopitisation of ultrama®c rocks. Although they do not
provide phlogopite analyses (and the Araxa mica studied by
Cruciani et al., 1995 comes from a sovite), the textural
features of the Araxa glimmerites resemble the metasomatic
phlogopitites from Tapira and CatalaÄo described in this
work.
Brigatti et al. (1996) studied the crystal chemistry of
Tapira phlogopites. They suggested that the formation
of tetra-ferriphlogopite (i.e. (IV)Fe31 substitution accom-
panied by inversion of pleochroism and depletion in Al
and Ti) is a magmatic process. In Tapira, it would
presumably increase with fractional crystallisation, in
the sequence: dunite) wehrlite) clinopyroxenite)bebedourite) garnet±magnetitite) perovskite±magne-
titite ) glimmerite) carbonatite. In fact they propose
the division of the Tapira complex in two magmatic
systems, one (alkaline-silicate system) including the ®rst
four rock types in the above sequence and the other (sili-
cate±carbonatite system) comprising the remaining three.
However, their petrographic descriptions suggest that the
studied rocks are mostly varieties of cumulates from the
ultrama®c magmas of the SPS. Perovskite- magnetite- and
garnet-rich cumulates, for instance, are commonly asso-
ciated with pyroxenites in Tapira and the very low TiO2
content of Tapira carbonatites (Brod, 1999) suggests that
the carbonatite liquid is unlikely to produce Ti-rich cumu-
lates. Furthermore, the specimens described by Brigatti et
al. (1996) do not include syenites or true carbonatites
(maximum CO2 content reported is 14.28%). Finally, they
report the formation of hematite rims associated with tetra-
ferriphlogopite, which could indicate a metasomatic origin.
The results of the present work Ð using a wider range of
petrographic types Ð indicate that magmatic mica in the
Tapira SPS evolves from phlogopite towards biotite, rather
than towards tetra-ferriphlogopite. In addition, Ti contents
of phlogopite increase with fractional crystallisation, albeit
in a scenario of multi-stage magmatic evolution. It was also
demonstrated in previous sections that magmatic crystalli-
sation of tetra-ferriphlogopite is generally restricted to true
(magmatic) carbonatites, the possible exceptions being
groundmass mica in some carbonate-rich phlogopite-
picrites. Whilst agreeing with the evolution proposed by
Brigatti et al. (1996) for the mica in ultrama®c rocks
(increasing Fe and Ti with decreasing Mg) and, indeed,
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296290
extending this model to the most differentiated rocks
(syenites), the results of the present research do not support
their model of magmatic progression from phlogopite to
tetra-ferriphlogopite in the Tapira SPS. Instead, the(IV)Fe31, (IV)Al substitution reported by Brigatti et al.
(1996) as magmatic, closely resembles the evolution of
the xenocrystic mica in Tapira carbonatites and the
metasomatic mica in phlogopitites from both the Tapira
and CatalaÄo complexes.
A summary of mica evolution in the carbonatite
complexes studied in this work is given in Fig. 21. The
main trends recognised are marked with bold arrows. The
evolution trends of metasomatic micas in all complexes
have been omitted for clarity. It is noteworthy that
magmatic micas from carbonatites and silicate rocks evolve
along independent trends in all cases.
Carbonatite micas at Jacupiranga contain much higher Al
than their Tapira and CatalaÄo equivalents. These micas
follow a trend of rapid decrease in Al with concomitant
increase in Mg, at roughly constant Fe contents. This
contrasts with Tapira and CatalaÄo, where magmatic carbo-
natite micas evolve independently of Al, although it is not
clear from textural evidence whether they vary from Fe-rich
towards Mg-rich or otherwise. We believe that this contrast-
ing behaviour is related to the Al content of the primitive
carbonatite magma. If the Jacupiranga carbonatite liquids
had a small amount of Al, crystallising phlogopite would
have acted as a Al scavenger, since it is the only Al-bearing
mineral recorded from these rocks.
At Tapira and CatalaÄo, on the other hand, it seems that the
carbonatite magma had been Al-poor since its origin. Brod
(1999) concluded that Tapira carbonatites formed by liquid
immiscibility, a process capable of depleting the carbonatite
liquid (and enriching the silicate counterpart) in Al, among
other elements. This explains why the only micas in textural
equilibrium with carbonatite in these complexes are
virtually Al-free.
In all studied cases, silicate-rock phlogopites evolve
through Fe/(Fe 1 Mg) increase. At Jacupiranga, the trend
for under-saturated rocks follows a curved line, with Al
decreasing initially and increasing again in the late stages.
Although progressing in the same general direction, the
saturate silicate rocks do not conform exactly to this
trend, appearing to evolve at higher Al contents. At Tapira
and CatalaÄo, the mica evolution trend runs along the
phlogopite-annite line, but several offshoots occur, towards
both Al-rich and Al-poor compositions. Brod (1999)
demonstrated that liquid immiscibility at Tapira was a
recurrent, multi-stage process, occurring at various differ-
entiation stages of an evolving carbonate-rich silicate
magma. This is consistent with the many phlogopite trends
offshooting from the phlogopite±annite line in Fig. 21.
6. Comparison with mica from other carbonatites andalkaline rocks
Mica composition and evolution in alkaline rocks is
strongly dependent on the geochemical af®nity of the
magma. Mitchell (1995a) summarised the following points
regarding compositional variation of phlogopite in potassic
alkaline rocks:
² Micas rich in both Ti and Ba are characteristic of leuci-
tites and melilitites.
² Micas from Group I kimberlites always have low TiO2
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 291
Al
Mg Fe(T)
Eastonite Siderophyllite
Phlogopite Annite
TFP Tetra-ferri-annite
Jacupiranga,Tapira,and CatalãoPhlogopites
Tapira andCatalão
Jacupiranga
Fig. 21. Comparison of phlogopite evolution trends from Jacupiranga, Tapira and CatalaÄo, in terms of Al, Mg and total Fe. The compositional ®elds of Tapira
and CatalaÄo micas largely overlap and are represented as a single ®eld. Bold arrows depict major evolution trends in carbonatites and silicate rocks. Light
arrows show offshooting trends from the phlogopite-annite series, observed in Tapira micas.
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296292
0 2 4 6 8 10 12
2
4
6
8
10
12
14
16
18
20
Minettes/RP
Field of kimberlite micas
Kimberlitegroundmass
Orangeites Kamafugites
Lamproites
MARID
APIPcarbonatitecomplexes
0 2 4 6 8 10 12
2
4
6
8
10
12
14
16
18
20
Paraguay Juquiá
Jacupiranga
Kamafugites
A
B Silicate-rockmicas (carbonatitecomplexes)
Carbonatitemicas
Al
O2
3A
lO
23
TiO2
TiO2
Fig. 22. (A) Al2O3 versus TiO2 for the studied micas, compared with the carbonatite complex of JuquiaÂ, alkaline rocks and carbonatites from Paraguay and
kamafugites. (B) Comparisoin of micas from carbonatite complexes (including carbonatites and silicate rocks), compared with other types of alkaline rocks.
Fields for MARID, minette/RP (Roman Province) and kimberlite micas are from Mitchell and Bergman (1991). Trends for kimberlites, lamproites, minettes
and orangeites after Mitchell (1995b). Field of kamafugite includes the Brazilian occurrences of Mata da Corda (Sgarbi and ValencËa, 1994; Leonardos et al.,
1991; Mitchell and Bergman, 1991) and AmorinoÂpolis (Danni and Gaspar, 1992), as well as kamafugites from Africa (Edgar, 1979), Italy (Conticelli and
Peccerillo, 1992) and Arizona (Laughlin et al., 1989). APIP carbonatite complexes comprise include data for Salitre (Lloyd and Bailey, 1991) and CatalaÄo
(ArauÂjo, 1996). Fields of alkaline rocks from Paraguay (Comin-Chiaramonti et al., 1992, 1996) and for the Brazilian carbonatite complexes of JuquiaÂ
(Beccaluva et al., 1992) and Jacupiranga (Gaspar and Wyllie, 1987; Gaspar, 1989) are also represented.
(,4%) and can evolve along two different trends, one of
increasing Al2O3 and BaO, and the other of decreasing
Al2O3 (towards low-Ti tetra-ferriphlogopite). The latter is
commonly interpreted as secondary in origin.
² Group II South African kimberlites (orangeites) also
show two separate trends, one involving Fe, Al substi-
tution (phlogopite±tetra-ferriphlogopite) and the other
between the phlogopite±annite end-members.
² Phlogopites from lamproites typically have high Ti, and
evolve towards Ti and Fe enrichment with (Ti-tetra-
ferriphlogopite) or without (Ti-biotite) Al depletion.
Thompson et al. (1997) described a high-TiO2 (8 wt%)
phlogopite with Al2O3 as low as 2 wt% in a lamproite
from Middle Park, Colorado.
² Micas from minettes and lamprophyres show a more vari-
able behaviour, but in minettes, they are usually Al-rich.
Phlogopite evolution in carbonatites has been described
as increasing Mg and Si and decreasing Al, Ti and Fe,
coupled with incorporation of Fe31 in tetrahedral sites
during the later stages (Heathcote and McCormick, 1989).
On the other hand, a similar trend seems to derive from
metasomatic processes. McCormick and Heathcote (1987)
interpreted tetra-ferriphlogopite in carbonatites from
Arkansas as formed during dolomitisation of previous
sovite, with ingress of Si, Mg, Fe31 and removal of Al,
Fe21 and Ti from the system.
Fig. 22 compares phlogopites from the studied localities
with other examples of carbonatites and alkaline rocks, in
terms of TiO2 and Al2O3 compositions. As a general rule,
primary phlogopite from Tapira (Silicate Plutonic Series,
cores of xenocrysts in carbonatites and cores of mica
crystals from dykes) and other APIP complexes (CatalaÄo,
Salitre) have Al2O3 ranging from 9 to 13% and limited TiO2
(up to < 5%). This roughly coincides with the ®eld for
MARID micas and with the low-TiO2 half of the ®eld for
kamafugite phlogopites. It is distinguished from minettes
(Roman Province), alkaline rocks from Paraguay and the
Brazilian carbonatites complexes of Jacupiranga and JuquiaÂ
by the lower Al2O3 and more restricted TiO2. Primary carbo-
natite micas from Tapira and CatalaÄo cluster near the origin
of the diagram, and cannot be distinguished in the adopted
scale.
In carbonatite complexes, a coupled depletion of TiO2 and
Al2O3, leading towards tetra-ferriphlogopite seems to be
typical of metasomatic and late stage micas. This variation
is similar to that observed in orangeites, and distinguished
from lamproitic tetra-ferriphlogopites, which are Ti-rich.
Fig. 22 also provides an insight into the relative differ-
ences between micas from the APIP and those from other
alkaline provinces emplaced around the margins of the
Parana Basin. The ®eld for the Jacupiranga complex
includes analyses of phlogopites from silicate rocks and
carbonatites, with the latter being restricted to the very-
low-TiO2/high-Al2O3 end of the diagram. The composi-
tional range occupied by the carbonatite phlogopites is
marked by a sudden in¯ection of the Jacupiranga ®eld
towards higher Al2O3. Indeed, Gaspar and Wyllie (1987)
stated that micas from the Jacupiranga carbonatites usually
show (VI)Al as high as, or higher than, Fe21. A similar
pattern is shown by the alkaline rocks from Paraguay.
Although the ®eld in Fig. 22 does not include carbonatite
micas, an in¯ection towards high Al2O3 still appears at low
TiO2. Gomes et al. (1996b) draw attention to the fact that
eastonite±siderophyllite is an important component of
micas in the Paraguay alkaline rocks.
The ®eld for the Juquia complex comprises only
phlogopites from silicate rocks, and shows Al2O3 compar-
able to Jacupiranga and Paraguay, for the more restricted
TiO2 range covered. Therefore, the ®elds of mica from Jacu-
piranga, Juquia and Paraguay are roughly coincident, and
characterised by high-Al2O3 and variable (very low to very
high) TiO2. The lowest TiO2 portions of the Jacupiranga and
Paraguay ®elds are marked by a sudden increase in Al2O3.
The TiO2/Al2O3 ratio is similar to that of minettes and
Roman Province. On the other hand, carbonatite complexes
of the APIP are characterised by lower Al2O3 (similar to that
of kamafugite and MARID micas), and a more restricted
TiO2 range. Additionally, carbonatite micas in the APIP
complexes are depleted in Al2O3, causing the ®elds to in¯ect
towards tetra-ferriphlogopite (thus in the opposite direction
to Jacupiranga 1 Juquia 1 Paraguay).
Brod et al. (2000) established the geochemical af®nity of
the primitive magmas (phlogopite-picrites) that originated
the Tapira and other APIP carbonatite complexes with
kamafugites. The high-TiO2 portion of the kamafugite
®eld in Fig. 22 consists entirely of micas from APIP kama-
fugites (Mata da Corda Group, not individualised). These
micas may represent the high-Ti equivalent of Tapira/
Salitre/CatalaÄo, de®ning a wider APIP ®eld, still at lower
Al2O3 than Paraguay/Jacupiranga/JuquiaÂ.
Finally, it should be noted that the alkaline magmatism in
the APIP is Upper-Cretaceous in age, and interpreted as the
result of the impact of the Trindade mantle plume in the
lithosphere under southeast Brazil (Gibson et al., 1995).
Jacupiranga and JuquiaÂ, in turn, are Early Cretaceous and
therefore contemporaneous with the Parana ¯ood-basalt
magmatism. The Jacupiranga complex has been associated
with the Tristan da Cunha mantle plume (Huang et al.,
1995). Therefore, mica compositions may be re¯ecting
two different tectono-magmatic situations. It must be
stressed that the alkaline rocks from Paraguay cover a
wider range of ages (240±39 Ma Ð Gomes et al., 1996a),
and the ®eld plotted in Fig. 22 is not age-selective.
7. Implications for mineral chemistry systematic ofalkaline rocks
It was shown in this work that micas from carbonatites
and their associated silicate rocks vary widely in composi-
tion. They may show various chemical characteristics tradi-
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 293
tionally attributed to phlogopites from minettes, orangeites
(Group II South African kimberlites), kamafugites,
lamproites, and mantle xenoliths, depending both on the
origin and on the magmatic or metasomatic processes acting
upon a particular carbonatite±silicate association. In parti-
cular, the onset of liquid immiscibility has the ability to
override or signi®cantly de¯ect the original evolution trends
of phlogopite. Liquid immiscibility can produce dramatic
changes in the amount of Si, Mg, Ti and Al available for
mica crystallisation, besides strongly in¯uencing the Fe21/
Fe31 ratio of magmatic systems.
This has important consequences for the use of phlogo-
pite mineral chemistry as a means to discriminate alkaline
rock-types. Whilst it could be argued that the carbonatite±
silicate association is easily recognisable in plutonic
complexes, this is not always the case in volcanic
sequences and dyke swarms. In the APIP, for instance,
the phlogopite picrites and their evolved equivalents are
typically associated with carbonatite plutonic complexes,
but may also occur as scattered dykes, with no obvious
geographic relationship to carbonatite. Special care should,
therefore, be taken when using the Ti and Al contents of
phlogopite as geochemical discriminants.
8. Summary
The results of this study can be summarised as follows:
² Primary phlogopites in carbonatites and silicate rocks
seem to evolve independently and follow divergent
trends.
² In the silicate plutonic rocks of the Brazilian Alto
ParanaõÂba Igneous Province (APIP) complexes, mica
evolves from high-Mg phlogopite in the least differen-
tiated rocks (dunites, wehrlites) to biotite in the more
evolved syenites. This variation follows a continuous
trend of Fe,Mg substitution. Ti and Mn increase with
differentiation. The Fe,Mg substitution is also the
main differentiation-related feature of phlogopite in the
Jacupiranga silicate rocks. However, in this case, Ti
decreases with magma evolution.
² Micas from Jacupiranga and from the APIP carbonatites
show a remarkable contrasting behaviour. At Jacupir-
anga, the micas evolve through Al decrease and Mg
increase at relatively constant Fe. In the APIP carbona-
tites, the magmatic phlogopite is virtually Al-free.
² Ti variation indicates a multi-stage evolution for the
Tapira silicate rocks. Decreasing pressure, increasing
oxygen fugacity, and/or sudden changes in Ti avail-
ability in the system (e.g. liquid immiscibility) are
the main factors controlling Ti increase in mica
from the ultrama®c rocks to syenites. Temperature,
in turn, seems to be the more signi®cant factor on
a local scale. In the Jacupiranga silicate rocks, the
behaviour of Ti in the mica is consistent with
progressive temperature decrease at relatively
constant pressure.
² Metasomatic micas are characterised by replacement of
the original phlogopite by rims and patches of tetra-
ferriphlogopite. The inversion of pleochroism coincides
with a sharp decrease in MgO, Al2O3 and TiO2 and
increase in FeO(T), and occurs at approximately 1.5 Al
p.f.u. There is no gradation zoning towards tetra-ferriph-
logopite within a single crystal. The chemical changes
involved are distinct from those associated with
magmatic differentiation.
² A distinct chemical behaviour is displayed by magmatic
tetra-ferriphlogopite in carbonatites, consisting basically
of 2(IV)Fe31, Si41 1 Mg21 substitution. These virtually
Al-free micas form euhedral crystals, often showing
oscillatory zoning.
² Textural and compositional evidence indicates that Al-
rich phlogopite is not likely to have crystallised directly
from the carbonatite magma in Tapira and CatalaÄo
complexes. If that were the case, a wider range in Fe31
and Al would be expected in the cores of phlogopites
from different carbonatite samples. This was not
observed in the studied carbonatites, where single mica
crystals seem to ªjumpº in composition from Al-phlogo-
pite to tetra-ferriphlogopite. The presence of this type of
mica in Tapira carbonatites probably results from assim-
ilation of phlogopite from silicate rocks. Assimilation
could have taken place either at the time of carbonatite
emplacement or during ascent through partially crystal-
lised ultrama®c pockets in the magma chamber.
² Micas in ®ne-grained ultrama®c rocks associated with
carbonatites (phlogopite-picrite and bebedourite dykes)
generally follow the compositional trend of the coarse-
grained plutonic silicate rocks, but in the APIP
complexes various offshoots from this main trend are
observed. Phlogopite-picrites have the more primitive,
and bebedourite dykes have the more evolved mica.
Late-stage sharp chemical changes in phlogopite from
the dykes may signal the onset of major chemical distur-
bances in the system, such as liquid immiscibility.
² Micas from the APIP carbonatite complexes (e.g. Tapira,
Salitre and CatalaÄo) are compositionally similar to each
other, and distinct from micas from the Jacupiranga and
Juquia alkaline-carbonatite complexes.
² Micas in the ultrama®c dykes and plutonic rocks asso-
ciated with APIP carbonatites have TiO2 and Al2O3
contents similar to MARID micas and kamafugites,
although the ®eld for the latter extends to higher TiO2.
Some specimens also show strong TiO2 enrichment
coupled with Al depletion, a feature that is considered
typical of lamproites. The trend of metasomatic micas
from the APIP complexes resemble the evolution of oran-
geites. Micas from carbonatite complexes such as Jacu-
piranga and Juquia are chemically similar to minettes.
² The widely variable chemical composition of phlogo-
pite in carbonatite±silicate associations has important
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296294
implications for the use of mineral chemistry systematics
to discriminate groups of alkaline rocks. Special care
should be taken when using Ti and Al contents of phlo-
gopite from rocks that may have undergone liquid immis-
cibility processes.
Acknowledgements
The authors gratefully acknowledge the Universities of
Brasilia (Brazil), Cambridge and Durham (UK) for granting
access to their analytical facilities. Dr Stephen Reed is
thanked for his help with microprobe analyses at
Cambridge. This research was funded by the Brazilian
Research and Science Development Council (CNPq, grant
no. 200449/94-0 to J.A.B.).
References
ArauÂjo, D.P., 1996. Metassomatismo no complexo carbonatõÂtico CatalaÄo-I:
implicacËoÄes para a composicËaÄo do magma carbonatõÂtico e para o
metassomatismo carbonatõÂtico no manto superior. Unpublished MSc
Thesis, University of Brasilia, Brasilia.
ArauÂjo, D.P., Gaspar, J.C., 1993. Fe31 no sõÂtio tetraeÂdrico de ¯ogopitas das
rochas do complexo carbonatõÂtico de CatalaÄo I, Brasil. Proceedings of
the 4th Congresso Brasileiro de GeoquõÂmica, Extended Abstracts,
SBGq, BrasõÂlia, pp. 62±63.
ArauÂjo, D.P., Gaspar, J.C., Garg, V.K., Souza, P.A. Jr., 1996. DeterminacËaÄo
MoÈssbauer em Cristal Simples de Tetraferri¯ogopita do Complexo
CarbonatõÂtico CatalaÄo-II: ComparacËaÄo com CaÂlculos EstequiomeÂtricos.
Proceedings of the 35th Congresso Brasileiro de Geologia, vol. 3, SBG,
pp. 22±25.
ArauÂjo, D.P., Gaspar, J.C., Garg, V.K., 1998. The complete phlogopite±
tetra-ferriphlogopite series in the CatalaÄo-I and CatalaÄo-II carbonatite
complexes, Brazil. Proceedings of the 7th International Kimberlite
Conference, Extended Abstracts, Cape Town, pp. 29±31.
Arima, M., Edgar, A.D., 1981. Substitution mechanisms and solubility of
titanium in phlogopites from rocks of probable mantle origin.
Contributions to Mineralogy and Petrology 77, 288±295.
Baecker, M.L., 1983. A mineralizacËaÄo de nioÂbio do solo residual laterõÂtico e
a petrogra®a das rochas ultrama®cas alcalinas do domo de CatalaÄo I,
GoiaÂs. Unpublished MSc thesis, University of Brasilia, Brasilia.
Bailey, S.W., 1984. Classi®cation and structures of the micas. In: Bailey,
S.W. (Ed.). Micas. Mineralogical Society of America, pp. 1±12.
Beccaluva, L., Barbieri, M., Born, H., Brotzu, P., Coltorti, M., Conte, A.,
Garbarino, C., Gomes, C.B., Macciotta, G., Morbidelli, L., Ruberti, E.,
Siena, F., Traversa, G., 1992. Fractional crystallization and liquid
immiscibility processses in the alkaline-carbonatite complex of JuquiaÂ
(SaÄo Paulo, Brazil). Journal of Petrology 33, 1371±1404.
Brigatti, M.F., Medici, L., Saccani, E., Vaccaro, C., 1996. Crystal-chem-
istry and petrologic signi®cance of Fe31-rich phlogopite from the
Tapira carbonatite complex, Brazil. American Mineralogist 81,
913±927.
Brod, J.A., 1999. Petrology and geochemistry of the Tapira alkaline
complex, Minas Gerais State, Brazil. Unpublished PhD thesis,
University of Durham, UK.
Brod, J.A., Gibson, S.A., Thompson, R.N., Junqueira-Brod, T.C., Seer,
H.J., Moraes, L.C., Boaventura, G.R., 2000. Kamafugite af®nity of
the Tapira alkaline-carbonatite complex (Minas Gerais, Brazil). Revista
Brasileira de GeocieÃncias 30, 404±408.
Carvalho, W.T., 1974. Aspectos geoloÂgicos e petrogra®cos do complexo
ultrama®co-alcalino de CatalaÄo I, GO. Proceedings of the 28th
Congresso Brasileiro de Geologia, vol. 5, Porto Alegre, Brazil, SBG,
pp. 107±123.
Carvalho, W.T., Bressan, S.R., 1981. DepoÂsitos minerais associados ao
Complexo ultrama®co-alcalino de CatalaÄo I Ð GoiaÂs. In: Schmaltz,
W.H. (Ed.). Os principais depoÂsitos minerais da RegiaÄo Centro Oeste,
vol. 6. DNPM, Brasilia, pp. 139±183.
Comin-Chiaramonti, P., Censi, P., Cundari, A., De Min, A., Gomes, C.B.,
Marzoli, A., Piccirillo, E.M., 1996. Petrochemistry of Early Cretaceous
potassic rocks from the Asuncion-Sapucai graben, Central-Eastern
Paraguay. In: Comin-Chiaramonti, P., Gomes, C.B. (Eds.). Alkaline
Magmatism in Central-Eastern Paraguay: Relationships with Coeval
Magmatism in Brazil. Edusp/Fapesp, SaÄo Paulo, pp. 123±149.
Comin-Chiaramonti, P., Cundari, A., Gomes, C.B., Piccirillo, E.M., Censi,
P., Demin, A., Bellieni, G., Velazquez, V.F., Orue, D., 1992. Potassic
dyke swarm in the Sapucai graben, eastern Paraguay Ð petrographic,
mineralogical and geochemical outlines. Lithos 28, 283±301.
Conticelli, S., Peccerillo, A., 1992. Petrology and geochemistry of potassic
and ultrapotassic volcanism in central Italy Ð petrogenesis and infer-
ences on the evolution of the mantle sources. Lithos 28, 221±240.
Cruciani, G., Zanazzi, P.F., Quartieri, S., 1995. Tetrahedral ferric iron in
phlogopite-Xanes and Mossbauer compared to single-crystal X-ray
data. European Journal of Mineralogy 7, 255±265.
Danni, J.C.M., Baecker, M.L., Ribeiro, C.C., 1991. The geology of the
CatalaÄo I carbonatite complex. In: Leonardos, O.H., Meyer, H.O.A.,
Gaspar, J.C. (Eds.). Field Guide Book of the 5th International
Kimberlite Conference. CPRM, AraxaÂ, pp. 25±30 (Special Publication
3/91).
Danni, J.C.M., Gaspar, J.C., 1992. Mineralogia e quõÂmica do katungito de
AmorinoÂpolis, GoiaÂs. Proceedings of the 37th Congresso Brasileiro de
Geologia, SaÄo Paulo, pp. 85±86.
Droop, G.T.R., 1987. A general equation for estimating Fe31 concentra-
tions in ferromagnesian silicates and oxides from microprobe analyses,
using stoichiometric criteria. Mineralogical Magazine 51, 431±435.
Dyar, M.D., 1987. A review of Mossbauer data on trioctahedral micas:
evidence for tetrahedral Fe31 and cation ordering. American
Mineralogist 72, 102±112.
Dymek, R.F., 1983. Titanium, aluminum and interlayer cation substitutions
in biotite from high-grade gneisses, West Greenland. American
Mineralogist 68, 880±899.
Edgar, A.D., 1979. Mineral chemistry and petrogenesis of an ultrapotassic±
ultrama®c volcanic rock. Contributions to Mineralogy and Petrology
71, 171±175.
Edgar, A.D., Arima, M., 1983. Conditions of phlogopite crystallization in
ultrapotassic volcanic rocks. Mineralogical Magazine 47, 11±19.
Farmer, G.L., Boetcher, A.L., 1981. Petrologic and crystal-chemical
signi®cance of some deep-seated phlogopites. American Mineralogist
66, 1154±1163.
Gaspar, J.C., 1989. Geologie et mineralogie du complexe carbonatitique de
Jacupiranga, Bresil. Unpublished PhD thesis, Universite d'Orleans.
Gaspar, J.C., 1992. On the use of the term jacupirangite. Proceedings of the
37th Congresso Brasileiro de Geologia, SBG, SaÄo Paulo, pp. 89±90.
Gaspar, J.C., ArauÂjo, D.P., 1995. Reaction products of carbonatite with
ultrama®c rocks in the CatalaÄo I complex, Brazil: possible implications
for mantle metasomatism. Proceedings of the 6th International
Kimberlite Conference, Extended Abstracts, Novosibirsk, pp. 181±183.
Gaspar, J.C., ArauÂjo, A.L.N., Melo, M.V.L.C., 1998. Olivine in carbonatitic
and silicate rocks in carbonatite complexes. Proceedings of the 7th
International Kimberlite Conference, Extended Abstracts, Cape
Town, pp. 239±241.
Gaspar, J.C., Silva, A.J.G.C., ArauÂjo, D.P., 1994. Composition of priderite
in phlogopitites from the CatalaÄo I carbonatite complex, Brazil.
Mineralogical Magazine 58, 409±415.
Gaspar, J.C., Wyllie, P.J., 1982. Barium phlogopite from the Jacupiranga
carbonatite, Brazil. American Mineralogist 67, 997±1000.
Gaspar, J.C., Wyllie, P.J., 1983a. Ilmenite (high Mg, Mn, Nb) In the
carbonatites from the Jacupiranga Complex, Brazil. American
Mineralogist 68, 960±971.
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296 295
Gaspar, J.C., Wyllie, P.J., 1983b. Magnetite in the carbonatites from the
Jacupiranga complex, Brazil. American Mineralogist 68, 195±213.
Gaspar, J.C., Wyllie, P.J., 1987. The phlogopites from the Jacupiranga
carbonatite intrusions. Mineralogy and Petrology 36, 121±134.
Gibson, S.A., Thompson, R.N., Leonardos, O.H., Dickin, A.P., Mitchell,
J.G., 1995. The Late Cretaceous impact of the Trindade mantle plume-
evidence from large-volume, ma®c, potassic magmatism in SE Brazil.
Journal of Petrology 36, 189±229.
Gierth, E., Baecker, M.L., 1986. A mineralizacËaÄo de nioÂbio e as rochas
alcalinas associadas no complexo CatalaÄo I, GoiaÂs. In: Schobbenhaus,
C. (Ed.). Principais depoÂsitos minerais do Brasil, vol. 2. MME/DNPM,
Brasilia, pp. 456±462.
Gomes, C.B., Comin-Chiaramonti, P., VelaÂzquez, V.F., OrueÂ, D., 1996a.
Alkaline magmatism in Paraguay: a review. In: Comin-Chiaramonti, P.,
Gomes, C.B. (Eds.). Alkaline Magmatism in Central-Eastern Paraguay:
Relationships with Coeval Magmatism in Brazil. Edusp/Fapesp, SaÄo
Paulo, pp. 31±56.
Gomes, C.B., Morbidelli, L., Ruberti, E., Comin-Chiaramonti, P., 1996b.
Comparative aspects between post-Palaeozoic alkaline rocks from the
western and eastern margins of the Parana Basin. In: Comin-
Chiaramonti, P., Gomes, C.B. (Eds.). Alkaline Magmatism in
Central-Eastern Paraguay: Relationships with Coeval Magmatism in
Brazil. Edusp/Fapesp, SaÄo Paulo, pp. 249±274.
Greenwood, J.C., 1998. Barian±titanian micas from Ilha da Trindade,
South Atlantic. Mineralogical Magazine 62, 687±695.
Hasui, Y., Cordani, U.G., 1968. Idades PotaÂssio-ArgoÃnio de rochas
eruptivas mesozoÂicas do oeste mineiro e sul de GoiaÂs. Proceedings of
the 22nd Congresso Brasileiro de Geologia, pp. 139±143.
Heathcote, R.C., McCormick, G.R., 1989. Major-cation substitution in
phlogopite and evolution of carbonatite in the Potash Sulfur Springs
complex, Garland County, Arkansas. American Mineralogist 74,
132±140.
Huang, Y.M., Hawkesworth, C.J., Vancalsteren, P., McDermott, F., 1995.
Geochemical characteristics and origin of the Jacupiranga carbonatites,
Brazil. Chemical Geology 119, 79±99.
Issa Filho, A., Lima, P.R.A.S., Souza, O.M., 1984. Aspectos da geologia do
complexo carbonatõÂtico do Barreiro, AraxaÂ, MG, Brasil. Complexos
Carbonatiticos do Brasil: Geologia. CBMM, SaÄo Paulo (pp. 20±44).
Lalonde, A.E., Rancourt, D.G., Chao, G.Y., 1996. Fe-bearing trioctahedral
micas from Mont Saint-Hilaire, Quebec, Canada. Mineralogical
Magazine 60, 447±460.
Laughlin, A.W., Charles, R.W., Aldrich, M.J., 1989. Heteromorphism and
crystallization paths of katungites, Navajo volcanic ®eld, Arizona,
USA. Kimberlites and Related Rocks, Geological Society of Australia
(Special Publication 14), pp. 582±591.
Leonardos, O.H., Ulbrich, M.N., Gaspar, J.C., 1991. The Mata da Corda
volcanic rocks. In: Leonardos, O.H., Meyer, H.O.A., Gaspar, J.C.
(Eds.). Field Guide Book of the 5th International Kimberlite
Conference. CPRM, AraxaÂ, pp. 17±24 (Special Publication 3/91).
Lloyd, F.E., Bailey, D.K., 1991. Complex mineral textures in bebedourite:
possible links with alkali clinopyroxenite xenoliths and kamafugitic
volcanism. In: Leonardos, O.H., Meyer, H.O.A., Gaspar, J.C. (Eds.).
Proceedings of the 5th International Kimberlite Conference: Extended
Abstracts. CPRM, AraxaÂ, pp. 263±269 (Special Publication 3/91).
Machado-Junior, D.L., 1992. Geologia do complexo alcalino-carbonatõÂtico
de CatalaÄo II (GO). Proceedings of the 37th Congresso Brasileiro de
Geologia. Extended Abstracts. SBG, SaÄo Paulo, pp. 94±95.
Machado-Junior, D.L., 1992. Idades Rb/Sr do complexo alcalino-
carbonatõÂtico de CatalaÄo II (GO). Proceedings of the 37th Congresso
Brasileiro de Geologia. Extended Abstracts. SBG, SaÄo Paulo,
pp. 91±93.
McCormick, G.R., Heathcote, R.C., 1987. Mineral chemistry and petrogen-
esis of carbonatite intrusions, Perry and Conway Counties, Arkansas.
American Mineralogist 72, 59±66.
McCormick, G.R., Le Bas, M.J., 1996. Phlogopite crystallization in carbo-
natitic magmas from Uganda. Canadian Mineralogist 34, 469±478.
Melo, M.V.L.C., 1999. QuõÂmica dos minerais das rochas do complexo
carbonatõÂtico de CatalaÄo II: implicacËoÄes petrogeneÂticas. Unpublished
MSc thesis, University of Brasilia, Brasilia.
Mitchell, R.H., 1978. Manganoan magnesian ilmenite and titanian
clinohumite from the Jacupiranga carbonatite, Sao Paulo, Brazil.
American Mineralogist 63, 544±547.
Mitchell, R.H., 1995. Compositional variation of micas in kimberlites,
orangeites, lamproites and lamprophyres. Proceedings of the 6th
International Kimberlite Conference. Extended Abstracts. Novosibirsk,
pp. 390±392.
Mitchell, R.H., 1995b. Kimberlites, Orangeites and Related Rocks. Plenum
Press, New York.
Mitchell, R.H., Bergman, S.C., 1991. Petrology of Lamproites. Plenum
Press, New York.
Morbidelli, L., Beccaluva, L., Brotzu, P., Conte, A.M., Garbarino, C.,
Gomes, C.B., Grossi-Sad, J.H., Riffel, B.F., Ruberti, E., Traversa, G.,
(1995). Aspectos mineraloÂgicos e petrogra®cos de rochas ultrama®cas e
carbonatitos do complexo alcalino de Salitre, GO. Proceedings of the
5th Congresso Brasileiro de GeoquõÂmica. SBGq, NiteroÂi (CD-ROM
edition).
Morikiyo, T., Hirano, H., Matsuhisa, Y., 1990. Carbon and oxygen isotopic
composition of the carbonates from the Jacupiranga and CatalaÄo I
carbonatite complexes, Brazil. Bulletin of the Geological Survey of
Japan 41, 619±626.
Rieder, M., Cavazzini, G., D'Yakonov, Y.S., Frank-Kamenetskii, V.A.,
Gottardi, G., Guggenheim, S., Koval, P.V., MuÈller, G., Neiva,
A.M.R., Radoslovich, E.W., Robert, J.L., Sassi, F.P., Takeda, H.,
Weiss, Z., Wones, D.R., 1998. Nomenclature of the micas. Canadian
Mineralogist 36, 905±912.
Robert, J.L., 1976. Titanium solubility in synthetic phlogopite solid
solutions. Chemical Geology 17, 213±227.
Roden, M.F., Rama Murthy, V., Gaspar, J.C., 1985. Sr and Nd isotopic
composition of the Jacupiranga carbonatite. Journal of Geology 93,
212±220.
Santos, R.V., Clayton, R.N., 1995. Variations of oxygen and carbon
isotopes in carbonatites Ð a study of Brazilian alkaline complexes.
Geochimica et Cosmochimica Acta 59, 1339±1352.
Seifert, W., Kampf, H., 1994. Ba-enrichment in phlogopite of a nephelinite
from Bohemia. European Journal of Mineralogy 6, 497±502.
Sgarbi, P.B.A., ValencËa, J.G., 1994. Mineral and rock chemistry of the
Mata da Corda kamafugitic rocks (MG State, Brazil). Proceedings of
the International Symposium on the Physics and Chemistry of the
Upper Mantle. Extended Abstracts. CPRM/FAPESP, SaÄo Paulo,
pp. 27±29.
Sonoki, I.K., Garda, G.M., 1988. Idades K-Ar de rochas alcalinas do Brasil
Meridional e Paraguai Oriental: compilacËaÄo e adaptacËaÄo as novas
constantes de decaimento. Boletim do IG-USP (SeÂrie Cientõ®ca) 19,
63±85.
Thompson, R.N., Velde, D., Leat, P.T., Morrison, M.A., Mitchell, J.G.,
Dickin, A.P., Gibson, S.A., 1997. Oligocene lamproite containing an
Al-poor, Ti-rich biotite, Middle Park, northwest Colorado, USA.
Mineralogical Magazine 61, 557±572.
Ulbrich, H.H.G.J., Gomes, C.B., 1981. Alkaline rocks from continental
Brazil. Earth Sciences Reviews 17, 135±154.
Wones, D.R., Eugster, H.P., 1965. Stability of biotite: experiment, theory,
and application. American Mineralogist 50, 1228±1272.
Zaitsev, A., Polezhaeva, L., 1994. Dolomite-calcite textures in early
carbonatites of the Kovdor ore deposit, Kola Peninsula, Russia: their
genesis and application for calcite-dolomite geothermometry. Contribu-
tions to Mineralogy and Petrology 115, 339±344.
J.A. Brod et al. / Journal of Asian Earth Sciences 19 (2001) 265±296296