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


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.

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


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

AraxaÂ 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


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 AraxaÂ 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


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.


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


Al

Mg Fe (T)


Phlogopite 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)


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


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).


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


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


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


Phlogopite

Al

Mg Fe(T)


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


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


.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


(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-


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.


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.


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 AraxaÂ (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

AraxaÂ complex as being of metasomatic origin, formed by

phlogopitisation of ultrama®c rocks. Although they do not

provide phlogopite analyses (and the AraxaÂ mica studied by

Cruciani et al., 1995 comes from a sovite), the textural

features of the AraxaÂ 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,


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


Al

Mg Fe(T)


Phlogopite 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.


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

ParanaÂ 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 JuquiaÂ 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, JuquiaÂ 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 JuquiaÂ 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 ParanaÂ ¯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-


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

JuquiaÂ 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 JuquiaÂ are chemically similar to minettes.

² The widely variable chemical composition of phlogo-

pite in carbonatite±silicate associations has important


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 ultramaÂ®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 petrograÂ®cos do complexo

ultramaÂ®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 ultramaÂ®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


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



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 ParanaÂ 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 petrograÂ®cos de rochas ultramaÂ®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


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 CientõÂ®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.


Phlogopite and tetra-ferriphlogopite from Brazilian carbonatite complexes: petrogenetic constraints...

Documents

Transcript of Phlogopite and tetra-ferriphlogopite from Brazilian carbonatite complexes: petrogenetic constraints...