Beruflich Dokumente
Kultur Dokumente
1093/petrology/egq058
Agpaitic nepheline syenites contain complex zirconium silicate min- eudialyte-dominated liquidus assemblages (and possibly further to
erals as primary magmatic phases; agpaitic liquidus mineral assem- hyperagpaitic residual liquids); (2) the increasing water trend, lead-
blages are controlled by compositional parameters such as alkalinity ing to catapleiite crystallization; (3) the increasing fluorine trend,
and concentrations of water and halogens in the magma. The giving transitionally agpaitic liquidus assemblages with wo« hlerite,
Larvik Plutonic Complex in the late Palaeozoic Oslo Rift, Norway lvenite and hiortdahlite, with or without zircon, but eventually
consists of hypersolvus monzonite (larvikite) and different varieties with magmatic fluorite; this type of evolution is applicable to
of nepheline syenite. The nepheline syenite members have miaskitic weakly agpaitic systems such as the Oslo Rift pegmatites.The evolu-
mineralogy (i.e. zircon stable), except for a suite of late, mildly tion of such magmas will most probably terminate at liquidus bound-
agpaitic pegmatites, which contain zirconium silicate minerals of aries also involving eudialyte and/or rosenbuschite, and they are
the wo« hlerite, rosenbuschite and eudialyte groups, in some cases to- unlikely to evolve to more strongly agpaitic compositions.
gether with zircon.These minerals form part of the liquidus mineral
assemblage of the nepheline syenite magma together with fluorite
and the rock-forming minerals alkali feldspar, aegirine(^augite), KEY WORDS: agpaitic rocks; Oslo Rift; nepheline syenite; zirconium
alkali amphibole, biotite, nepheline and sodalite. Low-variance min- silicate minerals
eral assemblages with two or three zirconium silicate minerals, fluor-
ite and the major rock-forming minerals can be modelled in terms of
phase equilibria in a multi-component system involving the sodium I N T RO D U C T I O N
disilicate component of the melt and the volatile components water, Zircon is an abundant accessory mineral, and in most
fluorine and chlorine. A semiquantitative isothermal and isobaric igneous rocks it is the main host mineral for zirconium.
petrogenetic grid in log activity space constructed from the observed Factors that control the crystallization of zircon from a sili-
mineral assemblages and the compositions of the coexisting minerals cate melt include the concentration of zirconium, the
indicate that there are at least three trends of evolution leading from major element composition and volatile content of the
a miaskitic to an agpaitic crystallization regime in peralkaline neph- magma, as well as temperature (e.g. Watson & Harrison,
eline syenites: (1) the increasing alkalis trend, typically leading to 1983; Baker et al., 2002). However, in some highly evolved
Table 1: Mineral formulas for Zr-silicates and other (e.g. Motzfeldt and Gardiner complexes, Greenland;
McGerrigle and Mont Saint-Hilaire, Canada; Oslo Rift,
important minerals from nepheline syenite pegmatites in Norway; Srensen, 1997, and references therein), and
the Oslo Rift agpaitic mineral assemblages have formed by late- to
post-magmatic, fluid-induced processes in miaskitic neph-
Eudialyte-group eline syenite in Tamazeght, Morocco (Schilling et al., 2009,
Eudialyte Na15Ca6(Fe2þ,Mn2þ)3Zr3(Si25O73)(O,OH,H2O)3 and references therein).
(OH,Cl)2 The factors controlling the transition from a miaskitic
Ferrokentbrooksite Na15Ca6(Fe,Mn)3Zr3NbSi25O73(O,OH,H2O)3Cl2 (i.e. zircon present) to an agpaitic liquidus mineralogy
2304
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES
multi-component space, which is used to derive an empir- rocks range in composition from alkaline basalt to granite
ical, semiquantitative petrogenetic grid relating the zirco- and nepheline syenite, with a large proportion of inter-
nium silicate liquidus mineralogy to alkalinity and mediate, monzonitic intrusive rocks (larvikite) and com-
volatile activities. This, in turn, provides constraints for positionally equivalent latitic lavas known as rhomb
the relative importance of the parameters controlling the porphyries (Neumann et al., 2004, and references therein).
transition from a miaskitic to an agpaitic crystallization The Larvik Plutonic Complex (Fig. 1b) is a large, compos-
regime in peralkaline nepheline syenite magma. ite intrusion making up the southernmost onshore part of
the Oslo Rift (Larsen et al., 2008; Dahlgren, 2010). The
complex is built up of 10 arc-shaped or circular intrusions
GEOLOGIC A L S ET T I NG of hypersolvus monzonite (larvikite) and nepheline syenite
The Oslo Rift in southeastern Norway is part of a system (Fig. 1b; Petersen, 1978). U^Pb dating of magmatic zircon
of late Carboniferous to Permian continental rifts in north- suggests a relatively short period of emplacement and crys-
ern Europe (Neumann et al., 2004). Rift-related igneous tallization (299^292 Ma, Dahlgren, 2010), overlapping in
2305
JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 NOVEMBER 2010
time with the main rifting episode and widespread rhomb pegmatites. This stage is usually characterized by extensive
porphyry volcanism (Larsen et al., 2008). Trace element zeolitization and alteration of the magmatic minerals, and
distribution patterns and radiogenic isotopic signatures of by crystallization of low-temperature hydroxides and
larvikite and rhomb porphyry point towards an origin hydrous silicates. Many of the rare REE- and beryllium-
from the lithospheric mantle (Neumann et al., 2004, and bearing minerals belong to this stage of pegmatite forma-
references therein). The rocks of the Larvik Plutonic tion. A few minerals (e.g. goethite, hisingerite, neotocite)
Complex range from mildly quartz normative in the have formed as the result of supergene processes (Larsen,
early, eastern ring intrusions, to strongly nepheline norma- 2010). The present study is concerned only with the first,
tive in the younger intrusions in the west and magmatic stage of pegmatite crystallization. The host lar-
2306
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES
silicate minerals (Fig. 2b). Zircon and the other zirconium lvenite has been replaced by symplectites consisting of
silicate minerals commonly occur as euhedral to subhedral fluorite, magnetite and zircon (Fig. 2k, arrows); according
crystals with well-developed crystal face terminations to Brgger (1890), this type of alteration is common for
against matrix feldspar, nepheline and sodalite, or lvenite in the area.
they are included in or intergrown with biotite and/or Rosenbuschite typically forms pale, acicular crystals
pyroxene. intergrown with hiortdahlite (Fig. 2f) and eudialyte
(Fig. 2g); in sample TRO, it replaces hiortdahlite (Fig. 2h).
Rosenbuschite is closely associated with fluorite, which is
Occurrence and textural relationships a general observation in these pegmatites (Brgger, 1890;
of the zirconium silicate minerals
2307
JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 NOVEMBER 2010
2308
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES
Table 2: Summary of critical mineral assemblages in pegmatites from the Larvik Plutonic Complex
1
Localities marked in Fig. 1.
2
The paragenetic relationship between catapleiite and the other minerals is unclear.
a1, a2, a3, etc.: successive mineral assemblages in the pegmatite.
(WDS) spectrometers. An accelerating voltage of 15 kV, Matrix corrections were made according to the PAP
beam current of 15 nA and a focused beam were generally procedure (Pouchou & Pichoir, 1984).
used. Beam-sensitive minerals (catapleiite and eudialyte) In XRD analysis, crystalline materials examined in this
were analyzed at 10 nA and a beam diameter of 10 mm. study were identified using a Siemens D5005 X-ray
Counting time was 10 s on peak (and 5 s on each back- powder diffractometer equipped with a ceramic X-ray
ground positions). Na, F, Cl and K were analyzed first. tube (2·2 kW) and a Cu-Ka-anode and Debye^Scherrer
Prior to analysis, full-range spectra were acquired on sev- configuration, combined with BRUKER software for
eral grains of each mineral species to identify all the elem- evaluation (EVA) with the latest PDF database. Several of
ents present and possible interferences between X-ray the specimens in the Oslo collection were originally identi-
lines. Background positions for each mineral type were fied by XRD using 9 cm Debye^Scherrer cameras and
then selected on detailed WDS scans acquired around the Fe-Ka radiation; the films are kept in the archives of the
X-ray lines of interest. Natural History Museum, University of Oslo.
Calibration standards and X-ray lines used were wollas-
tonite (Si Ka, Ca Ka), Al2O3 (Al Ka), pyrophanite (Ti
Ka, Mn Ka), Fe metal (Fe Ka), MgO (Mg Ka) orthoclase Mineral compositions
(K Ka), albite (Na Ka), fluorite (F Ka), synthetic alforsite Electron microprobe data are given in Electronic
(Cl Ka), Monastery Mine zircon (Zr La), Hf metal (Hf Appendix 3 (http://www.petrology.oxfordjournals.org).
Ma), Si^Al^Ca glass with 15% ThO2 (Th Ma), Nb2O5 Hiortdahlite, wo«hlerite, lvenite and rosenbuschite were
(Nb La), Ta2O5 (Ta Ma), scheelite (W Ma), synthetic recalculated to structural formulae assuming Si ¼ 4·00 cat-
orthophosphates (Jarosewich & Boatner, 1991) for REE ions. This is justified by the A6þxB2^x(Si2O7)2(F,O,OH)4
and Y (La La, Ce La, Nd Lb, Y La). stoichiometry of these minerals, but it should be noted
2309
JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 NOVEMBER 2010
2310
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES
2311
JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 NOVEMBER 2010
M O D E L L I N G T H E S TA B I L I T Y O F
Z I RC O N I U M S I L I C AT E L I Q U I D U S
A S S E M B L AG E S I N P E R A L K A L I N E
N E P H E L I N E SY E N I T E M AG M A
Primary and secondary zirconium silicate
mineral assemblages
The minerals of the wo«hlerite, rosenbuschite and eudialyte
groups are part of the magmatic mineral assemblage of
indicated by black dots have been taken from table 3 of Johnsen et al.
(2003). (a) Total Mn^Fe^Nb distribution, illustrating that Fe4Mn in
Fig. 4. Cation distributions in eudialyte s.l., from data given in eudialyte in these samples. (b) M3 cation distribution, showing
Electronic Appendix 3. Structural formulae have been calculated roughly equal Nb and Si occupancies in this position. (c) N-cation
assuming Si þ Al þTi þ Zr þ Hf þ Nb ¼ 29·000; cations have been distribution, illustrating the significant contribution of Ca þ REE in
assigned to structural positions as given by Johnsen et al. (2003) using eudialyte from these pegmatites. (d) Vacant and occupied N-pos-
an adaptation of the algorithm of Pfaff et al. (2010), which allows itions, shown in a NaN^(Ca þ REE)N^(H3Oþ þ œ)N diagram. The
excess Ca þ REE to be allocated to the N-positions in the structure sum of vacancies and H3Oþ ions in the N position has been estimated
without regard for other site occupancies. Endmember compositions assuming 16 N-positions.
2312
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES
Zircon forms part of early mineral assemblages in the ferric iron from the system, we disregard the effect of
Langodden and Aryskjrene pegmatites (a2, e1). oxygen fugacity. In these pegmatites, fO2 is, however, in-
Secondary fluorite þ zircon catapleiite assemblages are ternally controlled by reactions involving titanomagnetite,
found in several of the samples studied (assemblages a3, microcline and biotite.
d3, e3, k2; Table 2). Other late magmatic or secondary as- Stoichiometric endmember compositions for the major
semblages are made up by fluorite þ low-iron lvenite (as- rock-forming minerals microcline, albite, nepheline and
semblage c3), replacing hiortdahlite in the Trompetholmen sodalite are used in the model calculations (Table 3).
pegmatite (Fig. 2h) and wo«hlerite in LV15 (assemblage k2, Micas in the pegmatites are close to annite in composition
Fig. 2l). Lvenite is itself commonly replaced by pseudo- (Larsen, 1998), and we include a pure hydroxy annite end-
2313
JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 NOVEMBER 2010
Table 3: Composition (in moles per formula unit) of phases and species used in modelling
SiO2 AlO1·5 ZrO2 TiO2 NbO2·5 REEO1·5 FeO CaO NaO0·5 KO0·5 HO0·5 FO–0·5 ClO–0·5
1
Mineral compositions from Electronic Appendix 1.
2
Additional data from Larsen (2001).
(reaction (1) in Table 4), and thus independent of which þ pyrochlore þ two zirconium silicate minerals þ volatile-
zirconium silicate minerals are stable. bearing silicate liquid) define a formally divariant mineral
Whereas there is compelling evidence for the presence of assemblage in the chlorine-free sub-system, which can be
a free, aqueous fluid phase during the later, miarolitic and represented by a plane in 3D log activity space. For reac-
hydrothermal stages of pegmatite crystallization (Brgger, tions involving eudialyte, ClO^0·5 must be included as a
1890; Larsen, 2010), this is commonly not the case during component, and the number of phases present must be
initial, magmatic crystallization of agpaitic nepheline sy- increased to 10, adding sodalite. The divariant planes sep-
enite magma (Srensen, 1997), and we do not invoke the arate (trivariant) stability volumes for the single zirco-
presence of a separate non-silicate fluid in the following nium silicate minerals, pairs of planes intersect along
analysis of magmatic phase reactions. It should, however, univariant lines (10 phases in the chlorine-free subsystem)
be noted that Brgger (1890) observed abundant two-phase and three planes intersect in isobaric^isothermal invariant
aqueous(?) fluid inclusions in catapleiite. No systematic points represented by 11 (or 12, including sodalite) coexist-
search for fluid inclusions has been made in the present ing phases.
study, but observations of crystal-face limited microcavities A reaction between several phases and/or chemical
in catapleiite in some samples (Fig. 2c) support the hypoth- species in a multicomponent system can be written as
esis that at least some of the late catapleiite has crystallized X
in the presence of a free, aqueous fluid. ni X i ¼ 0
i
At constant temperature and pressure, with external buf-
fering of TiO2 and REE2O3, nine coexisting phases (e.g. where ni is the stoichiometric coefficient of the ith species
microcline þ albite þ nepheline þ biotite þ clinopyroxene involved in the reaction (Xi). By convention, ni is positive
2314
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES
Table 4: Divariant mineral reactions between zirconium silicate minerals, major pegmatite minerals and melt
Fluorite saturation
(1) 0·25 Ne þ 0·08 Kf þ 1·00 Hdcpx þ 2·00 HF ¼ 1·00 Flu þ 0·75 Ab þ 0·33 Annbi þ 0·67 H2O
Sodalite–nepheline–HCl reaction
(2) 0·38 Ab þ 0·88 Ne þ 0·50 H2O þ 0·50 Nds þ 1·00 HCl ¼ 1·00 Sod þ 0·88 Kf
Hiortdahlite volume
(3) 0·89 Hio þ 2·60 Ab þ 1·25 Annbi þ 0·26 H2O ¼ 1·00 Zrc þ 0·87 Ne þ 0·38 Kf þ 3·96 Hdcpx þ 0·08 TiO2 þ 0·05 Pcl þ 0·04 REE2O3 þ 0·78 Nds þ
Abbreviations used for idealized endmember minerals and dissolved components in the melt are given in Table 3.
2315
JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 NOVEMBER 2010
for products and negative for reactants. The equilibrium 3D activity space represent the phase equilibria appropri-
constant for the reaction (log K) is given by ate for the nepheline syenite pegmatites. Although both
X the fluorite-saturated and -undersaturated domains must
log K ¼ ni log ai be considered, it is most convenient to start with the distri-
i
bution of stability fields at fluorite saturation, as this
where ai is the activity of the ith species. reduces the problem to a case of 2D Schreinemakers ana-
In the present case, the reactants and products are: (1) lysis. Figure 5a shows five possible configurations of stabil-
solid phases of constant composition (endmember feldspar, ity fields in the fluorite-saturated, Cl-free subsystem
nepheline, sodalite, hydroxy-annite, zirconium silicate surrounding the hiortdahlite (Hio) field (shaded) in a pro-
þnNds log aNds þ nH2 O log aH2 O þ nHF log aHF : wo«hlerite and two of the minerals hiortdahlite, catapleiite
and rosenbuschite (scenario ii in Fig. 5). None of the critical
Only the three last terms include the activities of four-phase assemblages that would confirm the validity of
Na2Si2O5, H2O and HF in the pegmatite magma. In the either scenario i or ii have been observed. Although the
absence of thermodynamic data on the participating min- present observations do not support the existence of any
erals and oxide species in the silicate liquid, G for the magmatic mineral assemblage involving stable Woe þ Flu,
reaction, and thereby log K, cannot be determined. there is no conclusive evidence against. We therefore chose
Furthermore, the activity terms for the pyrochlore and scenario ii as the preferred alternative, but with a
clinopyroxene solid solution, and for the minor oxide spe- Woe þ Flu field that is small, and/or restricted to combin-
cies are (at best) poorly constrained. Nevertheless, all ations of H2O and Nds activities that are not realized in
these terms can be regarded as constant with respect to the nepheline syenite pegmatites. It is possible that other
the three activity terms of interest, and the equation geometrical configurations will be stabilized by changes
simplifies to in temperature, pressure or other parameters not con-
sidered here. For example, decreasing fO2 would be likely
nNds log aNds þ nH2 O log aH2 O þ nHF log aHF ¼ constant:
to expand the stability field of lvenite with essential Fe2þ
This is the equation of a plane in an orthogonal, 3D co- at the expense of the low-iron zirconium silicate minerals,
ordinate system with log aNds, log aH2O and log aHF as eventually stabilizing a configuration like scenario iv in
axes. Although the right-hand side of the equation (and Fig. 5a.
hence the absolute coordinates of the plane) remains un-
constrained, its orientation in activity space is uniquely The 3D petrogenetic grid
given by the ratios of the stoichiometric coefficients; for ex- The 2D configuration of the fluorite saturation plane
ample, the slope of the intersection of a plane with a sec- (Fig. 5a) constrains the geometry of the fluorite-
tion parallel with the log aNds and log aHF axes is given by undersaturated part of the 3D petrogenetic grid. The
^nNds/nHF, and similar for the slopes in the projections divariant fluorite saturation surface (reaction (1) in
defined by the two other combinations of orthogonal axes. Table 4) is parallel to the log aNds axis, and has a positive
slope of 0·33 in the log aH2O^log aHF projection (Fig. 6).
An internally consistent petrogenetic grid The stability field boundaries in Fig. 5 are the intersections
Using idealized phase compositions listed in Table 3, 20 of this plane with divariant surfaces in the fluorite-
formally divariant, balanced mineral^melt reactions can undersaturated system, the features of which are conveni-
be defined (Table 4). ently illustrated in a projection of the fluorite saturation
surface to the log aNds^log aH2O plane (Fig. 7a) and in sec-
Fluorite-saturated case tions parallel to the log aNds^log aHF and log aH2O^log aHF
For the fluorite-saturated case comparison with observed planes (Fig. 7b^g). The grid is semiquantitative, in
mineral assemblages is necessary to determine which of the sense that it is internally consistent and gives a
several possible configurations of divariant reactions in true representation of the relative slopes of the divariant
2316
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES
Fig. 5. (a) Possible configurations of stability fields of hiortdahlite (Hio, shaded), zircon (Zrc), catapleiite (Cat), rosenbuschite (Ros), lvenite
(Lav) and wo«hlerite (Woe) in the presence of fluorite, projected to the log aNds^log aH2O plane (Nds is the Na2Si2O5 component in the pegmatite
melt). Scenarios i^v show five mutually exclusive configurations of stability fields, which are all consistent with a 3D Schreinemakers analysis
of the divariant reactions in Table 4 in log aNds^log aH2O^log aHF space. (b) Two- and three-phase assemblages predicted from the geometrical
configurations in (a). For each matrix, the header row shows single phases expected to be stable (þ) or unstable (^) with fluorite. In the body
of the matrix, pairs of phases that are expected to occur stably with fluorite are indicated by (þ), and combinations that are not expected to
occur by (^). Grey shades represent predicted phase assemblages that have actually been observed in the pegmatites, and black fields ‘forbidden’
combinations that have been found in the samples. Observation of one or more ‘forbidden’ phase assemblages is evidence against that particular
stability field configuration.
2317
JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 NOVEMBER 2010
planes and univariant lines in 3D log activity space, and of (Fig. 7b and c). The wedge-shaped stability volumes for
the relative positions of the stability volumes of the single wo«hlerite, lvenite and rosenbuschite occur at higher aHF
zirconium silicate minerals. However, it gives no informa- levels, and expand towards increasing aNds, but are trun-
tion on the absolute values of component activities, and cated by the eudialyte volume at elevated HCl activity.
the relative sizes of the divariant fields and the stability The stability volume of hiortdahlite is bounded by the vol-
volumes of single phases are arbitrary. Quantitative cali- umes of zircon (at low aNds), catapleiite (high aH2O), rosen-
bration of the grid would require experimental data on buschite (high aNds), wo«hlerite (low aHF), lvenite (low
critical mineral assemblages, and/or thermodynamic data aH2O) and eudialyte (high aNds and aHCl). The relative
on the zirconium silicate minerals and on components in slopes of the divariant planes limiting the hiortdahlite
the volatile-bearing silicate melt, none of which is volume are such that the volume can never extend to
available. great depths within the fluorite-undersaturated part of the
The chlorine-free subsystem has stability volumes for diagram. Hiortdahlite is therefore stable in nepheline syen-
catapleiite, zircon, wo«hlerite, hiortdahlite, rosenbuschite ite magmas only at or close to saturation with fluorite.
and lvenite. The boundary between the zircon and cata- The same can be said about rosenbuschite, but in low-Cl
pleiite volumes is subparallel to the log aHF axis, and systems, the rosenbuschite volume and the Ros þWoe plane
at very low aHF, these two, and the stability volume of will extend to increasingly greater depths below fluorite
eudialyte are the only ones that need to be considered saturation at increasing aNds.
2318
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES
Fig. 7. Sections through the semiquantitative, 3D petrogenetic grid in the multicomponent system in log aNds^log aH2O^log aHF space, drawn
according to scenario ii in Fig. 5. Zrc, zircon; Cat, catapleiite; Ros, rosenbuschite; Lav, lvenite; Hio, hiortdahlite; Woe, wo«hlerite. Inferred crystal-
lization conditions for the pegmatites in this study are shown as circles and ellipses. Inferred evolution with time is indicated by arrows.
Continuous-line outlines and arrows reflect conditions during magmatic crystallization, dashed-line outlines and arrows post-magmatic,
fluid-induced reactions. The circled fields are linked to the corresponding mineral assemblages in Table 2 by lower-case letters. (a) The fluorite
saturation surface (Fig. 6), projected to the log aNds^log aH2O plane. Thin dashed lines identified by letters enclosed in diamonds give the pos-
itions of the sections shown in Fig. 7b^g. The extent of the eudialyte field is unconstrained; its maximum possible size is given by dash^dot line
E1, which touches at the Flu þ Zrc þ Hio þ Lav point. The shaded field limited by dash^dot line E2 is a more realistic field for Eud þ Flu in the
presence of Ne þ Sod, as it allows observed assemblages involving fluorite, hiortdahlite, rosenbuschite and wo«hlerite within a restricted range of
component activities, as illustrated by assemblages c2, d2, e2, f and g2. At sodalite-undersaturated conditions, the eudialyte field is shifted to
higher aNds (dash^dot line E3). (b, c) log aHF^log aH2O sections through the 3D petrogenetic grid along profiles indicated in (a). The bold line
represents the fluorite saturation surface (i.e. the maximum possible aHF in this system). The shaded field is the assumed extent of the eudialyte
field in the presence of nepheline and sodalite. (d^g) log aNds^log aHF sections, as indicated in (a).
2319
JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 NOVEMBER 2010
The HCl activity is not tied to the other components by hiortdahlite and catapleiite requires higher aHF than any
mixed volatile reactions. The maximum aHCl is limited by of the other hiortdahlite-bearing mineral assemblages
the simultaneous crystallization of nepheline and sodalite observed.
(reaction (2) in Table 4). In the Cl-bearing system, a The evolution from wo«hlerite-bearing assemblages
volume for eudialyte will appear at high aNds, and expand (assemblages c1, d1, g1) to hiortdahlite þ rosenbuschite
towards lower aNds with increasing aHCl. The non-existence eudialyte (c2, d2, g2) in the pegmatites from
of a stable Zrc þ Eud þ Flu assemblage indicates that the Trompetholmen, Risya and Skutesundskjr suggests a
eudialyte field does not extend as far into the diagram as trend at increasing activities of HF and Nds, ending up at
the Zrc þ Lav þ Hio þ Flu point; the dash^dot line E1 in the Hio þ Ros þ Eud þ Flu point, which may be the
2320
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES
minerals in the pegmatites, and fO2 is likely to be con- The most interesting behaviour is seen within the transi-
trolled by the reaction tional zone within the zircon primary field at low aHF in
Fig. 7. Within this area, the liquidus zirconium silicate min-
Annbi þ 05O2 ¼ Mag þ Kf þ H2 O
eral assemblage of a nepheline syenite magma can be
where Mag is the Fe3O4 component in titanomagnetite. forced to change from miaskitic with stable zircon, to ‘low
This is not a formally univariant oxygen buffer curve, as agpaitic’ assemblages with different successions of wo«hler-
it depends on mineral compositions and water activity, ite-, hiortdahlite- and lvenite-bearing assemblages by a
but as long as biotite and titanomagnetite remain stable, it simple increase in aHF without simultaneous changes in
will prevent wide excursions in oxygen fugacity level. This either water or alkali activity.
2321
JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 NOVEMBER 2010
Metasomatic eudialyte has not been observed in any of the and Khibina (Konnerup-Madsen, 2001; Ryabchikov &
present samples, but crosscutting catapleiite microveins in Kogarko, 2009, and references therein). The presence of
the TYHIO sample (Fig. 2c, i.e. assemblage a3) may have such fluids is a consequence of the low oxygen fugacity
a similar origin. and low solidus temperature of these magmatic systems
Hydrocarbon-bearing fluids have formed during crystal- (Ryabchikov & Kogargo, 2009), rather than a controlling
lization of large, agpaitic intrusions such as Il|¤ maussaq factor for the stability of zirconium silicate mineral
2322
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES
stability. Whether or not organic fluids were present in the concentration-dependent parameters in these highly
nepheline syenite pegmatites of the Larvik Plutonic evolved magmas.
Complex is not known. There are at least three types of process that may change
the liquidus assemblage of a nepheline syenite magma
Wider implications from miaskitic with zircon to agpaitic with complex zirco-
In the absence of thermodynamic data for the species nium silicate minerals: (1) an increase in peralkalinity,
involved in phase equilibria, the method developed here is leading to crystallization of eudialyte; (2) increasing water
semiquantitative. Its main importance is that differences activity (and eventually, saturation with an aqueous fluid
and trends in peralkalinity and volatile activities in a crys- phase), leading to the stabilization of catapleiite-bearing
2323
JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 NOVEMBER 2010
2324
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES
system using eudialyte textures and composition: a case study from Watson, E. B. & Harrison, M. T. (1983). Zircon saturation revisited:
Tamazeght, Morocco. Canadian Mineralogist 47, 351^365. temperature and composition effects in a variety of crustal magma
Sokolova, E. (2006). From structure to chemical composition. I. types. Earth and Planetary Science Letters 64, 295^304.
Structural hierarchy and stereochemistry in titanium disilicate Weibye, P. H. (1850). Neue Mineralien aus Norwegen, beschrieben
minerals. Canadian Mineralogist 44, 1273^1330. von P. H. Weibye; analysiert von N. J. Berlin, K. A. Sjo«gren und J.
Srensen, H. (ed.) (1974). The Alkaline Rocks. London: John Wiley, B. von Borck (Erster Theil). Annalen der Physik und Chemie 79,
622 p. 299^304.
Srensen, H. (1997). The agpaitic rocksçan overview. Mineralogical White, R. W., Powell, R. & Baldwin, J. A. (2008). Calculated phase
Magazine 61, 485^498. equilibria involving chemical potentials to investigate the textural
Srensen, H. (2006). The Il|¤ maussaq Alkaline Complex, South evolution of metamorphic rocks. Journal of Metamorphic Geology 26,
Greenlandçan overiview of 200 years of research and an outlook. 181^198.
2325