The Lambert-Amery rift is the principal tectonic feature in East Antarctica, forming a significant recess in the coastline between 65° and 75°E, and is occupied to a large extent by the Lambert Glacier and Amery Ice Shelf.It is flanked by mid-Proterozoic amphibolite-to granulite facies rocks (known as the Rayner Complex; Sheraton et al., 1980) forming the northern Prince Charles Mountains on the western side of the rift.Archean basement is found adjacent to the rift in the continental interior (southern Prince Charles Mountains; Tingey, 1982) and further from the rift in Enderby Land to the west and in the Vestfold Hills to the east.Magmatic activity associated with the rift occurred episodically over a prolonged period from 460 to 40 Ma (Andronikov et al., 1998).Mande peridotite xenoliths occur in ultramafic alkaline dykes and stocks of Jurassic age in the Jetty Peninsula area at the northwestern margin of the rift close to the Amery Ice Shelf, and in trachybasaltic volcanics of the Manning Massif in the northern Prince Charles Mountains.The present study concerns xenoliths from two intrusive bodies from the Jetty Peninsula area.

Acquiring knowledge about microstructures and textures is crucial for the improvement and development steel products, because these two characteristics are controlling factors for the properties of steel. Diffraction techniques using X-rays, electrons or neutrons are suitable to study microstructures (e.g. phase relationships) and textures (crystallographic orientations). X-ray diffraction (XRD) and electron backscatter diffraction (EBSD) are generally available techniques within an industrial research environment. Different examples from daily research within Corus RD&T are shown, where these diffraction techniques have been used. For the development of new high-tech multi-phase steels we study phase composition and microstructure to optimise product properties. Retained austenite may easily be detected and quantified by XRD, using the right approach. Moreover, using an area-sensitive 2D-detector, such as e.g. the GADDS system, the textures of ferrite and austenite components may be identified simultaneously during a single measurement. Micro XRD (beam diameter down to 100 mu m) allows us to study small-scale welding joints within TRIP steels and how these welds evolve with time. Even more interesting is the possibility to study recrystallisation and phase transformations in situ at high temperature in real time, e.g. to make phase transformations visible during inter-critical annealing. Another fascinating and beneficial attribute of XRD and EBSD is to study coatings/platings such as Ni in relation to the underlying steel substrate. Not only it is possible to identify the thickness of a plating or the phases in a coating, but textural and microstructural relations between the substrate and the surface layer can also be visualised. The understanding of microstructure opens the way to grain boundary engineering in coated products to help improve the corrosion properties.

Garnet and spinel peridotite xenoliths associated with the Phanerozoic Lambert–Amery Rift in eastern Antarctica contain evidence for several stages in the development of the mantle beneath the rift. Despite the fact that equilibria were only partly attained, a combination of petrography, whole-rock geochemistry, mineral chemistry and thermobarometry can be used to decipher four stages prior to entrainment of the xenoliths in the host magma during the initial stages of the breakup of Antarctica, India and Madagascar. The first chronological stage is represented by harzburgitic protoliths represented by rare occurrences of low-Ca olivines and orthopyroxenes in spinel lherzolites: these yield the lowest temperatures of 830–850°C, and are also characterized by distinct trace element contents; lower Ti, Cr, V and Zn in olivine and orthopyroxene, and additionally lower Cu, Ni, Ga and Li in orthopyroxene. Some garnets are subcalcic, indicating that the spinel–garnet lherzolites also formed from harzburgitic protoliths. The second stage is the formation of garnet due to a pressure increase probably related to collision at 1.1 Ga. The third stage is marked by the growth of clinopyroxene, demonstrably in cpx-poor spinel lherzolites but probably in all xenolith groups: equilibrium of clinopyroxene with olivine and orthopyroxene was not attained in all samples, so that the non-judicious use of thermobarometers can produce bewildering results. The fourth stage is an enrichment episode that affected all spinel–garnet peridotites and about half of the spinel peridotites. During this stage, reaction rims were produced on the clinopyroxenes that formed during stage 3, the modal content of olivine and Mg/(Mg+Fe) in the rocks was reduced, CaO, Al2O3 and trace elements were enriched, and garnets were almost completely transformed to kelyphites. A later stage is documented by interstitial glasses and films around spinels related to infiltration of melt from the host magma. These post-date, and are more enriched in alkalies than, partially melted rims on clinopyroxenes, demonstrating that all the three earlier episodes were pre-entrainment events. Pressures indicated by the spinel+garnet lherzolites are restricted to 20–24kbar at 1040–1180°C. Early harzburgitic assemblages are interpreted to represent an earlier, cooler geotherm, whereas the kelyphite assemblages indicate temperatures 180–200°C hotter than the main xenolith geotherm. This event also caused recrystallization of the clinopyroxene rims and is attributed to heating during rifting, but not due to the host magma itself. The preservation of evidence for three progressively hotter geotherms can be related to the upward movement of isotherms during the development of the sub-rift mantle.

To simulate metasomatism at convergent plate boundaries, the process of phengite dehydration, followed by reactions between the expelled fluid and a model mantle material consisting of forsterite + enstatite, was investigated at subduction zone conditions using a piston cylinder apparatus. An experimental technique has been devised in which the fluid, which should be set free during phengite decomposition at about 950°C and 3.5 GPa, would be triggered by a temperature gradient of about 50 K to infiltrate through a perforated Au-foil the hotter part of the capsule, containing the model mantle material. Crystallization of phlogopitic mica and Al-rich enstatite at the boundary phengite - (forsterite + enstatite), representing the „slab - mantle” contact region, indicates successful experimental simulation of mantle wedge metasomatism. Cs-Rb exchange coefficients K D , determined from compositions of newly formed phlogopitic micas, recrystallized phengite and of an amorphous phase, representing the quenched liquid, agree well with recently determined alkali-fractionation data between micas and fluid. The composition of the quenched liquid is rich in alkali, aluminium, silicon, and poor in magnesium, which is in agreement with fluid solubility data determined in earlier experimental studies at similar conditions.

Phlogopitic micas of the solid solution binaries KMg3[AlSi3O10](OH)(2) (phlogopite) - RbMg3[AlSi3O10](OH)(2) (Rb-phlogopite), phlogopite - CsMg3[AlSi3O10](OH)(2) (Cs-phlogopite), and phlogopite - BaMg3[Al2Si2O10](OH)(2) (kinoshitalite) have been synthesized at temperatures of 700 and 800degreesC and pressures of 0.2 and 2.0 GPa. The run products have been investigated by optical microscopy, X-ray powder diffraction, electron microprobe, and infrared spectroscopy. All runs yielded between 81 and 100 wt.% of phlogopitic micas, beside traces of quartz, sanidine, and in one run talc. Celsian and cymrite formed as additional phases in the runs of the (K-Ba)-series. The synthetic phlogopitic micas often consist of mixtures of the three polytypes 1M, 2M(1) and 2M(2), with 1M being the most abundant polytype. Based on electron microprobe analyses, interlayer vacancy concentrations of up to 0.29 (p.f.u.) were determined, indicating a significant talc component within the synthesized phlogopitic micas. In addition to the known characteristic phlogopite OH-stretching vibrational bands, the infrared spectra of the synthetic micas with incompletely filled interlayer sites exhibit a further OH-band, centered in the spectral range 3674 - 3678 cm(-1). The intensity of this band is correlated with the amount of vacancies. The vacancy concentration of phlogopitic micas was determined quantitatively from the intensity of this infrared band by using the intensity of the principal OH-band of synthetic talc (Mg-3[Si4O10](OH)(2)) as a standard. The vacancy concentration of the interlayer site as determined in such-a way by infrared spectroscopy corresponds to those independently derived by electron microprobe analyses.

Mesozoic melilite-bearing ultramafic lamprophyres are developed as sill, dyke and plug-like intrusive bodies in the East Antarctic Beaver Lake area. They consist of varying amounts of olivine, melilite, phlogopite, nepheline, titanomagnetite and perovskite as major phases, accompanied by minor amounts of apatite, carbonate, spinel, glass and, rarely, monticellite. The rocks are mineralogically and geochemically broadly similar to olivine melilitites, differing in higher CO2 and modal phlogopite and carbonate contents. The ultramafic lamprophyres are MgO-rich (13.4–20.5 wt%) and SiO2-poor (32.8–37.2 wt%), indicative of a near-primary nature. Major and trace element features are consistent with minor fractionation of olivine and Cr-spinel from melts originating at depths of 130–140 km.

In this study, phlogopite synthesis experiments were carried out at 800 degreesC and at 1000 and 2000 MPa for low concentrations of Rb (X(Rb)=Rb/(Rb+K)<0.1) in a concentrated (K,Rb)Cl brine+(Rb,K)Cl salt (X(H2O)(bulk)=0.3). For each experiment, solid run products consisted of approximately 70-80% phlogopite (X(Rb)=0.03-0.12) with both a minor celadonite component (approximate to10%) as well as a Cl component in the order of 5-12 M% relative to OH. The remaining 20-30% consisted of approximately equal amounts of newly grown sanidine and quartz. There is no physical evidence that any of these phases are metastable. Rb fractionated strongly into the phlogopite, whereas for the K-feldspar, Rb was almost equally distributed between the feldspar and fluid+salt. At 800 degreesC, the derived distribution coefficients K(D)(Phlg-fluid) (Rb-K) and K(D)(Snd-fluid) (Rb-K) for Rb (Henry's law) are 1.88 +/- 0.12 and 0.86 +/- 0.14, respectively, at 2000 MPa and 1.52 +/- 0.05 and 0.6 +/- 0.4, respectively, at 1000 MPa.Several conclusions may be drawn from these experimental results in conjunction with observations in the field. The first is that phlogopite, with a minor celadonite component and 5-12 M% Cl relative to OH, can form in a relatively dry system consisting of a concentrated (Rb,K)Cl brine+(K,Pb)Cl salt (X(H2O)(bulk) = 0.3) at high temperature and pressures, i.e., 800 degreesC and 1000-2000 MPa. The H(2)O presence of celadonite and Cl(-) apparently stabilises the phlogopite in the presence of quartz at low H(2)O activities as compared to end-member OH-phlogopite which breaks down to enstatite + K-feldspar in the presence of quartz. A second conclusion is that in concentrated KCl brines, Rb is preferentially partitioned into the phlogopite under these P-T and alpha(H2O) conditions. This agrees with similar studies involving 1 and 2 in (K,Rb)Cl salt solutions as well as a simple K-Rb exchange experiments between end-member id OH-phlogopite and a concentrated (K,Rb)Cl brine (X(H2O)(fluid) = 0.6, X(Rb)=0.05) at 800 degreesC and 1000 MPa. (C) 2002 Elsevier Science B.V. All rights reserved.

The distribution of Rb–K and Cs–K between phlogopite and 1- to 2-m aqueous (K,Rb,Cs)–chloride solutions was investigated at 800 °C and pressures of 0.2, 2 and 4 GPa. Phlogopite of the solid solution binaries KAlMg3Si3O10(OH)2 (phlogopite)–RbAlMg3Si3O10(OH)2 (Rb–phlogopite) and KAlMg3Si3O10(OH)2–CsAlMg3Si3O10(OH)2 (Cs–phlogopite) formed within the experiments according to the chemical exchange vectors XIIK−1+XIIRb1+ and XIIK−1+XIICs1+ involving the interlayer sites. The compositions of phlogopite and coexisting fluids were determined from EMP and ICP analyses, respectively. The following K–Rb and K–Cs exchange coefficients KDphl–fluid between phlogopite and fluid were derived by extrapolating the measured Rb and Cs concentrations to the parts per million range, for which Henry's law is valid: KDphl–fluid(Rb–K): 1.71±0.06 at 0.2 GPa, 2.73±0.10 at 2 GPa and 2.76±0.15 at 4 GPa and KDphl–fluid(Cs–K): 0.57±0.05 at 0.2 GPa, 0.73±0.09 at 2 GPa and 0.93±0.26 at 4 GPa. Based on the contrasting and slightly pressure-dependent fractionation behaviour of Rb and Cs relative to K between phlogopite and hydrous fluids—Rb preferentially partitions into phlogopite, whereas Cs fractionates into the fluid—variations in the large ion lithophile element (LILE) ratios within rocks and infiltrated fluids, as a result of formation or breakdown of phlogopite, are discussed by applying models of Rayleigh fractionation and ion exchange processes operating in a one-dimensional chromatographic rock column. Assuming significant amounts of metasomatically formed phlogopite within the mantle wedge, calculations using the chromatographic column model, led to Cs/K ratios significantly lower than those observed for island arc basalts (IAB). We therefore propose that metasomatic formation of phlogopite within the mantle wedge should be limited and does not significantly influence the LILE characteristics of IAB.

In order to constrain the large ion lithophile (LIL) element distribution for subduction-zone environments, exchange coefficients K-D(phe-fluid) = (X-Rb,Cs(phe).X-K(fluid))/(X-K(Phe).X-Rb,Cs(fluid)) for K, Rb, and Cs between aqueous fluids and phengite (phe) have been determined experimentally at 2.0 and 4.0 GPa, Derived KD values for the Rb-K exchange slightly increase from 1.62 +/- 0.10 at 2 GPa, 600 degrees C, to 1.84 +/- 0.15 at 4 GPa, 700 degrees C. For the Cs-K exchange, much lower K, values of 0.22 +/- 0.06 (at 2 GPa, 600 degrees C) and 0.37 +/- 0.10 (at 4 GPa, 700 degrees C) were determined. The results show that, for the pressure-temperature range investigated, Rb preferentially fractionates into phengite, whereas Cs partitions into the fluid. Assuming a one-step model of perfect Rayleigh fractionation For continuous decomposition of phengite during subduction, varying alkali ratios observed for island-are basalts as a function of slab depth may be explained by the LIL-fractionation behavior between fluids and phengites determined in this study, Our data indicate that previously derived models for metasomatic mass transfer during subduction processes need to be reconsidered.

Liquidus phase relations in the system diopside–kalsilite–forsterite–quartz with 3 wt% F were examined at 1 bar and the locations of important invariant points were determined at 18 kbar. At all pressures within this range a large liquidus field for fluorphlogopite (Phl) exists, and has a large influence on both melting and fractionation processes. One eutectic point was found to the silica-rich side of the plane Lc–Fo–Di at Di 1 Ks 30 Fo 2 Qz 67 , where a melt coexists with San, Qz, Phl and Di at 840 °C and 1 bar. Another eutectic point must exist in the silica-poor part of the system because the phase topology determines that thermal barriers must exist. At this point a feldspathoid, either Lc or Ks, must coexist with Fo, Phl and a Ca-bearing phase such as Di. The exact location and phase assemblage were not determined, but the equilibrium melt must have a composition rich in Di (>29 wt%) and extremely poor in Qz (<8 wt%). The composition of the first eutectic moves towards lower SiO 2 contents with increasing pressure (Di 3 Ks 40 Fo 1 Qz 56 at 18 kbar), whereas the second does not exist at 18 kbar due to the disappearance of Lc as a stable liquidus phase. Liquids which coexist with mafic minerals such as En, Fo, Phl and Di are important for the genesis of potassium-rich mafic rocks by partial melting in the mantle and for the early stages of fractional crystallisation. The equilibrium melt at the invariant point Fo + En + Phl + Di + L at 1125 °C is very poor in Fo and Di components at atmospheric pressure (Di 5 Ks 37 Fo 5 Qz 53 ), whereas at 18 kbar the melt contains large amounts of Fo and Di (Di 19 Ks 31 - Fo 28 Qz 21 ), and has a composition close to that of natural lamproites. Kamafugites do not correspond to melts in this system under any of the studied conditions, and appear to require CO 2 in the source. Fractionation processes from primitive potassic basanite melts are controlled principally by the size (and not the mere presence) of the liquidus phase field for phlogopite: at high pressures where the Phl field is large, olivine is eliminated early from the fractionating assemblage and Cpx + Phl fractionation may lead to relatively silica-rich rock differentiates such as trachytes. At low pressures, extensive olivine and restricted Phl crystallisation prevents silica enrichment in the melt, resulting in phonolitic differentiates. Later crystallisation of alkali feldspar accentuates the trends laid down in the early stages of fractionation.

Amphiboles of predominantly tremolite composition have been synthesized at 670-800 degrees C and 200-700 MPa in the system CaO-MgO-SiO2-H2O. The tremolite crystals were large enough to be analyzed by electron microprobe. The run products were also studied by scanning electron microscopy, high resolution electron microscopy, X-ray diffraction (Rietveld analysis) and IR spectroscopy. The electron microprobe analyses resulted in average compositions between tr(87)cum(13) and tr(99)cum(1). The cummingtonite contents determined by electron microprobe varied by up to +/-7 mol% within each run. Investigations using high resolution electron microscopy revealed that the tremolite-cummingtonite solid solutions were highly ordered and chain multiplicity faults were rare. Although the lattice parameters a, b and beta are expected to be a function of the cummingtonite content, no correlations were observed between the compositions derived by electron microprobe and the lattice parameters. The energy and absorbance of the OH stretching vibration is a function of M4 site occupancy. Therefore, the cummingtonite content was determined precisely by IR spectroscopy. At least 3 different bands at distinct energies were identified: 3674.6 cm(-1) for the CaCa-CaCa and 3669.3 and 3672.2 cm(-1) for the MgCa-CaCa and CaCa-MgCa configurations, respectively. The relative integral absorbances showed that all synthesized amphiboles had cummingtonite contents lower than 6 mol%. In some cases nearly pure tremolite was synthesized (tr(99)cum(1)). This amount is considerably less than the 10 mol% cummingtonite content (tr(90)cum(10)) often claimed for synthetic tremolite. The compositions derived by IR spectroscopy correlate with the lattice parameters. Using these correlations the lattice parameters of pure tremolite were extrapolated to be a = 9.8353(18) Angstrom, b = 18.0562(14) Angstrom, c = 5.2768(6) Angstrom, beta = 104.74(2)degrees, V = 906.3(2) Angstrom(3).

The distribution of Rb-Na and Rb-K between richterite and a 2-molal aqueous (Na, K, Rb)-chloride solution has been investigated with hydrothermal experiments at 800∘C and 200 MPa. Experiments were performed as syntheses in which amphiboles grew in the presence of an excess fluid containing the exchangeable cations Na+-Rb+ or Na+-K+-Rb+. The obtained amphiboles were large enough (up to 20 m in width) for reliable EMP analysis. They were chemically homogeneous and HRTEM investigations showed that they were structurally well ordered. The Rb, Na, K, Ca and Mg concentrations in coexisting fluids were measured by ICP-AES. According to the possible incorporation of Na, K and Rb on the A-site, solid solutions in the ternary Na(NaCa) Mg5[Si8O22/(OH)2] (richterite)-K(NaCa)Mg5[Si8O22/(OH)2] (K-richterite)-Rb(NaCa)Mg5[Si8O22/(OH)2] (Rb-richterite) were expected. However, Rb-rich richterites always had significant amounts of A-site vacancy concentrations (X□amph=□ A /(RbA+KA +NaA+□A) of up to 0.42 in the K-free (Na,Rb)-richterites and of up to 0.67 in the (Na, K, Rb)-richterites which corresponds to the same content of tremolite+cummingtonite-component. Amphiboles containing practically only Rb besides vacancies and no Na and/or K on the A-site were also synthesized, however. The Rb-Na and Rb-K exchange coefficients between fluid and richterites are similar. Rubidium always fractionated strongly into the fluid phase. For low Rb-concentrations in richterite (XRbamph<0.1) a linear correlation between XRbfluid and XRbamph exists. In this concentration range, the derived exchange coefficients KD(Rb−K)amph−fluid and KD(Rb−Na)amph−fluid were 0.08 ± 0.04 and 0.04 ± 0.02, respectively. These low exchange coefficients show that significant amounts of Rb in amphiboles require a Rb-rich fluid phase. The results indicate that K-Rb fractionation between alkali amphiboles and fluids is significantly different from K-Rb fractionation between alkali feldspar/ phlogopite and fluid, with KDs of about 0.5 and 1.2, respectively. Formation of richterites will drastically alter the K/Rb-ratios of fluids or melts. These results may have important implications for the genetical interpretation of various geological settings, e.g., MARID-type rocks.

The liquidus surface of the system quartz-kaisilite-forsterite (SiO2-KAL-SiO4-Mg2SiO4) with 3 and 6 wt % F has been investigated at 1 arm pressure, and the locations of the most important invariant points have been determined at 10 and 18 kbar. Two different fluorine contents, 3 and 6 wt %, were used for this study and fluorine was added by direct exchange for oxygen (F2O-1). Melting relations in this system differ greatly from those of the F-free system due to the presence of a large liquidus field for fluorphlogopite (Phl), the size of which correlates with the amount of F in the system. The locations of most invariant points are very similar for these different F-contents except for the reaction points Fo + L <----> En + Phl and En + L <----> Qz + Phl. For both F-contents there are two eutectic points: one on the silica-side at a melt composition of Qz(67)Ks(32)Fo(1) coexists with San, Qz and Phl at 850 degrees C, and one on the kalsilite-side of the join San-Fo, where a melt of composition Qz(12)Ks(75)Fo(13) coexists with Ks, Lc and Phl at 1225 degrees C. The composition of the first eutectic point moves towards lower SiO2- contents at increasing pressures (Qz(59)Ks(75)Fo(1) at 18 kbar), whereas the second disappears at 10 kbar for 6 wt % F and 18 kbar for 3 wt % F due to the absence of Lc as a stable liquidus phase at high pressures. We observe a new eutectic point San + Ks + Phl + L (Qz(31)Ks(66)Fo(3) at 10 kbar) for which the location is determined approximately at 18 kbar. The eutectic point San + Qz + Phl + L and the reaction point Qz + Phl <----> En + L (for which equilibrium temperatures lie between 1150 degrees C at atmospheric pressure and 1250 degrees C at 18 kbar) represent model assemblages for melting processes in crustal rocks. The location of the reaction point Qz + Phl <----> En + L varies from Qz(72)Ks(26)F(o2) at 1 bar through Qz(68)Ks(28)Fo(4) at 10kbar to Qz(67)Ks(30)Fo(3) at 18 kbar; with 3 wt % F, and from Qz(76)Ks(20)Fo(4) at 1 bar through Qz(71)Ks(26)Fo(3) at 10kbar to Qz(67)Ks(30)Fo(3) at 18 kbar with 6 wt % F.Liquids which coexist with mafic minerals such as En, Fo and Phi are important for understanding the genesis of potassium rich rocks by partial melting of mantle rocks and their subsequent evolution by fractionation. The peritectic point Fo + En + Phl + L shows large variations in equilibrium liquid composition and temperature for the two different F-contents and pressures (Qz(19)Ks(37)Fo(14) at 1250 degrees C and atmospheric pressure, Qz(18)Ks(46)Fo(26) at 1390 degrees C and 18 kbar for 3 wt % F; Qz(42)Ks(19)Fo(39) al 1295 degrees C and atmospheric pressure, Qz(22)Ks(30)Fo(48) at > 1390 degrees C at 18 kbar for 6 wt % F). Melts produced at this point have lower SiO2-contents and evolve to more SiO2-poor compositions by fractional crystallization than in the H2O-bearing system at comparable pressures. F-Phl varies considerably in chemical composition and the Si-content of Phl correlates positively with the Si-content of the melt.