Jean-Marc Montel

A model for monazite/melt equilibrium and application to the generation of granitic magmas

Abstract In order to model the behaviour of Zr, Hf, Ti, P, U, Th, Y and the rare-earth elements (REE) in the continental crust, the parameters which control the equilibria between accessory minerals and granitic melts must be determined. This paper proposes equations for the equilibrium between monazite and Ca-poor felsic melts. In a first step, all the light REE are considered as a single component and an equation describing monazite solubility is deduced from the available experimental data. In a second step, the fractionation of REE by monazite is investigated on the basis of a natural example of coexisting monazite and obsidian from Macusani, Peru. The solubility of monazite can be used as a thermometer, based on the total REE content of natural magmatic rocks. Application of REE thermometry to various peraluminous rock series suggests that it is a reliable tool for determining temperatures of magmatic processes. The equation governing REE fractionation by monazite is used to discuss genetic relationships between a melt and a presumed parental rock, using the composition of the monazite in the parental rocks and the REE pattern of the daughter rock. This method has been applied successfully to High Himalayan Manaslu and Tibetan Slab granites. It is suggested that detailed petrographical work on monazite including electron microprobe analyses and SEM imaging of its internal structures, combined with the model presented here, can be a useful tool for deciphering the genesis and the differentiation of peraluminous granitic magmas in the continental crust.

Experimental resetting of the U–Th–Pb systems in monazite

Abraded fragments (200–400 µm) of a large, chemically homogeneous, and non-metamict Brazilian monazite crystal, characterised by a concordant U–Pb ages of 474 +/- 1 Ma (208Pb/206Pb = 19.5), were hydrothermally treated at varying temperatures with solutions of different compositions. Product monazites were analysed with Scanning Electron Microscope (SEM), Electron Microprobe (EMP), Secondary Ion Mass Spectrometer (SIMS) and Isotope Dilution–Thermal Ionisation Mass Spectrometer (ID-TIMS). Experiments with pure water over a temperature range of 800–1200°C, at 700 MPa and durations ranging from 5 to 60 days showed that even at 1200°C any dissolution and recrystallization of new monazite is confined to the outermost surface of the grain. Neither Pb diffusion at the EMP scale, nor significant discordancy were observed. We performed experiments at 800 and 1000°C for different durations using different fluid compositions at quartz saturation: a 10 wt.% CaCl2 fluid, a 10 wt.% SrCl2 fluid, a 10 wt.% NaCl fluid and a fluid containing NBS 982 Pb standard (208Pb/206Pb = 1). For all runs, EMP traverses revealed no Pb-diffusion profiles. Significant overgrowths of newly formed monazite are documented by SEM analyses. They occurred only in the 1000°C experiments when either CaCl2 or Pb-bearing fluids were present. In the CaCl2 experiment, two zones could be distinguished within the crystal: a core possessing the initial monazite composition and a rim consisting of newly formed monazite produced by dissolution/precipitation, which was enriched in Ca and Pb-free. ID-TIMS dating of single grains treated with SrCl2 and CaCl2 solutions at 1000°C are significantly discordant. Experiments employing the NBS Pb-standard produced sub-concordant monazite, for which the 207Pb/206Pb apparent age has become older than prior to the experiment (544 Ma at 800°C and 495 Ma at 1000°C). The newly grown monazite rim had obviously incorporated Pb from the fluid. None of our reaction products contained a detectable diffusion profile. The only resetting mechanism we detected involved dissolution/precipitation. The extent of the dissolution/precipitation process depends on fluid composition and is a more efficient mechanism than diffusion for controlling the resetting of monazite in natural rocks.

Electron microprobe dating of monazites from high-grade gneisses and pegmatites of the Kerala Khondalite Belt, southern India

Monazites of five samples (one leptynitic garnet–biotite gneiss, one khondalite, one augen gneiss and two pegmatites) from the central and northern part (Ponmudi Unit) of the Kerala Khondalite Belt (KKB) in southern India were analyzed with the electron microprobe dating technique. Monazites in the augen gneiss and the pegmatites yield grain sizes between 200–800 μm, Th abundances are rather low (<10 wt.%) and the distribution of Th, U and Pb within single grains is fairly homogeneous. Contrasting to this, monazites in the leptynitic gneiss and the khondalite are small (<150 μm). They often display very complex Th–U–Pb patterns and contain high Th concentrations up to 20 wt.%. The statistical treatment of individual ages from the investigated samples revealed three populations of Lower Proterozoic (∼1.9 Ga), Upper Proterozoic (∼580 Ma) and Ordovician age (∼470 Ma) as well as Mid Proterozoic ages between 0.8–1.7 Ga which are not regarded to be of geological significance. Lower Proterozoic ages are preserved in the cores of monazites from leptynitic gneisses and khondalites. They fairly agree with Sm–Nd model ages for similar rocks of the KKB and give a minimum age for first monazite growth or complete homogenization. The prominent Pan-African population with mean values between 540 and 580 Ma is present in the leptynitic gneiss, the khondalite and the augen gneiss and in line with published isotope ages for the KKB. The Ordovician population finally marks the emplacement of granitic pegmatites subsequent to the Pan-African high-grade metamorphic event. There is an obvious discrepancy between khondalites and leptynitic gneisses on the one hand and augen gneisses on the other concerning the presence of Lower Proterozoic ages. While these are abundant in the former, often rimmed by Upper Proterozoic ages, they are completely absent in the latter. It appears unlikely that Lower Proterozoic ages were completely reset during a Pan-African event exclusively in the augen gneisses while they were preserved in leptynitic gneisses and khondalites. It is further concluded that the augen gneisses are of magmatic origin and were derived from porphyritic granites. Thus, the Upper Proterozoic age of 605±37 Ma calculated for the investigated sample approximates the time of magma emplacement, which slightly precedes the peak stage of Pan-African high-grade metamorphism in the KKB, and of monazite crystallization from the granitic melt. A characteristic feature of the investigated monazites is the resetting of Lower Proterozoic and Pan-African ages to significantly younger values due to partial Pb loss. Monazites not shielded by other minerals (e.g., garnet) suffered selective mobilization of Pb along fractures or at their rims while Th and U concentrations remained almost unchanged. The results presented in this study indicate that this was mainly due to fluid-rock interaction. It is concluded that thermal diffusion of Pb even at the suggested temperatures of 900°C only had minor influence on the Th–U–Pb composition in monazite and that the closure temperature for this system must be significantly higher than previously assumed (∼750°C).

Apatite solubility in peraluminous liquids: Experimental data and an extension of the Harrison-Watson model

Abstract New experimental data at 750–1000°C, 2–5 kbar, and for peraluminous melt compositions show a dramatic increase of apatite solubilities and phosphorus melt contents when compared to metaluminous or peralkaline compositions ( Harrison and Watson , 1984). The formation of alumino-phosphate units in peraluminous melts qualitatively explains our data. A new apatite solubility model, incorporating the data for the peraluminous region, is presented. Application of this model to examples of peraluminous felsic suites is discussed.

X-ray diffraction study of brabantite-monazite solid solutions

Abstract We synthesized compounds with stoichiometry of natural brabantite M2+Th(PO4)2 with M2+=Ca, Cd, Sr, Pb, Ba at 1 bar, 1200 °C and 2.5 kbar, 700 °C. Those compounds were studied by powder X-ray diffraction and electron microprobe. For Ca, Sr, and Pb, we obtained always crystals with the monazite structure. For Cd, the monazite structure is obtained at 1 bar, but not at 2.5 kbar. For Ba, we obtained the monazite structure only at 2.5 kbar. The unit-cell parameters of the compounds with monazite structure vary regularly with the size of the M2+ ion. We studied also the solid-solution LaPO4–M2+Th(PO4)2, at 1200 °C, 1 bar. The solid solutions for Cd, Ca, and Sr are continuous, and the unit-cell parameters vary linearly with the degree of substitution. For Pb, the solid solution is continuous, but the unit-cell parameters do not vary linearly. For Ba, there is a miscibility gap, with a maximum of about 50 mol% of BaTh(PO4)2 in LaPO4. These results have several consequences for U–Th–Pb geochronology of monazite, and for utilization of monazite as a nuclear-waste ceramic.

Phosphates and Nuclear Waste Storage

A significant effort has been made by the scientific community to evaluate the potential of phosphate minerals and glasses as nuclear waste storage hosts. Radioactive waste-bearing phosphates, including monazites, apatites, and glasses, can be readily synthesized in the laboratory. Because of their low solubilities and slow dissolution rates, these phosphates are more resistant to corrosion by geological fluids than many other potential nuclear waste storage hosts, including borosilicate glass. Phosphates are, however, not currently being used for nuclear waste storage, in part because their synthesis at the industrial scale is relatively labor intensive, often requiring the separation of the waste into distinct fractions of elements. Such limitations may be overcome by adding phosphate amendments to backfill material, which could provoke the precipitation of stable radiactive waste-bearing phosphate minerals in situ.

Experimental determination of Thorium partitioning between monazite and xenotime using analytical electron microscopy and X-ray diffraction Rietveld analysis

The Thorium distribution between monazite and xenotime has been determined experimentally using the coupled substitution Th + Si=REE + P. Experiments have been conducted in standard cold seal hydrothermal and internally heated pressure vessels at 200MPa in the range of 600-1100°C. Starting mixtures were prepared from gels composed of equal amounts of CePO4 and YPO4 with addition of 10, 20 and 50 mol%ThSiO4. The grain sizes of the run products were in the range of a few microns. Analytical electron microscopy (AEM) methods were applied to obtain reliable chemical compositions of the reaction products. Lattice parameters of run products were determined using Rietveld analysis. For runs with 10 and 20 mol% ThSiO4 component in the bulk the ThSiO4 component distributes almost exclusively into monazite at all temperatures. The amount of the YPO4 component in monazite increases relative to the Th-free system if significant amounts of ThSiO4 are present within the structure. ThSiO4 favours incorporation of YPO4 resulting in a shift of the monazite limb and the shrinkage of the monazite-xenotime miscibility gap in the CePO4-YPO4-ThSiO4 ternary diagram. Thermometric calculations based on monazite-xenotime equilibria must be corrected for this effect. For runs with 50 mol% ThSiO4 in the bulk, thorite formed as an additional phase at 600 to 900°C but was absent at higher temperatures. At high Xbulk ThSiO4 and low T, the system is three-phase. The three-phase stability field strongly shrinks with increasing temperature. A tentative phase diagram of the ternary system CePO4-YPO4-ThSiO4 is proposed and discussed in the light of monazite-xenotime-thorite-bearing assemblages in natural rocks.

Importance of late-magmatic and hydrothermal fluids on the Sm–Nd isotope mineral systematics of hypersolvus granites

Abstract Despite the significant impact of the Sm–Nd radiometric system in various fields of isotope geochemistry, very few attempts have been made to use it at the mineral scale in granite studies. This approach has been used on two granites from the anorogenic province of Corsica (SE France). Two statistically acceptable alignments were obtained in an isochron diagram for the hypersolvus fayalite-bearing granite of Mantelluccio. Alkali–feldspar, amphibole, inclusion-free zircon, fluorite and fergusonite give an age of 331±10 Ma, whereas allanite, whole-rock, amphibole and inclusion-free zircon yield a significantly younger age at 291±13 Ma. The latter is validated by a zircon U–Pb age of 283±1 Ma as the magmatic emplacement date of the granite. The other alignment, giving a 50-Ma older age, can be interpreted either as being inherited or, more likely, as the result of hydrothermal convection interacting with the country rocks at the end of granite crystallization. The Sm–Nd systematics of epidote records a second fluid–rock event, which probably occurred at least 200 Ma after the emplacement of the granite. Extremely low 143 Nd/ 144 Nd ratio indicates that the epidote crystallized from a fluid that carried lanthanides in the crust over a distance of at least 1 km. Zircons with inclusions have a lower initial 143 Nd/ 144 Nd ratio than inclusion-free zircons. This feature is attributed to apatite and/or zircon core inheritance, through contamination, during the ascent of the granitic melt in the crust. Two isochrons have been also obtained for the hypersolvus peralkaline granite of Evisa. Amphibole, fluorite, sphene and fergusonite yield an age of 259±6 Ma, whereas the whole-rock, alkali–feldspar and metamict zircon give an age at 209±14 Ma. In this case, the former value is roughly in agreement with a possibly slightly rejuvenated whole-rock Rb–Sr date of 249±3 Ma, suggesting that it documents the magmatic crystallization age of the granite. Sm–Nd isotope systematics of the alkali–feldspar, metamict zircon and possibly allanite, were re-equilibrated during an important hydrothermal event that affected Western Europe in Early Jurassic times. In both examples, however, it appears that the whole-rock Nd isotopic signatures were not strongly affected by the various fluid–rock events, indicating that they can still be used for petrogenetic investigations. In contrast, these results show that, like K–Ar and Rb–Sr, the Sm–Nd systematics of minerals with weak resistance to post-magmatic processes can be used to trace and date hydrothermal events.

SOLUTION MECHANISMS OF PHOSPHORUS IN QUENCHED HYDROUS AND ANHYDROUS GRANITIC GLASS AS A FUNCTION OF PERALUMINOSITY

Solution mechanisms of P in metaluminous to peraluminous quenched, hydrous (≈9 wt% H2O) and anhydrous glasses in the system CaONa2OK2OAl2O3SiO2P2O5 have been examined with microRaman spectroscopy. The principal aim was to examine relative stability of phosphate complexes as a function of bulk chemical composition. Increasing peraluminosity was accomplished by increasing Al3+ and Ca2+ proportions with constant SiO2 content. The molar ratio Al2O3/(CaO + Na2O + K2O) (A/ CNK) ranged from 1 (metaluminous) to ∼1.3 (peraluminous). In all compositions P5+ is bonded to Al3+ to form AIPO4 complexes. The principal solution mechanism is one where depolymerized species (Q3), involving Al3+ both within and outside the aluminosilicate network, interact with P to form the AIPO4 complex together with Q4 species. The mechanism does not involve alkali metals or alkaline earths. In anhydrous compositions, the spectra are interpreted to suggest SiOP cross-linking in the structure. In hydrous compositions, evidence for SiOP bonding is less evident. In such glasses, there is, however, possible spectroscopic evidence for SiOH bonding and possibly POH bonding resulting from breakage of cross-linking SiOP bonds existing in the anhydrous glasses. Therefore, the water content of peraluminous aluminosilicate melts is likely to affect the solubility behavior of P, and conversely, the solubility behavior of H2O is affected by P in such melts.

Nano-petrographic investigation of a mafic xenolith (maar de Beaunit, Massif Central, France)

The thermal history of a mafic xenolith from the Beaunit maar (Massif Central, France) is reconstructedatthe basisofa transmissionelectronmicroscopystudy. The protolith isa meta-microgabbro(opx- 1,cpx-1, pl-1)sampledinthelowercontinentalcrust(T=870-970°C,P 0.7-0.8 GPa).The incorporation inthe basalticmagmaproducedfivereactionsaround orthopyroxene: opx-1 ficpx-2fi cpx-3(augite+highpigeonite) fi liq fi cpx-4. The final reaction is the transformation of residual cpx-3 (augite + high pigeonite) into cpx-5 (augite+lowpigeonite).The calculationofthetimerequiredforeachtransformationyieldsaminimumresidence time of the enclave in the host magma of 16 hours and a magma ascent velocity of 1.8 km.h -1 . Exsolutions are produced by pressure decreaseas the xenolith is brought up to the surfacein the host basalt. Fracturesobserved in primary minerals are interpreted as a consequence of xenolith shocks against the wall of the magma conduit.