Mineral analysis and mineral formulae recalculation

Mineral analyses and elemental X-ray maps were performed with a JEOL-Superprobe JXA-8900M microprobe equipped with five spectrometers at the ICTS-National Electronic Microscopy Centre at the Complutense University of Madrid (Spain; The operating parameters for punctual analyses were: 15kV accelerating voltage, 20 nA beam current, between 1 and 5 µm beam diameter (1 µm for the microinclusions) and 10 s counting time on peak for each element. X-ray maps were operated at 20 kV and 150 nA. Representative analyses of selected minerals are listed in Tables 2-5. Mineral formulae have been calculated using the software AX (http:/ The amount of ferric iron was calculated from stoichiometric constraints. These stoichiometric constraints for the specific minerals are those performed by the software AX, which executes standard mineral recalculations, with attempts at ferric iron estimation. For example, for amphiboles it uses the method of Holland and Blundy (1993). Mineral abbreviations are those used by THERMOCALC (Holland and Powell, 1998): ab-albite, act-actinolite, bi-biotite, chl-chlorite, ep-epidote, g-garnet, gl-glaucophane, hb-hornblende, hem-hematite, ilm-ilmenite, jd-jadeite, law-lawsonite, mt-magnetite, mu-muscovite, o-omphacite, pa-paragonite, pl-plagioclase, q-quartz, ru-rutile and sph-titanite (sphene). Other abbreviations: bar-barroisite, carb-carbonates, sul-sulphides, win-winchite. Compositional variables: XFe = Fe2+/(Fe2++Mg); XFe3+(ep) = Fe3+/(Fe3++Al–2); XNa(mu, pa) = Na/(Na+K); amphiboles: y = XAlM2; z = XNaM4; a = XNaA; c = XCaM4; f = XFe3+M2; garnet: Alm = Fe/(Fe+Mg+Ca+Mn), Prp = Mg/(Fe+Mg+Ca+Mn), Grs = Ca/(Fe+Mg+Ca+Mn), Sps = Mn/(Fe+Mg+Ca+Mn); feldspars: Ab = Na/(Na+Ca+K), An = Ca/(Na+Ca+K), Or = K/(Na+Ca+K); Other symbols: pfu – per formula unit; wt. % – weight percent; “” denotes core-to-rim evolution. “/” indicates partial replacement; “*” specifies phases texturally inferred from petrography or from petrologic modelling that have not been identified petrographically. “#” indicates phases described by other authors that have not been recognized in this study.

40Ar/39Ar Geochronology

Mineral characterization and sample preparation

All sample preparation was performed at the Complutense University of Madrid. Samples were crushed and sieved, and single pristine grains of muscovite (samples MT1 and LM) and hornblende (sample CA) were separated using conventional magnetic and gravimetric methods, followed by hand-picking using a binocular microscope.

The following K-bearing minerals were separated from the samples: 1) lens-shaped crystals of muscovite in the matrix of sample MT1, which range in size from 100 to 1000 μm and have 3.25–3.35 Si pfu and XK =0.94–0.98 (Table 6). The selected fraction consists of the crystals in the interval 100-200 μm because they typically have the highest K2O contents. 2) Crystals of unzoned phengitic muscovite (300–500 μm) from the S2 matrix foliation of sample LM, which have Si ~ 3.50 pfu and XK =0.93–0.99, typically close to 1 (Table 6). It was possible to separate individual unzoned grains larger than 1500 μm. A chemically homogeneous population of S2-micas in the matrix can be distinguished from the S3-micas because the latter usually appear in mixed grains together with paragonite, biotite and chlorite and have slightly lower Si content (Si = 3.4-3.45 pfu). In sample CA, hornblende crystals (Si= 7.036.96 pfu; XFe = 0.380.41; z = 0.140.20; c = 0.800.74; f = 0.160.18) from the matrix foliation in the grain-size fraction between 200 and 300 μm, were selected (Table 2).

Analytical techniques

All analyses were carried out in the 40Ar-39Ar Geochronology Research Laboratory of Queen's University (Kingston, Canada). All samples were ultrasonically rinsed several times in distilled water, wrapped in pure aluminium foil, and stacked vertically in an Al canister, which was then irradiated at the McMaster University Nuclear Reactor in Hamilton, Canada with the 40Ar-39Ar flux monitor - Hb3gr hornblende [1072 ± 11 Ma (2σ)] (Roddick, 1983). Following irradiation, the samples and monitors were placed in small pits, ~2 mm in diameter, drilled in a Cu sample holder. The number of grains picked per fraction was variable. Particularly, for the analysed samples, inside each pit a total of four grains were selected for dating the muscovite concentrates, from sample LM, and 6 grains for those in sample MT1. This was placed inside a small, bakeable, stainless steel chamber with a ZnSe viewport connected to an ultra-high vacuum purification system. Monitors were fused in a single step, using a focused New Wave MIR-10 30-watt CO2 laser.

For the step-heating experiments, the laser beam was defocused to heat the entire sample until fusion as a glass bead in the final step. Samples were heated for ~3 minutes with increasing power increments. The evolved gases were purified using a SAES C50 getter for ~5 minutes. Argon isotopes were measured using a MAP 216 mass spectrometer, with a Bäur Signer source and an electron multiplier. All data were corrected for blanks, atmospheric contamination, and neutron-induced interferences (Onstott and Peacock, 1987; Roddick, 1983). All errors are reported as ±2σ, unless otherwise noted, and dates were calculated using the decay constants recommended by Steiger and Jager (1977).