Appendix 2 - Analytical Methods

Appendix 2 - Analytical Methods

Geochronology

Samples were examined in thin section and then crushed in an effort to recover feldspar (particularly sanidine) crystals for analysis. Individual feldspar crystals were then handpicked, avoiding altered or cloudy grains.

The whole rock samples were crushed, sieved (250– 500 μm fractions) and washed in dilute nitric acid, de-ionised water and acetone, before final handpicking to exclude altered and xenocrystic material. The feldspar and whole rock samples were packaged in separate irradiation batches (UM#16; UM#24/25) and irradiated in position 5C (cadmium-lined) of the McMaster University reactor, Hamilton, Ontario, Canada. 40Ar/39Ar analyses were conducted at the University of Melbourne Noble Gas Geochronology and Geochemistry Laboratory, using analytical procedures analogous to those described by Matchan and Phillips (2011)andPhillips et al. (2007). Step-heating analyses of feldspar and whole rock samples were carried out using a tantalum furnace connected to a VG3600 mass spectrometer equipped with a Daly detector. The neutron fluence monitors used were GA1550 biotite (UM#16; 98.8 ± 0.5 Ma, 1σ; Renne et al., 1998), which yielded J-values (±1σ) ranging from 0.000478 ± 0.000002 (0.465%) to 0.000482 ± 0.000003 (0.6%) and Fish Canyon Tuff (FCT) sanidine (UM#24/25; 28.02 ± 0.14 Ma, 1σ; Renne et al., 1998), which gave J-values of 0.000703 ± 0.000001 (0.17%) to 0.000721 ± 0.000003 (0.4%). Mass discrimination (UM#16: 1.0053 ± 0.11%; UM#24: 1.0075 ± 0.20%; UM#25: 1.0081 ± 0.05%; per atomic mass unit) was monitored by measuring air aliquots from a Doerflinger pipette. Correction factors (±1σ) for interfering isotopes were (39Ar/37Ar)Ca = 0.000679 ± 0.000005 (0.7%), (36Ar/37Ar)Ca = 0.000270 ± 0.000005 (1.7%), (40Ar/39Ar)K = 0.0005 ± 0.0002 (39%) and (38Ar/39Ar)K = 0.01232 ± 0.00002 (0.2%) for UM#16; and (39Ar/37Ar)Ca = 0.000680 ± 0.000020 (2.9%), (36Ar/37Ar)Ca = 0.000289 ± 0.000005 (1.7%), (40Ar/39Ar)K = 0.0004 ± 0.0004 (100%) and (38Ar/39Ar)K = 0.013 ± 0.005 (39%) for UM#24/25. Plateau ages are defined as including at least 50% of 39Ar, distributed over a minimum of 3 consecutive steps, coincident at the 95% confidence level (Fig. 9). Plateau ages are given at the 95% confidence level and calculated using the mean of all plateau steps, each weighted by the inverse variance of their individual analytical error. Calculated uncertainties associated with mean and plateau ages include errors in the J-values, but exclude errors associated with the age of the monitor and the decay constant (internal errors only; see discussion by Min et al., 2000). Decay constants and isotopic ratios are those of Steiger and Jager (1977). Note that new 40Ar/39Ar decay constants and standard ages have recently been suggested by Renne et al. (2010). However, as these values have yet to be sanctioned by the Subcommission of Geochronology, the values of Steiger and Jager (1977) are retained in this study. Note however, that use of the Renne et al. (2010) values increases the calculated ages by only 1%, which is well within the quoted uncertainties for the current analyses.

Whole-Rock Geochemistry

Rock samples were carefully selected, cleaned in distilled water and all enclaves were removed during preliminary crushing operations. Rock chips were then powdered in an agate mill and analysed at Activation Laboratories Ontario, for major, minor, trace and REE elements (code WRA+TRACE 4Lithoresearch) by ICP-OES and ICP-MS methodologies. For major elements the precision is estimated better than 2% for values higher than 5 wt% and better than 5% in the range 0.1 – 5 wt% (Table 2). Mineral phases were analysed using polarized microscopy on a Nikon Eclipse E600 POL instrument and by combined WDS EDS techniques using a CAMECA SX50 electron microprobe at the “Centro di Studi per il Quaternario e l’Evoluzione ambientale – CNR” (Rome). Minerals were analysed using a 15 KeV accelerating voltage, 20nA beam current, 5µm beam size (up to 8µm for albitic feldspars) and counting times of 100s. Data were reduced using the ZAF4/FLS software by Link Analytical. Representative analyses are shown in Table 3a-e. Structural formulae for feldspars, clinopyroxenes, and olivines were calculated using CALCMIN software (Brandelik, 2009). For amphiboles the structural formulae and a preliminary IMA-04 classification were determined using WINAMPHCAL software (Yavuz, 2007).