⁴⁰Ar-³⁹Ar DATA

Volcanic rock grains with a brownish to blackish appearance and rounded shapes were carefully concentrated by hand-picking under a binocular microscope, cleaned ultrasonically, and dried with pure acetone. The concentrates were wrapped in aluminium foil and irradiated in the TRIGA reactor at the University of Pavia (Italy) for 8 hr along with the biotite standard FCT-3 with an age of 27.95 Ma (Baksi et al., 1996). Values of the irradiation parameter J for individual sample packages were calculated by parabolic interpolation between the analyzed standards. Estimated uncertainty (one standard deviation) is 0.3%. Sample separation and argon step-heating experiments were carried out at the IGGI-CNR laboratory (Pisa). Because of the small grain size of volcanic fragments, they were not suitable for single-grain total fusion analyses. After irradiation, milligram-sized fractions (only for Samples 178-1097A-27R-1, 35-58 cm, and 178-1103A-36R-3, 4-8 cm) were spread on the bottom of 7-mm-diameter holes of a copper holder placed into a ultra high vacuum laser port and baked overnight at ~200°C. The samples were incrementally heated by a multimode laser beam generated by a continuous Nd:YAG (Nd-doped yttrium-aluminum-garnet) laser defocused to ~3-mm spot, and homogeneous heating was obtained by slowly rastering the laser beam (at 0.2 mm/s) by a computer-controlled x-y stage. In addition, with the aim of resolving as much as possible the true sample heterogeneity, various small fractions from the same irradiation packages of all three samples, each consisting of 10 grains for Sample 178-1097A-27R-1, 35-58 cm, and 30 grains for Samples 178-1103A-31R-2, 0-4 cm, and 36R-3, 4-8 cm, were incrementally laser heated over three steps. Gases extracted, gettered for 15 min (including about 5 min of lasering) for the milligram-sized samples and 10 min for the small fractions (including 1 min of lasering), were equilibrated via automated valves into a MAP215-50 noble gas mass spectrometer fitted with a Balzers SEV217 secondary electron multiplier. Argon isotope peak intensities were measured 10 times for a total time of ~20 min. Blanks were analyzed after every two to four runs. Data corrected for postirradiation decay, mass discrimination effects, isotopes derived from interfering neutron reactions, and blanks are listed in Tables T2 and T3. Errors are 2 and do not include the uncertainty in the J value that was included in the total fusion ages. However, for analyses on small fractions the age errors are dominated by the experimental uncertainty. The most relevant correction factors used were as follows: ⁴⁰Ar/³⁹Ar_(K) = 0.0096, ³⁶Ar/³⁷Ar_(Ca) = 0.00024, and ³⁹Ar/³⁷Ar_(Ca) = 0.00075. Raw data reduction and age calculations were made using the program ArArCALC (v.2.0) (see Koppers, 1998).

Figure F6 shows the age spectra and Ca/K and Cl/K variations for the milligram-sized fractions of Samples 178-1097A-27R-1, 35-58 cm, and 178-1103A-36R-3, 4-8 cm; data are listed in Table T2. Both samples display irregularly discordant apparent age profiles. Sample 178-1097A-27R-1, 35-58 cm, after a first step at ~110 Ma, shows high apparent ages up to ~200 Ma that progressively decline to a minimum segment at ~100 Ma, with a slightly older final step (114 Ma) (Table T2). The total fusion age is 122.1 ± 1.0 Ma. Sample 178-1103A-36R-3, 4-8 cm, exhibits a different apparent age pattern and significantly younger dates, characterized by high age in the first step (66 Ma) followed by a steady increase from 16 to 27 Ma. The total fusion age is 22.15 ± 0.24 Ma. For both samples, the Ca/K ratio displays a monotonic increase starting from values significantly lower than the bulk data in the first steps (Table T2; Fig. F6). The Ca/K ratios largely overlap in the two samples but, in Sample 178-1097A-27R-1, 35-58 cm, we observe a larger range (0.27-7.3 against 1.3-6.1) and significantly lower values in the low temperature steps. In Sample 178-1103A-36R-3, 4-8 cm, the Cl/K ratio shows a steady decline from 0.086 to 0.038 with increasing temperature, whereas in Sample 178-1097A-27R-1, 35-58 cm, it displays an irregular pattern in the first steps followed by a steady increase from the 0.35-W step. It is worthy of note that Sample 178-1103A-36R-3, 4-8 cm, Cl/K ratios show narrow variation and significantly higher values (~10 times higher) when compared to Sample 178-1097A-27R-1, 35-58 cm.

Figure F7 illustrates the age spectra and Ca/K variations of the step-heating experiments on the small fractions, and data are listed in Table T3. Most of the spectra display discordant apparent ages for the three steps. The use of small fractions strongly enhances the heterogeneity in age as observed in the milligram-sized fractions. Both samples exhibit a younger lower limit: 75 and 7.6 Ma for Samples 178-1097A-27R-1, 35-58 cm, and 178-1103A-36R-3, 4-8 cm, respectively. By contrast, the Ca/K ratios largely overlap with the ranges obtained from incremental heating of the milligram-sized fractions. Sample 178-1103A-31R-2, 0-4 cm, which was only analyzed as small fractions, shows a younger apparent age range (18-57 Ma) than the sample from Site 1097 (75-173 Ma) but older than the other sample from Site 1103 (7.6-50 Ma). The Cl/K ratios are higher than the sample from Site 1097 but similar to those of Sample 178-1103A-36R-3, 4-8 cm. It is to be noted that in Sample 178-1097A-27R-1, 35-58 cm, all the small fractions analyzed are characterized by marked increase in the Ca/K ratios from the first (with Ca/K << 1) to the last step. By contrast, both samples from Site 1103 commonly display a narrower variation (starting from values generally >>1). Note that in most instances the second and the third steps of both samples from Site 1103 have Ca/K ratios that overlap within errors (Fig. F7).