Tephra layers were selected for analysis on the basis of shipboard petrographic studies of smear slides, with emphasis placed on unaltered, biotite- and/or feldspar-bearing layers. Most of the layers are only partly consolidated and were easily disaggregated in an ultrasonic bath. The carbonate fraction was removed with mild acetic acid. The remaining glass, smectite, and phenocryst residue was wet-sieved. Under the stereo microscope, biotite phenocrysts were often conspicuous in the residue and were generally hand picked from the >63- or >125-µm size fractions (Fig. 3). Some biotites show signs of minor alteration, with development of chlorite at the edges of grains. Such grains were discarded whenever possible. Feldspar phenocrysts are smaller and often difficult to distinguish from glass shards. In this case, feldspars were separated from glass shards using a lithium polytungstate heavy liquid.
The freshest, unfractured single grains were handpicked, packed in aluminum foil and sent for irradiation at either the Risö National Laboratory in Denmark or the research reactor at McMaster University, Canada. The age determinations were carried out at the Ar-Ar laboratory at the Earth Sciences Department, Open University, England. The laboratory is equipped with a Mass Analyzer Products 215-50 noble gas mass spectrometer with an automated extraction line. The system is equipped with a Johnston multiplier detector, dedicated to laser Ar-Ar dating, using a continuous infra-red (Nd-YAG) laser focused through a customized Leica Metallux 3 microscope. It is designed to measure small samples, either as single grains heated for 10-300 s or in situ in thick rock sections using millisecond pulses (Kelley, 1995). The laser is capable of delivering up to 70 W using the wider multimode beam, sufficient to melt and degas even sanidine. Upon return from irradiation, individual grains are loaded into small 2-mm diameter, 2-mm deep holes in an aluminum plate that is loaded into a ultra-high vacuum laser port with a fused silica window. Grains are heated individually or in small groups in 60-s heating steps using a defocused multimode laser.
Gas samples are accumulated for a minimum of 5 min and equilibrated into the MAP 215-50 noble gas mass spectrometer. All argon peaks are scanned 10 times, and peak heights are extrapolated back to the inlet time to take account of argon buildup and memory effects. The mean blank plus mass spectrometer backgrounds are currently 2.4, 0.02, and 0.3 × 10-12 cm3 STP for 40Ar, 39Ar, and 36Ar respectively. Raw data are corrected for mass spectrometer discrimination and irradiation interference reactions, which are monitored by irradiating calcium and potassium salts with the samples. A J-value is assigned to the samples, using the international biotite mineral standard GA1550 and its assigned age of 97.9 Ma (McDougall and Roksandic, 1974), and an internal biotite standard. Accuracy is difficult to determine, since there is controversy among Ar-Ar specialists about the precise age of the international standards. However, the largest uncertainty in the absolute accuracy is 1.25% at the 1 sigma level. Generally, the precision on the ages is better than 1% at the 2 sigma (95% confidence level). In the case of biotites dating from the Neogene, say 20 Ma, the error is probably 0.2 Ma or better, depending upon the amount of secondary chlorite in the biotite grains.