The most important results came from the thermal ionization experiments. The thermal ionization mass spectrometer used in this work is similar to others described previously (3). It has a single magnet with 90° deflection and a 30-cm central radius of curvature. It is equipped with a pulse-counting detection system to allow complete isotopic analyses to be made on small quantities(<1 ng) of suitable elements ionized from a single filament. The filaments, made of V-shaped rhenium foil 0.64 cm long and 0.08 cm deep (4), were baked out at 2000°C before loading the zircons. Ions are formed by resistive heating of the filament; typical temperatures for this work were 1400° to 1470°C (uncorrected pyrometer readings).

Previous work done to develop a technique for analyzing small lead samples led to the use of silica gel to enhance ionization efficiency (5). Because individual zircons are chemically somewhat similar to silica, we decided to try to analyze lead from individual zircons loaded directly on the rhenium filament. Such a technique would have several advantages over traditional methods: contamination would be essentially eliminated because no chemical separation would be required and, since the zircons are small (~ 50 μm in diameter), they would provide an approximate point source of ions, which is known to optimize ion-optical conditions in the mass spectrometer (6).

Test experiments with zircons from other localities (7) were uniformly successful; ion signals were observed at masses (m) 206, 207, and 208 which could definitely be ascribed to Pb isotopes. To help ensure that we were at the correct ion lens conditions, we focused on the ¹³⁸BaO⁺ peak (the zircons contained some Ba), which was reasonably intense at 1200°C. Surficial residues left on the zircons after the acetone wash burned off before the operating temperature of 1450°C, where the lead signal was measured. Great care had to be exercised to avoid making the temperature too high; very rapid evaporation of the lead occurred only a little above the operating temperature. Typical count rates were 100 to 3000 counts per second for ²⁰⁶Pb⁺. Traces of thallium (m = 203 and 205) were sometimes observed, but burned out more rapidly than the lead. Other than thallium, lead gave the only substantive peaks in the range m = 202 to 210. There was, however, a general background generated by the sample; chemically unseparated samples such as these zircons almost always yield such backgrounds. This background has little effect on the 206, 207, and 208 peaks, but made precise measurement of the ²⁰⁴Pb signal, which was very small, impossible. For example, in an analysis typical of these experiments, 1.6 × l0⁵ counts from ²⁰⁶Pb were collected; the background correction was about 40 counts and, after correction, 18 counts remained at mass 204. Although these counts are listed as ²⁰⁴Pb counts in Table 2, more work is needed to determine how much may be uncompensated background.

Table 2 shows the results of our mass analyses of filaments loaded with single and multiple zircons from five granite cores. The range of ²⁰⁶Pb/²⁰⁸Pb values reflects the fact that this ratio varied from one group of zircons to another, and sometimes varied during measurements on a single zircon. These variations are not surprising in view of the ion microprobe analyses, which showed significant U/Th variations at different points on a single zircon (²³²Th decays to ²⁰⁸Pb and ²³⁸U decays to ²⁰⁶Pb). These variable ²⁰⁶Pb/²⁰⁸Pb ratios do not furnish any direct information on differential Pb retention in these zircons. For that purpose, it is generally accepted that the Radiogenic ²⁰⁶Pb/²⁰⁷Pb ratios derived from ²³⁸U/²³⁵U decay are more specific. We note that Zartman's (8) isotopic measurements of Pb, which was chemically extracted from zircons taken from the GT-2 core at 2900 m, yield an adjusted ²⁰⁶Pb/²⁰⁷Pb ratio (9) that approximates our ratios.

In a conventional chemical extraction of lead from zircons, the lead measured in the mass analysis is considered to be a combination of radiogenic lead (from U and Th decay) and nonradiogenic lead (from common lead contamination and from some initial lead in the zircon). The radiogenic component is obtained by subtracting out a nonradiogenic component proportional to the amount of ²⁰⁴Pb. In our experiments, however, the direct loading procedure virtually eliminated the common lead contamination, and we circumvented the need to make adjustments for initial lead in the zircons by accepting only analyses (10) showing a ratio of ²⁰⁴Pb to total Pb of less than 2 × 10⁻³. Thus the ²⁰⁶Pb/²⁰⁷Pb ratios shown in Table 2 represent highly radiogenic lead and hence are potential indicators of Pb retention.

We consider that the most important observations on the data in Table 2 are: (i) the fact that the ²⁰⁶Pb/²⁰⁷Pb ratios on single zircons closely approximate the ratio obtained when a group of similar zircons was loaded simultaneously on a single filament, (ii) the relative uniformity of the ²⁰⁶Pb/²⁰⁷Pb ratios for zircons from all depths, and (iii) the fact that the total number of Pb counts per zircon (the counts in column 4 of Table 2 divided by the product of columns 2 and 3) shows no systematic decrease with depth, as would be expected if differential Pb loss had occurred at higher temperatures. Taken together, items (ii) and (iii) provide strong evidence for high Pb retention in zircons even for a prolonged period in an environment at an elevated temperature. These results have possible implications for long-term nuclear waste disposal.