Considerations for double dating zircon in secular disequilibrium with protracted crystallisation histories

Zircon double-dating utilises (U–Th)/He dating coupled with U–Th disequilibrium or U–Pb dating to determine eruption ages for volcanic rocks between ca. 2 ka to 1 Ma. This approach depends on understanding the crystallisation history of each zircon crystal analysed. For lack of better constraints, zircon crystallisation is generally assumed to be represented by a single crystallisation age, which is routinely determined on the rim of the grain by spot analyses. While zircon crystallisation is often protracted, interrogating the crystallisation history of a zircon crystal usually requires grinding the grain, which can introduce uncertainty to the alpha ejection (Ft) correction, critical for accurate (U–Th)/He ages. Grinding a zircon crystal to exactly 50% of its original width, to a plane of symmetry, leaves the Ft correction factor unchanged relative to that of the whole crystal. This is verified by a new computer program – GriFt, which also allows the calculation of accurate Ft correction factors for a range of different grinding depths, opening the opportunity to measure both the core and rim crystallisation ages and integrate these into a more robust disequilibrium correction of (U–Th)/He data. The feasibility of this approach is tested here in a case study of zircon crystals with protracted crystallisation histories from the Shikotsu-Toya volcanic field in Hokkaido, Japan. A maximum of 15% difference in overall eruption age is calculated between rim- and core-corrected (U–Th)/He ages. Eruption ages were determined for two tephras – Kimobetsu 1 (59–79 ka) and Kimobetsu 2 (96 ± 5 ka, 2σ). The geological implication from these dates is that a regionally important tephra, Toya, may be younger (<96 ± 5 ka) than previously reported (109 ± 3 ka). In addition, the maximum eruption ages determined from crystallisation age distributions calculated for samples from eruptions at Shikotsu and Kuttara (48 ± 17 and 49 ± 21 ka, respectively) are within uncertainty of previous measurements (44–41 ka and >43 ka, respectively).