39Ar and 37Ar recoil loss during neutron irradiation of sanidine and plagioclase

Abstract

The 40Ar/ 39Ar dating technique requires the activation of 39Ar via neutron irradiation. The energy produced by the reaction is transferred to the daughter atom as kinetic energy and triggers its displacement, known as the recoil effect. Significant amounts of 39Ar and 37Ar can be lost from minerals leading to spurious ages and biased age spectra. Through two experiments, we present direct measurement of the recoil-induced 39Ar and 37Ar losses on Fish Canyon sanidine and plagioclase. We use multi-grain populations with discrete sizes ranging from 210 to <5 μm. One population consists of a mixture between sanidine and plagioclase, and the other includes pure sanidine. We show that 39Ar loss (depletion factor) for sanidine is ∼3% for the smallest fraction. Age spectra of fractions smaller than ∼50 μm show slight departure from flat plateau-age spectrum usually observed for large sanidine. This departure is roughly proportional to the size of the grain but does not show typical 39Ar loss age spectra. The calculated thickness of the total depletion layer d 0(sanidine) is 0.035 ± 0.012 (2 σ). This is equivalent to a mean depth of the partial depletion layer ( x 0) of 0.070 ± 0.024 μm. The latter value is indistinguishable from previous values of ∼0.07–0.09 μm obtained by argon implantation experiments and simulation results. We show that it is possible to adequately correct ages from 39Ar ejection loss provided that the d 0-value and the size range of the minerals are sufficiently constrained. As exemplified by similar calculations performed on results obtained in a similar study of GA1550 biotite [Paine J. H., Nomade S., and Renne P. R. (2006) Quantification of 39Ar recoil ejection from GA1550 biotite during neutron irradiation as a function of grain dimensions. Geochim. Cosmochim. Acta 70, 1507–1517.], the d 0(biotite) is 0.46 ± 0.06 μm. The significant difference between empirical results on biotite and sanidine, along with different simulation results, suggests that for biotite, crystal structures and lattice defects of the stopping medium and possibly subsequent thermal degassing (due to ∼150–200 °C temperature in the reactor or extraction line bake out) must play an important role in 39Ar loss. The second experiment suggests that 37Ar recoil can substantially affect the age via the interference corrections with results that suggest up to ∼98% of 37Ar can be ejected from the ∼5 μm grain dimension. Further investigation of silicates of various compositions and structures are required to better understand (and correct) the recoil and recoil-induced effects on both 39Ar and 37Ar and their influences on 40Ar/ 39Ar dating.