L'isotope cosmogénique Cl-36 dans les minéraux riches en Ca et en K : développements analytiques, calibrations des taux de production et inter-calibration avec le He-3 et le Ne-21

摘要

Published cosmogenic 36Cl SLHL production rates from Ca and K spallation differ by almost 50% (Gosse and Phillips, 2001). The main difficulty in calibrating 36Cl production rates is to constrain the relative contribution of the various production pathways, which depend on the chemical composition of the rock, particularly on the Cl content. Whole rock 36Cl exposure ages were compared with 36Cl exposure ages evaluated in Ca-rich plagioclases in the same independently dated 10 ± 3 ka lava sample taken from Mt. Etna (Sicily, 38° N). Sequential dissolution experiments showed that high Cl concentrations in plagioclase grains could be significantly reduced after 16% dissolution yielding 36Cl exposure ages in agreement with the independent age. Stepwise dissolution of whole rock grains, on the other hand, is not as effective in reducing high Cl concentrations as it is for the plagioclase. 330 ppm Cl still remains after 85% dissolution. The 36Cl exposure ages are systematically about 30% higher than the ages calculated from the plagioclase. We could exclude contamination by atmospheric or magmatic 36Cl as an explanation for this overestimate. High Cl contents in the calibration samples used for several previous production rate studies are most probably the reason for overestimated spallation production rates from Ca and K. This is due to a poorly constrained nature of 36Cl production from low-energy neutrons. We used separated minerals, very low in Cl, to calibrate the production rates from Ca and K. 36Cl was measured in Ca-plagioclases collected from 4 lava flows at Mt. Etna (38° N, Italy, altitudes between 500 and 2000 m), and in K-feldspars from one flow at Payun Matru volcano (36° S, Argentina, altitudes 2300 and 2500 m). The flows were independently dated between 0.4 and 32 ka. Scaling factors were calculated using five different published scaling models resulting in five calibration data sets. Using a Bayesian statistical model allowed including the major inherent uncertainties. The inferred SLHL spallation production rates from Ca and K are 42.2 ± 4.8 atoms 36Cl (g Ca)-1 a-1 and 124.9 ± 8.1 atoms 36Cl (g K)-1 a-1 scaled with Stone (2000). Using the other scaling methods results in very similar values. These results are in agreement with previous production rate estimations both for Ca and K calibrated with low Cl samples. Moreover, although the exposure durations of our samples are very different and the altitude range is large, the ages recalculated with our production rates are mostly in agreement, within uncertainties, with the independent ages no matter which scaling method is used. However, scaling factors derived from the various scaling methods differ significantly. Cosmic ray flux is sensitive to elevation and its energy spectrum increases considerably with increasing altitude and latitude. To evaluate whether various TCN production rates change differently with altitude and latitude and if nuclide-specific or even target-element-specific scaling factors are required, cosmogenic 36Cl, 3He and 21Ne concentration were determined in pyroxenes over an altitude transect between 1000 and 4300 m at Kilimanjaro volcano (3° S). No altitude-dependency of the nuclide ratios could be observed, suggesting that no nuclide-specific scaling factors be needed for the studied nuclides.