Helium data

Method description of helium analysis. All the OREX samples were analyzed during the CRPG-CNRS session with the Helix MC mass spectrometer, from November 2023 to April 2024. This file describes the flowchart of neon data processing. The table gives the number of lines, the label for each line, the sample #, the 4He signal in fA (lines 1-81) and in mol/g (lines 85-125), its standard deviation, the 3He signal in fA (lines 1-81) and in mol/g (lines 85-125), its standard deviation, the 3He/4He ratio, its standard deviation, the percentage of blank relative to the sample’s signal for 4He, the sample’s mass, the laser power percentage relative to full power, and the laser power in Watt. - Lines 1- 25 (label “standard) display data for 24 standard gas runs - Lines 27 and 28 display the mean and the standard deviation divided by the square root of n value for 4He, 3He and 3He/4He - Lines 30-36 display data for 5 blank runs, i.e., same procedure as standard and sample runs without gas. These runs yield the amount of gas released by the analytical system - Lines 41-81display samples’ raw data - Line 83 is a copy of the column labels - Lines 85-125 display samples’ processed data. The correction procedure consists in: - (1) correcting for blank by removing the mean blank value from the raw values; - (2) computing the 3He and 4He concentrations using the sensitivity of the mass spectrometer determined with the amount of standard gas, in mole; - (3) dividing the concentration by the mass of sample, in mole/g; - (3) correcting for mass discrimination using the mean standard value. The 3He/4He ratio is divided by the mean standard ratio and multiplied by the 3He/4He ratio of the standard gas, which is HESJ (R/Ra) = 20.63 ± 0.10; Matsuda et al Geochemical Journal, 36, 191–195, where Ra is the atmospheric 3He/4He ratio; Ra = 1.39 x 10-6) Errors are propagated accordingly. For samples, the error is taken as a quadratic sum of error on standard composition, standard error/SQRT(n) which represents variability of n standards, and standard error/SQRT, which is a proxy for error on sample measurement given that only one run was done.