3. Isotope Measurements
Methane was extracted from the stainless-steel storage cylinders on a vacuum system at UCLA and purified by gas chromatography using procedures described previously by Young et al. (2017). Gas aliquots of a given sample were opened to a silica gel-filled stainless-steel U-trap cooled with liquid nitrogen. Helium carrier gas was used to flush the sample to the gas chromatograph. Separation was accomplished with a 3m long, 1/8 in. OD stainless steel column packed with 5 Å molecular sieve, followed in series by a 2m long ⋅ 1/8 in. OD stainless steel column packed with HayeSep D porous polymer. Peaks from the gas chromatograph were detected with an in-line thermal conductivity detector (TCD). Because total pressures were > 1 bar in the reaction vessels, and because samples were dominated by H2, purification on the vacuum line had to be performed in sequence of multiple aliquots. Once collection of a purified aliquot of methane was complete, this methane was transferred to a glass vial filled with silica-gel at liquid nitrogen temperature. Another sample aliquot would then be purified and added to the same glass vial, until a given reaction vessel had been entirely processed. For the cylinders to be emptied, 6 to 8 aliquots were processed in this way on the vacuum line. Samples were then transferred to the inlet of the mass spectrometer where they were warmed and expanded into the mass spectrometer using a silica-gel filled cold finger.
Mass spectrometry measurements were done on a Nu Instruments Panorama with a mass resolving power that allows the simultaneous measurement of ion currents for resolved12CH4+,13CH4+,12CH3D+,13CH3D+ and12CH2D2+ion beams, as described in Young et al., (2016). The analyte peaks are in most cases baseline resolved from surrounding interferences, and where there is the potential for a tail effect (in the case of13CH5+ tailing into12CH2D2+), previous work shows that the effects on our final results are less than quoted uncertainties (Young et al., 2016). The measured ratios of these ion currents yield values for both Δ13CH3D and Δ12CH2D2 as well as for bulk 13C/12C and D/H. The methane quantities available for isotopic measurements were relatively low. For standards and samples available in quantities > 80 μmol, relative abundances of13CH3D+ and12CH2D2+are measured for up to 10 and 20 hours, respectively. In these cases, typical propagated internal uncertainties for the calculation of Δ13CH3D and Δ12CH2D2 are generally ±0.1‰ and ±0.5‰, respectively. In our study, often only 35 to 60 μmol of CH4 were available (Table 1), and uncertainties are larger, reflecting less favorable counting statistics. The internal uncertainties range from 0.2‰ and 0.5‰ for Δ13CH3D, and between 1.1‰ and 3.2‰ for Δ12CH2D2.