A sealed-tube method for offline δ 13 C analysis of CO 2 via a Gas Bench II continuous flow isotope ratio mass spectrometer.

Rationale The isotopic measurement of environmental sample CO2 via isotope ratio mass spectrometry (IRMS) can present many analytical challenges. In many offline applications, exceedingly few samples can be prepared per day. In such applications, long-term storage (months) of sample CO2 is desirable, in order to accumulate enough samples to warrant a day of isotopic measurements. Conversely, traditional sample tube cracker systems for dual-inlet IRMS offer a capacity for only 6-8 tubes and thus limit throughput. Here we present a simple method to alleviate these concerns using a Gas Bench II gas handling device coupled with continuous-flow IRMS. Methods Sample preparation entails the cryogenic purification and quantification of CO2 on a vacuum line. Sample CO2 splits are expanded from a known volume to several sample ports and allowed to isotopically equilibrate (homogenize). Equilibrated CO2 splits are frozen into 3 mm OD Pyrex break-seals and sealed under vacuum with a torch to a length of 5.5 cm. Sample break-seals are scored, placed into 12-mL Labco Exetainer® vials, purged with ultra-high purity helium (He), cracked inside the capped He-flushed vial and subsequently measured via a Gas Bench equipped Isotope Ratio Mass Spectrometer using a CTC Analytics PAL autosampler. Results Our d13 C results from NIST and internal isotopic standards, measured over a time period of several years, indicate that the sealed-tube method produces accurate δ13 C values to a precision of ±0.1‰ for samples containing 10-35 μgC. The tube cracking technique within Exetainer® vials has been optimized over a period of 10 years, resulting in decreased sample failure rates from 5-10% to Conclusions This technique offers an alternative method for δ13 C analyses of CO2 where off-line isolation and long-term storage are desired. The method features a much higher sample throughput than traditional dual-inlet IRMS cracker setups at similar precision (±0.1‰).