An automated sampler for collection of atmospheric trace gas samples for stable isotope analyses

Research focused on the isotopic composition of CO2 exchange between terrestrial ecosystems and the atmosphere has been historically constrained by the need for personnel to be present at remote field sites for sample collection. In practice, this has limited sampling frequency and duration, and potentially even biases sampling events to fair weather periods. We have developed an automated sampling system that can be installed and used for unattended collection of 100-ml air samples in remote areas. The sampler was designed with the primary goal of collecting samples for analysis of CO2 concentration and its isotopic composition in ecosystem-atmosphere flux research, but several other potential applications are also discussed. Laboratory tests examined potential artifacts associated with sampler components. These tests included evaluation of potential isotopic exchange between atmospheric CO2 and sampler component materials and the effects of sample exposure to these materials for up to 5 days and under a wide range of temperatures (10-50degreesC). Some of the rejected component materials influenced either CO2 mole fraction or CO2 isotopic content. Exposure of air at subambient CO2 concentrations to all sampler components in an intact system for 5 days resulted in a [CO2] value that was 0.9 mumol mol(- 1) higher than for an equivalent sample collected by the sampler but not stored. Associated exposure-induced errors in delta(13)C Of CO2 were generally small, ranging between 0.03 and 0.17parts per thousand for 0 day versus 5 days exposure, respectively. These error values were within the sampling precision associated with a PreCon continuous flow mass spectrometer analysis. A more substantial exposure-induced error was observed for delta(18)O in CO2 (0.29 and 0.88parts per thousand, respectively). The potential for isotopic exchange between CO2 and sampler components increased under a combination of elevated temperature and multiple-day storage treatments. These errors were small and of similar magnitude between 10 and 40degreesC, but unacceptably large at 50degreesC. Finally, we compared Keeling plots created with samples collected by the sampler with those collected simultaneously by a manual method and found no detectable differences between the two approaches. Based on these results, we conclude that sampler induced isotopic exchange for air samples held up to 5 days between 10 and 40degreesC is largely within the overall precision limits of a PreCon continuous flow mass spectrometer measurement. (C) 2003 Elsevier Science B.V. All rights reserved.