跳转到主要内容

The Journey of a Thousand Good Measurements Begins with One Replicate: Optimizing CRDS Measurements for Capturing Day-to-Day Variations in Triple Oxygen Isotopes of Rainfall

Literature Reference
Peer Reviewed Literature
日期
十二月 1st, 2021
Abstract

The demanding precision of triple oxygen isotope (17O) measurements in water has restricted its measurement to dual-inlet mass spectrometry until the recent development of commercially available infrared-laser analyzers, including cavity ringdown spectrometers (CRDS). 17O is now increasingly measured by laboratories in order to better constrain the source and history of meteoric waters. However, in practice, these measurements are subject to large analytical errors, many of which are not yet well-documented in the literature or by instrument manufacturers, which can restrict their confident application to settings where 17O variations are relatively large (~ 25-60 per meg). We present our operating method of a Picarro L2140-i CRDS during the analysis of low-latitude rainwaters where confidently resolving daily variations in 17O (differences of ~10-20 per meg) is desired. Our approach, optimized over ~ 2 years, uses a combination of published best-practices (some modified) plus additional steps to combat spectral contamination by trace amounts of dissolved organics, which emerges as a much more substantial problem than previously documented even in pure rainwater samples. We present our suite of corrections for sample memory, drift, and two-point scale normalization. We resolve the extreme sensitivity of the 17O measurement to organics through their removal via the Picarro micro-combustion module, whose performance is evaluated each sequence using an alcohol-spiked standard. In addition to competitive performance for traditional metrics, we report a long-term, control standard root-mean-square error for 17O of 11 per meg, and an overall performance that is comparable to mass spectrometry (17O error of ~ 7 per meg, calculated by averaging 3 replicates spread across distinct, independently calibrated sequences) and requires only 6.3 h per unknown sample. We demonstrate the impact of each correction using a rainfall dataset from Uganda and offer recommendations for other community efforts that aim to measure meteoric 17O via CRDS.