Measurements of volatile organic compounds in ambient air by gas-chromatography and real-time Vocus PTR-TOF-MS: calibrations, instrument background corrections, and introducing a PTR Data Toolkit

Abstract

Abstract. Volatile organic compound (VOC) emissions and subsequent oxidation contribute to the formation of secondary pollutants and poor air quality in general. As more VOCs at lower mixing ratios have become the target of air quality investigations, their quantification has been aided by technological advancements in proton-transfer-reaction time-of-flight mass spectrometry (PTR-TOF-MS). However, such quantification requires appropriate instrument background measurements and calibrations, particularly for VOCs without calibration standards. This study utilized a Vocus PTR-TOF-MS coupled with a gas chromatograph for real-time and speciated measurements of ambient VOCs in Boulder, Colorado, during spring 2021. The aim of these measurements was to understand and characterize instrument response and temporal variability as to inform the quantification of a broader range of detected VOCs. Fast, frequent calibrations were made every 2 h in addition to daily multipoint calibrations. Sensitivities derived from the fast calibrations were 5 ± 6 % (average and 1 standard deviation) lower than those derived from the multipoint calibrations due to an offset between the calibrations and instrument background measurement. This offset was caused, in part, by incomplete mixing of the standard with diluent. These fast calibrations were used in place of a normalization correction to account for variability in instrument response and accounted for non-constant reactor conditions caused by a gradual obstruction of the sample inlet. One symptom of these non-constant conditions was a trend in fragmentation, although the greatest observed variability was 6 % (1 relative standard deviation) for isoprene. A PTR Data Toolkit (PTR-DT) was developed to assess instrument performance and rapidly estimate the sensitivities of VOCs which could not be directly calibrated on the timescale of the fast calibrations using the measured sensitivities of standards, molecular properties, and simple reaction kinetics. Through this toolkit, the standards' sensitivities were recreated within 1 ± 8 % of the measured values. Three clean-air sources were compared: a hydrocarbon trap, zero-grade air and ultra-high purity nitrogen, and a catalytic zero-air generator. The catalytic zero-air generator yielded the lowest instrument background signals for the majority of ions, followed by the hydrocarbon trap. Depending on the ionization efficiency, product ion fragmentation, ion transmission, and instrument background, standards' limits of detection (5 s measurement integration) derived from the catalytic zero-air generator and the fast calibration sensitivities ranged from 2 ppbv (methanol) to 1 pptv (decamethylcyclopentasiloxane; D5 siloxane) with most standards having detection limits below 20 pptv. Finally, applications of measurements with low detection limits are considered for a few low-signal species including sub-parts-per-trillion by volume (pptv) enhancements of icosanal (and isomers; 1 min average) in a plume of cooking emissions, and sub-parts-per-trillion by volume enhancements in dimethyl disulfide in plumes containing other organosulfur compounds. Additionally, chromatograms of hexamethylcyclotrisiloxane, octamethylcyclotetrasiloxane, and decamethylcyclopentasiloxane (D3, D4, and D5 siloxanes, respectively), combined with high sensitivity, suggest that online measurements can reasonably be associated with the individual isomers.