Passive sampling capabilities for ultra-trace quantitation of atmospheric nitric acid (HNO3) in remote environments

Abstract Custom-built HNO 3(g) passive diffusion-based samplers employing nylon filters were comprehensively quality assurance and quality control ( QA/QC) tested, including analysis of the PTFE protective filters for potential bias. The passive samplers routinely quantified HNO 3(g) mixing ratiosin the Canadian borealup to a factor of 20 lower compared to previous passive sampling dose-response determination and measurements. The use of appropriate blanks alongside field collections resulted in the detection of time-weighted mixing ratiosas low as 5 parts per trillion by volume (pptv) over a monthly sampling period or 121 pptv over a 24-h sampling period. The nylon filters were tested for a single reuse to reduce resources required and no statistical difference between the measured mean value of reused and new filters between five replicates across four intercomparisons was found. This showed that a single nylon filter reuse in remote regions is sound as long as replicates are employed. The custom-built passive samplers were installed along the Newfoundland and Labrador – Boreal EcosystemLatitudinal Transect (NL-BELT) to quantify mixing ratiosof HNO 3(g) in these forests. Samples generated monthly time-weighted average mixing ratiosat the NL-BELT field regions. Samples were collected monthly across 15 months from the summer of 2015 through to the fall of 2016 on the order of tens to a hundred pptv. Across this time period the HNO 3(g) mixing ratiosranged from 9 to 200 pptv – after incorporation of comprehensive field QA/QC– with seasonal maxima occurring during the early summer months. No consistent latitudinal trends were observed along the transect, except that the most northern region always corresponded to the lowest mixing ratiosof HNO 3(g) for a given observation period. Extracts of the protective PTFE filter and collocated measurements of temperature and humidity determined that previously uncertain effects on sampling accuracy likely result in a negative sampling biasof 10–30%. These unpowered devices are now demonstrated to be sensitive, affordable, simple, and redeployable for the continuous ultra-trace detection of HNO 3(g) in regions such as the remote Canadian borealforest.