Comparison of annular denuders and filter packs for atmospheric sampling

Abstract

Filter packs have commonly been used to sample atmospheric gases and particles. However, reactions between the gases and particles can occur, particularly during summer sampling. Recently, annular denuder systems (ADS) have been developed, consisting of denuder tubes that strip the reactive gases from the air, thus leaving the particles to be collected on the ADS filters. We compared the ADS to filter packs in the winter, when the filter packs are most reliable. The species sampled included SO₂, HNO₃, HCl and particulate NO₃⁻, SO₄²⁻, Na⁺ and Cl⁻. HNO₂ was also sampled with the ADS.

Filter packs are far less expensive than ADS and simpler to use in the field. The usual filter pack problem, dissociation of particulate NO₃⁻ to gaseous HNO₃, was not apparent during winter sampling. However, the open-faced filter packs are more exposed to the elements than the ADS, thus leading to high and variable blank levels—particularly for NaCl where blank levels averaged 40% of measured concentrations. In addition, up to 50% of the SO₂ was collected on the backup collector, indicating occasional poor collection efficiency. In contrast, the ADS had low blanks and high collection efficiencies with less than 3% of the SO₂ on the backup collector.

Measurements between filter packs and the ADS agreed within 10% for particles and hydrochloric acid. However, due to losses of HNO₃ in the filter pack and small losses of particles in the denuder sections, HNO₃ concentrations appear greater with the ADS than with filter packs. To the extent that the particle loss in the ADS is due to bounce-off from the impactor frit, it can be corrected in future studies. Sulfur dioxide is also 14% greater with the ADS than with the filter pack and reasons for the difference are considered.