Intercomparison of NO2 measurement techniques

Abstract

An intercomparison was made near Niwot Ridge, Colorado, of three different instruments for measuring NO2 at low concentrations in ambient air: (1) the photolysis/chemiluminescence (PC) instrument, (2) the tunable diode laser absorption spectrometer (TDLAS), and (3) the Luminox instrument. Calibrated mixtures of NO2 in air and NO2 with possible interferants (HNO3, peroxyacetyl nitrate (PAN), H2O2, n‐propyl nitrate, and O3) were provided in simultaneous tests. In addition, ambient air measurements were made using the three instruments. Blind procedures were followed in preparing all results. Several conclusions were reached concerning the performance of these instruments during this intercomparison: (1) For NO2 levels above 2 parts per billion by volume (ppbv), similar results were obtained for all instruments; (2) Below 2 ppbv, the expected interferences from ozone and PAN influenced the NO2 measurements made using the Luminox instruments. Those interferences were sufficiently consistent that they could be corrected for by using the measured values Of O3 and PAN down to about 0.3 ppbv NO2; (3) The ozone interference on the Luminox instruments was removed by an ozone scrubber placed in the sampled air stream of the Luminox instrument. However, this did not remove PAN. In addition, the scrubber appeared to remove about 50% of the NO2 as well; (4) Although no interferences were identified for the TDLAS technique, care must be taken in the data analysis near (or below) the detection limit for the instrument. At these levels the data reduction program provided with the TDLAS will tend to find background noise that is correlated with the reference NO2 spectrum and calculate levels of NO2 that are too high; (5) No interferences or artifacts were found for the final results reported by the PC technique. However, these results for ambient measurements were corrected by subtracting an artifact that averaged 5 parts per trillion by volume (pptv) and by calculating a correction for the effect of ambient ozone. This latter correction averaged 1.0% in magnitude.