Detailed characterizations of a Comparative Reactivity Method (CRM) instrument: experiments vs. modelling

Abstract

The Hydroxyl radical (OH) is an important oxidant in the daytime troposphere that controls the lifetime of most trace gases, whose oxidation leads to the formation of harmful secondary pollutants such as ozone (O 3) and Secondary Organic Aerosols (SOA). In spite of the importance of OH, uncertainties remain concerning its atmospheric budget and integrated measurements of the total sink of OH can help reducing these uncertainties. In this context, several methods have been developed to measure the first-order loss rate of ambient OH, called total OH reactivity. Among these techniques, the Comparative Reactivity Method (CRM) is promising and has already been widely used in the field and in atmospheric simulation chambers. This technique relies on monitoring competitive OH reactions between a reference molecule (pyrrole) and compounds present in ambient air inside a sampling reactor. However, artefacts and interferences exist for this method and a thorough characterization of the CRM technique is needed. In this study, we present a detailed characterization of a CRM instrument, assessing the corrections that need to be applied on ambient measurements. The main corrections are, in the order of their integration in the data processing: (1) a correction for a change in relative humidity between zero air and ambient air, (2) a correction for the formation of spurious OH when artificially produced HO 2 react with NO in the sampling reactor, and (3) a correction for a deviation from pseudo first-order kinetics. The dependences of these artefacts to various measurable parameters, such as the pyrrole-to-OH ratio or the bimolecular reaction rate constants of ambient trace gases with OH are also studied. From these dependences, parameterizations are proposed to correct the OH reactivity measurements from the abovementioned artefacts. A comparison of experimental and simulation results is then discussed. The simulations were performed using a 0-D box model including either (1) a simple chemical mechanism, taking into account the inorganic chemistry from IUPAC 2001 and a simple organic chemistry scheme including only a generic RO 2 compounds for all oxidized organic trace gases; and (2) a more exhaustive chemical mechanism, based on the Master Chemical Mechanism (MCM), including the chemistry of the different trace gases used during laboratory experiments. Both mechanisms take into account self- and cross-reactions of radical species. The simulations using these mechanisms allow reproducing the magnitude of the corrections needed to account for NO interferences and a deviation from pseudo first-order kinetics, as well as their dependence on the Pyrrole-to-OH ratio and on bimolecular reaction rate constants of trace gases. The reasonable agreement found between laboratory experiments and model simulations gives confidence in the parameterizations proposed to correct the Total OH reactivity measured by CRM. However, it must be noted that the parameterizations presented in this paper are suitable for the CRM instrument used during the laboratory characterization and may be not appropriate for other CRM instruments, even if similar behaviours should be observed. It is therefore recommended that each group characterizes its own instrument following the recommendations given in this study. Finally, the assessment of the limit of detection and total uncertainties is discussed and an example of field deployment of this CRM instrument is presented.