On the use of data from commercial NOx analyzers for air pollution studies

Highlights

• Commercial NOx analyzers are in common use.

• Substantial interferences have been documented.

• Some recent studies appear to neglect these interferences.

• Results add urgency to deployment of true NOx.

Abstract

Reactive oxidized nitrogen species play a central role in environmental pollution, and long term monitoring is widespread. But conventional NOx (defined as NO + NO₂) analyzers employing heated converters respond to many species in addition to the compounds NO and NO₂. The response of these instruments to nitric acid, peroxyacetyl nitrate, alkyl nitrates and other oxidized nitrogen species (the sum of these plus NOx is defined as NOy) is well established, but the ratio of NOx to NOy varies widely in time and space making the accuracy of commercial NOx monitors uncertain. Care must be taken when comparing spectroscopic measurements of NO₂ or numerical models to output from commercial NOx monitors. Correction factors can be developed for specific conditions, and long term trends can be meaningful. Recent studies comparing modeled NOx to measurements with large interferences can involve errors of a factor of two or more and produce misleading guidance on science and policy; the need for rigorous model evaluation adds urgency to the deployment of “true NOx” monitors.

Introduction

Excess reactive oxidized nitrogen, concentrations above the natural atmospheric background in the atmosphere, lead to a variety of environmental problems, especially ground level ozone, a recalcitrant air quality problem that poses a threat to human health and climate. Ozone production rates respond nonlinearly to NOx concentrations (Chameides et al., 1992; Crutzen, 1973; Sillman et al., 1990) making it vitally important to measure NOx (NO + NO₂) specifically and to represent NOx concentrations accurately in chemical transport models e.g., (Lelieveld et al., 2015). If the modeled concentration of NOx is wrong, simulations can give misleading guidance on pollution control policy. Because ozone production rates fall off at higher NOx levels, model overestimates in NOx concentrations can contribute to underestimates in the benefit to ozone from a given cut in NOx emissions e.g., (Gilliland et al., 2008).

Oxides of nitrogen are most frequently monitored via a chemiluminescent reaction with ozone (Fontijn et al., 1970) with NO measured directly and higher oxides following reduction to NO (Fehsenfeld et al., 1987; Winer et al., 1974). Because NO₂ is designated one of the criteria pollutants by the United States Environmental Protection Agency (USEPA), States are required to demonstrate attainment of the standard, and chemiluminescence NOx analyzers are in common use. Most monitors reporting to the EPA's Air Quality System, AQS (https://www.epa.gov/aqs), employ commercial NOx analyzers with hot molybdenum NO₂ converters – one purpose of this paper is to alert users these do not measure true NOx, because they suffer substantial interferences from other reactive nitrogen species. These interfering species include nitrous acid (HONO), nitric acid (HNO₃), nitric acid anhydride (N₂O₅), organic nitrogen peroxides, alkyl nitrates (RONO₂), nitryl chloride (ClNO₂) and other important air pollutants. These instruments more nearly measure NOy than NOx; in the US, a good approximation is NOy = NOx + HONO + HNO₃ + 2XN₂O₅ + PANs + RONO₂, where PANs represents the family of peroxyacetyl nitrates and RONO₂ represents the family of alkyl nitrates. When nitrate aerosol passes through the inlet it can likewise be detected as NOx, and some species such as NH₄NO₃ thermally decompose to NH₃ and HNO₃ that can cause interferences. Although NH₃ oxidation on heated molybdenum is usually small at ambient humidity, under certain circumstances ammonia and amines can also cause substantial interferences (Saylor et al., 2010; Suzuki et al., 2011). The difference between NOy and NOx is sometimes referred to as NOz, and gives an indication of the aging of the air parcel e.g., (Gaudel et al., 2018; Kleinman et al., 2002). Commercial “NOx” instruments are in common use because the interferences do not cause a problem when such monitors are deployed to demonstrate attainment with NO₂ standards – they provide an upper bound on NO₂ and NOx concentrations. So while these monitors can be useful, they generate numbers with variable, and often severe, high bias.

The high efficiency of hot molybdenum converters for NOy has been known for some time (Fehsenfeld et al., 1987; McClenny et al., 2002) and many investigations have documented or quantified substantial interferences in commercial instruments deployed in Europe, Asia, South America, as well as North America (Dunlea et al., 2007; Geddes and Murphy, 2014; Hassler et al., 2016; Lamsal et al., 2008; Leston and Ollison, 2017; Luke et al., 1998; Ordonez et al., 2006; Piters et al., 2012; Poulida et al., 1994; Reed et al., 2016; Steinbacher et al., 2007; Suzuki et al., 2011; Villena et al., 2012; Wild et al., 2014; Xu et al., 2013). In this invited paper we will discuss the appropriate comparison of remotely sensed NO₂ to surface-based measurements as well as models and measurements of reactive nitrogen, and show how uncertain high bias or assumptions of equivalence between NOx and NOy can lead to misleading results.