The existence of isocyanic acid (HNCO) in the atmosphere has been established only recently (Roberts et al., 2011; Wentzell et al., 2013) despite its molecular structure and chemical synthesis being first discovered in the 19th century (Liebig and Wöhler, 1830). HNCO can form H-bonded clusters (Zabardasti et al., 2009, 2010; Zabardasti and Solimannejad, 2007) and in pure form appreciably polymerizes to other species (Belson and Strachan, 1982) but becomes relatively stable in the gaseous phase (ppm level) under ambient temperature conditions (Roberts et al., 2010). It is thus near-exclusively present as a monomer in the gaseous phase under ambient temperature conditions (Fischer et al., 2002; Roberts et al., 2010). The background ambient mixing ratios of HNCO as determined by Young et al. (2012) using a global chemistry transport model vary in the range of a few parts per trillion by volume (pptv) over the ocean and remote Southern Hemisphere to tens of pptv over land. In urban regions, HNCO mixing ratio increases from tens of pptv to hundreds of pptv (Roberts et al., 2014; Wentzell et al., 2013). Peak levels can reach up to a few parts per billion by volume under the conditions impacted by direct emissions (Chandra and Sinha, 2016).

HNCO has been linked to adverse health effects such as cataracts, cardiovascular disease, and rheumatoid arthritis via a process called protein carbamylation (see Leslie et al., 2019; Roberts et al., 2011; Suarez-Bertoa and Astorga, 2016; SUVA, 2016; Wang et al., 2007, and references therein). To our knowledge, no past studies have been performed to provide a direct link between inhalation exposure and related adverse health effects. However, human exposure to HNCO concentrations of 1 ppbv is estimated to be potentially sufficient to start the process of protein carbamylation (Roberts et al., 2011). Unfortunately, an air quality standard for HNCO does not exist in most of the countries, whereas an occupational exposure limit has been established by law in only a few countries, including the Swedish Work Environment Authority (SWEA, 2011) and the Swiss National Accident Insurance Fund (SUVA, 2016). For example, the Swedish Work Environment Authority sets the level limit value (LLV) for HNCO at about 0.018 mgm-3, i.e. 10 ppbv (SWEA, 2011). The potential negative impact on health makes it important to assess the atmospheric sources and sinks of HNCO to determine its fate and lifetime.

HNCO emission into the atmosphere is driven primarily by combustion processes based on both natural and anthropogenic activities (see Leslie et al., 2019, and references therein), where the pyrolysis of nitrogen-containing biomass materials during the events of wildfires and agricultural fires leads to the emission of HNCO into the atmosphere. The presence of HNCO in cigarette smoke has been established via the pyrolysis of urea used as a cigarette additive (Baker and Bishop, 2004), oxidation of nicotine (Borduas et al., 2016a), and oxidation of formamide (Barnes et al., 2010; Borduas et al., 2015; Bunkan et al., 2015). Even the combustion of almost all sorts of common household materials, including fibre glass, rubber, wood, PVC-based carpet, and cables (Blomqvist et al., 2003), and polyurethane-based foam (Blomqvist et al., 2003; Jankowski et al., 2014), leads to HNCO emissions along with other isocyanates (Leslie et al., 2019). HNCO emissions from traffic are originating mainly from usage of recent catalytic converters in the exhaust systems of gasoline-based (Brady et al., 2014) and diesel-based (Heeb et al., 2011) vehicles. These converters are implemented to control the emission of primary pollutants such as hydrocarbons, carbon monoxide, particulate matter, and nitrogen oxides. However, these implementations have promoted (Suarez-Bertoa and Astorga, 2016) the formation and emissions of HNCO via surface-bound chain reactions at different stages of the flue gas exhaust and additionally due to emission of unreacted HNCO in the most commonly used urea-based SCR (selective catalytic reduction) system (Heeb et al., 2011). The usages of catalytic converters in modern vehicles potentially give rise to the emission of HNCO especially in urban regions with a growing density of vehicles. A few studies also reported a direct formation of HNCO in the diesel engines during fuel combustion without any after-treatments (Heeb et al., 2011; Jathar et al., 2017). A tabular overview of past studies for HNCO emissions related to gasoline or diesel exhausts can be found in Wren et al. (2018) and Leslie et al. (2019). HNCO emissions via fossil fuel usage are not limited to on-road activity. Off-road fossil fuel activities (e.g. tar sands) also contribute to significant HNCO emissions on regional scales (Liggio et al., 2017). Finally, secondary HNCO formation in the atmosphere is also known through the oxidation of amines and amides (e.g. Borduas et al., 2016a; Parandaman et al., 2017).

The number of studies examining HNCO gas-phase chemistry is limited and mostly focused on its role in the chemistry in NOx mitigation strategies in combustion systems. The scarce data suggest that HNCO destruction in the atmosphere by typical pathways such as reactions with oxidizing agents or by photolysis is ineffective. We give a short overview here to supplement a recent review (Leslie et al., 2019). The reaction of HNCO with the hydroxyl radical (OH), the most important daytime oxidizing agent, has only been studied experimentally at temperatures between 620 and 2500 K (Baulch et al., 2005; Mertens et al., 1992; Tsang, 1992; Tully et al., 1989; Wooldridge et al., 1996), where the extrapolated rate expressions lead to an estimated rate coefficient of 5–12×10-16 cm3molecule-1s-1 at 298 K, i.e. a HNCO lifetime towards OH of over 25 years when assuming a typical OH concentration of 1×106 moleculecm-3. Early theoretical work by Sengupta and Nguyen (1997) at temperatures ≥500 K showed that the mechanism proceeds predominantly by H abstraction, forming NCO+H2O, with an energy barrier of ∼6 kcalmol-1. Wooldridge et al. (1996) determined an upper limit of ≤0.1 for the fraction of CO2+NH2 formation. To our knowledge, no experimental or theoretical data are available on HNCO reactions with other dominant atmospheric oxidants, including the nitrate radicals (NO3), chlorine atoms (Cl), or ozone (O3). Some data are available for H- and O-atom co-reactants of importance in combustion, as well as estimates for HCO and CN (Baulch et al., 2005; Tsang, 1992), but these are not reviewed here. There is no direct measurement for the dry deposition of HNCO. In a global chemical-transport-model-based study, the deposition velocity was considered to be similar to formic acid, yielding a HNCO lifetime of 1–3 d (over the ocean) to 1–2 weeks (over vegetation) (Young et al., 2012). The UV absorption for HNCO is only reported at wavelengths <262 nm, and photolysis is mostly reported for energies at wavelengths below 240 nm by excitation to the first singlet excited states, forming H+NCO or NH+CO (Keller-Rudek et al., 2013; Okabe, 1970; Spiglanin et al., 1987; Spiglanin and Chandler, 1987; Uno et al., 1990; Vatsa and Volpp, 2001). In the troposphere photolysis occurs only at the UV absorption wavelength band >290 nm due to filtering of shorter-wavelength radiation (Hofzumahaus et al., 2002). DrozGeorget et al. (1997) have reported the photolysis of HNCO forming NH(a1Δ)+CO(X1∑+) at 332.4 nm, but the HNCO absorption cross section at this wavelength would lead to a lifetime of months (Roberts et al., 2011). Therefore, HNCO loss due to photo-dissociation appears to be negligible in the lower atmosphere. HNCO has absorption bands in the infrared (Sharpe et al., 2004), but at these wavelengths the photon energy is generally too limited for photo-dissociation (Hofzumahaus et al., 2002). The main atmospheric loss processes are considered to be the transfer to the liquid phase, followed by hydrolysis, and deposition. This process depends on the varying atmospheric liquid water contents, relevant temperatures, and pH of cloud droplets. Therefore, the gas-to-liquid partitioning, in the varying atmospheric properties, i.e. water content, temperature, and pH of cloud droplets, becomes important to determine the atmospheric fate of HNCO (Leslie et al., 2019). The gas-to-liquid partitioning has been described by the Henry's law coefficient KH (ranging from 20 to 26±2 Matm-1) and related parameters by a handful of studies (Borduas et al., 2016b; Roberts et al., 2011; Roberts and Liu, 2019). Based on a recent study (Barth et al., 2013), the lifetime of HNCO due to heterogeneous processes is known to be of the order of a few hours (in-cloud reactions) to weeks (aerosol deposition).

The emissions and sources of HNCO have been focused on by many past studies, but there remain large uncertainties in our understanding of HNCO removal processes, especially in gas-phase chemistry. This missing information on HNCO removal processes limits global models to predict HNCO with confidence. To alleviate the dearth of direct data and therefore improve the representation of HNCO in global models, we first provide a theoretical analysis of the chemical reactions of HNCO with the dominant atmospheric oxidants: OH and NO3 radicals, Cl atoms, and O3 molecules, including the prediction of each rate coefficient at atmospheric conditions. In a second step, these results are included in a global numerical chemistry and climate model to assess the impact of chemical loss of HNCO in competition against hydrolysis within cloud droplets and against deposition to the Earth's surface. Additionally, the model is used to provide an estimate of the relative importance of primary and secondary HNCO sources.

The radical reactions characterized above proceed by H abstraction, forming the NCO radical with an H2O, HNO3, or HCl co-product. Likewise, the ozonolysis reaction proceeds for a large part by H abstraction, forming NCO with a HO3 co-product that readily dissociates to OH+O2. Though NCO radical formation through these reactions is found to be negligibly slow in atmospheric conditions, this radical remains of interest in other environments. Examples include combustion chemistry, where it can be formed directly from nitrogen-containing fuels and where it is a critical radical intermediate in, for example, the RAPRENOx (RAPid REmoval of nitrogen oxides) NOx mitigation strategy which employs HNCO introduced in the combustion mixture through (HOCN)3 (cyanuric acid) injection (Fenimore, 1971; Gardiner, 2000). The NCO radical has also been observed in space (Marcelino et al., 2018). There is extensive experimental and theoretical information on the reactions of NCO radicals, e.g. tabulated in Tsang (1992), Baulch et al. (2005), and other works. To our knowledge, the rate coefficients of the reactions of NCO radicals with H2O, HNO3, and HCl have not been determined before, but Tsang (1992) has estimated a rate coefficient k(NCO+H2O)=3.9×10-19T2.1exp⁡(-3046K/T) cm3molecule-1s-1 based on the equilibrium constant and rate coefficient of the HNCO+OH reaction. Since the H–N bond in HNCO is quite strong, with a bond energy of ∼110 kcalmol-1 (Ruscic, 2014; Ruscic and Bross, 2019), it is expected that NCO can readily abstract a hydrogen atom from most hydrogen-bearing species to produce HNCO, and that H abstraction is the main reaction channel. Hence, despite that our potential energy surfaces do not include an exhaustive search of all possible reaction channels in the NCO radical chemistry, we expect that the single-channel H-abstraction rate predictions for NCO from H2O, HNO3, and HCl are sufficiently dominant that these rates are fair estimates of the total rate coefficients including all possible channels for each of these reactions.

The energy barriers for the NCO radical reactions with H2O, HNO3, and HCl, being 14, 7, and 4 kcalmol-1, respectively (see Fig. 1), follow the bond strength trend in these reactants, with D0(H-OH)=118 kcalmol-1, D0(H-NO3)=104 kcalmol-1, and D0(H-Cl)=103 kcalmol-1 (Luo, 2007; Ruscic et al., 2002). Figure 1 also shows that the NCO+H2O reaction is endothermic by 8 kcalmol-1, while the HNO3 and HCl paths are exothermic by -5 and -7 kcalmol-1, respectively. The predicted rate coefficients are then the following: kNCO+H2O(300K)=1.36×10-21cm3molecule-1s-1,kNCO+HNO3(300K)=3.37×10-17cm3molecule-1s-1,kNCO+HCl(300K)=1.39×10-14cm3molecule-1s-1,kNCO+H2O(300-3000K)=4.59×10-24T3.63exp⁡(-4530K/T)cm3molecule-1s-1,kNCO+HNO3(300-3000K)=7.18×10-26T4.21exp⁡(-1273K/T)cm3molecule-1s-1,kNCO+HCl(300-3000K)=3.73×10-20T2.63exp⁡(-662K/T)cm3molecule-1s-1. The indirect estimate of Tsang (1992) compares well to our prediction for NCO+H2O, reproducing our values within a factor of 15 at 1000 K and a factor of 3 at 2000 K, i.e. within the stated uncertainties. An analysis of the impact of the NCO reactions in combustion or non-terrestrial environments is well outside the scope of this paper, and reactions with other co-reactants not discussed in this paper are likely to be of higher importance, e.g. H abstraction from organic compounds or recombination with other radicals. In atmospheric conditions, the fate of the NCO radical is likely recombination with an O2 molecule, with a rate coefficient of k(298K)=1.3×10-12 cm3molecule-1s-1 (Manion et al., 2020; Schacke et al., 1974), leaving H2O, HNO3, and HCl as negligible co-reactants. Hence, the NCO radical will not affect the atmospheric fate of any of these compounds to any extent. Subsequent chemistry of the ⚫OONCO radical is assumed to be conversion to an ⚫ONCO alkoxy radical through reactions with NO, HO2, or RO2, followed by dissociation to NO+CO.

Global atmospheric simulations allow us to gain insights into the significance of the chemical loss processes of HNCO and its distribution. Table 1 shows the corresponding HNCO budget for both performed simulations. The full kinetic model including our theoretically predicted gas-phase chemical reactions of HNCO is detailed in Tables S1 and S2 of the Supplement. Figure 4 shows the mean seasonal surface mixing ratio of HNCO using the biomass burning emission factors by Koss et al. (2018). It can be observed that high levels persist in each season. In general, high HNCO levels occur in regions associated with frequent biomass burning activities. Regions with no biomass burning activities have low HNCO concentrations, mainly caused by free tropospheric entrainment from regions with higher concentrations. The global vertical profile of HNCO is well illustrated by that for January as given in Fig. 5, showing that the free troposphere contains about 81 % of the total HNCO mass. The gas-phase production via formamide differs greatly depending on the biomass burning emissions used. In the case of Kumar et al. (2018), significantly more formamide is emitted, leading to a higher production of HNCO in the gas phase. The hydrolysis of HNCO produces ∼120 Tgyr-1 of ammonia and thus contributing little to the global ammonia budget. Our estimate is a factor of 5–6 lower than the upper limit estimated by Leslie et al. (2019).

The model predictions for local OH radical concentrations range from 1.15×100 to 1.56×107 moleculecm-3, with a weighted atmospheric global average of 1.14×106 moleculecm-3; in the air parcel where the highest OH concentration is found this leads to a HNCO lifetime towards OH of more than 500 years when accounting for the temperature-dependent rate coefficient (∼276 K). In the planetary boundary layer, the highest OH concentration predicted is 7.6×106 moleculecm-3 at a temperature of 297.8 K, leading to a HNCO lifetime to OH of ∼6 years in that air parcel. The calculated average OH concentration of 1.2×106 moleculecm-3 in the boundary layer leads to lifetimes towards OH of about 40 years near the surface. For O3, Cl, and NO3 – with maximum oxidant concentrations of 1.0×1013, 7.8×105, 1.5×109 moleculecm-3 and atmospheric average concentrations of 1.0×1012, 2.0×103, and 1.1×107 moleculecm-3, respectively – even longer temperature-dependent lifetimes are found, exceeding 5000 years even in the air parcels with the most favourable co-reactant concentration and temperature. The relative contributions of the different co-reactants varies locally and temporally, and shorter lifetimes might occur locally when co-reactant concentration and temperature are at their most favourable, but it is clear that gas-phase chemical losses of HNCO are small. Only the reaction of HNCO with OH leads to some destruction of HNCO, while the other chemical sinks (O3, NO3, and Cl) are negligible. When compared to the major loss processes, however, all these loss processes are negligible on a global scale (see Table 1). Young et al. (2012) have a somewhat higher chemical loss via OH compared to our result, which is due to the higher rate constant used. Figure 2 shows the rate coefficient that would be required to allow for the gas-phase loss of HNCO by reaction with OH radicals to contribute 10 % of the total atmospheric sink, which is well outside the expected uncertainty of the theoretical kinetic rate predictions. It can therefore be robustly concluded that the gas-phase chemical sinks predicted and assessed in this study (OH, Cl, NO3, O3) are insignificant when compared to heterogeneous loss processes, confirming earlier assumptions. This is independent of the high uncertainty in the available biomass burning emission factors or missing road emission datasets.

As seen in Table 1 the major sinks are dry deposition and scavenging (heterogeneous losses), where the former contributes between 2520 and 2890 Ggyr-1 and the latter from 274 to 377 Ggyr-1, when using the emission factors by Koss et al. (2018) and Kumar et al. (2018), respectively. The results in this study are in a similar range as the modelling study by Young et al. (2012). These authors had lower total HNCO emissions and did not include formamide as a secondary source of HNCO. The lower total HNCO emissions could be explained by a different year simulated in that study and different biomass burning emission model approaches used. Young et al. (2012) also scaled their HNCO emissions to the HCN emissions by a factor of 0.3, whereas in this study actual measured emission factors are used. In our study, formamide contributes between 17 % and 70 % of the total HNCO emissions when using the biomass burning emission factors by Koss et al. (2018) and Kumar et al. (2018) respectively. Young et al. (2012) find a higher HNCO lifetime due to generally lower total heterogeneous loss terms (dry and wet deposition). The total dry deposition varies slightly depending on the biomass burning emission factor used (see Table 1). In both scenarios, most HNCO is deposited over the ocean. For biomass burning emission factors from Koss et al. (2018), this contribution 53 %, is significantly lower when compared to the simulation using emission factors from Kumar et al. (2018), where about 62 % of the total HNCO deposition is deposited over the ocean. The larger fraction of computed HNCO deposition over the ocean is a consequence of the much larger secondary HNCO production from formamide far from its source regions (continents). Young et al. (2012) found that the importance of both heterogeneous loss processes depends on the cloud pH. In the SCAV submodel, as used in this work, cloud droplet pH is calculated online and includes an explicit hydrolysis scheme for HNCO, whereas Young et al. (2012) used a simplified approach. The relative importance of dry deposition is higher in the simulation in which Young et al. (2012) calculated pH online, when compared to the findings in this study.

The atmospheric lifetime of HNCO is dominated by its heterogeneous loss processes, leading to an atmospheric lifetime of multiple weeks when accounting for all HNCO losses (chemical and heterogeneous), as opposed to a gas-phase lifetime in the free troposphere of about 50 years when calculated solely based on the chemical losses towards the four chemical oxidants described in this study. This long gas-phase lifetime and the fact that mainly surface sources are relevant indicate that atmospheric HNCO distribution is significantly affected by transport processes. Our simulations even show that HNCO is transported from the surface into the UTLS and that about 10 % of the total atmospheric HNCO mass is located in the stratosphere (see Fig. 5), with modelled concentrations of HNCO in the lower stratosphere of typically tens of parts per trillion by volume but reaching up to hundreds of parts per trillion by volume in tropical regions. In the chemical model, photolysis in the stratosphere was not taken into account. Thus, OH is the only significant stratospheric sink included, resulting in a stratospheric lifetime of more than 330 years. During the monsoon period, the total stratospheric HNCO mass increases from 15 Gg before to 20 Gg at the end of the monsoon season. Pumphrey et al. (2018) demonstrated that in 2015 and 2016, elevated levels of stratospheric hydrogen cyanide (HCN) can be linked to biomass burning emissions from Indonesian fires. Figure 5 shows the vertical profiles of HCN and HNCO over South East Asia well before (January) and after (November) the Indian monsoon. It becomes evident that, similar to HNCO in our simulations, tropospheric and stratospheric concentrations of HCN increase during the Indian monsoon period. In the performed simulations, the ratio between stratospheric HCN and HNCO is very similar throughout the year, indicating that HCN and HNCO are similarly affected by transport processes within this period. The combination of strong biomass burning events and strong vertical transport during the monsoon period leads to high HNCO concentrations in the UTLS, indicating that pollutants from biomass burning events could potentially influence stratospheric chemistry.

Figure 6 shows the number of days exceeding a daily mean HNCO concentration of 1 ppbv. Mainly regions impacted by biomass burning events have frequent concentrations above this threshold. When using 10 ppbv as a limit for toxic concentrations of HNCO, as proposed by the Swedish Work Environment Authority (SWEA, 2011), only a few days can be observed in which this limit is exceeded. The maximum number of days exceeding 10 ppbv is 10 d over Africa, compared to 120 d above 1 ppbv. It is important to take into account that this analysis is limited by the computational output available in this study, which has only daily averages. Therefore, it is expected that areas which frequently exceed daily averages of 1 ppbv are potentially areas in which peak HNCO can be observed above 10 ppbv throughout the day.

No correlation exists between the number of days exceeding 1 or 10 ppbv and road traffic emissions. This becomes evident since typical areas of high road traffic activities (i.e. USA and Europe) do not exceed daily averages of 1 ppbv (see Fig. 6). Road traffic activities occur on a smaller spatial scale than biomass burning events. The EMAC model used is not capable of representing, for example, inner-city road traffic activities, due to the spatial resolution of the model used (1.875∘ by 1.875∘ in latitude and longitude). Therefore, we are not capable of drawing any conclusion if 10 ppbv is exceeded regionally in densely populated areas, impacted by high traffic emissions.

The isocyanic acid molecule, HNCO, is found to be chemically unreactive towards the dominant atmospheric gas-phase oxidants, i.e. OH and NO3 radicals, Cl atoms, and O3 molecules. The reactions all remove HNCO predominantly by H abstraction and have low rates of reactions with k(298K)≤7×10-16 cm3molecule-1s-1, leading to chemical gas-phase lifetimes of decades to centuries. Yearly loss of HNCO towards these reactants is only ∼5 Ggyr-1 out of ∼ 3000 Ggyr-1 total losses. Removal of HNCO by clouds and precipitation (“scavenging”), with hydrolysis to ammonia, is also implemented in the global model and was found to contribute significantly more, ∼300 Ggyr-1, than the gas-phase loss processes. Still, these combined processes are overwhelmed by the loss of HNCO by dry deposition, which is removing ∼2700 Ggyr-1. These conclusions are robust against modifications of the emission scenarios, where two distinct sets of emission factors were used, incorporating HNCO formation from biomass burning, as well as anthropogenic sources such as formamide oxidation and road traffic. The inefficiency of gas-phase chemical loss processes confirms earlier assumptions; inclusion of the gas-phase chemical loss processes in kinetic models appears superfluous except in specific experimental conditions with very high co-reactant concentrations. The long gas-phase chemical lifetime (multiple decades to centuries) allows HNCO to be transported efficiently into the upper troposphere lower stratosphere (UTLS) demonstrating that surface emissions may impact the upper troposphere. Further research is necessary to identify the importance of strong biomass burning events coupled to strong vertical transport processes (i.e. monsoon systems) on the chemical composition of the UTLS.

On a global scale, the daily-average concentrations of HNCO rarely exceed 10 ppbv, which is the threshold assumed here for toxicity; the exceedances are mainly located in regions with strong biomass burning emissions. Average daily concentrations of the order of 1 ppbv are encountered more frequently, with about one-third of the year exceeding this limit. This suggests that local concentrations might peak to much higher values, e.g. in urban environments where road traffic emissions are highest, or in the downwind plume of biomass burning events, and could impact regional air quality. Such regional effects were not studied in the current work, as the resolution of the global model used here is not sufficiently fine grained.

Though not important for the atmosphere, we briefly examined the reactions of the NCO radical formed in the chemical reactions studied. The rate coefficients of the H-abstraction reactions with H2O, HNO3, and HCl suggest that these reactions might contribute in high-temperature environments, such as combustion processes.