Sources of atmospheric nitrous acid: State of the science, current research needs, and future prospects

Abstract

Nitrous acid (HONO) plays a key role in tropospheric photochemistry, primarily due to its role as a source of hydroxyl (OH) radicals via its rapid photolysis. OH radicals are involved in photooxidation processes, such as the formation of tropospheric O₃ and other secondary atmospheric pollutants (peroxyacetyl nitrate [PAN] and secondary particles). Recent field and modeling studies have postulated the occurrence of a strong and unknown daytime HONO source, but there are still many significant uncertainties concerning the identification and formation mechanisms of these unknown sources. Up to now, five HONO formation pathways are known: direct emission, homogeneous gas-phase reactions, heterogeneous reactions, surface photolysis, and biological processes. In this review paper, the HONO sources proposed to explain the observed HONO budget, especially during daytime, are discussed, highlighting the knowledge gaps that need further investigation. In this framework, it is crucial to have available accurate and reliable measurements of atmospheric HONO concentrations; thus, a short description of HONO measurement techniques currently available is also reported. The techniques are divided into three basic categories: spectroscopic techniques, wet chemical techniques, and off-line methods.

Implications:

As important OH radical precursor, HONO plays a key role in the atmosphere. OH radicals are involved in photo-oxidation processes, resulting in the formation of tropospheric O₃ and secondary pollutants (PAN and secondary particles). The identification of species affecting the oxidation capacity of the atmosphere is crucial for the understanding of the tropospheric chemistry and in determining effective pollution control strategies. Current models underestimate HONO, especially during daytime, since sources are lacking. Although significant uncertainties still exist, some additional sources contributing to HONO budget have been discussed. Advantages and limitations of the currently HONO measurement techniques are also reported.

Introduction

Since the late 1970s, nitrous acid (HONO) has been identified as atmospheric key species due to its role of direct source of hydroxyl (OH) radicals (Perner and Platt, 1979). The OH radicals are involved in the formation of ozone (O₃) and peroxyacetylnitrates (PANs), causing the so-called “photochemical smog” in polluted regions and contribute to the oxidation of volatile organic compounds (VOCs), forming secondary oxygenated gaseous and particulate species. In the past, the HONO photolysis (R1; 300 < λ < 405 nm) was considered a dominant source of OH radicals in the early morning, when other primary HO_x (OH + HO₂) sources, such as the photolysis of O₃ and formaldehyde (HCHO), are still weak.

R1

Recent field and modeling studies have shown that R1 contributes significantly to the OH budget not only in the early morning, but also during daytime, accounting on average up to 60% of OH production in the boundary layer in some cases (Alicke et al., 2002; Kleffmann et al., 2005; Kleffmann, 2007; Elshorbany et al., 2009). HONO concentrations show a vertical gradient (depending on its sources); thus, HONO contribution to OH production is affected by altitude. Modeling studies indicated that HONO contribution to HO_x production is more relevant near the surfaces (where HONO is mainly formed), whereas O₃ contribution is higher at elevated levels. (Czader et al., 2013; Rappenglück et al., 2014).

Alvarez et al. (2013) showed that the HONO photolysis may be an important indoor OH source under certain conditions (such as direct solar irradiation inside the room), causing indoor OH concentrations similar to the urban outdoor OH levels.

HONO is also proposed as a toxic and harmful pollutant. Short-term exposures to elevated HONO concentrations may damage mucous membranes, the respiratory system of asthmatics (Rasmussen et al., 1995). HONO is also precursor of the mutagenic and carcinogenic nitrosamines by its reaction with secondary and tertiary amines (Pitts et al., 1978; Sleiman et al., 2010).

In the last few decades, field experiments have been carried out at remote, rural, and urban locations, reporting atmospheric HONO concentrations to range from a few ppt in remote and clean areas (Beine et al., 2001; Zhou et al., 2001; Honrath et al., 2002; Liao et al., 2006; Villena et al., 2011b) up to 15 ppb in a polluted urban site (Elshorbany et al., 2009, 2010; Spataro et al., 2013). Recent studies highlighted a daytime unknown HONO source, having strength in the order of 90, 500–600, and 2 ppb hr⁻¹ in polar, rural, and urban environments, respectively (Kleffmann et al., 2005; Acker et al., 2006; Villena et al., 2011b; Wong et al., 2011; Li et al., 2012; Spataro et al., 2013). Rappenglück et al. (2014) daytime HONO source strength reaching up to 10 ppb hr⁻¹ near snow surface at relatively low altitudes and under extreme conditions.

The formation mechanisms of HONO in the atmosphere are still under discussions. Kleffmann (2007) reported that HONO is mainly produced by heterogeneous processes, such as the NO₂ conversion on different surfaces. However, the HONO sources, especially during daytime, are still elusive.

This article reviews the recent progresses in understanding of HONO formation and sink processes in the troposphere, especially during daytime. The contribution of different reactions to the HONO budget is discussed, evaluating the cognitive gaps related to this pollutant. Finally, recommendations of further researches needed to better understand and define the HONO budget are also reported.

HONO Formation Rate and Photostationary State (PPS)

The atmospheric HONO concentrations are affected by several processes, which can cause its formation or removal in the troposphere. The HONO formation rate can be expressed by the following equation, including HONO source/production (P_i) and losses (L_i) processes (Spataro et al., 2013):

E1

E1 includes three terms. The first one, ,

indicates the HONO sources including the known HONO sources (P_{i, known}) and the missing daytime HONO sources (P_unknown). The second one, , indicates the HONO losses including three processes: the R1 (L_photo) and R2 (L_HONO+OH) reactions and the dry HONO deposition (L_dep). The third term, , includes the vertical (T_V) and horizontal (T_h) transport processes (Su et al., 2008b, Sörgel et al., 2011, Spataro et al., 2013). According to the modeling analysis by Czader et al. (2012), T_v and T_h can reach appreciable amounts. T_V and T_h can be sources or sinks depending on the HONO concentration in advected air compared with the HONO concentration at the measurement site (and height) (Sörgel et al., 2011). If HONO is formed at the ground or near surface aerosol, T_V means a sink because vertical mixing dilutes HONO formed near the ground.

R2

R1 and R2 affect the daytime HONO lifetime being in the order of 10–20 min under typical conditions.

L_dep can be estimated by dividing the dry deposition velocity (Harrison et al., 1996) with the boundary layer height. T_V can be estimated as in Dillon et al. (2002). During daytime, when more turbulence occurs in the atmosphere and L_photo is high, both L_dep and T_V are small compared with L_photo and they can be neglected (Su et al., 2008b, Sörgel et al., 2011, Spataro et al., 2013). During nighttime, the boundary layer is stable and L_photo is small. Both L_dep and T_V cannot be neglected, because this would result in the P_unknown underestimation (Sörgel et al., 2011). Su et al. (2008b) reported that T_h can be ignored assuming weak horizontal transport effects of HONO. This last simplification is reliable in a relatively homogeneous atmosphere with low wind speeds (Su et al., 2008b, Sörgel et al., 2011).

Assuming that the daytime HONO concentrations reach an instantaneous photoequilibrium around midday, as a photostationary state (PSS approach; d[HONO]/dt ≈ 0), the HONO production rate related to the unknown source (P_unknown) can be determined as follows:

E2

The missing HONO concentration ([HONO]_unknown) can be calculated by the following equation (E3):

E3

J_HONO is the photolysis frequency of HONO, k₂ (6.0 × 10⁻¹² cm³ sec⁻¹) is the rate constant of R2 at 298 K and pressure = 1010 hPa (Atkinson et al., 2004), and [OH] is the atmospheric OH concentration.

Great care must be exercised using PSS method to quantify the unexplained secondary HONO source strength. Since the model input data can vary significantly on short time scale, the PSS approach might not be correct (Kleffmann et al., 2005; Lee et al., 2013).

Formation Pathways of HONO

Despite the importance of HONO for atmospheric chemistry and its potential effects on the human health, the HONO sources are still not completely identified. There are five HONO formation pathways actually known: direct emission, homogeneous gas-phase reactions, heterogeneous reactions, surface photolysis, and biological processes (Kleffmann, 2007; Su et al., 2011). The major HONO source and loss pathways are reported in Figure 1. In the following, a short description and discussion of the presently established formation pathways is reported.

Figure 1. A schematic sketch over the chemical processes acting as sources and sink for nitrous acid (HONO) in the lower atmosphere. The processes occurring only in presence of solar radiation are reported in the yellow section, whereas the processes producing and removing HONO mainly during daytime are showed in the blue section. Finally, the gray section includes the sources and sinks of HONO operating similarly both during daytime and nighttime.

Direct emission

HONO can be directly emitted into the troposphere by combustion processes such as biomass burning, vehicle exhaust, domestic heating, and industrial burn. The quantitative determination of direct HONO emission was mainly focused to quantify the contribution related to vehicle tailpipes.

Pitts et al. (1984) found that cars with three-way catalyst emitted lower HONO and NO levels than station wagons equipped with leaded gasoline engine. The HONO to nitrogen oxides (NO_x) ratios (HONO/NO_x) ranged between 1 × 10⁻³ and 8 × 10⁻³, with higher ratios associated with older light-duty motor vehicles without emission control devices.

Kichstetter et al. (1996) determined HONO/NO_x ratios ranging between 2.4 × 10⁻³ and 3.4 × 10⁻³ at the Caldecott Tunnel (San Francisco Bay). These values were comparable to the HONO/NO_x ratios reported by Pitts et al. (1984). Similarly, Kurtenbach et al. (2001) measured simultaneously HONO, nitrogen dioxide (NO₂), nitrogen monoxide (NO), and carbon dioxide (CO₂) in the Wuppertal Kiesbergtunnel (Germany), determining the direct HONO emission from both single vehicles (such as a truck, a diesel passenger car, or a gasoline passenger car) and a traffic fleet. Higher HONO concentrations were measured during higher density traffic and were related to the using conditions of the single vehicles. The heterogeneous HONO formation on tunnel walls was investigated and determined to be strongly dependent on the residence time of an air parcel and on the surface-to-volume (S/V) ratio in the tunnel. The authors concluded that during high traffic density, direct HONO emission was more important than heterogeneous NO₂ into HONO conversion on tunnel walls. In contrast, during low traffic density, the HONO concentration was dominated by heterogeneous NO₂ conversion on tunnel walls. The calculated average HONO/NO_x ratio of (8 ± 1) × 10⁻³ was higher than that obtained by Kirchstetter et al. (1996), (2.9 ± 0.3) × 10⁻³. The difference can be related to the type of vehicles, which in the Kiesbergtunnel was mainly composed by diesel-fueled vehicles.

These traffic tunnel studies showed that on average only 0.3–0.8% of the total traffic-induced NO_x (NO + NO₂) can be attributed to direct emitted HONO, suggesting that direct traffic emission can contribute significantly to the tropospheric HONO budget only in areas with high traffic volume. The atmospheric HONO/NO_x ratios observed in the atmosphere are usually higher than those calculated for direct emission, indicating that HONO is mostly secondarily formed. Rappenglück et al. (2013, 2014) reported that current mobile emission models underestimate the HONO/NO_x ratios and found HONO traffic emissions up 1.7% of the total traffic NO_x emissions, likely due to heavy-duty diesel (HDD) vehicles, which previous tunnel studies may not have included, as there might have been some restrictions for HDD vehicles in those tunnels.

Homogeneous gas-phase reactions

Several gas-phase reactions have been proposed as HONO formation mechanisms, but kinetic studies indicated them to be not important HONO sources due to their small rate constants and competing daytime photolysis (Stockwell and Calvert, 1983; Tyndall et al., 1995; Mochida and Finlayson-Pitts, 2000; Finlayson-Pitts et al., 2003).

Up to now, only few homogeneous gas-phase mechanisms have been accounted for the atmospheric HONO concentrations. The most important homogeneous gas-phase reaction producing HONO is the reaction R3:

R3

R3 contributes mainly to the HONO budget during the daytime, when OH and NO concentrations are both high (Pagsberg et al., 1997; Alicke et al., 2002; Qin et al., 2009; Li et al., 2012), but it competes with HONO photolysis (R1).

Lee et al. (2013) used the PSS to explain the observed HONO, NO, and NO₂ concentrations at Houston and showed that the comparison between the modeled and measured concentrations might not be reliable. The PSS approach assumes that HONO concentrations reach an instantaneous photoequilibrium; however, the chemistry and dilution of boundary layer can affect the assumptions necessary to reach PSS. If the transport time for the pollutants from the source to the measurement site is lower than the time required to reach PSS, discrepancies between measured and modeled values can be recognized (Lee et al., 2013). The HONO concentration associated with this PSS can be determined by the following equation (E4), assuming d[HONO]/dt ˜ 0:

E4

E4 assumes R3 (P_OH+NO) as the only daytime HONO source, whereas R1 (L_photo) and R2 (L_HONO+OH) are HONO loss processes. Assuming the PSS approach, the HONO concentration at PSS ([HONO]_PSS) can be calculated (E5):

E5

[OH] and [NO] are the atmospheric OH and NO concentrations, respectively, J_HONO is the photolysis frequency of HONO, whereas k₂ (6.0 × 10⁻¹² cm³ sec⁻¹) and k₃ (9.8 × 10⁻¹² cm³ sec⁻¹) are the rate constants of R2 and R3, respectively, at 298 K and pressure = 1010 hPa (Atkinson et al., 2004). Kleffmann (2007) calculated the theoretical [HONO]_PSS of low ppt (<10 ppt) during the daytime. Field experiments reported daytime HONO concentrations in urban and rural areas to be substantially higher than the theoretical [HONO]_PSS (Kleffmann et al., 2005; Acker et al., 2006; Li et al., 2012; Spataro et al., 2013). Since this discrepancy could not be related to measurement and/or model errors, R3 can explain partially the observed daytime HONO and that other HONO sources, not included in E5, occur in the atmosphere.

Very high OH levels have been observed in Chinese environments, with daily maxima ranging between 15 × 10⁶ molecules cm⁻³ (0.610 ppt) and 26 × 10⁶ molecules cm⁻³ (1.057 ppt), and nighttime values ranging between 1 × 10⁶ molecules cm⁻³ (0.041 ppt) and 5 × 10⁶ molecules cm⁻³ (0.203 ppt) during the summer period. Therefore, in highly polluted Chinese environments, with significant amount of nighttime OH and NO concentrations, the contribution of R3 to the nocturnal HONO formation cannot be negligible (Qin et al., 2006; Hofzumahaus et al., 2009; Li et al., 2012; Lu et al., 2012, 2013; Zhang et al., 2012; Spataro et al., 2013).

Bejan et al. (2006) studied the HONO formation by the photolysis of different gaseous nitrophenols using two glass flow-tube reactors with significantly different surface-to-volume (S/V) ratios and volumes. HONO formation (at ppb levels) was observed when mixtures of nitrophenols (at ppm levels) were irradiated by six ultraviolet-visible (UV-Vis) lamps (300–500 nm). This HONO production was linearly correlated with the light intensity, nitrophenol concentration, and photolysis time, indicating that any intermolecular reaction between two nitrophenol molecules occurred (because a quadratic concentration dependency should result). Any dependence of HONO formation from S/V ratio was recognized, supporting that the process proceeded in the gas phase. Assuming that the ortho-nitrophenol concentration of 1 ppb was representative of an urban site, Bejan et al. (2006) estimated the HONO production rate of 100 ppt hr⁻¹. Thus, the photolysis of ortho-nitrophenols can contribute to the daytime HONO budget, but its strength source is not enough to explain the observed HONO concentrations. These results were obtained by extrapolating the laboratory data; thus, further field investigations should be needed to verify the reliability of this estimate.

Li et al. (2008a) suggested the reaction between electronically excited NO₂ (λ > 420 nm) and water vapor (H₂O) to form HONO and OH (R4–R6):

R4

R5

R6

When NO₂ absorbs radiation at wavelengths longer than ˜420 nm (λ > 420 nm), it forms electronically excited NO₂ (NO₂*) (R4), having fluorescence lifetimes of 40–60 msec. Most of this NO₂* will be quenched by collisions with any nonreactive molecule (M) that causes NO₂* to relax to a lower energy state (R5). The molecule M can be N₂ or O₂ or water vapor (H₂O) for which the quenching rate constants are known to be 2.7 × 10⁻¹¹, 3.0 × 10⁻¹¹, and ˜1.7 × 10⁻¹⁰ cm³ molecule⁻¹ sec⁻¹, respectively. Since the atmosphere is mainly composed by N₂ and O₂, most of the NO₂* will be quenched by collisions with these species. However, H₂O is a relatively abundant trace gas in the troposphere and some of NO₂* can collide with it. NO₂* can also react directly with H₂O producing HONO and OH (R6). Li et al. (2008a) determined a reaction rate for R6 of 1.7 × 10⁻¹³ cm³ molecule⁻¹ sec⁻¹. The HONO production rate, which is the same as OH production rate, can be expressed by the equation E6:

E6

where J₄ is rate constant for no dissociative excitation of NO₂, k₅ is the total collisional quenching rate constant of NO₂* by air, and k₆ is the rate constant for R6. Modeling studies, based on the rate constants reported by Li et al. (2008a), concluded that NO₂* chemistry can enhance the formation of oxidants in the urban areas (Sarwar et al., 2009; Ensberg et al., 2010).

Three- and one-dimensional modeling studies (Li et al., 2010, 2011; Czander et al., 2012, 2013; Gonçalves et al., 2012; Zhang et al., 2012) confirmed that R3 cannot explain the observed daytime HONO levels and that the comparison between measured and simulated HONO concentrations improved when additional HONO sources (direct emission, heterogeneous NO₂ conversion, and the reactions R4–R6) were included into the model. The results indicated that the HONO budget was only slightly affected by NO₂*, suggesting that it was not an important HONO source.

In conclusion, homogeneous gas-phase processes, such as R3 and the reaction between NO₂* and water vapor, act as atmospheric HONO sources, but they cannot explain the observed HONO concentrations. Laboratory investigations focused on the determination of the rate constant k₆ should be needed. Finally, the real contribution of photolysis of nitrophenols to the daytime HONO budget should be quantified.

Heterogeneous reactions

It is currently accepted that heterogeneous processes involving conversion of NO₂ on humid surfaces (R7) are the predominant formation pathways of nocturnal HONO:

R7

Laboratory studies indicate that the kinetics of R7 is likely to be first order in NO₂ and depends on various parameters, including the gaseous NO₂ concentration, the surface adsorbed water, relative humidity, surface-to-volume ratio (S/V), and surface properties (Jenkin et al., 1988; Kleffmann et al., 1998a; Finlayson-Pitts et al., 2003, and reference therein; Stutz et al., 2002, 2004). The surfaces available for reaction include airborne particles, mineral dust, soils, and urban surfaces such as glass, concrete, and foliage. These laboratory observations of NO₂ hydrolysis have in common the formation of three gas-phase products: HONO (the main gas-phase product), NO, and nitrogen protoxide (N₂O, in small amounts) (Finlayson-Pitts et al., 2003, and references therein). Nevertheless, the investigations on R7, its detailed reaction mechanism, and its dependence on surface water content, chemical properties (such as chemical composition and pH), and type of surface are still under discussions. Table 1 reports the main reaction mechanisms that have been proposed for R7.

Table 1. Main reaction mechanisms proposed for the heterogeneous NO₂ hydrolysis

Surface	Reaction Mechanism	Reference
Pyrex glass		Jenkin et al., 1988
Borosilicate glass		Finlyson-Pitts et al., 2003
Mineral dust		Gustafsson et al., 2009
Borosilicate glass		Ramazan et al., 2004

Jenkin et al. (1988) suggested that the reaction between NO₂ and water vapor, both adsorbed on the surface, formed the water complex NO₂·H₂O. NO₂·H₂O reacted with gaseous NO₂, producing gaseous HONO and adsorbed HNO₃. However, this mechanism was not consistent with other experimental results. Finlayson-Pitts et al. (2003) proposed the formation of the dimer of nitrogen dioxide (N₂O₄) on the reactive surface. N₂O₄ isomerizes to the asymmetric form, ONONO₂, which either can autoionize to generate the complex NO⁺NO₃⁻, or can react with gaseous NO₂ to form symmetric N₂O₄. The complex NO⁺NO₃⁻, reacts with H₂O, producing HNO₃ and HONO. HNO₃ on the surface generates NO₂⁺, which reacts with HONO, forming NO. HONO either can be released by the surface or can undergo secondary reactions to produce NO, NO₂, and small amounts of nitrous oxide (N₂O). Studies performed by Gustafsson et al. (2009) indicated that the HONO formation acid on mineral surfaces through reaction R7 does not involve an N₂O₄ intermediate. Ramazan et al. (2004) investigated the HONO formation from NO₂ hydrolysis in a borosilicate glass cell in the presence and absence of UV radiation (320–400 nm). They proposed a similar mechanism to Finlayson-Pitts et al. (2003), eliminating the NO₂ adsorption on the surface (not observed) and including two more reactions: the competitive adsorption between water vapor and HONO on the surface, which caused HONO desorption from the surface, and the reaction between HONO and HNO₃, producing NO₂. The results reported by Ramazan et al. (2004) highlighted that R7 is not photoenhanced; thus, it proceeds through the same mechanism in light and dark conditions. Both Finalyson-Pitts et al. (2003) and Ramazan et al. (2004) performed their experiments at high gaseous NO₂ concentrations (40–100 ppm), but their proposed mechanism could not explain the observations at low NO₂ concentrations (Kleffmann et al., 1998a; Kleffamann, 2007). Kleffmann et al. (1998a) investigated the heterogeneous NO₂ conversion on different humid surfaces and measured the observed uptake coefficient (γ), which is defined as the net probability that a molecule undergoing a gas-kinetic collision with a surface is actually taken up at the surface. The γ value is function of the mass accommodation, gas- and liquid-phase diffusion, Henry’s law solubility, and liquid and surface chemistry (Crowley et al., 2010). Kleffmann et al. (1998a, 1998b) determined γ values for NO₂ of about 10⁻⁶ and 10⁻⁷ for surface with adsorbed pure water and acidic solutions, respectively.

In order to account R7 for the HONO budget, the average nighttime conversion frequency from NO₂ into HONO (C_HONO) can be estimated by the following equation (E7), where is the average NO₂ concentration in the time interval of (t₂ − t₁) (Alicke et al., 2002, 2003; Su et al., 2008a; Li et al., 2012):

E7

Assuming that the conversion frequency of NO₂ into HONO is the same in the night, the daytime HONO production rate (expressed as ppb hr⁻¹) can be determined by the equation E8:

E8

Su et al. (2008a) suggested combinations of scaling methods in deriving the C_HONO to reduce the uncertainties in emissions and diffusion processes. Modeling studies, including homogeneous gas-phase reactions, direct emission, and heterogeneous reactions at the aerosol and ground surfaces (R7) as HONO sources, showed that R7 and direct emission could explain the observed nighttime HONO concentration, but other HONO sources prevailed during daytime (Vogel et al., 2003; Wong et al., 2011). There are also discussions about the prevalence of aerosol or ground surfaces in the heterogeneous HONO production. Although the γ values cannot be transferred directly from laboratory to other conditions, Kleffmann et al. (1998a) used γ ≈ 10⁻⁶ to estimate the heterogeneous HONO formation from NO₂ conversion on atmospheric aerosol, by the equation E9:

E9

where S/V is the surface-to-volume ratio (9 × 10⁻⁶ cm² cm⁻³ for normally polluted areas and 3 × 10⁻⁴ cm² cm⁻³ for highly polluted areas) and is the mean NO₂ molecular velocity. The HONO formation rate () was calculated as half (considering the stoichiometry of R7) of NO₂ conversion rate (%NO₂/hr), resulting of 0.5% hr⁻¹ and 1.5%/hr⁻¹ for normally and highly polluted areas, respectively. Thus, in the case of high aerosol load, significant HONO amount can be formed by heterogeneous NO₂ conversion on aerosol surface. Field measurements of vertical gradients and fluxes of HONO suggested that HONO can be significantly produced not only close to the ground, but also at higher altitudes of the boundary layer (Harrison and Kitto, 1994; Harrison et al., 1996; Reisinger, 2000; Kleffmann et al., 2003; Stutz et al., 2002; Villena et al., 2011a, 2011b; Wong et al., 2011, 2012; VandenBoer et al., 2013).

The reaction of NO and NO₂ on humid surfaces (R8) was also proposed as HONO source, but laboratory studies showed that its contribution to the atmospheric HONO budget cannot be significant (Kleffmann et al., 1998a):

R8

Later, the heterogeneous reaction between HNO₃ and NO on glass surfaces (R9) was suggested as atmospheric HONO source and as of importance for the renoxification of the boundary layer:

R9

The HONO and NO₂ formations were not observed when HNO₃ and NO concentrations were close to atmospheric levels (Kleffmann et al., 2004). The upper limit of the γ value for NO resulted to be <4 × 10⁻¹¹, indicating that R9 cannot be an important HONO source. These results confirmed previous field and laboratory studies indicating that the heterogeneous HONO formation is unlikely to involve high NO concentrations (Harrison and Kitto, 1994; Kleffmann et al., 1998a).

Ammann et al. (1998) proposed the interaction between NO₂ and soot particles to explain the observed HONO concentrations in air masses where combustion sources contribute to soot and NO_x emissions. The authors reported a γ value of 3.3 × 10⁻⁴ for the first 30 sec of the reaction. Kleffmann et al. (1999) investigated the heterogeneous NO₂ conversion on different carbonaceous surfaces (freshly prepared flame soot and commercial soot treated or not treated with sulfuric acid) and showed that the mean initial γ value of 10⁻⁶ decreased to <10⁻⁸ with NO₂ consumption increasing. The interaction between NO₂ and the soot surface caused the decrease of the number of active sites on the surface; thus, without any recycling mechanism atmospheric HONO formation on soot surfaces is not of major importance. The discrepancy between the results obtained by Kleffmann et al. (1999) and Ammann et al. (1998) was likely related to the different time scale of the experiments. Kleffmann et al. (1999) reported also that the HONO formation required the presence of water, since the γ values increased with increasing water vapor pressure. Aubin and Abbatt (2007) summarized the literature data for the initial uptake coefficient (γ₀) of NO₂ on various black carbon substrates, showing that they ranged from 0.1 to 10⁻⁸, but most of them were of the order of 10⁻⁵. The variance among these measurements can be related to initial NO₂ concentration, type of soot, the different time scale of the laboratory experiments, and the different method employed for the determination of the uptake coefficient. Indeed, the uptake coefficient can be calculated by the following equation (E10), where r is the radius of the flow reactor, k_w is the wall-loss rate constant, is the mean molecular speed of NO₂, A_SSA is the specific surface area of the sample, and A_geo is the geometric surface area:

E10

Many studies accounted only the geometric surface area in the γ₀ calculation, and since A_geo can be >3 orders of magnitude lower than A_SSA, the term γ₀ was overestimated.

Han et al. (2013a) investigated the composition of soot aerosol and its influence on NO₂ uptake coefficient and HONO yield, showing that (1) preheated samples exhibited a greater decrease in NO₂ uptake coefficient and HONO yield than the fresh soot samples, due to the removal of organic carbon from soot; and (2) ozonized soot showed a lower reactivity toward NO₂ than fresh soot, due to the increase of oxidation state of the soot surface. When fresh soot is emitted in the atmosphere, it undergoes aging processes by the uptake of reactive gases, leading to an increase of soot surface oxidation and thus to a lower reactivity of soot towards NO₂.

The semivolatile and/or water-soluble organic species in diesel exhaust were also suggested to be significantly involved in secondary HONO formation (R10; Gutzwiller et al., 2002a):

R10

{C–H}_red indicates surface-bound species or class of compounds. Diesel exhaust contains a water-soluble class of compounds, condensing below 100 °C and able to reduce NO₂ to nitrite (NO₂⁻) in the aqueous phase or to HONO on a surface. {C–H}_red indicates surface-bound species or class of compounds. Diesel exhaust contains a water-soluble class of compounds, condensing below 100 °C and able to reduce NO₂ to nitrite (NO₂⁻) in the aqueous phase or to HONO on a surface. Kirchstetter et al., 1996; Kurtenbach et al., 2001). The NO₂ reduction in aqueous solutions by organic species (resorcinol, 2,7-dihydroxynaphthalene, guaiacol, syringol, and catechol) was recognized to be enhanced in alkaline conditions (Gutzwiller et al., 2002b; Ammann et al., 2005). Although R10 seems to contribute significantly to the atmospheric HONO, there are questions to be resolved. First, laboratory experiments focused on the determination of dependence of the reaction rate from the surface water content is still lacking. Then, field studies aimed to identify the organic compounds, which are involved in NO₂ reduction and to assess its real contribution of reaction R10 to the HONO budget, should be performed.

Ziemba et al. (2010) used PSS assumption, suggesting heterogeneous HONO production due to the HNO₃ depletion on hydrocarbon-like organic aerosol (HOA) surface. However, Lee et al. (2013) showed that the PSS assumption might not be valid and might lead to overestimation of secondary HONO source strength. Additionally, different air mass, vertical mixing, and parameters such as HONO/NO_x ratio, deposition velocity, and uptake coefficients could contribute to the uncertainty of PSS calculation. Rutter et al. (2014) showed that HONO formation rate, due to the HNO₃ reduction into HONO by volatile organic compounds (VOCs), ranged between 0.1 and 0.5 ppb hr⁻¹, but that it was not affected by the increasing of surface area. These results indicated that the reaction does not occur on aerosol surface and that, likely, HNO₃ reacts homogeneously with VOC emitted by the aerosol surface (R11):

R11

The correlation between HONO concentration and aerosol surface, recognized by Ziemba et al. (2010), can be explained by the simultaneous emission of VOCs and aerosol by vehicle exhaust. However, more laboratory investigations are required to assess if the HONO formation observed by Ziemba et al. (2010) was heterogeneous or homogeneous in nature. Laboratory experiments involving aerosol loading comparable to that observed by Ziemba et al. (2010) and evaluating the dependence of the HNO₃ conversion from relative humidity and water content of aerosol surface should be needed. Up to now, it is not possible to exclude that the HNO₃ into HONO conversion might occur also on the organic aerosol surface.

In conclusion, heterogeneous processes are important atmospheric HONO sources. The ground is mainly involved in the HONO production, but high aerosol loading in the atmosphere can affect significantly the HONO budget. R7 can explain the nighttime HONO concentrations, but more research focused on the identification of this reaction mechanism as well as its dependence on surface water content, on chemical properties (such as chemical composition and pH), and on type of surface is needed. However, R7 cannot explain the observed daytime HONO levels. R10 is also a possible HONO source, but the following items should be evaluated: dependence from the surface water content, the involved organic compounds, and the real contribution of R10 to the HONO budget. Finally, the homogeneous and/or heterogeneous nature of R11 should be clarified.

Photolysis/photoenhanced surface reactions

Since direct emission and homogeneous gas-phase and heterogeneous reactions could not fully explain the observed daytime HONO concentrations, the scientific efforts have been focused on the photochemical/photoenhanced processes. In this framework, several studies have been performed on the photoenhanced NO₂ or HNO₃ conversion on organic and inorganic surfaces.

Photoenhanced reactions on soot/organic surfaces.

Heterogeneous reactions on organic aerosol surfaces have been widely reported in literature and several carbonaceous surfaces, such as soot, hydrocarbons, diesel soot, and humic acid, have been studied.

George et al. (2005) investigated the effect of light with 300 < λ <500 nm on the uptake kinetic of NO₂ of different organic films (such as benzophenone, catechol, anthracene, anthrarobin, and their mixtures) occurring ubiquitously on aerosol and ground surfaces. The NO₂ uptake coefficient determined under UV light for the different organic films are reported in Table 2 and ranged between 10⁻⁷ and 10⁻⁶, with HONO yields ranging between 50% and 100% under UV irradiation comparable to the irradiance in the wavelength interval of 300–420 nm at the Earth surface under 0° zenith angle.

Table 2. NO₂ uptake coefficients (γ_NO2) of different organic surfaces under UV irradiation

	RH	λ	[NO2]
Surface	(%)	(nm)	(ppb)	γNO2	Reference
Humic acid (soil)	20	300–700	20	2 × 10^−5	Stemmler et al., 2006
Humic acid (coating)	26	400–750	˜10	˜1.5 × 10^−5	Stemmler et al., 2007
Humic acid (aerosol)	26	400–750	˜10	˜5.5 × 10^−6	Stemmler et al., 2007
Benzophenone/catechol	76	300–420	20	5.1 × 10^−6	George et al., 2005
Pyrene/KNO3		300–420	30	2.7 × 10^−6	Ammar et al., 2010
Benzophenone/catechol	14	300–420	20	2.5 × 10^−6	George et al., 2005
Benzophenone/catechol	Dry	300–420	20	2.4 × 10^−6	George et al., 2005
Anthracene	56	300–420	20	2.3 × 10^−6	George et al., 2005
Fluoranthene surface	<5	300–420	25	2.0 × 10^−6	Cazoir et al., 2014
Catechol	42	300–420	20	1.8 × 10^−6	George et al., 2005
Catechol	Dry	300–420	20	1.4 × 10^−6	George et al., 2005
Catechol/anthracene	46	300–420	20	1.3 × 10^−6	George et al., 2005
Anthracene	Dry	300–420	20	1.2 × 10^−6	George et al., 2005
Benzophenone	Dry	300–420	20	6.5 × 10^−7	George et al., 2005
Soot surface	30	300–420	150	5.0 × 10^−7	Monge et al., 2010a

Notes: RH = relative humidity; [NO₂] = concentration of NO₂; λ = wavelength.

George et al. (2005) concluded that this process may contribute to the daytime HONO budget and noted that the proposed mechanism can occur continuously during daytime. Similarly, Stemmler et al. (2006, 2007) proposed the photosensitized NO₂ reduction on humic acid, representing the complex unsaturated organic materials ubiquitously present in the environment (R12–R14):

R12

R13

R14

The mechanism proposed for this process involves the photochemical activation of reductive sites (A^red) on the organic surface (R12). The reductive sites may undergo two different fates: their deactivation (R13) or the reaction with gaseous NO₂ to form HONO. The X species is an oxidant formed on the organic surface. However, it is not clear the exact chemical nature of the reductive species formed on the humic acid surface. The upper limit of NO₂ uptake coefficients (γ_NO2) for humic acid coating and aerosol were lower at higher NO₂ concentrations, and when the relative humidity (RH) was lower than 20% and higher than 60% (Stemmler et al., 2007), the γ_NO2 values for humic acid coatings were 3 times higher than those observed for humic acid aerosol, suggesting that HONO formation on ground surface is higher than those on aerosol surface. The daytime HONO production rate from soils containing humic acid (700 ppt hr⁻¹ within a boundary layer of 100 m height) was estimated to be comparable to the unknown HONO sources (90, 500–600, and 2 ppb hr⁻¹ for polar, rural, and urban environments, respectively) (Stemmler et al., 2006). Monge et al. (2010b) showed the persistence of the photoenhanced NO₂ into HONO conversion on soot over long periods. Thus, soot transported away from a polluted environment might provide a local photochemical source of NO and HONO. However, the initial γ_NO2 coefficients of 5 × 10⁻⁷ and 5 × 10⁻⁸ were determined at initial NO₂ concentrations of 16 and 120 ppb, respectively. More recently, Han et al. (2013b) showed that simulated sunlight enhanced the aging process of soot toward NO₂, resulting in the production of nitro compounds on soot surface. The photolysis of nitro compounds resulted in carbonyl groups and gaseous NO and HONO. Soot samples with higher organic carbon concentration exhibited higher photochemical reactivity. This fraction of organic carbon was composed by polycyclic aromatic hydrocarbons (PAHs) with 3–5 rings and some unidentified components. However, the low γ_NO2 coefficients indicated that light-induced NO₂ conversion on soot aerosol surface cannot be an important HONO source even if the soot surface may remain chemically active for long periods. These laboratory experiments investigated the interaction between NO₂ and soot under UV-A radiation, whereas the effect of visible and UV-B as well as the changes of the chemical and physical properties of soot surface have not been evaluated yet.

Ammar et al. (2010) performed laboratory experiments on the heterogeneous NO₂ reaction with solid pyrene/KNO₃ films (as a proxy of urban grime), determining λ_NO2 coefficient of 2.67 × 10⁻⁶ under near-UV irradiation (300–420 nm) and estimating the HONO production rate of 130 ppt hr⁻¹.

Recently, Cazoir et al. (2014) studied the heterogeneous loss kinetics of gaseous NO₂ on solid fluoranthene films deposited on a Pyrex substrate under UV-A (range 300–420 nm) radiation. The uptake coefficients decreased from 2 × 10⁻⁶ to 7.0 × 10⁻⁷, for initial NO₂ concentrations of 20 and 150 ppb, respectively, and they were not affected by temperature and RH. These results suggested the photoenhanced NO₂ reactivity towards fluoranthene and possibly other PAHs. It would be interesting to extend this study to other PAH compounds to evaluate the real influence of this process on HONO formation and its atmospheric implications.

In conclusion, the photoenhanced NO₂ conversion on organic surfaces or humic acids can fully explain the observed daytime HONO concentrations. However, more research, focused on the organic chemicals participating to the HONO formation and the reaction mechanisms, should be performed. The effect of both UV-A and UV-B radiations on HONO production should be also assessed. Finally, field experiments evaluating the real contribution of these processes to the HONO budget are needed.

Photolysis reactions on inorganic surfaces.

During the last 30 years, several studies have been focused on the photolysis reactions on inorganic surfaces resulting in HONO production. The photolysis of adsorbed HNO₃ or nitrate (NO₃⁻) (λ ˜300 nm) was identified as a daytime HONO and NO_x source in rural environments, and the following mechanism has been identified (reactions R15–R17):

R15

R16

R17

R17 results in the formation of NO₂, which in turn can react with the water adsorbed on the surface forming gaseous HONO and HNO₃ adsorbed on the surface (R7). This process is 1–2 orders of magnitude faster than that in the gas phase and aqueous phase (Zhou et al., 2002, 2003).

Zhou et al. (2011) performed direct measurements of HONO fluxes over a rural forest canopy in Michigan, showing that the daytime HONO flux was positively correlated with the product of leaf surface nitrate loading and the rate constant of nitrate photolysis. They suggested that the photolysis of HNO₃ on forest canopies is a significant daytime HONO source to the troposphere in rural environments and an important pathway for the remobilization of deposited HNO₃. The authors explained that the remobilization of deposited HNO₃ causes the overlying troposphere to be more photochemically reactive than generally predicted, and the long-range transport of reactive nitrogen to impact potentially on greater distances than previously thought.

In contrast to heterogeneous HONO formation that mainly accelerates morning ozone formation, inclusion of HONO photochemical sources influences ozone throughout the day, affecting its peak concentration (Sarwar et al., 2008; Czader et al., 2012; Rappenglück et al., 2014). Li et al. (2012) showed that the photolysis of HNO₃ adsorbed on ground surface was related to HONO formation in the Pearl River Delta region, during the Program of Regional Integrated Experiments of Air Quality over Pearl River Delta 2006 (PRIDE-PRD2006) campaign. The importance of this mechanism is associated with its potential role of pathway for the remobilization of deposited HNO₃. A similar mechanism has been recognized at high polar latitudes during springtime (Zhou et al., 2001; Dibb et al., 2002; Honrath et al., 2002). This process starts with the deposition and migration of NO₃⁻ ion to the snow pack surface, followed by the photoreduction of NO₃⁻ to form NO, NO₂ and nitrite (NO₂⁻) ion, and the NO₂⁻ acidification producing HONO. Several studies showed that NO_x and HONO released from the snow surface at high and mid latitudes can alter significantly the chemistry in the remote boundary layer, because the radical pool is altered (Ianniello et al., 2002; Beine et al., 2003; Rappenglück et al., 2014). However, more field experiments, including the analysis in situ of surface sample (chemical and physical properties) and vertical fluxes of HONO, NO, and NO₂, should be performed in order to better understand the impact of the reactions R15–R17 on the atmospheric HONO budget.

Research studies highlighted the simultaneous increase of NO₃⁻ in PM₁₀ and PM_2.5 (particulate matter with aerodynamic diameters of <2.5 and <10 μm, respectively) and gaseous HONO concentrations during Saharan dust events (Wang et al., 2003; Saliba and Chamseddine, 2012), suggesting that mineral dust affects the conversion of nitrogen species into HONO in the troposphere. Laboratory studies suggest that HONO can be formed by the UV-irradiated interaction of reactive nitrogen compounds (NO₂, NO, HNO₃) with mineral-containing substrates such as titanium dioxide (TiO₂), silicon oxide (SiO₂), aluminium oxide (Al₂O₃), calcium carbonate (CaCO₃), and real Saharan and Asian dust (Saliba et al., 2001; Usher et al., 2003, Cwiertny et al., 2008, Ndour et al., 2008, Ma et al., 2013). However, these studies were not conclusive regarding the understanding of reaction mechanisms and their dependence on several parameters: surface chemical properties (acidity, water content, and composition), relative humidity, concentration and type of nitrogen species, and wavelength and intensity of radiation. Ramazan et al. (2004) investigated the effect of UV radiation on the heterogeneous NO₂ hydrolysis on borosilicate glass surface, and their results showed that R7 is not photoenhanced. Their experiments indicated also that HNO₃, resulting from R7, can exist either as nitric acid–water complexes (HNO₃·H₂O) or as NO₃⁻ ion. HNO₃·H₂O can undergo photolysis (λ < 741 nm), producing adsorbed HONO and hydrogen peroxide (H₂O₂). The release of this HONO depends on the water vapor concentration, which triggers a competitive adsorption process (Ramazan et al., 2006; Ma et al., 2013):

R18

R19

Since SiO₂ is the major component of mineral dust, laboratory and field experiments aimed to verify the heterogeneous NO₂ hydrolysis on mineral dust and the following photolysis of HNO₃·H₂O or NO₃⁻ ion are still lacking. The laboratory experiments should evaluate also how gaseous HONO production from mineral dust surface is affected by RH and surface water content.

Many research studies have been focused on the NO₂ removal by titanium oxide (TiO₂) surfaces. Although TiO₂ is found in dust particles at mass concentration ranging from 0.1% to 10%, this species has attracted great interest due its well-known photocatalytic properties. The irradiation of TiO₂ molecules by UV lights cause the photoproduction of excess electrons in the conduction band (e⁻_cb) and holes in the valence band (h⁺_vb) (reaction R20). The hole oxidizes water vapor (R21), whereas the electron reduces the oxygen (R22):

R20

R21

R22

R23

R24

When gas-phase NO₂ diffuses on TiO₂ surface, it can react with OH radical to form HNO₃, or with the oxygen activated species (O₂⁻) to produce NO₂⁻ ion (Ndour et al., 2008; Monge et al., 2010a). Ndour et al. (2008) performed laboratory studies, exposing mixed TiO₂/SiO₂ and Saharan and Arizona test dust to gaseous NO₂. TiO₂/SiO₂ mixtures (with TiO₂ and SiO₂ contents of 1 wt% and 5 wt%, respectively) were related to uptake coefficients ranging from 1.2 × 10⁻⁷ to 1.9 × 10⁻⁶. The small TiO₂ amount was enough to make the surface photochemically active. Higher uptake coefficients were related to increased TiO₂ content in the TiO₂/SiO₂ mixtures (Monge et al., 2010a). These uptake coefficients ranged from (2.0 ± 0.6) × 10⁻⁷ to (1.5 ± 0.4) × 10⁻⁶ for TiO₂ content of 10 wt% and 100 wt%, respectively. The results showed that the presence of O₂ in the carrier gas affects the HONO production yields in the heterogeneous reaction between NO₂ and TiO₂ as well as the products of the renoxification process. However, more kinetic data should be collected in order to improve understanding of this process, and to include it into modeling studies. The influence of parameters such as RH, surface water content, intensity and wavelength of radiation, and changes on the substrate have to be considered for future studies.

Biological processes

Su et al. (2011) proposed that the unknown HONO source was related to soil biological processes. They reported that the dominant source of N(III) in soil are biological nitrification and denitrification processes. The nitrification process converts ammonium ion (NH₄⁺) to NO₂⁻, which in turn is converted to NO₃⁻ resulting in the accumulation of both NO₂⁻ and NO₃⁻ in the soil. During the denitrification, NO₃⁻ is reduced to gaseous NO, which can be released from the soil surface (Su et al., 2011; Pilegaard, 2013; Oswald et al., 2013). Since NO₂⁻ is highly water soluble, it can be protonated by H⁺ (reaction R25) resulting in adsorbed HONO. Then, HONO can undergo the competitive adsorption with water vapor (R26), resulting in the release of gaseous HONO from the soil.

R25

R26

The equilibrium gas-phase HONO concentration over an aqueous solution ([HONO]*) can be determined by the following equation (E11), where [H⁺] is the hydrogen ion concentration, [N(III)] is the total nitrite concentration ([HNO₂]+[NO₂⁻], K_a,HNO2 is acid dissociation constant for nitrous acid, and H_HONO is Henry’s law coefficient of nitrous acid:

E11

[HONO]* controls the gaseous HONO amount that can be released to the atmosphere, and its value is affected by pH of the soil, nitrite concentration in the soil, and temperature (which affects both K_a,HNO2 and H_HONO). If [HONO]* is higher than gaseous HONO concentration, then HONO will be released from the soil. Thus, the potential impact of this process is related to the use of the soil, such as agriculture, fertilization, aquatic transport, deposition of atmospheric pollutants, waste, etc. Su et al. (2011) evaluated the functional dependence of [HONO]* on [N(III)] and pH, suggesting that neutral or alkaline soils (high pH) were characterized by higher N(III) concentrations, whereas acid soils had low N(III) concentrations. These results were in agreement with their proposed mechanism, where acid soils favor the release of HONO. Su et al. (2011) used the following equation (E12) to determine the HONO emission fluxes from soil nitrite. ν_t is the transfer velocity, affected by meteorological data and soil conditions:

E12

The obtained fluxes were used to estimate atmospheric HONO production rate from soil (P_soil):

E13

where a is a factor converting ppb units into ng m⁻³ units (a = 572 ng m⁻³ ppb⁻¹ at 101,325 Pa 132 and 298 K) and H is the boundary layer height. Their results indicated that even low soil nitrite concentrations can cause emissive HONO fluxes of ˜1 to 1000 ng m⁻² sec⁻¹ for boundary layer heights of ˜100 to 1000 m.

Oswald et al. (2013) studied the relation between soil HONO emissions and biogenic soil NO emissions, using the dynamic chamber method and investigating soil samples, with a wide range of soil pH, organic matter, and nutrient contents. The results showed that maximum HONO (about 13 ng m⁻² sec⁻¹) and NO (about 18 ng m⁻² sec⁻¹) fluxes from soil samples were comparable and were affected by water content, pH, and temperature. However, unexpectedly high HONO and NO emissions were observed for neutral and alkaline soils, which was characterized by higher NH₄⁺ and NO₂⁻ concentrations. These results were in disagreement with the acid-base equilibria proposed by Su et al. (2011), where acid pH favors release of gaseous HONO. The authors suggested that beside the acid-base equilibria, ammonia-oxidizing bacteria (AOB) can dominate the formation of HONO and NO within the soil. Therefore, HONO is formed and emitted during nitrification, which is favored at lower water holding capacity (WHC). Maximum HONO and NO fluxes were measured within 0–40% of WHC.

HONO emission from soil can be considered as another potential source of this species in the troposphere, which is consistent with the observations that ground surfaces are dominant for HONO production. Further studies are recommended to assess the real impact of HONO release from soil nitrite on in the biogeochemical cycling of N in both agricultural and natural environments.

HONO Measurement Techniques

Accurate HONO measurements are difficult to perform. Certified reference gases for HONO do not exist, because it is unstable. HONO reactivity and solubility can cause sampling losses and artifacts. Consequently, intercomparisons between different techniques may exhibit significant discrepancies. During the international FIONA (Formal Intercomparisons of Observations of Nitrous Acid) campaign, a set of experiments was performed in the EUPHORE (European PHOtoREactor) chambers, simulating urban and semirural scenarios and intercomparing almost all types of instruments used to detect gaseous HONO. Good agreement among the techniques was found, but deviations for some measurements were also revealed. These differences were related to the difficult adaptation of the instruments to the EUPHORE chambers and to calibrations issues (Ródenas et al., 2013). However, the FIONA results are only preliminary and analyses are still ongoing.

Pinto et al. (2014) compared the HONO measurements made by different instruments during the SHARP (Study of Houston Atmospheric Radical Precursors) camping. The HONO temporal patterns obtained by these techniques were in good agreement. The largest differences among the measurements were related to (1) high pollution conditions causing positive and/or negative interference in in situ techniques; (2) the sampling site, where heterogeneous reactions can affect the HONO measurement; (3) the position of the instrumental inlet; and (4) other experiments performed at the sampling site causing differences among instruments.

In the following sections, an overview of HONO measurement techniques currently available in the literature is reported. They are divided in three basic categories: spectroscopic techniques, wet chemical techniques, and off-line methods.

Optical spectroscopic methods

The optical spectroscopic techniques measure atmospheric trace gases without the need for chemical extraction, with calibrations that are based on known absorption cross-sections (line strengths) and specificity that can be confirmed by spectral identification. However, absorption spectroscopic analytical methods tend to be expensive and to have bulky system components, and for many trace gases the fundamental sensitivity is relatively low, requiring either long absorption paths or increased signal averaging time (Kleffmann, 2007).

Differential optical absorption spectroscopy (DOAS).

The DOAS technique, schematically showed in Figure 2, has been widely used to detect atmospheric HONO in urban, rural, and semirural environments (Perner and Platt, 1979; Pitts et al., 1984; Andrés-Hernández et al., 1996; Reisinger, 2000; Stutz et al., 2002, 2004, 2010; Acker et al., 2006).

Figure 2. Schematic sketch of the major components related to the differential optical absorption spectroscopy (DOAS) system.

The DOAS system (Figure 2) includes a light source, emitting light radiation, which travels across the atmosphere for distances ranging from few hundred meters to several kilometers (Alicke et al., 2002; Kleffmann et al., 2006; Stutz et al., 2010; Zhou et al., 2013). At the end of this path, the radiation is then collected and detected (Platt and Stutz, 2008). The active DOAS is equipped with a broadband source, such as a Xenon arc lamp (e.g., long-path DOAS), whereas the passive DOAS uses sunlight or scattered sunlight as source (e.g., multiaxis DOAS). Since the radiation moves through the atmosphere, its intensity is reduced by absorption of a specific trace gas. Simultaneously, the light intensity can be also absorbed by other trace gases (extinction), or can be scattered by gaseous and particulate, and these processes cause interferences. The broad bands are related to instrument effects and turbulences, whereas the high-frequency narrow band cross-section is related to characteristic absorption lines or bands of trace gases. HONO is detected by its narrow-band UV absorption between 340 and 370 nm and is quantified by Beer-Lambert law. Detection limits (LOD) of 100 and 200 ppt related to path length of 750 and 6 m, respectively, have been reported (Kleffmann et al., 2006; Qin et al., 2006). Stutz et al. (2010) achieved the LOD of 16 ppt related to path length of 4–5 km. DOAS is an accurate and sensitive technique, allowing the simultaneous detection of different species (Platt and Stutz, 2008; Stutz et al., 2010). Since HONO measurements by DOAS are artifact-free with respect to the sampling procedures, DOAS is often used as reference during HONO intercomparison. However, the simultaneous measurement of different species at high concentration, especially NO₂, can cause large uncertainty for HONO quantification by DOAS (Kleffmann et al., 2006).

Cavity ring-down spectroscopy (CRDS).

The CRDS is a direct spectroscopic technique (detects directly the HONO molecule), which measures the rate of decay of the radiation due to the strong absorption of HONO at the wavelength of 354.2 nm. Two highly reflective mirrors (R > 99.9%) are used to trap tunable laser pulses inside an optical cavity, and the trapped light makes about 1000 rounds in the cavity. The decay rate (1/τ) can be determined by the following equation (E14), where (1/τ₀) is the background decay rate due to the mirror loss and the scattering, α is the absorption coefficient, n and σ are the number density and absorption cross-section of the absorbing sample, respectively:

E14

The detection limit (LOD) was determined to be in the order of 5 ppb, with a time resolution of 15 sec, but LOD in the order of 0.1 ppb could be reached by upgrading optics, a longer cavity, or other improvements. Absorptions in the cavity by other trace gases (NO₂, SO₂, or others) can affect the HONO determination by CDRS. This problem can be solved by the use of differential CDRS. However, no field application of this technique has been reported (Wang and Zhang, 2000).

Tunable infrared laser absorption spectroscopy (TIRLS).

TIRLS is based on absorption spectroscopy to measure directly atmospheric HONO without chemical extraction. Ambient air is sampled into a low-pressure multipass cell, where it interacts with light from an infrared (IR) laser. The calibrations are based on known absorption cross-sections (line strengths) and on spectral identification (Lee et al., 2011). Li et al. (2008b) used tunable diode laser absorption spectroscopy (TDLS) with a light absorption path of ˜150 m to detect HONO by its strong absorption feature at 1713.511 cm⁻¹. The detection limit was of ˜200 ppt, with a time resolution of 1 sec. Lee et al. (2011, 2013) developed a tunable IR laser differential absorption spectrometer equipped with continuous-wave quantum-cascade lasers. This instrument was able to measure both HONO and NO₂ concentrations. The continuous-wave quantum-cascade lasers allowed much more mode stability, higher laser power output, and the ability to operate both lasers and detectors near room temperature for long-term field measurements. Detection limits of 40 and 10 ppt for HONO and NO₂, respectively, were reported with 1-hr spectral averaging.

Photofragmentation (PF)/laser-induced fluorescence (LIF).

The determination of HONO by PF/LIF technique consists in its UV photofragmentation (λ = 355 nm) producing NO and OH, followed by the quantifications of OH by laser-induced fluorescence at an excitation wavelength of 282 nm and an emission wavelength of 309 nm. The detection limit is of 2–15 ppt, with a time resolution of 1 min. The excitation wavelength of 282 nm allows minimizing the interferences due to O₃. In particular, at high O₃ concentrations and high relative humidity, this interference needs to be quantified and corrected (Kleffmann, 2007). The detection limit is of 2–15 ppt, with a time resolution of 1 min. Liao et al. (2006) used PF/LIF to measure HONO at the South Pole (Liao et al., 2006), but this technique has not been used in more than 10 years.

Wet chemical methods

The wet chemical techniques operate by collecting gaseous HONO by several devices (wet effluent denuder, rotated denuder, bubbler), in which inner walls are wetted with suitable solutions. These sampling solutions are then analyzed by analytical techniques such as ion chromatography (IC), high-performance liquid chromatography (HPLC), spectrophotometry, etc. The wet techniques are simpler and less expensive compared with the spectroscopic techniques, providing higher sensitivities with detection limit in the order of few ppt. Although the wet chemical methods can be very sensitive, the need to scrub HONO into solution may introduce chemical interferences caused by, e.g., NO₂ and phenol reaction (Gutzwiller et al., 2002a, 2002b), NO₂ and SO₂ (Spindler et al., 2003), NO₂ and aromatic amines (Saltzman, 1954), or PAN hydrolysis (Frenzel et al., 2000). Thus, attention has to be paid to minimize the chemical interferences when wet chemical techniques are used to measure HONO concentrations.

Wet denuder.

A widely used wet denuder for HONO measurement is the wet effluent diffusion denuder (WEDD). It is set up vertically, in which the purified water (absorbing solution) is continuously pumped into the tube at the top forming a thin aqueous film on the inner wall surface. The adsorbing solution collects gaseous HONO, which is highly soluble in water, converting it into nitrite and analyzing by IC. The detection limit is of few ppt, with a time resolution of 5–30 min depending on the run time of IC (Simon and Dasgupta, 1993; Zellweger et al., 1999; Acker et al., 2004, 2006).

HONO can be also measured by rotating wet annular denuder (RWAN), set up horizontally. In this case, the absorbing solution is an alkaline solution, which coats the inner walls of the denuder. HONO is collected as nitrite, and the absorbing solution is then analyzed by IC. The detection limit is of 12 ppt, with a time resolution of 5–60 min depending on the run time of IC (Trebs et al., 2004).

Mist chamber–IC.

The mist chamber was used for sampling of HONO and other water-soluble species. This technique operates by nebulizing deionized water (scrubbing solution) into fine mist by the high flow of ambient air drawn into the chamber. The water droplets collect gaseous HONO, forming nitrite, which is analyzed by IC. The detection limit is of few ppt, with a time resolution depending on the chromatographic run (5–30 min). Aerosol particles are removed by a filter at the inlet of the chamber, to avoid interferences. The inlet is also maintained in insulated and dark conditions to avoid possible HONO formation due to heterogeneous and photolytic processes (Kleffmann and Wiesen, 2008).

HPLC/UV system.

HONO can be also determined by reverse-phase HPLC analysis with UV absorption detection at 309 nm (Zhou et al., 1999). A schematic sketch of HPLC/UV system is shown in Figure 3.

Figure 3. Schematic sketch of the semiautomatic HONO instrument based on reverse-phase HPLC analysis, followed by UV absorption detection.

Gaseous HONO is trapped quantitatively in a coil sampler using buffer solution. The scrubbing solution is derivatized with sulfanilamine (SA)/N-(1-naphtyl)ethylendiamine dihydrochloride (NED), analyzed by HPLC, and detected by UV-Vis absorption. The LOD is <0.8 ppt, with a time resolution of about 10 min. This method suffers from chemical interferences caused by, e.g., NO₂ and phenol reaction or by NO₂ and SO₂, which cannot be quantified. To correct the data, interference from gaseous and particulate species could be removed operating with a two-channel system. The first channel collects ambient air directly, whereas the second channel is equipped with a sodium carbonated (Na₂CO₃)-coated denuder to remove interfering compounds. However, during the last 10 years, HONO measurements have been performed by a technique based on long-path absorption photometer.

Long-path absorption photometric technique.

As HPLC/UV method, this method operates by collecting HONO by aqueous coil scrubbing, followed by acid derivatization (SA/NED mixture) of NO₂⁻, forming the highly light-absorbing azo dye, which is the detected by long-path absorption photometer (Heland et al., 2001; Kleffmann et al., 2002; He et al., 2006; Kleffmann and Wiesen, 2008, Ren et al., 2011). Up to now, many variants of this method have been reported. Ren et al. (2011) used long-path absorption photometric technique with a single channel to measure HONO concentrations at different heights in California. Zhou et al. (2002) performed simultaneous HONO and HNO₃ measurements by a two-channel long-path absorption photometric system. In particular, HONO was measured in the first channel as described above, whereas HNO₃ was measured by the same technique after first converting it to NO₂⁻ by a hydrazine reduction method. Helland et al. (2001) and Kleffmann et al. (2002) used long-path absorption photometric technique operating with two coil samplers in series (Figure 4): the first coil collects HONO and interfering species (such as NO_x and PAN), whereas the second coil collects the same amount of interfering species but no HONO.

Figure 4. Schematic sketch of the long-path absorption photometric (LOPAP) technique.

Thus, the correct HONO concentration can be determined by the difference between the signals of the first coil and the second coil. The detection limit is 1 ppt, with a time resolution of 1.5–4 min. Acidic sampling conditions allows minimizing some interferences (NO₂ and SO₂, NO₂ and phenols, NO₂ and aromatic amines, or peroxyacetyl nitrate [PAN] hydrolysis) (Heland et al., 2001; Kleffmann et al., 2002). Sampling artifacts (heterogeneous formation on surfaces by different NO₂ reactions, by photolysis, or by condensation of analytes on inlet) can be avoided by placing the stripping reagent directly in the atmosphere (Kleffmann and Wiesen, 2008).

Off-line methods

Besides the on-line techniques indicated above, HONO concentrations can be determined by off-line methods. They are based on the HONO sampling by devices such as dry diffusion denuders (Febo et al., 1993, 1996), filter pack (Sickles and Hodson, 1989), and passive samplers (Bytnerowicz et al., 2005), followed by analysis of extracted nitrite. Today, dry diffusion denuders and passive samplers are still used for HONO measurement.

The HONO sampling by dry diffusion denuder is performed by two Na₂CO₃-coated denuders (1% Na₂CO₃ + 1% glycerol in 1:1 ethanol/water solution). Gaseous HONO diffuses and is absorbed on the Na₂CO₃-coated surface, forming nitrite, which is analyzed by IC.

The HONO determination by carbonate-coated denuders can be affected by both positive interferences from nitrite forming species, such as NO₂ and PAN, and by negative interferences from ozone (O₃) and other oxidants. Corrections for positive interferences can be accounted for by operating two denuders in series (absolute differential technique), assuming that NO₂ and PAN are collected with the same efficiencies on the second Na₂CO₃-coated denuder (Febo et al., 1993, 1996; Ianniello et al., 2007). Gaseous HNO₃ should be selectively removed from the sampling air by means of sodium fluoride (NaF)- or sodium chloride (NaCl)-coated denuders in order to avoid the interfering nitrate amount in the HONO measurements (negative interference). Nitrate ions generated on Na₂CO₃-coated denuders by the oxidation of nitrite (from HONO) is indistinguishable from nitrate ions generated by the HNO₃ collection unless this species is removed from the incoming air stream. Thus, the atmospheric HONO concentrations are determined by subtracting the sum of nitrite and nitrate (expressed as nitrite) concentrations in the second Na₂CO₃-coated denuder from the sum of nitrite and nitrate (expressed as nitrite) concentrations in the first Na₂CO₃-coated denuder (Febo et al., 1993, 1996; Ianniello et al., 2007, Spataro et al., 2013). The diffusion denuder technique is accurate and selective but also labor-intensive, and requires long sampling time. The collection efficiency for HONO is about 99%, whereas the detection limit is 0.88 ppt, with a time resolution of 24 hr.

Finally, Bytnerowicz et al. (2005) developed a passive sampler for the simultaneous collection of gaseous HNO₃ and HONO. This device is based on a diffusion of ambient air through Tefion membrane and absorption of the pollutants on Nylasorb nylonfilter. After the sampling, the passive sampler is analyzed by IC, and HNO₃ and HONO are determined as nitrate and nitrite, respectively. Passive sampler is simple, easy to make, inexpensive, resistant to harsh weather conditions and does not require power supply. It can measure wide ranges of ambient HONO and HNO₃ concentrations for extended periods of time and can be used for regional-scale monitoring of the pollutants. The detection limit of 0.2 ppb over the 6-day sampling period is not suitable to evaluate the diurnal variation of HONO concentrations (Lee et al., 2002).

Conclusions and Future Outlook

This paper reports and discusses the HONO sources proposed to explain the HONO budget, especially during daytime. Five HONO formation pathways can be recognized: direct emission, homogeneous gas-phase reactions, heterogeneous reactions, photolysis/photoenhanced reactions, and biological processes.

Direct HONO emission (from vehicles) ranges between 0.3% and 1.7% of the total traffic NO_x emission. Vehicular traffic can be important HONO source only in heavily polluted urban environment. Since the atmospheric [HONO]/[NO_x] ratios are usually higher than those related to the direct emission, HONO is mostly secondarily produced. However, more evaluations of HONO, NO, and NO₂ emission are required to quantify the emitted amount by different vehicles (in particular light- and heavy-duty diesel vehicles, etc.).

The homogeneous gas-phase R3 and R6 reactions may act as atmospheric HONO sources, but they cannot explain the observed daytime HONO levels. Bejan et al. (2006) proposed the photolysis of different gaseous nitrophenols, estimating the HONO production rate of 100 ppt hr⁻¹ for a typical urban site. However, there are not yet field studies investigating this process in the real atmosphere and verifying its estimated HONO production rate.

Heterogeneous reactions, involving the NO₂ conversion on humid surfaces (R7), have been accepted to be the dominant nighttime HONO source. However, the detailed mechanism of R7 and its dependence on surface water content, chemical properties (such as chemical composition and pH), and type of surface is still not understood. The ground surface mainly affects the HONO production, but high aerosol loading in the atmosphere may increase the heterogeneous HONO production. Thus, more data, related to the role of aerosol and ground surface into the heterogeneous HONO production, are needed to clarify this question. The reactions R8 and R9 and the NO₂ conversion on soot surfaces cannot be important HONO sources. The interaction between NO₂ and diesel exhaust can also cause HONO formation. Ziemba et al. (2010) proposed that VOCs affect the HNO₃ conversion into HONO. Since VOCs are emitted in both gas and particulate phases by vehicular exhausts, it is not clear the homogeneous and/or heterogeneous nature of this process as well as the dependence of the HONO production rate to the surface water content and to the chemical composition of substrate.

Photolysis/photoenhanced surface processes were considered to explain the observed daytime HONO concentrations. Three processes were identified. The first one, the photolysis of adsorbed HNO₃/NO₃⁻ (λ ˜ 300 nm) results in the daytime HONO and NO_x production and is a possible pathway remobilizing deposited HNO₃ in rural and polar environments. However, there are still few studies quantifying its contribution to the atmospheric HONO budget. The second one is the photoenhanced NO₂ into HONO conversion on mineral dust substrates. The mineral powder (TiO₂, SiO₂, and CaCO₃) are likely involved in the HONO production, but more studies are required to clarify the mechanisms and their dependence on several parameters: surface chemical properties (acidity, water content, and composition), relative humidity, and concentration and type of nitrogen species. The last one, the photoenhanced NO₂ conversion on the organic and humic acid surfaces, was related to the HONO production rate of 700 ppt hr⁻¹ within a boundary layer of 100 m height. However, the organic chemicals participating in the HONO formation and the reaction mechanism of NO₂ with organic materials should be investigated. Laboratory researches on the photoenhanced HONO production considered only UV-A radiation, whereas the effect of visible and UV-B as well as the changes of the chemical and physical properties of surfaces were not evaluated.

Recently, biological nitrification and denitrification processes in the soil were proposed as the unknown HONO sources. This mechanism is consistent with the observations that ground surfaces are dominant for HONO production, but further studies are recommended to assess the actual impact of HONO release from soil nitrite on in the biogeochemical cycling of N in both agricultural and natural environments.

The knowledge of atmospheric HONO sources is crucial to obtain a more rigorous representation of its sources and their relative contribution to the HONO budget. Field measurements of HONO, NO, and NO₂ concentrations and fluxes in the lower troposphere are needed. These studies are crucial to fully characterize air quality and to integrate the current models.

Accurate and fast HONO measurement is the first step to investigate its sources. During the last 30 years, several techniques have been developed to determine HONO levels, and each of them is characterized by specific disadvantages and advantages. Wet methods suffer from sampling artifacts and some, such as NO₂ and SO₂, NO₂ and phenols, NO₂ and aromatic amines, or peroxyacetyl nitrate (PAN) hydrolysis. However, some versions of the long-path absorption photometric method claim to have minimized and quantified the known interferences.

Summary

HONO plays a key role in the tropospheric photochemistry, acting as a relevant source of OH radicals by its rapid photolysis (R1). OH radicals are involved in photooxidation processes, such as the formation of tropospheric O₃ and other secondary atmospheric pollutants (PAN and secondary particles). Recent field and modeling studies showed that HONO photolysis contributes significantly to the OH production throughout the day and that a strong and unknown daytime HONO source should exist. Up to now, there are five HONO formation pathways known: direct emission, homogeneous gas-phase reactions, heterogeneous reactions, surface photolysis, and biological processes. In this review paper, the HONO sources proposed to explain the HONO budget, especially during daytime, have been discussed. Since there are many techniques to detect HONO that do not fully satisfy the need for sensitivity, selectivity, and fast time response, a short description of HONO measurement techniques currently available in literature is reported. They are divided in three basic categories: spectroscopic techniques, wet chemical techniques, and off-line methods.

Acknowledgment

The authors would like to thank all participants of the field measurements focused on HONO sampling in the troposphere, and in particular the authors are grateful to Mr. Giulio Esposito for helpful discussions.

Funding

This work was financially supported by Institute for the Atmospheric Pollution Research (CNR-IIA).

Additional information

Notes on contributors

Francesca Spataro

Francesca Spataro and Antonietta Ianniello are both researchers at the CNR-Institute of Atmospheric Pollution Research (CNR-IIA), Rome, Italy.

Antonietta Ianniello

Francesca Spataro and Antonietta Ianniello are both researchers at the CNR-Institute of Atmospheric Pollution Research (CNR-IIA), Rome, Italy.