Environmental Science and Pollution Research

, Volume 21, Issue 15, pp 9259–9269

Formation of indoor nitrous acid (HONO) by light-induced NO2 heterogeneous reactions with white wall paint

  • Vincent Bartolomei
  • Matthias Sörgel
  • Sasho Gligorovski
  • Elena Gómez Alvarez
  • Adrien Gandolfo
  • Rafal Strekowski
  • Etienne Quivet
  • Andreas Held
  • Cornelius Zetzsch
  • Henri Wortham
Research Article

DOI: 10.1007/s11356-014-2836-5

Cite this article as:
Bartolomei, V., Sörgel, M., Gligorovski, S. et al. Environ Sci Pollut Res (2014) 21: 9259. doi:10.1007/s11356-014-2836-5

Abstract

Gaseous nitrogen dioxide (NO2) represents an oxidant that is present in relatively high concentrations in various indoor settings. Remarkably increased NO2 levels up to 1.5 ppm are associated with homes using gas stoves. The heterogeneous reactions of NO2 with adsorbed water on surfaces lead to the generation of nitrous acid (HONO). Here, we present a HONO source induced by heterogeneous reactions of NO2 with selected indoor paint surfaces in the presence of light (300 nm < λ < 400 nm). We demonstrate that the formation of HONO is much more pronounced at elevated relative humidity. In the presence of light (5.5 W m−2), an increase of HONO production rate of up to 8.6 · 109 molecules cm−2 s−1 was observed at [NO2] = 60 ppb and 50 % relative humidity (RH). At higher light intensity of 10.6 (W m−2), the HONO production rate increased to 2.1 · 1010 molecules cm−2 s−1. A high NO2 to HONO conversion yield of up to 84 % was observed. This result strongly suggests that a light-driven process of indoor HONO production is operational. This work highlights the potential of paint surfaces to generate HONO within indoor environments by light-induced NO2 heterogeneous reactions.

Keywords

HONO Indoor Light Heterogeneous reactions Paint 

Introduction

It is now recognized that nonindustrial indoor environments, such as private residences, offices, schools, and commercial and public buildings, are important places of exposure to air pollutants (Weschler 2009). First detected indoor by Pitts and co-workers, nitrous acid (HONO) is now known to be one of the most important pollutants within indoor environments (Finlayson-Pitts and Pitts 2000). HONO is known to form carcinogenic nitrosamines upon reaction with secondary amines (Pitts et al. 1978). In addition, the duplex DNA is directly cross-linked in the reaction of HONO with the exocyclic amino groups of the nucleosides, leading to mutagenic properties (Harwood et al. 2000). Further, the surface reactions of HONO with adsorbed nicotine lead to the formation of the so-called third-hand tobacco smoke hazard (Sleiman et al. 2010). The measured indoor HONO mixing ratios may vary between 2 and 25 parts per billion (ppb) (Spengler et al. 1993; Febo and Perrino 1991; Leaderer et al. 1999; Febo and Perrino 1995; Večeřa and Dasgupta 1994; Pitts et al. 1989; Lee et al. 2002) and are much higher than those observed in outdoor environments (Finlayson-Pitts and Pitts; 2000 Febo and Perrino 1995).

HONO can be directly emitted as a primary pollutant in the indoor air through combustion processes (gas stoves, burning candles, and open fire places). In addition, HONO can be formed via heterogeneous reactions of NO2 with sorbed water on various indoor surfaces (Pitts et al. 1984; Sakamaki et al. 1983; Jenkin et al. 1988; Akimoto et al. 1987). The measured indoor mixing ratios of NO2 are observed to be in the range between 15 and 200 ppb (Weschler 2009; Lee et al. 2002; Wainman et al. 2001). In some indoor environments, such as industrial workplaces and homes equipped with gas stoves, peak NO2 mixing ratios may reach at 1 to 2 ppm with a 24-h average NO2 mixing ratio of up to 0.5 ppm (Monn 2001).

Weschler and Shields (1997) stated that heterogeneous chemistry plays an important role in the indoor environment. The surface-to-volume ratios (S/V) (m−1) are much higher within indoor settings (Wainman et al. 2001; Weschler 2004) increasing the importance of NO2 heterogeneous reactions. Interestingly, given the potential importance of indoor heterogeneous reactions, the literature data remains limited. In particular, the heterogeneous reactions in the presence of UV light (300 nm < λ < 400 nm) were not studied at all (Gómez Alvarez et al. 2012) and their impact on indoor air quality remains to be explored (Gligorovski and Weschler 2013).

Carslaw (2007) using her MCM model examined the effect of solar light in the near-UV range on indoor HONO formation. Carslaw model indicates that indoor OH production rate via HONO photolysis is very sensitive to the values used for the attenuation of light from outdoors to indoors.

Since HONO absorbs light (300 nm < λ < 405 nm) (Stutz et al. 2000) in the wavelength region of the solar spectrum that is readily available within indoor environments (λ > 340 nm), it becomes a major precursor of indoor hydroxyl radicals (OH) (Gómez Alvarez et al. 2013). Indeed, Gómez Alvarez et al. (2013) employing a LIF-FAGE instrument directly measured high concentrations of OH radicals of up to 1.8 · 106 radicals cm−3 in a school classroom which were directly associated with the photolysis of HONO. This value is two orders of magnitude higher in comparison to the modelled values of 5 · 104 radicals cm−3 (Sarwar et al. 2002) which were obtained considering the reactions of ozone with alkenes and NO with HO2 as the source of OH radicals indoors. The degradation of organic compounds induced by photocatalytic indoor paints and their influence on the indoor air quality was studied by Salthammer and Fuhrmann (2007) and Auvinen and Wirtanen (2008). Laufs et al. (2010) observed a degradation of HONO under light irradiation of photocatalytic paint (with TiO2 nano-particles) aimed for outdoor building facades.

To the best of the authors’ knowledge, this is the first study which focused on NO2 to HONO conversion on indoor white wall paint under light irradiation. Using an experimental approach that couples a horizontal flow reactor to a Long Path Optical Absorption (LOPAP) analyzer, the NO2 → HONO conversion is investigated as a function of light intensity and relative humidity.

Experimental (materials and methods)

Description of the setup

The flow tube reactor is a cylindrical borosilicate glass tube inserted into another double-wall reactor with connections to a thermostatic bath which allowed operation at 295 K.

The experimental setup is represented in Fig. 1a. The dimensions of the flow reactor were chosen to ensure conditions for gas-phase laminar flow. Figure 1b presents a drawing of the half reactor with the applied paint on the surface. By using this design, we could ensure a homogeneous coverage and irradiation of the surface. The half reactor was inserted into the cylindrical flow reactor prior to experiments.
Fig. 1

a Schematic illustration of the experimental setup. Mass flow controller (a), needle valve (b), relative humidity (RH) system (c), on/off valve (d), injector (e), reactor (f), thermostated bath (g), hygrometer (h), RH probe (i), NOx monitor (j), and LOPAP (k). b Schematic representation of the semi-cylindrical reactor used to study NO2 reactivity on selected paint surfaces. The geometry of the experimental setup allows homogeneous irradiation of the selected paint that is uniformly coated on the inner wall of the half reactor located inside a cylindrical reactor

This flow tube is surrounded with four UV–VIS fluorescent lamps (Philips TL-D 18 W, 340–400 nm, λmax = 368 nm, length = 60 cm) which were installed symmetrically around the flow tube and could be operated individually. The flow tube and the lamps were placed in a stainless steel box. All tubings, connected to the flow tube, were wrapped in aluminum foil to protect them from light and prevent photolysis of the different gaseous compounds. The values of NO2 provided were those read by the NOx analyzer which was appropriately calibrated beforehand according to manufacturer recommendations and procedure and were corrected for dilution. The NOx analyzer was an Eco Physics model (CLD 88p) associated with an Eco Physics photolytic (metal-halide lamp; 180 W) converter (PLC 860), which allows simultaneous measurements of NOx, NO2, and NO concentrations. This instrument is very sensitive with a detection limit of 50 ppt. Nitrogen oxides (NOx = NO + NO2) were measured by chemiluminescence (reaction between NO and O3) with alternative detection of NO and then detection of NOx by selective photo-conversion of the NO2 to NO. This is particularly important because the gas-phase HONO quantitatively interferes in commercial NOx instruments using molybdenum converters. The NO channel was directly calibrated by diluted standard NO calibration mixtures.

A Dynacal® permeation tube (VICI, AG International) inserted into a gas-phase calibration unit (UPK model 1001, Germany) was used to generate stable gas-phase NO2. The carrier gas was air (zero air generator). Inside the reactor, variable mixing ratios of gas-phase NO2 in a total of 50 ml min−1 of air-carrier-gas are injected through a mobile injector, which also enables varying the surface of the reactor in contact with NO2 for kinetic studies.

Sheath flow of 800 ml min−1 was introduced in the reactor. Humidity was adjusted by splitting this sheath flow in two fluxes, one of dry air flow controlled by mass flow controller (Brooks SLA Series mass flow controller, in the range 0–1 L min−1 ± 1 % accuracy) and the other one humidified by bubbling in deionized water was controlled by a needle valve. A mixing of these two flows at different ratios generated a carrier gas at controlled relative humidities. Downstream of this device, a hygrometer “Hygrolog NT2” (Rotronic) with a “HygroClip SC04” probe measured the resulting humidity on-line (±1.5 % relative humidity (RH) accuracy).

At the reactor exit, the effluent of the reactor was mixed with a flow of air (1.45 l min−1) controlled by a Brooks SLA Series mass flow controller (2000 sccm, ±1 % accuracy, Brooks) to make up the effluent of the reactor to feed the NOx analyzer (minimum required flow of 730 ml min−1) and the LOPAP instrument (minimum required flow 1 l min−1) giving a total flow of 2.3 l min−1.

The gas-phase HONO was measured using a “LOPAP”, after its conversion into an azo dye in the measuring instrument. The LOPAP instrument has a detection limit of 2–3 ppt and a total accuracy of ±10 % with an actual time resolution of about 30 s under the operation conditions applied. A detailed description of its operation principle and procedure can be found in the literature (Kleffmann et al. 2002; Kleffmann and Wiesen 2008).

The spectral irradiance emitted by four lamps surrounding the flow tube reactor was measured with a spectroradiometer (SpectralEvolution SR-501) (range 250 and 1100 nm, with an accuracy of 0.5 nm, and a reproducibility of 0.1 nm). The spectroradiometer is factory calibrated for radiance and/or irradiance using NIST-traceable source. It has a radiometric calibration accuracy specification for the SR-501 of ±5 % at 400 nm and ±4 % at 700 nm. Prior to each measurement, we registered the signal under dark conditions.

The white paint of brand name Luxens (certified A+ by the European ecolabel), France, was applied on the surface of the half cylindrical reactor.

Please note that the chemical properties of the paint are not often made public. The chemical characterization of this commercial formulation was beyond the scope of this study.

Results and discussion

Spectral irradiance

We measured the spectral irradiance of the light which penetrates indoors and compared it with the outdoor light irradiance measured with a few minutes of difference. The integrated irradiance over the range of wavelengths between 300 and 400 nm was 9.7 and 22.2 (W m−2) for the solar light which penetrates indoors through the windows and the solar light outdoors, respectively. In addition, we measured the spectral irradiance of the lamps irradiating the flow tube reactor by inserting the fiber optics of the spectroradiometer into the reactor where the heterogeneous reactions take place. The comparison of the spectra obtained is presented in Fig. 2.
Fig. 2

Comparison of spectral irradiance of the fluorescent lamps used in the present study with solar natural radiation. Pink line, green line, and blue line: Spectral irradiance of two, three, and four lamps, respectively, λ = 340–400 nm (90 % of the total light power), λmax = 368 nm. Solar spectral irradiance were taken on 6 May 2013 at 16:00 in the city center of Marseille, France (clear sky) on a horizontal surface at 10-cm distance from the window obtained with open windows (outdoors) (black line) and closed windows (indoors) (red line). The spectra were taken in the building inside the university laboratory in Marseille, France oriented north/south. As a comparison, a standard solar spectral irradiance for solar zenith of 48º and clear sky (yellow line) is shown (American Society for Testing and Materials (ASTM)). This spectrum corresponds to the total irradiance on a surface tilted towards the sun (37º)

The integrated spectral irradiance in the near UV range of the solar spectrum (300 nm < λ < 400 nm) for two, three, and four lamps is 5.5, 8.7, and 10.6 (W m−2), respectively. Assuming an isotropic radiation in the reactor due to the geometry of the four lamps surrounding the flow tube at the same distance, the actinic flux was estimated. With actinic flux obtained and using the IUPAC recommendations (Atkinson et al. 2004) for cross-sections of NO2 and HONO for the region (300 nm < λ < 400 nm), we calculated the corresponding photolysis frequency (J) as 1.2 · 10−2 and 2.3 · 10−3 s−1, respectively.

NO2 photodissociates in the wavelength region (340–400 nm) emitted by the lamps. However, according to the estimated values of J(NO2), the NO2 loss in the gas phase was negligible for the residence time of 4.24 s used in our experiments.

Kinetic analysis

NO2 uptake measurements

We tested the reaction of NO2 with the paint under dark conditions. In the absence of light, the gas-phase NO2 only physically adsorbs on the paint surface without any subsequent surface reaction. As a next step, we performed the experiments in the presence of light to test its influence on the NO2 heterogeneous reactivity with the paint. Before starting the experiment represented in Fig. 3, lights were turned on so that all the steps described were performed under light irradiation. The kinetic measurements were performed as follows. First, a stable concentration of NO2 was established with the injector pushed to the end of the reactor. In this way, the coated paint is not exposed to the flow of gaseous NO2. This step corresponds to region from A to B in Fig. 3.
Fig. 3

Typical NO2 uptake signal and signal of build-up of HONO. Positions A to H and A′ to F′ are explained in the core of the paper. Experimental conditions: [NO2] = 60 ppb, integrated irradiance for (300 nm < λ < 400 nm) is 10.6 (W m−2), 70 % RH, T = 298 K. Error bars in the profile of HONO correspond to the uncertainties of 10 % in the measurements of HONO by the LOPAP

Once a steady-state NO2 concentration is reached, the injector was withdrawn (position B in Fig. 3) to the first position of the reactor to expose the first section of the reactor covered with paint to NO2. The drop of the NO2 signal from B to C corresponds to the mixing of NO2 with the air stream and the adsorption of NO2 on the surface. Then, the NO2 signal returns to position D. The difference between B and D (ΔNO2) corresponds to the concentration of NO2 consumed due to the reaction with the exposed paint surface. From position D to E, a stable signal of NO2 can be observed. Once the NO2 concentration was stable, the injector was pushed to the end (position E) of the reactor, and desorption was observed as an increase of the NO2 signal from E to F followed by a decrease of the NO2 signal by reaching the steady-state NO2 concentration from G to H (position equivalent to the initial state from A to B). This experimental procedure was repeated for three different positions of the injector, i.e., at three different contact times/surfaces.

The uptake coefficient of NO2 can be estimated in function of the gas/surface exposure time. The experimental data points correspond to the different positions of the movable injector. The linear fit of the natural logarithm of c0/ci (Eq. 1) at three exposure times, corresponding to three different positions of the injector inside the flow tube, was used to derive an apparent pseudo-first-order coefficient (k1st) for the NO2 uptake as follows:
$$ 1\mathrm{n}\frac{c_0}{c_{\mathrm{i}}}={k}_{1\mathrm{st}}t $$
(1)
where c0 is the concentration of NO2 at the flow tube entrance (points A to B), ci is the NO2 concentration after the reaction (points D to E), and t is the residence time.
As shown in Fig. 4, the observed loss as a function of the exposure time, i.e., exposed surface, corresponds to first-order kinetics with respect to the gas-phase concentration of NO2.
Fig. 4

Loss of gas-phase NO2 as a function of exposure time to a white paint in the presence of light (10.6 W m−2). Black circle, 50 % RH; black triangle, 70 % RH. [NO2]0 = 60 ppb and T = 298 K. Errors bars are the propagation errors

The slope of the regression line in Fig. 4 gives the first-order rate constant (k1st). The uptake coefficient of NO2 was estimated by Eq. 2 based on the derived first-order rate constant:
$$ \gamma =\frac{4{k}_{1\mathrm{st}}}{\overline{v}}\kern0.5em \frac{V}{S} $$
(2)
where γ is the dimensionless uptake coefficient of NO2 which defines the probability that a collision of NO2 with the paint surface leads to HONO conversion, \( \overline{v} \) is the mean molecular velocity of NO2, V/S represents the volume-to-surface ratio of the reactor, and number 4 comes from the solution of the derived equation which describes the number of collisions of gas-phase species with a surface per unit time and per unit area. As explained in the “Experimental (materials and methods)” section, the paint was applied to the surface of a half cylindrical reactor that was inserted in the flow tube reactor. Considering the different geometries and radii of both the half cylindrical and full cylindrical reactor (the flow tube), the ratio V/S in Eq. 2 becomes as follows:
$$ \frac{V}{S}=\frac{\pi {r}_2^2L}{\pi {r}_2L}=1.04\kern0.5em \mathrm{cm} $$
(3)
where r1 = 0.6 cm is the radius of the half cylindrical tube, r2 = 0.79 cm is the radius of the flow tube reactor, and L = 34 cm is the length of the flow tube reactor. Replacing the ratio V/S from Eq. 3 into Eq. 2, the uptake coefficient can be estimated in the following manner:
$$ \gamma =\frac{4.16\left[\mathrm{cm}\right]{k}_{1\mathrm{st}}}{\overline{v}} $$
(4)

The Cooney-Kim-Davis method was used to check if the measured uptake coefficients are limited by the gas-phase diffusion processes in the flow tube reactor. This method was described in detail by Behnke et al. (1997). The estimated corrections were less than 1 % implying that the obtained uptake coefficients are not limited by the gas-phase diffusion processes under the given experimental conditions.

The relative humidity is an important parameter influencing the heterogeneous reactions of NO2 with organic surfaces (George et al. 2005). Indeed, the water molecules can adsorb on the paint surface modifying its properties and its ability to react with NO2. Figure 4 shows that at 70 % of RH, the first-order rate constant (k1st) is about three times higher in comparison to the obtained k1st at RH of 50 %. The estimated uptake coefficients (γ) are 1 · 10−7 and 4 · 10−7 for the heterogeneous reactions of NO2 (60 ppb) with paint surface under light irradiation at 50 and 70 % RH, respectively. These values are in the same order of magnitude as the uptake coefficients for the photosensitized reactions of NO2 with benzophenone (γ = 6.5 · 10−7) under dry conditions or with anthrarobin (γ = 3.4 · 10−7) in the presence of 14 % of RH (George et al. 2005).

The observed humidity dependence in this study at 50 % and 70 % RH is in agreement with the literature data showing that increase of RH leads to elevated uptake coefficients (Sosedova et al. 2011; Arens et al. 2002). These facts demonstrate that the adsorbed water molecules on the organic surface can enhance the surface reactivity towards NO2.

Contrary to our results, Brigante et al. (2008) have shown that there is no dependence of the NO2 uptake coefficients on a pyrene surface over the broad range of RH 10–90 %. However, it has been shown that surface-bound water might be involved in the reaction of NO2 with adsorbed phenolic species through hydrolysis of OH groups (Alfassi et al. 1986). In the same manner, Stemmler et al. (2007) have shown strong humidity dependence (20–80 %) of NO2 uptakes on solid anthrarobin film.

In a study performed by Sosedova et al. (2011), it has been shown that the NO2 reactions with solid films of polyphenolic substances such as tannic acid and gallic acid in the presence of light are strongly influenced by the RH. These authors claimed that water seems to be involved in this photochemical reaction through stabilizing the intermediary semiquinone radical.

Clearly, further research is necessary to reveal the influence of RH in the heterogeneous NO2 reactions with adsorbed organic compounds.

Note that no uptake coefficients were observed under dark conditions. George et al. (2005) carried out a comprehensive study of heterogeneous loss reactions of gas-phase NO2 with a number of organic compounds coated on the glass surface. They demonstrated that under dark conditions, the obtained NO2 uptake coefficients on pure catechol surface and on a surface made of anthrarobin are extremely slow under dry conditions concluding that NO2 only physically adsorbs on those surfaces. Brigante et al. (2008) have studied the heterogeneous reactions of NO2 with coated pyrene surface, and the obtained uptake coefficients under dark conditions were also very small in the range between 2.3 · 10−7 and 5.8 · 10−8 over the range 20–119 ppb of NO2.

It should be noted that we do not observe any NO2 uptake coefficients for the heterogeneous reactivity of NO2 with paint under light irradiation at 25 % RH. Also, under such experimental conditions, the LOPAP instrument that has much lower detection limit (few ppt) than the NOx analyzer did not detect any release of HONO (see the section below).

HONO measurements

We did not observe any HONO formation under dark conditions. In presence of light, simultaneously with the loss of NO2, a formation of gaseous HONO occurs due to the surface reaction between the adsorbed NO2 and the paint surface. A typical signal describing the build-up of HONO due to the light-induced heterogeneous reactions of 60 ppb of NO2 with the painted surface of the reactor at 50 % of RH is given in Fig. 3.

The region from A′ to B′ in Fig. 3 corresponds to the background concentrations of HONO formed in the tubings prior to the reactor by heterogeneous reactions of NO2 with the sorbed water on the Teflon surface (Finlayson-Pitts et al. 2003). This reaction occurs in the dark since as mentioned before, the Teflon tubings were wrapped with the aluminum foil during the experiments. Such background formation of HONO prior to the reactor was also reported in the literature (Stemmler et al. 2007). In position B′ (Fig. 3), the injector has been retracted and the paint surface exposed to the gaseous NO2. From B′ to C′, we observed an increase of the HONO signal (ΔHONO) up to about 1 ppb. A simple comparison of this value with the NO2 loss observed of about 1 ppb in Fig. 3 implies that the NO2 → HONO conversion yield is very high (84 %). The high conversion yield indicates that NO2 is reduced by photochemically activated electron donors (Stemmler et al. 2006) which are most likely present in the paint. The plateau formed from C′ to D′ corresponds to the stabilized level of HONO signal. When the injector is pushed again to the end of the reactor, it causes a decrease of the HONO signal back to the baseline, i.e., region E′ to F′ in Fig. 3.

The HONO formation rate was studied at three different light intensities. Figure 5 shows the linear dependence between the light-enhanced HONO production and different integrated spectral irradiances in the range 300 nm < λ < 400 nm, at 60 ppb of NO2 and 50 % RH. Typically, the time of simultaneous exposure to NO2 and light is about 40 min as shown in Fig. 3.
Fig. 5

Black circle: HONO production rate as a function of the total integrated irradiance for wavelength between 300 and 400 nm. The intercept of the regression line corresponds to the HONO uptake under dark conditions. Error bars correspond to the uncertainties in the measurements of HONO by the LOPAP. Black square: NO2 consumed by the paint surface in function of the integrated spectral irradiance (300 nm < λ < 400 nm). The error bars correspond to the RSD of three repeated experiments. [NO2]0 = 60 ppb, RH = 50 %, and T = 298 K

In Fig. 5, it can be appreciated that the generation of HONO increases linearly with the irradiation intensity of the paint surface. The intercept of the linear regression gives an HONO uptake of 3.0 · 109 molecules cm−2 s−1 under dark conditions. This implies that the background HONO formed prior to the reactor is adsorbed on the paint under dark conditions. In Fig. 5, it can be seen that there is an increase of surface HONO production rates up to 2.1 · 1010 molecules cm−2 s−1, at 60 ppb of NO2 and 50 % RH under maximum light irradiation of the surface, 10.6 W m−2 in the range of wavelengths between 300 and 400 nm. This linear dependence clearly indicates a light-driven process of HONO production under light irradiation.

Figure 5 also shows the linear dependence of correspondingly consumed NO2 with the light intensities. From the two plots, the NO2 → HONO conversion yields were derived at different light intensities. The NO2 to HONO conversion yields are very high in order of about 60 to 90 %. Such high conversion yields were observed for outdoor surfaces (Stemmler et al. 2006), and they are associated with photosensitized reactions (Gómez Alvarez et al. 2012).

Concerning the outdoor air, a linear dependence of the HONO production rate on the solar irradiance has been observed as well (Bartels-Rausch et al. 2010). These authors have also reported that the linear dependence leads to a plateau at higher values of spectral irradiance which was ascribed to the deactivation mechanism involving photochemically produced transient species. In our study, we did not observe such stabilization of HONO production with increasing light intensities, probably due to the lower light intensity employed. This is reasonable since our study concerns indoor air where the near UV fraction of the solar light in the region between 300 and 400 nm is attenuated about two times in comparison to outdoors (Fig. 2).

The ultimate goal of this study was to demonstrate that real indoor surfaces such as wall paint could be responsible for high concentrations of HONO indoors, while the chemical characterization of the paint was beyond the scope in this work. The paint contains many organic and inorganic compounds; hence, it would be extremely difficult to assign the contribution of the compounds present in the paint to the HONO production. A detailed description of the formulation of the paint would not help an unequivocal attribution of the photosensitization properties to a certain compound or even a mixture of them.

The mechanisms leading to HONO formation still constitute a field where a significant amount of unknowns remain. Photosensitization processes have contributed to account partially for the missing sources of HONO (Stemmler et al., 2006). Presumably, the light-absorbing organic compounds, exhibiting photosensitizing properties present in the paint, could be responsible for the photo-enhanced HONO formation observed in this work.

In the generation of HONO, different factors play a role, and one is the nature of the surfaces: roughness, porosity, pH, and adsorption characteristics, and these features are linked to the ensemble of components, i.e., to the formulation itself and not to the individual components.

The possible influence of the bare glass, when it is covered with the paint, to the production of HONO was discarded as most unlikely. Perraudin et al. (2005) reported that the NO2-induced loss rate of selected PAHs does not depend on the nature of the substrate upon which they were deposited indicating that the glass itself while coated with chemicals does not have any influence on the formation of HONO.

The formation of HONO due to the light-induced heterogeneous reactions with the paint surface was also studied at three different relative humidities (RH %) of indoor relevance.

At RH of 25 %, the detected quantity of HONO is indistinguishable from the LOPAP noise. This is not surprising since we were not able to estimate any reliable NO2 uptake coefficients at 25 % RH (see the section above).

Figure 6 shows the increase of HONO production with RH due to the light-induced heterogeneous reaction of NO2 with the paint surface.
Fig. 6

Black circle: Production rate of HONO as a function of relative humidity. Error bars correspond to the uncertainties in the measurements of HONO by the LOPAP. Black square: Linear dependency of NO2 with the relative humidity. The error bars correspond to the RSD of three repeated experiments. [NO2]0 = 60 ppb, integrated irradiance = 10.6 W m−2, and T = 298 K. Experimental conditions: Integrated spectral irradiance (300 nm < λ < 400 nm) = 10.6 W m−2 and T = 298 K

In Fig. 6, it can be appreciated that the HONO production rate is strongly dependent on the RH. At 70 % of RH, the HONO production rate increases up to 6.8 · 1010 molecules cm−2 s−1. Arens et al. (2002) studied the reaction between gas-phase NO2 and solid anthrarobin and found that HONO yield increased with increasing RH. These authors claimed that the presence of adsorbed water could promote hydrolysis of the hydroxyl groups of the substrate, which then became reactive towards NO2. The dependence of HONO formation on RH presented in Fig. 6 suggests that water molecules are most likely involved during the NO2 to HONO conversion on the paint surface under light irradiation.

Also, Wainman et al. (2001) studied the generation of HONO in their Teflon chamber induced by heterogeneous reactions of NO2 with carpet and wallpaper at 50 and 70 % RH. These authors have shown that highest and prolonged release of HONO concentrations occurs at elevated RH.

Atmospheric implications

To estimate the atmospheric implications, we assume an average-sized room of 2.5 m high, 5 m wide, and 4 m long, hence with a total volume of 50 m3. One side of the room represents a large window with the surface of 10 m2 and the other side of the room is a wall with the surface of 10 m2. The other two walls are with surface of 25 m2 (each of 12.5 m2). Assuming that the typical indoor-integrated spectral irradiance is 9.7 W m−2 (Fig. 2) for the wavelength region 300 < λ < 400 nm, then the HONO surface production rate corresponds to 1.86 · 1010 molecules cm−2 s−1. In order to simulate more realistic conditions, we minimize the area of the walls which can be exposed to the direct sunlight. Therefore, we assume that only one half of one of the side walls (S = 6 m2) is irradiated with direct sunlight penetrating through the window yielding to the total indoor HONO emission rate of 0.3 mg h−1. The estimation provided assumes an average NO2 mixing ratio to be 60 ppb and 50 % RH. The indoor HONO emission rate was calculated using the near UV fraction (300 < λ < 400 nm) of the transmitted solar light; hence, the described photo-enhanced HONO formation can be attributed exclusively to the fraction of solar light that penetrates into the room through the window, i.e., not considering indoor electrical lighting such as compact fluorescent lamps (CFLs) or light-emitting diodes (LEDs). Further, the solar light which penetrates indoors depends on many factors such as the geographic location, the period of the year, orientation of the room, the size of the windows, and the nature of the glass (Gligorovski and Weschler 2013).

Assuming an average air exchange rate of 0.56 h−1 (Wallace et al. 2002), we can estimate the concentration of HONO produced by the paint in function of time by the procedure explained below.

The time-dependent HONO concentration can be described by the mass conservation relationship:
$$ \frac{d{c}_{\mathrm{a}}(t)}{ dt}=\frac{E_{\mathrm{R}}}{V}+{k}_{{\mathrm{AER}}^C\mathrm{out}}(t)-\left({k}_{\mathrm{AER}}+k^{\prime}\right){c}_{{}_{\mathrm{a}}}(t) $$
(5)
where t is the time (h), ca(t) is the gas-phase concentration of HONO in the room (mg m−3), kAER is the air exchange rate constant (h−1), k′ is the kinetic rate constant that describes the loss mechanisms (h−1), for example chemical degradation or photolysis, ER is the mass emission rate of HONO induced by the paint (mg h−1), cout(t) represents the gas-phase HONO concentration coming from outdoors (mg m−3), and V is the total volume of the room (50 m3). The concentrations of HONO outdoors are much lower compared to indoors (Febo and Perrino 1995) which implies that HONO coming from outdoor is negligible; therefore, the term kAERcout(t) can be excluded from (Eq. 5). Assuming that the emission rate of HONO from the paint is constant over time, the analytical solution of (Eq. 5) can be expressed as follows:
$$ {c}_{\mathrm{a}}(t)=\frac{E_{\mathrm{R}}}{\left({k}_{\mathrm{AER}}+k^{\prime}\right)V}+\left({c}_0-\frac{E_{\mathrm{R}}}{\left({k}_{\mathrm{AER}}+k^{\prime}\right)V}\right){e}^{-\left({k}_{\mathrm{AER}}+k\prime \right)t} $$
(6)
where c0 = 0 is the initial concentration of HONO (mg m−3).
The estimated production rates of HONO mixing ratios in function of time are depicted in Fig. 7.
Fig. 7

The estimated (Eq. 7) mixing ratios of HONO released by paint in function of time

The HONO values obtained by this model (Eq. 6) correspond to the HONO levels which could be found in various indoor environments (Spengler et al. 1993; Febo and Perrino 1991; Leaderer et al. 1999; Febo and Perrino 1995; Večeřa and Dasgupta 1994; Pitts et al. 1989; Lee et al. 2002). This model considers generic degradation; we show that photolysis is very important (Gómez Alvarez et al. 2013), then we assume that most HONO degradation is carried out by photolysis and place k′ = J(HONO) = 2.1 · 10−3 s−1.

Then, Eq. 6 can be rewritten as follows:
$$ {c}_{\mathrm{a}}(t)=\frac{E_{\mathrm{R}}}{k_{\mathrm{AER}}V+J\left(\mathrm{HONO}\right){V}_1}+\left({c}_0-\frac{E_{\mathrm{R}}}{k_{\mathrm{AER}}V+J\left(\mathrm{HONO}\right){V}_1}\right){e}^{-\left({k}_{\mathrm{AER}}+J\left(\mathrm{HONO}\right)\right)t} $$
(7)
where V1 represents the volume of the room that is irradiated with the near UV fraction (300 < λ < 400 nm) of the solar light which penetrates indoors. Under assumption that one third of the volume of the room was illuminated during certain period of the day, we obtain an average mixing ratio of 2 (μg m−3), i.e., 1 ppb of HONO released by the paint. Considering that the volume of indoor air is much smaller and confined in comparison to the troposphere, the value of 1 ppb of HONO is very high with respect to the indoor air quality and the consequent health impact. The paint surface represents only one source of indoor HONO. Further studies focused on light-induced NO2 heterogeneous reactions with various indoor surfaces will reveal the total HONO production indoors.

Clearly, the light-induced heterogeneous reactions of NO2 with white wall paint give an important contribution to indoor HONO generation pathways. Considering that nowadays the architectural trend tends towards houses oriented north–south with large glass windows and glass wall buildings, the impact of indoor photochemistry becomes increasingly important. The photodissociation of indoor HONO gives high levels of OH radicals (Gómez Alvarez et al. 2013). OH radicals can induce degradation of typical indoor pollutants (e.g., pollutants emitted by furniture). Ultimately, it can oxidize all hydrocarbons in the indoor atmosphere and form CO2 and H2O. This process, however, is complicated and usually the reactive intermediates in radical chain reactions are formed. The reaction products may produce irritants and play a role in so-called sick building syndrome. To adequately describe and evaluate the space and time distribution of such intermediates, indoor OH radical sources and concentration levels need to be studied in detail (Gligorovski and Weschler, 2013). Considering that human beings in developed countries typically spend 85 to 90 % of their time indoors, the emerged results from this study on behalf of HONO formation becomes of great importance.

Acknowledgments

The authors gratefully acknowledge the project “Les surfaces internes à l'habitat, une source probable d'acide nitreux” and the French program PRIMEQUAL.

This work is also a contribution to the LABEX SERENADE (no. ANR-11-LABX-0064) funded by the “Investissements d’Avenir,” French Government program of the French National Research Agency (ANR) through the A*Midex project (no. ANR-11-IDEX-0001-02). Financial support by the German Science foundation (DFG project HE5214/4-1) is gratefully acknowledged.

Copyright information

© Springer-Verlag Berlin Heidelberg 2014

Authors and Affiliations

  • Vincent Bartolomei
    • 1
  • Matthias Sörgel
    • 2
    • 3
  • Sasho Gligorovski
    • 1
  • Elena Gómez Alvarez
    • 1
  • Adrien Gandolfo
    • 1
  • Rafal Strekowski
    • 1
  • Etienne Quivet
    • 1
  • Andreas Held
    • 2
  • Cornelius Zetzsch
    • 2
  • Henri Wortham
    • 1
  1. 1.Aix Marseille UniversityMarseille Cedex 3France
  2. 2.Forschungsstelle für Atmosphärische ChemieUniversität BayreuthBayreuthGermany
  3. 3.Biogeochemistry DepartmentMax Planck Institute for ChemistryMainzGermany

Personalised recommendations