Environ. Sci. Technol. 2007, 41, 3898-3903
Photochemical Loss of Nitric Acid on Organic Films: a Possible Recycling Mechanism for NOx SUSANNAH R. HANDLEY,† DANIEL CLIFFORD,† AND D . J . D O N A L D S O N * ,†,‡ Department of Chemistry, University of Toronto, Toronto, Ontario, Canada M5S 3H6, and Department of Physical and Environmental Sciences, University of Toronto at Scarborough, Toronto, Ontario, Canada
The films coating urban impervious surfaces have been found to be comprised of about 7% inorganic nitrate and ∼10% organic compounds (by mass). A simple steady-state analysis of the lifetime of the nitrate in the film suggests the existence of a loss process(es) in addition to washout by rainfall. We show here that gas-phase nitric acid can be taken up in organic films and lower the film pH. Photolysis of nitrated films using actinic illumination causes loss both of protons and of nitrate anion. We argue that this is possibly due to a combination of direct and indirect (photosensitized) photochemistry involving nitrate ions, yielding gas-phase HONO and/or NO2.
Introduction The possibility of “renoxification” of urban air and potential mechanisms for it have received increasing attention of late (1-5). Recent field measurements of HONO in the boundary layer find up to an order of magnitude higher daytime concentrations present than expected on the basis of the known formation and destruction routes (6-10). Urban HOx (dOH + HO2) concentrations are coupled to that of HONO, via the latter compound’s photolysis to OH + NO; this reaction is believed to be the primary source of boundary layer OH at dawn, when the solar zenith angle is too high to allow ozone-mediated OH production and may also be a significant source of HOx during the day (9, 10). It has been generally accepted that HONO is formed predominantly in a heterogeneous reduction of NO2, mainly through the hydrolysis reaction:
2NO2(g) + H2O(surf) f HONO(g) + HNO3(surf)
(1)
The observation of higher-than expected HONO concentrations suggests that current models are lacking an important source for this compound. Two novel sources have recently been proposed. In one, photochemical reduction of NO2 on various organic substrates has been observed (11, 12). It is believed that the reduction proceeds via adsorption onto the substrate, followed first by a one-electron transfer to the adsorbed nitrogen dioxide, then proton transfer, forming HONO. This process occurs efficiently using visible wavelengths of light, * Corresponding author phone: 416-978-3603; fax: 416-978-8775; e-mail:
[email protected]. † Department of Chemistry, University of Toronto. ‡ Department of Physical and Environmental Sciences, University of Toronto. 3898
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 11, 2007
and could thus play a role in HONO generation throughout the day. In the second proposed mechanism, nitric acid adsorbs onto surfaces which have an existing thin layer of water present and the hydrated or dissolved nitrate anion undergoes known aqueous phase photochemistry in the nearultraviolet region of the spectrum (1, 13-15) to yield HONO (and perhaps other “active nitrogen” products such as NO or NO2 as well). This route was proposed initially by Zhou et al. (4) and has been observed experimentally using ATR and Si crystals and borosilicate glass as the substrate (2, 5). Because the primary pathway for removal of inorganic nitrate (nitric acid or ammonium nitrate) from the atmosphere is by wet (i.e., uptake by water droplets) or dry deposition, followed by rainout/wash off to the ground, this route provides a mechanism to recycle nitrate back to the gas phase as “active” nitrogen oxides (HONO, NO2, or NO). Recently, the work of Diamond and co-workers (16-19) has highlighted the potential role of urban surface films (i.e., “window grime”) in controlling air quality. Such films are known to be present on urban impervious surfaces (windows, roadways, etc.) and to be composed of a complex mixture of chemical compounds, both organic and inorganic (18, 20-23). The primary constituents by mass are inorganic compounds, mainly nitrate and sulfate (21, 23); the organic fraction, which comprises 5-10% by mass of the total, contains a wide array of naturally occurring and anthropogenic chemicals, including carbohydrates, aliphatics, and aromatics (18, 21-24). The presence of such organic material in urban films suggests that these films could present a significant heterogeneous medium for atmospheric chemical reactions (25) and potentially play an important role both in sequestering gas-phase compounds and in providing a reactive medium for heterogeneous chemistry (26). We have begun to explore this possibility by studying the heterogeneous ozonation kinetics of a series of polycyclic aromatic hydrocarbons (PAHs) incorporated in organic films (27). Similar to our earlier work on such reactions at the air-water interface (26, 28), the ozonation was seen to take place at the organic film surface at a rate orders of magnitude more rapid than the corresponding gas-phase reactions. Modeling studies which incorporate these kinetics suggest that reactions on urban film surfaces can be extremely important in governing the fates of compounds which may undergo gas-surface partitioning and be present in these films (25). A class of such reactions which has received little attention to date is photochemical transformations taking place in urban organic films. These reactions could be significant not only to the fates of organics residing on such films, but could also affect the gas-phase chemistry by releasing reactive species from the film. The high (approximately 7% by mass) nitrate concentrations in urban films suggest that photochemistry involving this compound could be important in urban films. This point was suggested as well by Ramazan et al. (5), who pointed out the potential for HONO generation from nitric acid associated with urban surfaces. Taking a conservative estimate of 6.4 × 106 m2 for the coated area of impermeable urban surface in a city such as Toronto, Ontario, Canada (25), with empirical estimates for the film thickness (500 nm) and density (0.82 g cm-3) yields 3000 mol of nitrate, for a film which is 7% nitrate by mass (21). Assuming a steady state in nitrate concentration in the film and a dry deposition velocity of 5 cm s-1 for nitric acid (29) (at 1 ppb abundance) gives a lifetime of ∼3 days for nitrate on the film. If other sources of nitrate (i.e., ammonium nitrate in particles) are also included, the film lifetime 10.1021/es062044z CCC: $37.00
2007 American Chemical Society Published on Web 04/27/2007
diminishes. The 30 year average for days with rainfall over 5 mm in Toronto is 44 year-1 (30), giving a lifetime against wash off from surfaces of ∼8 days. This brief calculation suggests that washout by rainfall may not be the only loss process for film-associated nitrate. In the following we present experimental results which imply that urban film surfaces may act as effective media for photochemical renoxification; in particular, that HONO and NO2 could be photogenerated from urban film surfaces. Hydrophobic organic films take up nitric acid vapor which exists in the film at least partially in its ionic form (H+ + NO3-). The near-UV spectrum of nitrate dissolved in octanol is similar to that dissolved in water, but somewhat more intense. Both the acidic proton and the nitrate anion are lost during actinic irradiation of nitrated organic films, under illumination conditions similar to those which yield NO2 and HONO from nitric acid adsorbed on ice surfaces (31). We infer that the nitrate observed in urban films could arise from nitric acid uptake, or by deposition of particle-associated nitrate and that, once deposited, it may take part in photochemical reactions which may yield gas-phase HONO and/or NO2.
Experimental Methods Photochemistry Studies. These experiments used the fluorescent pH indicator acridine to monitor the presence of protons in the organic film. Acridine is a three-ringed heterocyclic PAH whose central nitrogen acts as a Lewis base, with pKa ) 5.4. When excited at 337 nm (nitrogen laser) or 355 nm (frequency-tripled Nd:YAG laser) both protonated and neutral forms exhibit broad fluorescence spectra centered at ∼430 nm and ∼470 nm, respectively. Figure 1a displays the absorption (measured as excitation) and fluorescence spectra measured in neutral (black) and acidic (red) aqueous solution. The photophysics underlying the fluorescence spectra is summarized very nicely by Sayer et al. (32); here, we merely record the resolved fluorescence spectra to determine the relative acidity of the sample. The fraction of acridine (pKa ) 5.5 at 293 K) (33) that becomes protonated in aqueous solution depends on the pH of the solution. One can thus relate the fluorescence spectrum of acridine to the pH of the sample. This is illustrated in Figure 1b, which shows the ratio of the fluorescence intensity at 430 nm to that at 470 nm (hereafter called the 430/470 intensity ratio) as a function of solution pH, as measured using a pH meter. The result is essentially identical to that reported by Sayer et al. (32) and also to that used by us to measure pH changes at the aqueous-air interface (34). Experiments were carried out using a modification of the apparatus described in Kahan et al. (27) A microscope slide was greased with ∼1 mg Dow Corning silicone stopcock grease and 0.02 mL of a solution of 1.25 × 10-4 M acridine in 1-octanol was smeared onto the greased surface. The slide was mounted at the bottom of a Teflon chamber whose lid was removed for most of the work described here, allowing the sample to be exposed to ambient temperature and humidity conditions. The 355 nm output of a frequencytripled Nd:YAG laser impinged upon the sample at ∼750 from the surface normal. Fluorescence excited by the laser was collected by a 7 mm diameter liquid light guide positioned 3 cm above the sample and fed into a monochromator, then detected by a photomultiplier tube. The PMT signal was routed into a digital oscilloscope whose output was captured by a computer. An initial fluorescence spectrum of the sample was measured by manually scanning the monochromator in 4 nm increments over the range 390-414 nm and in 2 nm increments from 414 to 510 nm. The spectrum obtained in this manner was essentially identical to that observed from the neutral acridine species in aqueous solution. Following
FIGURE 1. (a) Absorption (dashed lines; measured as excitation) and fluorescence (solid lines) spectra of acridine in neutral (black curves) and acidic (red curves; pH 3.4) aqueous solutions. The fluorescence spectra were excited at 355 nm. Note the large red shift and somewhat increased intensity in the protonated form. (b) The ratio of the acridine fluorescence intensity at 430 nm to that at 470 nm measured in aqueous solution using 355 nm excitation. this wavelength scan, the monochromator wavelength was set to 580 nm, where there is no fluorescence from neutral or protonated acridine, and a baseline intensity value for the run was obtained. Approximately 5 mL of concentrated HNO3 or HCl solution was placed in a 50 mL Erlenmeyer flask and topped with a rubber septum. The acid was swirled and allowed to stand for at least 5 min. A plastic insulin syringe was inserted through the septum approximately 8 cm above the surface of the liquid and ∼0.5 mL of vapor was drawn up, then 0.2 mL was “puffed” onto a wetted piece of pH paper to test for the presence of acid. Generally a pH level of ∼3 was read from the paper. The remaining ∼0.3 mL of vapor was ejected from the syringe about 2 mm above the sample. A fluorescence spectrum of the sample was then measured using the same parameters as those used in recording the initial spectrum. Following these spectral measurements, the sample was irradiated for 15 min using either the full output of a 75 W Xe lamp, or the output filtered through one of a series of long-pass optical cutoff filters. Filters having 5% transmission at 295 nm (50% transmission at 305 nm) and with 5% transmission at 340 nm (50% at 355 nm) were used. The lamp was then turned off and a spectrum was measured as per the two previous spectra. The sample was irradiated a further 30 min and another spectrum was taken. This procedure of irradiation followed by measurement of a fluorescence spectrum was continued for up to a total of 2 h irradiation time. The temperature was not controlled; VOL. 41, NO. 11, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3899
however, its rise during the course of an experiment is estimated to be 295 nm. However, all samples showed much slower deprotonation rates than the corresponding samples exposed to nitric acid vapor.
Discussion The key results are that hydrophobic organic films take up gas-phase nitric acid which dissociates, at least to some extent, in the film to form H+ and NO3- and that both the proton and nitrate are removed when the film is exposed to actinic radiation (λ > 295 nm). This is not a consequence of a temperature increase during illumination, as the temperature rise was insignificant. Ramazan et al. (5) also report that NOx photoproduction from nitric acid adsorbed on glass surfaces is not a consequence of thermal desorption. Films exposed to HCl vapor also show deprotonation under actinic irradiation, but slower by about a factor of 5 than the nitric acid samples. Illuminating samples with the full output of the Xe lamp gives more rapid deprotonation; nitric acid deprotonation is, again, ∼3× faster than hydrochloric acid. The wavelength dependence, the lack of any significant temperature rise and the different results for HNO3 and HCl suggest that the (dissociated) nitric acid in the film is lost due to some photochemical mechanism. We argue below that it is recycled to the gas phase as NO2 and/or HONO. When the octanol-grease film was exposed to gas-phase nitric (or hydrochloric) acid, the acridine spectrum clearly showed the presence of protons in the film, indicating that the acids had dissociated, at least to some extent. The number of nitrate anions detected in the film was comparable to the number of acridine molecules; the 430/470 intensity ratios indicated a proton concentration equivalent to a pH of e5 in aqueous solution. Given the amounts of acid and acridine in the film, this suggests that a considerable fraction of the nitric acid (or HCl) could be present in dissociated form. Figure 3 displays an electronic absorption spectrum of ammonium nitrate, dissolved in octanol. Spectra of the same compound dissolved in water and also of acridine in octanol are shown as well. The intensities of the three spectra are scaled to the same solution concentration. The nitrate spectra in aqueous and octanol solution are essentially identical, except for a small enhancement in intensity of the octanol spectrum. This is further evidence that anionic nitrate does exist in octanol, a reasonably hydrophobic medium. These findings thus suggest that the nitrate, which is found in urban organic films (21), could arise from dry deposition of gasphase nitric acid and further, could be present in ionic form, implying acidification of the film. This could have important consequences for chemical reactions occurring in the film. Deprotonation and denitration are both observed to occur when the nitric-acid dosed film is irradiated with the Xe lamp output. After 2 h of irradiation through a 295 nm long-pass filter, ∼40% of the nitrate is lost; examination of Figure 2b indicates that the 430/470 intensity ratio has recovered to ∼60% of its initial value. Although this ratio is not a linear
FIGURE 3. Absorption spectra of ammonium nitrate dissolved in water (red trace) and in 1-octanol (black trace) and acridine dissolved in 1-octanol (blue trace; the intensity is reduced by a factor of 5 × 10-4). Also shown are the transmission spectra of the 295 nm long-pass filter (dot-dash line) and the 340 nm cutoff filter (dotted line). function of proton concentration, clearly there is significant loss of both H+ and total nitrate measured under the same conditions. Because the temperature rise is insignificant (
