Article pubs.acs.org/est
Effect of NO2 Concentration on Dimethylnitronaphthalene Yields and Isomer Distribution Patterns from the Gas-Phase OH Radical-Initiated Reactions of Selected Dimethylnaphthalenes Kathryn Zimmermann,†,‡ Roger Atkinson,†,‡,§,∥ and Janet Arey*,†,‡,§ †
Air Pollution Research Center, University of California, Riverside, California 92521, United States Environmental Toxicology Graduate Program, University of California, Riverside, California 92521, United States § Department of Environmental Sciences, University of California, Riverside, California 92521, United States ∥ Department of Chemistry, University of California, Riverside, California 92521, United States ‡
S Supporting Information *
ABSTRACT: Dimethylnitronaphthalene (DMNN) formation yields from the reactions of 1,7- and 2,7- dimethylnaphthalene (DMN) with OH radicals were measured over the NO2 concentration range 0.04−1.4 ppmv. The measured DMNN formation yields under conditions that the OH-DMN adducts reacted solely with NO2 were 0.252 ± 0.094% for Σ1,7-DMNNs and 0.010 ± 0.005% for Σ2,7-DMNNs. 1,7-DM-5-NN was the major isomer formed, with a limiting high-NO2 concentration yield of 0.212 ± 0.080% and with equal reactions of the adduct with NO2 and O2 occurring in air at 60 ± 39 ppbv of NO2. The reactions of the OH-DMN adducts with NO2 must therefore result in products other than DMNNs. Although the yields of the DMNNs are low, ≤0.3%, the DMNN (and ethylnitronaphthalene) profiles from chamber experiments match well with those observed in polluted urban areas under conditions where OH radical-initiated chemistry is dominant. Daytime OH radical and nighttime NO3 radical reactions appear to account for the alkylnitronaphthalenes formed and their observed profiles under most urban atmospheric conditions, with profiles reflecting daytime OH chemistry modified by contributions from isomers formed by any NO3 radical chemistry that had occurred. Since the formation yields and NO2 dependencies for the formation of a number of alkylnitronaphthalenes have now been measured, the effect of NOx emissions control strategies on their atmospheric formation can be quantitatively assessed, and the decrease in formation of these genotoxic species may provide a previously unrecognized health benefit of NOx control.
■
INTRODUCTION Dimethyl- and ethyl-naphthalenes (DMNs and ENs, respectively) have been measured in ambient air samples in southern California1,2 along with their corresponding nitro-derivatives.2 Nitro-polycyclic aromatic hydrocarbons (nitro-PAHs), including methylnitronaphthalenes (MNNs),3 have been shown to be direct-acting mutagens and are considered as probable human carcinogens,4 and it is thus important to study their sources and formation pathways. Sources of ambient nitro-PAHs include direct emissions and/or atmospheric gas-phase radical-initiated reactions of parent PAHs with OH (daytime) or NO3 (evening and nighttime) radicals,5 and in most ambient atmospheres radical-initiated formation has been found to dominate over direct emission.3,6,7 It has been shown that nitro-PAH isomer profiles from the gas-phase OH radical-initiated reactions of methylnaphthalenes5 and DMNs and ENs8 are distinct from the profiles resulting from gas-phase NO3 radical-initiated reactions, and ratios of 2-methyl-4-nitronaphthalene/1-methyl-5-nitronaphthalene and 2,7-dimethyl-4-nitronaphthalene/1,7-dimethyl-5© 2012 American Chemical Society
nitronaphthalene have been proposed as sensitive markers of nitro-PAH formation from NO3 versus OH radical chemistry.8 While the dimethylnitronaphthalene and ethylnitronaphthalene (DMNN and ENN, respectively) yields and NO2 dependence of the NO3 radical-initiated reactions of DMNs and ENs have been measured,8,9 DMNN yields from the OH radical-initiated reactions and whether or not formation of DMNNs by OH reaction of DMNs can occur at ambient NO2 concentrations remain to be determined. The formation of nitro-PAHs from the gas-phase OH radical reaction with DMNs and ENs is postulated to proceed by initial addition of the OH radical to the aromatic ring, forming a hydroxycyclohexadienyl-type radical (hereafter referred to as an OH-PAH adduct) as shown for 1,7-DMN in Scheme 1.5,10,11 The OH-DMN adduct can subsequently react with NO2 (to Received: Revised: Accepted: Published: 7535
March 12, 2012 June 20, 2012 June 21, 2012 July 3, 2012 dx.doi.org/10.1021/es3009826 | Environ. Sci. Technol. 2012, 46, 7535−7542
Environmental Science & Technology
■
Scheme 1. Potential Gas-Phase Mechanism for the Reaction of OH + 1,7-DMNa
Article
EXPERIMENTAL METHODS
Experiments were performed at ∼296 K and ∼735 Torr total pressure of dry purified air in a ∼7000 L collapsible Teflon chamber equipped with parallel banks of black lamps and a Teflon-coated fan for the rapid mixing of reagents upon their introduction into the chamber. The DMNs and ENs were introduced into the chamber by flushing known amounts of compound(s) from a heated Pyrex bulb with N2 gas, with initial mixing ratios in the chamber of 200−400 ppbv (1 ppbv =2.40 × 1010 molecules cm−3 at 296 K and 735 Torr). Hydroxyl radicals were generated by the photolysis of CH3ONO at wavelengths >300 nm, with NO being present to avoid formation of O3 and hence of NO3 radicals. Irradiations were carried out, for either 3 or 4 min, at a light intensity corresponding to an NO2 photolysis rate, J(NO2), of 0.14 min−1. These reactions resulted in 6−55% consumption of the initially present 1,7-, 2,7- and 1,6-DMN. The initial concentrations of CH3ONO and NO were equal in each experiment, ranging from 0.1 to 5 ppmv (see Supporting Information (SI) Table S1 for details of the specific experiments). The NO and initial NO2 concentrations were monitored using a chemiluminescence NO-NO2−NOx analyzer. Since these chemiluminescence analyzers read NO2 as a combination of NO2, CH3ONO, and any organic nitrates present in the system, the NO2 concentrations during the reactions were calculated assuming that NO + NO2 remained constant during each experiment,14 with this assumption being validated by computer modeling using a detailed chemical mechanism for irradiated CH3ONO−NO−toluene−air mixtures.17 DMNs and ENs were analyzed by gas-chromatography with flame ionization detection (GC-FID), with two or more replicate analyses prior to and after each reaction. Samples of 100 cm3 volume were collected from the chamber onto Tenax-TA solid adsorbent, with subsequent thermal desorption at 270 °C onto a 30 m DB-5 megabore column, initially held at 40 °C and then programmed at 8 °C min−1. GC-FID response factors for the DMNs were determined by adding a known concentration of authentic standards in methanol onto Tenax cartridges, followed by GC-FID analyses. Replicate analyses of the DMNs and ENs in the chamber in the dark agreed to within 6% and showed no evidence for dark decays. DMNNs and ENNs were sampled from the chamber for 3.6 min using a modified high volume sampling apparatus consisting of two polyurethane foam (PUF) plugs in series upstream of a Teflon-impregnated glass fiber filter. The chamber volumes sampled were calculated from the NOx concentrations measured before sampling and again after sampling and backfilling with purified air, and ranged from ∼4380−5510 L (SI Table S1). Sampling for DMNNs and ENNs was carried out ∼45 min after the completion of each reaction, with ≤15% wall loss being expected based on previous data.9 For a number of experiments, after sampling was complete the chamber was refilled with dry purified air and equilibrated for 30 min, followed by another sampling. The results showed no evidence for any desorption of DMNNs from the walls. The PUF plugs and filter from each experiment were Soxhlet extracted in dichloromethane (DCM) and 2,6-DM-1-NN was added as an internal standard prior to extraction. Extracts of the PUF plugs and filter were analyzed separately in an initial experiment, which showed ∼23% of the 1,7-DMNNs were present on the filter. In all subsequent experiments, the PUF
a
The rate constants k1-k3 correspond to reactions 1-3 in the text. Although the adduct shown is expected to lead to the formation of 1,7DM-5-NN, the most abundant 1,7-DMNN observed, the “other adducts” may be equally or more important, and ring-opened dicarbonyls from these other adducts have been reported as major products from the OH reaction with 1,7-DMN.12 Note also that k2 and k3 need not be the same for all of the OH-1,7-DMN adducts.
form DMNNs and other products) or with O2 (Scheme 1).5 Studies have shown that for the monocyclic aromatic hydrocarbons benzene and toluene, the OH-aromatic adducts react in air equally with O2 and NO2 at NO2 mixing ratios of 1.2 and 3.3 ppmv, respectively,13 concentrations far exceeding those expected in ambient atmospheres. Therefore, nitrobenzene and nitrotoluenes are not expected to be formed to any significant extent under atmospheric conditions,14 and for monocyclic aromatics the yields of nitro-products determined from laboratory experiments performed at high NO 2 concentrations are not representative of what occurs in the atmosphere.5 However, for naphthalene it has been shown that reactions of the OH-naphthalene adducts with NO2 and O2 are of equal importance in air at ∼60 ppbv of NO2,14 a concentration often reached in polluted urban atmospheres.15,16 In this study, we have measured the DMNN formation yields from the reactions of 1,7- and 2,7-DMN with OH radicals over the NO2 concentration range 0.04−1.4 ppmv. These DMNs were chosen because, as noted, their nitro-products are expected to be useful in distinguishing between OH and NO3 radical formation.8 In addition, the DMNN yields were measured from 1,6-DMN at 0.06 and 1.2 ppmv NO2. This study also examined the isomer distribution profiles of alkylnitronaphthalenes formed in the laboratory from OH radical-initiated reactions of a mixture of ENs and DMNs as a function of NO2 concentration (0.05−1.4 ppmv) and compared these nitro-PAH isomer profiles to those of ambient gas-phase samples from Mexico City. 7536
dx.doi.org/10.1021/es3009826 | Environ. Sci. Technol. 2012, 46, 7535−7542
Environmental Science & Technology
■
plug and filter extracts were combined for analysis. Extracts were concentrated and eluted through a silica solid phase extraction column (Supelco Discovery DSC-Si, 6 mL, 1 g) with DCM. The extracts were then concentrated by rotary evaporation and fractionated using normal-phase high performance liquid chromatography (HPLC) as described previously.18 An experiment in which 2,6-DM-1-NN was subjected to solid phase extraction and HPLC fractionation showed 73% recovery. DMNNs and ENNs were analyzed with gas chromatography−mass spectrometry with negative chemical ionization (GC-MS/NCI) in select ion monitoring (SIM) mode with methane as the reagent gas, using an Agilent Technologies 6890N GC interfaced to a 5975XL quadrupole mass detector with a 30 m HP-5MS capillary column (0.25 mm i.d., 0.25 μm phase). Reaction products were also analyzed by GC-MS with electron impact ionization (EI) in total ion count mode. DMNNs and ENNs were identified based on the retention times of selected synthesized standards, as well as their EI fragmentation patterns.19 NCI response factors of the 1,7-, 2,7-, and 1,6-DMNN isomers were calculated relative to that of the internal standard, 2,6-DM-1-NN, by comparing their responses from GC-FID analyses, which are assumed to be identical for a series of isomers,20 with those observed in NCI analyses. The NCI responses of individual isomers, relative to that of 2,6-DM-1-NN, ranged from 0.3 to 2.5. Hydroxyl radicals were also generated from the dark reaction of 2,3-dimethyl-2-butene with O3,21 in the presence of NO2 (SI Table S2; see also Experiments using O3 + 2,3-dimethyl-2butene to generate OH radicals in the Supporting Information). Reaction of ENs + DMNs Mixture with OH Radicals. A mixture of ENs and DMNs, designed to mimic relative ambient concentrations observed in southern California,2 was introduced into the chamber as a solution in methanol. Relative concentrations in this surrogate ambient mixture were; 1-EN, 0.07; 2-EN, 0.16; 1,2-DMN, 0.04; 1,3-DMN, 0.06; 1,4-DMN, 0.05; 1,5-DMN, 0.04; 1,6-DMN, 0.18; 1,7-DMN, 0.10; 2,3DMN, 0.04; 2,6-DMN, 0.13; and 2,7-DMN, 0.13. 1,8-DMN was not present in this mixture since this isomer has not been observed from ambient samples1,2 or in diesel fuel.22 OH radicals were generated as described above and the experimental details, including the average NO2 mixing ratios (0.05−1.4 ppmv), irradiation times, and the percent of the Σ(DMNs + ENs) reacted, are given in SI Table S1. Samples were collected from the chamber, extracted, and analyzed as described above, without the addition of an internal standard. Ambient Measurements of Ethyl- and Dimethylnitronaphthalenes. Archived data from the analyses of HPLC fractions of gas-phase samples collected on PUF plugs in Mexico City during the MCMA campaign (April 25−30, 2003),7,23 and in Riverside CA (August 26−30, 2002)2 were reanalyzed for comparison with profiles obtained from the chamber experiments. Chemicals. The chemicals used, and their started purities were as follows: 1-EN (98+%), 2-EN (99+%), 1,2-DMN (98%), 1,3-DMN (96%), 1,4-DMN (95%), 1,5-DMN (98%), 1,6-DMN (99%), 1,7-DMN (99%), 2,3-DMN (98%), 2,6DMN (99%), 2,7-DMN (99%), and 2,3-dimethyl-2-butene (98%), Aldrich; 2,6-DM-1-NN, synthesized;19 methanol, hexanes, dichloromethane and acetonitrile (all Optima grade), Fisher Scientific; and NO (≥99.0%), Matheson Gas Products. CH3ONO was prepared and stored as described previously.17
Article
RESULTS AND DISCUSSION Yields of Dimethylnitronaphthalenes from the Reactions of Individual DMNs with OH Radicals in the Presence of NOx. To ensure complete chromatographic resolution of the DMNN products, a series of experiments were conducted with 1,7-DMN and 2,7-DMN individually at various average NO2 mixing ratios. Additional experiments were then carried out with mixtures of 2,7- + 1,7-DMN, and 2,7- + 1,6DMN. There are six possible 1,7-DMNN isomers and all six were observed from the 1,7-DMN reaction, but with 1,7-DM-5NN accounting for 80−90% of the total 1,7-DMNNs. The measured yields of 1,7-DM-5-NN, the three 2,7-DMNN isomers, and the three major 1,6-DMNN isomers (75−85% of the total 1,6-DMNNs) are given in SI Table S3, and the isomer distributions are given in SI Table S4. The DMNNs are expected to be less reactive than their parent DMNs toward OH radicals10 and, based upon the photolysis rates measured for the MNNs,24 the light intensity and irradiation times used here would result in only minor photolysis losses of the DMNNs. Hence secondary reactions were assumed to be of minor or negligible importance. By analogy with alkylbenzenes,10 H-atom abstraction from the DMNs and ENs is expected to account for ≤10% of the overall OH radical reaction. The reactions of DMNs with the OH radical to form DMNNs proceed as follows: OH + DMN → α OH‐DMN adduct
(1)
OH‐DMN adduct + NO2 → β DMNN + other products (2)
OH‐DMN adduct + O2 → products
(3)
where α is the yield of a specific OH-DMN adduct from reaction 1 and β is the formation yield of a specific DMNN isomer from reaction 2. Therefore, the measured DMNN yield of each specific DMNN isomer is given by ⎛ ⎞ ⎛ [NO2 ] ⎞ k 2[NO2 ] Y (%) = γ ⎜ ⎟ = γ⎜ ⎟ ⎝ [NO2 ] + b ⎠ ⎝ k 2[NO2 ] + k 3[O2 ] ⎠
(I)
in which Y is the measured yield of the DMNN isomer(s), k2 and k3 are the rate constants for reactions 2 and 3, respectively, γ = αβ, and b = k3[O2]/k2 and is the NO2 concentration at which reactions 2 and 3 are equally important. The measured yields of 1,7-DM-5-NN as a function of NO2 concentration are plotted in Figure 1 (smaller filled symbols). The yield increases with NO2 concentration, approaching a maximum where reaction 2 dominates over reaction 3. The dashed line shown is from eq I with γ and b obtained through a least-squares analysis of 1/Y against 1/[NO2] for 1,7-DM-5NN (SI Figure S1), where 1/γ is the intercept and b/γ is the slope. The yield of 1,7-DM-5-NN from the OH radical reaction with 1,7-DMN is γ = 0.21%. Thus, even when reactions of the OH-1,7-DMN adducts with NO2 dominate, 1,7-DM-5-NN is only a minor product. The calculated NO2 concentration for which the OH-1,7-DMN adduct leading to 1,7-DM-5-NN reacts equally with NO2 and O2 is b ∼60 ppbv (Table 1), the same value as measured for the OH-naphthalene adduct.14 Concentrations of NO2 in this range can be found in ambient polluted atmospheres,15,16 confirming that formation of DMNNs and ENNs via OH radical-initiated reaction will occur in these polluted, generally urban, atmospheres. Calculated γ- and b-values for the major DMNN isomers and for the sum of the DMNN isomers formed from 1,7-, 2,7-, and 7537
dx.doi.org/10.1021/es3009826 | Environ. Sci. Technol. 2012, 46, 7535−7542
Environmental Science & Technology
Article
Interestingly 2,7-DM-4-NN appears to have a yield nearly independent of NO2, suggesting that for this isomer b ≤ 36 ppbv NO2 (but see below). The relatively high yield of 1,7DM-5-NN (Table1) from reaction of the OH-1,7-DMN adducts with NO2 is consistent with ambient profiles of DMNNs and ENNs in which 1,7-DM-5-NN dominates in ambient samples characterized by OH chemistry,8 as discussed below. ENN and DMNN Profiles from Reactions of the Surrogate Ambient Mixture. Of the possible 53 ENN and DMNN isomers formed from these reactions (since 1,8-DMN was not included), only 33 resolved chromatographic peaks were observed resulting in many coeluting isomers. SI Figure S3 shows the excellent reproducibility in isomer profiles between replicate reactions of the surrogate ambient DMNs + ENs mixture at high NO2 concentrations, and SI Table S5 gives the identifications of the DMNN and ENN isomers numbered on SI Figure S3. Figure 2 shows chromatograms of the surrogate ambient mixture reacted at three different average NO2 concentrations and, for comparison, an ambient morning sample collected in Mexico City. Clearly 1,7-DM-5-NN is the most abundant isomer in each trace; however the relative abundances of the other isomers compared with 1,7-DM-5-NN vary. Table 2 gives the fraction of the 1,7-DM-5-NN peak area to the total area of the DMNNs + ENNs from the chamber reactions. At 1.3−1.4 ppmv NO2 the 1,7-DM-5-NN peak represents ∼20% of the total area. At lower NO2 concentrations, 1,7-DM-5-NN represents 30−40% of the total area. This suggests that the NO2 concentration at which the OH-1,7DMN adduct leading to 1,7-DM-5-NN reacts equally with O2 and with NO2 is lower than for many other isomers. This is consistent with the lower value of b for 1,7-DM-5-NN than for 2,7-DM-1-NN and 1,6-DM-5-NN (Table 1). These latter isomers are labeled in Figure 2 and have no coelutions with other DMNN or ENN isomers.19 Therefore, for exposures of a mixture including a constant ratio of 1,7-, 2,7-, and 1,6-DMN, the ratios of 1,7-DM-5-NN to 2,7-DM-1-NN and 1,6-DM-5NN can be obtained, and these are also given in Table 2. At the lower NO2 concentrations, the 1,7-DM-5-NN isomer is more dominant, consistent with this isomer having a low b value. As previously noted, the ratio of 1,7-DM-5-NN/2,7-DM-4NN has been suggested as a potential marker of OH versus NO3 chemistry8 and as shown in Table 2 the ratio obtained from the DMNs + ENs mixture reactions is independent of the NO2 concentration. Predicted ratios of 1,7-DM-5-NN/2,7-DM4-NN are also given in Table 2, using the individual values of γ and b obtained from experiments with 1,7- and 2,7-DMN (Table 1), the OH radical reaction rate constants for 1,7- and 2,7-DMN, and their relative abundances (see Table 3). The predicted ratio at 50 ppbv of NO2 is about a factor of 2 higher than that observed from the ENs + DMNs mixtures, and the predicted ratios increase with increasing NO2. The NO2independent ratios observed from the ENs + DMNs mixtures suggest that the NO2 dependencies for formation of 1,7-DM-5NN and 2,7-DM-4-NN are similar and that our value of b for 2,7-DM-4-NN in Table 1 is too low. Mechanistic Implications. In Scheme 1, the formation of nitro-PAH (1,7-DM-5-NN in this particular case) is shown as involving NO2 addition to the OH-PAH adduct to form the intermediate A, followed by elimination of H2O to generate the nitro-PAH. Note that formation of certain isomers through this mechanism, for example 1,7-DM-8-NN, would require ipso
Figure 1. Plots of the measured formation yields of 1,7-DM-5-NN, the dominant isomer formed from the 1,7-DMN + OH reaction, Σ1,7DMNNs, Σ2,7-DMNNs, and Σ1,6-DMNNs formed from reactions of 1,7-, 2,7-, and 1,6-DMN, respectively, with OH radicals against the average NO2 concentration. The dashed line is a fitted curve based on eq I with b = 60 ppbv and a yield of 1,7-DM-5-NN from reactions (1) and (2) of γ = 0.212%. ■, experiments conducted with only 1,7-DMN present; ●, experiment conducted with 1,7-DMN + 2,7-DMN. The solid lines are fitted curves based on eq I with the values of γ and b calculated for the sum of the isomers (Table 1; see SI Table S4 for the isomer distributions). □, experiments conducted with only 1,7-DMN present; Δ, experiments with only 2,7-DMN present; ○ and ◊ are data for Σ1,7-DMNNs and Σ2,7-DMNNs, respectively, from an experiment conducted with 1,7- + 2,7-DMN. ▽ and ▼ are data for Σ2,7-DMNNs and Σ1,6-DMNNs, respectively, from experiments conducted with 2,7+ 1,6-DMN.
Table 1. Calculated Values of γ and b for the 1,7-, 2,7- and 1,6-DMN Reactions with OH in the Presence of NOx, Using Photolysis of CH3ONO to Generate OH Radicals 1,7-DM-5-NN Σ1,7-DMNN 2,7-DM-1-NN 2,7-DM-3-NN 2,7-DM-4-NN Σ2,7-DMNN 1,6-DM-3-NN 1,6-DM-4-NN 1,6-DM-5-NN Σ1,6-DMNN
yield γ (%)a
b (ppbv)a
0.212 ± 0.080 0.252 ± 0.094 0.0037 ± 0.0025 0.0010 ± 0.0004 0.0046 ± 0.0018 0.0103 ± 0.0050 0.016b 0.012b 0.0054b 0.036b
60 ± 39 66 ± 40 127 ± 103 107 ± 47 ≤36 (4 ± 32)c 44 ± 46 585b 187b 105b 187b
a
Indicated errors are two least-squares standard deviations. bValues of γ and b for the 1,6-DMNN isomers were calculated from only two data points. cLeast-squares values are given in parentheses.
1,6-DMN are listed in Table 1. Figure 1 also shows plots of the yields of Σ1,7-DMNNs, Σ 1,6-DMNNs and Σ 2,7-DMNNs as a function of NO2, clearly showing that Σ1,7-DMNNs (∼90% being 1,7-DM-5-NN) has the highest yield of the three DMNs studied. Individual DMNN isomers have different measured values of b and SI Figure S2 shows plots of the measured yields as a function of NO2 for the three 2,7-DMNN isomers. As seen from Table 1, although the errors are large, 2,7-DM-1-NN seems to have a higher b value than 1,7-DM-5-NN. 7538
dx.doi.org/10.1021/es3009826 | Environ. Sci. Technol. 2012, 46, 7535−7542
Environmental Science & Technology
Article
Figure 2. GC-MS/NCI mass chromatograms of the molecular weight 201 nitro-PAHs from: reactions of the ENs + DMNs surrogate ambient mixture with OH radicals at average NO2 mixing ratios of (top left) 1.4 ppmv; (top right) 0.3 ppmv; (bottom left) 0.05 ppmv; and (bottom right) from analysis of the nitro-PAH-containing HPLC fraction from an ambient gas-phase sample collected in Mexico City on April 25, 2003 from 0700 to 1100 h local time. Peak identifications are (1) 2,7-DM-1-NN; (2) 1,6-DM-5-NN; (3) 2,7-DM-4-NN; and (4), 1,7-DM-5-NN. Note that in these analyses, 2,7-DM-4-NN coelutes with 1,7-DM-2-NN (with 2,7-DM-4-NN/1,7-DM-2-NN ∼2.5 at 1.3 ppmv NO2) and 1,7-DM-5-NN coelutes with 1,6-DM-2-NN (with 1,7-DM-5-NN/1,6-DM-2-NN ∼80 at 1.3 ppmv NO2).
Table 2. Ratios of the 1,7-DM-5-NN Formed to the Σ(ENNs + DMNNs) and to Selected Isomers Formed from the OH Reactions with the DMNs + ENs Mixture at Varying Average NO2 Concentrations experiment no. ITC ITC ITC ITC
4042 4046 4049 4047
average NO2 (ppmv)
1,7-DM-5-NN/ Σ(ENNs + DMNNs)a
1,7-DM-5-NN/2,7DM-1-NNb
1,7-DM-5-NN/1,6DM-5-NNb
1,7-DM-5-NN/2,7DM-4-NNb
predictedc 1,7-DM-5-NN/ 2,7-DM-4-NN
1.3 1.4 0.3 0.05
0.19 0.21 0.36 0.33
20 23 44 86
22 23 38 42
7.9 8.0 8.3 8.3
34 34 30 17
a
Ratios are based on area counts with no correction for differences in NCI response factors. bThe measured NCI response factors relative to that for 2,6-DM-1-NN were: 1,7-DM-5-NN, 0.6; 2,7-DM-1-NN, 1.2; 1,6-DM-5-NN, 2.5; 2,7-DM-4-NN, 0.3. cSee text for full discussion.
addition12,13 of the OH (in this case to C7). If the OH-PAH adducts react with NO2 exclusively to form intermediates analogous to A followed by elimination of H2O to yield the nitro-PAH (pathway I), then at sufficiently high NO 2 concentrations such that all OH-PAH adducts react with NO2, the sum of the yields of the individual nitro-PAH isomers
would be 100%. In general, the yields of nitro-PAH from the OH + PAH reactions at high NO2 concentrations, where the OH-PAH adducts + NO2 reaction dominates, are low or very low, ranging from a high of ∼10% for the OH-benzene, OHtoluene and OH-biphenyl adducts14,27 down to ∼1% for the OH-naphthalene adducts14 and to ≤0.3% for the OH-DMN 7539
dx.doi.org/10.1021/es3009826 | Environ. Sci. Technol. 2012, 46, 7535−7542
Environmental Science & Technology
Article
Table 3. Parameters Used to Calculate the Formation Rates of 1,7-DM-5-NN and 2,7-DM-4-NN from Reactions of 1,7-DMN and 2,7-DMN with OH Radicals (Formation Rate = kY[OH][DMN]) and NO3 Radicals (Formation Rate = k[NO2] Y[NO3][DMN]), where Y is the DMNN Yield at the NO2 Concentration Considered relative abundances of parent DMN OH radical reactiona k (cm3 molecule−1 s−1) γ (%) b ppbv formation rate at 40 ppb NO2c formation rate at 100 ppb NO2c NO3 radical reactiond k (cm6 molecule−2 s−1)e γ (%)f b ppbvf formation rate at 40 ppb NO2c formation rate at 100 ppb NO2c
1,7-DM-5-NN formation
2,7-DM-4-NN formation
1.0
1.3
6.79 × 10−11 0.212 60 1.15 × 10−7 [1,7-DMN] 1.80 × 10−7 [1,7-DMN]
6.87 × 10−11 0.0046 4b 7.47 × 10−9 [1,7-DMN] 7.90 × 10−9 [1,7-DMN]
1.35 × 10−27 6.18 330 4.33 × 10−9 [1,7-DMN] 2.33 × 10−8 [1,7-DMN]
2.10 × 10−27 18.2 130 5.61 × 10−8 [1,7-DMN] 2.59 × 10−7 [1,7-DMN]
a Values of γ and b (see eq I and associated text) are from this work. The rate constants for the OH radical reactions are from ref 25. A daytime OH radical concentration of 2 × 106 molecules cm−3 was used.26 bThis value of b is that from a least-squares analysis of the data for 2,7-DM-4-NN in the form 1/Y against 1/[NO2], as for 1,7-DM-5-NN in SI Figure S1, with b = 4 ± 32 ppbv. cFormation rate in molecules cm−3 s−1. dA nighttime NO3 radical concentration of 5 × 108 molecules cm−3 was used.26 eBecause of rapid back-reaction of the NO3−PAH adducts to reactants, the effective rate constant for reaction of PAH with NO3 radicals is given by keffective = k[NO2] cm3 molecule−1 s−1.5 The rate constants for the NO3 radical reactions are from ref 1. fThe value of b for 2,7-DM-4-NN is that cited in ref 9. The value of b for 1,7-DM-5-NN and the values of γ are obtained from the ratios of the DMNN yields at 0.23 and 1.87 ppmv NO29 [thereby allowing the value of b for 1,7-DM-5-NN to be derived] and the measured 1,7DM-5-NN and 2,7-DM-4-NN yields at 1 ppmv NO2 given by ref 8 [thereby allowing γ to be derived using eq I and these values of b].
may be epoxides.12 The formation of ring-opened dicarbonyls from the OH-DMN + NO2 reactions must therefore arise from other pathways of intermediate A (or analogous intermediates containing >C(OH)−C(NO2)< moieties) (pathway (II) in Scheme 1) or from other intermediates B containing different structural units than intermediate A (such as, for example, OHDMN-O• alkoxy radicals). Additionally, OH radical addition can also occur at other positions, for example at C2 next to the methyl on the other ring and in this case the analogous pathway I gives a 1,7-DM-3-NN yield lower than that of 1,7-DM-5-NN. In the absence of NO, reactions of the OH-PAH adducts with species other than O2 and NO2, such as HO2 and/or organic peroxy radicals,34 O3, and N2O5 and/or NO3 radicals, may be significant. These additional reactions of the OH-PAH adducts may explain, at least in part, the present observations of variable nitro-isomer distributions from the DMNs and from the DMNs + ENs mixture when using the O3 + 2,3-dimethyl-2butene reaction to generate OH radicals (see Supporting Information, including Figures S4−S7, for details), as well as those of Nishino et al.14 of elevated 1- and 2-nitronaphthalene yields from the OH radical-initiated reaction of naphthalene under experimental conditions where N2O5 formation may have occurred when using photolysis of CH3ONO or the dark O3 + 2-methyl-2-butene reaction to generate OH radicals. Implications for Ambient Nitro-PAH Formation. It has previously been observed that the ENNs + DMNNs profile formed from the NO3 radical-initiated reaction does not show significant variations with NO2 concentrations ranging from 0.2 to 1.9 ppmv.9 Relative formation rates of 1,7-DM-5-NN and 2,7-DM-4-NN (the latter being the isomer formed in highest yield from the NO3 radical reaction8,9) can now be derived for both OH and NO3 radical-initiated chemistry. Table 3 gives the rate constants1,25 and other parameters used to calculate these formation rates at 40 and 100 ppbv NO2. The calculated ratios of 1,7-DM-5-NN/2,7-DM-4-NN obtained from the respective formation rates (i.e., neglecting any differences in loss processes) are 15 and 23 for daytime OH radical reaction at
adducts studied here. Therefore, other pathways, via pathway II in Scheme 1 and/or through formation of other intermediates (denoted by B in Scheme 1), must also occur. Consistent with the formation of nitro-PAH by H2O elimination from intermediate(s) A being at most a minor channel of the overall OH-PAH adduct + NO2 reaction, the elimination process has a calculated barrier height of 46−47 kcal mol−1 for the intermediates formed after addition of the OH radical to the 1- or 2-positions of naphthalene.28,29 It is possible that nitro-PAH formation occurs through a chemically activated intermediate A (as indicated in Scheme 1) which initially will have at least 40−48 kcal mol−1 excess energy resulting from addition of NO2 to the OH-naphthalene adducts,28 ≤7 kcal mol−1 lower than the calculated barrier heights.28 With this interpretation, nitronaphthalene formation (and nitro-PAH formation in general) would occur via a “prompt” decomposition of the chemically activated intermediate A, in competition with collisional stabilization of A and/or other reactions of chemically activated A or other intermediates B to form products other than nitro-PAH. While it could be argued that H2O elimination from A occurs heterogeneously (for example, at chamber walls or on aerosol surfaces as occurs for the dehydration of the cyclic hemiacetals formed from 1,4-hydroxycarbonyls30,31), the observation of apparently gas-phase nitro-PAH such as 2-nitrofluoranthene during chamber OH radical-initiated reactions of fluoranthene (ref 7 and references therein) suggests that this is not the case. In the case of naphthalene, 2-formylcinnamaldehyde has been shown to be formed in 56% yield from the reaction of the OH-naphthalene adduct with NO2 and in 14% yield from reaction of the adduct with O2.32,33 While they have not been quantified, the corresponding ring-opened dicarbonyls of molecular weight 188 have been observed under high NOx conditions from the OH radical-initiated reactions of DMNs and ENs.12 Also observed were lower molecular weight products resulting from the loss of two β-carbons and associated alkyl groups, and ring-containing compounds that 7540
dx.doi.org/10.1021/es3009826 | Environ. Sci. Technol. 2012, 46, 7535−7542
Environmental Science & Technology
Article
NO2/HNO3 mixtures. Also, the 2-NFL/1-NPY ratios (which indicate the importance of atmospheric formation versus direct emissions7) given in SI Table S6 show that atmospheric formation of nitro-PAHs dominates over direct emissions in both Mexico City and Riverside.7 In composite evening and nighttime samples from Riverside where the 2-NFL/2-NPY ratios (100 and 87, respectively) indicate the occurrence of NO3 chemistry,2 the 1,7-DM-5-NN/ 2,7-DM-4-NN ratios were 2.1 and 1.7, respectively, significantly lower than those in Mexico City (SI Table S6), but not the
