Environ. Sci. Technol. 2007, 41, 8362–8369
Experimental Confirmation of the Dicarbonyl Route in the Photo-oxidation of Toluene and Benzene E . G Ó M E Z A L V A R E Z , * ,† J . V I I D A N O J A , †,§ A . M U Ñ O Z , † K . W I R T Z , †,# A N D J . H J O R T H * ,⊥ Fundación Centro de Estudios Ambientales del Mediterráneo (CEAM), C/Charles Darwin 14, 46980 Paterna, Valencia, Spain, and Institute for Environment and Sustainability (IES), The European Commission, Joint Research Centre (JRC), TP290, 21020 Ispra (VA), Italy
Received June 05, 2007. Revised manuscript received September 20, 2007. Accepted September 27, 2007.
The methodology of solid phase microextraction (SPME) with O-(2,3,4,5,6)-pentafluorobenzylhydroxylamine hydrochloride (PFBHA) on-fiber derivatization for the determination of carbonyls has been applied to the photo-oxidation of benzene and toluene carried out in the EUPHORE chambers. This work focuses on the results obtained for a number of highly reactive carbonyls, crucial in the determination of branching ratios and confirmation of the carbonylic route. The observed yields and kinetic behavior were compared to simulations with the Master Chemical Mechanism model, version 3.1 (MCMv3.1). The following yields were measured in the toluene system: glyoxal, (37 ( 2)%; methylglyoxal, (37 ( 2)%; 4-oxo-2-pentenal, >(13.8 ( 1.5)%; and total butenedial, (13 ( 7)% [cisbutenedial, (6 ( 3)%; trans-butenedial, (7 ( 4)%]. For benzene, the experimental glyoxal yields were (42 ( 3) and (36 ( 2)% for the two successive experiments (September 24 and 25, 2003), (17 ( 9)% for total butenedial [(8 ( 4)% cis-butenedial and (9 ( 5)% trans-butenedial (September 24, 2003)] and (15 ( 6)% total butenedial (September 25, 2003) [(7 ( 3) and (7 ( 3)% for the cis and trans isomers, respectively]. PTR-MS estimations for butenedial also allowed the two isomers of butenedial to be distinguished, but the measurements showed signs of interference from other products. The results presented confirm the fast ring cleavage and provide further experimental confirmation of the dicarbonylic route.
Introduction It is known that the photo-oxidation of toluene proceeds via several channels that result in either ring-retaining or ringopening products. An estimate of the contribution of the * Address correspondence to either author. Phone: 00-34-961318227 (E.G.A.); 00-39-0332-789076 (J.H.). Fax: 00-34-96-1318190 (E.G.A.); 00-39-0332-785837 (J.H.). E-mail:
[email protected] (E.G.A);
[email protected] (J.H.). † Fundación Centro de Estudios Ambientales del Mediterráneo (CEAM). § Present address: PerkinElmer Life and Analytical Sciences, P.O. Box 10, FIN-20101, Turku, Finland. # Present address: Umweltbundesamt, Wörlitzer Platz 1, 06844 Dessau, FG II5.4 Referenz Laboratory/Pilotstation, Germany. ⊥ Institute for Environment and Sustainability (IES). 8362
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 24, 2007
different pathways, based on the outcome of simulation chambers and other laboratory experiments, is provided by the Master Chemical Mechanism (MCM), which has been updated on the basis of results from the EXACT project (MCMv3.1) (1). The dicarbonyl and epoxides routes are described as the major ones with branching ratios of 0.65 and 0.10, respectively, although these figures are uncertain. Figure 1 presents a summary of the main routes predicted by MCMv3.1 in the oxidation of toluene. The central routes in the figure involve ring opening and dominate ozone production. In this study, we will focus on the products of the so-called dicarbonylic route (surrounded in the figure by a square). Figure 1 presents an overview of the representation in the MCM. However, it should be borne in mind that this illustration only summarizes some of its main aspects. Furthermore, the MCM also contains some simplification in representation and visualization. For the discussion that follows, a number of facts should be taken into consideration: (i) Three cresol isomers are actually formed, with the (major) o-cresol isomer depicted in the picture as a representative in MCM. (ii) Four possible isomeric hydroxycyclohexadienyl peroxy radicals are actually formed. The one shown is a representative visualization in the MCM, but is assumed to react to form the various product pairings in the dicarbonyl route based on consensus literature yields (strictly, the structure shown could yield methylglyoxal, but not glyoxal). (iii) Furanone coproducts are assumed to be partly formed with the R-dicarbonyls in MCM, to account for the reported inequality of R-dicarbonyl and unsaturated γ-dicarbonyl yields (as noted elsewhere in the present work). Although the two main ring-retaining routes, which form cresols and benzaldehyde as first-generation products (2–8) are well understood, a number of aspects remain uncertain regarding the chemistry of the ring-opening routes (7–16). This uncertainty is caused by a number of factors, among which the lack of good analytical techniques for the quantification of the γ-dicarbonyls, a lack of commercially available standards for many of the compounds, their very high reactivity (which means that they will be produced only in low concentrations), and probably also the importance of different experimental conditions, can be mentioned. The formation of carbonyls and phenol from the reaction of OH radicals with benzene is studied in the work by Berndt et al. (17). The experiments were conducted with a large excess of benzene [(8.9-118) × 1014 molecules cm-3]. For investigations in the presence of NOx, the conditions were chosen so that >95% of the OH/benzene adduct reacted with O2 even for the highest NO2 concentration occurring in the experiment. Detected carbonyls were glyoxal, cisbutenedial, and trans-butenedial with formation yields regarding converted benzene of 0.27 ( 0.06, 0.08 ( 0.02, and 0.023 ( 0.007, respectively, in the presence of NOx. Nondependence of the relative product yields of the carbonyls on the NOx concentration was reported. In the present work, two complementary techniques to measure the carbonyls are used that may aid in advancing the present knowledge of degradation mechanisms of aromatic compounds: SPME for identification and quantification of carbonyls and PTR-MS for high time resolution, which is particularly suitable for model comparison purposes. The majority of the data reported concern the toluene reaction. 10.1021/es0713274 CCC: $37.00
2007 American Chemical Society
Published on Web 11/16/2007
FIGURE 1. Toluene oxidation routes in MCMv3.1. The paths in the middle involve ring opening producing active photochemical intermediates and maximum ozone formation (courtesy of M. Pilling, University of Leeds).
TABLE 1. Summary of Initial Conditions of the Experiments date
precursor
concn (ppb)
HONO, initial (ppb)
NO, initial (ppb)
NO2, initial (ppb)
Sept 24, 2003 Sept 25, 2003 Sept 26, 2003
benzene benzene toluene
1998 ( 6 3912 ( 43 1672 ( 33
34 ( 3 64 ( 6 31 ( 3
6.3 ( 0.6 14.4 ( 1.4 115 ( 11
20 ( 2 37 ( 4 24 ( 2
Experimental Section The experiments were performed in the EUPHORE chambers, an outdoor chamber (i.e., irradiated by sunlight) of approximately 200 m3 in volume. Description of some aspects of the chambers can be found in refs 8 and 18. Floor and air temperatures inside the chamber were measured with two PT-100 thermocouples. Ozone, NOx, temperature, radiation, and humidity data were collected and saved in a data acquisition system. For a detailed technical description of the EUPHORE chambers, see the EUPHORE Reports (19–22). PDF copies of the reports can be downloaded from the Web site http://www.physchem.uni-wuppertal.de/PC_WWW_Site/ Publications/Publications.html. In each of the runs, a predetermined amount of a single aromatic compound (benzene or toluene) was introduced in the chamber using a calibrated syringe, heated, and carried by an air stream. Table 1 summarizes the conditions of the experiments. All concentrations provided were corrected for the actual temperature and pressure values at the moment of the introduction. Nitrous acid was generated by the aqueous-phase reaction of sodium nitrite (1.5%) with sulfuric acid (30%). The mixtures of HONO, NO, and NO2 produced by this reaction were transferred directly into the chamber by a stream of purified air. NO was added to the mixture on September 26, 2003 to delay the formation of ozone in the system.
The reactants were allowed to mix for at least 30 min before the chamber was exposed to sunlight. To keep the chamber inflated, purified dry air is added to compensate for system leakage. The dilution was measured following the decay of SF6, an initially added (25 ppbV) inert tracer, by means of FTIR. The decay of the parent aromatic compounds was monitored every 5 min by GC-FID (Hewlett-Packard 6890) and GC-PID (Fisons 8130) (see experimental conditions below). Calculation of parent compound lost by wall losses and dilution was performed from the precursor decay before any chemical reaction takes place, that is, before the chamber was opened. Prevalent trans-butenedial standard was synthesized by a collaborative work between University of Cork, Ireland (UCC), and University of NewCastle upon Tyne. The general procedure for the preparation involves oxidative ring opening of substituted furan compounds. Prevalent cis-butenedial was prepared from 2,5-dimethoxy-2,5-dihydroxyfuran according to the method of Hufford et al. (23). The product was characterized by NMR and IR. Twenty-four milligrams of 3H-furan-2-one (>90%, synthethised by A. Henderson) was introduced into the chamber to obtain its PTR-MS product ion pattern and calibration coefficient, because this is a likely product from butenedial photo-oxidation (24). VOL. 41, NO. 24, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8363
FIGURE 2. Carbonyl SPME profiles for glyoxal, methylglyoxal, benzaldehyde, butenedial, and 4-oxo-2-pentenal in the photo-oxidation of toluene in the presence of NOx (September 26, 2003). Toluene profile was provided by GC-PID. For SPME determinations, a 6890 gas chromatograph supplied by Hewlett-Packard Co.(Wilmington, DE) was used, equipped with a HP-5MS capillary column (cross-linked 5% PHME siloxane, 30 m × 0.25 µm × 0.25 µm), which was coupled to a flame ionization detector (FID) and an inlet liner with a narrow internal diameter, 0.75 mm i.d. Predrilled Thermogreen LB-2 septa for SPME were also purchased from Supelco (Lund, Sweden). The aldehydes were quantified through the formation of the corresponding oximes by reaction with O-(2,3,4,5,6)-pentafluorobenzylhydroxylamine hydrochloride (25, 26) (PFBHA; Fluka, Milano, Italy), which was loaded on the fiber. Sampling takes place by direct exposure of the SPME fiber for 5 min to the gaseous mixture in the chamber. The reaction takes place on the fiber surface. PDMS/DVB fibers (65 µm; Supelco) were used for sample acquisition throughout the experiment. Prior to their initial use, the fibers were conditioned at 250 °C for 30 min according to the manufacturer’s instructions. The following chromatographic conditions were applied in the analysis of SPME fibers: injector, 270 °C; detector, 300 °C; oven, initial temperature, 80 °C for 2 min, ramp 1, 20.0 °C/min to 280 °C, held for 3 min (total run time ) 15 min); column, constantpressure mode (25.3 psi). The following instrumentation was also used: JNO2 filter radiometers, an ozone monitor (Monitor Laboratories ML9810- SN2028, Englewood, CO), different NOx monitors (ML 9841A and ECO Physics (Duernten, Switzerland) CLD 770 AL with a photolytic converter PLC 760), an NOy monitor (API200 AU, Advanced Pollution Instrumentation, Inc., San Diego, CA), and a hygrometer (Walz TS-2, Heinz Walz GmbH, Eichenring, Germany). Air samples were collected automatically in a sampling loop and then injected into the chromatograph (GC Fisons 8160, San Jose, CA). The component peaks eluting from the column were identified and quantified using photoionization (PID) and flame ionization detection (FID) operated in isothermal conditions. The chromatograph, a Fisons GC8160, is equipped with a 30 m DB-624 (cyanopropylphenyl polysiloxane) fused silica capillary column (J&W Scientific, 0.32 mm i.d., 1.8 µm film). The rest of the chromatographic conditions were as follows: injector temperature, 80 °C; detector temperature, 250 °C; PHe, 25 kPa; PH2, 70 kPa; Pair, 100 kPa; P N2 (PID purge), 20 kPa; PID lamp, 10.6 eV; column 8364
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 24, 2007
flow, 1.6 mL/min (120 kPa); split, 1/75; splitless from 1 to 4 s (time of injection). The GC-ECD (PAN-GC Schmitt) was calibrated against FTIR for methylglyoxal. The chromatograph is equipped with a 7.5 m DB-5 (5% phenyl polysiloxane) fused silica capillary column (0.53 mm i.d., 5 µm film). This column is mounted in a thermostat-controlled oven, cooled by Peltier elements. Temperature fluctuations are
