Polycyclic aromatic hydrocarbon and nitroarene concentrations in ambient air during a wintertime high-NOx episode in the Los Angeles basin

Atmqkrk Enoiro~vnr F’rialed in OIut Britain.

Vol. 21. No. 6. pp. 1437-1444.

ooM4981/87 13.00+0.00 Q 1987 Perpmon Joumalr Ltd.

1987.

POLYCYCLIC AROMATIC HYDROCARBON AND NITROARENE CONCENTRATIONS IN AMBIENT AIR DURING A WINTERTIME HIGH-NO, EPISODE IN THE LOS ANGELES BASIN JANET AREY*,

BARBARA ZIELINSKA, ROGER ATKINSON

and

ARTHUR M. WINER

Statewide Air Pollution Research Center, University of California, Riverside, CA 92521. U.S.A. (First

received 29 August 1986 and infinolform 6 November 1986)

Abstract-The ambient concentrations of polycyclic aromatic hydrocarbons (PAH) (including biphenyl) and nitroarenes were measured during a wintertime, high-NO, episode at a location in Southern California. Daytime and night-time ambient air samples were collected using Hi-vol filters, polyurethane foam (PUF) plugs and Tcnax-GC solid adsorbent. 2-Nitrofluoranthenc was the most abundant particle-associated nitroarene, but higher concentrations of I- and 2-nitronaphthalcne, methylnitronaphthalcnes and 3nitrobiphcnyl were observed on the PUF plugs. Our data show that the ambient concentrations of the more volatile PAH and nitroarcnes can be far greater than those of the less volatile species,and suggestthat the most abundant nitroarenes in ambient air arise from atmospheric transformations of PAH emitted from combustion sources. Key word index: Nitroarenes. polycyclic aromatic hydrocarbons (PAH). nitronaphthalenes, nitrobiphcnyl, nitrofluoranthenes, nitropyrenes, ambient concentrations, atmospheric transformations, nitrogen oxides. INTRODUCTION Recently we have shown that 2-nitrofluoranthene and 2-nitropyrene are among the most abundant nitroarenes present in ambient particulate organic matter (POM) (Pitts et al., 1985a; Ramdahl et ol., 1986) and that it is likely that their presence in ambient POM is the result of gas-phase atmospheric reactions of fluoranthene and pyrene with N,OS and/or with the OH radical in the presence of NO, (Pitts et al., 1985a; Swcetman et al., 1986; Zielinska et al., 1986a, Arey et al., 1986). However, no data have been reported concerning the ambient atmospheric concentrations and formation routes of those nitroarenes which are more volatile than the nitrofluoranthenes and nitropyrenes. Depending on their molecular structure and the ambient temperature, the more volatile polycyclic aromatic hydrocarbons (PAH) (including biphenyl) and nitroarenes may occur totally or partially in the gas phase. It is recognized that during conventional high volume (Hi-vol) collection of ambient POM, PAH of the volatility of fluoranthene and pyrene will be largely ‘blown-or the filters at temperatures typical of those encountered in Southern California, even during winter months (Thrane and Mikalsen, 1981; Yamasaki et nl., 1982; Van Vaeck et al., 1984). Thus, in this work we have collected ambient air samples on Hi-vol filters, polyurethane foam (PUF) plugs and Tenax-CC solid adsorbent (Keller and Bidleman, 1984; Ligocki and Pankow, 1985) to measure the concentrations of several nitroarenes and their parent species in ambient

*Author IO whom correspondence should be addressed.

air during a wintertime, high-NO, episode at the Western end of the California South Coast Air Basin. Our current understanding of the formation routes leading to the observed nitroarenes is also discussed.

EXPERIMENTAL Sampling of ambient air was carried out from a roof (elevation - 9 m) at El Camino Community College in Torrance, CA on 24-25 February 1986 during a high-NO, smog episode. This sample site was located - 17 km NNE of Long Beach harbor, - 20 km S of central Los Angeles, and - 8 km E of the Pacific Ocean, with numerous mobile and stationary emission sources in the vicinity. The episode was characterized by low inversion heights with daytime clear-sky conditions. The two 12-h sampling periods were 18OO-0600 h 24-25 February and 0600-1800 h 25 February. The maximum pollutant levels observed were: NO, > 500 parts-perbillion (ppb) at 0800 h; NO = 400 ppb at 0800 h; and NOI = 250 ppb at 1000 h. as measured by a chemiluminescence analyzer; 0, = 90 ppb at 1600 h, measured by ultraviolet absorption. The maximum temperature during the daytime sampling period was 35°C at II00 h, with an approximately constant night-time temperature of 20°C. Concurrent measurementsof other gaseouspollutants, including HONO, HCHO, HNO, and NO3 radicals, were made by long pathlength u.v.-visible and Fourier transform infrared spectroscopy (H. W. Biermann, E. C. Tuazon and A. M. Wirier, unpublished data 1986), and these will be reported elsewhere.

Sump/e

collection.

exrracrion

and fractioncrrion

Tenox-CC cartridge. Gas-phase PAH were samoled usina 10 cm x 4 mm id. -Pyrex iubes packed with ’ 0.1 g of Tenax-GC and doped with 610 na of nanhthalene-d.. The Row rate was l.Ij/min-‘. yielding an’ -0.8m3 v&&k sampled for each 12-h sampling period. A second Tenax cartridge was placed in series downstream IO check for breakthrough from the first cartridge.

1437

1438

JANET AREY et al.

Thecartridges wereelutcd with 3 ml of ~thylethcr,and 60 and 67 ng of the internal standards p~~nth~n~~~ and anthraccne-drO, respectively, were added to the eluate which was then concentrated with a micro Snyder apparatus. The concentrated extracts were analyxed for PAH by gas chromatography/mass apcctrometry (GC/MS) with multiple ion detection (MID), as described &low. Hi-se/ Jitter f&owed by three PUF pkgs. A singk highvolume sampler system consisting of a Tern-operated glass fiber (TIGF) filter followed by three PUF plugs (each + 9 cm diameter x 5 cm thickness) was run at 30SCFM for 12-h intervals, yielding ambient samples collected from 610m’ of air. The TIGF filters and PUF plugs were precleaned by Soxhlct extraction with CH,Cl, and CH&OH. After sample collection, the filter and each PUF plug were separately Soxhkt extracted in CHsClr and analyzed. Deuterated internal standards were added to the filter, PUF No. I, PUF No. 2, and PUF No. 3, rcspactively, as follows (in w): naphthakne-d, (1.1, 5.3, 5.3, 5.3) phenanthrcne-d,, (2.0, 10.0, 10.0, 10.0); anthraccm-d,O (2.2, 11.2, 11.2, 11.2); fluoranthcne-d,, (1.2,11.2,1.2,0.0f;pyrcnc-& (1.2,11.0,1.2, O.Oh~~o~o~~ene~,~ (1.1,O-0,0.0,O.O); perykne-dt r (1.0, O.O,O.O, 0.0); and I-nitro~pht~i~c-d, (0.6,1.2,1.2,1.2). The amounts added to the filter and PUF plug extracts were similar to the nondeuteratcd PAH and nitroarcne concentrations observed. The extracts were concentrated by rotary evaporation and fractionated by high performance liquid chromatography (HPLC) using an Altex semi-preparative scale Ultrasphere Silica column~l cm x 25 cm). ihe_HPLC system consisted of a Socctra-Phvsics Model 8100 chromatoaraoh. Model 4100 computing integrator, Model 8400 u.v:/v~s detector and ISCO fraction collector. The mobik phase program employed was: n-hexane/CH,Cl,: 95/S for 10 min. then a linear gradient to 100%CH2C120ver 15 mitt, heldat 100%CH2C12 for 10 mitt, followed by a linear gradient to 100% CHSCN over 10 min. A PAH~n~~ning fraction was cofkcted from 4 min to 13 min and a nitroarenetontaining fraction from 13 min to 25 min. The fractions were concentrated by rotary evaporation, then taken just to dryness under a stream of nitronen. After dissolvinn in CH,C.%. the PAH and nitroareie fractions were analyzed by GC/MS-MID as described below. Hi-vois wizh TfGFfiirers. Inorder tocollect suflioient POM for analyses of nitrofluoranthencs and nitropyrcnes, up to four filters from simultaneous Hi-vol collections of POM were combined. The Hi-vols were run at either 30 or 40 SCFM and the total volumes of air sampled were: 18OO-O600h, 1960m3; 0600-1800 h, 319Om’. The filters ~mprising the ‘day’ and ‘night’ samples were Soxhlet cxtracted in pairs in CHrCl, and each doped proportionately so that the total amount of internal standard added for each

composite day and night sample was (in pg): 2nitrofluo~nthene~~ (0.6) l-nitropyrcne-d* (O.S),pyrene-d, 0 (6.1), pcrylens-d,, (4.3), 1,2dinitrofluoranthene-d, (0.6), dinitropyrenes-d, (0.9). The extracts were fractionated by HPLC as described above, except that the faction containing the nitroarenes was divided into two, the nitrofhroranthene and nitropyrene isomers being in the fraction eluting between 22 and 25 min. The nitrofluo~nthene and nitropyrene fraction was further cleaned prior to GC/MS-MID analysis by additional chromatography on an Altex semi-preparative Ultrasphere-ODS column (1 cm x 25 cm). The HPLC system used was a Beckman Model 334 equipped with a Beckman Model 164 u.v./vis detector. Isocratic elution by CHJOH/H,0:80/20 resulted in the nitrofluoranfhene and nitropyrene isomers (and the deuteratcd nitrofluoranthene and nitropyrene internal standards used) eluting between 22 and 32 min. Analysis Compound identifications and quantifications were made using a Finnigan 3200 quadrupole G&/MS operating in the ekctron impact mode (70 ev) and interfati to a Teknivent data system. The GC was equipped with a cool on-column injection system and a 30 m DB-5 fused silica capillary column (both from J % W Scientific, Inc.) eluting directly into the MS ion source. The quantitYcations of naphthalene, phenanthrene, anthraccne, flooranthen~ pyrene, ~nzo(a~yren~ and perylene were made by ratioing the area of the GC/MS-MID molecular ion peak to that of the corresponding deuterated internal standard. For the methylnaphthalene isomers and biphenyl an external calibration of the molecular ion peak to that of naphthalene-d, or phenanthrene-d,, was made, and for benzo(e)pyrene the deuteratcd benzo(a)pyrene internal standard ion was used. identification of the nitroarenes (apart from the methylnitronaphthalenes for which 14 isomers are possible) by GC/MS-MID was made on the basis of the presence, in correct abundance, of all or the major fragment ions (as given in Table I), as well as retention time matching. Authentic samples of all eight of the nitrofluoranthene and nitropyrene isomers and all thra of the nitrobipbenyl isomers were available for retention time and fragment ion abundance. comparisons. Quantifications for I-nitronaphthalene, 2-m trofluoranthcne and I-nitropyrene were made by comparison with the deuterated internal standards. 2-Nitronaphthalene, 3-nitrobiphenyl and 9-nitroanthracene were quantified by external calibration of the molecular ion peak to that of the lnitronaphthaIene-d, internal standard. 2-Nitropy~ne was quantified by external calibration of the molecular ion peak to that of the I-nitropyrene-d, internal standard.

Table 1. Molecular ions, characteristic fragment ions and relative abundances used for identification of nitroarenes by GC/M~MID m/z and relative abundance for authentic compound*

WI” 1-Nitronaphthalene 2-Nitro~pt~lene Methylnitro~pbt~len~~ 3-Nitrobiphenyl 9-Nitroauthracene 2.Nitrofluoranthene I-Nitropyrcne 2-Nitropyrenc

173 (0.4) 173 (0.5) 187 199 (0.7) 223 (0.7) 247 (0.7) 247 (0.5) 241 (0.51

[M-CO]+ 145 (0.1)

[M-NO]+ 143 (0.1) 143 (0.04) 157 169 (0.03) 193 (0.6) 217 (0.04) 217 (0.4) 217 (0.04)

[M-NOa]+

[M-HNOIJ+

127 (1.0) 127 (1.0) 141 153 (0.7) 177 (0.7) 201 (1.0) 201 (1.0) 201 (1.0)

126 (0.4) 126 (0.3) 140 152 (1.0) 176 (1.0) 200 (0.7) 200 (0.7) 200 (0.S)

Relative ebuudanccs arc given in puenthcscs. t 14 isomers are possibk, individual relative abundances not given.

l

[M-HINO1]+

151 (0.3)

[M-CN%l+ 115 (0.9) 115 (0.4) 129 141 (0.1) 165 (0.8) 189 (0.3) 189 (0.5) 189 (0.2)

1439

Polycychc aromatic hydrocarbon and nitroarene concentrations in ambient air Chemfcofs

The following chemicals were obtained from commercial sources: phenanthrenc-d,e, anthracene&. pyrene-dre, benxo@)pyrene-d,z and perylene-d,, (Cambridge Isotope Laboratories); fluomnthene-dab (MSD Isotopes 1nc.k napht~le~~*, i-nitronapht~l~e~,, biphenyl, I- and 2nitronaphthalene, 2-, 3- and 4-nitrobiphenyl and 9-nitroanthracenc (Aldrich Chemical Co.); I- and 2.methylnaphthalene (Chem Service); Standard Reference Material l&7. certified PAH (National Bureau of Standards). Co&ercially available i-nitropyrene (Pfaltzand Batter,Inc:) was purified according to the method described by PaputaPeck et at. (1983). Perdeuterated 2-nitrofiuoranthene and 1-nitropyrene were synthaixed as described by Zielinskaer al. (1986a)and Pitts et al. (1985b). 2-Nitropyrene was provided by Dr D. Schuetxle (Ford Motor Co.; Dearborn, Ml) and Cnitropyrene by Dr A. Berg(University of Aarhus, Denmark). The 1-,2-. 3-, 7-and Snitrofluomnthen~ were synthesixed as described previously (Ramdahl et of., 1985;Zielinska et crl.,1986a). RESULTS PAH

The measured con~ntratjons of PAH and their nitro-derivatives are given in Tables 2 and 3 for the night-time (1800-0600 h) and daytime (0600-1800 h) samples, respectively. The most abundant gas-phase PAH observed on the Tenax cartridges were naphthalene, the l- and 2-methylnapht~lenes and biphenyl, with lesser amounts of C~-naphthalenes, fluorene and phenanthrene also being observed. Breakthrough of naphthalene-da on the Tenax collec-

tion system was c 3%. showing that these PAH were quantitatively collected under the sampling conditions employed. On the first of the PUF plugs (sampling gas-phase PAH and those ‘blown-off’ the particles collected by the filter) phenanthrene was the most abundant PAH. Clearly, as seen from the data given in Tables 2 and 3 for PUF plugs No. 2 and No. 3 and the Tenax cartridges, the PAH more volatile than phenanthrene were not quantitatively collected on the PUF plugs. Thus, for example, in the night sample < 3% of the naphthalene, and in the day sample < 0.5% of this PAH, was retained on the three PUF plugs. For phenanthrene, there was reasonable agreement between the concentrations quantified on the Tenax cartridges and the sum of the phenanthrene concentrations observed on the filter plus the three PUF plugs, Clearly, at higher temperatures, Tenax sampling is necessary for quantitative collection of phenanthrene and other PAH and nitroarenes of similar or greater volatility. As expected, quanti~~tions of ambient concentrations of fluoranthene and pyrene based solely on extracts from the filter-collected POM were significantly low, but a single PUF plug was adequate to retain these PAH at the sampling temperatures encountered during this study. Benz(o)anthracene and chrysene were not quantized in this study due to a lack of internal standards, but only traces of these PAH (of M.W. 228) were observed on the first PUF plug.

Table 2. Ambient night-time co~ntrations

(ng m-‘) of PAH and nitroaren~ (ng m-‘)

Compound

M.W. PUF No. 1 PUF No. 2

PUF No. 3

Filter

Z(Filter+ PUFs)

Tenax

m Napht~iene 2Methylnaphthalene I-Methvhraohthalene Biphenyl ’ Phenanthrene Anthracene Fluoranthene Pyrcne Benxo(e)pyrene Benxo(a)pyrene Nitroarenes I-Nitronaphthalene 2-Nitro~pht~~ne 3-NitrobiphenyI 9-Nitroanthracene 2-Nitrofluoranthene I-Nitropyrene 2-Nitropyrene

128 142 142 154 178 178 202 202 252 252 252 173 173 :z 247 247 247

29

0.27

75

0.28 0.03 0.53 0.67 2.1 1.6 0.47

81 11 9.7 12 2.1

0.01 0.002 0.01 0.11 0.32

2.3 1.1

n.d.

0.03

K 014

n.d.

0.035

0.03

: t 70 K 8:s n.d. n.d. n.d. &I 0.38 n.d. 0.04

24 : t 9.4 1.5 0.06 2.7$

0.53 0.19 0.15

22 I 1.0 0.28 n.d.* n.d.

0.12 0.03 0.04

2800 1100 1200 62 90 7.2

K

0.03

Night-time sample collected at El Amino Community College, Torrance, CA, 24-25 February 1986.18~ h. l n.d. = None detected. t As was the case with naphthalene. each PUF retained a similar small amount of this compound and was, therefore, not useful for quantification purposes. $ Vlllue suspect; see, for example, Ruoranthene, this table, and fluoranthene and pyrene, Table 3. 8 Quantification by reverse phase HPLC.

1440

JANETAREYet al.

Table 3. Ambient daytime concentrations (ng m-‘) of PAH and nitroarenes (ng m-‘) Compound

M.W.

PUF No. 1

PLJF No. 2

PUF No. 3

Filter

X(Filter + PUFs)

Tenax

Naahthalene t-Methylnaphthalene I-Methylnaphthalene Biphenyl Phenanthrene Anthracene Fluoranthene Pyrene Benzo(e)pyrene Benzo(a)pyrene Perytene

128 142 142 154 178 178 202 202 252 252 252

4.1

4.5

4.9

0.45

15

: 78 6.1 8.0 8.0 2.1 0.6 0.2

3300 900 1100 60 78

Nitroarenes I-Nitronaohthalene 2-Nitronaphthalene 3-Nitrobiphenyl 9-Nitroanthracene 2-Nitrofluoranthene I-Nitropyrene 2-Nitropyrene

173 173 199 223 247 247 247

w :

:

t

t

t

33 2.3 7.1 7.0 n.d. n.d. n.d.

32 2.7 0.44 0.36

13 1.1 n.d.* n.d.

0.33 0.03 0.47 0.60 2.1 0.59 0.18

1.3

1.0 1.1 0.29

0.67 0.38 0.01

0.05 0.006 0.03 0.05 0.28 0.04 0.04

::: nd. 0.009 n.d. n.d.

3.01: 2.9’ 6.0 0.05 0.3 0.04 0.04

Daytime sample collected at El Camino Community College, Torrance, CA, 25 February 1986,060-1800 h. l n.d. = None detected. t As was the case with naphthalene, each PUF retained a similar small amount of this compound and was, therefore, not useful for quantification purposes. $ Lower limit, due to the observed breakthrough through the three PUFs used.

Consistent with this observation, none of the isomeric M.W. 252 compounds were observed on the PUF plugs. It is interesting to note that the ratios of the reactive PAH benzo(a)pyrene and perylene to the nonreactive benzo(e)pyrene (Nielsen, 1984) were lower in the POM collected during the day than during the night, suggesting the occurrence of daytime chemical loss processes for benzo(a)pyrene and perylene and/or of differing PAH emission sources during these sampling intervals. N itroarenes In agreement with our previous data (Pitts et al., 1985a; Ramdahl et al., 1986). f-nitrofluoranthene was the most abundant nitroarene observed in the ambient POM extracts. It is clear, however, from Tables 2 and 3 that in this ambient air sample, and perhaps in general, the most abundant nitroarenes are the more volatile species. Figure 1 shows the GC/MS-MID traces for extracts of the first PUF plug from the night-time (Fig. 1A) and daytime (Fig. 1B) samples for the molecular ions of the 1-nitronaphthalene-d, internal the standard, the l- and 2-nitronaphthalenes, methylnitronaphthalenes, and the nitrobiphenyls. The 1. and 2-nitronaphthalenesand 3-nitrobiphenyl were identified on the basis oftheir retention times and fragment ion abundances matching those of authentic standards. The presence of 5-nitroacenaphthene, an isomer of the nitrobiphenyls, in ambient POM has previously been reported by Tokiwa et 01. (1981). However, the recently reported GC retention index of 5-nitroacenaphthene (Robbat et al., 19864 clearly

shows that this nitroarene would be well resolved from any of the nitrobiphenyl isomers. Since it was determined using authentic standards that 2- and 4nitrobiphenyl, as well as 3-nitrobiphenyl, eluted in the HPLC nitroarene-fraction, it isclear from Fig. 1 that 2and Cnitrobiphenyl were not detected in these ambient air samples. The presence of several methylnitronaphthalene isomers is evident from the m/z = 187 ion traces shown in Fig. 1 (similar peaks were observed for all the methylnitronaphthalene fragment ions listed in Table 1). However, the lack of authentic standards for the 14 isomeric methylnitronaphthalenes made specific isomer identification impossible. It can be seen from Fig 1 that 3-nitrobiphenyl and certain of the methylnitronaphthalenes were more abundant during the daytime (Fig. IB) than during the night-time (Fig. 1A) sampling period. Although not obvious from Fig. 1, which only shows the extracts from the first PUF plugs, the I- and 2.nitronaphthalenes were also more abundant in the daytime sample [see the Z(Filter +PUFs) column, Tables 2 and 33. This diurnal variation was not observed for the less volatile nitroarenes such as 9-nitroanthracene and 2-nitrofluoranthene, which were present mainly in the POM.

DISCUSSION

Direct emission from combustion sources is still viewed as the primary source of nitroarenes in ambient air, as evidenced by the current review of Tokiwa and

Polycyclic aromatic hydrocarbon and nitroarenc concentrations in ambient air

I

10

11

12

RETENTION

I

14

13

TIME

m/t

187

15

(min)

10

11

12

RETENTION

13

TIME

14

1s

(min)

Fig. 1. GC/MS-MID traces showing nitroarcnc molecular ion peaks for: A: nitroarenc-containing HPLC fraction of extract from first PUF plug, night-time sample (18OO-0600h). B: Nitroarcne-containing HPLC fraction of extract from first PUF plug, daytime sample (0600-1800 h). m/z = 180. molecular ion of deutcrated I-nitronaphthalene (I-NN-d,) added to both extracts in equal amounts as an internal standard; m/z = 173,molecular ion of nitronaphthalenes (NN); m/z = 187, molecular ion of mcthylnitronaphthalenes; m/z = 199, molecular ion of nitrobiphcnyls (NBPh). GC separation on a 30 m DB-5 fused silica capillary column. Injection at WC, then programmed at 8°C min- ’ to 250°C. (Additional retention times: 2nitrobiphenyl, 12.2-12.3 min; Cnitrobiphenyl 14.4-14.5 min.) Ohnishi (1986). However, the present work and recent ambient air measurements and laboratory studies (Nielsen et ol., 1984; Pitts er al., 1985a; Arey et nl., 1986; Atkinson et al., 1987; Nielsen and Ramdahl, 1986; Ramdahl et al., 1986, Sweetman et al., 1986; Zielinska et al., 1986a) show that this conclusion must be significantly modified. The nitroarenes observed in what is certainly the best characterized combustion emission, namely diesel POM, have been those expected from electrophilic nitration of PAH (Schuetzle, 1983), presumably being formed after the combustion process but before extensive dilution after emission occurs (Kittelson et al, 1985). The specific isomers and the relative abundance of the nitroarenes present in diesel POM can be rationalized on the basis of the abundance of the parerit PAH and their reactivity toward electrophilic nitration. For example, 1-nitropyrene is the primary nitroarene observed in diesel POM, together with much lower amounts of 3- and 8-nitrofluoranthcnes (Paputa-Peck et ol., 1983; Liberti and Ciccioli, 1986; Robbat et al., 1986b). Fluoranthene and pyrene are among the most abundant PAH in diesel POM and are emitted in similar quantities (Schuetzle et al., 1981); however, pyrene is more reactive toward electrophilic nitration

than is fluoranthene (Nielsen, 1984). As shown in Tables 2 and 3, we observed that fluoranthene and pyrene were present in ambient air also at very similar concentrations. In addition to being the primary nitroarene in diesel POM, 1-nitropyrene has been reported to be present in particulate emissions from gasoline-fueled vehicles (Gibson, 1982; Nishioka et al., 1982) and coal-fired power plants (Harris et al., 1984). However, neither 2nitrofluoranthene nor 2-nitropyrene has been reported to be present in any of these emission sources,although both are present in ambient POM (Nielsen et al.. 1984; Pitts et al., 1985a; Nielsen and Ramdahl, 1986; Ramdahl et a!., 1986; Sweetman et al., 1986; Liberti and Ciccioli, 1986; Tables 2 and 3). Indeed, we have shown that 2-nitrofluoranthene is often the most abundant nitroarene in ambient POM. In addition to the present work, we have shown this to be true for POM collected at several locations in the U.S. and Europe, and by various collection methods, including Hi-vol filtration, electrostatic precipitation and bag-house c&action (Pitts er al.. 1985a; Arey ef al., 1986; Ramdahl ef al., 1986; Swectman et al., 1986). It should be noted, however, that a single report of the presence of 2nitrofluoranthene and 2-nitropyrene in an industrial

1442

JANET AREY cr al.

emission (from a plant in Italy manufacturing carbon electrodes) has very recently appeared (Liberti and Ciccioli, 1986). This particular emission source seems unlikely to be a major contributor to ambient 2nitrofluoranthene and 2nitropyrene concentrations, except perhaps on a very local scale, and no such plants exist in the California South Coast Air Basin (South Coast Air Quality Management District, 1986, Personal communication) where the present ambient measurements were conducted. The Hi-vol filters plus PUF plug sampling system employed in this study may also provide information concerning the origin(s) of the nitroarenes. Thus, the nitroarenes observed on the fiiter consist of those formed in the gas-phase which condense onto POM, those present in primary POM emissions, plus those formed in the adsorbed phase during transport and collection minus those blown-off during collection. In contrast, the nitroarenes observed on the PUF plug should consist of the gas-phase nitroarenes plus those blown-off the particles during sampling minus the gasphase nitroarenes which condensed on the particles during atmospheric transport and sampling. 9-Nitroanthracene has been reported to be present in diesel emissions (Pitts er al., 1982; Paputa-Peck er al., 1983; Draper, 1986; Robbat et al., 1986b) as well as in ambient POM (Ramdahl er al., 1982; Wise ef al., 1985). We observed 9nitroanthracene in the TICF filter extracts but none was observed on the PUF plugs, whereas 2-nitrofluoranthene (which is expected to be of a similar or lower volatility than 9-nitroanthracene) was observed both in the filter extracts and, in relatively trace amounts, on the first PUF plug. Thus, it can be speculated that the observed 9nitroanthracene is a direct emission, in contrast to 2nitrofluoranthene which is a gas-phase atmospheric transformation product of fluoranthene. For naphthalene and biphenyl, electrophilic nitration is expected to produce I- 9 2nitronaphthalene (Ruehle et al., i985) and 4- > 2- > 3-nitrobiphenyl (Merck Index, 1983), respectively. However as we report here, the observed species in ambient air are lnitronaphthalene h 2-nitronaphthalene w 3-nitrobiphenylg 2- and 4-nitrobiphenyl (these latter being below our detection limits). In the POM extracts, more I-nitronaphthalene than 2-nitronaphthalene and only traces of 3-nitrobiphenyl were observed (see Tables 2 and 3). Since 2nitronaphthalene and 3nitrobiphenyl exhibited less breakthrough to the second and third PUF plugs during sampling than did l-nitronaphthalene, they are presumably less volatile than lnitronaphthalene and their lower abundance on the filters cannot be explained by assuming differential ‘blow-off’ from the filters. However, if the observed Initronaphthalene was present in emitted POM, some of it would be incorporated inside the particles and hence be unavailable for ‘blow-off’ (though solvent extractable). This suggests that the 2-nitronaphthalene and $nitr&iphenyl are gas-phase atmospheric transformation products, while a certain fraction of the l-

nitronaphthalene is a direct emission (with the remainder being formed in the gas-phase during atmospheric transport; see below). Possible atmospheric formation

routes of nitroarenes

We have recently shown that there are at least two possible atmospheric transformation pathways which may lead to nitroarene species distinct from those formed from electrophilic nitration. Thus, 2-nitrofluoranthene is the sole mononitro-product of the gasphase reaction of fluoranthene with N205 (Sweetman et al., 1986; Zielinska et al., 1986a) and the major nitroproduct, together with minor amounts of 7- and 8nitrofluoranthene, from the reaction of OH radicals in the presence of NO, (Arey et al., 1986). 2-Nitropyrene is the sole mononitropyrene formed from the reaction of pyrene with OH radicals in the presence of NO, (Arey et al., 1986). but is not formed from the gasphase reaction of pyrene with N205 (Zielinska et crl.. 1986b). Additionally, we have shown that the gasphase reaction of naphthalene with N20, produces land 2-nitronaphthalene in a + 3 : 1 abundance (Pitts et al., 198%). Due to the presence of NO, which reacts rapidly with NO3 radicals, negligible concentrations of the NO, radicals, and thus also of NzOS, were present at this site during the night-time cohection period (H. W. Biermann, E. C. Tuazon, and A. M. Winer, unpublished data, 1986). Since the NO3 radical photolyzes rapidly during daylight hours, N,O, would not be present during the daytime collection period (for a discussion of ambient N205 concentrations, see Atkinson et al., 1986). Thus the gas-phase reactions of N,05 with fluoranthene to form 2-nitrofluoranthene and with naphthalene to form l- and 2-nitronaphthalene (and with other PAH to form their nitroderivatives) were of minor or negligible importance during the present study. However, the formation of both 2-nitrofluorantheneand 2-nitropyrene from reaction with OH radicals in the presence of NO, can explain the presence of these nitroarenes in the POM collected during this high-NO, episode, It is interesting to note that in this winter episode and location in the western end of the California South Coast Air Basin the ratio of 2-nitropyrene to 2-nitrofluoranthene was higher than previously observed at Riverside, CA (Pitts er al., 1985a) and Claremont, CA (Ramdahl et al., 1986) during summer photochemical pollution cpisodcs, in which both formation pathways for 2nitrofluoranthene may have been operative. Further evidence for the fact that the nitroarenes observed during this winter, high-NO, episode were largely the result of atmospheric reactions of the parent PAH with the OH radical arises from our recent environmental chamber exposures of naphthalene and biphenyl to OH radicals in the presence of NO,, which show that the nitro-products are I- and 2nitronaphthalene in nearly equal abundance and 3nitrobiphenyl, respectively (Atkinson ~1 al., 1987).

1443

Polycyclic aromatic hydrocarbon and nitroarcne concentrations in ambient air CONCLUSIONS

Previous analyses of PAH and, in particular, nitroarenes have primarily been made on extracts of ambient POM, and this has ied to an emphasis on the less volatile PAH and their nitro-derivatives. It is evident from the present work that the ambient concentrations of the more volatile PAH and nitroarenes can be far greater than those of the less volatile species present in ambient POM. Although the PAH which are carcinogenic (Jacob et al., 1986) are the higher molecular weight, mainly particle-associated, species (M.W. > ZOO),this may not be the case for the nitroarenes. For example, 2nitronaphthalene and 4nitrobiphenyl have both been shown to be carcinogenic in animal feeding studies (Tokiwa and Ohnishi, 1986). The abundance of the lower molecular weight PAH and the high abundance of the nitro-derivatives of these species (in particular of the nitroarenes such as 3nitrobiphenyl and 2-nitronaphthalene which are not the major expected electrophilic nitration products) suggests that the most abundant of the nitro-species in ambient air arise from the atmosphe~c transformations of combustion-emitted PAH. Clearly, further understanding of the atmospheric chemistry of PAW, and of the nitroarene isomers emitted from industrial sources, is necessary before the nitroarenes to which populations are exposed, and for which health effect data are needed, can be more fully assessed or predicted. Acknowledgements-The authors gratefully acknowledge the financial support of the California Air Resources Board Contract No. A4-08 l-32 (Dr Jack K. Sudcr, Project Monitor). The able technical assistance of MS Tricia McElroy, Mr Travis M. Dinoff, Mr Phillip C. Pelzel and Mr William D. Long is acknowledged. Drs Heinz W. Biermann and Ernest0 C. Tuaton are thanked for helpful discussions and Dr James N. Pitts, Jr for encouraging this work. We also thank the Administrators of El Camino Community College, and especially Mr J. Stimac and Mr H. Ledesma of the College Physical Plant, for their cooperation and assistance in siting this study, and Mr J. Arnold of the South Coast Air Quality Management District for providing air quality forecasts.

Draper W. M. (1986) Quantitation of nitro- and dinitropolycyclic aromatic hydrocarbons in diesel exhaust particulate matter, Chemosphere15.437-447. Gibson T. L. (1982) Nitroderivatives of polynuclear aromatic

hydrocarbons in airborne and source particulate matter. Arnwsp~eric Environment S6, 2037-2040. Harris W. R., Chess E. K., Okamoto D., Remsen J. F. and Later D. W. (1984) Contribution of nitropyrene to the mutagenic activity of coal fly ash. En&. Mutagenesis 6, 131-144. Jacob J., Karcher W. and Wagstaffe P. J. (1986) Polycyclic aromatic compounds of environmental and occupational importance--their occurrence, toxicity and the development of high purity certified reference materials-l. Fresenius Z. Anufyt. Chem. 317, IOI-I 14. Keller C. D. and Bidleman T. F. (1984) Collection of airborne polycyclic aromatic hydrocarbons and other organics with a glass fiber filter-polyurethane foam system. Atmospheric Environmenr 18, 837-845. Kittelson D. B., Du C.-J. and Zweidinger R. B. (1985) Measurements of polycyclic aromatic compounds in the cylinder of an operating diesel engine. EPA-600/D-85/012, January. Liberti A. and Ciccioli P. (1986) High resolution chromatographic techniques for the evaluation of atmospheric pollutants. J. High Resolution Chromatogr. Chromotogr. Commun. 9,492-501. Ligocki M. P. and Pankow J. F. (1985) Assessment of adsorption/solvent extraction with ~lyurethane foam and adsorption/thermal desorption with Tenax-GC for the collection and analysis of ambient organic vapors. Analyr. Chem. 57, 1138-I 144. Merck Index (1983) (edited by Windholz M.), p. 944. Merck and Co., Inc., Rahway, NJ. Nielsen T. (1984) Reactivity of polycyclic aromatic hydrocarbons toward nitrating species. Enuir. Sri. Technol.18, 157-163. Nielsen T. and Ramdahl T. (1986) Discussion on “Detcrmination of 2-nitro~uoranthene and 2-nitropyrene in ambient particulate matter: evidence for atmospheric reactions.” Atmospheric Environment 20, 1507. Nielsen T., Seitx B. and Ramdahl T. (1984) Occurrence of nitro-PAH in the atmosphere in a ruraf area. Atmospheric Environment

I 8,2 159-2

t 65.

Nishioka M. G., Petersen B. A. and Lewtas J. (1982) Comparison of nitro-aromatic content and direct-acting mutagenicity of diesel emissions. in Po/ynuc~eur Aromatic ~y~r~ar~ns:

Physical

and Biological

Chemistry

(edited

by Cooke M., Dennis A. J. and Fisher G. L.), pp. 663-613. Battelle Press, Columbus, Ohio. Paputa-Peck M. C., Marano R. S., Schuetzle D., Riley T. L., Hampton C. V., Prater T. J., Skewes L. M., Jensen T. B., Ruehle P. H., Bosch L. C. and Duncan W. P. (1983) Dctcrmination of nitrated polynuclear aromatic hydrocarbons in particulate extracts by capillary column gas

chromatography with nitrogen selective detection. An&t. Chem. 55,

REFERENCES

Arty J., Zielinska B., Atkinson R., Wincr A. M., Ramdahl T. and Pitts J. N., Jr (1986) The formation of nitro-PAH from the gas-phase reactions of fluomnth~ea~ pyrenc with the OH radical in the presence of NO,. Atmospheric Environment 20, 2339-2345. Atkinson R., Amy J., Zielinska B. and Aschmann S. M. (1987) Kinetics and products of the gas-phase reactions of OH radicals and N,Os with ~pht~lene and biphenyl E&r. Sci. Teehnol. (submitted for publication). Atkinson R., Winer A. M. and Pitts J. N., Jr (1986) Estimation of the night-time N,Os concentrations from ambient NO, and NO, radical concentrations and the role of NIOs in night-time chemistry. ~r~spberjc Enviro~~t 20,331-339.

1946-1954.

Pitts J. N., Jr, Lokensgard D. M., Harger W., Fisher T. S., Mejia V., Schuler J. J., Scorxiell G. M. and Katzenstein Y. A. (1982) Mutagens in diesel exhaust particulate-identification and direct activities of 6-nitrobenzo(a)pyrene, 9.

.nitroanth~ne, I-nitropyrene and 5H-ph~~~~thro(4,5 bed)pyran&one. Murat. Res. 103.241-249. Pitts J. N., Jr, Sweetman J. A., Zielinska B., Winer A. M. and Atkinson R. (1985a) Determination of 2-nitrofluoranthenc and 2.nitropyrene in ambient particulate organic matter: evidence for atmospheric reactions. Atmosoheric Environment

19,

I 60I- 1408.

Pitts J. N., Jr. Sweetman J. A., Zielinska B., Atkinson R., Winer A. M. and Harger W. P. (1985b) Formation of nitroarenes from the reaction of polycyclicaromatic hydrocarbons with dinitrogen pentaoxide. Eauir. Sci. ~ec~no~.19, 1115-1121.

JANET AREY et al.

1444

Pins 1. N., Jr, Atkinson R., Sweetman J. A. and Zielinska B. (19%) The gas-phasereaction of naphthalenc with N,O, to form nitronaphthalenes. Atmospheric Enuironmenr 19, 701-705. Ramdahl T., Bechcr G. and Bj#rseth A. (1982) Nitrated polycyclic aromatic hydrocarbons in urban air particles. Envir. Sci. Techtwl. 16, 861-865. Ramdahl T, Sweetman J. A., Ziclinska B., Atkinson R., Winer A. M. and Pitts J. N., Jr (1985) Analysis of mononitroisomers of fluoranthenc and pyrene by high resolution capillary gas chromatography/mass spectrometry. J. High Resolut.

Chromat.

Chromor.

Commun.

8, 849-852.

Ramdahl T., Zielinska B., Arey J., Atkinson R., Winer A. M. and Pitts J. N.. Jr (19861 Ubiauitous occurrence of 2nitrofluoranthene and 2-nitropyiene in air. Nature 321, 425-421. Robbat A., Jr, Corso N. P., Doherty P. J. and Marshall D. (1986a) Multivariate relationrhips between gas chromatographic retention index and molecular connectivity of mononitrated polycyclic aromatic hydrocarbons. Anolyt. Chem.

58,2072-2077.

Robbat A., Jr, Corso N. P., Doherty P. J. and Wolf M. H. (1986b) Gas chromatographic chcmiluminescent detection and evaluation of predictive models for identifying nitrated polycyclic aromatic hydrocarbons in a diesel fuel particulate extract. Anolyr. Chem. 58, 2078-2084. Ruehle P. H.. Bosch L. C. and Duncan W. P. (1985) Synthesis of nitrated polycyclic aromatic hydrocarbons. In Nitrated Pofycyclic &o&&c Hydrocarbo& (edited by White C. M.), DD. 169-235. Dr Alfred Huethin. Heidelberg. Schuetzle D. (1983) Sampling of v&icle emis.&ns for chemical analysis and biological testing. Enoir. HIrh Perspec. 47, 65-80.

Schuetzle D., Lee F. S.-C., Prater T. J. and Tejada S. B. (1981) The identification of polynuclear aromatic hydrocarbon (PAH) derivatives in mutagenic fractions of diesel particulate extracts. Jnr. J. Enuir. Analyr. Chem. 9, 93-144. Sweetman J. A., Zielinska B.. Atkinson R., Ramdahl T., Winer

A. M. and Pitts J. N., Jr (1986) A possible formation pathway for the 2nitrofluoranthene observed in ambient particulate organic matter. Atmospheric Enoironmenr 20. 235-238.

Thrane K. E. and Mikalsen A. (1981) High-volume sampling of airborne polycyclic aromatic hydrocarbons using glass fibre filters and polyurethane foam. Atmospheric Enuironment 15,909-918. Tokiwa H., Nakagawa R., Morita K. and Ohnishi Y. (1981) Mutagenicity of nitro derivatives induced by exposure of aromatic compounds to nitrogen dioxide. Muror. Rex 85, 195-205. Tokiwa H. and Ohnishi Y. (1986) Mutagenicity and carcinogenicity of nitroarenes and their sources in the environment, CRC Critical Reviews in Toxicology 17, 23-60. Van Vaeck L., Van Cauwenberghe K. and JanssensJ. (1984) The gas-particle distribution of organic aerosol constituents: measurement of the volatilisation artefact in Hivol cascade impactor sampling. Atmospheric Enoironmenr l8,417-430.

Wise S. A., Chesler S. N., Hilpert L. R., May W. E., Rebbert R. E., Vogt C. R., Nishioka M. G., Austin A. and Lewtas J. (1985) Quantification ofpolycyclicaromatic hydrocarbons and nitro-substituted polycyclic aromatic hydrocarbons and mutagenicity testing for the characterization of ambient air particulate matter. Enuir. Int. 11, 147-160. Yamasaki H.. Kuwata K. and Miyamoto H. (1982) Effects of ambient temperature on aspects of airborne polycyclic aromatic hvdrocarbons. Enuir. Sci. Technol. 16, 189-194. Zielinska B., krey J., Atkinson R., Ramdahl T., Winer A. M. and Pitts J. N., Jr (1986a) Reaction of dinitrogen pentoxide with fluoranthene. J. Am. C/tern. Sot. 108, 4126-4132. Zielinska B., Arey J., Atkinson R., Winer A. M. and Pitts J. N., Jr (1986b) The gas-phase reaction of polycyclic aromatic hydrocarbons with dinitrogen pentoxide. Presented at the 192nd Natl. Am. Chem. Sot. Meeting, Anaheim, CA, 7-12 Sept.