The Science o f the Total Em,ironment, ! 03 ( i 99 i ) !- 17 Elsevier Science Publishers B.V., Amsterdam

Polycyclic aromatic hydrocarbon (PAH) concentrations in ambient airborne particles from local traffic and distant sources; variation of the PAH profile W.K. de Raat and F.A. de Meijere institute o1" Environmental Sciences. TNO P.O. Box 6011. 2600 JA Delft. Nettu, tlamls

(Received May 10th, 1990; accepted July 18th, 1991))

ABSTRACT The temporal and spatial dependence of the PAH profile, i.e. the relative concentrations of polycyclic aromatic hydrocarbons, was investigated !br ambient airborne particles during a period with moderate photochemical air pollution. The concentrations of 14 PAH were measured; they differed in volatility, sensitivity to atmospheric chemical conversion and contributing sources. Multivariate analysis (principal-component analysis and factor analysis) revealed that temporal dependence was predominantly determined by live factors clearly linked with volatility, reactivity and sources of the PAH, the first being by far the most important. The results, therefore, indicate that volatilization, conversion and a varying contribution of local sources were the major causes of the variation of the profile with time. The contribution of local sources was investigated by comparison of samples that were taken simultaneously at three different sites, one a background site lind two sites downwind of traffic. A Wl~:|w.kt,~|site dependence was found. The comparison suggested that the differences were not only determined by sources, but also by volatilization and/or conversion during residence of the particles in the air.

INTRODUCTION

This paper is concerned with an analysis of the temporal and spatial dependence of the relative concentrations of polycyclic aromalic hydrocarbons in ambient airborne particles, i.e. th : PAH profile of the particles. Such an analysis can yield information on the sources contributing to the PAH in the particles and the chemical and physical processes affecting the concentrations of the PAH after emission. Sources are char~,cte"ized by a particular profile and variations in the contribution of individual :~ources may be detected in the variation of the profile in the ambient air (Cre~ney e~ a!., !985: Daisey et al., 1986; de Raat et al., 1987a). Chemical reaction of PAH after

J/

emission will affect the concentrations of reactive PAH more than those of inert PAH and will therefore result in a decrease of the relative concentrations of the reactive PAH (de Raat et al., 1987a; Nielsen, 1988). Likewise, evaporation of PAH from, and condensation on, particles will be reflected by a variation of the proportion of the more volatile PAH in the profile (de Raat et al., 1987a). In addition, sampling of the particles may induce an artificial change of the profile. The almost inevitable pressure drop across the filter used in high-volume samplers will accelerate evaporation of the more volatile PAH (Puppet al., 1974; van Vaeck and van Cauwenberghe, 1978; van Cauwenberghe et al., 1980; K6nig et al., 1980). The exposure of a particle on a filter surtace to other particles and an air stream containing reactive gaseous components may result in chemical conversion (Lee et al., 1980; Peters and Seil'ert, 1980; Brorstr6m et al., 1983; Grosjean, 1983; Alfheim and Lindskog, i984; Miguel and Andrade, 1986; Lindsko~ el al. 1987). In the present study, we mvestigated the variation of the PAH profiles of a set of samples Collected during a period with moderate photochemical air pollution in the western part of the Netherlands. The aim was to determine whether or not sources, chea-iical conversion and volatilization could be detected fiom the variation of the profile. We collected at least four 3-h samples during the day to evaluate the possible influence of the diurnal trend of the reactive gaseous pollutants typical of photochemical episodes (Finlayson-Pitts and Pitts, 1986). During part of the study, samples were simultaneously collected at three locations, thereby allowing the influence of local trattic o~er distant sources to be singled out. METHODS AN,.~ MATERIALS

Sanlpling Airborne particles were collected with the aid of Sartorius HVI00 highvolume samplers on Sartorius SMI3400 glass-fibre filters. Prior to sampling the filters were cleaned by Soxhlet extraction (24 h, methanol). The particlebearing filters were wrapped in aluminium foil and stored at - 80° C. The flow rate was 100 m 3 h-~, equivalent to a linear air speed of 0.54 m s-~. During the first 8 days, samples were taken on the premises of TNO (TNO-ZP) in the southeastern corner of Delft, about 100 m west of the busy motorway between the Hague and Rotterdam, at ground level. During the last 4 days, additional samples were taken on the roof of the Fire Service building near Delft city centre, about 20m above ground level; and at Delfgauw, 1.5km east of TNO-ZP in the grounds of a market gardener, at ground level. For the locations, see Fig. I. Sampling began at 08.00 or 05.00 h and continued for 24 h; filters were

~,~

.,,,,

i$

ua

,

8 4 ~

Fig. I. Map of Delft: locations TNO-ZP (T), Delfgauw (DG) and Delft (DE) are indicated.

changed at 08.00 (if sampling was started at 05.00 h). i 1.00, 14.00, 17.00, 20.00 and on some sampling days at 23.00h. Sampling dates and schedules are shown in Fig. 2. The decision of whether or not to take samples was based on the weather forecast issued by the Royal Dutch Meteorological Institute: when weather favouring photochemical air pollution was predicted (high temperatures, low wind speeds from easterly directions and a clear sky) sampling was commenced. The relevant meteorological conditions are listed in Table 1, which shows that conditions favouring photochemical air pollution occurred on most sampling days; no rain occurred on samplir, g days. Extraction

After sampling, the filters were extracted with methanol (Rathburn, HPLC grade) in a Soxhlet apparatus for 8 h (20 cycles) in an atmosphere of nitrogen in the dark. The volume of solvent was reduced to 5 mi with the aid of a rotatory evaporator at 30 ° C in the dark. Dimethyl sulphoxide (Merck, A.R.) was added to the concentrated extract and tbe remaining methanol was evaporated. The extracts were tested for mutagenicity within a few days (see De Raat and De Meijere, 1988, for details and results). The remainder of the

Date of first day

> time. h

u- s -'82

f

f

t

t

i

t

~4-S-'82

~

t

t

. t

f

t

Ts- s -'e2

,

~

t

t

~

t

~6- s-'st

T

f

t

t

W

f

2- 6-'at

f

V

t

f

~1 .

t

4- 6-'82

T

f

T

f

~

f

8- 7 -'82

f

t

V

t

9.7.'8t

*

*

t

i

~

f

,. 7 -'~

f

f

t

t

..f

f

2-8-'82

V

f

t

,t

f

,t

~

~ ,t

t

........... ,t

Fig. 2. Date and sampling schedule. (J,) Replacement of filter.

extracts was stored for ~ 3 years at - 8 0 ° C in glass scintillation vials wrapped in aluminium foil. Although the storage time is rather long, significant conversion of PAH under these storage conditions is highly improbable. It was expected that the very low temperature would have slowed down chemical reactions and microbiological degradation of the PAH to a negligible level. In addition, the absence of light would have prevented any photochemical conversion of the PAH.

Determination of PAH The PAH listed in Table 2 were determined with the aid of reversed-phase HPLC. The extracts were injected directly onto the column without prior purification. The dimethyl sulphoxide had no effect on the outcome of the analysis. Stationary phase: Supelco-PAH (250 × 4.6ram). Mobile phase: 75w/w % methanol in water to 100% methanol (Rathburn, HPLC grade). Detection: based on fluorescence, excitation at 250nm and emission at > 390 nm. The PAH were identified and their concentrations determined with the aid of a mixture of external reference compounds consisting of PAH purchased from various commercial firms. The identities and concentrations of 10 PAH in this mixture were thoroughly checked with a standard mixture

,m,

~.~

,m,

,m,

r~

r~

,u

,u

,m,

r~

~

~S L.

¢,zJ

1,==,

,ml

oO

'~" ill

~"

,.-,

C~l

Q

~=,

,m,

E

~J

E