Aerosol-Bound Emissions of Polycyclic Aromatic Hydrocarbons and

Apr 4, 2003 - Citation data is made available by participants in CrossRef's Cited-by Linking service. For a more comprehensive list of citations to th...
0 downloads 0 Views 111KB Size
Environ. Sci. Technol. 2003, 37, 2109-2113

Aerosol-Bound Emissions of Polycyclic Aromatic Hydrocarbons and Sterols from Aeration Tanks of a Municipal Waste Water Treatment Plant MICHAEL RADKE* AND REIMER HERRMANN Chair of Hydrology, University of Bayreuth, 95440 Bayreuth, Germany

Aeration tanks of wastewater treatment plants (WWTP) are a potential source of atmospheric aerosol particles. Several groups of organic compounds (sterols, polycyclic aromatic hydrocarbons, estrogens) were analyzed in aerosol particles sampled at a municipal WWTP, and the particle size distribution was measured directly with optical particle counters. Aerosol emissions from an activated treatment tank equipped with fine bubble diffusers were low; however, at the preaeration tank equipped with coarse bubble diffusers, sterol concentrations up to 14 ng m-3 were measured. Directly next to the tank, sterols were associated mainly to particles with aerodynamic diameter >1.35 µm. The results suggest that coprostanol could be a useful tracer for monitoring the emission of aerosol particles from WWTPs. Moreover, wastewater treatment could contribute substantially to the atmospheric concentrations of cholesterol and 24-ethylcholesterol. Aeration tanks with fine bubble diffusers are no major source of atmospheric aerosol particles, whereas coarse bubbling devices seem to emit considerable amounts of aerosol particles.

Introduction Bursting air bubbles at the sea surface are responsible for the formation of most of the atmospheric sea salt particles, the so-called sea spray. At wastewater treatment plants (WWTP), a similar process occurs when air is blown into the wastewater, especially at the preaeration and the activated treatment tanks. It has been shown that this process leads to the emission of bacteria from wastewater treatment tanks (1-4). To date, however, little is known about the aerosol bound emissions of organic constituents from wastewater. Schmid et al. (5) determined rather low emissions of particulate organic carbon (POC) above a pilot scale aeration tank equipped with diaphragm blowers. In contrast, Lepri et al. (6). founds besides other groups of chemicalssconsiderable amounts of polycyclic aromatic hydrocarbons (PAHs) in aerosol particles sampled directly above an aeration tank equipped with turbine aerators. To evaluate the influence of different aeration systems on the aerosol formation, the objective of this study was to quantify the emissions of several organic compounds from an activated treatment tank * Corresponding author phone: +49/921/552297; fax: +49/921/ 552366; e-mail: [email protected]. 10.1021/es025926g CCC: $25.00 Published on Web 04/04/2003

 2003 American Chemical Society

equipped with fine bubble diffusers and a preaeration tank with coarse bubble aeration devices. Wastewater contains high concentrations of nonpolar compounds, such as fats and oils, which form surface films at the air-water interface. Hydrophobic, nonpolar organic trace compounds are enriched in these surface films up to a factor >100 compared to the bulk phase (7, 8). These compounds may comprise not only typical wastewater constituents such as aliphatic hydrocarbons, estrogenic, or steroidal compounds but also pollutants such as PAHs or pharmaceutical compounds. Aerosol particles are released from a bursting bubble in two steps: first the so-called film-droplets and second the so-called jet-droplets are formed (9). The film-droplets are ejected into the air immediately after the collapse of the air bubble at the water surface. They are formed from the surface film at the air-water interface around the bubble. The particle diameter (dp) of film-droplets is within the range of 0.08 µm < dp < 500 µm (10-12). After that, a jet of liquid is ejected from beneath the air bubble, leading to the formation of jet-droplets, where dp is typically in the range 100-200 µm (13, 14). In laboratory experiments with bacterial suspensions, it has been shown that a bursting bubble produces up to six jet-droplets. The first droplet was highly enriched in bacteria scavenged from the bulk water, whereas the following ones showed nearly no enrichment (15). The ejection height of jet droplets decreased with each consecutive drop (15), progressively reducing the importance for particle emission into the atmosphere. The number and the size of aerosol droplets formed also depend on bubble numbers per volume and bubble size (2, 13, 16). In light of the high build-up potential of surface films on wastewater and the subsequent importance of surface films for the aerosol formation, we investigated the potential for particle-bound emissions of nonpolar wastewater constituents at aerated treatment tanks of a WWTP. This was addressed using a high-volume and sizesegregated aerosol particle sampling scheme. We focused on representatives of three groups of organic compounds: faecal sterols (cholesterol, coprostanol, and 24ethylcholesterol), estrogens (estrone, β-estradiol), and PAHs. PAHs are quite ubiquitous in the environment and were therefore analyzed as a class of major contaminants in the urban atmosphere. The two other groups were tested for their suitability as tracers for sewage-bound aerosol particles. Cholesterol and 24-ethylcholesterol are present in epicuticular waxes of higher plants (17); cholesterol is also present in animal fats and oils (18). These compounds can be emitted by dispersion of plant wax detritus, wood combustion (19, 20), cooking processes (18, 21), and tobacco smoke (22). Unlike these compounds, coprostanol is not naturally occurring (23). It is formed by bacterial reduction of cholesterol and other sterols in the digestive tract of man and higher animals (24) and accounts for approximately 60% of the total sterol content in human faeces (25). Coprostanol is widely used as marker for sewage contamination in surface and seawater (26-28). To our knowledge, no potential atmospheric sources of coprostanol and estrogens exist other than wastewater treatment.

Experimental Methods Sample Collection. The experiments were carried out at two treatment tanks of the WWTP of Bayreuth, Germany, a plant with a capacity of about 300 000 population equivalents. Mean inflow during the sampling period was about 23 000 m3 d-1. In the preaeration tank (area approximately 220 m2) VOL. 37, NO. 10, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2109

FIGURE 1. (a) Schematic drawing of the area around the preaeration tank. HV1, HV2, HV3: locations of high volume samplers. LPI: location of LPI. (b) Prevailing wind directions during the sampling periods at the preaeration tank. located directly after the trash rack at the inflow, rather outdated coarse bubble aeration equipment (details about diffusers not available) was used. In contrast, the activated treatment tanks were equipped with fine bubble diffusers (Brandol 60, USF Schumacher, Crailsheim, Germany; bubble size: 2-3 mm, total length: 3270 m, tank area: 4270 m2). All tanks were aerated with pressurized air and were operated continuously. We sampled size-segregated aerosol particles with a 4-stage low-pressure impactor (LPI, Berner-type, Hauke KG, Austria). The impaction plates were covered with aluminum foil. Glass-fiber filters (GF 9, Schleicher & Schuell, Dassel, Germany) were used as backup filters. The cutoff diameters of the impactor stages were 12.2, 4.05, 1.35, 0.45, and 0.15 µm, respectively. The air flow through the LPI was set to 8.89 m3 h-1 by a critical nozzle. To avoid aspiration of urban aerosol particles and to obtain a well-defined source area, samples from the activated sludge tanks were taken with a tent consisting of plastic tubes and polyethylene foil. The pyramidal construction, with an area of 9.5 m2 and a height of 2 m, was floating on the wastewater. A heated manifold (30 °C) was used to avoid condensation of water during sampling (5). The manifold and the impactor were mounted directly on top of the tent to minimize particle losses due to impaction in tubes. Aerosol particles were collected for periods of 5-7 days. The tent could not be used to sample the preaeration tank because of strong turbulence in the wastewater. Instead, several high-volume samples were taken simultaneously at different distances from the tank (Figure 1) in order to investigate the spatial distribution of aerosol concentrations. Self-constructed samplers were used to collect particles on glass-fiber filters (GF 6, Schleicher und Schuell) for periods between 24 and 96 h at air-flow rates between 25 and 40 m3 h-1. Additionally, samples with the 4-stage LPI were taken directly next to the tank (approximately 1 m above the water surface). After collection, the samples were taken to the laboratory and stored at -20 °C until extraction. During sampling at the preaeration tank, meteorological parameters (wind speed, wind direction, temperature and relative humidity) were measured near HV2 (Figure 1). For calculating the prevailing wind direction, only measurements with wind speed >0.5 m s-1 were used. Sample Extraction and Analysis. Prior to extraction, the aerosol samples were spiked with a solution containing approximately 1 µg of β-estradiol 17-acetate (all sterols were purchased from Sigma-Aldrich, Taufkirchen, Germany) as recovery standard. The samples were extracted for 24 h using a Soxhlet-apparatus with 150 mL of hexane:acetone (1:1, v:v). 2110

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 10, 2003

All solvents used were nano- or ultragrade solvents purchased from LGC Promochem (Wesel, Germany). After evaporation to approximately 1 mL, the samples were separated into three fractions using alox-silica-columns (3 g of alox III over 5 g of silica I). The first fraction (elution with 35 mL of hexane) was discarded. The second fraction (30 mL of hexane: dichloromethane, 3:1, v:v) contained the PAHs, and the third fraction (25 mL of toluene:acetone, 4:1, v:v) contained the sterols and estrogens. The fractions of interest were evaporated to dryness with a rotary evaporator. Prior to analysis, the samples were redissolved in 500 µL of hexane. PAHs were analyzed by HPLC-FD/UV using a C18-PAHcolumn (150 mm × 3 mm; Macherey & Nagel, Du ¨ ren, Germany) and gradient elution with acetonitrile-watermethanol mixtures. Concentrations were determined using external standards (compounds purchased from Promochem). The sterolic and estrogenic compounds were derivatized with 50 µL of Sylon BFT (Sigma-Aldrich) for 1 h at 60 °C, evaporated to dryness, redissolved in 500 µL of a hexane solution containing the injection standards, and analyzed by GC-MS (Hewlett-Packard GC 5890, MSD 5970B). Three Polychlorinated Biphenyls (congeners 20, 46, and 205; Promochem) were used as injection standards for the GCMS measurements. The GC was equipped with a DB-5 column (30 m × 0.25 mm × 0.25 µm; J&W Scientific, Folsom, U.S.A.) with helium as carrier gas. The temperature program started at 80 °C for 1 min, then heating with 15 °C min-1 to 250 °C and holding for 5 min, and finally heating with a ramp of 10 °C min-1 to 300 °C. The injector was set to 290 °C, the transfer line to 300 °C. Injection of 1 µL was done in splitless mode. The measurements were performed in SIM mode. Three characteristic fragment ions were recorded for each compound. Wastewater was sampled as grab samples. Particulate and liquid fractions were separated by filtration with glass fiber filters (GF 6). The particulate fraction was freeze-dried and extracted with a Soxhlet-extraction method similar to the one used for aerosol particles. The liquid fraction was extracted three times with hexanes (29). Due to small amounts of sample, we could not determine analytical replicates of all samples. Based on data of the recovery standards and on experiments with spiked samples, our method was able to quantify the compounds within (15%. Recovery rates for all compounds were determined experimentally to be in the range from 80% to 100%. Laboratory blanks of aluminum foil and glass fiber filters were determined routinely for each extraction series. Optical Measurements. We performed several measurements with the optical particle counter LAS-X (PMS Inc., Boulder, CO) at the outlet of the floating tent on the activated treatment tank. The particles were collected with a short tube (length approximately 1 m) from the interior of the tent. Several measurements were carried out directly next to the activated treatment tanks. To get reference particle size distributions of urban aerosol, several measurements were performed at the campus of Bayreuth University. The LAS-X counts the particle numbers in 15 size classes in the range 0.1 µm < dp < 2.75 µm. Particle volume distributions were calculated using the geometric mean diameter of each class. A HC-15 spectrometer (Polytec, Waldbronn, Germany), which counts particles with dp in the range 0.59 µm to 21.6 µm in 127 classes, was used to detect particles above the upper limit of the LAS-X range.

Results and Discussion Optical Measurements The volume distribution (Figure 2) measured with LAS-X next to the activated treatment tank showed a distinct maximum in the accumulation mode which is typical for urban aerosol. The distribution at the university site, which is not influenced by the WWTP, was very similar.

TABLE 1. Aerosol Concentrations at the Outlet of the Tent on the Activated Treatment Tanka aerosol concentration (pg m-3) compound

activated treatment tank (n ) 2)

urban aerosol (n ) 3)

phenanthrene fluoranthene pyrene benzo[a]anthracene chrysene benzo[b]fluoranthene benzo[k]fluoranthene benzo[a]pyrene dibenz[a,h]anthracene benzo[g,h,i]perylene indeno[1,2,3-c,d]pyrene cholesterol coprostanol

0.7-1.0