Article pubs.acs.org/est
Comparison of Tire and Road Wear Particle Concentrations in Sediment for Watersheds in France, Japan, and the United States by Quantitative Pyrolysis GC/MS Analysis Ken M. Unice,* Marisa L. Kreider, and Julie M. Panko Cardno ChemRisk, 20 Stanwix Sreet, Suite 505, Pittsburgh, Pennsylvania 15222, United States S Supporting Information *
ABSTRACT: Impacts of surface runoff to aquatic species are an ongoing area of concern. Tire and road wear particles (TRWP) are a constituent of runoff, and determining accurate TRWP concentrations in sediment is necessary in order to evaluate the likelihood that these particles present a risk to the aquatic environment. TRWP consist of approximately equal mass fractions of tire tread rubber and road surface mineral encrustations. Sampling was completed in the Seine (France), Chesapeake (U.S.), and Yodo-Lake Biwa (Japan) watersheds to quantify TRWP in the surficial sediment of watersheds characterized by a wide diversity of population densities and land uses. By using a novel quantitative pyrolysis-GC/MS analysis for rubber polymer, we detected TRWP in 97% of the 149 sediment samples collected. The mean concentrations of TRWP were 4500 (n = 49; range = 62−11 600), 910 (n = 50; range = 50−4400) and 770 (n = 50; range = 26−4600) μg/g d.w. for the characterized portions of the Seine, Chesapeake and Yodo-Lake Biwa watersheds, respectively. A subset of samples from the watersheds (n = 45) was pooled to evaluate TRWP metals, grain size and organic carbon correlations by principal components analysis (PCA), which indicated that four components explain 90% of the variance. The PCA components appeared to correspond to (1) metal alloys possibly from brake wear (primarily Cu, Pb, Zn), (2) crustal minerals (primarily Al, V, Fe), (3) metals mediated by microbial immobilization (primarily Co, Mn, Fe with TOC), and (4) TRWP and other particulate deposition (primarily TRWP with grain size and TOC). This study should provide useful information for assessing potential aquatic effects related to tire service life.
■
INTRODUCTION Pollution from disperse anthropogenic sources has become a concern with respect to contamination of the aquatic environment, including the impact of roadway surface water runoff on sediment. Road runoff contains dust from road surfaces, crustal minerals, antifreeze and oils from automobiles, pesticides and fertilizers, and particulate associated with vehicle traffic, such as brake dust, exhaust emissions, and tire and road wear particles (TRWP).1 These materials wash into receiving water bodies during rainfall events, as well as during dry weather flow in urban settings, and thus have the potential to impact the aquatic environment. Additionally, fine road dust or soluble metals resuspended by wind and traffic can subsequently be deposited throughout a watershed and transported to sediment by overland flow.2−4 TRWP are formed at the frictional interface between the tire and the road, and contain contributions from both the tire tread and the pavement.5 The enrichment of the tread particles with mineral encrustations from the roadway surface decreases the polymer content in TRWP by a factor of 2 as compared to tread, such that TRWP consists of approximately equal parts by © 2013 American Chemical Society
mass of tread polymer and mineral encrustations from the roadway.5 TRWP are characterized by low toxicity to both water- and sediment-dwelling species based on the absence of observed adverse effects on test organisms from seven species at TRWP concentrations up to 10 000 mg TRWP/L sediment elutriate or 10 000 μg TRWP/g sediment d.w in both acute (Pseudokirchneriella subcapitata, Daphnia magna, and Pimphales promelas) and chronic (Chironomus dilutus, Hyalella azteca, Pimephales promelas, and Ceriodaphnia dubia) aquatic toxicity studies, respectively.6,7 In order to estimate the margin of exposure between toxicity test concentrations and environmental concentrations, an accurate estimate of TRWP concentration in environmental media is necessary. Based on their density and size distribution, TRWP are expected to primarily partition into the sediment and soil following road runoff, with a small fraction partitioning to air.5,8,9 Received: Revised: Accepted: Published: 8138
February 26, 2013 May 24, 2013 June 24, 2013 July 10, 2013 dx.doi.org/10.1021/es400871j | Environ. Sci. Technol. 2013, 47, 8138−8147
Environmental Science & Technology
Article
Table 1. Sediment Concentrations of Tire-Related Wear Particles Presented on Tread and TRWP Basis area France Troyes
St. Dizier Paris
Rouen United States Harrisburg
Washington, D.C
Japan Shiga
Kyoto/Mie
Hyogo Osaka
water body
n
mean TRWP (μg/g)a
locationb
La Seine River La Seine River La Boderonne River La Barse River Lac d’Orient Lac du Der La Marne River La Seine River La Seine River La Seine River La Seine. River La Seine. River
7 2 2 1 1 4 6 3 1 9 4 9
4800 (2800−7000) 350 (72−620) 4600 (4000−5200) 6800 1600 5600 (1100−11000) 3500 (64−9800) 6100 (1100−10000) 7400 6000 (160−11600) 3200 (300−8000) 3500 (62−9400)
U D U U U U C/D U C D U D
proximity to highway, national roads and regional park proximity to national road inside national regional park; near major highway proximity to national road; inside national regional park regional recreational lake; close to national road lake is a national bird refuge area proximity to an environmental protection area urbanized metropolitan area; high traffic density city center of major metropolitan area high traffic density; proximity to environmental protected areas urbanized metropolitan area; high traffic density high traffic density; locations inside national regional park
various creeks Susquehanna River Susquehanna Riverc Susquehanna River upstream creeks Potomac River Potomac Riverc Potomac River
4 2 4 3 7 5 13 12
670 (100−1400) 480 (380−580) 810 (110−1900) 1200 (380−1800) 990 (220−2200) 590 (180−1200) 640 (50−3000) 1400 (96−4400)
U U C D U U C D
regional and local parks; confluence with main channel main river channel upstream of urbanized area dense urban area; sediment depositional areas dense urban area; recreational fishing piers Remote areas adjacent to urbanized district. main river channel upstream of Washington, DC major metropolitan area; includes many national parks main river channel downstream of major metropolitan area
Lake Biwa Seta River Yasu River Uji River Yamashina River Katsura River Higashi-Takase River Kizu River Nabari River Ina River Mo River Yodo River Oh River Dojima River Akuta River
8 3 2 6 1 5 1 8 1 4 1 6 1 1 2
2800 (680−4600) 1400 (380−2200) 46 (28−64) 150 (32−500) 98 300 (28−1100) 900 500 (98−1800) 480 230 (94−360) 160 96 (28−340) 1100 1400 36 (26−46)
U C/D C/U C/U C C C C/U C C C D C C C
drinking water source for 16 million people mix of commercial to residential areas; mountainous areas rural/agricultural area; tributary of Lake Biwa continuation of Seta; rural and agricultural areas; Amagase Dam primarily residential; agricultural; tributary of Uji River range from remote mountainous area to Kyoto; endangered fish Kyoto metropolitan area; tributary of Uji River range from remote agricultural area to residential areas residential area; tributary of Kizu River range from residential mountainous area to Hyogo urban area Hyogo metropolitan area downstream of Kyoto metropolitan area urban Osaka area; tributary of lower Yodo River urban Osaka area; tributary of lower Yodo River residential district in metropolitan area; tributary of upper Yodo River
location selection rationale
a
Range indicates minimum and maximum result. bRelative location of samples within channel indicated as upstream of city center (U), center city (C) or downstream of city center (D). cIncludes samples collected from tributary and main channel.
environment.15,17,19−21 In the case of the org-Zn marker,22 recent research has shown that the dichloromethane extraction does not reliably exclude inorganic ZnO from partitioning from tread polymer into the solvent.23 Inorganic Zn, while found in tread polymer, is also found in deicing salts, combustion exhaust, galvanized metal in automobiles, road barriers, fuel, oil, and brake linings, all of which can contribute to road runoff.10,24−26 In one detailed study of an urban highway site with average daily traffic count of 150 000 vehicles, total zinc was found to be present primarily in the dissolved phase,27 further suggesting that zinc-related markers are unlikely to be reliable markers for TRWP in the environment, despite the relatively high content of Zn in tread polymer (about 1%).5 Recently, a marker using pyrolysis products of tread polymers was developed for quantifying TRWP in environmental media.23,28,29 In pyrolysis analysis, a sample is thermally decomposed under high temperature, and the characteristic fragments of tread polymers are quantified by a mass selective detector, such as gas chromatography−mass spectroscopy
The need for a TRWP marker in sediment is widely recognized,10,11 and researchers have attempted to quantify TRWP in sediment by using various markers for rubber, including 2,4-morpholinyl-benzothiazole (2,4MoBT), N-cyclohexyl-2-benzothiazolamine (NCBA), benzothiazole (BT), 2hydroxybenzothiazole (HOBT), and organic zinc (orgZn)12−17 (see Table S1 in the Supporting Information (SI)). However, while these markers may be present in tires, several limitations in their application make estimating tread polymer or TRWP concentration uncertain. For example, Voparil, et al.18 estimated that in heavy traffic areas, sediment could contain as much as 15% tread polymer. This estimate, however, was based on a study relying on the marker 2,4MoBT in sediment where an internal standard was not used, and analyte recovery was not evaluated (ref 2 in SI Table S1).12,14,18 Further, benzothiazoles are not specific to tread polymer, which may result in overestimating TRWP in the environment, and conversely, can solubilize from tread polymer into the water column, which may lead to underestimating TRWP in the 8139
dx.doi.org/10.1021/es400871j | Environ. Sci. Technol. 2013, 47, 8138−8147
Environmental Science & Technology
Article
between 14 to 17 million, and population densities range from a low of 96 persons/km2 in the Chesapeake watershed to a high of 1700 persons/km2 in the Yodo watershed.35−37 Sample Location Selection. The primary objective of the study was to quantify TRWP concentrations in a wide variety of potential habitats of aquatic and sediment dwelling organisms, using a stratified approach to incorporate a representative diversity of areas including main channel locations upstream and downstream of urban areas, within environmental protection areas, and in tributaries and lakes within the watersheds (see Table 1). Sample location segments were selected in collaboration with local field teams using geographical information system (GIS) databases and descriptive data, including traffic load, population density, and land use. All locations were required to have an identifiable vehicular traffic source, and final sample locations were based on the ability to obtain access to the sampling site and on the presence of surficial sediment in the channel bottom. The precise location of sample collection was randomly selected by the field team. A total of 149 sediment samples were collected from three watersheds (see Table 1 and Figure 1), and detailed maps of the sampling areas are presented in SI Figure S1 panels “a” through “q”. In total, 93 samples were collected downstream or within population centers, 44 samples were collected upstream of populated areas, and 12 samples were collected from lakes. The assignment of “upstream” and “downstream” attributes to each sample was relative to the local area, and several areas identified as “upstream” receive flow impacted by downstream flow from urban areas. In aggregate, 89 samples were collected from areas with a population density less than or equal to 3000 persons/km2, and 60 samples were collected from areas with population density exceeding 3000 persons/km2. Daily average traffic counts for roadways situated near the sampling areas ranged from fewer than 10 000 vehicles/day to more than 250 000 vehicles/day based on the traffic data available from the sources listed in SI Table S2. In the Seine watershed, samples were collected upstream and downstream (or within) the population centers of Troyes, St. Dizier, Paris, and Rouen, with human populations varying from 30 000 persons in St. Dizier to 12 million in the Paris metropolitan area. In the Chesapeake watershed, sampling was focused on Harrisburg and the Washington, DC metropolitan areas with populations of 480 00 and 5.5 million persons, respectively. In the Yodo watershed, sampling focused on Lake Biwa, which is an important regional water resource. Extensive sampling was also performed in prefectures downstream of Lake Biwa, including Shiga, Kyoto, Mia, Hyogo, and Osaka. Human populations in these prefectures range from a low of 1 million in Shiga to a high of 9 million in Osaka. Sample Collection. Sediment samples were collected by experienced environmental field technicians using stainless steel hand trowels, coring tool, or clamshell dredge. Depending on site accessibility, samples were collected either from the shore or by boat. Samples were oven-dried at 105 °C for 24 h after collection. Samples were collected between October, 2010 and May, 2011 in accordance with local, regional, and national regulations, and permits were obtained when necessary. Samples were packaged in the country of collection, and shipped to Chemical Evaluation Research Institute (CERI) (Tokyo, Japan) for pyrolysis-GC/MS analysis. Pyrolysis GC/MS Analysis. Samples were analyzed using methods previously described23 Briefly, approximately 20 mg of dry-sieved (
