Environ. Sci. Technol. 1997, 31, 2847-2853
Environmental Chemistry of Benzothiazoles Derived from Rubber CHRISTOPHER M. REDDY* AND JAMES G. QUINN Graduate School of Oceanography, University of Rhode Island, Narragansett, Rhode Island 02882-1197
Benzothiazole (BT), 2-hydroxybenzothiazole (HOBT), and 2-(4-morpholino)benzothiazole (24MoBT) leach from crumb rubber material (CRM) and asphalt containing 1-3% CRM. To determine whether benzothiazoles would be an environmental problem if roads containing CRM-modified asphalt (CMA) were built in the state of Rhode Island, the source and fate of these compounds were investigated. Benzothiazoles enter the environment from a number of sources such as the leaching of rubber products, fine particles of automobile tires, and antifreeze. Compared to the fluxes of benzothiazoles currently entering rivers from urban runoff, CMA roads may initially deliver substantially more benzothiazoles to the environment; however, with time this source may diminish as the road ages. Because the benzothiazoles are water soluble, it is unlikely that they will sorb to particles, settle to sediments, or be bioaccumulated. In addition, BT can be volatilized, and BT and HOBT can be microbially degraded. Therefore, the environmental chemistry of these compounds suggests that the inputs of benzothiazoles from CMA should not be harmful.
Introduction In a previous study (1), we found that benzothiazole (BT), 2-hydroxybenzothiazole (HOBT), and 2-(4-morpholino)benzothiazole (24MoBT) were the major organic compounds that
leach from rubber and asphalt containing 1-3% rubber. These compounds are breakdown products (BT and HOBT) in some of the vulcanization accelerators and antioxidants added to rubber during manufacturing (2, 3) and an impurity (24MoBT) in these materials (4). Crumb rubber material (CRM) is prepared by shredding automobile tires and removing their steel belts; the resulting material can be mixed with hot asphalt to produce CRM-modified asphalt (CMA) (5). The Rhode Island Department of Transportation (RIDOT) funded us to identify organic contaminants that leach from CRM and CMA because they wanted to know if they could safely comply with Section 1038(b) of the Intermodal Surface Transportation Efficiency Act of 1991 (6), which mandated that at least 20% of all federally funded highways must be built with CMA by the year 2000. Although this bill is currently unfunded and stalled in Congress, many states, such as Arizona and California, have miles of roads built with CMA (7). To determine whether the BT, HOBT, and 24MoBT leached from CMA would be a problem, we investigated the envi* Corresponding author: Telephone: (401) 874-6612; Fax: (401) 874-6811; e-mail:
[email protected].
S0013-936X(97)00078-3 CCC: $14.00
1997 American Chemical Society
ronmental chemistry of these compounds. First, we determined the quantifiable sources of these compounds to the environment. Second, we measured the amounts of BT, HOBT, and 24MoBT already in the environment to serve as a baseline against which new inputs from CMA could be compared. Specifically, we analyzed samples of urban runoff, road dust, urban particulate matter, river and pond water, and sediments from rivers, a pond, and an estuary. Third, we used simple calculations and laboratory experiments to predict the major processes that would affect the fate of CMAleached BT, HOBT, and 24MoBT in the environment. We considered photolysis, microbial degradation, volatilization, uptake in tissue, and sorption. We were particularly concerned about the sedimentary fate of these compounds because there had been conflicting reports in the literature. Spies et al. (4) found traces of BT and 24MoBT in San Francisco Bay, with the concentrations of 24MoBT as high as 500 ppb, and suggested that these compounds could be molecular markers for urban runoff; Brownlee et al. (8) suggested that, because these compounds are very water soluble, they should not be found in sediments.
Methods Organic solvents were purchased from Burdick & Jackson and were of pesticide grade. Pure BT, HOBT, and 2-methylbenzothiazole (MeBT) were obtained from Aldrich Chemical Co. 24MoBT was graciously donated to us by Dr. Hideshige Takada of the Tokyo University of Agriculture and Technology. Extracts were analyzed on a Hewlett-Packard 5890 Series II gas chromatograph equipped with a Hewlett-Packard 5971 mass selective detector (GC-MSD) with a splitless injection. Compounds were separated on a 30-M J&W Scientific DB5ms or DB-XLB fused-silica capillary column (25 mm i.d. and 0.25 µm film thickness), and the MSD was operated in the selected ion monitoring (SIM) mode. Data were collected with Hewlett-Packard ChemStation software. Compounds were quantified with calibration curves derived from their responses of their quantification ions relative to the internal standard MeBT. The quantification (boldface) and quality ions for each compound were BT (m/z 135, 108, 69), HOBT (m/z 151, 123, 96), 24MoBT (m/z 220, 163, 135), and MeBT (m/z 149, 108, 69). Concentrations in aqueous samples are expressed as parts per trillion (ppt), parts per billion (ppb), and parts per million (ppm) for ng L-1, µg L-1, and mg L-1, respectively. For solid samples, the dry weight concentrations are expressed as parts per billion (ppb) and parts per million (ppm) for ng g-1 and µg g-1, respectively. Analysis of CRM (Recycled Automobile Tires). To determine the total amount of BT, HOBT, and 24MoBT in CRM, a sample (∼150 µm diameter) was spiked with MeBT, refluxed in methanol for 2 h, cooled, and then filtered through a precombusted Whatman GF/C glass-fiber filter (particle retention of 1.2 µm). [This sample of CRM, brand name GF80A, was obtained from Rouse Rubber, St. Louis, MO, and was 1 of 12 different samples of CRM investigated previously (1)]. The filtrate was diluted with preextracted deionized water and extracted three times with methylene chloride. The extracts were combined, rotary-evaporated to ∼1 mL, and solvent-exchanged into hexane. The hexane extract was chromatographed, using nitrogen pressure, on a 0.5-cm (i.d.) × 15-cm column containing activated silica gel (Grace grade 922, 200-325 mesh size), and two fractions (F1 and F2) were collected. The first fraction (F1) was obtained by elution with 15 mL of a 50/50 mixture of hexane/methylene chloride. The second fraction (F2), which contained the benzothiazoles, was eluted with 15 mL of an 80/20 mixture of methylene
VOL. 31, NO. 10, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2847
chloride/acetonitrile, rotary-evaporated to a small volume, spiked with an external standard (o-terphenyl), and analyzed by GC-MSD. Leaching of CRM. The leaching of benzothiazoles from CRM was investigated by shaking a ∼4-gram sample with 100 mL of deionized water in a sealed centrifuge tube at 25 °C for 24 h. (In earlier experiments, we had found that equilibrium was reached in 20 h). The pH of the mixture was ∼5. The mixture was then filtered through a precombusted glass-fiber filter. The CRM was returned to the centrifuge tube, another 100 mL of deionized water was added, and the mixture was leached for another 24 h. The last step was repeated three more times. Each of the five filtrates was spiked with MeBT and extracted three times with 30 mL of methylene chloride. The methylene chloride extracts were combined, rotary-evaporated, and analyzed by GC-MSD. Antifreeze. Samples of used antifreeze were collected from the radiators of five cars. One sample of unused antifreeze was taken from its plastic container. Each sample (0.1-2 mL) was diluted with 15-30 mL of a 90/10 mixture of deionized water/methanol. To protonate and increase the extraction efficiency of HOBT [pKa ∼8 (9)], the pH was adjusted to ∼4 by adding a few drops of 4 N hydrochloric acid. (For the rest of this paper only protonated HOBT will be considered. Also, all other extractions for all samples were done at slightly acidic pH values, 4-6,which was an optimum pH range for extracting both BT and HOBT). Each unfiltered solution was spiked with MeBT and extracted three times with methylene chloride. The extracts were combined and analyzed by GC-MSD. Urban Runoff. Eleven samples (∼4 L) of urban runoff draining into the Pawtuxet, Woonasquatucket, and Moshassuck Rivers were collected during storms in February, May, and September of 1995 and filtered through a precombusted glass-fiber filter. A complete description of the sampling has been published (10). All samples were from roads built with only asphalt. Except for the February samples, in which five sequential samples were collected, all other samples were collected during the first 20 min of each storm. The filters (particulate phases) were frozen until analyzed (see below). The filtrate (dissolved phase) was spiked with MeBT and extracted three times with methylene chloride. The extracts were combined, fractionated with silica gel chromatography, and analyzed by GC-MSD. After being thawed, the filters were cut into 10-20 pieces with solvent-rinsed scissors, spiked with MeBT, refluxed in methanol for 2 h, cooled, and then filtered through a precombusted glass-fiber filter. The filtrate was then diluted with deionized water and extracted three times with methylene chloride. The extracts were combined, fractionated with silica gel chromatography, and analyzed by GC-MSD. For the February storm, the filters were analyzed for the 23 National Oceanic and Atmospheric Administration (NOAA) Status and Trends polycyclic aromatic hydrocarbons (PAHs) (11) but not benzothiazoles. Briefly, the filters were spiked with perdeurated PAH internal standards, refluxed in acetonitrile, and filtered. The filtrate was diluted with an equal portion of deionized water and extracted three times with hexane. The hexane extracts were combined and fractionated with silica gel chromatography, and the F1 fraction was analyzed by GC-MSD. Water from Highway Settling-Pond and the Pawtuxet River. Water (∼ 4 L) from a highway-settling pond (intersection of RI Routes 102 and 4) and from the Pawtuxet River (under RI Route 114 bridge) was collected and filtered through a precombusted glass-fiber filter. The dissolved phases were preserved with methlyene chloride and then analyzed according to the same method used for urban runoff. The particulate phases were not analyzed. Sediments, Road Dust, and Urban Particulate Matter. In August 1995, 12 surface sediments (top 2-3 cm) were
2848
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 10, 1997
collectedstwo from the highway settling-pond, 1 from the Pawtuxet River, and 9 from mid to lower Narragansett Bay. The bay sediments were collected by diver. Road and highway dust was collected by scooping it into a glass jar; the contents of the jar were sieved, and the
