Hazardous Chemicals in Synthetic Turf Materials and Their Bioaccessibility in Digestive ﬂuids

Home , Potassium chloride, Sodium chloride

www.nature.com/jes

Hazardous chemicals in synthetic turf materials and their bioaccessibility in digestive ﬂuids

JUNFENG (JIM) ZHANGa, IN-KYU HANa,b,LINZHANGa AND WILLIAM CRAINc aSchool of Public Health, University of Medicine and Dentistry of New Jersey, 683 Hoes Lane West, Piscataway, New Jersey, USA bJohns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA cThe City College of New York, New York, New York, USA

Many synthetic turf fields consist of not only artificial grass but also rubber granules that are used as infill. The public concerns about toxic chemicals possibly contained in either artificial (polyethylene) grass fibers or rubber granules have been escalating but are based on very limited information available to date. The aim of this research was to obtain data that will help assess potential health risks associated with chemical exposure. In this small- scale study, we collected seven samples of rubber granules and one sample of artificial grass fiber from synthetic turf fields at different ages of the fields. We analyzed these samples to determine the contents (maximum concentrations) of polycyclic aromatic hydrocarbons (PAHs) and several metals (Zn, Cr, As, Cd, and Pb). We also analyzed these samples to determine their bioaccessible fractions of PAHs and metals in synthetic digestive fluids including saliva, gastric fluid, and intestinal fluid through a laboratory simulation technique. Our findings include: (1) rubber granules often, especially when the synthetic turf fields were newer, contained PAHs at levels above health-based soil standards. The levels of PAHs generally appear to decline as the field ages. However, the decay trend may be complicated by adding new rubber granules to compensate for the loss of the material. (2) PAHs contained in rubber granules had zero or near-zero bioaccessibility in the synthetic digestive fluids. (3) The zinc contents were found to far exceed the soil limit. (4) Except one sample with a moderate lead content of 53 p.p.m., the other samples had relatively low concentrations of lead (3.12–5.76 p.p.m.), accordingto soil standards. However, 24.7–44.2% of the lead in the rubber granules was bioaccessible in the synthetic gastric fluid. (5) The artificial grass fiber sample showed a chromium content of 3.93 p.p.m., and 34.6% and 54.0% bioaccessibility of lead in the synthetic gastric and intestinal fluids, respectively. Journal of Exposure Science and Environmental Epidemiology (2008) 18, 600–607; doi:10.1038/jes.2008.55; published online 27 August 2008

Keywords: synthetic turf, PAHs, lead, heavy metals, bioaccessibility.

Introduction individual legislators in Minnesota, New Jersey, and New York have called for moratoria on new synthetic turf Across the United States, Canada, and Europe, parks and installations until more research is conducted. A bill in athletic facilities are installing a ‘‘new generation’’ synthetic California calls for more research on the risks, and Connecti- turf. It is springier than the old AstroTurf and feels more like cut has allotted a modest amount of money for their study. natural grass. There are about 3500 full-sized synthetic turf In April 2008, concern intensified when the State of New fields in the United States, with 900–1000 new fields being Jersey Department of Health and Senior Services reported installed each year (McCarthy and Berkowitz, 2008). But high concentrations of lead in the plastic grass fibers in some despite the new turf’s popularity, local residents and of the old-variety AstroTurf fields (NJDHSS, 2008). Follow- government officials are increasingly concerned that it might up simulations in our laboratory indicated that the lead could contain toxic chemicals. Pilot studies reported in newsletters be absorbed by the human digestive tract (NBC, 2008). and on websites have raised the possibility that the tiny (0.5– These findings prompted the federal Centers for Disease 3 mm in size) rubber granules, which are made from discarded Control (CDC) to issue a June 18 health advisory (CDC, tires and contribute to the turf’s resiliency, might contain 2008). The CDC recommended precautions such as washing metals, polycyclic aromatic hydrocarbons (PAHs), or volatile hands after playing on synthetic turf fields, but the CDC organic compounds with toxic potential (see Brown, 2007; acknowledged the paucity of research information on the Claudio, 2008). These studies are very preliminary, but hazards. To assess the magnitude of the health risks of the ‘‘new generation’’ synthetic turf, public health officials need 1. Address all correspondence to: Dr. Junfeng (Jim) Zhang, UMDNJ- systematic information on the concentrations and variability SPH, 683 Hoes Lane West, Piscataway, NJ 08854, USA. ofanytoxicantsinit.Ourresearchstrivestobuildonthefew Tel.: 1 732 235 5405. Fax: 1 732 235 4004. þ þ studies (e.g., Plesser and Lund, 2004, in Norway), which E-mail: [email protected] Received 19 July 2008; accepted 4 August 2008; published online 27 have contributed such information with respect to PAHs and August 2008 selected metals. Hazardous chemicals and digestive fluids Zhang et al.

More critically, data are needed not only on the content of All the samples were collected by hand on artificial turf fields toxicants but also on their bioavailabilityFwhether they can be and immediately placed into clean glass bottles and tightly absorbed into the body through exposure routes such as caped. The bottles were precleaned to be suitable for the ingestion, inhalation, and dermal contact. Our work explores the collection of water samples and urine samples in our other ingestion route, which we see as a particularly likely route for the studies. Sample information is summarized in Table 1. All the absorption of metals such as lead, chromium, and cadmium. eight samples were analyzed for their contents of PAHs. Previous studies on ingestion have been incomplete. In a Except samples 5, 6, and 7, all the other five samples were widely cited article, Birkholz et al. (2003) evaluated the analyzed for selected toxic metals. In addition, samples 3, 4, possibility that ingested tire rubber material from flat and 8 were analyzed for the bioaccessibility of PAHs and playground surfaces produces cancer. On the basis of the metals in digestive fluids including synthetic human saliva, results of in vitro genotoxicity assays, the investigators gastric fluid, and intestinal fluid. The methods for determin- concluded that the risk is negligible. However, the investiga- ing the contents and bioaccessible fractions of PAHs and tors did not specify the potentially harmful chemicals they metals are described below. tested. The U.S. Consumer Product Safety Commission All the samples were stored in a refrigerator until 24 h (CPSC, 2008) estimated the amount of lead picked up by before analysis. One day before the extraction procedure, hand contact with artificial grass fibers, which might be each sample was placed in a dessicator at room temperature. transferred to the mouth and ingested. However, the CPSC We were not able to analyze all the eight samples for metals didn’t directly study the lead concentrations that can be and for bioaccessibility tests due to equipment and budgetary absorbed in the digestive tract. In a report to the State of constraints. California, the Office of Environmental Health Hazard Assessment (OEHHA, 2007) summarized its study of Extraction for Determining PAHs Content simulated digestive tract absorption of heavy metals and We used a Soxhlet apparatus for the extraction of PAHs volatile organic compounds from tire shreds. OEHHA in the seven samples of rubber granules and the sample of concluded that the risks are minimal, but it only simulated artificial grass fiber. This method or similar methods absorption within the stomach environment. In this study, we have been used previously to determine PAHs contents in estimated the fractions of PAHs and metals that can be food and PAHs concentrations in air (Howard et al., 1986; dissolved into three synthetic digestive fluids: saliva, gastric Smith and Harrison, 1996; Han et al., 2007). In this fluid, and intestinal fluid. These fractions are referred as procedure, an aliquot of each sample (0.5–2.5 g, typically bioaccessible fractions (i.e., the fractions that are dissolved in 1.0 g) was placed in the Soxhlet apparatus filled with 150 ml the biofluids) not exactly bioavailable fractions (i.e., the of dichloromethane for B16 h at 601C. After the solution fractions that are absorbed by the body) (Yu et al., 2006), hadbeencooleddowntoroomtemperature,itwas although the use of the terminologies may be mixed in the concentrated to approximately 3–5 ml by a rotary evapora- literature. We examined the contents and bioaccessibility of tion apparatus at 401C. This concentrated solution was then bothPAHsandmetalsfromrubbergranulesaswellasfrom transferredintoa5mlconicalglasstubeandfurther plastic fibers. concentrated to B0.1 ml with pure nitrogen gas. The resulting solution was made to exactly 1 ml by adding Methods acetonitrile and then filtered through a 0.2 mm pore-size Sep- Pak C18 cartridge (Waters, Milford, MA, USA). This final We collected seven samples of rubber granules (artificial turf extract solution was then ready for analysis of PAHs using an infill) and one sample of artificial (polyethylene) grass fiber. HPLC system.

Ta bl e 1 . Sample information.

Sample Location of turf ﬁeld Turf brand Turf installation Sample collection Age of turf when sampled no. name date date

1 Riverside Park, Manhattan, NY A-turf March 2006 May 2006 B2months 2 Riverside Park, Manhattan, NY A-turf March 2006 June 2006 B3months 3 Riverside Park, Manhattan, NY A-turf March 2006 January 2008 B22 months 4a Riverside Park, Manhattan, NY A-turf March 2006 April 2008 B25 months 5 Parade Grounds, Brooklyn, NY FieldTurf June 2001 October 2006 B5 years and 4 months 6 Parade Grounds, Brooklyn, NY FieldTurf June 2001 October 2006 B5 years and 4 months 7 Sara Roosevelt Park, Manhattan, NY FieldTurf August 2006 January 2007 B5months 8 East Rochester High School, Rochester, NY Astroplay 2001 March 2008 B7years aFibers were collected instead rubber granules.

Journal of Exposure Science and Environmental Epidemiology (2008) 18(6) 601 Zhang et al. Hazardous chemicals and digestive ﬂuids

The extraction was performed in a hood covered with Three of the samples (3, 4, and 8) were analyzed for aluminum foils to prevent photo-degradation and laboratory bioaccessible PAHs and metals in the three synthetic fluids. contamination. In addition, lab blanks and solvent blanks This was performed using a method of sequential extraction were analyzed to examine any lab contaminations. In the that, in principle, mimics the human digestive tract (Ellickson event of contamination, the extraction and analysis were et al., 2002; Yu et al., 2006). In the analysis of saliva re-performed, according to standard quality assurance bioaccessibility, we mixed 2.5 g of each rubber granule or procedures established in our laboratory for analyzing artificial grass fiber sample with 8 ml of the synthetic saliva. samples of air, food, and soil. The mixture was incubated in a water bath at body temperature (371C) with constant shaking at 90 r.p.m. for Extraction for Determining Metals Content 15 min. In the analysis of gastric bioaccessibility, we mixed We used a microwave-assisted digestion method that had 2.5 g of each sample with 8 ml of the synthetic saliva and been used for extracting heavy metals in food, soils, and 100 ml of the synthetic gastric fluid. The mixture was particles collected on filters (e.g., Yu et al., 2006). We incubated in a water bath at 371C for 2 h. In the analysis of first placed B100 mg aliquot of each sample on a tray intestinal bioaccessibility, we first followed the same incuba- prewashed with nitric acid, then transferred the sample into tion procedure for the gastric bioaccessibility analysis, then an HP-500 microwave oven digestion vessel, and then added 100 ml of the synthetic intestinal fluid to the mixture, added 2 ml of high-purity nitric acid (Fischer optima grade) and then incubated the mixture at 371C for another 2 h. into the vessel. After the vessels containing the samples and After cooling down at room temperature, the mixtures nitric acid had been placed inside a hood at room were split into two aliquots. One aliquot was used for PAHs temperature for B24 h, they were digested at 80% of analysis and the other one for metals analysis. The metals 300 W power for 20 min using a Marsx microwave oven aliquots were centrifuged for 20 min at 3400 r.p.m. and then (CEM Corp., NC, USA). After cooling down at room filtered through a Whatman 0.45 mm syringe filter. The temperature, the digested samples were diluted with deionized PAHs aliquots were extracted successively for three times water to make B5% nitric acid solutions. These solutions with 50 ml methylene chloride for the gastric and intestinal were transferred into 50 ml conical tubes and centrifuged at extracts or 10 ml methylene chloride for the saliva extracts.

4000 r.p.m. for 20 min to separate solids and liquids. The The extracts were passed through anhydrous Na2SO4 and liquids were filtered with 0.45mm Whatman cellulose GD/X then concentrated to 1–2 ml using a rotary evaporator. The syringe filters, ready for analysis of metals using an extracts were further concentrated to 0.1 ml under gentle pure inductively couple plasma mass spectrometry (ICP-MS) nitrogen gas, then made to 1 ml volume with acetonitrile, and system. then filtered through 0.2 mm Sep-Pak C18 cartridges. These final solutions were analyzed for PAHs. Extraction in Digestive Fluids In the bioaccessibility analysis for metals, we conducted We used three synthetic digestive fluids in simulating the extraction and subsequent metals analysis in duplicates. If bioaccessibility of PAHs and metals in the human digestive one of the duplicates for each sample had a value below the tract. These include synthetic saliva, gastric fluid, and limit of detection, we regarded the sample as non-detectable. intestinal fluid. The composition of these fluids was reported Otherwise, we took the average of the duplicates. (The in the previous studies of dust and soil bioaccessibility sample number was too small to quantify the reproducibility, (Ellickson et al., 2002; Yu et al., 2006). The synthetic saliva which showed a large variation among different chemical was prepared by mixing 0.004 M of calcium chloride species and across different synthetic biofluids.) However, we dehydrate (ACS grade, 99.0–105.0%; EMD Chemicals did not conduct duplicate extraction and analysis for PAHs, Inc., NJ, USA), 0.4% (w/v) of gastric mucin (MP due to time and budgetary constraints. Biomedicals, OH, USA), 0.005 M of potassium chloride (99.999%; Aldrich, MO, USA), 0.007 M of sodium chloride Analysis of PAHs (99.999%), 0.004 M of sodium phosphate dibasic A Waters HPLC 2695 system equipped with a Waters 2475 (499.999%; Fluka, MO, USA), and 0.007 M of urea fluorescence detector (Waters) was used for PAHs analysis. (ACS grade, 99.0–100.5%; Sigma-Aldrich, MO, USA). The analytical column was a Supelco PAH column The synthetic gastric fluid was prepared following the (5 mm Â 250 mm, 2.1 mm). The mobile phase program shown recommendations of the US Pharmacopoeia for drug in Table 2 was used at a flow rate of 1 ml/min. Analytical dissolution studies. The gastric fluid was prepared by mixing detection limits were determined using 1 ml of acetonitrile 0.03 M of sodium chloride (99.999%), 0.084 M of hydro- solvent. Recovery and analytical precision are also summar- chloric acid, and 0.32% (w/v) of pepsin (purified grade; izedinTable2.Thismethodwasusedtoquantifythe15 J.T. Baker, NJ, USA) (Ellickson et al., 2002). The synthetic PAHs listed in Table 3. These 15 PAHs were on the US intestinal fluid was a 0.2 M solution of sodium bicarbonate Environmental Protection Agency’s list of priority hazardous (499.5%; Sigma-Aldrich) (Yu et al., 2006). chemicals.

602 Journal of Exposure Science and Environmental Epidemiology (2008) 18(6) Hazardous chemicals and digestive ﬂuids Zhang et al.

Ta bl e 2. Gradient mobile phase program and ﬂuorescence detector program used for PAH analysis.

Mobile phase Detector

Time 50% acetonitrile 100% acetonitrile Time Excitation, Emission, (min) in DI water (%) (%) (min) l (nm) l (nm)

0 100 0 0 270 350 2 100 0 21 250 400 35 0 100 30–60 280 425 50 0 100 55 100 0 60 100 0

Abbreviations: DI, deionized; PAH, polycyclic aromatic hydrocarbon.

Ta bl e 3 . LOD and RSD for repeated sample analysis (N ¼ 6). concentrations may be considered PAH contents of the sampled materials, because the extract procedure was PAHs LOD RSD (%) intended for obtaining maximum amounts of PAHs out of the samples. The results are shown in Table 4, exhibiting a (ng/ml extract) mg/kg (p.p.b.)a range of total PAHs (the sum of the 15 target PAHs) from Naphthalene 0.860 0.860 20 4.40 p.p.m. (i.e., mg/kg) to 38.15 p.p.m. across the seven Acenaphthene 0.470 0.470 16 samples of rubber granules, higher than 0.40 p.p.m. for the Fluorene 1.174 1.174 5 sole artificial grass fiber sample (sample 4). Phenanthrene 0.234 0.234 34 Anthracene 0.020 0.020 0 Using nitric acid digestion under heat and pressure, we Fluoranthene 0.718 0.718 45 determined metals contents of the sampled materials. Sample Pyrene 0.735 0.735 18 1 and 2 had zinc (Zn) concentrations at 5710 and Benzo(a)anthracene 0.280 0.280 5 9988 p.p.m., respectively (Table 5). Five of the eight samples Chrysene 0.696 0.696 12 were analyzed for the following four metals: chromium (Cr), Benzo(b)fluoranthene 0.134 0.134 2 Benzo(k)fluoranthene 0.009 0.009 7 arsenic (As), cadmium (Cd), and lead (Pb). As shown in Benzo(a)pyrene 0.023 0.023 16 Table 5, Cr and Pb were detected in all the five samples, As Dibenzo(a,h)anthracene 0.269 0.269 5 was detected in all but one rubber granule sample (sample 3), Benzo(ghi)perylene 0.249 0.249 5 and Cd was detected in all four rubber granule samples but Abbreviations: LOD, limits of detection; PAH; polycyclic aromatic the artificial grass fiber sample. Interestingly, the fiber sample hydrocarbon; RSD, relative standard deviation. had the highest Cr concentration among the five samples. aThis is the limit of detection in mg per kg of samples, based on extracting In the bioaccessibility analysis of PAHs, sample 4 (the 1 g of rubber granules and artificial turf fibers. artificial grass fiber) had none of the PAHs that were detected in any synthetic digestive fluid extracts. Naphthalene was Analysis of Metals detected in both the saliva extract and the gastric fluid extract The analysis of metals in the nitric acid digestion solutions of sample 3. Naphthalene was also detected in both the and biofluid extracts was achieved using the ICP-MS system gastric extract and the intestinal extract of sample 8. We also that was housed in Environmental and Occupational Health detected benzo(a)pyrene and benzo(ghi)perylene in the Sciences Institute. The ICP-MS system was a Thermo- gastric extract of sample 3. elemental X5 model (ThermoFischer Scientific Inc.) and was Among the metals (Cr, As, Cd, and Pb) targeted in the calibrated with standard solutions prepared with a commer- bioaccessibility analysis, As and Cd were detected in none of cial stock solution diluted with 5% nitric acid. Laboratory the three synthetic fluids and none of the three samples (3, 4, (solvent and instrument) blanks were subtracted from sample and 8); Cr was detected in the saliva extract of sample 4 and concentrations. More details of the method can be found in a the gastric extract of sample 8; and Pb was detected in the previous publication (Xie et al., 2007). gastric extracts of all the three samples. In addition, Pb was detected in the intestinal extract of sample 4. Table 6 summarizes the fractions (%) of bioaccessible Results metals and PAHs that were detected in any of the simulated ingestion extracts. The most consistent results were Pb in the Using the Soxhlet extraction with a non-polar organic gastric fluid, ranging from 24.7% to 44.3%. Pb was not solvent (dichloromethane) for B16 h at 601C, we determined detected in the intestinal extracts of the two rubber granules concentrations of PAHs in the eight samples. These samples but had a high bioaccessible fraction (54.0%) in the

Journal of Exposure Science and Environmental Epidemiology (2008) 18(6) 603 Zhang et al. Hazardous chemicals and digestive ﬂuids

Ta bl e 4 . Concentrations (mg/kg) of PAHs in the samples of rubber granules and synthetic grass ﬁber.

PAHs Sample no.

1234a 5678

Naphthalene ND 0.10 0.40 0.20 0.03 0.03 ND 0.86 Acenaphthene ND 0.03 ND ND 0.16 0.09 ND ND Fluorene 0.76 0.77 ND ND 0.50 0.45 ND ND Phenanthrene 0.06 4.35 ND ND 2.19 1.46 ND ND Anthracene 0.03 0.17 ND 0.01 0.03 0.03 ND ND Fluoranthene 0.11 0.37 ND ND 5.08 3.54 25.4 ND Pyrene 3.73 8.76 ND ND 6.24 9.61 2.45 13.5 Benzo(a)anthracene 1.23 1.26 0.31 ND 0.29 0.98 0.06 ND Chrysene 1.32 7.55 ND ND 1.96 1.34 0.06 4.90 Benzo(b)ﬂuoranthene 3.39 2.19 ND ND 1.08 0.58 0.20 0.43 Benzo(k)ﬂuoranthene 7.29 1.78 0.17 ND 0.14 0.18 0.10 0.99 Benzo(a)pyrene 8.58 3.56 0.78 0.08 0.61 0.25 0.06 0.41 Dibenzo(a,h)anthracene 3.52 1.55 ND ND 0.71 0.52 1.43 ND Benzo(ghi)perylene 7.75 2.61 2.73 0.11 0.85 0.46 2.03 ND Indeno(1,2,3-cd)pyrene 0.40 0.37 ND ND ND ND ND ND

Total (sum of all the detected compounds) 38.15 35.4 4.40 0.40 19.9 19.5 31.8 21.1

Abbreviations: ND, non-detectable; PAH, polycyclic aromatic hydrocarbon. aArtiﬁcial grass ﬁber sample.

Ta bl e 5 . Concentrations (mg/kg) of metals in the samples of rubber observation is as expected because PAHs are semivolatile granules and synthetic grass fiber. compounds and are prone to photo- and thermal-degradation. Sample 7, collected in a different 5-month-old synthetic Metals Sample no. turf field, had a total PAH content comparable to the total 1234a 8 PAH contents of samples 1 and 2 (from a 2- to 3-month-old field). However, the total PAH content of sample 7 was Cr 0.87 1.68 0.69 3.93 0.93 mainly driven by one compound (fluoranthene). Zn 5710 9988 NA NA NA Samples 5 and 6 can be considered as ‘‘duplicate’’ samples, As 3.55 1.57 ND 0.28 0.28 Cd 0.21 0.41 0.37 ND 0.22 as they were collected in the same field at the same time. The Pb 5.76 53.5 4.63 2.80 3.12 total PAH content was very similar between the two samples; and individual PAHs were also generally consistent (Table 4). Abbreviations: NA, not analyzed for; ND, non-detectable. aArtificial grass fiber sample. These two samples were collected 5 years and 4 months after the field installation. However, their PAHs contents were still high (at B20 p.p.m. as compared with 4.4 p.p.m. for sample intestinal extract of the grass fiber sample. Except for 3 that was about 2 years after the field installation). We have naphthalene, bioaccessibility of PAHs in any of the synthetic learned that rubber granule infill needs to be refilled on a digestive fluids was very low (mostly below detection limits or periodic basis to compensate the loss due to degradation, o3%). run-off, and tracking-away by people. It is likely that a new batch of rubber granules had been applied to the field before the sample collection. Unfortunately, we were not able to get Discussion the information on refilling. A high total PAH content (21.1 p.p.m.) was also observed in sample 8 that was PAHs collected on a 7-year-old field. However, one PAH (pyrene) Samples 1, 2, and 3 were collected from the same synthetic contributed 64% of the total PAH in this sample. turf field 2 months, 3 months, and almost 2 years after the PAHs are formed through incomplete combustion of fuels field was installed, respectively. PAHs contents of rubber and materials and especially during the pyrolysis process. granules collected from this field appeared to decline as the Rubber granules used as infill materials in synthetic turf fields field aged (Table 4). The sampling interval between samples 1 are typically made from disposed tires. Tire making process and 2 was only 1 month, and we saw a slight decrease in total involves pyrolysis; and thus it is not surprising to find PAHs PAHs (from 38.15 to 35.4 p.p.m.). In contrast, we saw a in rubber granules. In contrast, artificial grass fibers are substantial decrease in sample 3 (only 4.40 p.p.m.). This plastic materials (e.g., polyethylene, nylon, and combina-

604 Journal of Exposure Science and Environmental Epidemiology (2008) 18(6) Hazardous chemicals and digestive ﬂuids Zhang et al.

Ta bl e 6 . Bioaccessible fractionsa (%) of metals and PAHs following a simulated ingestion by the synthetic saliva, the synthetic gastric ﬂuid, and the synthetic intestinal ﬂuid.

Metals/PAHs Sample no. 3 Sample no. 4b Sample no. 8

SalivaGastricSalivaGastricIntestinalGastricIntestinal

Cr 0 0 1.2 0 0 23.3 0 Pb 0 44.2 0 34.6 54.0 24.7 0 Naphthalene 45.7 6.38 0 0 0 43.3 50.9 Benzo(a)pyrene 0 2.95 0 0000 Benzo(ghi)perylene 0 1.22 0 0000

Abbreviation: PAH, polycyclic aromatic hydrocarbon. aThe percentage was determined by dividing the concentration in the synthetic biofluid by the total content (i.e., the concentration in the nitric aciddigestion extracts for metals or the concentration in the Soxhlet dichloromethane extracts for PAHs. bArtificial grass fiber. tions) that do not undergo the pyrolysis process; their PAHs the PAHs sampled in various studies, chrysene appears most contents are expected to be low. Our sole sample of artificial consistently (Plesser and Lund, 2004; OEHHA, 2007). (polyethylene) grass fiber indeed had non-detectable levels of All the PAHs that we found at or above the DEC safety almost all the 15 target PAHs (Table 4). When discarded tires limits are known, probable, or possible human carcinogens, were crushed into small granules and then applied to the as defined by International Agency for Research on Cancer synthetic turf fields, PAHs originally contained in the bulk (IARC, 2006). However, our bioaccessibility simulation materials (tires) had a large potential to be released out due study suggests that the absorption of these PAHs through to increased surface areas, volatilization (especially on hot ingestion is unlikely. This finding will undoubtedly provide a days), and photo-degradation. Our results suggest that the degree of comfort to health officials and the public. To health age of synthetic field, or more precisely, how long rubber scientists, this finding would provoke little surprise because granules had been applied, is important in determining PAH PAHs are non-polar organic compounds that have very concentrations in rubber granules. It is reasonable to expect limited solubility in water-based digestive fluids. The results that many other factors, such as rubber infill/synthetic turf of our digestive tract simulations will therefore confirm many brand, weather conditions, and turf field surface conditions, investigators’ expectations, and will direct them to other will affect PAH concentrations. However, a larger-scale and exposure routes. systematic study design is needed to examine within- and However, the fundamental limitation of our simulation between-brand variabilities and major factors that affect method should be recognized. The human digestive tract is PAH concentrations. not as simple as a glass vessel containing synthetic biofluids. In the absence of a better set of health-based standards as Its surfaces contain lipids that can enhance the absorption of reference, we compared PAH concentrations in the rubber lipophilic PAHs. Once ingested, PAHs in rubber granules granule samples to the PAH concentration levels that the would interact with foods, which may increase PAHs New York State Department of Environmental Conserva- bioavailability. tion (DEC) considers sufficiently hazardous to require the Historically, significant exposures to PAHs via dermal removal from contaminated soil sites (DEC, 2006). Chrysene contact were reported in workers handling used engine oils was found to be above the DEC residential contaminated soil (Moen et al., 1996; Nilsson et al., 2004) and chimney sweepers limit of 1.0 p.p.m. in five of the seven rubber granule samples (Boffetta et al., 1997; Bostrom et al., 2002; Armstrong et al., (1, 2, 5, 6, and 8). Dibenzo(a,h)anthracene was also found 2004), because both lipophilic and hydrophilic compounds can above the DEC limit of 0.33 p.p.m. in five of the seven be absorbed through the human skin. Considering children samples (1, 2, 5, 6, and 7). Benzo(b)fluoranthene and and athletes have frequent skin contact with the surface and benzo(k)fluoranthene were each at or above the DEC limit of rubber infill of synthetic turf field, exposure through dermal 1.0 p.p.m. in three samples (1, 2, and 5 for the former contact cannot be ignored until further investigation of this chemical; 1, 2, and 8 for the latter). Benzo(a)anthracene and exposure route shows it insignificant. benzo(a)pyrene were found to be above their corresponding PAHs are semivolatile compounds (boiling points from DEC limits (1.0 p.p.m. for each) in samples 1 and 2. Our B2401CtoB4001C). PAHs contained in solid rubber findings with respect to the PAHs that appear above or at granules are expected to evaporate into the atmosphere DEC safety levels are fairly consistent with findings of Plesser especially when ambient temperature is high. Out results and Lund (2004) in Norway, who also conducted a small suggested a relatively rapid decay in PAHs contents, survey of PAHs in rubber synthetic turf granules. Among all suggesting a potential for inhalation exposure to occur when

Journal of Exposure Science and Environmental Epidemiology (2008) 18(6) 605 Zhang et al. Hazardous chemicals and digestive fluids children and athletes are close to the turf surface and In this study, we found that Pb in rubber granules especially when the respiration rate is high during heavy (samples 3 and 8) was only bioaccessible in the synthetic exercise activities. gastric fluid (44.2% bioaccessible for sample 3 and 24.7% for sample 8. While these fractions were substantial, they Metals were not as high as the 52.4 to 77.2% gastric bioaccessibility Because Zn concentrations in rubber granules have been Pb values that (Yu et al., 2006) found in household dust. rather thoroughly studied, we only analyzed two samples (1 Although the gastric bioaccessible fraction of Pb in the and 2) for zinc (Zn) and found Zn levels exceeded the DEC artificial grass fiber sample (34.6%) was lower than in residential soil standard of 2200 p.p.m. Our results for Zn household dust, the intestinal bioaccessible fraction (54.0%) contents are consistent with other findings (Plesser and Lund, was higher than the values reported by Yu et al. for 2004; OEHHA, 2007). High Zn concentrations result from household dust (4.9–32.1%). the manufacture of tire rubber, from which the synthetic turf granules derive. Zn is added to tires, at about 1.5%, to strengthen the rubber (Snyder, 1998). Conclusion Runoff with high Zn from synthetic turf fields may Many synthetic turf fields consist of not only artificial grass produce adverse effects to plants and aquatic life (Tucker, but also rubber granules that are used as infill. On the basis of 1997). This is of particular concern given that the leaching results from seven samples of rubber granules collected in rate of Zn from rubber granules can be up to 20 times greater three fields at different ages of the fields, we reach the than the leaching rate of Zn from agricultural applications of following main conclusions: (1) Rubber granules often, manure and pesticides (Verschoor, 2007). especially when the synthetic turf fields were newer, contained Unlike PAHs, three metals (Cr, Cd, and Pb) did not show PAHs at levels above health-based soil standards. PAH levels a clear decay with age of the turf field, based on the results generally appear to decline as the field ages. However, the from samples 1, 2, and 3. In contrast, As appeared to decline. decay trend may be complicated by adding new rubber Nevertheless, As and Cd levels in all the tested samples were granules to compensate for the loss of the material. (2) PAHs below the DEC’s residential standards of 16 p.p.m. for As contained in rubber granules had low bioaccessibility (i.e., and 2.5 for Cd and may reflect ‘‘contamination’’ from hardly dissolved) in synthetic digestive fluids including salvia, underneath and nearby soils. Concentrations of Pb in the gastric fluid, and intestinal fluid. (3) The zinc contents were rubber granule samples 1, 3, and 8 were also low (5.76, 4.63, found to far exceed the soil limit. (4) Lead contents were low and 3.12 p.p.m., respectively); sample 2 had a Pb concentra- ( 53 p.p.m.) in all the samples in reference to soil standards. tion of 53.5 p.p.m., which is below the DEC standard of o However, the lead in the rubber granules was highly 400 p.p.m. for residential use but close to 63 p.p.m. for bioaccessible in the synthetic gastric fluid. The analysis unrestricted use. The range in Pb concentrations we observed of one artificial grass fiber sample showed a slightly is consistent with that of other studies (Plesser and Lund, worrisome chromium content (3.93 p.p.m.) and high bio- 2004). A recent report showed unacceptable Pb concentra- accessible fractions of lead in both the synthetic gastric and tions (in thousands of p.p.m.) in ‘‘old version’’ AstroTurf intestinal fluids. artificial grass fibers (NJDHSS, 2008). However, Pb concentration in the only ‘‘new generation’’ fiber sample that we tested (sample 4) was low (2.8 p.p.m.). To keep these Acknowledgements values in perspective, however, we should note that some health scientists believe that any Pb is harmful to children’s We thank Dr. Brian Buckley and Ms. Elizabeth McCandlish neurocognitive development, and that no new Pb should be for their assistance in the analysis of metals and Drs. Paul added to their surroundings (Canfield et al., 2003). Lioy and Thomas Wainman for their inspiring discussions on Our artificial grass fiber sample had Cr level at 3.93 p.p.m. public concerns related to the use of artificial turf. The study This is total concentration of Cr at all oxidation states is funded in part by a gift from William and Ellen Crain and (valences). The analytical method we used could not a gift from Mr. Patrick Barnard through the Foundation of differentiate hexavalent and trivalent Cr. If the hexavalent UMDNJ. Dr. Zhang is in part supported by an NIEHS form dominated the total Cr concentration, there would be a Center grant to Environmental and Occupational Health greater risk, as hexavalent Cr is substantially more toxic than Sciences Institute (P30 ES05022). trivalent Cr. (The NYS DEC soil limits for hexavalent and trivalent Cr are 22 and 36 p.p.m., respectively and are 1 and 30 p.p.m., respectively, for unrestricted use.) The concentra- References tions of Cr in all the rubber granule samples were lower than Armstrong B., Hutchinson E., Unwin J., and Fletcher T. Lung cancer risk after in the fiber sample, suggesting a possible source of Cr in exposure to polycyclic aromatic hydrocarbons: a review and meta-analysis. artificial grass fibers, perhaps chromium-containing dyes. Environ Health Perspect 2004: 112(9): 970–978.

606 Journal of Exposure Science and Environmental Epidemiology (2008) 18(6) Hazardous chemicals and digestive ﬂuids Zhang et al.

Birkholz D.A., Belton K.L., and Guldotti T.L. Toxicological evaluation of the Moen B.E., Nilsson R., Nordlinder R., Ovrebo S., Bleie K., Skorve A.H., and hazard assessment of tire crumb for use in public playgrounds. JAirWaste Hollund B.E. Assessment of exposure to polycyclic aromatic hydrocarbons in Manage Assoc 2003: 53: 903–907. engine rooms by measurement of urinary 1-hydroxypyrene. Occup Environ Boffetta P., Jourenkova N., and Gustavsson P. Cancer risk from occupational and Med 1996: 53(10): 692–696. environmental exposure to polycyclic aromatic hydrocarbons. Cancer Causes NBC. NJ Agency Releases Toxic Turf Report: Parents Should be Concerned.June Control 1997: 8(3): 444–472. 2008. Available at http://www.nbc10.com/health/16477245/detail.html Bostrom C.E., Gerde P., Hanberg A., Jernstrom B., Johansson C., Kyrklund T., accessed on 18 July 2008. et al. Cancer risk assessment, indicators, and guidelines for polycyclic aromatic Nilsson R., Nordlinder R., Moen B.E., Ovrebo S., Bleie K., Skorve A.H., hydrocarbons in the ambient air. Environ Health Perspect 2002: 110(Suppl 3): Hollund B.E., and Tagesson C. Increased urinary excretion of 8-hydro- 451–488. xydeoxyguanosine in engine room personnel exposed to polycyclic aromatic Brown D. Exposures to Recycled Tire Rubber Crumbs used on Synthetic Turf hydrocarbons. Occup Environ Med 2004: 61(8): 692–696. Fields, Playgrounds and as Gardening Mulch 2007. Available at NJDHSS (State of New Jersey Department of Health and Senior Services). DHSS http://www.ehhi.org/reports/turf/turf_report07.pdf accessed on 18 July 2008. Calls for Federal Action on Potential Lead Hazard Posed by Artificial Turf Canfield R.L., Henderson C.R., Cory-Slechta D.A., Cox C., Jusko T.A., April 2008. Available at http://www.state.nj.us/cgi-bin/dhss/njnewsline/view_ and Lanphear B.P. Intellectual impairment in children with blood lead article.pl?id ¼ 3174 accessed on 18 July 2008. concentrations below 120 micrograms per deciliter. NEnglJMed2003: Office of Environmental Health Hazard Assessment (OEHHA). Evaluation of 348(16): 1417–1526. Health Effects of Recycled Waste Tires in Playground and Track Products 2007. Centers for Disease Control and Prevention (CDC). Potential Exposure to Lead in Available at http://www.ciwmb.ca.gov/Publications/Tires/62206013.pdf Artificial Turf: Public Health Issues, Actions, and Recommendations.Available accessed on 18 July 2008. at http://www2a.cdc.gov/HAN/ArchiveSys/ViewMsgV.asp?AlertNum ¼ 00275 Plesser T.S.W., and Lund O.J. Potential health and environmental effect linked to accessed on 18 July 2008. artificial turf systemsFfinal report. NorwegianBuildingResearchInstitute Claudio L. Synthetic turf: health debate takes root. Environ Health Perspect 2008: Report to the Norwegian Football Association 2004. Available at http://www./ 116(3): A116–A122. sss.de/conferences/Dresden%202006/Technial/NBI%20Engelsk.pdf accessed Department of Environmental Conservation of New York State (DEC). 6 on 18 July 2008. NYCRR Part 375, Environmental Remediation Program,DraftRevised14 Smith D.J.T., and Harrison R.M. Concentrations, trends and vehicle source June 2006, Table 375-6.8 (a) and (b). Available at http://www.dec.ny.gov/ profile of polynuclear aromatic hydrocarbons in the UK atmosphere. Atmos regs/15507.html accessed on 17 July 2008. Environ 1996: 30(14): 2513–2525. Ellickson K.M., Schopfer C.J., and Lioy P.J. The bioaccessibility of low level Snyder R.H. Scrap Tires: Disposal and Reuse: Society of Automotive Engineers. radionuclides from two Savannah River Site soils. Health Phys 2002: 83: Warrendale, PA, 1998. 476–484. Tucker M.R. Ground Rubber: Potential Toxicity for Plants.NorthCarolina Han I.K., Duan X., Zhang L., Yang H., Rhoads G.G., WeiF., et al. Department of Agriculture & Consumer Services, Media Notes for North 1-Hydroxypyrene concentrations in first morning voids and 24-h composite Carolina Growers, April 1997. Available at http://www.ncagr.com/agronomi/ urine: intra- and inter-individual comparisons. J Expo Sci Environ Epidemiol pdffiles/rubber.pdf accessed on 14 December 2007. 2007: doi:10.1038/sj.jes.7500639. U.S. Consumer Product Safety Commission (CPSC). CPSC finds synthetic turf Howard J.W., Fazio T., White R.H., and Klimech B.A. Extraction and fieldsoktoinstall,oktoplayon. Washington D.C. July 30, 2008. estimation of polycyclic aromatic hydrocarbons in total diet. AOAC 1986: Verschoor A. Leaching of Zinc from Rubber Infill on Artificial Turf (Football 51: 122–129. Pitches). RIVM report 601774001/2007. Bilthoven, Netherlands, 2007. International Agency for Research on Cancer (IARC). Monographs on the Xie R., Johnson W., Rodriguez L., Gounder M., Hall G., and Buckley B. A study Evaluation of Carcinogenic Risk to Humans, PAHs, Vol. 92. 2006 Available at of the interactions between carboplatin and blood plasma proteins using size http://monographs.iarc.fr/ENG/Meetings/92-pahs.pdf accessed on 18 July exclusion chromatography coupled to Inductively Coupled Plasma Mass 2008. Spectrometry, Anal and Bioana. Chem 2007: 387: 2815–2822. McCarthy M., and Berkowitz S. Artificial turf: health hazard. USA Today.May8 Yu C.H., Yiin L.M., and Lioy P.J. The bioaccessibility of lead (Pb) from 2008. vacuumed house dust on carpets in urban residences. Risk Anal 2006: 26: 125–134.

Journal of Exposure Science and Environmental Epidemiology (2008) 18(6) 607