Environ. Sci. Technol. 2009, 43, 941–947
Respiratory and Dermal Exposure to Organophosphorus Flame Retardants and Tetrabromobisphenol A at Five Work Environments ¨ K I N E N , * ,† MAIJA S. E. MA ¨ KINEN,‡ MILJA R. A. MA J A A N A T . B . K O I S T I N E N , †,§ A N N A - L I I S A P A S A N E N , ‡,| PERTTI O. PASANEN,† PENTTI J. KALLIOKOSKI,† AND ANNE M. KORPI† Department of Environmental Science, University of Kuopio, P.O. Box 1627, 70211 Kuopio, Finland
Received September 12, 2008. Revised manuscript received December 2, 2008. Accepted December 2, 2008.
Organophosphorus compounds (OPs) and tetrabromobisphenol A (TBBPA) are widely utilized as flame retardants (FRs) in plastics, textiles, rubbers, and building materials. Eight OPs and TBBPA were quantified by GC/MS from air samples collected from a furniture workshop, a circuit board factory, two electronics dismantling facilities, a computer classroom, and offices and social premises. In addition, dermal exposure was assessed with patch and hand wash samples at some workplaces. Triphenyl phosphate, tris(2-chloroethyl) phosphate, and tris(2-chloroisopropyl) phosphate were typical contaminants of the workplaces, whereas TBBPA, tricresyl phosphate, trin-butyl phosphate, and tris(2-ethylhexyl) phosphate were rather site-specific. The highest geometric mean of totalFRs in the air samples was measured in personal samples at the electronics dismantling facilities (2.9 and 3.8 µg/m3), whereas the stationary sample results from the other environments ranged between 90 and 720 ng/m3. Stationary samplings underestimated the personal exposure at three out of four work places where comparisons were made. Dermal exposure was shown for the first time at these occupational settings. The geometric mean of totalFR levels in patch samples ranged between 1.5 and 24 ng/cm2 and in hand wash samples between 3.5 and 34 µg/ two hands. The health effects of the measured FR levels remain unknown.
Introduction Flame retardants (FRs), such as organophosphorus compounds (OPs) and tetrabromobisphenol A (TBBPA) are widely used to improve the fire safety of different products. In 2004, nearly 300 000 tons of OPs were used worldwide (1) in a large variety of materials such as PVC plastics, polyurethane foams, paints and lacquers, wall papers, textiles, and electronic * Corresponding author phone: +358 40 355 2893; fax: +358 17 163 191; e-mail:
[email protected]. † University of Kuopio. ‡ Finnish Institute of Occupational Health, Kuopio, Finland. § Current address: University of Helsinki, Tva¨rminne Zoological Station, Hanko, Finland. | Current address: Finnish Institute of Occupational Health, Helsinki, Finland. 10.1021/es802593t CCC: $40.75
Published on Web 01/09/2009
2009 American Chemical Society
equipment. Many of these compounds are also used as plasticizers and lubricants (2-7). TBBPA is the most widely used brominated flame retardant, the global TBBPA market being in the range of 170 000 tons in 2004 (8). FRs can be released from the products into the indoor air when products are used (7, 9) or being recycled (10-12). Consequently, people can be exposed to these compounds with potentially harmful health effects (2-6) via inhalation and ingestion of dust and via dermal contact. A study in Tokyo showed that OPs are more predominant in indoor environments (homes and offices) than brominated compounds (7). Exposure is likely to occur in occupational settings as well. In fact, exposure to OPs and TBBPA in various indoor environments, including occupational settings, has been investigated by air monitoring studies that have occurred in Sweden, Norway, Japan, and Switzerland (10-23). However, OP and TBBPA measurements have not been conducted at Finnish workplaces thus far even though occupational exposure limits for tributyl phosphate (TBP), triphenyl phosphate (TPP), and the ortho-isomer of tricresyl phosphate (TCP) have been established for 8 h of exposure in Finland, those limits being 5, 3, and 0.1 mg/m3, respectively (24). The EU’s risk assessment draft proposes the following occupational guidelines for tris(2-chloroethyl) phosphate (TCEP): 0.2 mg/m3 for inhalation exposure and 2 mg/person/day for dermal exposure (25). Exposure to FRs has previously been assessed with stationary air and dust samples, whereas personal exposure has only seldom been evaluated, and data on dermal exposure to OPs and TBBPA are lacking. Dermal exposure may be especially relevant in occupational settings, since at least four OPs, that is, TBP, tris(2-butoxyethyl) phosphate (TBEP), TCP, and tris(1,3-dichloro-2-propyl) phosphate (TDCPP), can be absorbed through the skin (2, 4, 6, 26). In addition, a few case studies suggest that TPP may cause skin sensitization in humans (27). There are several methods for assessing dermal exposure, of which many originate from pesticide studies. Direct dermal exposure can be measured for instance with patch samples or with hand wash samples (28). Stapleton et al. (29) assessed the exposure to brominated diphenyl ethers (PBDE) from hand to mouth transfer with hand wipe samples. Contamination of hands may also lead to gastrointestinal exposure, and poor hygiene may contribute to it. The aim of this study was to estimate the occupational exposure to eight OPs and TBBPA via air samplings at four different occupational settings where FR-containing materials were used or processed and at two types of nonindustrial occupational settings. In addition, dermal exposure to FRs was assessed by two methods. The sampling sites were chosen to represent both the beginning and the end of the products’ life cycle.
Experimental Section Chemicals. The reference substances (eight OP compounds, one brominated compound, and three analytical standards) are listed in Table 1. Dichloromethane was purchased from Labscan, methanol from J. T. Baker, toluene and sodium sulfate from Riedel de Hae¨n, n-hexane from Rathburn Ltd., ethanol from Altia Oyj, and nonane from Fluka. Diazomethane was prepared by a laboratory technician (30). Cleaning Procedures of Samplers and Devices before Sampling and Extraction. Since OP compounds have been found as ubiquitous indoor pollutants, extensive cleaning procedures were applied. Before sampling, the glass fiber filters (Ø 25 mm, Gelman Sciences) and alfacellulose paper VOL. 43, NO. 3, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
941
TABLE 1. Abbreviations, CAS Numbers, and Manufacturers of Reference Compounds compound (purity, when available)
abbreviation
CAS number
manufacturera
tripropyl phosphate (internal standard) (99%) tri-n-butyl phosphate triphenyl phosphate (>99%) tris(2-butoxyethyl) phosphate (94%) tricresyl phosphate (90%), a mixture of three isomers tris(2-ethylhexyl) phosphate (97%) tris(2-chloroethyl) phosphate tris(2-chloroisopropyl) phosphate, a mixture of three isomers tris(1,3-dichloro-2-propyl) phosphate 2,4,6-tribromophenol (internal standard) tetrabromobisphenol A phenanthrene-d10 (recovery standard) (98%)
TPrP TBP TPP TBEP TCP TEHP TCEP TCPP TDCPP TBrP TBBPA PHEN
513-08-6 126-73-8 115-86-6 78-51-3 1330-78-5 78-42-2 115-96-8 13674-84-5 13674-87-8 118-79-6 79-94-7 1517-22-2
SA1 SA2 SA2 SA1 SA1 SA2 SA2 E E E SA1 SA
Abbreviations: SA1 ) Sigma Aldrich (St. Louis, MO); SA2 ) Sigma Aldrich (Steinheim, Germany); E ) Ehrenstorfer (Augsburg, Germany). a
(technical filter paper 0860, 74 g/m2, Schleicher & Schu ¨ ll or grade 211, Whatman, Inc.) used in patch samples, and filter paper (40 Ashless, Ø 70 mm, Whatman International Ltd.) used in filtration of hand wash samples (see below) were ultrasonicated for 15 min in dichloromethane and dried in a nitrogen flow. Thereafter, these materials were stored wrapped in aluminum foil. OVS samplers (see below) were not precleaned. Extraction thimbles (25 × 80 mm, Schleicher & Schu ¨ ll) and cotton were cleaned with a Soxhlet apparatus for >16 h before the extraction of samples. The solvent (toluene) was replaced once during this precleaning extraction. All the glassware were heated to 450 °C for 4 h before use. Metal instruments (e.g., tweezers and spoons) were rinsed with methanol before use. Sampling and Extraction. The levels of FRs were studied in four industrial workplaces, in one computer classroom, in three offices, and in one coffee room (social premises). One of the industrial workplaces was a circuit board (CB) factory, where the work processes during the sampling included the gold-plating, printing, machine tooling, surface treatment, and inspection of CBs, and working in clean room. Another industrial workplace was a furniture workshop, where the studied work phases were upholstering, sewing, seaming, and bundling the fabrics. At the two electronics dismantling facilities, workers dismantled and sorted the electronics and controlled the crushing process of the electronics. Sampling sites are depicted in Table S1 in the Supporting Information. Air Samples. Two different samplers were used: a glass fiber filter placed in an IOM sampler (SKC Ltd.) and an OVS sampler (SKC Ltd.). The IOM sampler collects the inhalable fraction of particles. The OVS sampler contains a filter to collect particles, XAD resin, and polyurethane foam for the sampling of compounds present in the gas phase. In the analysis, the gas and particle phases were not separated. Altogether, 17 and 27 air samples were taken by the IOM sampler and OVS sampler, respectively. Some of the samples were collected simultaneously with both samplers. Batteryoperated personal sampler pumps (SKC 224, SKC Ltd.) were used. The sampling rates were 0.98-1.05 LPM for the OVS sampler and 1.95-2.08 LPM for the IOM sampler. Sampling equipment was calibrated before and after every sampling. The sampling time varied between 119 and 663 min. Accordingly, the total air volumes collected ranged from 0.20 to 1.31 m3. The samples were transported refrigerated (4 °C) to the laboratory and stored for no more than 15 days in a refrigerator before extraction. For extraction, the samples were treated with internal standards (absolute amounts of TPrP 248 ng and TBrP 583 ng) and Soxhlet extracted for 16-22 h with toluene (V ) 150 mL). To prevent the decomposition of TBBPA, which is 942
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 3, 2009
photolytically decomposed when exposed to UV light (31), the samplers, glassware, and Soxhlet extractor were covered with aluminum foil and the lighting in the laboratory was kept as low as possible. After extraction, nonane (ca. 300 µL) was added to the samples to prevent sample losses during the evaporation of extra solvent with rotavapor. The solvent was exchanged to hexane (V ) 10 mL), and extra solvent was thereafter evaporated in a gentle nitrogen flow. To detect TBBPA, the samples were derivatized with diazomethane. The recovery standard (phenanthrene-d10 (PHEN), 99 ng) was added to the total sample volume of 250 µL, and the vials were stored in -20 °C until the analysis. Patch Samples. Adsorbent patches attached onto workers’ clothing were used to evaluate potential body exposure (method modified from ref 32). At the electronics dismantling facilities and at the circuit board factory, one patch was attached on the worker’s chest, and two other patches were attached on one arm and one thigh. At the furniture workshop, a single patch on the chest was used (Table S1 in the Supporting Information). The exposed area of each patch was 25 cm2. After the working day (approximately 6 h), the patches were removed and placed in glass tubes with 25 mL of ethanol. With each set of samples, a blank patch sample was analyzed. In the laboratory, the ethanol was decanted into an evaporating flask, 25 mL of fresh ethanol was poured on the patch, and the glass tube was shaken for 15 min. The extracts were then combined, internal standards were added, and the samples were concentrated in a rotavapor. Thereafter, the samples were treated similarly to the air samples. Hand Wash Samples. Hand exposure of the workers to OP compounds and TBBPA was studied with hand wash samples. These samples were taken from two workers at the furniture workshop and in the circuit board factory. The washing was performed according to the procedure described in EN 1499 (33). Ethanol used as a washing solution was poured onto the workers’ hands above a beaker for 30 s, while the workers rubbed their hands in a specific way as described in the standard. The hand wash samples were transferred to a glass bottle. The beaker was washed with 10 mL of ethanol, which was then added to the glass bottle. The cleaning and drying of the hand wash samples was conducted mostly as described in Kangas et al. (32). Briefly, the samples were treated with internal standards (TPrP 248 ng and TBrP 583 ng), evaporated to 2-3 mL with a rotavapor, and transferred to a separating funnel with 25 mL of 2% Na2SO4-water solution. Ten milliliters of dichloromethane was added to the separating funnel. This mixture was shaken for 10 min, after which the dichloromethane layer was filtered into an evaporating flask. The same procedure was repeated first with 10 mL and
then with 5 mL of dichloromethane. Eventually the filter paper was rinsed with 3 × 2 mL dichloromethane. The extracts were combined, concentrated, and treated in a manner similar to the air sample treatment. Analysis. Analyses were performed with a gas chromatograph (GC) connected to a mass spectrometer (MS) operating in EI-mode. The GC/MS (Agilent 6890N GC + Agilent 5973 inert MS) was operated in single ion monitoring mode using pulsed splitless injection, 2 µL, under the following conditions: injection port 275 °C, split 1 min, helium with a flow rate of 1 mL/min as the carrier gas, 30 m × 0.25 mm (i.d.) × 0.15 µm DB dioxin column, and temperature program 50 °C for 2 min, ramp 15 °C/min to 180 °C, ramp 5 °C/min to 270 °C, hold 11.33 min. The following ions were used for quantifying and identifying the analytes: TBP (m/z 99, 155, 211), TCEP (m/z 249, 251), TCPP (m/z 99, 125, 277, 279), TEHP (m/z 99, 113, 211), TBEP (m/z 125, 199, 299), TPP (m/z 325, 326), TDCPP (m/z 209, 211, 381), TCP (m/z 367, 368), and TBBPA (m/z 555, 557, 570, 572). The liner used in GC was treated with silyl reagent (Sylon CT, Sigma Aldrich) to prevent the analytes from being adsorbed onto the liner’s surface. The liner was changed at the beginning of all sample sequences. Calibration, Quantification, and Quality Control. A stock solution was prepared, containing the eight OPs and TBBPA to be analyzed, with individual concentrations ranging between 0.001 and 14.6 µg/mL. To check the linearity of the GC/MS responses, calibration curves were established for each compound using a dilution series of the stock solution. All the compounds showed linearity, with correlation coefficients ranging from 0.73 to 0.98. The repeatability was tested by analyzing all the samples three times, and for the majority of compounds and tested concentrations the RSD% was between 0.1 and 5%, but occasionally RSD% ranging from 10 to 18.5% was detected. The limit of detection for each compound was defined as the concentration with a GC/MS response distinguishable from the background (signal-tonoise ratio ) 3) and reliable identification. The recoveries were determined by spiking a glass fiber filter with the analytes, whereafter the filter was Soxhlet extracted, analyzed as described above, and quantified with a standard mixture dissolved in hexane. The recoveries of TPrP, TBP, TEHP, TCPP, TBrP, and TBBPA were between 93 and 211%, and for TPP, TBEP, TCP, TCEP, and TDCPP the recoveries were higher, ranging from 318 to 611%. Because of this observed matrix effect, the quantification of the field samples was made with the standard prepared in the sampler, instead of a mere standard solution without sampler, and following extraction procedures. Recoveries of the internal standards were calculated for each sample. Recoveries between 70 and 130% were considered acceptable. Otherwise, quantification was based on the recovery standard. Field blanks were collected at each site, and the FR concentrations therein were subtracted from the samples. Geometric mean (GM) concentrations were calculated for personal air samples, stationary air samples, hand wash samples, and patch samples. In the text, it is referred to GM when the mean is indicated. If the concentration was below the detection limit or the compound was not detected, the concentration was set at the detection limit divided by two. Instead of expressing the GM for FR concentrations with more than 50% of the values below the detection limit, only the highest detection limit was indicated. The sums of eight OPs (TotalOP) and eight OPs and TBBPA (TotalFR) were calculated from individual results.
Results and Discussion To the best of our knowledge, this study may be the first systematic approach estimating occupational exposure to FR from the personal exposure standpoint. In most studies, OP and TBBPA levels were reported in indoor environments from samples collected at stationary sites. Thus far, only Sanchez et al. (21) collected two personal samples during half a working day in a laboratory, and Solbu et al. (23) measured exposure of aviation mechanics during regular aircraft maintenance (n ) 2). This is also the first study to report on dermal exposure to OPs and TBBPA. Air Samples. FRs were detected in all the air samples. The mean FR levels varied typically from a few nanograms per cubic meter up to hundreds of nanograms per cubic meter, whereas the individual FR concentrations measured at the same workplace ranged from less than the detection limit up to 14 600 ng/m3, as was the case for TBBPA (Table 2). In general, the highest mean FR levels were detected at the electronics dismantling facilities, except for TBP, for which the highest mean concentration was measured at the furniture workshop. Results from IOM and OVS samplers are combined in Table 2. This is because there was no clear trend between FR concentrations of the parallel samples collected with different samplers, even though the collection efficiencies for gaseous and particulate forms of the compounds are assumed to be different between the two samplers. However, the individual results are presented in the Supporting Information (Tables S2-S7) specified by sampler type and work description. Because of the low number of samples per work task, no comparisons of the exposure at a given workplace were made. Generally, the mean FR levels were higher in personal samples than in stationary samples (Figure S1 in the Supporting Information). However, the reverse was found for TBP and TCEP at the furniture workshop [B] and TBP at the other electronics dismantling facility [D] (Table 2). The mean sum of FR/OP was also higher in personal air samples at other work places except for site [B]. In addition, at the electronics dismantling facilities, the mean totalOP levels in the personal air samples exceeded those at the circuit board factory or the furniture workshop by more than eight times and totalFR levels exceeded those at the circuit board factory or the furniture workshop by more than 15 times (Figure S1 in the Supporting Information). The individual FRs contributing to this result were especially TPP and TBBPA, but also TCP and TCEP. Regarding the computer classroom where no personal samples were collected, the mean stationary sample TCEP and TCPP levels were greater than in the other environments. The profiles and ranges of the FRs were different between the sampling sites. TPP, TCEP, and TCPP were universal contaminants of the work air and were detected at all the workplaces, whereas TBBPA, TCP, TBP, and TEHP were rather site-specific FRs. TPP, TCEP, and TCPP were present in more than 75% of the air samples collected at sites [C] and [E]. In addition, TPP was detected in more than 50% of the air samples collected at sites [A] and [D], and TCEP at [B] and [D], and TCPP at [A] and [F] (Table 2). These compounds were also occasionally present at the remaining work environments. In addition, TCP and TBBPA were typical compounds at the electronics dismantling facilities, and TBP was typical at sites [A] and [B]. TEHP was persistent only in the computer classroom [E]. The compounds that were very seldom detected (i.e., in 0-50% of the air samples at each site) were TBEP and TDCPP. As expected, the highest number of compounds and the highest concentrations were measured at the electronics dismantling facilities, especially for TPP and TBBPA. Similarly, the highest individual OP level measured at electronics VOL. 43, NO. 3, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
943
TABLE 2. FR Concentrations and Percentage of Samples Containing Detectable Amounts of FRs at Circuit Board Factory, Furniture Workshop, Electronics Dismantling Facilities, Computer Classroom, Offices, and Social Premises geometric mean (range) of flame retardant concentrationsb [frequency of detects, %] compound TBP
TPP
TBEP
TCP (sum of three isomers)
TEHP
TCEP
TCPP (sum of three isomers)
TDCPP
TBBPA
totalOPd
totalFRd
total number of samples
sampling sitea
personal air samples, ng/m3
stationary air samples, ng/m3
patch samples, ng/cm2
hand wash samplesc, ng/hands
A B C D E F A B C D E F A B C D E F A B C D E F A B C D E F A B C D E F A B C D E F A B C D E F A B C D E F A B C D E F A B C D E F A B C D E F
10 (
