High Levels of Organophosphate Flame Retardants in the Great Lakes

Sep 19, 2013 - We sampled at five sites in the North American Great Lakes basin every 12 days from March 2012 to December 2012 (inclusive). This is th...
0 downloads 0 Views 549KB Size
Letter pubs.acs.org/journal/estlcu

High Levels of Organophosphate Flame Retardants in the Great Lakes Atmosphere Amina Salamova, Yuning Ma, Marta Venier, and Ronald A. Hites* School of Public and Environmental Affairs, Indiana University, Bloomington, Indiana 47405, United States S Supporting Information *

ABSTRACT: Levels of 12 organophosphate flame retardants (OPs) were measured in particle phase samples collected at five sites in the North American Great Lakes basin from March 2012 to December 2012 (inclusive). The target compounds were three chlorinated OPs [tris(2-chloroethyl) phosphate (TCEP), tris(1-chloro-2-propyl) phosphate (TCPP), and tris(1,3-dichloro-2-propyl) phosphate (TDCPP)], three alkyl phosphates [tri-n-butyl phosphate (TnBP), tris(butoxyethyl) phosphate (TBEP), and tris(2-ethylhexyl) phosphate (TEHP)], and six aryl phosphates [triphenyl phosphate (TPP), tri-o-tolyl phosphate (TOTP), tri-p-tolyl phosphate (TPTP), tris(3,5-dimethylphenyl) phosphate (TDMPP), tris(2-isopropylphenyl) phosphate (TIPPP), and tris(4-butylphenyl) phosphate (TBPP)]. Total OP (ΣOP) atmospheric concentrations ranged from 120 ± 18 to 2100 ± 400 pg/m3 at the five sites, with the higher ΣOP levels detected at Cleveland and Chicago. ΣOP concentrations at these urban sites were dominated by the chlorinated OPs (TCEP, TCPP, and TDCPP), with the sum of these three compounds comprising 51 ± 6 and 65 ± 12% of ΣOP concentrations at these two sites, respectively. Nonhalogenated OP compounds were major contributors to ΣOP concentrations at the remote sites, with the sum of all nine nonhalogenated OP concentrations comprising 70 ± 21 and 85 ± 13% of the ΣOP concentrations at Eagle Harbor and Sleeping Bear Dunes, respectively. On average, these ΣOP concentrations are about 2−3 orders of magnitude higher than the concentrations of brominated flame retardants in similar samples.



(up to 1950 ng/g),24 Finnish remote air (up to 12 ng/m3),25 and volcanic lakes in central Italy (up to 951 ng/L)26 are also due to long-range atmospheric transport. OPs were also detected in precipitation, suggesting atmospheric deposition of these chemicals onto terrestrial systems.25,27−29 More recently, Möller et al. reported OP concentrations of up to 3 ng/m3 in airborne particles over the Pacific, Indian, Arctic, and Southern Oceans,30 as well as over the North Sea,31 demonstrating the widespread global occurrence of these chemicals. In this study, we present our measurements of several halogenated and nonhalogenated OPs in particle samples collected as part of the Integrated Atmospheric Deposition Network (IADN). We sampled at five sites in the North American Great Lakes basin every 12 days from March 2012 to December 2012 (inclusive). This is the first comprehensive report on atmospheric OP concentrations in the United States, and in the Great Lakes basin specifically, and on the spatial and temporal variations of these concentrations. Only a few previous reports presented OP measurements in the atmosphere and water samples collected on the Canadian side of the Great Lakes.32−34 The findings presented here are important because OPs are high-production volume chemicals and because their level of consumption is expected to increase

INTRODUCTION Organophosphate esters (OPs) are widely used as flame retardants in various consumer and industrial products, such as plastics, electronic equipment, furniture, textiles, and building materials.1 In addition, some of these chemicals, mainly the nonchlorinated alkyl phosphates, are used as plasticizers, as well as antifoaming agents in lacquers, hydraulic fluids, and floor polishes.2 These compounds were first detected in the environment in the 1970s. Tachikawa,3 Meijers and Van der Leer,4 and Sheldon and Hites5 reported several phosphate esters in river water, seawater, and sediments. Subsequent studies in the 1980s focused on the bioaccumulation and biodegradation of these chemicals;6−9 the conclusion was that most of these compounds degrade in the environment,2 and this led to abandoning further such studies. Scientific interest in organophosphate esters re-emerged in late 1990s when Carlsson et al. reported tris(2-chloroethyl) phosphate levels of up to 250 ng/m3 in indoor air.10 Soon measurements confirmed the presence of these compounds throughout the environment. They are present in water,11,12 sediment,13 indoor air and dust,14−16 fish and biota,17−19 and human blood and milk.18,20−22 There have been only a handful of measurements of these chemicals in the ambient atmosphere, but the data that exist suggest that these compounds can undergo long-range atmospheric transport.23−26 Specifically, OPs were detected at concentrations of ∼1 ng/m3 in aerosols from Antarctica in 1994.23 It has been suggested that the elevated levels of OPs in pine needles from the Sierra Nevada Mountains, United States © 2013 American Chemical Society

Received: Revised: Accepted: Published: 8

August 13, 2013 September 11, 2013 September 11, 2013 September 19, 2013 dx.doi.org/10.1021/ez400034n | Environ. Sci. Technol. Lett. 2014, 1, 8−14

9

ΣOP

ΣnonCl-OP

ΣCl-OP

TDMPP TBPP

TIPPP

TPTP

TOTP

TEHP

TBEP

TPP

TDCPP

TCPP

TCEP

12 ± 2.6 7.7 790 ± 110 576 750 ± 79 690 1500 ± 170 1390

a

a

a

a

a

a

a

a

a

ab

a

100

100

100

4 63

33

63

11

100

93

100

74

100

93

100

a

% detected

a

concn

250 ± 53 176 180 ± 25 118 530 ± 80 407 120 ± 46 79 140 ± 22 108 320 ± 36 262 41 ± 6 42 0.93 ± 0.76 0.24 13 ± 5 5.5 4.0 ± 1.1 2.7 17 ± 4 7.9 1400 ± 390 407 750 ± 89 609 2100 ± 400 1306

5.4 ± 1.4 2.9 5.3 ± 0.7 4.8

150 ± 23 125 120 ± 41 104 850 ± 300 322 520 ± 220 106 200 ± 27 181 330 ± 49 227 66 ± 9 57

concn

a

a

a

a

a

a

a

a

a

a

a

a

a

100

100

100

0 73

82

64

0

96

96

100

96

96

68

100

% detected

Cleveland (N = 22)

270 ± 73 140 110 ± 22 86 340 ± 85 207

3.6 ± 0.4 3.6

34 ± 7 32 130 ± 27 152 170 ± 54 72 28 ± 3 28 43 ± 10 34 76 ± 13 77 8.5 ± 2.8 8.1

concn

b

b

b

a

b

b

b

b

b

a

b

100

100

100

0 0

13

0

0

38

44

100

44

81

63

88

% detected

Sturgeon Point (N = 16)

17 ± 5 7.7 100 ± 15 84 120 ± 18 109

42 ± 9 44 67 ± 13 58 4.7 ± 0.7 4.7

34 ± 8 28 11 ± 2 7.7 25 ± 7 27

concn

b

b

c

b

b

b

b

b

b

100

100

100

0 6

0

0

0

13

50

94

0

25

100

75

% detected

Sleeping Bear Dunes (N = 16)

53 ± 25 25 150 ± 45 89 170 ± 52 100

180 ± 60 61 5.5 ± 0.9 5.5 32 ± 9 29 52 ± 19 32 55 ± 13 31 68 ± 15 51 8.7 ± 2.2 8.6

concn

b

b

c

b

b

b

b

b

b

a

100

100

50

0 0

4

0

4

15

31

100

39

15

27

39

% detected

Eagle Harbor (N = 26)

The concentrations are averaged over the period from March 2012 to December 2012 (July 2012 to December 2012 for Sturgeon Point, Sleeping Bear Dunes, and Eagle Harbor) (inclusive). ANOVA results for the logarithmically transformed concentrations are given; the concentrations are not significantly different at the P < 0.05 level for those locations sharing the same letter. N is the sample size for a given site. Blank cells indicate that the compound was not detected in samples from that site.

a

compound

TnBP

Chicago (N = 27)

Table 1. Atmospheric Particle Concentrations [average ± standard error and median (pg/m3)] and Detection Percentages of the OPs Analyzed at the Five IADN Sitesa

Environmental Science & Technology Letters Letter

dx.doi.org/10.1021/ez400034n | Environ. Sci. Technol. Lett. 2014, 1, 8−14

Environmental Science & Technology Letters

Letter

Figure 1. Average atmospheric concentrations of individual OPs (and their standard errors) in the particle phase at the five United States IADN sampling sites around the Great Lakes.

extracted for 24 h with 50% (v/v) acetone in hexane. The extract was reduced in volume by rotary evaporation, and the solvent was exchanged with hexane and fractionated on a column containing 3.5% (w/w) water-deactivated silica gel. The column was eluted with 25 mL of hexane, 25 mL of 50% (v/v) hexane in dichloromethane, and 25 mL of 70% (v/v) acetone in dichloromethane. The OPs eluted in the third fraction. After N2 blow down, the samples were spiked with the quantitation internal standards (d10-anthracene, d12-benz[a]anthracene, and d12-perylene). The samples were analyzed by electron impact gas chromatographic mass spectrometry for three chlorinated OPs [tris(2-chloroethyl) phosphate (TCEP), tris(1-chloro-2-propyl) phosphate (TCPP), and tris(1,3dichloro-2-propyl) phosphate (TDCPP)], three alkyl phosphates [tri-n-butyl phosphate (TnBP), tris(butoxyethyl) phosphate (TBEP), and tris(2-ethylhexyl) phosphate (TEHP)], and six aryl phosphates [triphenyl phosphate (TPP), tri-o-tolyl phosphate (TOTP), tri-p-tolyl phosphate (TPTP), tris(3,5-dimethylphenyl) phosphate (TDMPP), tris(2-isopropylphenyl) phosphate (TIPPP), and tris(4-butylphenyl) phosphate (TBPP)] (see Figure S2 of the Supporting Information for their structures). The details of the instrumental analysis,38 chemicals used in this study, and quality control and quality assurance procedures are provided in the Supporting Information.

due to production and use restrictions placed on brominated flame retardants. For example, the production volumes for tris(1,3-dichloro-2-propyl) phosphate (TDCPP), triphenyl phosphate (TPP), and tris(2-chloro-isopropyl) phosphate (TCPP) in the United States increased from 1−10 million pounds in 1990 to 10−50 million pounds in 2006 [after the socalled pentabrominated diphenyl ether (Penta-BDE) mixture was phased out].35 In Western Europe, the level of consumption of OPs increased by 2.5% from 2001 to 2005 and 7.1% from 2005 to 2006.2 Data on the environmental persistence and toxicity of OPs are limited, but some of these chemicals (especially the halogenated ones) are known to be relatively persistent, mutagenic, carcinogenic, and neurotoxic; they are also developmental and reproductive toxins and skin irritants.2



EXPERIMENTAL SECTION Sampling Information. Atmospheric particle samples were collected at the five United States IADN sampling sites (see the map in Figure S1 of the Supporting Information). The locations include urban sites in Chicago, IL (41.8344° N, −87.6247° W), and Cleveland, OH (41.4921° N, −81.6785° W), a rural site at Sturgeon Point, NY (42.6931° N, −79.0550° W), and remote sites at Sleeping Bear Dunes, MI (44.7611° N, −86.0586° W), and Eagle Harbor, MI (47.4631° N, −88.1497° W). The IADN website provides detailed information about air sampling procedures and site operations (http://www.msc.ec. gc.ca/iadn). The samples discussed here were collected during the period from March 2012 to December 2012 (inclusive) for Chicago, Cleveland, and Eagle Harbor and from July 2012 to December 2012 (inclusive) for Sleeping Bear Dunes and Sturgeon Point. A modified Anderson high-volume air sampler (General Metal Works, model GS2310) was used to collect air samples for 24 h every 12 days at a flow rate giving a total sample volume of ∼820 m3. The air stream was first passed through Whatman quartz fiber filters (QM-A, 20.3 cm × 25.4 cm) to collect the particles and then through Amberlite XAD-2 resin (Supelco, 20−60 mesh) held in stainless steel cartridges to collect the vapor phase components. Details of the sampling procedures and site operations can be found elsewhere.36 A detailed description of the sample treatment and chemical analysis procedures for the particle samples has been given elsewhere.37 In summary, the samples were spiked with known amounts of d12-tris(2-chloroethyl) phosphate and [13C18]triphenyl phosphate as surrogate standards and Soxhlet



RESULTS AND DISCUSSION Concentrations. A brief screening study of 11 vapor samples collected in Chicago and Cleveland during March 2012 and April 2012 showed that total OP (ΣOP) concentrations associated with the particle phase comprised 95 ± 2% of the vapor and particle ΣOP concentrations. The only OPs present in the vapor phase were TnBP (3.8 ± 1.4%), TCEP (3.8 ± 1.4%), TCPP (3.4 ± 1.5%), and TPP (10.2 ± 2.6%), where these percentages are the concentrations in the vapor phase relative to the concentrations in the vapor and particle phases added together. On the basis of these results, further vapor phase samples were not analyzed for OPs, and all concentrations in this study are particle phase OP concentrations only. OPs analyzed in air samples by a few previous studies were predominantly found in the particle phase, as well.30,31 However, further research of seasonal variations in the vapor phase levels of some of these compounds may be warranted; for example, the vapor pressure of TCEP is 1.1 × 10−4 Torr,2 which suggests that its vapor phase concentration in 10

dx.doi.org/10.1021/ez400034n | Environ. Sci. Technol. Lett. 2014, 1, 8−14

Environmental Science & Technology Letters

Letter

widely used as polymer and foam additives and as possible replacements for the discontinued Penta-BDE formulation.16 Interestingly, the composition profile for ΣOP concentrations was different at Sleeping Bear Dunes and Eagle Harbor, the remote sites. Nonhalogenated OP compounds were major contributors to ΣOP concentrations at these two sites, with the sum of all nine nonhalogenated OP concentrations comprising 70 ± 21 and 85 ± 13% of the ΣOP concentrations at Eagle Harbor and Sleeping Bear Dunes, respectively (see Table 1). In addition, the three chlorinated OPs measured in this study (TCEP, TCPP, and TDCPP) were detected in Eagle Harbor samples only ∼50% of the time, and TDCPP was not detected at Sleeping Bear Dunes at all. In comparison, the nonhalogenated OPs contributed only 35 ± 7 and 49 ± 5% to the ΣOP concentrations at Cleveland and Chicago, respectively. The most abundant nonhalogenated OPs at Eagle Harbor and Sleeping Bear Dunes were TnBP, TPP, and TBEP, with the highest levels for TPP and TBEP detected at Chicago and Cleveland. The exception was TnBP, which was detected at particularly high levels at Eagle Harbor (detection rate of 39%), which were statistically indistinguishable from those at Chicago and Cleveland. This is surprising considering that Eagle Harbor is a remote site and the calculated atmospheric half-life of TnBP is