This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.
Article pubs.acs.org/est
Flame Retardant Chemicals in College Dormitories: Flammability Standards Influence Dust Concentrations Robin E. Dodson,*,† Kathryn M. Rodgers,† Gale Carey,‡ Jose Guillermo Cedeno Laurent,§ Adrian Covaci,∥ Giulia Poma,∥ Govindan Malarvannan,∥ John D. Spengler,§ Ruthann A. Rudel,† and Joseph G. Allen§ †
Silent Spring Institute, 320 Nevada Street, Suite 302, Newton, Massachusetts 02460, United States University of New Hampshire, Rudman Hall, 46 College Road, Durham, New Hampshire 03824, United States § Harvard T.H. Chan School of Public Health, 401 Park Drive, Boston, Massachusetts 02115, United States ∥ Toxicological Centre, University of Antwerp, Universiteitsplein 1, 2610 Wilrijk-Antwerp, Belgium ‡
S Supporting Information *
ABSTRACT: Furniture flammability standards are typically met with chemical flame retardants (FRs). FRs can migrate out of products into dust and are linked to cancer, neurological impairment, and endocrine disruption. We collected 95 dust samples from dormitory common areas and student rooms on two U.S. college campuses adhering to two different furniture flammability standards: Technical Bulletin 117 (TB117) and Technical Bulletin 133 (TB133). Because TB133 requires furniture to withstand a much-more-demanding test flame than TB117, we hypothesized that spaces with TB133 furniture would have higher levels of FRs in dust. We found all 47 targeted FRs, including 12 polybrominated diphenyl ether (PBDE) congeners, 19 other brominated FRs, 11 phosphorus FRs (PFRs), 2 Dechlorane-Plus (DP) isomers, and 3 hexabromocyclododecane (HBCDD) isomers in the 95 dust samples. We measured the highest reported U.S. concentrations for a number of FRs, including BDE 209 (up to 990 000 ng/g), which may be used to meet the TB133 standard. We prioritized 16 FRs and analyzed levels in relation to flammability standard as well as presence and age of furniture and electronics. Adherence to TB133 was associated with higher concentrations of BDE 209, decabromodiphenylethane (DBDPE), DPs, and HBCDD compared to adherence to TB117 in univariate models (p < 0.05). Student dormitory rooms tended to have higher levels of some FRs compared to common rooms, likely a result of the density of furniture and electronics. As flammability standards are updated, it is critical to understand their impact on exposure and health risks.
■
INTRODUCTION Increasing evidence shows that flammability standards, although intended to improve fire safety, lead to exposure to harmful chemical flame retardants (FRs) that may impact human health.1 Furniture flammability standards have typically been met by adding FRs to foam, textiles, or other furniture components.2 FRs are used to raise the barrier to ignition, slow the spread of fire, and test parameters of each performancebased standard and properties of upholstery materials influence the types and amounts of FRs used.3 In the absence of national U.S. furniture flammability standards, California’s furniture flammability standards have become the de facto national standards. California’s Technical Bulletin 117 (TB117) was designed to slow fire spread from a small open flame in residential furniture, and manufacturers have met the standard by adding FRs to the foam interior.4 California’s Technical Bulletin 133 (TB133) was designed to slow fire spread from larger open flames in furniture in public spaces, and manufacturers meet this standard by adding FRs to furniture fabric and filling.5 In 2013, California updated TB117 to TB © 2017 American Chemical Society
117-2013, a standard that is designed for furniture to resist a smoldering ignition source and that can be met without added FRs. Furniture manufacturing practices have been shifting away from FRs since the new standard’s enforcement in 2015. However, most furniture currently in place at U.S. colleges and universities presumably meet either TB117 or TB133, the predominant standards furniture manufacturers build to meet and contains the FRs associated with those standards. Previous studies, including ours, have demonstrated that Californians have higher exposures to and body burdens of FRs, presumably as a result of California’s flammability standards.6−8 FRs can migrate out of furniture and end up in dust, and dust concentrations of polybrominated diphenyl ethers (PBDEs) are correlated with serum levels.9,10 The FRs used in furniture have evolved because some classes of FRs have been phased out due Received: Revised: Accepted: Published: 4860
January 23, 2017 March 23, 2017 March 31, 2017 April 13, 2017 DOI: 10.1021/acs.est.7b00429 Environ. Sci. Technol. 2017, 51, 4860−4869
Article
Environmental Science & Technology
dormitory rooms. The Institute of Medicine and National Research Council have identified young adults (ages 18−26) as a separate group from adolescents and adults that are experiencing a critical period of development and recommends targeted preventative health practices in this age group.29 Engaging college students in environmental health research also provides opportunities to educate the next generation of consumers and professionals about the role that chemical exposures play in disease. To provide data on chemical FR concentrations measured on college campuses and to consider the relationship between measured chemical levels and different flammability standards in force on each campus, we targeted 47 FRs in vacuum dust samples collected from 95 student residential spaces on two university campuses located in the northeastern United States. These two campuses adhered to two different flammability furniture flammability standards: TB117 and TB133. Because TB133 requires furniture to withstand a much larger and longer test flame than TB117, we hypothesize that spaces with TB133 furniture have different FR profiles and potentially higher levels of FRs in dust compared to the spaces with TB117 furniture. We compared dust concentrations between the campuses, and we also looked for associations between FR concentrations and time since last renovation and number and types of furniture and electronics in the dormitory rooms. Our objective was to identify the factors that are most influential in contributing to chemical FR exposures in college dormitories, considering flammability standards as well as the age of the furniture and room furnishings. These data will be used to inform purchasing decisions in higher education and evaluate flammability standards.
to health concerns. For example, in 2005, manufacturers phased out PentaBDE, a FR mixture commonly used in upholstered furniture, which was slated to be banned by California11 and tightly restricted by the United States Environmental Protection Agency (U.S. EPA).12 This mixture was replaced with Firemaster 550 as well as other brominated FRs (BFRs) and phosphorus FRs (PFRs).13,14 The impact of this change in manufacturing is reflected in house dust concentrations, with PentaBDE levels generally decreasing and levels of Firemaster 550 and other FRs generally increasing since 2005.15 This trend was also seen in a 2012 study that found that PBDEs were the most commonly detected FR in couches bought before 2005, whereas tris 1,3-dichloro-isopropyl phosphate (TDCIPP; chlorinated “tris”) and Firemaster 550 FRs were the most commonly detected FRs after 2005.2 In addition to upholstered furniture, FRs are also used in building insulation, electronics, other textiles, and plastics.15−17 The efficacy of FRs in upholstered furniture has been called into question,1,3 while studies of their toxicity continue to emerge. FRs include classes of chemicals that have been associated with reproductive and developmental toxicity, altered metabolism, neurotoxicity, and cancer.18,19 The mostly phased-out, yet persistent, PBDEs are associated with thyroid disease and neurological impairment.18,20,21 Firemaster 550 components and TDCIPP have been associated with altered reproductive and thyroid hormone levels in animals and humans,22 including decreased free thyroxine and sperm concentration in men.23 Some halogenated PFRs are classified as carcinogenic,24 and structural similarities to organophosphate pesticides also suggest the potential for neurological impairment.25,26 Residence halls on college campuses are unique locations for studying exposures to FRs. Common areas and student lounges are typically furnished and controlled by the institution, and information about the furnishings, including the date of most recent furnishing, is available. FR levels in student dormitory rooms are influenced by furnishings and electronics, also sources of FRs, both provided by the college as well as brought in by students. To our knowledge, only one study has investigated the impact of furniture flammability standards on concentrations of FRs in dust in college residences.27 Keimowitz et al. analyzed 20 dust samples collected before and after the introduction of new upholstered furniture in student housing in which furniture mostly meeting TB117 was replaced with TB133 furniture. A total of three phosphate FRs (TDCIPP, tris(2-chloroethyl) phosphate (TCEP), and triphenyl phosphate (TPHP)) were detected in dust collected from residences that reportedly mostly met TB117, whereas TPHP, but not TDCIPP or TCEP, was detected in residences with new TB133 furniture. However, limited time elapsed between measurements, and the semivolatile FRs likely had not yet reached steady-state in the indoor spaces. This present study is part of a larger initiative called the Healthy Green Campus project,28 aimed at incorporating environmental chemicals and their relationship to health within sustainability programs in higher education. Because many sustainability programs in higher education are generating and tracking large amounts of data on energy use, greenhouse gas emissions, water use, and waste production, there is also an opportunity to measure environmental chemical levels in the campus community. Environmental chemical measurements have not been well-characterized in these settings, and this study provides some of the first data on FRs in student
■
MATERIALS AND METHODS Sample Collection. Dust samples were collected during the spring of 2015 (March−May) in dormitory common areas (i.e., student-accessible areas in residence halls furnished by the campus) and student rooms on two New England university campuses adhering to different furniture flammability standards. We confirmed with the purchasing offices on each campus that they furnished campus spaces according to local fire codes, which we also confirmed. One campus purchased institutional furniture to meet California’s TB117, while the other campus purchased furniture to meet California’s TB133. Trained field staff collected dust from residential spaces using vacuum cleaners fitted with a custom aluminum crevice tool following previously established protocols (Dyson, Inc., Chicago, IL).30−32 Field staff slowly dragged the crevice tool over all surfaces, including floors, mattresses, furniture, and window sills, for approximately 30 min to collect dust in cellulose extraction thimbles (19 × 90 mm). To ensure an adequate dust sample for subsequent analysis, some spaces required multiple thimbles, which were stored together in quality certified precleaned glass jars (Environmental Sampling Supply, Dallas, TX). Jars with thimbles were stored at 0.98) (Figure SI2). The brominated components of Firemaster 550 (EH-TBB and BEH-TEBP) were also strongly correlated (ML estimate = 0.76); however, these components were only weakly, although significantly, correlated with TPHP, another component of Firemaster 550, suggesting additional uses of TPHP, including as a plasticizer. We observed relatively little correlation between the brominated components of Firemaster 550 and PentaBDE congeners, which does not support the assumption that Firemaster 550 is the primary replacement for the once widely used PentaBDE mixture in upholstered furniture. Also, there may be other potential uses of the brominated components in addition to their presence in Firemaster 550. TDCIPP and TCIPP, also often used as PentaBDE substitutes in foam, were not correlated with PentaBDE congeners either, suggesting their use in other products. We expected to see a negative correlation between the recently phased-out DecaBDE and its replacement, DBDPE. Instead, we observed a significant weak positive correlation between the two, suggesting they potentially co-occur or are used in different products in the space. Impact of Flammability Standards on Concentrations. We first compared concentrations in common areas on the two campuses (TB117 versus TB133) in a univariate regression model. Of the 16 prioritized FRs, 4 FRs were found at higher modeled geometric mean (GM) concentrations on the TB133 campus (Figure 1 and Table 1). BDE 209, DBDPE, DP, and HBCDD were significantly higher on the TB133 campus compared to the TB117 campus (adjusted p value of < 0.05). Modeled GM BDE 209 concentrations were 6 times higher on the TB133 (GM = 6000 ng/g) versus TB117 (GM = 970 ng/g) campus, DBDPE concentrations 10 times higher (TB133 GM = 300 ng/g versus TB117 GM = 29 ng/g), and DP and HBCDD isomers were 4 and 9 times higher on the TB133 campus, respectively. BDE 209 and its presumed replacement DBDPE are used in textile back-coatings to meet severe upholstered furniture flammability standards, like TB133.54 BEH-TEBP and TCIPP were marginally significantly higher on the TB133 campus (adjusted p value of 0.099 and 0.077, respectively). Studies that include samples collected from spaces adhering to TB133 are limited. A study of office dust, air, and surface wipes from the Boston area, which likely included TB133 samples, as was specified by the Boston fire code at the time, found relatively similar GM dust concentrations of the major PBDE congeners (47, 99, 100, and 209) as our overall study medians and modeled GM levels, with the exception of BDE 209, which had a higher modeled GM on the TB133 campus than the Boston office study.55 Because FR use patterns vary over time as chemicals are banned or restricted, we expected dust concentrations to reflect the flammability standards in place at the time the space was
found in 19 other studies of FRs in dust in North America since 2006,15,23,27,35−50 with higher median levels than have previously been reported for TCIPP, EH-TBB, BEH-TEBP, anti-DP, and syn-DP. We found higher maximum or 95th percentile concentrations than have previously been reported for several FRs: BDE 47, BDE 99, BDE 100, BDE 209, EHTBB, TBBPA, DBDPE, TCIPP, anti-DP, syn-DP, and sum HBCDD. Comparisons are listed in Table SI2. We detected all priority PBDE congeners (Table 1). The maximum BDE 209 level in this study was the highest reported at 990 000 ng/g, with the next highest at 106 204 ng/g from dust collected in 2009 in Boston offices.37 We also found higher maxima of BDE 47 (130 000 ng/g), BDE 99 (140 000 ng/g), and BDE 100 (69 000 ng/g), which were more than 5,35 3,35 and 6 times greater15 than the next-highest levels reported, respectively. Our samples were collected 3 to 4 years after these samples with the next-highest levels, indicating PentaBDE’s persistence, its potential continued use in the marketplace, even after it was phased out of production in 2005,13,14 or that older, more-worn furniture may release PentaBDE to a larger degree than newer furniture does. Interestingly, some of the highest concentrations of these PentaBDE congeners were measured in student dormitory rooms on the TB117 campus, two of which were from the same building; an exception was a highly furnished lounge on the TB133 campus. A comparison of furniture and electronic items in these rooms did not reveal commonalities that would explain these findings. We recognize that one possible reason for observing the highest levels of some of these FRs is our large sample size, which increases the probability of observing higher levels. We found four priority PFRs (TCEP, TCIPP, TDCIPP, and TPHP) in 100% of dust samples (Table 1). Of these, median concentrations of TCEP, TDCIPP, and TPHP were lower or similar to other studies, with the exception of TCIPP, which was over 2 times higher (at 5400 ng/g) than the previous highest median (2200 ng/g) reported in California house dust.15 The maximum TDCIPP in this study (170 000 ng/g) is among the highest reported in U.S. dust, exceeded only by a small study of PFRs in student housing.27 We found Firemaster 550 components, EH-TBB and BEHTEBP, in nearly all samples (Table 1). Only median dust levels collected from a California fire station44 were higher than the median EH-TBB (1200 ng/g) and BEH-TBEP (1200 ng/g) levels we report. Our maximum EH-TBB concentration (130 000 ng/g) was about 1.5 times greater than the nexthighest level (86 007 ng/g),38 though our highest BEH-TBEP concentration (37 000 ng/g) was about 1.2 times lower than the highest level (47 110 ng/g) reported from Stapleton et al.’s 2009 study of vacuum bag house dust collected from Boston homes from 2002 to 2007.42 Samples in that study were collected before the presumed widespread introduction of Firemaster 550, adding support to previous reports that BEHTBEP was used in other flame-retardant mixtures before introduction of Firemaster 550.51 The median concentrations of other priority BFRs, such as TBBPA, BTBPE, and DBDPE, were lower than other studies, although we found the highest maxima reported for TBBPA and DBDPE (66 000 and 15 000 ng/g, respectively). Median concentrations of DP isomers in this study (anti-DP: 13 ng/g and syn-DP: 5.7 ng/g) were the highest median levels reported, and the median concentration of sum HBCDD (380 ng/g) was among the highest, with Stapleton et al. (2008)47 finding a similar median level (350 ng/g). Both DPs and sum 4863
DOI: 10.1021/acs.est.7b00429 Environ. Sci. Technol. 2017, 51, 4860−4869
4864
common area student dorm
HBCDDg
100 100
100 100
100 100 100 100 100 100 100 100
100 96 100 100 100 100 53 65 87 100
100 100 100 100 100 100 100 100 100 100
% > LOQ
240 320
13 11
97 810 2200 6200 6200 4400 5000 8200
2700 620 850 960 140 260 8.2 8.7 33 100
80 220 270 330 21 52 91 20 890 540
median
380 930
15 19
720 2300 4600 9100 15 000 10 000 6000 12 000
6100 11 000 1300 3400 540 340 12 37 55 680
6700 460 8300 780 1500 120 160 58 1200 770
mean
980 1600
31 69
3100 6800 14 000 30 000 55 000 42 000 11 000 36 000
24 000 46 000 2900 12 000 2200 1200 43 73 200 580
28 000 2000 34 000 2800 6500 480 530 230 2900 2500
95th percentile
% > LOQ
median
polybrominated diphenyl ethers 43 000 100 420 3000 100 580 52 000 100 640 4500 100 850 11 000 100 100 770 100 130 540 92 11 530 95 13 4800 100 4200 3400 100 1200 other brominated flame retardants 25 000 100 2000 110 000 100 1300 3900 100 2400 28 000 100 1300 4400 100 560 1400 100 120 52 75 10 620 81 10 220 100 250 14 000 100 140 phosphate flame retardants 8600 100 240 32 000 100 620 33 000 100 7600 50 000 100 6100 120 000 100 2200 66 000 100 3300 13 000 100 5800 41 000 100 8500 Dechlorane Plus 38 100 42 130 100 37 hexabromocyclododecane 1400 100 1000 14 000 100 660
max
61 000 4300
340 140
410 1000 23 000 11 000 5500 10 000 7200 12 000
15 000 4900 5300 2800 9600 480 42 22 1700 1100
1300 9500 1600 11 000 300 3200 75 55 89 000 2500
mean
TB133b
300 000 6800
1600 390
960 2300 96 000 37 000 18 000 31 000 16 000 24 000
68 000 22 000 20 000 7300 44 000 1300 150 60 7500 6000
5500 55 000 6100 60 000 1200 13 000 370 140 460 000 7900
95th percentile
300 000 130 000
2800 1900
1000 7400 150 000 78 000 28 000 170 000 21 000 110 000
130 000 39 000 37 000 30 000 66 000 6100 170 270 12 000 15 000
11 000 13 0000 12 000 140 000 2400 69 000 510 1100 990 000 21 000
max
280
14
4800
6400
2700
150
29
6.2
150
1000
2400
970
57
120
780
500
TB117 GMc
2500i
61i
4900
3000
8000 j
290
300i
14
490
2300j
2200
6000i
15
92
570
390
TB133 GMd
670 560
26 29
200h 710 4400 6400 4600 4100 4900 j 8100
2300 1400 1500 1600 240 180 9 10 83 170
450 600 680 950 110 150 32 18 2200i 1000
GMe
N common area = 15 dust samples; N student dorm room = 26 dust samples. bN common area = 12 dust samples (TBBPA and HBCDD = 10 dust samples); N student dorm room = 42 dust samples (TBBPA and HBCDD = 41 dust samples). cModeled geometric mean (GM) for TB117 common area. Model with natural logarithm of the concentration as the outcome and flammability standard as the sole predictor. dModeled GM for TB133 common area. Model with natural logarithm of the concentration as the outcome and flammability standard as the sole predictor. eModeled GMs for common area and student dormitory rooms across both campuses. Model with natural logarithm of concentration as the outcome and room type as the sole predictor. fIncludes anti-DP and syn-DP isomers. g Includes α-HBCDD, β-HBCDD, and γ-HBCDD isomers. hAdjusted p value of
