- Email: [email protected]
Concentrations, sources and human exposure implications of organophosphate esters in indoor dust from South Africa
Chemosphere 230 (2019) 239e247
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Concentrations, sources and human exposure implications of organophosphate esters in indoor dust from South Africa Ovokeroye A. Abafe 1, Bice S. Martincigh* School of Chemistry and Physics, University of KwaZulu-Natal, Westville Campus, Private Bag X54001, Durban, 4000, South Africa
h i g h l i g h t s Organophosphate ester (OPE) concentrations in 50 indoor dust samples were measured. Results indicate widespread use of OPEs in Durban, South Africa. Cars contained the highest levels of OPEs. TDCIPP and TCEP, known carcinogens, contributed majorly to human exposure levels. This is the first report of OPEs in the indoor environment in South Africa.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 February 2019 Received in revised form 15 April 2019 Accepted 21 April 2019 Available online 3 May 2019
The concentrations of four organophosphate esters (OPEs) were measured in 50 dust samples from homes (n ¼ 10), offices (n ¼ 9), university computer laboratories (n ¼ 12) and cars (n ¼ 19) in Durban, South Africa. The median concentrations Sn¼4 OPEs were 22940, 26930, 19565 and 49010 ng g⁻1 in homes, offices, university computer laboratories and cars respectively. OPEs were detected in all samples with the exception of one car and one computer laboratory sample in which TDCIPP was not detected. Significant association of indoor characteristics with OPE concentrations was observed. OPEs positively correlated (r ¼ 0.22, p value ¼ 0.4862) with electronics and correlated (r ¼ 0.522, p value ¼ 0.0675) with foams and furniture in homes. By employing the median concentrations and an average dust intake rate, the exposure doses (ng d1) were found to be 169 (TCEP), 74 (TCIPP), 162 (TDCIPP) and 55 (TPHP) for adults; 159 (TCEP), 70 (TCIPP), 108 (TDCIPP) and 57 (TPHP) for teenagers; 317 (TCEP), 152 (TCIPP), 334 (TDCIPP) and 94 (TPHP) for toddlers. The predominance and exposure magnitude of OPEs in the South African environment require further investigations to determine cumulative human health effects arising from mixtures of these compounds through multiple exposure routes. © 2019 Elsevier Ltd. All rights reserved.
Handling Editor: Jerzy Falandysz Keywords: Organophosphate esters Phosphorus flame retardants Plasticizers Indoor dust Sources Human exposure South Africa
1. Introduction Organophophate esters (OPEs) are used in a variety of applications. The chlorinated alkylphosphates, such as, tris(1-chloro-2propyl) phosphate (TCIPP), tris(1,3-dichloro-2-propyl) phosphate (TDCIPP) and tris(2-chloroethyl) phosphate (TCEP), are used mainly as phosphorus flame retardants (PFRs) in polyurethane foams (Andresen and Bester, 2006), electronic equipment, textiles, plastic
* Corresponding author. E-mail address: [email protected] (B.S. Martincigh). 1 Present address: Residue Laboratory, Agricultural Research CouncilOnderstepoort Veterinary Research, 100 Old Soutpan Road, Onderstepoort, Pretoria, 0110, South Africa. https://doi.org/10.1016/j.chemosphere.2019.04.175 0045-6535/© 2019 Elsevier Ltd. All rights reserved.
and building materials (Salamova et al., 2014), whilst the nonderivatized organophosphates, such as triphenyl phosphate (TPHP), are majorly used as plasticizers, lubricants, varnishes, glues, airplane hydraulic fluids and to regulate pore sizes, for example, in concrete (Brandsma et al., 2014). They are sometimes used as substitute flame retardants for the halogenated compounds, for example, the use of TPHP in electronic devices (Andresen and Bester, 2006). There has been a great increase in the use of PFRs to meet fire safety regulations in consumer products since restrictions were imposed on the use of polybrominated diphenyl ethers (PBDEs), in particular, penta-, octa- and deca-BDE technical mixtures (Brandsma et al., 2014). Since PFRs are not covalently bound to materials but are used as additive flame retardants, they may offgas and leach from products through volatilization and/or
240
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
abrasion into the environment (Brandsma et al., 2014; Abdallah and Covaci, 2014). Little is known of the toxicity of OPEs, however, studies have reported harmful effects of OPEs to include altered hormone levels and decreased sperm concentrations (Meeker and Stapleton, 2009); neurotoxic, mutagenic and carcinogenic effects in rats and mice; haemolytic and reproductive effects in humans (Van Der Veen and De Boer, 2012); and potential human carcinogenicity (Abdallah and Covaci, 2014). PFRs have been detected in various environmental media including air (Marklund et al., 2003, 2005; Carlsson et al., 1997; €ck et al., 2006; Bjo € rklund et al., 2004; Hartmann et al., 2004); Tollba surface and drinking waters, and sediments (Martínez-Carballo et al., 2007; Bacaloni et al., 2007); biota (Campone et al., 2010; Sundkvist et al., 2010) and indoor dust of various microenvironments in several locations worldwide (Brandsma et al., 2014; Abdallah and Covaci, 2014; Marklund et al., 2003; Brommer et al., 2012; Kajiwara et al., 2011; Bergh et al., 2011; Stapleton et al., 2009; Wu et al., 2016). Much attention has been given to the significance of indoor dust as a pathway of human exposure to OPEs. The relationship between dust and human body burdens is strongly implied by the association of PFRs in household dusts and human semen quality and hormone levels (Meeker and Stapleton, 2009). Nothing is known on the production, use, distribution and fate of OPEs in South Africa. Moreover, despite the increasing proof of the significant implications of indoor dust for human exposure to OPEs, attempts to link indoor contaminants with probable source items has had limited success. A dearth of information also exists for human exposure and pathways to OPEs. To bridge these gaps, the aim of the present study was to: (i) investigate the contamination of indoor dust by four OPEs (TCEP, TCIPP, TDCIPP and TPHP) in multiple indoor environments in South Africa as no data is currently available, (ii) compare the profiles of OPEs in the different microenvironments (cars, homes, offices and university computer laboratories), (iii) establish the relationships between various household products and the concentrations of OPEs in dust in order to identify their possible sources in the indoor environment, and (iv) estimate the exposure magnitude of OPEs among different population groups, by utilizing various exposure scenarios. 2. Material and methods 2.1. Chemicals Analytical standards (of known purity) of TCEP, TCIPP, TDCIPP, and TPHP were purchased from Sigma-Aldrich, South Africa. The internal standard, 13C12elabelled decachlorobiphenyl (13C12 PCB209), was obtained from Wellington Laboratories, Guelph, Ontario, Canada. Anhydrous sodium sulfate was from Associated Chemical Enterprises (ACE), Johannesburg, South Africa. Silica gel 90 was from Sigma-Aldrich, South Africa. A Restek Rtx®-1614 fused silica (5% diphenyl 95% dimethyl polysiloxane) capillary column was obtained as a gift from Restek Corporation, Bellefonte, PA, USA. All solvents were high performance liquid chromatography grade purchased from Sigma-Aldrich, South Africa. 2.2. Sampling A total of 50 dust samples were collected from homes (n ¼ 10), university students’ computer laboratories (n ¼ 12), and university staff offices (n ¼ 9), between August and October 2012 in Durban, South Africa. Similarly, dust samples (n ¼ 19) were collected between January and March 2013 from personal and previously owned cars available for resale. The previously owned cars were
sampled at a dealership in Durban, South Africa. All the vehicles from the dealership had been through a thorough cleaning process on arrival at the dealership prior to resale. Personal cars sampled had not undergone any form of cleaning at least three days before sampling. Information on the sampling protocol can be found elsewhere (Abafe and Martincigh, 2015). Detailed questionnaires were used to obtain pertinent information on cars, homes, offices and computer laboratories. This information included location, time since floor was last vacuumed, type of ventilation and flooring, and the number and types of electronic/electrical devices and furniture, model and manufacturer of car, year of manufacture, etc. Interviews were also conducted to obtain further information on building ages and to determine if, and when, any renovations were carried out. Since only 50 and 89% of the questionnaires were returned for homes and offices, respectively, correlation analyses were based on those samples for which questionnaires were returned. 2.3. Extraction and clean-up Samples were extracted following a previously reported method (Abafe and Martincigh, 2015) with slight modifications. Briefly, approximately 1.0 g of each sample was quantitatively weighed into a glass test tube and spiked with 2 mg 13C12 PCB-209. A volume of 12 mL n-hexane:methanol (1:3 v/v) was added. Samples were mixed in an orbital shaker for 15 min and then extracted in an ultrasonic water bath at 40 C for 30 min. The mixing and extraction was repeated for a second time without addition of fresh solvent. The samples were then centrifuged at 3500 rpm for 10 min and the supernatants were stored at <4 C prior to clean-up. Representative samples (n ¼ 5) from each microenvironment were analysed in triplicate. For the clean-up of extracts, silica gel column chromatography was used. Silica gel was activated at 130 C for 16 h and anhydrous sodium sulfate was baked at 450 C for 5 h before use. Silica gel and anhydrous sodium sulfate were subsequently cooled in a desiccator. A 30 cm 1 cm glass column was packed with 4 g deactivated silica gel. Each column was topped with 2 g of anhydrous sodium sulfate and then 30 mL of the extraction solvent was passed through the column. Extracts were loaded onto columns just before the exposure of the sodium sulfate layer. OPEs were eluted with 30 mL nhexane. This was kept as fraction 1; columns were further eluted with 30 mL of diethyl ether:n-hexane (50:50 v/v), and kept as fraction 2; both fractions were mixed together. Finally, columns were eluted with 30 mL acetone:dichloromethane (1:1 v/v) as fraction 3. The column flow rates were maintained at approximately 0.5 mL min1. Eluates were reduced in volume and concentrated to approximately 250 mL in a rotary evaporator at 55 C and stored in 1.5 mL amber glass GC/MS vials. All extracts were stored at <4 C until instrumental analysis. 2.4. GC-EI/MS analysis The analysis of OPEs was performed on an Agilent 6890 (Palo Alto, CA, USA) gas chromatograph (GC), coupled to a 5973 N series mass spectrometer (MS) operated in the electron impact (EI) ionization mode. A Restek Rtx® e 1614 fused silica (5% diphenyl, 95% dimethyl polysiloxane) capillary column (15 m 250 mm 0.1 mm) was used to effect separation, and the MS was operated in the selected ion monitoring (SIM) mode. The injections were made in the pulsed splitless mode with the injector temperature set at 250 C. The injection volume was 2 mL. The GC oven temperature programme started at 90 C (held for 2 min), then increased at 20 C min1 to 270 C and held for 1 min, and finally increased at 10 C min⁻1 to 290 C and held for a minute. Helium was used as the
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
carrier gas at a flow rate of 1.2 mL min1 and a constant linear velocity of 37 cm s1. For the MS, the ion source and transfer line temperatures were 230 and 350 C, respectively; and the ionization energy was 70 eV. OPE mass spectra were obtained in full scan mode to select prominent ions that were utilized in the SIM mode (Table 1). Quantification was carried out by means of a multiple point internal standard method. 13C12-labelled PCB-209 was employed as an internal standard for all OPEs studied. The response factors were determined from the slope of a plot of the ratio of peak areas against the ratio of the concentrations. The values for the plots were obtained from a 5 to 6 point triplicate analysis of the OPE standard solutions diluted to fall within a concentration range of 2e10 mg mL1. 2.5. Quality control/assurance The recovery for OPEs in dust was determined from spiked anhydrous sodium sulfate at different spike concentrations (Table 2). Samples were left to stand for at least 21 days at 10 C. Spiked samples were extracted and cleaned-up following the procedure for real samples. Method blanks were analysed with every batch of ten samples. For the method blank, dust samples were replaced with anhydrous sodium sulfate and passed through all the analytical procedure carried out for real samples. TPHP concentrations in samples were corrected for the blank concentrations as values found in the method blank (n ¼ 5) were as high as 1.6% of the concentrations in samples. Solvent blanks were injected after the analysis of at most three samples. All glassware was cleaned with laboratory wash solutions, rinsed with distilled water and then with organic solvents. Non-volumetric glassware was oven-dried prior to use. Direct ultraviolet light and plasticware was avoided throughout the analysis. The instrument responses to pure OPE analytical standards employed for calibrations were linear with r2 > 0.99 for all OPEs studied. Limits of detection (LODs) ranged from 0.6 ng g1 for TCEP to 100 ng g1 for TPHP; limits of quantitation (LOQs) were in the range 1.8 ng g1 (TCEP) and 301 ng g1 (TPHP) (see Table 1). Matrix spike recovery samples were prepared by adding 200e10000 ng g1 of each OPE to glass test tubes filled with 1.0 g anhydrous sodium sulfate for each concentration listed in Table 2. Matrix spikes were subjected to all analytical protocols employed for real samples. The average recovery and the percentage relative standard deviation (%RSD) for each of the OPEs at the respective spiked concentrations are shown in Table 2. 2.6. Statistics The distribution of OPE concentrations in the different microenvironments was tested with the Shapiro-Wilk test of normality. Descriptive statistics such as sum, mean, median, minimum, maximum and parametric statistics such as t-test and analysis of variance (ANOVA) were calculated by using Microsoft Excel® 2010.
241
Non-parametric statistics such as Wilcoxon Signed-Ranks test, Kendall tau test and Spearman rank correlation were performed with Analyse-it® software in Microsoft Excel 2010. The KruskalWallis test was employed to test for differences in location by using XLSTAT 2014 software. Principal component analysis (PCA) was performed with SIMCA version 13 statistical software. Limits of detection (LODs) and quantitation (LOQs) were estimated following Thomsen et al. (2003). Samples below the detection limit were treated as zero throughout the statistical analysis. 3. Results and discussion 3.1. Concentrations of OPEs in house dust The four organophosphate esters were detected in all the dust samples collected from homes in South Africa (Fig. 1). A full list of the OPE concentrations in home samples is presented in Supplementary Material Table S1. Descriptive statistics for the distribution of these OPEs are summarized in Table 3. The strong positive correlation of OPEs in this microenvironment (Supplementary Material Table S5) is indicative of similar sources for the PFRs in South African homes. Nothing is known on the production and use of organophosphate flame retardants in South Africa. However, due to industrialization in South Africa comparable to most countries in the European Union, and because of the very high levels of these OPEs found in indoor dust of two industrial sites e a textile and polyurethane industry, we hypothesize the possible indoor contamination of OPEs from the products of these industries and others which eventually end up indoors. The OPE profile in the house dust was TDCIPP > TCEP > TCIPP > TPHP. The distribution pattern of OPEs in this study is similar to those reported in Swedish homes (Bergh et al., 2011), German homes (Brommer et al., 2012) and Egyptian homes (Abdallah and Covaci, 2014); in which TDCIPP dominated in the house dust. Despite the replacement of TCEP by TCIPP, the house dust concentrations of TCEP in this study overwhelmed those of TCIPP, and TCEP was found to be the second most abundant OPE in South African homes similar to the observations of Bergh et al. (2011) in Swedish homes. However, the concentrations of TCEP strongly correlated (r ¼ 0.82) with TCIPP in this microenvironment (Supplementary Material Table S5). A statistically significant difference (p ¼ 0.022) was observed among the concentrations of the individual OPEs in South African homes. The high levels of TCEP in these house dust samples suggests a high volume use of this PFR and possibly the domineering Chinese flame retardant V6 [2,2ebis (chloromethyl)propane-1,3-diyltetrakis(2-chloroethyl)biphosphate] which is applied to polyurethane foam commonly present in furniture and car foam (Fang et al., 2013). The high vapour pressure (1.1 10⁻4 mmHg) of TCEP is likely responsible for its easy migration from treated products during the product's useful life. The elevated TCEP concentrations in South African homes should be of particular interest due to the reported carcinogenicity of TCEP (Fang et al., 2013; Stapleton et al., 2009). The OPE concentrations
Table 1 Qualifying and quantitation ions, r2, limits of detection and quantitation for TCEP, TCIPP, TDCIPP and TPHP. Analyte
Qualifier and quantifier ion
R2
LODa/ng g1
LOQa/ng g1
TCEP TCIPP TDCIPP TPHP 13 C12-PCB 209b
205, 201, 321, 326, 510
0.999 0.998 0.995 0.992
0.6 51 91 100
1.8 152 274 301
a b
Thomsen et al. (2003). Internal standard.
249 277 381 325
242
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
Table 2 Recovery of individual OPEs. Analyte
Spiked concentration (n ¼ 4)/mg g1
Average concentration found (n ¼ 4)/mg g1
Standard deviation
Recovery/%
Relative standard deviation/%
TCEP
0.20 1.00 2.00 6.00 0.20 1.00 2.00 6.00 0.60 2.00 10.0 0.60 2.00 10.0
0.19 1.22 1.74 5.16 0.19 0.74 1.63 5.92 0.57 2.71 12.87 0.61 2.03 10.79
0.05 0.002 0.03 0.44 0.09 0.10 0.15 0.22 0.04 0.76 0.82 0.10 0.05 0.16
95 122 87 86 95 74 82 99 95 136 129 102 102 108
25 0.2 2 8.5 47 14 9 4 7 28 6 16 3 2
TCIPP
TDCIPP
TPHP
did not correlate well (r ¼ 0.04) with the PBDE concentrations in the corresponding samples reported earlier (Abafe and Martincigh, 2015). This suggests differences in sources and/or release mechanisms of these flame retardants in the house dust samples. The P correlations of OPE concentrations in dust and the number of electronics (r ¼ 0.22, p value of 0.4862) and the number of foams and furniture (r ¼ 0.522, p value of 0.0675) in these homes, strongly implicate these household items as sources of OPEs.
3.2. Concentrations of OPEs in office dust All four OPEs were detected in 100% of the dust samples from offices in South Africa (Fig. 1). A full list of the concentrations of the individual OPEs in this microenvironment can be found in Supplementary Material Table S2 and the descriptive statistics are given in Table 3. Unlike the distribution of OPEs in home samples where TDCIPP was predominant, TCEP was most abundant in the office dust samples. The relative abundance of OPEs were TCEP > TPHP > TDCIPP > TCIPP. Limited data are available on indoor dust TCEP concentrations worldwide. The concentrations of TCEP were higher than the reported concentrations in two offices in Sweden (i.e. the minimum concentrations of TCEP in the South African offices were much higher than the maximum concentrations of TCEP reported in the Swedish offices) (Marklund et al., 2003). The high levels of TCEP in the office dust could be associated with indoor characteristics; however, there were no correlations between the concentrations of OPEs and office characteristics such as electronics or furniture in these offices. It should be noted that the Swedish study by Marklund et al. (2003) was conducted prior to the ban of commercial penta- and octa-PBDE mixtures, which had wider application in foam and furniture compared with organophosphate flame retardants. The abundance of TCEP and TPHP may be indicative of the possible use of V6 and Firemaster 550 (FM550) in flame retarding various office products in South Africa. TCEP is a component of V6 and TPHP is a major component of FM550 (Stapleton et al., 2009). Fang et al. (2013) showed a significant relationship between V6 and TCEP in dust samples and concluded that V6 is an important source of TCEP in the environment. This observation is supported by the poor correlation of TCEP with the other studied OPEs, whilst TCIPP and TDCIPP concentrations are correlated (r ¼ 0.49); similar to the strong correlation (r ¼ 0.73) of TDCIPP and TPHP concentrations in the offices (Supplementary Material Table S7). These relationships suggest similar sources for TCIPP, TDCIPP and TPHP in the office dust samples. We cannot exclusively associate the presence of TPHP in this microenvironment to the flame retardant FM550 since TPHP has other
applications as a plasticizer in various indoor products which may also account for a source of TPHP in the office dust (Brandsma et al., 2014).
3.3. Concentrations of OPEs in university computer laboratories The percentage contribution of individual OPEs in this microenvironment is presented in Fig. 1. The descriptive statistics of OPEs in the university computer laboratories are presented in Table 3. As observed for OPEs in offices, TCEP was the major OPE found in these laboratories. The concentration profiles of the OPEs were in the order: TCEP > TDCIPP > TPHP > TCIPP. The dominance of TCEP in this microenvironment is not surprising as a previous study had reported high levels of TCEP in the indoor air of a lecture hall with computers as opposed to a hall without computers in which TCEP was not detected in the same study (Staaf and Ostman, 2005). The authors also reported TCEP as high as 10 ng m-3 in air of an electronic dismantling facility (Staaf and Ostman, 2005). The concentration of TCEP in these computer laboratories calls for concern owing to the fact that these microenvironments were refurbished between 2006 and 2007. All carpeting, furniture, blinds, computers and printers were replaced and the dividing walls in some of the rooms were also replaced with plasterboard. Although no report on OPEs in a similar microenvironment is available in the literature, the levels of TCEP in this study exceeded the levels reported in dust from homes in New Zealand (Ali et al., 2012) and dust from Belgian homes (Van den Eede et al., 2011). All OPEs were normally distributed in this microenvironment. The correlation matrix (Supplementary Material Table S8) indicates a good relationship between the concentrations of the studied OPEs. However, the concentrations of TCEP negatively correlated with those of TPHP. The correlation of TDCIPP and TPHP concentrations in the computer laboratories resembles that of the offices; hence providing further evidence for source similarities of both OPEs in these microenvironments. The concentrations of TPHP in these microenvironments are not surprising as TPHP concentrations in excess of 500 000 ng g1 have been reported in liquid crystal display (LCD) televisions and laptop computers in Japan (Kajiwara et al., 2011). In the same light, Brandsma et al. (2014) showed high contamination levels of TPHP in dust samples collected from electronic equipment consistent with the observations of Kajiwara et al. (2011) in which TPHP was the major OPE detected in electronic equipment from the Japanese market. TPHP is used as a flame retardant and plasticizer in a variety of products, which may also account for some of the sources of TPHP in this microenvironment. The high volatilities of TDCIPP and TPHP are most likely responsible for their concentrations in the indoor dust since they can easily be released from
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
Fig. 1. Percentage contribution of individual OPEs in dust collected from (a) homes, (b) offices, (c) university computer laboratories and (d) cars in South Africa.
products into the environment (van der Veen and de Boer, 2012). 3.4. Concentrations of OPEs in cars All four OPEs were detected in the cars sampled for this study with the exception of one car in which TDCIPP was not detected.
243
Fig. 1 shows the percentage contribution of each OPE in the car dust samples in South Africa. A full list of the concentrations of OPEs in this microenvironment can be found in Supplementary Material Table S4. The concentration profiles of OPEs in this microenvironment are similar to those of the house dust. The abundance of OPEs in these cars is in the order: TDCIPP > TCEP > TCIPP > TPHP. A summary of the descriptive statistics for OPEs in cars is given in Table 3. The concentration profiles of TDCIPP in these cars are similar to concentrations reported in car dust from the Netherlands (Brandsma et al., 2013) and Germany (Brommer et al., 2012). The high TDCIPP concentrations in car dust are in tandem with its use as a flame retardant in flexible and rigid polyurethane foams (PUF) (Marklund et al., 2003). Unlike the observations of Brommer et al. (2012), in which high TDCIPP concentrations were associated with older cars, the concentrations of TDCIPP did not show a particular trend with the year of manufacture of the vehicle in South Africa. The highest TDCIPP concentration reported in this study was found to be 697 100 ng g1 in a car manufactured in 2009. However, a similar car from the same manufacturer, of the same model and year, had a TDCIPP concentration of only 2600 ng g1. The cause of this discrepancy could not be ascertained, since information obtained showed that both cars were cleaned at approximately the same time (i.e. 90 days) prior to the date of sampling. However, Brommer et al. (2012) postulated that intensive car usage leads to greater abrasion of vehicle upholstery with an attendant increase in the release of flame retarded upholstery fabric fibres, which may be accountable together with external sources for the variations of TDCIPP levels in the two cars. No direct relationship was found between OPE concentrations and year of car manufacture. Whilst TDCIPP and TPHP showed weak positive correlations of r ¼ 0.14 and r ¼ 0.15 with car manufacture year; TCEP and TCIPP only showed weak negative correlations of r ¼ 0.12 and r ¼ 0.04 with car manufacture year (Supplementary Material Tables S15 e S18). The correlation matrix (Supplementary Material Table S9) shows the relationship among OPE concentrations in the car dust. TCEP, TCIPP and TDCIPP are weakly correlated while TDCIPP and TPHP showed a good positive correlation (r ¼ 0.43) in the car samples similar to observations in the three other microenvironments; further suggesting a peculiarity in the source of TDCIPP and TPHP in the South African indoor environment. A box plot (Fig. 2) of the concentrations of OPEs and cars grouped by manufacturer depicts a wide range in the concentrations of PFRs in cars manufactured by Honda and Audi. This observation may indicate the large volume use of OPEs (majorly TDCIPP) in some applications, such as flame retardants for flexible and rigid PUF in vehicles; the observation could also be reflective of a small sample size (i.e. n ¼ 3) for each of the two automobile manufacturers. Contrary to the observation above, the OPE concentrations showed a similar distribution pattern in cars (n ¼ 4) manufactured by Toyota. TCEP and TCIPP seemed to have a similar use pattern in cars made by Toyota, as a strong positive correlation (r ¼ 0.995) (Supplementary Material Table S10) was observed between the concentrations of TCEP and TCIPP in cars made by Toyota. Further, the cars in this study were grouped in terms of year of manufacture: the first group (n ¼ 4) reflects cars manufactured prior to 2004, that is, before the ban of penta-BDE commercial formulations in PUF; the second group are cars manufactured between 2005 and 2012 (n ¼ 15) reflecting cars manufactured after the replacement of penta-BDE with alternative flame retardants, such as OPEs in PUFs. The box plot (Fig. 2) indicates a wide variation in the concentrations of OPEs in the latter group. This could mean an increased use of OPEs in cars manufactured after 2004.
244
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
Table 3 Descriptive statistics for OPEs in indoor dust from several microenvironments in Durban, South Africa. Statistical descriptor P
n¼10 Homes Mean Median Minimum Maximum 5th percentile 95th percentile P n¼9 Offices Mean Median Minimum Maximum 5th percentile 95th percentile P n¼12 University Computer Laboratories Mean Median Minimum Maximum 5th percentile 95th percentile P n¼19 Cars Mean Median Minimum Maximum 5th percentile 95th percentile
P OPEs
Concentrations of OPEs/ng g1 TCEP
TCIPP
TDCIPP
TPHP
97650 9765 7390 2400 34250 3723 24013 88690 9854 10820 6900 11720 7364 11624 109750 9145 8420 5040 18350 5310 15050 667410 35126 10200 2000 245230 4664 131173
47860 4786 3545 220 12870 621 10665 38880 4320 3810 1240 10320 1288 9484 48420 4035 3485 830 11350 951 8435 191310 10069 5000 770 56250 2417 35424
327510 32751 7695 740 226200 965 147972 60690 6743 8320 1450 11680 1778 11312 58520 4877 3415
47510 4751 2025 760 26120 1062 16738 65420 7269 3730 650 18370 974 18042 49250 4104 3580 1420 9990 1503 8681 177210 9327 6170 670 34100 823 31562
520530 52053 22490 4790 299440 7850 194968 253680 28187 26930 14920 49210 15280 46210 265940 22162 19565 10070 43390 11676 37764 2730600 143716 49010 14020 773590 21625 431176
3.5. Comparison of OPE levels in different microenvironments in South Africa with levels reported in other countries The target OPEs (with the exception of TDCIPP, which was below the detection limit in one car and one university computer laboratory sample) were detected in all samples from the various microenvironments. The Kruskal-Wallis test showed statistical differences (p ¼ 0.003) in the concentrations of the OPEs in the various microenvironments. A comparison of the median concentrations and range (given in parentheses) of OPEs in the studied microenvironments with data reported in different countries is shown in Table 4. Similar to data from the USA (Fang et al., 2013), Germany (Brommer et al., 2012), and the Netherlands (Brandsma et al., 2014), cars contain the bulk of the OPEs compared with the other indoor microenvironments. TDCIPP dominates the phosphorus flame retardants contaminating South Africa's indoor environments in line with the profile in many countries worldwide (Table 4). 3.6. Implications for human exposure via accidental dust ingestion
Fig. 2. Box plot for OPE concentrations and (a) car manufacturer, and (b) cars grouped into periods of manufacture.
In evaluating human exposure via dust ingestion of contaminants, we assumed 100% absorption of OPEs from ingested dust in accordance with other studies (Abdallah and Covaci, 2014; Stapelton et al., 2009; Ali et al., 2013). Average dust intake rates of 20 and 50 mg d1 and high dust intakes of 50 and 200 mg d1 for adults, teenagers and toddlers were used as reported by Ali et al. (2013). Body weights of 70 kg and 12 kg were used for adults and toddlers respectively (Ali et al., 2013), and 52 kg for teenagers (Johnson-Restrepo and Kannan, 2009). Questionnaires were used to obtain the average number of hours spent per day by adults in offices, cars and homes. The average number of hours students spend in computer laboratories (for lectures, studies and assignments) and in their residences per day were obtained by interview. The amount of time spent per day in homes by toddlers (79.9%) was
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
245
Table 4 Comparison of OPE concentrations in the present study with other studies worldwide. Country
Microenvironment
Median Concentrations/ng g1 (range in parenthesis) TCEP
South Africa
Homes (n ¼ 10)
TCIPP
Reference TPHP
China USA USA USA USA Sweden Sweden Sweden Philippines Germany Germany Germany Germany Norway Egypt
2025 (760 e26120) Offices (n ¼ 9) 10820 (6900 3810 (1240e10320) 8320 (1450e11680) 3730 (650 e11720) e18730) University computer laboratories 8420 (5040e18350) 3485 (830e11350) 3415 (e 11450) 3580 (1420 (n ¼ 12) e9990) Cars(n ¼ 19) 10200 (2000 5000 (770e56250) 12770 (e 697100) 6170 (670 e245230) e34100) Homes (n ¼ 10) 2100 (e 33000) 1600 (700e11000) 10000 (2200e27000) 1200 (100e4200) Daycare centres (n ¼ 10) 30000 (2500 3100 (800e12000) 9100 (3900e150000) 1900 (300 e150000) e17000) Workplaces (n ¼ 10) 6700 (1300 19000 (3400 17000 (3300e91000) 5300 (900 e260000) e120000) e32000) House dust around electronics (n ¼ 8) 1300 (220e6900) 1300 (480e3800) 280 (70e3200) 820 (680e11000) House dust on electronics (n ¼ 8) 880 (520e2200) 1300 (580e4500) 680 (100e7400) 6500 (1600 e21000) Car dashboards (n ¼ 8) 2800 (1100e5700) 5700 (1800 17000 (6000e150000) 1700 (360 e110000) e14000) Car seats (n ¼ 8) 600 (240e5600) 4300 (1400 110000 (3800 2400 (670 e110000) e1100000) e43000) Hotels (n ¼ 8) (82e2300) (1000e9800) (69e18000) (110e2600) Houses (n ¼ 148) 5800 8700 2800 4500 Houses (n ¼ 48) <650 740 <590 870 Homes (n ¼ 15) 1140 (131e10400) 2290 (945e12300) 502 (
Egypt
Offices (n ¼ 20)
31 (
80 (
Egypt
Cars (n ¼ 20)
127 (
291 (
135 (26e1872)
Egypt
Public (n ¼ 11)
234 (
232 (
416 (
629 (116e2357)
Egypt Spain Belgium Canada Canada Romania Kuwait Kuwait Pakistan Pakistan Turkey United Kingdom United Kingdom United Kingdom United Kingdom
Houses (n ¼ 17) Houses (n ¼ 8) Houses (n ¼ 8) Houses (n ¼ 134) Houses (n ¼ 92) Houses (n ¼ 3) Houses (n ¼ 15) Cars (n ¼ 15) Houses (n ¼ 15) Cars (n ¼ 15) Houses (n ¼ 39) Living rooms (n ¼ 32)
143 (40e594) (250e9800) (75e1350) 800 2572 (136e196459) (40e1450) 710 1765
276 (41 -2018) (
191 (1.8e2628) 810
891 (136e4122) 21000
674 (0.95e16805) 710
388 (87e8957) (290e9500) (236e2640) 1700 3071 (59e8845) (105e3750) 430 1760 175 245 304 (1.7e1584) 3300
Offices (n ¼ 62)
870
33000
480
4300
Cars (n ¼ 21)
1230
53000
31000
3300
Classrooms (n ¼ 28)
860
16000
510
4100
South Africa South Africa South Africa Sweden Sweden Sweden Netherlands Netherlands Netherlands Netherlands Japan Japan Japan China China China China
LOQ refers to limit of quantitation.
7390 (2400e34250) 3545 (220e12870)
TDCIPP 7695 (740e226200)
49 (
73 (11e337)
This study This study This study This study Bergh et al. (2011) Bergh et al. (2011) Bergh et al. (2011) Brandsma et al. (2014) Brandsma et al. (2014) Brandsma et al. (2014) Brandsma et al. (2014) Takigami et al. (2009) Araki et al. (2014) Tajima et al. (2014) Sun et al. (2019) Tao et al. (2018) Wu et al. (2016) Wu et al. (2016) Wu et al. (2016) Stapleton et al. (2009) Fang et al. (2013) Fang et al. (2013) Allen et al. (2013) Marklund et al. (2003) Marklund et al. (2003) Marklund et al. (2003) Kim et al. (2013) Brommer et al. (2012) Brommer et al. (2012) Brommer et al. (2012) Fromme et al. (2014) Cequier et al. (2014) Abdallah and Covaci (2014) Abdallah and Covaci (2014) Abdallah and Covaci (2014) Abdallah and Covaci (2014) Shoeib et al. (2019) García et al. (2007) Van den Eede et al., 2012 Fan et al. (2014) Shoeib et al. (2019) Van den Eede et al., 2012 Ali et al. (2013) Ali et al. (2013) Ali et al. (2013) Ali et al. (2013) Shoeib et al. (2019) Brommer and Harrad (2015) Brommer and Harrad (2015) Brommer and Harrad (2015) Brommer and Harrad (2015)
246
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
Table 5 Summary of adult, teenager and toddler exposure to target OPEs (ng d1) via accidental indoor dust ingestion using different exposure scenarios. OPEs
Reference dose Adult
TCEP
Exposure Scenario Adult
Toddler
154000 44000
Mean High TCIPP 560000 160000 Mean High TDCIPP 105000 30000 Mean High TPHP 490000 140000 Mean High
Teenager
Toddler
Low end Median Average High end Low end Median Average High end Low end Median Average High end 100 249 18 46 26 64 20 51
169 423 74 185 162 406 55 138
217 543 97 243 529 1323 116 289
488 1219 226 566 2258 5646 356 890
the same as that of Ali et al. (2012). Thus, the average time an adult spends in the office, car and home were obtained as 33.3, 4.2, and 62.5%, respectively. For full-time undergraduate students (teenagers), it was estimated that they spend 54.2 and 45.8% of their time indoors in classrooms and in their residences respectively. Different exposure scenarios were calculated by using the 5th percentile (low end), median, mean and 95th percentile (high end) concentrations from homes, cars, offices and computer laboratory dusts. Thus, the daily dose of OPEs (SDED/ng d1) via dust ingestion was calculated from the following equations reported by Ali et al. (2013) with modifications. SDED/ng d1 ¼ [(CHDFH) þ (CofDFof) þ (CCDFC)]DIR/BW, for adult exposure estimation, SDED/ng d1 ¼ [(CHDFH) þ (CLDFL)]DIR/BW, for teenager exposure assessment, and SDED/ng d1 ¼ [(CHDFH) þ (CCDFC)]DIR/BW, for toddler exposure assessment, where CCD, CHD, CofD and CLD are dust concentrations in cars, homes, offices and computer laboratories (5th percentile, median, mean and 95th percentile) and FC, FH, Fof and FL are the fraction of time spent in cars, homes, offices and computer laboratories, respectively. DIR is the dust intake rate and BW is the body weight. P Table 5 shows the daily dose of individual OPEs and the OPEs among different population groups in South Africa compared with the toxicological reference dose (RfD) values for each OPE. The results indicate that the high end exposure (i.e. using the 95th percentile concentration) may cause significant health implications for toddlers. At this exposure scenario, young children ingest as P much as 36859 ng of n¼4OPEs per day following inadvertent dust ingestion. For the individual OPEs, the worst-case scenario results in daily ingestion of 4939 ng TCEP, 2002 ng TCIPP, 26978 ng TDCIPP and 2940 ng TPHP by toddlers. The worst-case daily exposure dose of TDCIPP is at par with the toxicological reference dose of 30000 ng d1 (Abdallah and Covaci, 2014). The dominance of two chlorinated phosphorus flame retardants e TDCIPP and TCEP e in exposure doses of studied OPEs in the South African population may pose significant health challenges particularly among toddlers who are exposed to a higher magnitude of these flame retardant chemicals. Studies have reported that TDCIPP is mutagenic, carcinogenic in rats and humans, and a moderate hazard for reproductive and developmental effects (Stapleton et al., 2009). Similarly, TCEP is reportedly carcinogenic for animals, neurotoxic to rats and mice, induces adverse reproductive effects in rats, and exhibits haemolytic and reproductive effects such as reduced fertility, longer estrous cycle length, and reduced sperm motility and density, in humans. TCEP was associated with the acute death of dogs after ingestion of car seat cushions containing an enormous
92 229 16 40 15 37 26 65
159 398 70 176 108 269 57 143
189 472 88 219 353 883 88 220
383 958 189 473 1476 3691 248 619
159 634 30 119 43 171 44 177
317 1267 152 608 334 1337 94 375
464 1856 212 849 1496 5983 209 838
1235 4939 500 2002 6745 26978 735 2940
amount of TCEP (van der Veen and de Boer, 2012). Several toxicological effects have also been reported for TCIPP and TPHP (van der Veen and de Boer, 2012). 4. Conclusions We have reported for the first time the presence of organophosphate esters in the South African indoor environment. The results of the study indicate the wide use of organophosphate esters, particularly the chlorinated phosphorus flame retardants, at levels comparable to those in the European Union. The abundance of TCEP in these microenvironments suggests V6 as a possible replacement of penta-BDE in polyurethane foams in South Africa. Cars contained the highest levels of OPEs among the microenvironments and TDCIPP was the most abundant OPE found among the four studied OPEs. The relationships between OPEs and some indoor characteristics implicate furniture and electronics as reservoirs of OPEs in the indoor environment, particularly in homes. The result of the present study is reflective of a shift to OPEs as flame retardants, probably in response to international regulations. Generally, the current exposure of OPEs should be of interest as TDCIPP and TCEP, which are known mutagens and carcinogens, majorly contributed to the overall exposure of the South African population, particularly young children, to OPEs. The current levels, profiles and magnitude of exposure indicate further investigations are required on other exposure routes such as dietary and inhalation; as well as patterns and human health implications of exposure to mixtures of these organophosphate esters. Acknowledgements OAA acknowledges the University of KwaZulu-Natal for a doctoral scholarship. The authors acknowledge the kind donation of a Restek RTX 1614® capillary GC column from Restek Corporation. We sincerely thank all home, office, and car owners for permitting sample collections. We also acknowledge Burchmore Motor dealership for allowing us to sample cars available for resale. The National Research Foundation of South Africa is thanked for research support. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.chemosphere.2019.04.175. References Abafe, O.A., Martincigh, B.S., 2015. Polybrominated diphenyl ethers and polychlorinated biphenyls in indoor dust in Durban, South Africa. Indoor Air 25, 547e556. Abdallah, M.A.-E., Covaci, A., 2014. Organophosphate flame retardants in indoor dust from Egypt: implications for human exposure. Environ. Sci. Technol. 48,
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247 4782e4789. Ali, N., Ali, L., Mehdi, T., Dirtu, A.C., Al-Shammari, F., Neels, H., Covaci, A., 2013. Levels and profiles of organochlorines and flame retardants in car and house dust from Kuwait and Pakistan: implication for human exposure via dust ingestion. Environ. Int. 55, 62e70. Ali, N., Dirtu, A.C., Eede, N.V.D., Goosey, E., Harrad, S., Neels, H., T Mannetje, A., Coakley, J., Douwes, J., Covaci, A., 2012. Occurrence of alternative flame retardants in indoor dust from New Zealand: indoor sources and human exposure assessment. Chemosphere 88, 1276e1282. Allen, J., Stapleton, H., Vallarino, J., McNeely, E., McClean, M., Harrad, S., Rauert, C., Spengler, J., 2013. Exposure to flame retardant chemicals on commercial airplanes. Environ. Health 12, 17. Andresen, J., Bester, K., 2006. Elimination of organophosphate ester flame retardants and plasticizers in drinking water purification. Water Res. 40, 621e629. Araki, A., Saito, I., Kanazawa, A., Morimoto, K., Nakayama, K., Shibata, E., Tanaka, M., Takigawa, T., Yoshimura, T., Chikara, H., Saijo, Y., Kishi, R., 2014. Phosphorus flame retardants in indoor dust and their relation to asthma and allergies of inhabitants. Indoor Air 24, 3e15. , A., 2007. Liquid Bacaloni, A., Cavaliere, C., Foglia, P., Nazzari, M., Samperi, R., Lagana chromatography/tandem mass spectrometry determination of organophosphorus flame retardants and plasticizers in drinking and surface waters. Rapid Commun. Mass Spectrom. 21, 1123e1130. € Bergh, C., Torgrip, R., Emenius, G., Ostman, C., 2011. Organophosphate and phthalate esters in air and settled dust e a multi-location indoor study. Indoor Air 21, 67e76. € rklund, J., Isetun, S., Nilsson, U., 2004. Selective determination of organophosBjo phate flame retardants and plasticizers in indoor air by gas chromatography, positive-ion chemical ionization and collision-induced dissociation mass spectrometry. Rapid Commun. Mass Spectrom. 18, 3079e3083. Brandsma, S.H., de Boer, J., Leonards, P.E.G., Cofino, W.P., Covaci, A., Leonards, P.E.G., 2013. Organophosphorus flame-retardant and plasticizer analysis, including recommendations from the first worldwide interlaboratory study. Trac. Trends Anal. Chem. 43, 217e228. Brandsma, S.H., de Boer, J., van Velzen, M.J.M., Leonards, P.E.G., 2014. Organophosphorus flame retardants (PFRs) and plasticizers in house and car dust and the influence of electronic equipment. Chemosphere 116, 3e9. Brommer, S., Harrad, S., 2015. Sources and human exposure implications of concentrations of organophosphate flame retardants in dust from UK cars, classrooms, living rooms, and offices. Environ. Int. 83, 202e207. Brommer, S., Harrad, S., Van den Eede, N., Covaci, A., 2012. Concentrations of organophosphate esters and brominated flame retardants in German indoor dust samples. J. Environ. Monit. 14, 2482e2487. € Campone, L., Piccinelli, A., Ostman, C., Rastrelli, L., 2010. Determination of organophosphorous flame retardants in fish tissues by matrix solid-phase dispersion and gas chromatography. Anal. Bioanal. Chem. 397, 799e806. € Carlsson, H., Nilsson, U., Becker, G., Ostman, C., 1997. Organophosphate ester flame retardants and plasticizers in the indoor environment: Analytical methodology and occurrence. Environ. Sci. Technol. 31, 2931e2936. , R.M., Becher, G., Thomsen, C., 2014. Cequier, E., Ionas, A.C., Covaci, A., Marce Occurrence of a broad range of legacy and emerging flame retardants in indoor environments in Norway. Environ. Sci. Technol. 48, 6827e6835. Fan, X., Kubwabo, C., Rasmussen, P.E., Wu, F., 2014. Simultaneous determination of thirteen organophosphate esters in settled indoor house dust and a comparison between two sampling techniques. Sci. Total Environ. 491e492, 80e86. Fang, M., Webster, T.F., Gooden, D., Cooper, E.M., McClean, M.D., Carignan, C., Makey, C., Stapleton, H.M., 2013. Investigating a novel flame retardant known as V6: measurements in baby products, house dust, and car dust. Environ. Sci. Technol. 47, 4449e4454. Fromme, H., Lahrz, T., Kraft, M., Fembacher, L., Mach, C., Dietrich, S., Burkardt, R., € lkel, W., Go € en, T., 2014. Organophosphate flame retardants and plasticizers in Vo the air and dust in German daycare centers and human biomonitoring in visiting children (LUPE 3). Environ. Int. 71, 158e163. García, M., Rodríguez, I., Cela, R., 2007. Microwave-assisted extraction of organophosphate flame retardants and plasticizers from indoor dust samples. J. Chromatogr. A 1152, 280e286. Hartmann, P.C., Bürgi, D., Giger, W., 2004. Organophosphate flame retardants and plasticizers in indoor air. Chemosphere 57, 781e787. Johnson-Restrepo, B., Kannan, K., 2009. An assessment of sources and pathways of
247
human exposure to polybrominated diphenyl ethers in the United States. Chemosphere 76, 542e548. Kajiwara, N., Noma, Y., Takigami, H., 2011. Brominated and organophosphate flame retardants in selected consumer products on the Japanese market in 2008. J. Hazard Mater. 192, 1250e1259. Kim, J.-W., Isobe, T., Sudaryanto, A., Malarvannan, G., Chang, K.-H., Muto, M., Prudente, M., Tanabe, S., 2013. Organophosphorus flame retardants in house dust from the Philippines: occurrence and assessment of human exposure. Environ. Sci. Pollut. Control Ser. 20, 812e822. Marklund, A., Andersson, B., Haglund, P., 2003. Screening of organophosphorus compounds and their distribution in various indoor environments. Chemosphere 53, 1137e1146. Marklund, A., Andersson, B., Haglund, P., 2005. Organophosphorus flame retardants and plasticizers in air from various indoor environments. J. Environ. Monit. 7, 814e819. lez-Barreiro, C., Sitka, A., Scharf, S., Gans, O., 2007. Martínez-Carballo, E., Gonza Determination of selected organophosphate esters in the aquatic environment of Austria. Sci. Total Environ. 388, 290e299. Meeker, J., Stapleton, H., 2009. House dust concentrations of organophosphate flame retardants in relation to hormone levels and semen quality parameters. Environ. Health Perspect. 118, 318e323. Salamova, A., Ma, Y., Venier, M., Hites, R.A., 2014. High levels of organophosphate flame retardants in the Great Lakes atmosphere. Environ. Sci. Technol. Lett. 1, 8e14. Shoeib, T., Webster, G.M., Hassan, Y., Tepe, S., Yalcin, M., Turgut, C., KurtKarakus¸, P.B., Jantunen, L., 2019. Organophosphate esters in house dust: a comparative study between Canada, Turkey and Egypt. Sci. Total Environ. 650, 193e201. € Staaf, T., Ostman, C., 2005. Indoor air sampling of organophosphate triesters using solid phase extraction (SPE) adsorbents. J. Environ. Monit. 7, 344e348. Stapleton, H., Klosterhaus, S., Eagle, S., Fuh, J., Meeker, J., Blum, A., Webster, T., 2009. Detection of organophosphate flame retardants in furniture foam and U.S. house dust. Environ. Sci. Technol. 43, 7490e7495. Sun, Y., Liu, L.-Y., Sverko, E., Li, Y.-F., Li, H.-L., Huo, C.-Y., Ma, W.-L., Song, W.-W., Zhang, Z.-F., 2019. Organophosphate flame retardants in college dormitory dust of northern Chinese cities: occurrence, human exposure and risk assessment. Sci. Total Environ. 665, 731e738. Sundkvist, A.M., Olofsson, U., Haglund, P., 2010. Organophosphorus flame retardants and plasticizers in marine and fresh water biota and in human milk. J. Environ. Monit. 12, 943e951. Tajima, S., Araki, A., Kawai, T., Tsuboi, T., Ait Bamai, Y., Yoshioka, E., Kanazawa, A., Cong, S., Kishi, R., 2014. Detection and intake assessment of organophosphate flame retardants in house dust in Japanese dwellings. Sci. Total Environ. 478, 190e199. Takigami, H., Suzuki, G., Hirai, Y., Sakai, S.-I., 2009. Brominated flame retardants and other polyhalogenated compounds in indoor air and dust from two houses in Japan. Chemosphere 76, 270e277. Tao, Y., Shang, Y., Li, J., Feng, J., He, Z., Covaci, A., Wang, P., Luo, J., Mao, X., Shi, B., Hu, L., Luo, D., Mei, S., 2018. Exposure to organophosphate flame retardants of hotel room attendants in Wuhan City, China. Environ. Pollut. 236, 626e633. Thomsen, V., Schatzlein, D., Mercuro, D., 2003. Limits of detection in spectroscopy. Spectroscopy 18, 112e114. €ck, J., Tamburro, D., Crescenzi, C., Carlsson, H., 2006. Air sampling with Tollba Empore solid phase extraction membranes and online single-channel desorption/liquid chromatography/mass spectrometry analysis: determination of volatile and semi-volatile organophosphate esters. J. Chromatogr. A 1129, 1e8. Van den Eede, N., Dirtu, A., Ali, N., Neels, H., Covaci, A., 2012. Multi-residue method for the determination of brominated and organophosphate flame retardants in indoor dust. Talanta 89, 292e300. Van den Eede, N., Dirtu, A.C., Neels, H., Covaci, A., 2011. Analytical developments and preliminary assessment of human exposure to organophosphate flame retardants from indoor dust. Environ. Int. 37, 454e461. van der Veen, I., de Boer, J., 2012. Phosphorus flame retardants: properties, production, environmental occurrence, toxicity and analysis. Chemosphere 88, 1119e1153. Wu, M., Yu, G., Cao, Z., Wu, D., Liu, K., Deng, S., Huang, J., Wang, B., Wang, Y., 2016. Characterization and human exposure assessment of organophosphate flame retardants in indoor dust from several microenvironments of Beijing, China. Chemosphere 150, 465e471.
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Concentrations, sources and human exposure implications of organophosphate esters in indoor dust from South Africa Ovokeroye A. Abafe 1, Bice S. Martincigh* School of Chemistry and Physics, University of KwaZulu-Natal, Westville Campus, Private Bag X54001, Durban, 4000, South Africa
h i g h l i g h t s Organophosphate ester (OPE) concentrations in 50 indoor dust samples were measured. Results indicate widespread use of OPEs in Durban, South Africa. Cars contained the highest levels of OPEs. TDCIPP and TCEP, known carcinogens, contributed majorly to human exposure levels. This is the first report of OPEs in the indoor environment in South Africa.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 February 2019 Received in revised form 15 April 2019 Accepted 21 April 2019 Available online 3 May 2019
The concentrations of four organophosphate esters (OPEs) were measured in 50 dust samples from homes (n ¼ 10), offices (n ¼ 9), university computer laboratories (n ¼ 12) and cars (n ¼ 19) in Durban, South Africa. The median concentrations Sn¼4 OPEs were 22940, 26930, 19565 and 49010 ng g⁻1 in homes, offices, university computer laboratories and cars respectively. OPEs were detected in all samples with the exception of one car and one computer laboratory sample in which TDCIPP was not detected. Significant association of indoor characteristics with OPE concentrations was observed. OPEs positively correlated (r ¼ 0.22, p value ¼ 0.4862) with electronics and correlated (r ¼ 0.522, p value ¼ 0.0675) with foams and furniture in homes. By employing the median concentrations and an average dust intake rate, the exposure doses (ng d1) were found to be 169 (TCEP), 74 (TCIPP), 162 (TDCIPP) and 55 (TPHP) for adults; 159 (TCEP), 70 (TCIPP), 108 (TDCIPP) and 57 (TPHP) for teenagers; 317 (TCEP), 152 (TCIPP), 334 (TDCIPP) and 94 (TPHP) for toddlers. The predominance and exposure magnitude of OPEs in the South African environment require further investigations to determine cumulative human health effects arising from mixtures of these compounds through multiple exposure routes. © 2019 Elsevier Ltd. All rights reserved.
Handling Editor: Jerzy Falandysz Keywords: Organophosphate esters Phosphorus flame retardants Plasticizers Indoor dust Sources Human exposure South Africa
1. Introduction Organophophate esters (OPEs) are used in a variety of applications. The chlorinated alkylphosphates, such as, tris(1-chloro-2propyl) phosphate (TCIPP), tris(1,3-dichloro-2-propyl) phosphate (TDCIPP) and tris(2-chloroethyl) phosphate (TCEP), are used mainly as phosphorus flame retardants (PFRs) in polyurethane foams (Andresen and Bester, 2006), electronic equipment, textiles, plastic
* Corresponding author. E-mail address: [email protected] (B.S. Martincigh). 1 Present address: Residue Laboratory, Agricultural Research CouncilOnderstepoort Veterinary Research, 100 Old Soutpan Road, Onderstepoort, Pretoria, 0110, South Africa. https://doi.org/10.1016/j.chemosphere.2019.04.175 0045-6535/© 2019 Elsevier Ltd. All rights reserved.
and building materials (Salamova et al., 2014), whilst the nonderivatized organophosphates, such as triphenyl phosphate (TPHP), are majorly used as plasticizers, lubricants, varnishes, glues, airplane hydraulic fluids and to regulate pore sizes, for example, in concrete (Brandsma et al., 2014). They are sometimes used as substitute flame retardants for the halogenated compounds, for example, the use of TPHP in electronic devices (Andresen and Bester, 2006). There has been a great increase in the use of PFRs to meet fire safety regulations in consumer products since restrictions were imposed on the use of polybrominated diphenyl ethers (PBDEs), in particular, penta-, octa- and deca-BDE technical mixtures (Brandsma et al., 2014). Since PFRs are not covalently bound to materials but are used as additive flame retardants, they may offgas and leach from products through volatilization and/or
240
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
abrasion into the environment (Brandsma et al., 2014; Abdallah and Covaci, 2014). Little is known of the toxicity of OPEs, however, studies have reported harmful effects of OPEs to include altered hormone levels and decreased sperm concentrations (Meeker and Stapleton, 2009); neurotoxic, mutagenic and carcinogenic effects in rats and mice; haemolytic and reproductive effects in humans (Van Der Veen and De Boer, 2012); and potential human carcinogenicity (Abdallah and Covaci, 2014). PFRs have been detected in various environmental media including air (Marklund et al., 2003, 2005; Carlsson et al., 1997; €ck et al., 2006; Bjo € rklund et al., 2004; Hartmann et al., 2004); Tollba surface and drinking waters, and sediments (Martínez-Carballo et al., 2007; Bacaloni et al., 2007); biota (Campone et al., 2010; Sundkvist et al., 2010) and indoor dust of various microenvironments in several locations worldwide (Brandsma et al., 2014; Abdallah and Covaci, 2014; Marklund et al., 2003; Brommer et al., 2012; Kajiwara et al., 2011; Bergh et al., 2011; Stapleton et al., 2009; Wu et al., 2016). Much attention has been given to the significance of indoor dust as a pathway of human exposure to OPEs. The relationship between dust and human body burdens is strongly implied by the association of PFRs in household dusts and human semen quality and hormone levels (Meeker and Stapleton, 2009). Nothing is known on the production, use, distribution and fate of OPEs in South Africa. Moreover, despite the increasing proof of the significant implications of indoor dust for human exposure to OPEs, attempts to link indoor contaminants with probable source items has had limited success. A dearth of information also exists for human exposure and pathways to OPEs. To bridge these gaps, the aim of the present study was to: (i) investigate the contamination of indoor dust by four OPEs (TCEP, TCIPP, TDCIPP and TPHP) in multiple indoor environments in South Africa as no data is currently available, (ii) compare the profiles of OPEs in the different microenvironments (cars, homes, offices and university computer laboratories), (iii) establish the relationships between various household products and the concentrations of OPEs in dust in order to identify their possible sources in the indoor environment, and (iv) estimate the exposure magnitude of OPEs among different population groups, by utilizing various exposure scenarios. 2. Material and methods 2.1. Chemicals Analytical standards (of known purity) of TCEP, TCIPP, TDCIPP, and TPHP were purchased from Sigma-Aldrich, South Africa. The internal standard, 13C12elabelled decachlorobiphenyl (13C12 PCB209), was obtained from Wellington Laboratories, Guelph, Ontario, Canada. Anhydrous sodium sulfate was from Associated Chemical Enterprises (ACE), Johannesburg, South Africa. Silica gel 90 was from Sigma-Aldrich, South Africa. A Restek Rtx®-1614 fused silica (5% diphenyl 95% dimethyl polysiloxane) capillary column was obtained as a gift from Restek Corporation, Bellefonte, PA, USA. All solvents were high performance liquid chromatography grade purchased from Sigma-Aldrich, South Africa. 2.2. Sampling A total of 50 dust samples were collected from homes (n ¼ 10), university students’ computer laboratories (n ¼ 12), and university staff offices (n ¼ 9), between August and October 2012 in Durban, South Africa. Similarly, dust samples (n ¼ 19) were collected between January and March 2013 from personal and previously owned cars available for resale. The previously owned cars were
sampled at a dealership in Durban, South Africa. All the vehicles from the dealership had been through a thorough cleaning process on arrival at the dealership prior to resale. Personal cars sampled had not undergone any form of cleaning at least three days before sampling. Information on the sampling protocol can be found elsewhere (Abafe and Martincigh, 2015). Detailed questionnaires were used to obtain pertinent information on cars, homes, offices and computer laboratories. This information included location, time since floor was last vacuumed, type of ventilation and flooring, and the number and types of electronic/electrical devices and furniture, model and manufacturer of car, year of manufacture, etc. Interviews were also conducted to obtain further information on building ages and to determine if, and when, any renovations were carried out. Since only 50 and 89% of the questionnaires were returned for homes and offices, respectively, correlation analyses were based on those samples for which questionnaires were returned. 2.3. Extraction and clean-up Samples were extracted following a previously reported method (Abafe and Martincigh, 2015) with slight modifications. Briefly, approximately 1.0 g of each sample was quantitatively weighed into a glass test tube and spiked with 2 mg 13C12 PCB-209. A volume of 12 mL n-hexane:methanol (1:3 v/v) was added. Samples were mixed in an orbital shaker for 15 min and then extracted in an ultrasonic water bath at 40 C for 30 min. The mixing and extraction was repeated for a second time without addition of fresh solvent. The samples were then centrifuged at 3500 rpm for 10 min and the supernatants were stored at <4 C prior to clean-up. Representative samples (n ¼ 5) from each microenvironment were analysed in triplicate. For the clean-up of extracts, silica gel column chromatography was used. Silica gel was activated at 130 C for 16 h and anhydrous sodium sulfate was baked at 450 C for 5 h before use. Silica gel and anhydrous sodium sulfate were subsequently cooled in a desiccator. A 30 cm 1 cm glass column was packed with 4 g deactivated silica gel. Each column was topped with 2 g of anhydrous sodium sulfate and then 30 mL of the extraction solvent was passed through the column. Extracts were loaded onto columns just before the exposure of the sodium sulfate layer. OPEs were eluted with 30 mL nhexane. This was kept as fraction 1; columns were further eluted with 30 mL of diethyl ether:n-hexane (50:50 v/v), and kept as fraction 2; both fractions were mixed together. Finally, columns were eluted with 30 mL acetone:dichloromethane (1:1 v/v) as fraction 3. The column flow rates were maintained at approximately 0.5 mL min1. Eluates were reduced in volume and concentrated to approximately 250 mL in a rotary evaporator at 55 C and stored in 1.5 mL amber glass GC/MS vials. All extracts were stored at <4 C until instrumental analysis. 2.4. GC-EI/MS analysis The analysis of OPEs was performed on an Agilent 6890 (Palo Alto, CA, USA) gas chromatograph (GC), coupled to a 5973 N series mass spectrometer (MS) operated in the electron impact (EI) ionization mode. A Restek Rtx® e 1614 fused silica (5% diphenyl, 95% dimethyl polysiloxane) capillary column (15 m 250 mm 0.1 mm) was used to effect separation, and the MS was operated in the selected ion monitoring (SIM) mode. The injections were made in the pulsed splitless mode with the injector temperature set at 250 C. The injection volume was 2 mL. The GC oven temperature programme started at 90 C (held for 2 min), then increased at 20 C min1 to 270 C and held for 1 min, and finally increased at 10 C min⁻1 to 290 C and held for a minute. Helium was used as the
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
carrier gas at a flow rate of 1.2 mL min1 and a constant linear velocity of 37 cm s1. For the MS, the ion source and transfer line temperatures were 230 and 350 C, respectively; and the ionization energy was 70 eV. OPE mass spectra were obtained in full scan mode to select prominent ions that were utilized in the SIM mode (Table 1). Quantification was carried out by means of a multiple point internal standard method. 13C12-labelled PCB-209 was employed as an internal standard for all OPEs studied. The response factors were determined from the slope of a plot of the ratio of peak areas against the ratio of the concentrations. The values for the plots were obtained from a 5 to 6 point triplicate analysis of the OPE standard solutions diluted to fall within a concentration range of 2e10 mg mL1. 2.5. Quality control/assurance The recovery for OPEs in dust was determined from spiked anhydrous sodium sulfate at different spike concentrations (Table 2). Samples were left to stand for at least 21 days at 10 C. Spiked samples were extracted and cleaned-up following the procedure for real samples. Method blanks were analysed with every batch of ten samples. For the method blank, dust samples were replaced with anhydrous sodium sulfate and passed through all the analytical procedure carried out for real samples. TPHP concentrations in samples were corrected for the blank concentrations as values found in the method blank (n ¼ 5) were as high as 1.6% of the concentrations in samples. Solvent blanks were injected after the analysis of at most three samples. All glassware was cleaned with laboratory wash solutions, rinsed with distilled water and then with organic solvents. Non-volumetric glassware was oven-dried prior to use. Direct ultraviolet light and plasticware was avoided throughout the analysis. The instrument responses to pure OPE analytical standards employed for calibrations were linear with r2 > 0.99 for all OPEs studied. Limits of detection (LODs) ranged from 0.6 ng g1 for TCEP to 100 ng g1 for TPHP; limits of quantitation (LOQs) were in the range 1.8 ng g1 (TCEP) and 301 ng g1 (TPHP) (see Table 1). Matrix spike recovery samples were prepared by adding 200e10000 ng g1 of each OPE to glass test tubes filled with 1.0 g anhydrous sodium sulfate for each concentration listed in Table 2. Matrix spikes were subjected to all analytical protocols employed for real samples. The average recovery and the percentage relative standard deviation (%RSD) for each of the OPEs at the respective spiked concentrations are shown in Table 2. 2.6. Statistics The distribution of OPE concentrations in the different microenvironments was tested with the Shapiro-Wilk test of normality. Descriptive statistics such as sum, mean, median, minimum, maximum and parametric statistics such as t-test and analysis of variance (ANOVA) were calculated by using Microsoft Excel® 2010.
241
Non-parametric statistics such as Wilcoxon Signed-Ranks test, Kendall tau test and Spearman rank correlation were performed with Analyse-it® software in Microsoft Excel 2010. The KruskalWallis test was employed to test for differences in location by using XLSTAT 2014 software. Principal component analysis (PCA) was performed with SIMCA version 13 statistical software. Limits of detection (LODs) and quantitation (LOQs) were estimated following Thomsen et al. (2003). Samples below the detection limit were treated as zero throughout the statistical analysis. 3. Results and discussion 3.1. Concentrations of OPEs in house dust The four organophosphate esters were detected in all the dust samples collected from homes in South Africa (Fig. 1). A full list of the OPE concentrations in home samples is presented in Supplementary Material Table S1. Descriptive statistics for the distribution of these OPEs are summarized in Table 3. The strong positive correlation of OPEs in this microenvironment (Supplementary Material Table S5) is indicative of similar sources for the PFRs in South African homes. Nothing is known on the production and use of organophosphate flame retardants in South Africa. However, due to industrialization in South Africa comparable to most countries in the European Union, and because of the very high levels of these OPEs found in indoor dust of two industrial sites e a textile and polyurethane industry, we hypothesize the possible indoor contamination of OPEs from the products of these industries and others which eventually end up indoors. The OPE profile in the house dust was TDCIPP > TCEP > TCIPP > TPHP. The distribution pattern of OPEs in this study is similar to those reported in Swedish homes (Bergh et al., 2011), German homes (Brommer et al., 2012) and Egyptian homes (Abdallah and Covaci, 2014); in which TDCIPP dominated in the house dust. Despite the replacement of TCEP by TCIPP, the house dust concentrations of TCEP in this study overwhelmed those of TCIPP, and TCEP was found to be the second most abundant OPE in South African homes similar to the observations of Bergh et al. (2011) in Swedish homes. However, the concentrations of TCEP strongly correlated (r ¼ 0.82) with TCIPP in this microenvironment (Supplementary Material Table S5). A statistically significant difference (p ¼ 0.022) was observed among the concentrations of the individual OPEs in South African homes. The high levels of TCEP in these house dust samples suggests a high volume use of this PFR and possibly the domineering Chinese flame retardant V6 [2,2ebis (chloromethyl)propane-1,3-diyltetrakis(2-chloroethyl)biphosphate] which is applied to polyurethane foam commonly present in furniture and car foam (Fang et al., 2013). The high vapour pressure (1.1 10⁻4 mmHg) of TCEP is likely responsible for its easy migration from treated products during the product's useful life. The elevated TCEP concentrations in South African homes should be of particular interest due to the reported carcinogenicity of TCEP (Fang et al., 2013; Stapleton et al., 2009). The OPE concentrations
Table 1 Qualifying and quantitation ions, r2, limits of detection and quantitation for TCEP, TCIPP, TDCIPP and TPHP. Analyte
Qualifier and quantifier ion
R2
LODa/ng g1
LOQa/ng g1
TCEP TCIPP TDCIPP TPHP 13 C12-PCB 209b
205, 201, 321, 326, 510
0.999 0.998 0.995 0.992
0.6 51 91 100
1.8 152 274 301
a b
Thomsen et al. (2003). Internal standard.
249 277 381 325
242
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
Table 2 Recovery of individual OPEs. Analyte
Spiked concentration (n ¼ 4)/mg g1
Average concentration found (n ¼ 4)/mg g1
Standard deviation
Recovery/%
Relative standard deviation/%
TCEP
0.20 1.00 2.00 6.00 0.20 1.00 2.00 6.00 0.60 2.00 10.0 0.60 2.00 10.0
0.19 1.22 1.74 5.16 0.19 0.74 1.63 5.92 0.57 2.71 12.87 0.61 2.03 10.79
0.05 0.002 0.03 0.44 0.09 0.10 0.15 0.22 0.04 0.76 0.82 0.10 0.05 0.16
95 122 87 86 95 74 82 99 95 136 129 102 102 108
25 0.2 2 8.5 47 14 9 4 7 28 6 16 3 2
TCIPP
TDCIPP
TPHP
did not correlate well (r ¼ 0.04) with the PBDE concentrations in the corresponding samples reported earlier (Abafe and Martincigh, 2015). This suggests differences in sources and/or release mechanisms of these flame retardants in the house dust samples. The P correlations of OPE concentrations in dust and the number of electronics (r ¼ 0.22, p value of 0.4862) and the number of foams and furniture (r ¼ 0.522, p value of 0.0675) in these homes, strongly implicate these household items as sources of OPEs.
3.2. Concentrations of OPEs in office dust All four OPEs were detected in 100% of the dust samples from offices in South Africa (Fig. 1). A full list of the concentrations of the individual OPEs in this microenvironment can be found in Supplementary Material Table S2 and the descriptive statistics are given in Table 3. Unlike the distribution of OPEs in home samples where TDCIPP was predominant, TCEP was most abundant in the office dust samples. The relative abundance of OPEs were TCEP > TPHP > TDCIPP > TCIPP. Limited data are available on indoor dust TCEP concentrations worldwide. The concentrations of TCEP were higher than the reported concentrations in two offices in Sweden (i.e. the minimum concentrations of TCEP in the South African offices were much higher than the maximum concentrations of TCEP reported in the Swedish offices) (Marklund et al., 2003). The high levels of TCEP in the office dust could be associated with indoor characteristics; however, there were no correlations between the concentrations of OPEs and office characteristics such as electronics or furniture in these offices. It should be noted that the Swedish study by Marklund et al. (2003) was conducted prior to the ban of commercial penta- and octa-PBDE mixtures, which had wider application in foam and furniture compared with organophosphate flame retardants. The abundance of TCEP and TPHP may be indicative of the possible use of V6 and Firemaster 550 (FM550) in flame retarding various office products in South Africa. TCEP is a component of V6 and TPHP is a major component of FM550 (Stapleton et al., 2009). Fang et al. (2013) showed a significant relationship between V6 and TCEP in dust samples and concluded that V6 is an important source of TCEP in the environment. This observation is supported by the poor correlation of TCEP with the other studied OPEs, whilst TCIPP and TDCIPP concentrations are correlated (r ¼ 0.49); similar to the strong correlation (r ¼ 0.73) of TDCIPP and TPHP concentrations in the offices (Supplementary Material Table S7). These relationships suggest similar sources for TCIPP, TDCIPP and TPHP in the office dust samples. We cannot exclusively associate the presence of TPHP in this microenvironment to the flame retardant FM550 since TPHP has other
applications as a plasticizer in various indoor products which may also account for a source of TPHP in the office dust (Brandsma et al., 2014).
3.3. Concentrations of OPEs in university computer laboratories The percentage contribution of individual OPEs in this microenvironment is presented in Fig. 1. The descriptive statistics of OPEs in the university computer laboratories are presented in Table 3. As observed for OPEs in offices, TCEP was the major OPE found in these laboratories. The concentration profiles of the OPEs were in the order: TCEP > TDCIPP > TPHP > TCIPP. The dominance of TCEP in this microenvironment is not surprising as a previous study had reported high levels of TCEP in the indoor air of a lecture hall with computers as opposed to a hall without computers in which TCEP was not detected in the same study (Staaf and Ostman, 2005). The authors also reported TCEP as high as 10 ng m-3 in air of an electronic dismantling facility (Staaf and Ostman, 2005). The concentration of TCEP in these computer laboratories calls for concern owing to the fact that these microenvironments were refurbished between 2006 and 2007. All carpeting, furniture, blinds, computers and printers were replaced and the dividing walls in some of the rooms were also replaced with plasterboard. Although no report on OPEs in a similar microenvironment is available in the literature, the levels of TCEP in this study exceeded the levels reported in dust from homes in New Zealand (Ali et al., 2012) and dust from Belgian homes (Van den Eede et al., 2011). All OPEs were normally distributed in this microenvironment. The correlation matrix (Supplementary Material Table S8) indicates a good relationship between the concentrations of the studied OPEs. However, the concentrations of TCEP negatively correlated with those of TPHP. The correlation of TDCIPP and TPHP concentrations in the computer laboratories resembles that of the offices; hence providing further evidence for source similarities of both OPEs in these microenvironments. The concentrations of TPHP in these microenvironments are not surprising as TPHP concentrations in excess of 500 000 ng g1 have been reported in liquid crystal display (LCD) televisions and laptop computers in Japan (Kajiwara et al., 2011). In the same light, Brandsma et al. (2014) showed high contamination levels of TPHP in dust samples collected from electronic equipment consistent with the observations of Kajiwara et al. (2011) in which TPHP was the major OPE detected in electronic equipment from the Japanese market. TPHP is used as a flame retardant and plasticizer in a variety of products, which may also account for some of the sources of TPHP in this microenvironment. The high volatilities of TDCIPP and TPHP are most likely responsible for their concentrations in the indoor dust since they can easily be released from
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
Fig. 1. Percentage contribution of individual OPEs in dust collected from (a) homes, (b) offices, (c) university computer laboratories and (d) cars in South Africa.
products into the environment (van der Veen and de Boer, 2012). 3.4. Concentrations of OPEs in cars All four OPEs were detected in the cars sampled for this study with the exception of one car in which TDCIPP was not detected.
243
Fig. 1 shows the percentage contribution of each OPE in the car dust samples in South Africa. A full list of the concentrations of OPEs in this microenvironment can be found in Supplementary Material Table S4. The concentration profiles of OPEs in this microenvironment are similar to those of the house dust. The abundance of OPEs in these cars is in the order: TDCIPP > TCEP > TCIPP > TPHP. A summary of the descriptive statistics for OPEs in cars is given in Table 3. The concentration profiles of TDCIPP in these cars are similar to concentrations reported in car dust from the Netherlands (Brandsma et al., 2013) and Germany (Brommer et al., 2012). The high TDCIPP concentrations in car dust are in tandem with its use as a flame retardant in flexible and rigid polyurethane foams (PUF) (Marklund et al., 2003). Unlike the observations of Brommer et al. (2012), in which high TDCIPP concentrations were associated with older cars, the concentrations of TDCIPP did not show a particular trend with the year of manufacture of the vehicle in South Africa. The highest TDCIPP concentration reported in this study was found to be 697 100 ng g1 in a car manufactured in 2009. However, a similar car from the same manufacturer, of the same model and year, had a TDCIPP concentration of only 2600 ng g1. The cause of this discrepancy could not be ascertained, since information obtained showed that both cars were cleaned at approximately the same time (i.e. 90 days) prior to the date of sampling. However, Brommer et al. (2012) postulated that intensive car usage leads to greater abrasion of vehicle upholstery with an attendant increase in the release of flame retarded upholstery fabric fibres, which may be accountable together with external sources for the variations of TDCIPP levels in the two cars. No direct relationship was found between OPE concentrations and year of car manufacture. Whilst TDCIPP and TPHP showed weak positive correlations of r ¼ 0.14 and r ¼ 0.15 with car manufacture year; TCEP and TCIPP only showed weak negative correlations of r ¼ 0.12 and r ¼ 0.04 with car manufacture year (Supplementary Material Tables S15 e S18). The correlation matrix (Supplementary Material Table S9) shows the relationship among OPE concentrations in the car dust. TCEP, TCIPP and TDCIPP are weakly correlated while TDCIPP and TPHP showed a good positive correlation (r ¼ 0.43) in the car samples similar to observations in the three other microenvironments; further suggesting a peculiarity in the source of TDCIPP and TPHP in the South African indoor environment. A box plot (Fig. 2) of the concentrations of OPEs and cars grouped by manufacturer depicts a wide range in the concentrations of PFRs in cars manufactured by Honda and Audi. This observation may indicate the large volume use of OPEs (majorly TDCIPP) in some applications, such as flame retardants for flexible and rigid PUF in vehicles; the observation could also be reflective of a small sample size (i.e. n ¼ 3) for each of the two automobile manufacturers. Contrary to the observation above, the OPE concentrations showed a similar distribution pattern in cars (n ¼ 4) manufactured by Toyota. TCEP and TCIPP seemed to have a similar use pattern in cars made by Toyota, as a strong positive correlation (r ¼ 0.995) (Supplementary Material Table S10) was observed between the concentrations of TCEP and TCIPP in cars made by Toyota. Further, the cars in this study were grouped in terms of year of manufacture: the first group (n ¼ 4) reflects cars manufactured prior to 2004, that is, before the ban of penta-BDE commercial formulations in PUF; the second group are cars manufactured between 2005 and 2012 (n ¼ 15) reflecting cars manufactured after the replacement of penta-BDE with alternative flame retardants, such as OPEs in PUFs. The box plot (Fig. 2) indicates a wide variation in the concentrations of OPEs in the latter group. This could mean an increased use of OPEs in cars manufactured after 2004.
244
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
Table 3 Descriptive statistics for OPEs in indoor dust from several microenvironments in Durban, South Africa. Statistical descriptor P
n¼10 Homes Mean Median Minimum Maximum 5th percentile 95th percentile P n¼9 Offices Mean Median Minimum Maximum 5th percentile 95th percentile P n¼12 University Computer Laboratories Mean Median Minimum Maximum 5th percentile 95th percentile P n¼19 Cars Mean Median Minimum Maximum 5th percentile 95th percentile
P OPEs
Concentrations of OPEs/ng g1 TCEP
TCIPP
TDCIPP
TPHP
97650 9765 7390 2400 34250 3723 24013 88690 9854 10820 6900 11720 7364 11624 109750 9145 8420 5040 18350 5310 15050 667410 35126 10200 2000 245230 4664 131173
47860 4786 3545 220 12870 621 10665 38880 4320 3810 1240 10320 1288 9484 48420 4035 3485 830 11350 951 8435 191310 10069 5000 770 56250 2417 35424
327510 32751 7695 740 226200 965 147972 60690 6743 8320 1450 11680 1778 11312 58520 4877 3415
47510 4751 2025 760 26120 1062 16738 65420 7269 3730 650 18370 974 18042 49250 4104 3580 1420 9990 1503 8681 177210 9327 6170 670 34100 823 31562
520530 52053 22490 4790 299440 7850 194968 253680 28187 26930 14920 49210 15280 46210 265940 22162 19565 10070 43390 11676 37764 2730600 143716 49010 14020 773590 21625 431176
3.5. Comparison of OPE levels in different microenvironments in South Africa with levels reported in other countries The target OPEs (with the exception of TDCIPP, which was below the detection limit in one car and one university computer laboratory sample) were detected in all samples from the various microenvironments. The Kruskal-Wallis test showed statistical differences (p ¼ 0.003) in the concentrations of the OPEs in the various microenvironments. A comparison of the median concentrations and range (given in parentheses) of OPEs in the studied microenvironments with data reported in different countries is shown in Table 4. Similar to data from the USA (Fang et al., 2013), Germany (Brommer et al., 2012), and the Netherlands (Brandsma et al., 2014), cars contain the bulk of the OPEs compared with the other indoor microenvironments. TDCIPP dominates the phosphorus flame retardants contaminating South Africa's indoor environments in line with the profile in many countries worldwide (Table 4). 3.6. Implications for human exposure via accidental dust ingestion
Fig. 2. Box plot for OPE concentrations and (a) car manufacturer, and (b) cars grouped into periods of manufacture.
In evaluating human exposure via dust ingestion of contaminants, we assumed 100% absorption of OPEs from ingested dust in accordance with other studies (Abdallah and Covaci, 2014; Stapelton et al., 2009; Ali et al., 2013). Average dust intake rates of 20 and 50 mg d1 and high dust intakes of 50 and 200 mg d1 for adults, teenagers and toddlers were used as reported by Ali et al. (2013). Body weights of 70 kg and 12 kg were used for adults and toddlers respectively (Ali et al., 2013), and 52 kg for teenagers (Johnson-Restrepo and Kannan, 2009). Questionnaires were used to obtain the average number of hours spent per day by adults in offices, cars and homes. The average number of hours students spend in computer laboratories (for lectures, studies and assignments) and in their residences per day were obtained by interview. The amount of time spent per day in homes by toddlers (79.9%) was
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
245
Table 4 Comparison of OPE concentrations in the present study with other studies worldwide. Country
Microenvironment
Median Concentrations/ng g1 (range in parenthesis) TCEP
South Africa
Homes (n ¼ 10)
TCIPP
Reference TPHP
China USA USA USA USA Sweden Sweden Sweden Philippines Germany Germany Germany Germany Norway Egypt
2025 (760 e26120) Offices (n ¼ 9) 10820 (6900 3810 (1240e10320) 8320 (1450e11680) 3730 (650 e11720) e18730) University computer laboratories 8420 (5040e18350) 3485 (830e11350) 3415 (e 11450) 3580 (1420 (n ¼ 12) e9990) Cars(n ¼ 19) 10200 (2000 5000 (770e56250) 12770 (e 697100) 6170 (670 e245230) e34100) Homes (n ¼ 10) 2100 (e 33000) 1600 (700e11000) 10000 (2200e27000) 1200 (100e4200) Daycare centres (n ¼ 10) 30000 (2500 3100 (800e12000) 9100 (3900e150000) 1900 (300 e150000) e17000) Workplaces (n ¼ 10) 6700 (1300 19000 (3400 17000 (3300e91000) 5300 (900 e260000) e120000) e32000) House dust around electronics (n ¼ 8) 1300 (220e6900) 1300 (480e3800) 280 (70e3200) 820 (680e11000) House dust on electronics (n ¼ 8) 880 (520e2200) 1300 (580e4500) 680 (100e7400) 6500 (1600 e21000) Car dashboards (n ¼ 8) 2800 (1100e5700) 5700 (1800 17000 (6000e150000) 1700 (360 e110000) e14000) Car seats (n ¼ 8) 600 (240e5600) 4300 (1400 110000 (3800 2400 (670 e110000) e1100000) e43000) Hotels (n ¼ 8) (82e2300) (1000e9800) (69e18000) (110e2600) Houses (n ¼ 148) 5800 8700 2800 4500 Houses (n ¼ 48) <650 740 <590 870 Homes (n ¼ 15) 1140 (131e10400) 2290 (945e12300) 502 (
Egypt
Offices (n ¼ 20)
31 (
80 (
Egypt
Cars (n ¼ 20)
127 (
291 (
135 (26e1872)
Egypt
Public (n ¼ 11)
234 (
232 (
416 (
629 (116e2357)
Egypt Spain Belgium Canada Canada Romania Kuwait Kuwait Pakistan Pakistan Turkey United Kingdom United Kingdom United Kingdom United Kingdom
Houses (n ¼ 17) Houses (n ¼ 8) Houses (n ¼ 8) Houses (n ¼ 134) Houses (n ¼ 92) Houses (n ¼ 3) Houses (n ¼ 15) Cars (n ¼ 15) Houses (n ¼ 15) Cars (n ¼ 15) Houses (n ¼ 39) Living rooms (n ¼ 32)
143 (40e594) (250e9800) (75e1350) 800 2572 (136e196459) (40e1450) 710 1765
276 (41 -2018) (
191 (1.8e2628) 810
891 (136e4122) 21000
674 (0.95e16805) 710
388 (87e8957) (290e9500) (236e2640) 1700 3071 (59e8845) (105e3750) 430 1760 175 245 304 (1.7e1584) 3300
Offices (n ¼ 62)
870
33000
480
4300
Cars (n ¼ 21)
1230
53000
31000
3300
Classrooms (n ¼ 28)
860
16000
510
4100
South Africa South Africa South Africa Sweden Sweden Sweden Netherlands Netherlands Netherlands Netherlands Japan Japan Japan China China China China
LOQ refers to limit of quantitation.
7390 (2400e34250) 3545 (220e12870)
TDCIPP 7695 (740e226200)
49 (
73 (11e337)
This study This study This study This study Bergh et al. (2011) Bergh et al. (2011) Bergh et al. (2011) Brandsma et al. (2014) Brandsma et al. (2014) Brandsma et al. (2014) Brandsma et al. (2014) Takigami et al. (2009) Araki et al. (2014) Tajima et al. (2014) Sun et al. (2019) Tao et al. (2018) Wu et al. (2016) Wu et al. (2016) Wu et al. (2016) Stapleton et al. (2009) Fang et al. (2013) Fang et al. (2013) Allen et al. (2013) Marklund et al. (2003) Marklund et al. (2003) Marklund et al. (2003) Kim et al. (2013) Brommer et al. (2012) Brommer et al. (2012) Brommer et al. (2012) Fromme et al. (2014) Cequier et al. (2014) Abdallah and Covaci (2014) Abdallah and Covaci (2014) Abdallah and Covaci (2014) Abdallah and Covaci (2014) Shoeib et al. (2019) García et al. (2007) Van den Eede et al., 2012 Fan et al. (2014) Shoeib et al. (2019) Van den Eede et al., 2012 Ali et al. (2013) Ali et al. (2013) Ali et al. (2013) Ali et al. (2013) Shoeib et al. (2019) Brommer and Harrad (2015) Brommer and Harrad (2015) Brommer and Harrad (2015) Brommer and Harrad (2015)
246
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247
Table 5 Summary of adult, teenager and toddler exposure to target OPEs (ng d1) via accidental indoor dust ingestion using different exposure scenarios. OPEs
Reference dose Adult
TCEP
Exposure Scenario Adult
Toddler
154000 44000
Mean High TCIPP 560000 160000 Mean High TDCIPP 105000 30000 Mean High TPHP 490000 140000 Mean High
Teenager
Toddler
Low end Median Average High end Low end Median Average High end Low end Median Average High end 100 249 18 46 26 64 20 51
169 423 74 185 162 406 55 138
217 543 97 243 529 1323 116 289
488 1219 226 566 2258 5646 356 890
the same as that of Ali et al. (2012). Thus, the average time an adult spends in the office, car and home were obtained as 33.3, 4.2, and 62.5%, respectively. For full-time undergraduate students (teenagers), it was estimated that they spend 54.2 and 45.8% of their time indoors in classrooms and in their residences respectively. Different exposure scenarios were calculated by using the 5th percentile (low end), median, mean and 95th percentile (high end) concentrations from homes, cars, offices and computer laboratory dusts. Thus, the daily dose of OPEs (SDED/ng d1) via dust ingestion was calculated from the following equations reported by Ali et al. (2013) with modifications. SDED/ng d1 ¼ [(CHDFH) þ (CofDFof) þ (CCDFC)]DIR/BW, for adult exposure estimation, SDED/ng d1 ¼ [(CHDFH) þ (CLDFL)]DIR/BW, for teenager exposure assessment, and SDED/ng d1 ¼ [(CHDFH) þ (CCDFC)]DIR/BW, for toddler exposure assessment, where CCD, CHD, CofD and CLD are dust concentrations in cars, homes, offices and computer laboratories (5th percentile, median, mean and 95th percentile) and FC, FH, Fof and FL are the fraction of time spent in cars, homes, offices and computer laboratories, respectively. DIR is the dust intake rate and BW is the body weight. P Table 5 shows the daily dose of individual OPEs and the OPEs among different population groups in South Africa compared with the toxicological reference dose (RfD) values for each OPE. The results indicate that the high end exposure (i.e. using the 95th percentile concentration) may cause significant health implications for toddlers. At this exposure scenario, young children ingest as P much as 36859 ng of n¼4OPEs per day following inadvertent dust ingestion. For the individual OPEs, the worst-case scenario results in daily ingestion of 4939 ng TCEP, 2002 ng TCIPP, 26978 ng TDCIPP and 2940 ng TPHP by toddlers. The worst-case daily exposure dose of TDCIPP is at par with the toxicological reference dose of 30000 ng d1 (Abdallah and Covaci, 2014). The dominance of two chlorinated phosphorus flame retardants e TDCIPP and TCEP e in exposure doses of studied OPEs in the South African population may pose significant health challenges particularly among toddlers who are exposed to a higher magnitude of these flame retardant chemicals. Studies have reported that TDCIPP is mutagenic, carcinogenic in rats and humans, and a moderate hazard for reproductive and developmental effects (Stapleton et al., 2009). Similarly, TCEP is reportedly carcinogenic for animals, neurotoxic to rats and mice, induces adverse reproductive effects in rats, and exhibits haemolytic and reproductive effects such as reduced fertility, longer estrous cycle length, and reduced sperm motility and density, in humans. TCEP was associated with the acute death of dogs after ingestion of car seat cushions containing an enormous
92 229 16 40 15 37 26 65
159 398 70 176 108 269 57 143
189 472 88 219 353 883 88 220
383 958 189 473 1476 3691 248 619
159 634 30 119 43 171 44 177
317 1267 152 608 334 1337 94 375
464 1856 212 849 1496 5983 209 838
1235 4939 500 2002 6745 26978 735 2940
amount of TCEP (van der Veen and de Boer, 2012). Several toxicological effects have also been reported for TCIPP and TPHP (van der Veen and de Boer, 2012). 4. Conclusions We have reported for the first time the presence of organophosphate esters in the South African indoor environment. The results of the study indicate the wide use of organophosphate esters, particularly the chlorinated phosphorus flame retardants, at levels comparable to those in the European Union. The abundance of TCEP in these microenvironments suggests V6 as a possible replacement of penta-BDE in polyurethane foams in South Africa. Cars contained the highest levels of OPEs among the microenvironments and TDCIPP was the most abundant OPE found among the four studied OPEs. The relationships between OPEs and some indoor characteristics implicate furniture and electronics as reservoirs of OPEs in the indoor environment, particularly in homes. The result of the present study is reflective of a shift to OPEs as flame retardants, probably in response to international regulations. Generally, the current exposure of OPEs should be of interest as TDCIPP and TCEP, which are known mutagens and carcinogens, majorly contributed to the overall exposure of the South African population, particularly young children, to OPEs. The current levels, profiles and magnitude of exposure indicate further investigations are required on other exposure routes such as dietary and inhalation; as well as patterns and human health implications of exposure to mixtures of these organophosphate esters. Acknowledgements OAA acknowledges the University of KwaZulu-Natal for a doctoral scholarship. The authors acknowledge the kind donation of a Restek RTX 1614® capillary GC column from Restek Corporation. We sincerely thank all home, office, and car owners for permitting sample collections. We also acknowledge Burchmore Motor dealership for allowing us to sample cars available for resale. The National Research Foundation of South Africa is thanked for research support. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.chemosphere.2019.04.175. References Abafe, O.A., Martincigh, B.S., 2015. Polybrominated diphenyl ethers and polychlorinated biphenyls in indoor dust in Durban, South Africa. Indoor Air 25, 547e556. Abdallah, M.A.-E., Covaci, A., 2014. Organophosphate flame retardants in indoor dust from Egypt: implications for human exposure. Environ. Sci. Technol. 48,
O.A. Abafe, B.S. Martincigh / Chemosphere 230 (2019) 239e247 4782e4789. Ali, N., Ali, L., Mehdi, T., Dirtu, A.C., Al-Shammari, F., Neels, H., Covaci, A., 2013. Levels and profiles of organochlorines and flame retardants in car and house dust from Kuwait and Pakistan: implication for human exposure via dust ingestion. Environ. Int. 55, 62e70. Ali, N., Dirtu, A.C., Eede, N.V.D., Goosey, E., Harrad, S., Neels, H., T Mannetje, A., Coakley, J., Douwes, J., Covaci, A., 2012. Occurrence of alternative flame retardants in indoor dust from New Zealand: indoor sources and human exposure assessment. Chemosphere 88, 1276e1282. Allen, J., Stapleton, H., Vallarino, J., McNeely, E., McClean, M., Harrad, S., Rauert, C., Spengler, J., 2013. Exposure to flame retardant chemicals on commercial airplanes. Environ. Health 12, 17. Andresen, J., Bester, K., 2006. Elimination of organophosphate ester flame retardants and plasticizers in drinking water purification. Water Res. 40, 621e629. Araki, A., Saito, I., Kanazawa, A., Morimoto, K., Nakayama, K., Shibata, E., Tanaka, M., Takigawa, T., Yoshimura, T., Chikara, H., Saijo, Y., Kishi, R., 2014. Phosphorus flame retardants in indoor dust and their relation to asthma and allergies of inhabitants. Indoor Air 24, 3e15. , A., 2007. Liquid Bacaloni, A., Cavaliere, C., Foglia, P., Nazzari, M., Samperi, R., Lagana chromatography/tandem mass spectrometry determination of organophosphorus flame retardants and plasticizers in drinking and surface waters. Rapid Commun. Mass Spectrom. 21, 1123e1130. € Bergh, C., Torgrip, R., Emenius, G., Ostman, C., 2011. Organophosphate and phthalate esters in air and settled dust e a multi-location indoor study. Indoor Air 21, 67e76. € rklund, J., Isetun, S., Nilsson, U., 2004. Selective determination of organophosBjo phate flame retardants and plasticizers in indoor air by gas chromatography, positive-ion chemical ionization and collision-induced dissociation mass spectrometry. Rapid Commun. Mass Spectrom. 18, 3079e3083. Brandsma, S.H., de Boer, J., Leonards, P.E.G., Cofino, W.P., Covaci, A., Leonards, P.E.G., 2013. Organophosphorus flame-retardant and plasticizer analysis, including recommendations from the first worldwide interlaboratory study. Trac. Trends Anal. Chem. 43, 217e228. Brandsma, S.H., de Boer, J., van Velzen, M.J.M., Leonards, P.E.G., 2014. Organophosphorus flame retardants (PFRs) and plasticizers in house and car dust and the influence of electronic equipment. Chemosphere 116, 3e9. Brommer, S., Harrad, S., 2015. Sources and human exposure implications of concentrations of organophosphate flame retardants in dust from UK cars, classrooms, living rooms, and offices. Environ. Int. 83, 202e207. Brommer, S., Harrad, S., Van den Eede, N., Covaci, A., 2012. Concentrations of organophosphate esters and brominated flame retardants in German indoor dust samples. J. Environ. Monit. 14, 2482e2487. € Campone, L., Piccinelli, A., Ostman, C., Rastrelli, L., 2010. Determination of organophosphorous flame retardants in fish tissues by matrix solid-phase dispersion and gas chromatography. Anal. Bioanal. Chem. 397, 799e806. € Carlsson, H., Nilsson, U., Becker, G., Ostman, C., 1997. Organophosphate ester flame retardants and plasticizers in the indoor environment: Analytical methodology and occurrence. Environ. Sci. Technol. 31, 2931e2936. , R.M., Becher, G., Thomsen, C., 2014. Cequier, E., Ionas, A.C., Covaci, A., Marce Occurrence of a broad range of legacy and emerging flame retardants in indoor environments in Norway. Environ. Sci. Technol. 48, 6827e6835. Fan, X., Kubwabo, C., Rasmussen, P.E., Wu, F., 2014. Simultaneous determination of thirteen organophosphate esters in settled indoor house dust and a comparison between two sampling techniques. Sci. Total Environ. 491e492, 80e86. Fang, M., Webster, T.F., Gooden, D., Cooper, E.M., McClean, M.D., Carignan, C., Makey, C., Stapleton, H.M., 2013. Investigating a novel flame retardant known as V6: measurements in baby products, house dust, and car dust. Environ. Sci. Technol. 47, 4449e4454. Fromme, H., Lahrz, T., Kraft, M., Fembacher, L., Mach, C., Dietrich, S., Burkardt, R., € lkel, W., Go € en, T., 2014. Organophosphate flame retardants and plasticizers in Vo the air and dust in German daycare centers and human biomonitoring in visiting children (LUPE 3). Environ. Int. 71, 158e163. García, M., Rodríguez, I., Cela, R., 2007. Microwave-assisted extraction of organophosphate flame retardants and plasticizers from indoor dust samples. J. Chromatogr. A 1152, 280e286. Hartmann, P.C., Bürgi, D., Giger, W., 2004. Organophosphate flame retardants and plasticizers in indoor air. Chemosphere 57, 781e787. Johnson-Restrepo, B., Kannan, K., 2009. An assessment of sources and pathways of
247
human exposure to polybrominated diphenyl ethers in the United States. Chemosphere 76, 542e548. Kajiwara, N., Noma, Y., Takigami, H., 2011. Brominated and organophosphate flame retardants in selected consumer products on the Japanese market in 2008. J. Hazard Mater. 192, 1250e1259. Kim, J.-W., Isobe, T., Sudaryanto, A., Malarvannan, G., Chang, K.-H., Muto, M., Prudente, M., Tanabe, S., 2013. Organophosphorus flame retardants in house dust from the Philippines: occurrence and assessment of human exposure. Environ. Sci. Pollut. Control Ser. 20, 812e822. Marklund, A., Andersson, B., Haglund, P., 2003. Screening of organophosphorus compounds and their distribution in various indoor environments. Chemosphere 53, 1137e1146. Marklund, A., Andersson, B., Haglund, P., 2005. Organophosphorus flame retardants and plasticizers in air from various indoor environments. J. Environ. Monit. 7, 814e819. lez-Barreiro, C., Sitka, A., Scharf, S., Gans, O., 2007. Martínez-Carballo, E., Gonza Determination of selected organophosphate esters in the aquatic environment of Austria. Sci. Total Environ. 388, 290e299. Meeker, J., Stapleton, H., 2009. House dust concentrations of organophosphate flame retardants in relation to hormone levels and semen quality parameters. Environ. Health Perspect. 118, 318e323. Salamova, A., Ma, Y., Venier, M., Hites, R.A., 2014. High levels of organophosphate flame retardants in the Great Lakes atmosphere. Environ. Sci. Technol. Lett. 1, 8e14. Shoeib, T., Webster, G.M., Hassan, Y., Tepe, S., Yalcin, M., Turgut, C., KurtKarakus¸, P.B., Jantunen, L., 2019. Organophosphate esters in house dust: a comparative study between Canada, Turkey and Egypt. Sci. Total Environ. 650, 193e201. € Staaf, T., Ostman, C., 2005. Indoor air sampling of organophosphate triesters using solid phase extraction (SPE) adsorbents. J. Environ. Monit. 7, 344e348. Stapleton, H., Klosterhaus, S., Eagle, S., Fuh, J., Meeker, J., Blum, A., Webster, T., 2009. Detection of organophosphate flame retardants in furniture foam and U.S. house dust. Environ. Sci. Technol. 43, 7490e7495. Sun, Y., Liu, L.-Y., Sverko, E., Li, Y.-F., Li, H.-L., Huo, C.-Y., Ma, W.-L., Song, W.-W., Zhang, Z.-F., 2019. Organophosphate flame retardants in college dormitory dust of northern Chinese cities: occurrence, human exposure and risk assessment. Sci. Total Environ. 665, 731e738. Sundkvist, A.M., Olofsson, U., Haglund, P., 2010. Organophosphorus flame retardants and plasticizers in marine and fresh water biota and in human milk. J. Environ. Monit. 12, 943e951. Tajima, S., Araki, A., Kawai, T., Tsuboi, T., Ait Bamai, Y., Yoshioka, E., Kanazawa, A., Cong, S., Kishi, R., 2014. Detection and intake assessment of organophosphate flame retardants in house dust in Japanese dwellings. Sci. Total Environ. 478, 190e199. Takigami, H., Suzuki, G., Hirai, Y., Sakai, S.-I., 2009. Brominated flame retardants and other polyhalogenated compounds in indoor air and dust from two houses in Japan. Chemosphere 76, 270e277. Tao, Y., Shang, Y., Li, J., Feng, J., He, Z., Covaci, A., Wang, P., Luo, J., Mao, X., Shi, B., Hu, L., Luo, D., Mei, S., 2018. Exposure to organophosphate flame retardants of hotel room attendants in Wuhan City, China. Environ. Pollut. 236, 626e633. Thomsen, V., Schatzlein, D., Mercuro, D., 2003. Limits of detection in spectroscopy. Spectroscopy 18, 112e114. €ck, J., Tamburro, D., Crescenzi, C., Carlsson, H., 2006. Air sampling with Tollba Empore solid phase extraction membranes and online single-channel desorption/liquid chromatography/mass spectrometry analysis: determination of volatile and semi-volatile organophosphate esters. J. Chromatogr. A 1129, 1e8. Van den Eede, N., Dirtu, A., Ali, N., Neels, H., Covaci, A., 2012. Multi-residue method for the determination of brominated and organophosphate flame retardants in indoor dust. Talanta 89, 292e300. Van den Eede, N., Dirtu, A.C., Neels, H., Covaci, A., 2011. Analytical developments and preliminary assessment of human exposure to organophosphate flame retardants from indoor dust. Environ. Int. 37, 454e461. van der Veen, I., de Boer, J., 2012. Phosphorus flame retardants: properties, production, environmental occurrence, toxicity and analysis. Chemosphere 88, 1119e1153. Wu, M., Yu, G., Cao, Z., Wu, D., Liu, K., Deng, S., Huang, J., Wang, B., Wang, Y., 2016. Characterization and human exposure assessment of organophosphate flame retardants in indoor dust from several microenvironments of Beijing, China. Chemosphere 150, 465e471.