Partitioning and diffusion of PBDEs through an HDPE geomembrane

Waste Management xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Waste Management journal homepage: www.elsevier.com/locate/wasman

Partitioning and diffusion of PBDEs through an HDPE geomembrane R. Kerry Rowe a,⇑, Pooneh T. Saheli a, Allison Rutter b a b

Geotechnical and Geoenvironmental Engineering, GeoEngineering Centre at Queen’s-RMC, Queen’s University, Ellis Hall, Kingston, ON K7L 3N6, Canada School of Environmental Studies, Queen’s University, Kingston, ON K7L 3N6, Canada

a r t i c l e

i n f o

Article history: Received 16 October 2015 Revised 20 March 2016 Accepted 4 May 2016 Available online xxxx Keywords: Geosynthetics Polybrominated diphenyl ethers High density polyethylene Diffusion coefficient Partitioning coefficient MSW landfill Contaminants of emerging concern Brominated flame retardants

a b s t r a c t Polybrominated diphenyl ether (PBDE) has been measured in MSW landfill leachate and its migration through a modern landfill liner has not been investigated previously. To assure environmental protection, it is important to evaluate the efficacy of landfill liners for controlling the release of PBDE to the environment to a negligible level. The partitioning and diffusion of a commercial mixture of PBDEs (DE-71: predominantly containing six congeners) with respect to a high-density polyethylene (HDPE) geomembrane is examined. The results show that the partitioning coefficients of the six congeners in this mixture range from 700,000 to 7,500,000 and the diffusion coefficients range from 1.3 to 6.0
1015 m2/s depending on the congener. This combination of very high partitioning coefficients and very low diffusion coefficients suggest that a well constructed HDPE geomembrane liner will be an extremely effective barrier for PBDEs with respect to diffusion from a municipal solid waste landfill, as illustrated by an example. The results for pure diffusion scenario showed that the congeners investigated meet the guidelines by at least a factor of three for an effective geomembrane liner where diffusion is the controlling transport mechanism. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Many modern engineered landfills have a composite liner as part of a barrier system to contain contaminants and prevent their migration to the surrounding environment at a concentration that could have health or environmental impacts. High density polyethylene (HDPE) geomembranes have been commonly used in composite liners at the bottom of modern municipal solid waste (MSW) landfills and they have shown excellent performance in containing a broad range of chemicals (Rowe et al., 2004). Polybrominated diphenyl ethers (PBDEs) are organobromine compounds (C12H(10x)BrxO (x = 1, 2, . . ., 10) (Fig. 1a). Homolog groups (mono through deca) refer to the number of bromine atoms (represented by ‘x’ in the formula) replacing hydrogen on the diphenyl structure. There are 209 PBDE congeners depending on the number and location of bromines. Table 1 summarizes the PBDEs examined in this study. PBDEs have been widely used as flame retardants in products that are prone to ignition in fire situations (e.g., furniture) or items in which a fire can start (e.g., electrical devices). PBDEs allow more time either to extinguish the fire and/or to escape by slowing down ignition and fire growth (U.S. EPA, 2006). The release of PBDEs to the environment can arise

⇑ Corresponding author. E-mail addresses: [email protected] (R.K. Rowe), pooneh.saheli@queensu. ca (P.T. Saheli), [email protected] (A. Rutter).

from (i) the manufacturing of PBDEs, (ii) the manufacture, aging and wear of products containing PBDEs like sofas and electronics, and (iii) the recycling and disposal of such products at the end of their lives (U.S. EPA, 2006). One common source of PBDEs is household dust which has been reported to have an average PBDE concentration of more than 4600 parts per billion (ppb) (UNEP, 2004). PBDEs can enter the human body through high fat foods such as fatty fish (Szlinder-Richert et al., 2010), supermarket food (U.S. EPA, 2006) and breast milk (U.S. EPA, 2006), or by inhalation of PBDEs in household dust (UNEP, 2004; Sjödin et al., 2008). PBDEs have been shown to bioaccumulate in fat tissues, blood, and breast milk. Since commercial production of PBDE started the 1970s, the concentration of PBDEs in human blood and tissue has doubled every five years (UNEP, 2010). Studies have shown that PBDEs are endocrine disrupters which mimic the behavior of natural hormones in human body and disrupt the chemical signaling system and hence development of brain and reproductive systems (UNEP, 2010). Many wastes containing PBDEs find their way into landfills (e.g., household dust, old furniture, appliances, mobile phones, etc.). PBDEs have been detected in landfill leachate (Table 2) in different parts of the world (Osako et al., 2004; Danon-Schaffer et al., 2006; Haarstad and Borch, 2008; Odusanya et al., 2009) and in Canada at total concentrations up to 2.5 lg/L (personal communication with Environment Canada, 2009). Given their presence in MSW landfill

http://dx.doi.org/10.1016/j.wasman.2016.05.006 0956-053X/Ó 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

2

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx

Fig. 1. Structural formula of PBDEs: (a) generic (m + n = 1, 2, . . ., 10), (b) congener 47 (tetraBDE), (c) congener 99 (pentaBDE), and (d) congener 153 (hexaBDE).

Table 1 Molecular descriptors of main congeners in DE-71 used in this study in the order of molecular volume (adapted from Li et al., 2008). Congener (No. of Br)

Percent composition (%)

Log Kow

Molecular weight (g/mole)

Molecular volume (Â3)

Molecular polarizability (Â3)

q+H

q C

47 (4) 85 (5) 99 (5) 100 (5) 154 (6) 153 (6)

32.4 2.42 43.9 8.93 3.30 3.84

6.81 7.37 7.32 7.24 7.82 7.90

486 565 565 565 644 644

806.53 850.05 854.17 858.55 893.00 903.68

31.23 33.86 33.86 33.86 36.49 36.49

0.141 0.140 0.144 0.143 0.147 0.144

0.162 0.162 0.160 0.160 0.182 0.166

q O

q+Br

(atomic charge unit) 0.092 0.093 0.090 0.089 0.076 0.094

0.060 0.094 0.064 0.066 0.067 0.065

q+H, the most positive atomic net charges on a hydrogen atom. q C , the most negative atomic net charges on a carbon atom. q O , the atomic net charges on the oxygen atom. q+Br, the most positive atomic net charges on a bromine atom.

Table 2 Total PBDE concentrations reported in landfill leachate.

a b

Reference

Location

Number of landfills

Total PBDE (ng/L)b

Osako et al. (2004) Danon-Schaffer et al. (2006) Haarstad and Borch (2008) Odusanya et al. (2009) Environment Canada (2009)a

Japan BC, Canada Norway South Africa Canada

7 – 3 5 10

0.096–18.17 1470 0.02–11.1 8.4–54.7 7.6–2476

Personal communication. 1 ppb = 1000 ng/L.

leachate and the fact that most landfill regulations were developed before the potential health issues associated with PBDEs were recognized, it is important to investigate whether and, if so, under what circumstances, modern landfill liners can contain PBDEs sufficiently to prevent unacceptable impact. There have been many studies of the diffusion of various organic compounds in landfill leachate through different types of geomembranes including diffusion of a number of organic chemicals through HDPE geomembranes (Park and Nibras, 1993; Sangam and Rowe, 2001; Joo et al., 2005; Chao et al., 2007; Islam and Rowe, 2009), PVC and LLDPE (McWatters and Rowe, 2009) and coextruded geomembranes (McWatters and Rowe, 2010). Preliminary work from the present study was reported by Taghizadeh-Saheli et al. (2011), however, the diffusion of PBDE through an HDPE geomembrane has not been reported in the archival literature. The objective of this paper is to: (i) quantify partitioning of several common PBDE congeners to an HDPE geomembrane, (ii) evaluate the diffusion coefficient of several PBDE congeners in an HDPE geomembrane, (iii) compare the partitioning and diffusion coefficients of the different PBDE congeners, (iv) use the parame-

ters derived from the experimental work to calculate the transport of PBDE through a typical MSW landfill barrier system.

2. Background 2.1. Regulation of polybrominated diphenyl ether (PBDE) production With research showing the potential adverse effects of PBDEs on the human health and the environment, the European Parliament banned marketing and use of two out of three commercial mixtures of PBDE in the European Union (EU) in 2004 (U.S. EPA, 2006). The Bromine Science and Environmental Forum (BSEF) member companies, in cooperation with US EPA, voluntarily stopped the production and use of decaBDE for the US market by the end of 2012 (www.bsef.com). The Canadian Environmental Protection Act (CEPA 1999) prohibited the production of congener groups of tetraBDE and higher BDEs in 2008 and also bans the use, sale, offer for sale and import of tetraBDEs, pentaBDEs and hexaBDEs (Environment Canada, 2008). In 2009, the Stockholm Convention added two out of three commercial mixtures of PBDE

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx

to the list of Persistent Organic Pollutants (POPs) to stop their new production while allowing their recycling and reuse until 2030 (UNEP, 2010). 2.2. Diffusive transport through geomembrane Diffusive migration of a contaminant through a geomembrane occurs in three steps (Sangam and Rowe, 2001): adsorption to the geomembrane on the side in contact with contaminated fluid, diffusion through the geomembrane, and desorption from the geomembrane on the side away from the contaminated source. Sorption and desorption are controlled by the partitioning coefficient of a contaminant between geomembrane and the adjacent fluid, Sgf [–]:

cg ¼ Sgf cf

ð1aÞ

where Sgf is generally deduced from a batch test as the ratio of the final concentration of contaminant in the geomembrane, cg = cgF [ML3], to the final concentration in the fluid, cf = cfF [ML3] at equilibrium (Sangam and Rowe, 2001):

Sgf ¼ cgF =cfF

The evaluation of the partitioning and diffusion coefficients to be described subsequently required the extraction of these PBDEs from both the aqueous phase and the solid (geomembrane) phase to allow the quantification of the concentration of PBDE in the two phases (Saheli, 2016). 3.2.1. Extraction Liquid/liquid extraction was used to extract the PBDEs from the aqueous solutions associated with the partitioning and diffusion tests. The first stage of the extraction was into dichloromethane (DCM) followed by solvent exchange from DCM to hexane. The solution was concentrated to 500 lL and transferred to GC vials for analysis by a Gas Chromatograph/Electron Capture Detector (GC/ECD). The PBDEs were extracted from the geomembrane using a 24-h soxhlet extraction with DCM which was collected at the end of the soxhlet run and the DCM was exchanged to hexane and concentrated to 500 lL for final GC/ECD analysis. Decachlorobiphenyl was used as surrogate to correct for the efficiency of the extraction for both methods.

ð1bÞ

Contaminant diffuses through the geomembrane due to concentration gradient according to Fick’s first law:

f ¼ Dg

3

dcg dz

ð2Þ

where f is the mass flux [ML2 T1], Dg [L2 T1] is the diffusion coefficient (which depends on the contaminant and the type of the geomembrane), and dcg/dz is the concentration gradient in the geomembrane [ML4] parallel to the direction of transport. Assuming the diffusion coefficient is constant and considering the conservation of mass, the change in the concentration of the contaminant in the geomembrane with time t, can be expressed as (often called Fick’s second law):

@cg @ 2 cg ¼ Dg 2 @t @z

ð3Þ

3. Experimental investigation procedure 3.1. Materials The experiments described below examined a 1.5-mm-thick high density polyethylene (HDPE) geomembrane (Table 3) in contact with an aqueous solution containing one commercial PBDE mixture (DE-71; purchased from Wellington Laboratories). For analysis, BDE-MXE, a mixture of 27 congeners (1.0/2.0/5.0 ppm in nonane/toluene) was used as a standard (purchased from Wellington Laboratories; see Saheli, 2016). 3.2. Methods As will be described later, both partitioning and diffusion tests for the PBDEs in DE-71 were performed in an aqueous solution. Table 3 Properties of the HDPE geomembrane examined. Properties based on Ewais and Rowe (2014). Properties

Unit

Value

Nominal thickness (ASTM D5199) Resin density (ASTM D1505) Geomembrane density (ASTM D1505) Std-OIT (ASTM D3895) HP-OIT (ASTM D5885) Degree of crystallinity (ASTM793)

mm g/cm3 g/cm3 min min %

1.5 mm 0.936 0.946 175 ± 3 960 ± 17 48 ± 1.6

3.2.2. Analytical method The final 500 lL samples from either liquid/liquid extraction or soxhlet extraction in the GC vials were analyzed using an Agilent 6890 GC/ECD equipped with a Supelco SBP1 (30 m
250 lm
0.25 lm) capillary column. The initial oven temperature was at 100 °C for 1 min and then increased to 180 °C at the rate of 15 °C/min and to 300 °C at 3 °C/min. The carrier and make-up gases were helium and nitrogen, respectively. The concentration of six main congeners in DE-71 (47, 85, 99, 100, 154, and 153, Fig. 1b–d) were calculated by comparison to the standard and corrected for surrogate. The detection limit in a 500 mL sample was 10 ppt (0.010 ppb) for each individual congener (Saheli, 2016). 3.2.3. Lab quality control (QC) Control tests (water only spiked with PBDE) were conducted to check the stability of the solution, as a check for any mass loss during the tests and analysis, and as a check for repeatability as a reference for other tests. Blank tests (water with no PBDE in a similar container to that used the control test) were used to check for background concentration of PBDEs in the laboratory environment (including testing apparatus and sampling equipment such as glassware and syringes as well as the general environment). The samples from the control and blank tests went through the same process as other samples to provide quality control. Prior to use in these experiments, containers and sampling equipment were rinsed six times with DCM to minimize the potential for residual PBDE contamination. Possible sources of the small amounts of PBDE found in blank samples may have been from dust in the lab environment. Control and blank specimens were run as part of QA/QC for each run. 3.3. Conventional partitioning and diffusion tests Knowing the properties of PBDEs one may attempt to predicted diffusion coefficients of PBDEs to HDPE using methods such as that given by Tang et al. (2013) or Fries and Zarfl (2012); however, the calculated Dg values using these correlations were either unreasonable or extremely low compared to the measured values. Thus, it appears that regressions such as these deduced of contaminants with much lower log Kow would not have been useful for estimating the diffusion characteristics of contaminants such as PBDE with extremely high log Kow, highlighting the need to perform experiments rather than rely on correlations when substantial extrapolation would be involved.

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

4

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx

Table 4 Reported solubility of various PBDE congeners in water (lg/L, ppb). Homolog (No. of Br)

Congener

Palm et al. (2002)

Tittlemier et al. (2002)

Kuramochi et al. (2007)

2 3

15 17 28 47 99 85 153

– 26 – 9.32 1.02 0.08 –

130 – 70 15 9.4 6 0.87

216 – – 14.7 4.37 – 0.05

4 5 6

Preliminary partitioning tests were performed for the HDPE geomembrane and PBDEs by immersing 0.8 g of a 0.5-mm-thick HDPE geomembrane in 500 mL of 4 ppb aqueous solution of congener 7 (diBDE). The concentration in the solution dropped below the detection limit in less than 72 h, making it impossible to establish the equilibrium between the concentrations of congener 7 in the geomembrane and solution. Likewise, an attempt to perform conventional two-compartment diffusion test (Park et al., 1996; Sangam and Rowe, 2001; Joo et al., 2005; Islam and Rowe, 2009; McWatters and Rowe, 2009, 2010) failed because the concentration in the source dropped below the detection limit in less than a week and the concentration in the receptor never exceeded the detection limit, making it impossible to obtain a diffusion coefficient based on the typical approach of finding a relationship between the source and the receptor concentrations. These problems arose because of the very low solubility of PBDEs in water (limiting the maximum source concentration), combined with the incredibly high affinity of PBDEs for the HDPE geomembrane. To address these problems with using conventional methods, two very different approaches were developed and used together to establish the partitioning and diffusion coefficients as described below. 3.4. Methods adopted for assessing the partitioning and diffusion coefficients 3.4.1. Low solubility of PBDE in water The solubility of PBDE congeners in water decreases with an increase in the number of attached bromines (Table 4). The commercial mixture DE-71 (Table 1) used in this study has a reported solubility of 13.3 lg/L (ppb) in water (U.S. HHS, 2004) which means the initial concentration in aqueous solution would be too low to reach equilibrium before going below detection limit and hence it would not be possible to obtain partitioning and diffusive parameters. To increase the initial concentration, a pure commercial mixture (DE-71) was dissolved in 90% DD water-10% methanol solutions at four concentrations (10, 50, 100, 500 ppb) and monitored regularly. Based on the results, 100 ppb DE-71 was selected as the initial concentration for both partitioning and diffusion tests (Saheli, 2016). The approach adopted here can be expected to have no effect on the diffusion coefficient since diffusion through the geomembrane is independent of the source solution provided that the solution does not change the properties of the geomembrane (as in this case). However, the presence of 10% methanol in the base solution will likely decrease the value of Sgf compared to that expected if the PBDE was only in water; thus the Sgf values reported in this paper are expected to be conservative (i.e., underestimating partitioning) and hence the impacts calculated in the example case may be too high. 3.4.2. PBDE (DE-71) partitioning to HDPE geomembrane Since the preliminary tests showed that the PBDE in the solution partitioned to 0.8 g of 0.5-mm-thick HDPE geomembrane

and did not remain in the solution at a detectable level, partitioning tests were performed with only a very small mass of geomembrane added to 500 mL of the 100 ppb DE-71 solution (hereafter referred to as the ‘‘spiked solution” solution), so that equilibrium concentrations above the detection limit could be reached. A 2 cm
2 cm piece of 1.5-mm-thick HDPE geomembrane (Table 3) was shaved into thin layers (25 lm) using a microtome and then a small mass (0.005 g) of this shaved geomembrane was added to 500 mL of the spiked solution and the bottle was placed in a tumbler for the duration of the test. At the end of the test, the geomembrane pieces and the solution were separated and both were analyzed for PBDEs. Several sets of partitioning tests were performed for different time durations until equilibrium was reached. Each set of tests consisted of: (1) three or four partitioning tests, (2) one or two control tests, and (3) a blank test. 3.4.3. PBDE (DE-71) diffusion through HDPE geomembrane A new approach, developed by Jones (2016) for PCBs, was used to estimate PBDE diffusion through an HDPE geomembrane. The following procedure was adopted for each diffusion test: i. Immersion phase: A 1.5 cm
7 cm (1.5 g) sample of 1.5-mm-thick HDPE geomembrane was immersed for six days in 500 mL of spiked solution (100 ppb DE-71). After 6 days, more than 94% of the DE-71 in the solution had partitioned into the geomembrane. ii. Pure diffusion phase: The geomembrane sample was removed from the solution, air dried, and stored in a glass jar to allow the sorbed PBDE to diffuse through the geomembrane. iii. Monitoring: At specific times, a geomembrane sample was removed from the jar, the edges of the sample were removed to a distance greater than the geomembrane thickness to eliminate edge effects, and then the samples were divided into eight smaller specimens (1.2 cm
0.7 cm, 0.15 g) and a microtome was used to obtain 25-lm-thick geomembrane layers at three different distances from the surface of the geomembrane to monitor the diffusion profile through the geomembrane. iv. The shavings from different depths were weighed and analyzed for PBDEs. The concentration of PBDEs for each layer (depth) was calculated by dividing the measured mass of each PBDE congener by the mass of geomembrane analyzed. Each set of tests typically involved three components: (1) three replicate diffusion tests, (2) one or two control tests, and, (3) a blank test. 4. Results 4.1. Partitioning test The partitioning tests were performed for 1, 7, 13, 32, 57, 133, 162 and 186 days to ensure that equilibrium had been reached (Fig. 2). Both geomembrane and solution were analyzed for the total mass of the six main congeners in DE-71 at the end of each test. The measured DE-71 mass of each test was normalized with respect to the mass of DE-71 in the control cell (mo). At equilibrium, which was reached after about 80 days, more than 94% of the initial mass of DE-71 had partitioned to the HDPE geomembrane (Fig. 2). The data for each individual DE-71 congener was averaged for three (or more) parallel tests at each sampling time and the average congener concentrations at these times (Table 5) were

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx

5

The best estimate of the partitioning coefficients for individual congeners and DE-71 mixture based on an average of the partitioning coefficients from six tests (three 133-day and three 186-day tests) at equilibrium are given in Table 6. They ranged from a very high value of 700,000 for BDE-47 to an extremely large value of 7,500,000 for BDE-153 with an overall mixture average of 1,800,000. The range of values based on the mean ± one standard deviation (from the 6 tests) is also given. These results demonstrate that if these PBDEs are in an aqueous solution that comes into contact with HDPE, they will predominantly partition to the HDPE out of solution; rapidly depleting the PBDEs in the solution (e.g., leachate). 4.2. Diffusion test and numerical model

Fig. 2. Partitioning of DE-71 into HDPE geomembrane (GMB) with time.

Table 5 Partitioning of individual DE-71 congeners to the HDPE geomembrane with time, based on the average concentration of at least three tests (percentage of the initial mass of the congener remaining in the solution). Time (day)

1 7 13 32 57 133 162 186

Congeners

DE-71

47 (4Br)

85 (5Br)

99 (5Br)

100 (5Br)

154 (6Br)

153 (6Br)

65.4 48.1 32.5 29.9 18.4 10.5 11.4 11.0

73.7 60.0 46.4 30.7 10.9 4.0 4.9 3.8

65.1 49.8 37.3 22.2 7.4 1.9 1.8 1.6

70.2 61.2 48.0 28.9 10.0 3.7 3.7 3.2

79.4 64.4 62.2 40.3 12.9 2.2 2.2 1.2

80.5 61.5 59.1 40.4 12.3 1.7 1.7 1.0

66.9 51.4 38.1 26.6 11.7 4.6 4.8 4.3

examined. At equilibrium (the last three rows in Table 5) more than 88% of the initial mass of congener 47 (tetraBDE), more than 95% of congeners 85, 99 and 100 (pentaBDE) and more than 97% of congeners 154 and 153 (hexaBDE) were partitioned to the HDPE geomembrane. Of the six congeners in DE-71, the lowest brominated congener 47 (tetraBDE) migrated more quickly into the geomembrane (35% decrease from the solution into the geomembrane after 1 day) implying faster diffusion but it also reached equilibrium at a higher concentration (11% remaining in the solution) implying that it had the lowest partitioning coefficient (Sgf) of the congeners. In contrast, the most highly brominated congeners 153 and 154 (hexaBDE) migrated more slowly into geomembrane (20% decrease in the solution in 1 day) but reached the lowest equilibrium concentration in the aqueous solution (1% to 2% in the solution) implying the highest partitioning coefficient of these congeners (Table 5). The three pentaBDEs (congeners 85, 99 and 100) gave intermediate results between the tetraBDE and hexaBDE congeners. Congener 99 is the dominant congener in DE-71 representing 43.9% of the total BDE mass (Table 1) implying greater availability for sorption and diffusion. In contrast, congener 85 only represented 2.4% of available BDE which means much less availability. It is unknown whether the fact that congener 99 exhibited the greatest sorption of the three pentaBDEs (reaching the lowest normalized equilibrium concentration at 1.8% of the initial value compared to 4.2% for congener 85 and 3.5% for congener 100; Table 5) was due to the higher concentration of congener or was a consequence of its particular structure, although it is suspected it may be the former since there is no apparent correlations between the degree of partitioning and the parameters given in Table 1.

For each diffusion sample, the concentration of DE-71 was measured in the three layers between surface and the middle of the geomembrane. The measured concentration for each layer was plotted in the middle of that specific layer with the error bars showing the depth range of each layer (Figs. 3–6). Fig. 3 shows the diffusion profile of bulk DE-71 (all congeners) through the 1.5-mm-thick HDPE geomembrane at different times. However, while it is convenient to represent DE-71 as a single contaminant and to estimate bulk Sgf and Dg of the mixture, in fact the different congeners each have a different molecular volume (Table 1) and solubility (Table 4) and thus their diffusive characteristics (and hence Dg and Sgf) are quite different as illustrated by the plots for congener 47 (which has the smallest molecular volume and highest solubility and represents 32% of DE-71; Fig. 4), 99 (the predominate congener representing 43.9% of DE-71; Fig. 5) and 153 (with the largest molecular volume and lowest solubility; Fig. 6). Since the DE-71 partitions and diffuses from both sides of the geomembrane, the middle of the geomembrane is a zero flux boundary and so it is sufficient to model only half of the thickness of the geomembrane. To estimate the diffusion coefficient based on the experimental results, the computer code POLLUTE (Rowe and Booker, 2004) was used to model both the 6-day immersion phase and subsequent pure diffusion stages of the experiments. For the first stage, the partitioning of DE-71 to the 1.5-mmthick geomembrane in 500 mL of spiked solution was modelled together with the diffusion that could occur in the 6 days of this stage to establish the initial mass and concentration profile in the sample at the start of the second stage. Half the thickness of the geomembrane (0.75 mm) was modelled with a finite mass boundary condition (Rowe and Booker, 1985) for the source solution (e.g., total co = 94.8 lg/L for the six congeners in DE-71 that were investigated) and a zero flux boundary condition for the centre of the sample. The model was run for a range of Dg (0.5
1015–7
1015 m2/s) and Sgf (700,000–7,500,000) (Table 6) to capture the partitioning and diffusion of the relevant congeners. For the second stage, again half of the geomembrane thickness was modelled but this time with zero flux boundaries at the surface and mid-geomembrane and the calculated concentration at day six from the first stage was used as the initial concentration profile for the second stage of the modelling, using the corresponding Dg and Sgf. The model was used to calculate the diffusion profile for days 105, 161, 248 and 405 and the results were compared with the experimental data. The same procedure was adopted for the sum of the six congeners of DE-71 and for each individual congener in the mixture to estimate Dg for each. The R2 was calculated for the fit to each set of experimental data at the four times examined (105, 161, 248 and 405 days). The Dg with the best average R2s for the fit to the experimental data at the four times was selected as the best estimate. Since this represents an overall best estimate based on all times, the corresponding theoretical fit at different

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

6

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx

Table 6 Sgf and Dg deduced for individual congeners in the order of molecular volume.

a b

Congener (No. of Br)

Percent composition

Best estimatea Sgf (–)

Rangeb Sgf
106 (–)

Best estimate Dg (m2/s)

Range Dg
1015 (m2/s)

47 (4) 85 (5) 99 (5) 100 (5) 154 (6) 153 (6) All 6 congeners

32.4 2.42 43.9 8.93 3.30 3.84 94.79

0.7
106 2.3
106 4.9
106 2.6
106 6.0
106 7.5
106 1.8
106

0.63–0.77 2.1–2.5 4.2–5.6 2.4–2.8 4.3–7.7 5.5–9.5 1.7–1.9

5.7
1015 3.7
1015 3.3
1015 2.0
1015 1.5
1015 2.5
1015 3.7
1015

5.5–6.0 3.0–4.5 3.0–3.7 1.5–2.5 1.3–2.0 2.0–3.0 3.0–4.5

average of six data points. Range of Sgf is the average Sgf ± 1 standard deviation based on six data points.

Fig. 3. DE-71 diffusion through HDPE geomembrane (model and experimental data), error bars show the analyzed depth of geomembrane.

times is variable (Figs. 3–6). A better fit could be obtained at any given time by selecting a Dg to give the best fit at that specific time; however, there really should only be one diffusion coefficient applicable at all the times and the variance from the average best fit Dg theoretical curve at different times reflects uncertainty in the parameter related to the approximations associated with the microtoming and analytical variability. To reflect this uncertainty, a range of Dg values was also deduced (Table 6). The Dg values were very low for all congeners, with best estimate values ranging from 1.5
1015 m2/s (BDE-154) to 5.7
1015 m2/s (BDE-47) and a best estimate of 3.7
1015 m2/s for the overall mixture. Generally, the diffusion coefficient of the congeners decreased with an increase in the number of attached bromines. This may be due to an increase in the molecular volume and weight but other molecule characteristics, such as the arrangement of atomic charges, may have also affected the diffusion coefficient. For example, congener 153 had a slightly larger Dg compared to congeners 154 and 100 which both have a small molecular volumes but congener 153

also has a different most negative atomic net charges on oxygen and carbon atoms. Irrespective of minor experimental uncertainty with respect to Dg and Sgf (Table 6), a comparison of the partitioning and diffusion coefficients for PBDE with those for BTEX or even PCBs (Table 7) shows that the bulk diffusion coefficient for the six DE-71 congeners investigated (which represents 95% of the total DE-71 and referred to hereafter as the diffusion coefficient for the DE-71 mixture) is almost two orders of magnitude smaller than that for BTEX and even about one order of magnitude smaller than that for PCBs. The difference is even more profound for partitioning coefficient for which the DE-71 was one order of magnitude higher than that for PCBs and almost four orders of magnitude higher than that for BTEX. There are many factors affecting the partitioning and diffusion coefficients of a compound with respect to HDPE geomembrane. For example, for organic compounds increasing in molecular weight and hydrophobicity increases Sgf and increasing in the weight and size of the compounds decreases the diffusion

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx

7

Fig. 4. Congener 47 (tetraBDE) diffusion through HDPE geomembrane (co in the solution = 32.4 lg/L), error bars show the analyzed depth of geomembrane.

Fig. 5. Congener 99 (pentaBDE) diffusion through HDPE geomembrane (co in the solution = 43.9 lg/L), error bars show the analyzed depth of geomembrane.

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

8

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx

Fig. 6. Congener 153 (hexaBDE) diffusion through HDPE geomembrane (co in the solution = 3.84 lg/L), error bars show the analyzed depth of geomembrane.

Table 7 Reported Sgf and Dg for various contaminants for an HDPE geomembrane. 2

Compound name

Dg (m /s)

Sgf (–)

Reference

BTEX PCB PBDE (DE-71 mixture)

2
1013 1
1014 4
1015

50–450 150,000 1,800,000

Islam and Rowe (2009) Jones (2016) This study

coefficient (Rowe et al., 2004). BTEX and DCM are lighter molecules (78–106 g/mole) compared to PCBs (188–499 g/mole depending on the congener) and PBDEs (249–959 g/mole depending on the congener). Although the estimated permeation coefficients for PBDE suggest that an HDPE geomembrane would be an effective diffusive barrier, the question remains; how effective would a composite liner be for controlling the escape of PBDEs? The answer to this question is case specific, but could be estimated by modelling using the values of Sgf and Dg deduced from the experiments described above (Table 6) to calculate the migration of the BDEs in DE-71 through a landfill barrier system and hence the potential impact on an underlying water resource. This question will be examined by a specific example in the following section. 5. Estimation of PBDE transfer through landfill barrier system To illustrate the implications of the parameters given in Table 6, the migration of DE-71 through a typical composite liner consisting of geomembrane and either a geosynthetic clay liner (GCL) or compacted clay liner (CCL) was modelled using POLLUTEv5 (Rowe and Booker, 2004) for a hypothetical landfill. The composite liner was underlain by an attenuation layer (AL) such that the minimum distance from the underside of the geomembrane to the aquifer was 3.75 m in accordance with the regulatory require-

ments of MOE (1998). The aquifer was assumed to be 3 m thick (Fig. 7). The geomembrane partitioning and diffusion coefficients were selected based on the best estimates from the experimental data (Table 6). In the absence of other data, the diffusion coefficients for the PBDEs through the GCL, CCL and AL, were taken to be the same as those for chloride as reported in the literature (Table 8) because all these values are available in the literature and have been used in landfill design; these values are considered to be conservative because large molecules tend to diffuse slower than chloride (Rowe et al., 2004) and PBDEs are likely to partition to any organic matter in the soil and this could be considered in a site specific evaluation, however consistent with the approach adopted in developing the generic designs in MOE (1998), no sorption is assumed here (since the organic matter is unknown in a generic case). The initial source concentration of each congener of PBDEs in the landfill leachate was obtained from personal communication with Environment Canada (Table 9) and this presently represents the best available data. Calculations such as those discussed below can be repeated with improved data as it becomes available. To establish the most critical congener, the ratio (comax/cguideline) of the maximum initial concentration in the landfill, comax, to the allowable concentration in water, cguideline, was calculated for the four congeners where guidelines existed (Table 9). Based on this, congener 100 represented the most critical case since an almost 1000-fold reduction in concentration would be required before it would meet the guideline and hence it was considered the critical contaminant for containment by the landfill barrier system. Since there was no information on the mass of PBDEs in the landfill waste in the literature, it was (likely conservatively) assumed that the ratio of the mass of PBDEs per unit mass of waste

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

9

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx

Fig. 7. A hypothetical landfill with two typical configurations of the barrier system.

Table 8 Properties assumed for contaminant transport through the barrier system examined.

a b c

Layer

Thickness (H) (m)

Density (q) (g/cm3)

Porositya (n) (–)

Hydraulic conductivitya (k) (m/s)

Diffusion coefficient (Dg) (m2/a)

Dispersivity (a) (m)

Geomembrane GCL CCL AL

0.0015 0.007 0.75 3.743b/3.0c

0.946 1.9 1.9 1.9

1.0 0.7 0.4 0.3

– 2.0
1010 1.0
109 1.0
107

Varies 5.0
103a 2.0
102a 2.0
102a

– 0.4 0.4 0.4

Rowe and Brachman (2004). For GCL case. For CCL case.

Table 9 Concentrations of DE-71 congeners in landfill leachate at 10 landfills across Canada (personal communication with Environment Canada 2009).

a

Congener (No. of Br)

Minimum (ng/L)

Maximum (ng/L)

Average (ng/L)

Guideline (ng/L)a

comax/cguidelines

47 (4) 85 (5) 99 (5) 100 (5) 154 (6) 153 (6) Total

0.63 0.04 1.23 0.29 0.19 0.21 2.59

558 47.1 1000 210 119 160 2094

65.7 4.6 100.6 27.4 15.7 18.8 233

24 – 3.9 0.23 – 120 –

23 – 256 913 – 1.3 –

to the initial concentration was similar to that of chloride (based on MOE, 1998), although more research is required to establish the likely mass of PBDEs in landfills. It was assumed that the landfill had an average of 28 m of waste with a density of 1000 kg/m3 (Table 10). For both cases (either using a GCL or CCL in the composite liner – Fig. 7), six scenarios were modelled to calculate the peak concentration of each of the congeners in DE-71 (with a guideline limit) in the aquifer: (1) pure diffusion neglecting the diffusive resistance of the geomembrane but assuming no leakage (i.e., no holes) as a reference case,

Environment Canada (2013).

Table 10 Assumptions and input data for numerical modelling of DE-71 transfer through landfill barrier system. Parameter Landfill length Landfill width Infiltration Waste height Waste density Reference height of leachatea

a b

L W q(in) Hw

Unit

Value 1000 500 0.15 28 1000 20

Hr

m m m/a m kg/m3 m

Initial source concentrationb

Congener Congener Congener Congener

47 99 100 153

co

lg/L (mg/m3)

0.56 1.00 0.21 0.16

Total mass in waste per unit area

Congener Congener Congener Congener

47 99 100 153

mTC

g/m2

0.011 0.020 0.004 0.003

qw

Based on Rowe et al. (2004) as described in Saheli (2016). Based on personal communication with Environment Canada (2009).

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

10

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx

Table 11 Calculated leakage (in litres per hectare per day, lphd) through a composite liner for cases examined with 5 holes/ha in the geomembrane (rounded to 2 significant figures). Leakage (lphd) No wrinkle GCL

No wrinkle CCL

10 m wrinkle GCL

10 m wrinkle CCL

100 m wrinkle GCL

100 m wrinkle CCL

1000 m wrinkle GCL

1000 m wrinkle CCL

0.014

2.6

1.5

4.6

15

27

150

250

Note: wrinkle compressed width 2b = 0.1 m (Gudina and Brachman, 2006) and hole’s radius ro = 5.7 mm (Rowe and Brachman, 2004).

Table 12 Normalized model output for peak concentration of each congener in the aquifer to guideline value using GCL in the composite liner (GMB = geomembrane). Congener

47 99 100 153

cpeak/cguideline Diffusion no GMB

Diffusion GMB

5 holes/ha, no wrinkle

5 holes/ha, 10 m wrinkle

5 holes/ha, 100 m wrinkle

5 holes/ha, 1000 m wrinkle

2.6 29 103 0.2

0.03 0.04 0.3 2
104

0.03 0.05 0.3 2
104

0.1 1.1 4.0 0.01

0.9 10 36 0.05

6.2 68 243 0.4

Table 13 Normalized model output for peak concentration of each congener in the aquifer to guideline value using CCL in the composite liner (GMB = geomembrane). Congener

47 99 100 153

cpeak/cguideline Diffusion no GMB

Diffusion GMB

5 holes/ha, no wrinkle

5 holes/ha, 10 m wrinkle

5 holes/ha, 100 m wrinkle

5 holes/ha, 1000 m wrinkle

2.8 31 109 0.2

0.03 0.05 0.3 2
104

0.2 1.9 6.6 0.01

0.3 3.2 12 0.02

1.6 18 64 0.09

8.3 91 323 0.5

(2) pure diffusion through the composite liner with a geomembrane (no holes), (3) diffusion and leakage through 5 holes/ha in a geomembrane with no wrinkle, (4) diffusion and leakage through 5 holes/ha where 4 holes/ha were on planar geomembrane (i.e., in direct contact with the GCL) and 1 hole/ha was coincident with a 10 m long wrinkle in the geomembrane, (5) diffusion and leakage through 5 holes/ha where 4 holes/ha were on planar geomembrane and 1 hole/ha was coincident with a 100 m long wrinkle in the geomembrane, (6) diffusion and leakage through 5 holes/ha where 4 holes/ha were on planar geomembrane and 1 hole/ha was coincident with a 1000 m long wrinkle in the geomembrane. The leakage calculated for the cases with holes in the geomembrane is given in Table 11. For any given case, the leakage through a composite liner with a CCL was greater than with a GCL. This was primarily due to the lower GCL/clay liner interface transmissivity (as discussed by Rowe (2012)). The calculated peak concentrations in the aquifer for the four congeners of DE-71 with available guideline concentrations in water were calculated and normalized to the guideline concentrations (Tables 12 and 13). Any value exceeding unity may be considered as potentially unacceptable since it exceeds the guideline (for the case considered). As expected, congener 100 was the most critical of the six congeners because of the much higher value of co/cguideline in the leachate. The breakthrough curve in the aquifer for all the congeners showed a similar trend but here only those for the critical congener 100 are presented (Figs. 8 and 9). As the wrinkle length, and accordingly the leakage through the hole on the wrinkle increased (Table 11), the peak concentration of PBDEs in the aquifer occurred at earlier times (Figs. 8 and 9) and the peak concentration in the aquifer normalized to the guideline concentration

Fig. 8. Calculated concentration in aquifer normalized to guideline concentration for congener 100 for the composite liner with GCL for the six scenarios considered.

increased. The effectiveness of the geomembrane as a diffusion barrier to the PBDE is evident from considering the two pure diffusion cases (i.e., neglecting and considering the diffusive resistance of the geomembrane) in Tables 12 and 13. Ignoring the diffusive resistance of the geomembrane (i.e., not including geomembrane in the composite liner), congeners 47, 99 and 100 all exceed the guideline in the aquifer (by as much as two orders of magnitude for BDE-100). In contrast, considering the geomembrane’s observed diffusive resistance, all congeners meet the guidelines by at least a factor of three (worst case being BDE-100) and much more for three of the four congeners. Also, the partitioning of PBDEs to the geomembrane substantially slows down the PBDE migration through the barrier system which allows more PBDE mass to be removed by leachate collection system and slows the release; thus, the concentration in the aquifer reaches its peak after

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx

Fig. 9. Calculated concentration in aquifer normalized to guideline concentration for congener 100 for the composite liner with CCL for the six scenarios considered.

1500 years with the geomembrane compared to 300 years for the case ignoring the geomembrane resistance (Figs. 8 and 9). It is well recognized that even with a good geomembrane and good construction quality assurance, a geomembrane may have a few holes (typically 2.5–5 holes/ha is often assumed; Giroud and Bonaparte, 2001; Rowe, 2012). Modelling of advective-diffusive transport using commonly used computer programs is relatively straightforward for most contaminants including typical volatile organic compounds (e.g., benzene). However, the very low diffusion coefficient and extremely high partitioning of PBDEs (Table 7) makes the modelling of both mechanisms far more challenging for PBDEs and the modelling requires very careful consideration when there is a hole in wrinkle (Rowe, 1998, 2005, 2012; Rowe et al., 2012; Chappel et al., 2012). Modelling the leakage through 5 holes/ha for a geomembrane in intimate contact with an underlying GCL gave a combined advective-diffusive impact on the aquifer only marginally greater than for pure diffusion alone (Table 12). The situation was not so good for a composite liner with a compacted clay liner (Table 13). Because of the higher calculated leakage (due to the higher interface transmissivity; Rowe, 2012) with the CCL and the geomembrane, peak concentration in the aquifer for both congener 100 and 99 were predicted to exceed the guideline. A comparison of the two cases shows that the composite liner with a GCL is much better at minimizing PBDE migration. Considering a hole in even one wrinkle per hectare substantially increased the leakage (Table 11). For a 10 m long wrinkle/ha with a hole, congeners 99 and 100 were predicted to give unacceptable impacts for both composite liners. With a 100 m wrinkle/ha congeners 99 and 100 were predicted to exceed the guideline for both composite liners and also congener 47 with a CCL (Tables 12 and 13). For a 1000 m wrinkle congeners 47, 99 and 100 all exceeded the guideline for both composite liners for the parameters considered (Figs. 8 and 9 and Tables 12 and 13). In the example above, once the concentration in the leachate and pore water below the geomembrane drops to a low level, this means that there can also be desorption of PBDE from the geomembrane (and presumably the leachate collection pipes and other plastics in the landfill) and release, albeit at low concentrations, to the adjacent pore water over a long period of time. The release from the geomembrane was considered in the analysis described above (assuming the same partition coefficient, Sgf, for partitioning and departitioning, Sgf,’). However, this may also mean the HDPE liner material (and other plastics) may be regarded as hazardous material and that the waste cannot be simply dug up

11

and used for other purpose in the future without very careful consideration of this accumulation of PBDE in the plastics. Many assumptions have been made in the analyses reported herein, including: (i) the initial congener concentrations were the maximum reported (Table 9), (ii) the same mass to concentration ratio as chloride in the waste, (iii) no sorption to clay or in the attenuation layer, (iv) a relatively conservative hydraulic conductivity for the GCL and CCL, and (v) a relatively low (but not unrealistic) Darcy flux in the aquifer and a thin aquifer (i.e., relatively little dilution). Of particular note is the assumption that there was no sorption to the soil which is likely very conservative if there is any organic matter in the soil to which the PBDE could sorb. PBDE partitioning to clay has not been reported but the PBDE concentration in soil and sediment samples has been measured in various locations (Oros et al., 2005; Labadie et al., 2010; Jin et al., 2011; Yuan et al., 2012; Sun et al., 2015). A wide range of values for PBDE partitioning to soil and sediments has been reported (30–12,000 L/kg) (Wang et al., 2011; Olshansky et al., 2011). Since partitioning depends on many factors including dissolved organic carbon (DOC) in pore water (Wang et al., 2011), the amount, structure and properties of the soil organic matter (Olshansky et al., 2011), and the clay content (Sun et al., 2015), the assumption here of no sorption is likely extremely conservative but also appropriate since any alternative value would need to be on a site specific basis considering the factors just noted. In specific cases substantial sorption may be expected (although data is needed to confirm this on a site specific basis). Given the number of (likely) conservative assumptions, it is probable that the barrier systems would perform better than predicted here for containing PBDEs in a municipal solid waste landfill. Nevertheless, the results also suggest that more consideration needs to be given to PBDEs as contaminants of emerging concern in municipal solid waste landfills unless they can be constructed with essentially no holes in wrinkles (which is unlikely in most cases). 6. Conclusions Laboratory experiments for a commercial mixture of PBDEs (DE-71) were performed at room temperature to estimate its partitioning and diffusion coefficients for an HDPE geomembrane. For the specific conditions considered, it is concluded that:  The partitioning coefficient for the DE-71 mixture was 1.8
106 and was 0.7
106, 2.3
106, 4.9
106, 2.6
106, 6.0
106, and 7.5
106 for congeners 47, 85, 99, 100, 154, and 153 respectively.  As a general trend, the partitioning coefficients for the congeners in DE-71 increased with the number of attached bromines. This can be attributed to an increase in the molecular weight and volume and a decrease in water solubility.  The PBDEs in DE-71 had very substantially higher partitioning coefficients to the HDPE geomembrane than other organic compounds being around 3–5 orders of magnitude higher than for BTEX or even about two orders of magnitude higher than for PCBs.  The diffusion coefficients for the congeners in DE-71 through an HDPE geomembrane had estimated values of 3.7
1015 m2/s for the mixture and 5.7
1015 m2/s, 3.7
1015 m2/s, 3.3
1015 m2/s, 2.0
1015 m2/s, 1.5
1015 m2/s, and 2.5
1015 m2/s for congeners 47, 85, 99, 100, 154, and 153, respectively.  Generally, an increase in the number of attached bromine atoms decreased the diffusion coefficient of the DE-71 congeners through the geomembrane.

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

12

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx

 The diffusion coefficient of the DE-71 mixture was one order of magnitude smaller than that for PCB and two orders of magnitude smaller than that for BTEX which implies much slower migration of DE-71 through an HDPE geomembrane compared to these other organic compounds.  The best estimate partitioning and diffusion coefficients obtained from the experiments were used to perform analyses for a typical composite liner with two configurations (using either GCL or CCL) and six diffusion/leakage scenarios. The results indicated that the HDPE geomembrane significantly helped contain DE-71 inside the barrier system but that leakage through a hole in an interconnected wrinkle can significantly increase the transfer of PBDE through the composite liner as the wrinkle length increases. These results suggest that the construction details for composite liners can greatly affect their performance at containing PBDEs and that more attention needs to be paid to potential PBDE release in landfill design.

Acknowledgements Funding for the development of the research infrastructure was provided by the Canada Foundation for Innovation, the Ontario Innovation Trust, the Ontario Research Fund Award and Queen’s University. The research was being funded by the Natural Sciences and Engineering Research Council of Canada through Grant A1007 and by the Ontario Ministry of the Environment through the BestIn-Science program. The support of the Killam Trust in the form of a Killam Fellowship to Dr. Rowe is gratefully acknowledged. This investigation into the long-term performance of geosynthetic liner systems is being done in partnership with the Ontario Ministry of the Environment, Terrafix Geosynthetics Inc., Solmax International Inc., AMEC Earth and Environmental, Gartner Lee, Golder Associates, and CTT Group. We would also like to thank the Analytical Services Unit (ASU) of Queen’s University, Kingston, Canada for their support and use of their laboratory facilities. References Chao, K.P., Wang, P., Wang, Y.T., 2007. Diffusion and solubility coefficients determined by permeation and immersion experiments for organic solvents in HDPE geomembrane. J. Hazard. Mater. 142, 227–235. Chappel, M.J., Rowe, R.K., Brachman, R.W.I., Take, W.A., 2012. A comparison of geomembrane wrinkles for nine field cases. Geosynth. Int. 19 (6), 453–469. Danon-Schaffer, M.N., Grace, J.R., Wenning, R.J., Ikonomou, M.G., Luksemburg, W.J., 2006. PBDEs in landfill leachate and potential for transfer from electronic waste. Organohalogen Compd. 68, 1759–1762. Environment Canada, 2008. Technical Information on the Polybrominated Diphenyl Ethers Regulations. Environment Canada, 2013. Canadian Environmental Protection Act, 1999; Federal Environmental Quality Guidelines; Polybrominated Diphenyl Ethers (PBDEs). Ewais, A.M., Rowe, R.K., 2014. Effects of blown film process on initial properties of HDPE geomembranes of different thicknesses. Geosynth. Int. 21 (1), 62–82. Fries, E., Zarfl, C., 2012. Sorption of polycyclic aromatic hydrocarbons (PAHs) to low and high density polyethylene (PE). Environ. Sci. Pollut. Res. 19, 1296–1304. Giroud, J.P., Bonaparte, R., 2001. Geosynthetics in liquid containing structures. In: Geotechnical and Geoenvironmental Engineering Handbook. Kluwer Academic Publishing, Norwell, Mass, pp. 789–824. Gudina, S., Brachman, R.W.I., 2006. Physical response of geomembrane wrinkles overlying compacted clay. J. Geotech. Geoenviron. Eng. 132, 1346–1353. Haarstad, K., Borch, H., 2008. Halogenated compounds, PCB and pesticides in landfill leachate, downstream lake sediments and fish. J. Environ. Sci. Health, Part A 43 (12), 1346–1352. Islam, M.Z., Rowe, R.K., 2009. Permeation of BTEX through unaged and aged HDPE geomembranes. J. Geotech. Geoenviron. Eng. 135 (8), 1130–1140. Jin, J., Wang, Y., Liu, W., Yang, C., Hu, J., Cui, J., 2011. Polybrominated diphenyl ethers in atmosphere and soil of a production area in China: levels and partitioning. J. Environ. Sci. 23 (3), 427–433. Jones, D.D., 2016. Containment of organic contaminants using geosynthetics, PhD Thesis. Queen’s University, 562p. Joo, J.C., Nam, K., Kim, J.Y., 2005. Estimation of mass transport parameters of organic compounds through high density polyethylene geomembranes using a modified double-compartment apparatus. J. Environ. Eng. 131 (5), 790–799.

Kuramochi, H., Maeda, K., Kawamoto, K., 2007. Physicochemical properties of selected polybrominated diphenyl ethers and extension of the UNIFAC model to brominated aromatic compounds. Chemosphere 67, 1858–1865. Labadie, P., Tlili, K., Alliot, F., Bourges, C., Desportes, A., Chevreuil, M., 2010. Development of analytical procedures for trace-level determination of polybrominated diphenyl ethers and tetrabromobisphenol A in river water and sediment. Anal. Bioanal. Chem. 396, 865–875. Li, L., Xie, S., Cai, H., Bai, X., Xue, Z., 2008. Quantitative structure–property relationships for octanol–water partition coefficients of polybrominated diphenyl ethers. Chemosphere 72, 1602–1606. McWatters, R.S., Rowe, R.K., 2009. Transport of volatile organic compounds through PVC and LLDPE geomembranes from both aqueous and vapour phases. Geosynth. Int. 16 (6), 468–481. McWatters, R.S., Rowe, R.K., 2010. Diffusive transport of VOCs through LLDPE and two coextruded geomembranes. J. Geotech. Geoenviron. Eng. 136 (9), 1167– 1177. MOE, 1998. Ministry of Environment and Climate Change, Ontario Regulation 232/ 98, Landfilling sites. Environmental Protection Act. Odusanya, D.O., Okonkwo, J.O., Botha, B., 2009. Polybrominated diphenyl ethers (PBDEs) in leachates from selected landfill sites in South Africa. Waste Manage. 29 (1), 96–102. Olshansky, Y., Polubesova, T., Vetter, W., Chefetz, B., 2011. Sorption-desorption behavior of polybrominated diphenyl ethers in soils. Environ. Pollut. 159, 2375– 2379. Oros, D.R., Hoover, D., Rodigari, F., Crane, D., Sericano, J., 2005. Levels and distribution of polybrominated diphenyl ethers in water, surface sediments, and bivalves from the San Francisco estuary. Environ. Sci. Technol. 39, 33–41. Osako, K., Kim, Y., Sakai, S., 2004. Leaching of brominated flame retardants in leachate from landfills in Japan. Chemosphere 57 (10), 1571–1579. Palm, A., Cousins, I.T., Mackay, D., Tysklind, M., Metcalfe, C., Alaee, M., 2002. Assessing the environmental fate of chemicals of emerging concern: a case study of the polybrominated diphenyl ethers. Environ. Pollut. 117, 195–213. Park, J.K., Nibras, M., 1993. Mass flux of organic chemicals through polyethylene geomembranes. Water Environ. Res. 65 (3), 227–237. Park, J.K., Sakti, J.P., Hoopes, J.A., 1996. Transport of organic compounds in thermoplastic geomembranes I: mathematical model. J. Environ. Eng. 122 (9), 800–806. Rowe, R.K., 1998. Geosynthetics and the minimization of contaminant migration through barrier systems beneath solid waste. 6th International Conference on Geosynthetics, Atlanta, Georgia USA, vol. 1, pp. 27–103. Rowe, R.K., 2005. Long-term performance of contaminant barrier systems, 45th Rankine Lecture. Geotechnique 55 (9), 631–678. Rowe, R.K., 2012. Short and long-term leakage through composite liners, The 7th Arthur Casagrande Lecture. Can. Geotech. J. 49 (2), 141–169. Rowe, R.K., Booker, J.R., 1985. 1-D pollutant migration in soils of finite depth. J. Geotech. Eng. 111 (4), 479–499. Rowe, R.K., Booker, J.R., 2004. POLLUTE V. 7-1D Pollutant Migration Through a NonHomogeneous Soil. GAEA Environmental Engineering Ltd. Rowe, R.K., Brachman, R.W.I., 2004. Assessment of equivalence of composite liners. Geosynth. Int. 11 (4), 273–286. Rowe, R.K., Chappel, M.J., Brachman, R.W.I., Take, W.A., 2012. Field monitoring of geomembrane wrinkles at a composite liner test site. Can. Geotech. J. 49 (10), 1196–1211. Rowe, R.K., Quigley, R.M., Brachman, R.W., Booker, J.R., 2004. Barrier Systems for Waste Disposal Facilities, second ed. Spon Press, New York. Saheli, P.T., 2016. Transport of PBDE, MWCNTs, and BPA (contaminants of emerging concern) through modern landfill liners, PhD Thesis. Queen’s University, Kingston, Ontario, Canada, 387p. Sangam, H.P., Rowe, R.K., 2001. Migration of dilute aqueous organic pollutants through HDPE geomembranes. Geotext. Geomembr. 19, 329–357. Sjödin, A., Päpke, P., McGahee, E., Focant, J.-F., Jones, R.S., Pless-Mulloli, T., Leontjew Toms, L.-M., Herrmann, T., Müller, J., Needham, L.L., Patterson Jr., D.G., 2008. Concentration of polybrominated diphenyl ethers (PBDEs) in household dust from various countries. Chemosphere 73, S131–S136. Sun, Y., Yuan, G.-L., Li, J., Li, J.-C., Wang, G.-H., 2015. Polybrominated diphenyl ethers in surface soils near the Changwengluozha Glacier of Central Tibetan Plateau, China. Sci. Total Environ. 511, 399–406. Szlinder-Richert, J., Barska, I., Usydus, Z., Grabic, R., 2010. Polybrominated diphenyl ethers (PBDEs) in selected fish species from the southern Baltic Sea. Chemosphere 78, 695–700. Taghizadeh-Saheli, P., Rowe, R.K., Rutter, A., Brachman, R.W.I., 2011. Diffusive transport of PBDE through an HDPE geomembrane. Geo-Frontiers (March), 1141–1151, Dallas. Tang, J., Gallagher, D.L., Dietrich, A.M., 2013. Predicting permeation of organic contaminants into polyethylenes. J. Environ. Eng. 139 (2), 205–212. Tittlemier, S.A., Halldorson, T., Stern, G.A., Tomy, G.T., 2002. Vapor pressures, aqueous solubilities, and Henry’s law constants of some brominated flame retardants. Environ. Toxicol. Chem. 21 (9), 1804–1810. U.S. EPA, 2006. Polybrominated Diphenyl Ethers (PBDEs) Project Plan. U.S. Environmental Protection Agency. U.S. HHS, 2004. Toxicological Profile for Polybrominated Bisphenyls and Polybrominated Diphenyl Ethers. U.S. Department of health and human

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006

R.K. Rowe et al. / Waste Management xxx (2016) xxx–xxx services, Public Health Services Agency for Toxic Substances and Disease Registry. UNEP, 2004. The Environment in the News. United Nations Environment Programme. UNEP, 2010. New Science and Development in Our Changing Environment. United Nations Environment Programme Year Book.

13

Wang, W., Delgado-Moreno, L., Ye, Q., Gan, J., 2011. Improved measurements of partition coefficients for polybrominated diphenyl ethers. Environ. Sci. Technol. 45, 1521–1527. Yuan, G.-L., Han, P., Xie, W., Che, X.-C., Wang, G.-W., 2012. Altitudinal distribution of polybrominated diphenyl ethers (PBDEs) in the soil along Central Tibetan Plateau, China. Sci. Total Environ. 433, 44–49.

Please cite this article in press as: Rowe, R.K., et al. Partitioning and diffusion of PBDEs through an HDPE geomembrane. Waste Management (2016), http:// dx.doi.org/10.1016/j.wasman.2016.05.006