Equivalent Transport Parameters for Volatile Organic Compounds in Coextruded Geomembrane–Containing Ethylene-Vinyl Alcohol

Eun, Jongwan; Tinjum, James M.; Benson, Craig H.; Edil, Tuncer B.

doi:10.1061/(ASCE)GT.1943-5606.0001888

Open access

Technical Papers

May 10, 2018

Equivalent Transport Parameters for Volatile Organic Compounds in Coextruded Geomembrane–Containing Ethylene-Vinyl Alcohol

Authors: Jongwan Eun, M.ASCE [email protected], James M. Tinjum, M.ASCE [email protected], Craig H. Benson, F.ASCE [email protected], and Tuncer B. Edil, Dist.M.ASCE [email protected]Author Affiliations

Publication: Journal of Geotechnical and Geoenvironmental Engineering

Volume 144, Issue 7

https://doi.org/10.1061/(ASCE)GT.1943-5606.0001888

PDF

Abstract

A method to estimate the equivalent diffusion coefficient and partition coefficient of a coextruded ethylene-vinyl alcohol (EVOH) geomembrane (GM) is presented. The method is based on the weighted harmonic mean of diffusion coefficients of each layer and the partition coefficient of the outer layers. The method was evaluated using experimental data obtained with five volatile organic compounds (VOCs) typically found in landfill leachate [methylene chloride (MC), methyl tertiary butyl ether (MTBE), trichloroethylene (TCE), toluene (TOL), and chlorobenzene (CB)] and a coextruded GM composed of outer layers of low-density polyethylene (LLDPE) or high-density polyethylene (HDPE), an adhesive tie layer (maleic anhydride), and a core of EVOH. Equivalent diffusion coefficients computed with the recommended equation were found to be no different statistically from equivalent diffusion coefficients measured using a modified double-compartment apparatus. VOC transport through a composite liner using a coextruded EVOH GM was predicted using a finite-difference model with the equivalent transport parameters as input. Good agreement was obtained between predictions from the model and measured concentrations from a column test simulating a composite liner.

Introduction

Coextruded geomembranes (GMs) with an ethylene-vinyl alcohol (EVOH) layer sandwiched between polyethylene (PE) layers have been introduced as a means to reduce the flux of nonpolar organic contaminants and odiferous gases in barrier systems (Sangam and Rowe 2005; Edil 2007; McWatters and Rowe 2010, 2015; Eun et al. 2016). Experiments conducted by McWatters and Rowe (2010, 2011, 2015) showed that some organic compounds are transmitted at a significantly lower rate when a coextruded EVOH GM is used as a barrier relative to a relative conventional GM composed of a single polymer, such as linear low-density polyethylene (LLDPE), high-density polyethylene (HDPE), or polyvinyl chloride (PVC).

EVOH is a random copolymer of ethylene and vinyl alcohol that includes polar oxygen–hydrogen (-OH) groups. The chemical formulation of EVOH is

{({CH}_{2} - {CH}_{2})}_{m} {({CH}_{2} - CH)}_{n} - \begin{matrix} OH \end{matrix}

(1)

EVOH polymer is typically a combination of highly ordered crystalline structures interspersed within disordered amorphous regions and solvents (Zhang et al. 1999; McWatters and Rowe 2011). EVOH has high resistance to diffusion of nonpolar gases such as oxygen, nitrogen, volatile compounds, and helium due to the polarity of the alcohol group in the polymer (Zhang et al. 1999; Byun et al. 2007; McWatters and Rowe 2011, 2015). EVOH copolymer is graded by the mole fraction of ethylene, which affects the transport properties for organic contaminants and the flexibility of the polymer after extrusion.

Commercially available coextruded EVOH GMs have a five-layer structure [Figs. 1(a and b)] with an EVOH layer (0.05 mm) as the middle core and outer layers of LLDPE or HDPE. An adhesive polymer layer approximately 0.05 mm thick, referred to as a tie sheet, bonds the outer PE layers to the EVOH (Armstrong 2011). The maleic anhydride (

C_{4} H_{2} O_{3}

) used for the tie sheet consists of acid anhydride and maleic acid. C-H groups in the maleic anhydride graft to the ethylene in the HDPE or LLDPE outer layer, and an oxidation reaction grafts maleic anhydride to the OH group of EVOH [Fig. 1(c)].

Fig. 1. Layer of coextruded EVOH GM: (a) cross-sectional schematic of 1.5-mm coextruded EVOH GM (not to scale); (b) photograph of 1.5-mm coextruded EVOH GM; and (c) chemical formula for EVOH, tie, and outer layers in coextruded EVOH GM.

Partition and diffusion coefficients for GMs are needed to predict concentrations and solute fluxes in barrier systems (Foose et al. 2001, 2002; Chen et al. 2009). For simplicity and expediency, batch test methodologies are preferred for measuring partition and diffusion coefficients for GMs composed of a single polymer (Park et al. 2012a). However, coextruded GMs are composites composed of multiple polymers, and batch testing of the composite GM may not be appropriate for the determination of transport parameters because solute migration through the multilayer system is not represented properly (Dingemans et al. 2008; McWatters and Rowe 2010; Eun et al. 2014, 2017).

This paper illustrates a method to estimate equivalent transport parameters for coextruded EVOH GMs based on outcomes of a series of batch tests conducted on each component in the multilayer system. An equation for estimating the equivalent diffusion coefficient of coextruded EVOH GMs is presented and recommendations are made for assigning a partition coefficient to the equivalent system. Equivalent transport parameters for coextruded EVOH GMs estimated using this method are compared to equivalent transport parameters measured directly using a modified double-compartment apparatus (MDCA). In addition, experimental data from diffusion column tests simulating composite liners are compared to predictions of a VOC transport model using a numerical model using equivalent transport parameters computed with the method developed herein.

Equivalent Diffusive Transport through a Multilayer Geomembrane

Solute transport through a composite liner consisting of a multilayer system has been described by Foose et al. (2001), Huysmans and Dassargues (2007), Haijian et al. (2009), Li and Cleall (2010), and McWatters and Rowe (2010, 2015). The equivalent diffusion coefficient for a composite liner can be derived based on the principle of mass conservation in a manner similar to the calculation of the equivalent hydraulic conductivity of a layered system (Liu et al. 1998; Foose et al. 2001; Guyonnet et al. 2001; Huysmans and Dassargues 2007; Friedmann 2008; Li and Cleall 2010).

A schematic of the transport process through a coextruded GM composed of

n

layers is shown in Fig. 2. The steady state flux of solute through each layer is identical

J = J_{1} = J_{2} = \dots = J_{i} = \dots = J_{n}

(2)

where

J

= solute flux for the coextruded EVOH GM; and

J_{i}

= solute flux in the

i

th layer [

L^{- 2} T^{- 1}

]. According to Fick’s first law, the steady-state flux for the entire coextruded EVOH GM can be rewritten as

J = D_{eq} \frac{Δ C_{total}}{L_{g t}} = D_{eq} \frac{C_{1, in} - C_{n, out}}{L_{g t}}

(3)

where

D_{e q}

= equivalent diffusion coefficient [

L^{2} T^{- 1}

];

Δ C_{total}

= difference in concentration across the entire coextruded GM [

{ML}^{- 3}

];

L_{g t}

= total thickness of the coextruded GM [L];

C_{1, in}

= inward-gradient concentration at the upper surface of the first layer in the GM; and

C_{n, out}

= outward-gradient concentration at the lower surface of the

n

th layer in the GM [

{ML}^{- 3}

]. Partitioning between the solution and the surface of the outer layers of the GM is described by the partition coefficient (

K_{g}

) between the GM and the solute (Leo et al. 1971; Foose et al. 2001; Park et al. 2012a, b)

K_{g} = \frac{C_{g}}{C_{w}}

(4)

where

C_{w}

= equilibrium concentration in liquid in the contact with the GM (leachate or pore water) [

M / L^{3}

]; and

C_{g}

= equilibrium concentration of the organic compound in the polymer at the surface of the GM [M/M] (Step A and D in Fig. 2). From Eq. (4),

C_{1, in}

and

C_{n, out}

in the GM can be defined as

C_{1, in} = K_{g 1} C_{w, in}

(5)

C_{n, out} = K_{g n} C_{w, out}

(6)

where

C_{w, in}

and

C_{w, out}

= inward-gradient and outward-gradient concentrations [

{ML}^{- 3}

] in the leachate for inward-gradient and outward-gradient of the GM; and

K_{g 1}

and

K_{g n}

= partition coefficients between the solution and the first and

n

th layer of the GM, respectively.

Fig. 2. Schematic of solute transport through a multilayered GM composed of $n$ layers with continuity in concentration assumed at internal interfaces.

Similar to the flux for the entire coextruded GM, the flux from the

i

th layer (Step B in Fig. 2) is

J_{i} = \frac{D_{g i}}{L_{g i}} (C_{i, in} - C_{i, out}) = \frac{D_{g i}}{L_{g i}} Δ C_{i}

(7)

where

D_{g i}

= diffusion coefficient for the

i

th layer [

L^{2} T^{- 1}

];

C_{i, in}

= inward-gradient concentration for the

i

th layer [

{ML}^{- 3}

];

C_{i, out}

= outward-gradient concentration for the

i

th layer [

{ML}^{- 3}

];

L_{g i}

= thickness of the

i

th layer [L]; and

Δ C_{i}

= difference in concentration between the upper and lower surfaces of the

i

th layer [

{ML}^{- 3}

].

Rewriting Eq. (7), the difference in concentration at the

i

th layer can be presented in terms of flux from the

i

th layer, thickness, and diffusion coefficient of the

i

th layer

J_{i} = \frac{C_{i, in} - C_{i, out}}{L_{g i} / D_{g i}}

(8)

By rearranging Eq. (8), the difference between the inward-gradient and outward-gradient concentrations of the

i

th layer is obtained

C_{i, in} - C_{i, out} = J_{i} \frac{L_{g i}}{D_{g i}}

(9)

Combining Eqs. (2), (3), and (9) yields

(C_{1, in} - C_{1, out}) \dots + (C_{i, in} - C_{i, out}) + \dots (C_{n, in} - C_{n, out}) = J (\frac{L_{g 1}}{D_{g 1}} + \frac{L_{g 2}}{D_{g 2}} + \dots + \frac{L_{g i}}{D_{g i}} + \dots + \frac{L_{g n}}{D_{g n}})

(10)

Accordingly, the total difference in

i

th concentration across the GM can be written as

Δ C_{total} = \sum_{i = 1}^{n} (C_{i, in} - C_{i, out}) = J_{n} \sum_{i = 1}^{n} (\frac{L_{g i}}{D_{g i}})

(11)

If the concentration between the

i

th and (

i + 1

)th layers at the internal interfaces in the multilayered system (e.g., coextruded GM) (Step C in Fig. 2) is assumed to be continuous, i.e.

C_{i, out} |_{L g i} = C_{i + 1, in} |_{L g i}

(12)

then conservation of mass at the interface yields

J_{n} = D_{i} \frac{\partial C_{i, out}}{\partial L_{g i}} |_{L g i} = D_{i + 1} \frac{\partial C_{i + 1, in}}{\partial L_{g i + 1}} |_{L g i}

(13)

This formulation implies that partitioning, or preference for one polymeric material relative to the other, is insignificant at the internal interfaces. Though partitioning between adjacent layers in a solid does exist, the preference for one material over the other is diminished by strong chemical bonds between layers [see previous discussion of Fig. 1(c)]. Thus, partitioning between adjacent coextruded polymers is modest compared to the sharp at the interface between water and a crystalline polymer or mineral solid (Zhang and Niu 2004).

With Eq. (12), Eq. (11) can be rewritten as

(C_{1, in} - C_{n, out}) = J_{n} \sum_{i = 1}^{n} (\frac{L_{g i}}{D_{g i}})

(14)

and Eq. (14) can be rearranged in terms of the summation of flux from the first to

n

th layer as

J_{n} = \frac{(C_{1, in} - C_{n, out})}{\sum_{i = 1}^{n} (L_{g i} / D_{g i})}

(15)

The flux for the composite system [Eq. (3)] is identical to flux for each layer [Eq. (15)] [

J = J_{1} = \dots = J_{i}

, Eq. (2)]. Combining Eqs. (3) and (15) yields

D_{eq} \frac{(C_{1, in} - C_{n, out})}{L_{g t}} = \frac{(C_{1, in} - C_{n, out})}{\sum_{i = 1}^{n} (L_{g i} / D_{g i})}

(16)

By rearranging Eq. (16), the equivalent diffusion coefficient for a coextruded GM consisting of

n

layers is

D_{e q} = \frac{L_{g t}}{\sum_{i = 1}^{n} (L_{g i} / D_{g i})}

(17)

For the assumptions made, Eq. (17) shows that

D_{e q}

for a multilayer GM is influenced by the thickness and diffusion coefficient of each layer within the GM in much the same way that the equivalent vertical hydraulic conductivity of a horizontally layered medium is influenced by the thickness and hydraulic conductivity of each layer.

McWatters and Rowe (2015) proposed an equation analogous to Eq. (17) to estimate an equivalent permeation coefficient (product of partition and diffusion coefficient) for a coextruded EVOH GM using the harmonic mean. In their analysis, they considered the outer polyethylene layers and the EVOH core, but ignored the maleic anhydride tie sheet. Because their equation used a permeation coefficient, partitioning at the internal interfaces of the coextruded EVOH GM was accounted for explicitly. However, the permeation coefficients used in their analysis corresponded to a water–polymer system in which substantial partitioning occurs due to preference for one material over another. In contrast, in a coextruded GM, transport between layers occurs in a polymer–polymer system, with strong bonds between the layers creating a transition zone that minimizes partitioning [e.g., as suggested in Friedmann (2008)]. The tie sheet in a coextruded EVOH geomembrane further enhances the transition between materials, thereby minimizing preference and partitioning.

Materials and Methods

Volatile Organic Compounds

Five VOCs were used: methylene chloride (MC), methyl tertiary butyl ether (MTBE), trichloroethylene (TCE), toluene (TOL), and chlorobenzene (CB). These VOCs represent each of the types of VOCs (i.e., alkanes, ethers, alkenes, aromatic hydrocarbons, and halogenated aromatic hydrocarbons) among the 31 VOCs detected in lysimeters beneath 34 municipal solid waste (MSW) landfills in Wisconsin (Klett et al. 2005). These VOCs have been used extensively in previous studies (Park et al. 2012a, b; Eun et al. 2014, 2017). General properties of these VOCs are described in Table 1.

Table 1. Properties of VOCs used in experiments

Property	MC	MTBE	TCE	TOL	CB
Chemical formula	${CH}_{2} {Cl}_{2}$	${CH}_{3} ─ O ─ C ({CH}_{3})_{3}$	$CHCl = {CCl}_{2}$	$C_{6} H_{5} ─ {CH}_{3}$	$C_{6} H_{5} Cl$
Type of compound	Alkane	Ether	Alkene	Arene	Arene
Molecular weight ( $g / mol$ )	84.93	88.15	131.39	92.14	112.56
Density ( $g / mL$ )	1.33	0.740	1.46	0.867	1.11
$\log K_{o w}$	1.31	0.94	2.42	2.69	2.78
Solubility ( $mg / L$ )	20,000	48,000	1,100	515	500
Vapor pressure (kPa)	57.5	32.4	10.0	3.72	1.59
Dielectric constant	8.9	4.5	3.4	2.4	2.7
Melting point (°C)	$- 97.2$	$- 108.6$	$- 84.7$	$- 94.95$	$- 45.31$
Boiling point (°C)	40	55	87.21	110.63	131.72

Sources: Data from Schwarzenbach et al. (2003); Lake and Rowe (2005).

Note: CB = chlorobenzene; $K_{o w}$ = partition coefficient between octanol and water; MC = methylene chloride; MTBE = methyl tertiary butyl ether; TCE = trichloroethylene; and TOL = toluene.

VOC solutions were prepared by filling a 1-L flask with distilled and deionized (DDI) water. Sodium azide (0.05%) was added to prevent microbial activity. The flask was filled completely to minimize loss of VOCs in the headspace. Weight-volume calculations were used to compute the mass required to achieve the desired VOC concentration, and liquid VOC was injected into the water using a 100-mL gas-tight syringe. When using multiple solutes, the compounds were injected in order of decreasing density; however, MTBE was injected first because of its high solubility and ability to act as a co-solvent for the other VOCs. The flasks were immediately covered with a glass stopper and sealed with Teflon tape. Sealed flasks were placed on a magnetic stir plate for 24 h before use (Park et al. 2012a, b; Eun et al. 2014, 2017).

VOC concentrations were measured using a Shimadzu GC-2010 (Shimadzu, Kyoto, Japan) gas chromatograph (GC) equipped with an auto sampler, flame ionization detector (FID), and Restek RTX-624 (Restek, Bellefonte, Pennsylvania) column (

length = 30 m

, inner

diameter = 0.32 mm

, and film

thickness = 1.80 μm

). Temperatures of the injection port and the FID were set at 280°C. The sample split ratio was 3.0 and the injection volume was 0.5 mL. The column had an initial column temperature of 35°C and hold time of 5 min. The column was then heated to 100°C at a rate of

10 ° C / \min

, held at 100°C for 3 min, heated to 220°C at a rate of

40 ° C / \min

, and held at 220°C. The total run time for each injection was 20 min (Park et al. 2012a, b; Eun et al. 2014, 2017). The detection limit was

0.82 mg / L

for MC,

0.53 mg / L

for MTBE,

0.34 mg / L

for TCE,

0.31 mg / L

for TOL, and

0.22 mg / L

for CB.

Geomembrane

Two coextruded EVOH GMs (1.0 mm and 1.5 mm thick) with different outer layers (LLDPE for 1.0 mm GM and HDPE for 1.5 mm GM) were evaluated (henceforth labeled as 1.0-mm-EVOH GM and 1.5-mm-EVOH GM, respectively). The coextruded GM was produced by Raven Industries (Sioux Falls, South Dakota) using EVOH with

3.5 % {CaCO}_{3}

inert filler supplied by Kuraray America (Pasadena, Texas). The barrier core layer in all cases was approximately 0.04 to 0.05 mm thick and was composed of 42 mol% of ethylene copolymer EVOH. A smooth 1.5-mm-thick black HDPE GM (henceforth labeled as 1.5-mm HDPE GM supplied by Raven Industries) was also tested for comparative purposes.

Tests were also conducted on 0.5-mm-thick layers of pure EVOH and 0.5-mm-thick maleic anhydride tie sheets to identify the transport parameters (i.e., partition and diffusion coefficients) of each layer. Properties of the GM and each layer are in Table 2.

Table 2. Engineering properties of the coextruded EVOH GM

Property	Method	Unit	HDPE	LLDPE	Maleic anhydride	EVOH sheet
Thickness	ASTM D5944	mm	1.5	1.0	0.5	0.5
Puncture resistance	ASTM D4833	N	360	280	130	280
Tear resistance	ASTM D1004	N	$98 \times 98$	$72 \times 72$	$40 \times 40$	$45 \times 50$
Load at break	ASTM D6693	N	12,000	10,300	6,500	8,500
Elongation at break	ASTM D6693	%	950	600	400	180
Tensile strength	ASTM D6693	$kN / m$	16.7	13.5	8.3	14.5

Equilibrium Batch Test

Equilibrium batch tests were conducted to obtain partition coefficients of the VOCs for the coextruded GM, conventional HDPE GM, pure EVOH sheet, and tie maleic anhydride adhesive sheet. The test procedure followed that of Park et al. (2012a). Membrane samples were cut into strips (

17 \times 80 mm

), 1–5 strips (1.4–6.2 g of GM) were placed in an amber glass vial (40 mL), and solution was added to fill the vial without head space (solid:liquid

ratio = 0.04 ∶ 0.15

). The vial was sealed with a screw cap and polytetrafluoroethylene (PTFE)-coated septum. As a control, two vials containing VOC alone (i.e., without GM strips) were prepared, tested, and processed in parallel with the equilibrium GM tests as controls. For these tests, the VOC concentration was measured at the start and on completion of the tests. Data from the control tests were used to correct for losses, although the corrections were small to negligible (

< 3 %

) (Park et al. 2012a).

Solutions having initial concentrations of 10, 40, 70, and

100 mg / L

of each VOC were used for tests conducted with multisolute mixtures of VOCs. The solutions were transferred to each vial from 1-L flasks using a peristaltic pump. When VOC solution was moved into each vial from the flask, the PTFE tube on the pump was attached to the bottom of the vial to minimize loss of VOC (Parker and Britt 2012). Filled and sealed vials were tumbled in a rotator at 30 rpm for 8 days at 23.5°C, which is sufficient time to reach equilibrium (Park et al. 2012a). After tumbling, the vials were centrifuged at 2000 rpm (429 g) for 15 min to ensure VOCs were mixed homogeneously. The supernatant was transferred to autosampler vials using a glass syringe without opening the cap of the vial. An analyses were conducted using the GC as described previously. Three replicates were conducted for each concentration. Partition coefficients were computed by fitting a linear isotherm [Eq. (4)] to the sorption data using least-squares regression.

Kinetic Batch Test

Kinetic batch tests were performed on 1.5-mm-thick coextruded EVOH GM and single polymer layers (conventional HDPE GM, pure EVOH, and tie adhesive) to obtain the diffusion coefficient of each material making up the coextruded EVOH GM. The testing method was adopted from Sangam and Rowe (2001a), Joo et al. (2004), and Park et al. (2012a). Membrane samples were cut into strips (

17 \times 80 mm

) similar to those in the equilibrium batch tests. Approximately 1.4–6.2 g of GM (1–5 strips) was placed in an amber glass vial (40 mL), and solution (

100 mg / L

for all five VOCs) was added to fill the vial (solid:liquid

ratio = 0.04 ∶ 0.15

). The vial was then sealed with a screw cap and PTFE-coated septum. Control tests were also conducted without GM specimens.

The vials were prepared and handled using the same method used for the batch tests. Duplicate vials were prepared (including controls) for periodic decommissioning during the experiment. Samples were decommissioned and sampled after 1, 2, 3, 5, 7, and 9 days from the start of the test. The supernatant was transferred to autosampler vials using a glass syringe (without opening the cap of the vial) and was analyzed by GC using the methods described previously. Concentrations were adjusted for losses using the data from the controls as done for the equilibrium batch tests.

The VOC concentration data were analyzed assuming a planar sheet of geomembrane initially devoid of solute that is suspended with both sides exposed in a well-stirred solution of fixed volume and mass. The analytical solution for this system is given by Crank (1975) as follows:

\frac{C_{t}}{C_{0}} = \exp (\frac{D_{g} t {(K_{g})}^{2}}{A^{2}}) erfc {(\frac{D_{g} t {(K_{g})}^{2}}{A^{2}})}^{0.5}

(18)

where

C_{t}

= concentration of the VOC in the solution [

{ML}^{- 3}

] at time

t

[T];

C_{0}

= initial concentration of the organic compound in the solution [

{ML}^{- 3}

];

D_{g}

= diffusion coefficient [

L^{2} T^{- 1}

] in the GM;

K_{g}

= partition coefficient (dimensionless); and

A

= half thickness of the solution in contact with both sides of the GM [L], which is calculated by dividing the volume of the solution by the area of the GM. Eq. (18) was fit to the data using nonlinear least-squares regression.

Modified Double-Compartment Apparatus Test

The modified double-compartment apparatus test was conducted to obtain partition and diffusion coefficients for the coextruded GM as shown in Fig. 3. The general procedure for the test followed previous studies (Park et al. 1996; Joo et al. 2005; Park et al. 2012a). A brass testing column (200 mm height, and 155 mm diameter) was used that was composed of upper and lower compartments, with the GM being evaluated placed at the joint between the compartments. The MDCA experiments had

S : L = 0.025

, which is slightly less than the S:L for the batch experiments (0.04–0.15). Therefore, solute in the MDCA experiments had comparable or greater availability to the GM surface as in the batch experiment. Thus, the variation in S:L between the MDCA and batch tests is believed to have had a negligible effect on sorption and diffusion. Park et al. (2012a) report that diffusion and partition coefficients obtained from batch tests are unaffected by S:L ratios over a similar range as those used in the present study.

Fig. 3. Modified double-compartment apparatus used to obtain partition and diffusion coefficients for the coextruded EVOH GM.

In a previous version of the apparatus (Park et al. 1996, 2012a), VOC losses occurred at the flange of the compartments. In this study, a double-sealing joint was used to minimize losses (Fig. 3). The double-sealing joint had two PTFE gaskets installed on either side of the GM. This mechanical and material modification of the seal minimized losses of VOCs (Eun et al. 2017). PTFE gaskets were also used to seal the column to the top and bottom plates. A control test using a stainless steel plate in lieu of the GM was conducted simultaneously. No VOCs were detected in the effluent (lower) reservoir during the control test, and the change in concentration in the upper reservoir was less than 2.3% over a 40-d monitoring period. Data from the control tests were used to make corrections for losses.

The influent (upper) compartment was filled with VOC solution having an initial concentration of

100 mg / L

for all five VOCs. The effluent (lower) compartment was filled with DDI water spiked with 0.05% sodium azide. During the test, the solution in both reservoirs was stirred continuously to ensure a well-mixed solution. Samples were collected periodically from the sampling ports to monitor concentrations in the influent and effluent reservoirs. The sampling ports included a brass connector and cap with two septa to extract samples with minimal loss of VOCs during sampling.

Partition and diffusion coefficients for each VOC were determined by inversion using the 1-D finite-difference model from Park et al. (2012a) and least-squares regression. This model solves Fick’s second law [Eq. (19)] for

0 < z < L_{g}

, where

L_{g}

is the thickness of the GM [L]. Eq. (4) is used to define the relationship between

C_{g}

at the upper and lower surfaces of the GM in Fick’s second law [Eq. (19)] and the aqueous phase concentrations measured in the reservoirs. For the initial condition at

t = 0

, the concentration of the bottom reservoir and the GM (

0 < z < L_{g}

) was set at

0 mg / L

.

Experimental Results

Equilibrium Batch Tests for Coextruded EVOH GM and HDPE GM

Sorption isotherms obtained from equilibrium batch tests for the 1.5-mm EVOH GM, 1.0-mm EVOH GM, and 1.5-mm HDPE GM are shown in Fig. 4. Partition coefficients were computed by fitting the linear isotherm [Eq. (4)] to the sorption data using least-squares regression. The partition coefficients are summarized in Table 3. The isotherms are approximately linear, which is consistent with previous studies on VOC sorption to geomembranes for concentrations less than

100 mg / L

(Edil et al. 1995; Park et al. 1996; Sangam and Rowe 2005; Park et al. 2012a).

Fig. 4. Sorption isotherms of VOCs for 1.5-mm and 1.0-mm coextruded EVOH GMs with HDPE or LLDPE outer layer, respectively, and conventional 1.5-mm-HDPE GM: (a) MC; (b) MTBE; (c) TCE; (d) TOL; and (e) CB.

Table 3. Partition coefficients (

K_{g}

) for 1.5-mm HDPE GM, 1.5-mm EVOH GM, and 1.0-mm EVOH GM from equilibrium batch testing

Type	$K_{g}$ ( $L / kg$ )
Type	MC	MTBE	TCE	TOL	CB
1.5-mm HDPE GM	3.38	0.82	69.02	88.55	112.03
1.5-mm EVOH GM with HDPE outer layer	3.08	0.74	62.08	84.34	107.10
1.0-mm EVOH GM with LLDPE outer layer	5.80	1.90	157.85	195.70	171.09

Sorption of VOCs to the HDPE outer layer of the 1.5-mm EVOH and the 1.5-mm HDPE GM were similar for each batch test (Fig. 4), and paired

F

-tests showed that the partition coefficients were not statistically different (

p = 0.46 > 0.05

). For example, the partition coefficients of TOL for 1.5-mm EVOH and 1.5-mm HDPE GM were 88.55 [

L / kg

] and 84.34 [

L / kg

], respectively. The EVOH film in the coextruded GM was thin relative to the HDPE outer layer (less than 8% of total thickness) and therefore had minimal impact on sorption.

In contrast, the 1.0-mm EVOH GM having an LLDPE outer layer had a much higher partition coefficient (

1.9 - 195.7 L / kg

) compared to other GMs. For example, the partition coefficients of MC and TOL for the LLDPE GM were 1.7 and 2.2 times higher than those for the HDPE GM, respectively. The lower crystallinity of LLDPE allows VOCs to sorb more readily to the GM, resulting in higher

K_{g}

compared to the more crystalline HDPE (Park et al. 1996).

Equilibrium Batch Tests for Pure EVOH, Tie Sheet, and HDPE GM

Sorption isotherms obtained from equilibrium batch tests are shown in Fig. 5 for the 0.5-mm pure EVOH, tie sheet, and HDPE outer layer. Partition coefficients for the VOCs are summarized in Table 4. The HDPE had the highest partition coefficient for the nonpolar VOCs (e.g., TCE, TOL, and CB) (

69.2 - 112.0 L / kg

), and the pure EVOH sheet had the lowest partition coefficient for the nonpolar VOCs (e.g., TCE, TOL, and CB) (

5.5 - 5.6 L / kg

). In contrast, the pure EVOH sheet had the highest partition coefficient for the more polar VOCs (e.g., MC and MTBE). In these batch tests, the EVOH sheet had unfettered access to water and could swell, whereas less water for swelling would be available when the EVOH was encased in polyethylene in a coextruded GM, potentially affecting the partition coefficient.

Fig. 5. Sorption isotherms of VOCs for HDPE outer layer, maleic anhydride (tie), and pure EVOH sheet: (a) MC; (b) MTBE; (c) TCE; (d) TOL; and (e) CB.

Table 4. Partition coefficients (

K_{g}

) of 0.5-mm pure EVOH, tie sheet, and 1.5-mm HDPE GM making up the coextruded EVOH GM from equilibrium batch tests

Type	$K_{g}$ ( $L / kg$ )
Type	MC	MTBE	TCE	TOL	CB
Pure EVOH	7.08	7.50	5.61	5.52	5.63
Maleic anhydride (tie)	5.61	3.80	19.90	29.20	24.84
1.5-mm HDPE GM	3.38	0.82	69.02	88.56	112.03

The partition coefficient for VOCs is highly dependent on their polarity (Park and Nibras 1993; Sangam and Rowe 2001b; Joo et al. 2004; Nefso and Burns 2007; Park et al. 2012a). Chemical properties related to the polarity of VOCs, including the octanol-water partition coefficient (

K_{o w}

) and aqueous solubility (A.S.), were investigated to evaluate how they affected sorption to each material making up the coextruded GM. Relationships between the chemical properties and partition coefficients are shown in Fig. 6. An empirical relationship between the partition coefficient and

K_{o w}

for HDPE GM has been reported by others (Park and Nibras 1993; Sangam and Rowe 2001b; Joo et al. 2004; Nefso and Burns 2007; Park et al. 2012a). In most polymer-penetrant systems, partition coefficients generally increase with an increase in similarity between the components, according to “like dissolves like” (August and Taztky 1984; Sangam and Rowe 2001b). In general, the permeation affinity has the following order: alcohols < acids < nitro derivatives < aldehydes < ketones < esters < ethers < aromatic hydrocarbons < halogenated aromatic hydrocarbons (August and Taztky 1984; Sangam and Rowe 2001a).

Fig. 6. (a) Relationship between the partition coefficient and octanol-water partition coefficient (log $K_{o w}$ ); and (b) A.S. of the VOCs for the LLDPE outer layer, HDPE outer layer, maleic anhydride tie sheet, and EVOH core.

Increasing

\log K_{o w}

is indicative of greater hydrophobicity, increased preference for nonpolar polymers (HDPE, LLDPE), and decreased preference for the more polar EVOH. The

\log K_{o w} - \log K_{g}

relationships in Fig. 6(a) illustrate this affinity for HDPE and

LLDPE : CB > TOL > TCE > MTBE > MC

. For the EVOH sheet, the relationship between the partition coefficient and

K_{o w}

is slightly negative and statistically significant (

F = 126.1 ≫ 1.0

, correlation coefficient = 0.988). Because of the alcohol group (-OH) in EVOH, the chemical hydrophobia decreases and the VOC has lower affinity for the EVOH film (lower partition coefficient) with increasing

\log K_{o w}

. Maleic anhydride, which is the acid anhydride of maleic acid, exhibits an intermediate behavior between EVOH and HDPE and LLDPE.

A strong relationship between

\log K_{g}

versus logA.S. also exists [Fig. 6(b)]. The increasing affinity for HDPE and LLDPE with decreasing A.S. reflects the increasing hydrophobicity of the VOCs. The positive slope for the EVOH data is statistically significant (

F value = 170.23 ≫ 1.0

, correlation coefficient = 0.991). As with

\log K_{o w}

, maleic anhydride exhibits a slope between that of EVOH and HDPE and LLDPE.

Kinetic Batch Tests

Typical solution concentrations from the kinetic batch tests are shown in Fig. 7, with fits of Eq. (18) shown as dashed lines. Tests on the pure EVOH sheet and maleic anhydride tie sheet did not reach equilibrium due to the slow rate of diffusion into these materials. McWatters and Rowe (2015) also observed a slow rate of diffusion in tests conducted with EVOH. For both materials, however, a systematic monotonic reduction in concentration occurred in the aqueous phase, providing the transient behavior necessary to compute the diffusion coefficient, as described in Park et al. (2012a). Partition coefficients obtained from the equilibrium batch tests (Table 4) were input to Eq. (18), which was then fit to the kinetic batch test data by adjusting only the diffusion coefficient (

D_{g}

) as recommended by Park et al. (2012a). The fits were made using nonlinear least-squares regression with Microsoft Excel Solver using the generalized reduced gradient algorithm. Because the tests were not run to equilibrium, the analysis implicitly assumed that the behavior observed during the test continued throughout the duration of the test.

Fig. 7. Solution concentrations versus time from kinetic batch tests on pure EVOH and tie sheets along with fits of Eq. (18): (a) MC; and (b) MTBE.

Diffusion coefficients for the 1.5-mm HDPE GM obtained using this method were compared to diffusion coefficients measured by Park et al. (2012a) for the same VOCs (Table 5). For all five compounds, the diffusion coefficients found in this study were within 12% of those obtained by Park et al. (2012a), except MTBE (= 29%). On average, the diffusion coefficient differed by only 15.4%. Paired

F

-tests indicated no statistically significant difference (

p = 0.18 > 0.05

) between the diffusion coefficients obtained from this study and those from Park et al. (2012a).

Table 5. Diffusion coefficients (

D_{g}

) for 1.5-mm HDPE GM from kinetic batch tests

Type	$D_{g}$ ( $\times 10^{- 13} m^{2} / s$ )
Type	MC	MTBE	TCE	TOL	CB
HDPE	10.1	11.05	5.17	4.06	2.86
HDPE (Park et al. 2012a)	8.86	7.74	5.45	3.77	3.96

Diffusion coefficients of each material, obtained from the kinetic batch tests, are summarized in Table 6. The outer layer of the HDPE GM had the highest diffusion coefficient (

2.86 - 10.1 \times 10^{- 13} m^{2} / s

), and the pure EVOH sheet had the lowest diffusion coefficient (

0.0049 - 0.0261 \times 10^{- 13} m^{2} / s

). The diffusion coefficient for the EVOH sheet was two and three orders of magnitude lower than diffusion coefficient for the tie sheet, HDPE, and coextruded EVOH GM.

Table 6. Diffusion coefficients (

D_{g}

) of materials making up the 1.5-mm EVOH GM and the 1.5-mm HDPE GM from kinetic batch tests

Type	$D_{g}$ ( $\times 10^{- 13} m^{2} / s$ )
Type	MC	MTBE	TCE	TOL	CB
Pure EVOH	0.0232	0.0261	0.0068	0.0049	0.0059
Maleic anhydride (tie)	1.5	2.18	1.16	0.92	1.27
1.5-mm HDPE GM	10.1	11.05	5.17	4.06	2.86
1.5-mm coextruded EVOH GM with HDPE outer layer	9.67	10.24	5.15	3.24	3.63

Similar to the partition coefficient, diffusion coefficients for the VOCs from the batch tests are similar for the 1.5-mm-thick coextruded EVOH and HDPE GM (Table 6). Paired

F

-tests showed no statistical difference between the diffusion coefficients for the EVOH GM and HDPE GM. In the batch test, the thickness of the EVOH film exposed to the solution is negligible in comparison to the thicker HDPE surrounding the EVOH film in the coextruded GM (McWatters and Rowe 2010; Eun et al. 2014, 2017).

Relationships between the diffusion coefficients and

\log K_{o w}

and logA.S. are shown in Fig. 8. The diffusion coefficients reflect the hydrophobicity of the VOCs. Diffusion coefficients for the HDPE, maleic anhydride tie sheet, and the EVOH GM decrease with an increase in

K_{o w}

or a decrease in A.S. (increasing hydrophobicity). The more polar EVOH core exhibits the opposite trend.

Fig. 8. Relationship between diffusion coefficients and chemical properties of VOCs for HDPE outer layer, maleic anhydride tie sheet, 1.5-mm EVOH GM, and pure EVOH sheet: (a) octanol-water partition coefficient ( $\log K_{o w}$ ); and (b) aqueous solubility.

Modified Double-Compartment Apparatus Tests

Typical relative concentrations of MC and CB in the influent (upper) and effluent (lower) reservoirs of a MDCA test over time are shown in Fig. 9. Similar data were obtained for the other compounds. The tests were conducted for approximately 160 days. Equilibrium was not established during the test period due to the very slow diffusion through the EVOH GM. Computations indicated that testing for up to a decade may have been required to achieve equality between the influent effluent concentrations. Testing for this long duration was beyond the scope of the study, but the transient behavior of the concentration was sufficient for determination of partition and diffusion coefficients.

Fig. 9. Relative concentrations in the influent (upper reservoir, UR) and effluent (lower reservoir, LR) reservoirs in modified double-compartment apparatus (MDCA) test for 1.5-mm coextruded EVOH GM along with fits obtained with the finite-difference model: (a) MC; and (b) CB.

Concentrations of VOCs with lower affinity for HDPE (MC and MTBE) decreased in the upper reservoir more slowly [Fig. 9(a)], whereas concentrations of VOCs with higher affinity for HDPE (CB, TOL, and TCE) decreased rapidly [e.g., Fig. 9(b)]. In contrast, for all VOCs, concentrations in the lower reservoir increased slowly due to a very low rate of diffusion through the coextruded GM because the solute diffused minimally through the EVOH film in the GM. Low concentrations (

< 5 mg / L

) existed in the effluent reservoir for the entire testing time (160 days).

The dashed lines in Fig. 9 correspond to fits with the finite difference model (Joo et al. 2005; Park et al. 2012a). Good correspondence exists between the predicted and measured influent concentrations, whereas the effluent concentrations were underpredicted slightly, indicating that the diffusion coefficient may have been underestimated. Partition and diffusion coefficients fitted to the experimental data simultaneously using least-squares regression are summarized in Table 7.

Table 7. Partition (

K_{g}

) and diffusion coefficients (

D_{g}

) obtained from MDCA tests on 1.5-mm EVOH GM

Type	Unit	MC	MTBE	TCE	TOL	CB
$K_{g}$ of 1.5-mm EVOH GM	$L / kg$	3.77	0.91	77.4	97.4	117.1
$D_{g}$ of 1.5-mm EVOH GM	$\times 10^{- 13} m^{2} / s$	0.587	0.697	0.182	0.140	0.161

Partition coefficients for the 1.5-mm EVOH GM obtained from the MDCA test and the equilibrium batch tests are compared in Fig. 10(a). All of the data fall slightly above the

1 ∶ 1

line (

< 10 %

), indicating that the MDCA test yields a higher partition coefficient than the equilibrium batch test. However, a paired

F

-test indicated that the two sets of partition coefficients are not statistically different (

p = 0.41 > 0.05

).

Fig. 10. Comparison of transport parameters: (a) partition coefficient ( $K_{g}$ ) for 1.5-mm HDPE GM measured using double-compartment (DC) test by Park et al. (2012a) versus $K_{g}$ from equilibrium batch tests in this study along with $K_{g}$ for 1.5-mm EVOH GM measured using MDCA tests and equilibrium batch tests in this study; (b) diffusion coefficient ( $D_{g}$ ) for HDPE GM measured using DC by Park et al. (2012a) versus $D_{g}$ for HDPE measured with the MDCA in this study along with $D_{g}$ for 1.5-mm HDPE GM versus $D_{g}$ for the EVOH GM with HDPE outer layer in this study using the MDCA.

Diffusion coefficients for the coextruded EVOH GM obtained from the MDCA test are compared to those of HDPE GM obtained from kinetic batch tests in Fig. 10(b). Diffusion coefficients for the 1.5-mm EVOH GM (

0.14 - 0.59 \times 10^{- 13} m^{2} / s

) are approximately 16–29 times lower than those of the 1.5-mm HDPE GM (

2.86 - 11.05 \times 10^{- 13} m^{2} / s

), illustrating the role of the EVOH film in reducing diffusive transport of VOCs.

Validation of Equivalent Diffusion Coefficient of a Coextruded EVOH Geomembrane

Calculated and Measured Diffusion Coefficients

Equivalent diffusion coefficients (

D_{e q}

) estimated by Eq. (17) for a 1.5-mm coextruded EVOH GM are compared to diffusion coefficients for the same GM measured using MDCA tests in Fig. 11. Partition and diffusion coefficients of each layer (i.e., HDPE outer layer, maleic anhydride tie, pure EVOH film) measured individually using equilibrium and kinetic batch tests (Tables 4 and 6) were used to calculate equivalent diffusion coefficients using Eq. (17).

Fig. 11. Comparison of equivalent diffusion coefficients ( $D_{e q}$ ) estimated using Eq. (17) and measured using the modified double-compartment apparatus (MDCA) for a 1.5-mm coextruded EVOH GM.

The diffusion coefficients for TCE, TOL, and CB estimated using Eq. (17) and measured with the MDCA differ by less than 4%, whereas the estimated and measured diffusion coefficients for MC and MTBE differed by less 7.5%, on average. A paired

F

-test indicated that the diffusion coefficients estimated with Eq. (17) are not statistically different from those measured with the MDCA test (

p = 0.39 > 0.05

).

The smaller decrease in concentration of the more polar MC and MTBE during the MDCA tests may have contributed to greater uncertainty in the diffusion coefficients for these compounds computed from the MDCA data (Fig. 9), resulting in the greater difference between the estimated and measured diffusion coefficients in Fig. 11. Another factor may have been reduced swelling of the EVOH in the coextruded GM in the MDCA apparatus relative to the batch test due to less availability of water with the EVOH encased in polyethylene. Swelling of the EVOH can alter the diffusion coefficient of the EVOH layer (McWatters and Rowe 2015), potentially contributing to the larger differences between the estimated and measured diffusion coefficients for MC and MTBE.

Measured and Predicted VOC Transport through Composite Liner

Concentrations predicted from a numerical simulation of the composite liner experiment in Eun et al. (2017) are shown in Fig. 12 along with data from the column experiments. The predictions were made using equivalent diffusion coefficients predicted with Eq. (17) and the partition coefficient for the outer later obtained from the batch tests. The polar VOCs (MC, MTBE) are shown in Fig. 12(a), and the nonpolar VOCs (TCE, TOL, and CB) are shown in Fig. 12(b). The composite liner in the tests consisted of a 1.5-mm coextruded EVOH GM underlain by a 120-mm-thick clay liner. The initial concentration of VOCs in these column tests was

100 mg / L

, and the concentration in the influent reservoir was kept constant by injecting VOCs periodically. Concentrations were measured in the clay liner at 20 mm below the lower surface of the coextruded EVOH GM. Additional details of the column test are described in Eun et al. (2014, 2017).

Fig. 12. Measured and predicted concentrations in the clay liner component of column tests simulating a composite liner described in Eun et al. (2017). Concentrations measured in clay liner 20 mm below GM. Predictions made with finite-difference model using equivalent diffusion coefficient estimated with Eq. (17) and partition coefficient for the outer layer from the equilibrium batch test: (a) data from MC and MTBE (more polar VOCs); and (b) TCE, TOL, and CB (nonpolar VOCs).

The physical mechanism of transport through composite liners includes three steps: (1) adsorption of the permeating species at homogeneous GM surface, (2) diffusion through the GM in the direction of the lower chemical potential, and (3) desorption of the permeating species from the GM surface into the pore water in the soil beneath the GM (Park and Nibras 1993; Park et al. 1996; Sangam and Rowe 2001b; Joo et al. 2004, 2005; Park et al. 2012b). Sorption between outer layers of the GM from the contaminant solution is described by Eq. (4). Diffusion in the GM is described in 1D using Fick’s second law (Park and Nibras 1993; Park et al. 1996; Sangam and Rowe 2001b; Joo et al. 2004, 2005; Park et al. 2012a, b)

\frac{\partial C_{g}}{\partial t} = D_{g} \frac{\partial^{2} C_{g}}{\partial z^{2}}, - L_{g} < z < 0

(19)

where

t

= time [T];

D_{g}

= diffusion coefficient of the organic compound in the GM [

L^{2} / T

];

z

= spatial coordinate in the direction of diffusion [L]; and

L_{g}

= thickness of the GM [L]. Eq. (19) is valid if the diffusion coefficient in the GM is spatially and temporally invariant (Foose 2002; Foose et al. 2002).

For a saturated soil layer, 1-D mass transport of a nondecaying solute via diffusion can be expressed as (Hashimoto et al. 1964; Freeze and Cherry 1979)

\frac{\partial C_{s}}{\partial t} = \frac{D^{*}}{R} \frac{\partial^{2} C_{s}}{\partial z^{2}}, z > 0

(20)

where

C_{s}

= concentration of the organic compound in the pore water of the soil layer [

{ML}^{- 3}

];

z

= spatial coordinate in the direction of mass transport [L];

t

= elapsed time [T];,

R

= retardation factor; and

D^{*}

= effective diffusion coefficient [

L^{2} / T

].

Eqs. (4), (19), and (20) were used to simulate VOC transport through the composite liner in the column test via the finite difference method (Foose et al. 2001; Park et al. 2012b; Eun et al. 2017) using the Crank-Nicholson procedure (Foose et al. 2001, 2002; Eun 2014). The upper boundary condition was assumed to be constant concentration (

100 mg / L

), and the lower boundary at the base of the cell was assumed to be no flux. Zero concentration throughout the profile was stipulated as the initial condition, and continuity of concentration at the interface between the GM and the clay liner was imposed (Foose et al. 2002). Parameters input for each layer of the composite liner are in Table 8. The GM was simulated as a homogeneous layer using the equivalent diffusion coefficient estimated by Eq. (17) and the partition coefficient of the HDPE outer layer obtained from equilibrium batch tests.

Table 8. Transport parameter input for simulation of diffusion column test with a composite liner containing coextruded GM

Compound	Compacted clay (Eun 2014)		1.5-mm EVOH GM
Compound	Partition coefficient ( $K_{d}$ ) ( $L / kg$ )	Diffusion coefficient ( $D *$ ) ( $\times 10^{- 9} m^{2} / s$ )	Partition coefficient ( $K_{g}$ ) ( $L / kg$ ) from MDCA test	Diffusion coefficient ( $D_{g}$ ) ( $\times 10^{- 13} m^{2} / s$ ) from Eq. (17)
MC	0.097	0.14	3.77	0.648
MTBE	0.092	0.138	0.91	0.731
TCE	0.178	0.147	77.4	0.197
TOL	0.192	0.146	97.4	0.141
CB	0.217	0.139	117.1	0.168

Predicted and measured VOC concentrations for the column tests are shown in Fig. 12. Concentrations were measured in the clay liner beneath the coextruded EVOH GM, after the VOCs had diffused through the entire thickness of the GM. Concentrations predicted in the clay layer by the model are consistent with the measured concentrations. The only substantial deviation is for MC, which was predicted to break through faster than measured and to increase in concentration more slowly than measured. The diffusion coefficient for MC calculated by Eq. (17) also showed the greatest deviation from the diffusion coefficient obtained from the MDCA (Fig. 11).The good agreement between predicted and measured concentrations for the other VOCs is consistent with the good agreement between the estimated and measured effective diffusion coefficients shown in Fig. 11.

The good agreement between the predicted and measured concentrations in Fig. 12 suggests that a reliable prediction can be achieved when the coextruded GM is approximated as a single-layer material with the diffusion coefficient estimated with Eq. (17) and the partition coefficient assigned as the partition coefficient corresponding to the outermost material in the GM.

Summary and Conclusions

A methodology to estimate equivalent transport parameters of VOCs for multilayer coextruded geomembranes with an EVOH core was evaluated in this study. Experiments were conducted with five VOCs commonly found in leachates for municipal solid waste landfills using equilibrium batch tests, kinetic batch tests, and a modified double-compartment apparatus using a coextruded GM with a core of EVOH, an outer layer of HDPE or LLDPE, and tie layers of maleic anhydride. VOC concentrations in a simulated composite liner using the coextruded EVOH GM were predicted using a finite-difference model parameterized using equivalent transport parameters computed with the recommended estimation methodology using data from the batch tests. The predicted concentrations were compared to concentrations measured in a composite liner experiment.

The findings indicate that an equivalent diffusion coefficient for a multilayer coextruded EVOH GM can be estimated using the thickness-weighted harmonic mean [Eq. (17)] of the diffusion coefficients for each of the layers, and the partition coefficient of the outer layer can be assigned as partition coefficient of the multilayer coextruded GM. Equivalent diffusion coefficients and partition coefficients assigned using this approach generally agreed with diffusion coefficients and partition coefficients of the multilayer coextruded EVOH GM measured using the MDCA. Predictions of VOC concentrations in a composite liner using an EVOH GM made using a finite-difference model parameterized with the equivalent transport parameters generally agreed with measured concentrations from a composite liner experiment. The greatest deviation between predicted and measured concentrations was obtained for the VOC for which the largest difference was obtained between the estimated and measured equivalent diffusion coefficient.

Acknowledgments

Financial support for this study was provided by Kuraray Co., Ltd. and a research fellowship from the Geosynthetic Institute (GSI). The opinions, findings, and conclusions expressed herein are solely those of the authors and my not reflect those of Kuraray Co., Ltd. or the GSI.

References

Armstrong, R. B. 2011. “Improving performance of geosynthetics for containment of volatile organic compounds through the use of ethylene vinyl alcohol (EVOH).” In Geo-Frontier 2011. Reston, VA: ASCE.

Google Scholar

August, H., and R. Taztky. 1984. “Permeability of commercial available polymeric liners for hazardous landfill leachate organic constituents.” In Int. Conf. on Geomembrane, 151–156. Roseville, MN: IFAI.

Google Scholar

Byun, Y. J., S. I. Hong, K. B. Kim, D. H. Jeon, J. M. Kim, W. S. Whiteside, and H. J. Park. 2007. “Physical and chemical properties of γ-irradiated EVOH film.” Radiat. Phys. Chem. 76 (6): 974–981. https://doi.org/10.1016/j.radphyschem.2006.09.005.

Google Scholar

Chen, Y., H. Xie, H. Ke, and R. Chen. 2009. “An analytical solution for one-dimensional contaminant diffusion through multi-layered system and its applications.” Environ. Geol. 58 (5): 1083–1094. https://doi.org/10.1007/s00254-008-1587-3.

Google Scholar

Crank, J. 1975. The mathematics of diffusion, 2nd ed., 11–24. Oxford, UK: Clarendon.

Google Scholar

Dingemans, M., J. Dewulf, L. Braeckman, H. van Langenhove, K. Friess, V. Hynek, and M. Sipek. 2008. “Mass transfer characteristics for VOC permeation through flat sheet porous and composite membrane: The impact of the different membrane layers on the overall membrane resistance.” J. Membr. Sci. 322 (1): 234–242. https://doi.org/10.1016/j.memsci.2008.05.043.

Google Scholar

Edil, T. B. 2007. “Is aqueous-phase VOC transport from modern landfills a potential environmental problem?” In Proc., Sardinia 2007, 11th Int. Landfill Symp. Padova, Italy: International Waste Working Group.

Google Scholar

Edil, T. B., W. S. Wambold, and J. K. Park. 1995. “Partitioning of VOCs in clay liner materials.” In Geoenvironment 2000, ASCE GSP 46, 775–790. Reston, VA: ASCE.

Google Scholar

Eun, J. 2014. “Volatile organic compound (VOC) and methane transport through composite barrier with co-extruded geomembrane containing ethylene-vinyl alcohol (EVOH).” Ph.D. dissertation, Dept. of Civil and Environmental Engineering, Univ. of Wisconsin-Madison.

Google Scholar

Eun, J., J. M. Tinjum, C. H. Benson, and T. B. Edil. 2014. “Volatile organic compound (VOC) transport through composite liner with co-extruded geomembrane containing ethylene-vinyl alcohol (EVOH).” In Geo-Congress 2014. Atlanta: ASCE.

Google Scholar

Eun, J., J. M. Tinjum, C. H. Benson, and T. B. Edil. 2017. “Comparison of volatile organic compound (VOC) transport between composite liners with HDPE and ethylene-vinyl alcohol (EVOH) co-extruded geomembranes.” J. Geotech. Geoenviron. Eng. 143 (6): 04017010. https://doi.org/10.1061/(ASCE)GT.1943-5606.0001484.

Google Scholar

Eun, J., M. Yilmaz, J. M. Tinjum, C. H. Benson. 2016. “Hydrogen sulfide (

H_{2} S

) transport through simulated interim cover with conventional and co-extruded ethylene-vinyl alcohol (EVOH) geomembranes.” In Geo-Chicago 2016. Chicago: ASCE.

Google Scholar

Foose, G. J. 2002. “Transit-time design for diffusion through composite liners.” J. Geotech. Geoenviron. Eng. 128 (7): 590–601. https://doi.org/10.1061/(ASCE)1090-0241(2002)128:7(590).

Google Scholar

Foose, G. J., C. H. Benson, and T. B. Edil. 2001. “Equivalent of composite geosynthetic clay liners as barrier to volatile organic compounds.” In Geosynthetic ‘99 Conf. Roseville, MN: IFAI.

Google Scholar

Foose, G. J., C. H. Benson, and T. B. Edil. 2002. “Comparison of solute transport in three composite liners.” J. Geotech. Geoenviron. Eng. 128 (5): 391–403. https://doi.org/10.1061/(ASCE)1090-0241(2002)128:5(391).

Google Scholar

Freeze, R. A., and J. A. Cherry. 1979. Ground water. Englewood Cliffs, NJ: Prentice-Hall.

Crossref

Google Scholar

Friedmann, M. H. 2008. Principles and models of biological transport. Berlin, Germany: Springer.

Crossref

Google Scholar

Guyonnet, D., P. Perrochet, B. Come, J. J. Seguin, and A. Parriaux. 2001. “On the hydro-dispersive equivalence between multi-layered mineral barriers.” J. Contam. Hydrol. 51 (3–4): 215–231. https://doi.org/10.1016/S0169-7722(01)00127-9.

Google Scholar

Haijian, X., C. Yunmin, K. E. Han, T. Xiaowu, and C. Renpeng. 2009. “Analysis of diffusion-adsorption equivalency of landfill liner systems for organic contaminants.” J. Environ. Sci. 21 (4): 552–560. https://doi.org/10.1016/S1001-0742(08)62307-4.

Google Scholar

Hashimoto, I., K. B. Deshpande, and H. C. Thomas. 1964. “Peclet numbers and retardation factors for ion exchange columns.” Ind. Eng. Chem. Fundam. 3 (3): 213–218. https://doi.org/10.1021/i160011a007.

Google Scholar

Huysmans, M., and A. Dassargues. 2007. “Equivalent diffusion coefficient and equivalent diffusion accessible porosity of a stratified porous medium.” Transp. Porous Medium 66 (3): 421–438. https://doi.org/10.1007/s11242-006-0028-6.

Google Scholar

Joo, J. C., J. Y. Kim, and K. Nam. 2004. “Mass transfer of organic compounds in dilute aqueous solutions into high density polyethylene geomembranes.” J. Geotech. Geoenviron. Eng. 130 (2): 175–183. https://doi.org/10.1061/(ASCE)0733-9372(2004)130:2(175).

Google Scholar

Joo, J. C., K. Nam, and J. Y. Kim. 2005. “Estimation of mass transport parameters of organic compounds through high density polyethylene geomembranes using a modified double-compartment apparatus.” J. Geotech. Geoenviron. Eng. 131 (5): 790–799. https://doi.org/10.1061/(ASCE)0733-9372(2005)131:5(790).

Google Scholar

Klett, N., T. B. Edil, C. H. Benson, and J. Connelly. 2005. Evaluation of volatile organic compounds in Wisconsin landfill leachate and lysimeter samples. Madison, WI: Univ. of Wisconsin at Madison.

Google Scholar

Lake, C. B., and R. K. Rowe. 2005. “A comparative assessment of volatile organic compound (VOC) sorption to various types of potential GCL bentonites.” Geotext. Geomembr. 23 (4): 323–347. https://doi.org/10.1016/j.geotexmem.2005.01.001.

Google Scholar

Leo, A., C. Hansch, and D. Elkins. 1971. “Partition coefficients and their uses.” Chem. Rev. 71 (6): 525–616. https://doi.org/10.1021/cr60274a001.

Google Scholar

Li, Y., and P. J. Cleall. 2010. “Analytical solutions for contaminant diffusion in double-layered porous media.” J. Geotech. Geoenviron. Eng. 136 (11): 1542–1554. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000365.

Google Scholar

Liu, C., W. P. Ball, and J. H. Ellis. 1998. “An analytical solution to the one-dimensional solute advection-dispersion equation in multi-layer porous media.” Transp. Porous Media 30 (1): 25–43. https://doi.org/10.1023/A:1006596904771.

Google Scholar

McWatters, R. S., and R. K. Rowe. 2010. “Diffusive transport of VOCs through LLDPE and two coextruded geomembranes.” J. Geotech. Geoenviron. Eng. 136 (9): 1167–1177. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000345.

Google Scholar

McWatters, R. S., and R. K. Rowe. 2011. “Sorption and diffusion of BTEX through thin-film EVOH.” In Geo-Frontiers 2011, 2073–2083. Reston, VA: ASCE.

Google Scholar

McWatters, R. S., and R. K. Rowe. 2015. “Permeation of volatile organic compounds through thin film membranes and coextruded LLDPE/EVOH/LLDPE geomembranes.” J. Geotech. Geoenviron. Eng. 141 (2): 04014091. https://doi.org/10.1061/(ASCE)GT.1943-5606.0001209.

Google Scholar

Nefso, E. K., and S. E. Burns. 2007. “Comparison of the equilibrium sorption of five organic compounds to HDPE, PP, and PVC geomembranes.” Geotext. Geomembr. 25 (6): 360–365. https://doi.org/10.1016/j.geotexmem.2006.12.002.

Google Scholar

Park, J. K., and M. Nibras. 1993. “Mass flux of organic chemicals through polyethylene geomembranes.” Water Environ. Res. 65 (3): 227–237. https://doi.org/10.2175/WER.65.3.6.

Google Scholar

Park, J. K., J. P. Sakti, and J. A. Hoopes. 1996. “Transport of organic compounds in thermoplastic geomembranes. I: Mathematical model.” J. Environ. Eng. 122 (9): 800–806. https://doi.org/10.1061/(ASCE)0733-9372(1996)122:9(800).

Google Scholar

Park, M. G., C. H. Benson, and T. B. Edil. 2012a. “Comparison of batch and double compartment tests for measuring VOC transport parameters in geomembranes.” Geotext. Geomembr. 31 (1): 15–30. https://doi.org/10.1016/j.geotexmem.2011.09.001.

Google Scholar

Park, M. G., T. B. Edil, and C. H. Benson. 2012b. “Modeling volatile organic compound transport in composite liners.” J. Geotech. Geoenviron. Eng. 138 (6): 641–657. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000630.

Google Scholar

Parker, L., and S. Britt. 2012. “The effect of bottle fill rate and pour technique on the recovery of volatile organics.” Ground Water Monit. Rem. 32 (4): 78–86. https://doi.org/10.1111/j.1745-6592.2012.01401.x.

Google Scholar

Sangam, H., and R. Rowe. 2005. “Effect of surface fluorination on diffusion through a high density polyethylene geomembrane.” J. Geotech. Geoenviron. Eng. 131 (6): 694–704. https://doi.org/10.1061/(ASCE)1090-0241(2005)131:6(694).

Google Scholar

Sangam, H. P., and R. K. Rowe. 2001a. “Migration of dilute aqueous organic pollutants through HDPE geomembranes.” Geotext. Geomembr. 19 (6): 329–357. https://doi.org/10.1016/S0266-1144(01)00013-9.

Google Scholar

Sangam, H. P., and R. K. Rowe. 2001b. “The role of HDPE geomembranes in retarding the diffusive migration of organic contaminants through composite liner systems.” Proc., Sardinia 2001, 8th Int. Landfill Symp. Padova, Italy: International Waste Working Group.

Google Scholar

Schwarzenbach, R. P., P. M. Gschwend, and D. M. Imboden. 2003. Environmental organic chemistry. 2nd ed. Hoboken, NJ: Wiley-Interscience.

Google Scholar

Zhang, L. Z., and J. L. Niu. 2004. “Modeling VOCs emissions in a room with a single-zone multi-component multi-layer technique.” Build. Environ. 39 (5): 523–531. https://doi.org/10.1016/j.buildenv.2003.10.005.

Google Scholar

Zhang, Z., I. J. Britt, and M. A. Tung. 1999. “Water absorption in EVOH films and its influence on glass transition temperature.” J. Polym. Sci. Part B: Polym. Phys. 37 (7): 691–699. https://doi.org/10.1002/(SICI)1099-0488(19990401)37:7%3C691::AID-POLB20%3E3.0.CO;2-V.

Google Scholar

Information & Authors

Information

Published In

Journal of Geotechnical and Geoenvironmental Engineering

Volume 144 • Issue 7 • July 2018

Copyright

This work is made available under the terms of the Creative Commons Attribution 4.0 International license, http://creativecommons.org/licenses/by/4.0/.

History

Received: Sep 12, 2016

Accepted: Dec 11, 2017

Published online: May 10, 2018

Published in print: Jul 1, 2018

Discussion open until: Oct 10, 2018

Authors

Affiliations

Jongwan Eun, M.ASCE [email protected]

Assistant Professor, Dept. of Civil Engineering, Univ. of Nebraska–Lincoln, Lincoln, NE 68588 (corresponding author). Email: [email protected]

View all articles by this author

James M. Tinjum, M.ASCE [email protected]

Associate Professor, Engineering Professional Development and Geological Engineering, Univ. of Wisconsin–Madison, Madison, WI 53706. Email: [email protected]

View all articles by this author

Craig H. Benson, F.ASCE [email protected]

Dean, School of Engineering and Hamilton Endowed Chair in Engineering, Univ. of Virginia, Charlottesville, VA 22904. Email: [email protected]

View all articles by this author

Tuncer B. Edil, Dist.M.ASCE [email protected]

Professor Emeritus, Dept. of Civil and Environmental Engineering Geological Engineering, Univ. of Wisconsin–Madison, Madison, WI 53706. Email: [email protected]

View all articles by this author

Metrics & Citations

Metrics

Citations

Download citation

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

Abstract

Introduction

Equivalent Diffusive Transport through a Multilayer Geomembrane

Materials and Methods

Volatile Organic Compounds

Geomembrane

Equilibrium Batch Test

Kinetic Batch Test

Modified Double-Compartment Apparatus Test

Experimental Results

Equilibrium Batch Tests for Coextruded EVOH GM and HDPE GM

Equilibrium Batch Tests for Pure EVOH, Tie Sheet, and HDPE GM

Kinetic Batch Tests

Modified Double-Compartment Apparatus Tests

Validation of Equivalent Diffusion Coefficient of a Coextruded EVOH Geomembrane

Calculated and Measured Diffusion Coefficients

Measured and Predicted VOC Transport through Composite Liner

Summary and Conclusions

Acknowledgments

References

Information

Published In

Copyright

History

Authors

Affiliations

Metrics

Citations

Download citation

Cited by

Figures

Other

Share

Copy the content Link

Share with email

Share

Request Username

Create a new account

Change Password

Password Changed Successfully

Verify Phone

Congrats!