at SciVerse ScienceDirect

Geotextiles and Geomembranes 33 (2012) 7e14

Contents lists available

Geotextiles and Geomembranes

journal homepage: www.elsevier .com/locate/geotexmem

Forensic examination of field GCL performance in landfill capping and miningcontainment applications

John Buckley a,*, Will P. Gates b, Daniel T. Gibbs a

aGeosynthetic Centre of Excellence, 11 Production Avenue, Molendinar, Queensland 4214, Australiab Smectech Research Consulting, 24 Chapel Road, Moorabbin, Victoria 3189, Australia

a r t i c l e i n f o

Article history:Received 18 March 2011Received in revised form7 November 2011Accepted 7 February 2012Available online 24 April 2012

Keywords:Geosynthetic clay linersLandfill final coverLandfill cappingMining containmentExhumationField performance

* Corresponding author. Tel.: þ61 7 5594 8600; faxE-mail address: j.buckley@geofabrics.com.au (J. Bu

0266-1144/$ e see front matter Crown Copyright � 2doi:10.1016/j.geotexmem.2012.02.006

a b s t r a c t

Geosynthetic Clay Liners (GCL’s) have been used extensively in landfill capping and mine containmentapplications in the Australian environment, since 1996, and while they have been widely accepted byregulatory authorities and design engineers over this time, some questions remain over their long-termin-field service life and performance. To better evaluate the field performance of GCL’s in terms ofhydraulic performance, changes to bentonite mineralogy and physical characteristics, an on-going studyof the in-field performance of GCL’s exhumed from landfill capping and mine containment sites aroundAustralia has been initiated. Our preliminary results presented in this paper support previous researchindicating that superior field hydraulic performance in GCL’s are related to high RMD values, whichprovide a low risk of bentonite calcium for sodium exchange. However, for one site with marked wet/drycycling, calcium for sodium exchange due to low RMD values, coupled with low moisture content, hardlyinfluenced the GCL hydraulic performance after 6 years.

Crown Copyright � 2012 Published by Elsevier Ltd. All rights reserved.

1. Introduction

Geosynthetic clay liners (GCL’s) originally evolved from thewaterproofing industry in the form of bentonite-impregnatedcardboard panels. Over the years, the product characteristicschanged (adhesive bonded, membrane backed, stitch bonded) tomake them more suitable for use in civil and environmental engi-neering applications (Bouazza et al., 2006). Needle-punched GCL’swere developed for the global lining market in Europe in 1987 andsubsequently, production of these products began in Australia in1996. Since then, demand for needle-punched GCL’s has grownwith their recognition as cost effective alternatives to traditionalcompacted clay liners in many applications (Koerner, 2005).

Most GCL’s consist of a layer of sodium bentonite clay securedbetween two geotextiles. Egloffstein (2002) distinguished betweennatural sodium bentonites and “mixed occupation” sodiumbentonites which may contain significant concentrations of othercations, such as calcium, magnesium and/or potassium. Theswelling and hydraulic performance of these “mixed occupation”sodium bentonites are improved by activating with soda-ash(sodium carbonate) to replace the other cations with sodium

: þ61 7 5563 3727.ckley).

012 Published by Elsevier Ltd. All

cations. The primary function of the bentonite component is tolimit the migration of fluids (Rowe, 2001; Bouazza et al., 2006). Thegeotextile component is essentially the carrier/reinforcementnetwork which allows the placement of a uniform barrier layer ofprocessed bentonite (generally w5e8 mm thick) and reinforce-ment of the bentonite layer to restrict free swelling and improvehydrated shear strength characteristics.

Typical liner applications include landfill (municipal andhazardous), water reservoirs, tailings facilities and mine wasteresidue impoundments (Gates et al., 2009; Hornsey et al., 2010).The performance parameters (e.g., swell index, hydraulic conduc-tivity, shear strength) for these applications are generally highlyvariable across GCLs, due primarily to differences in manufacturingandmaterial sourcing, and these need to be thoroughly understoodin order to accurately predict field behaviour.

The development of GCL’s has been driven primarily from twopoints of reference. Firstly, as an alternative to compacted clayliners (e.g. Francisca and Glastein, 2010), which because they aremade up of locally available clayey soils, are normally only specifiedin terms of thickness and hydraulic conductivity. Secondly, devel-opment derived from quality control testing carried out by GCLmanufacturers. This type of testing requires fast turn around andgenerally is used to confirm that product quality remains withinmanufacturing tolerances. The reliance on manufacturing testprocedures derived from the textile industry has resulted in

rights reserved.

J. Buckley et al. / Geotextiles and Geomembranes 33 (2012) 7e148

a specification focus on strength performance attributes, which inmost cases are of secondary importance to the necessary hydraulicperformance of the product.

GCL field hydraulic performance can be affected by a variety offactors, such as compatibility with overlying soils (cationexchange), presence of roots from vegetation, humidity of theenvironment (GCLmoisture content) and age of the installation (e.gDobras and Elzea, 1993; James et al., 1997; Meer and Benson, 2007;Benson et al., 2007). Over the years, various investigations havebeen conducted into field performance of GCL’s. Dobras and Elzea(1993) found calcium for sodium exchange had occurred in thesodium-bentontie component of a GCL due to leachates from theoverlying fill material. James et al. (1997) examined the perfor-mance of GCL’s installed as roof sealing elements to potable waterreservoirs, where micro-cracking of the bentonite was observed.According to the researchers, these observations indicatedevidence of calcium for sodium ion exchange, probably caused bya combination of calcite contained in the bentonite and calcareoussoils higher in the cover soil profile. A more recent study conductedby Benson et al. (2007) observed higher than anticipated percola-tion rates 4e15 months post installation of GCL caps over a coal ashlandfill that had been covered with 760 mm silty sand vegetationmedium. The increased permeability (tow5�10�7 m/s) was foundto be due to elevated levels of calcium and/or magnesium in thepercolating fluids, which induced exchange for sodium on thebentonite during desiccation events. This desiccation inducedcation exchange effect was also demonstrated on a series of GCLsexhumed from landfills in Florida andWisconsin by Benson and co-authors (Meer and Benson, 2007; Benson et al., 2007, 2010; Scaliaand Benson,2011). The exchange was quantified (Meer andBenson, 2007; Benson et al., 2007) using the ratio of monovalentto divalent (RMD) cations in the leachate, where lower values (i.e.below 0.14M1/2 indicate a predilection for such exchange reactions.This index is now an accepted parameter indicative for the poten-tial of leachates in equilibrium with cover soils to initiate poten-tially catastrophic cation exchange reactions. In the worse cases,hydraulic conductivity increased by as much as 4 orders ofmagnitude (Benson et al., 2010), but in some instances hydraulicconductivity was not significantly greater than pre-installed GCLsdespite nearly complete exchange of Na by divalent cations andmeasured swell index values at less than half the original values(Scalia and Benson, 2011).

There exists some evidence that the exchange-induced barrierfailure can be reversed. In one of the first studies investigatingcauses of failure to GCLs, Dobras and Elzea (1993)s found thatpermeability of a GCL was returned to its initial status by re-dopingthe cover soil with soda-ash. In capping applications, this mayprove to be a useful method for restoring barrier performance, atleast in the short term.

Historically in Australia, there has existed a knowledge gapregarding the in-field behaviour of GCL’s, mainly due to difficulties

Table 1Details of GCL exhumation sites.

Site GCLgrade

Yearconstructed

Type of installat

WA-1 Type B 02 Landfill CapQLD-1 Type C Sep 01 Landfill CapWA-2a Type B May 06 Mineral Process

Recovery DamWA-2b Type B May 04 Mineral Process

Settling DamSA-1 Type C Jul 00 Cellulose WasteNSW-1 Type A Late 00/Early 01 Landfill Cap

a Mean annual rainfall is based on the average of all historical rainfall data at the stat

associated with accessing sites for exhumation and, until recently,an almost universal acceptance by industry that the key GCLperformance parameters are geotextile related. This paper outlinesan exhumation program of GCL’s from landfill capping and miningtailing pond applications within Australia and presents preliminaryresults from selected sites in various states around the country.Given that Australia has a diverse range of climatic conditions e

from continental desert to Mediterranean to sub tropical e it isexpected that some of the results will be applicable to similarclimates elsewhere. The aim is to relate manufacturing specifica-tion parameters to in-field attributes of GCLs, which will providea basis to improve both GCL performance and testing.

These results may be used by designers to gauge GCL perfor-mance for similar applications. Recommendations as to bentonitespecifications required within a GCL can be made based on thefindings to provide designers with confidence in the performanceof the liner system.

2. Background

2.1. Site selection

Site selection for each exhumation was influenced by factorssuch as age of installation, average annual rainfall and accessibilityof GCL, and at least one site was selected in each state withinAustralia. The details of each site are shown in Table 1.

2.2. Description of GCLs

The GCL grades throughout the exhumation program aregenerally designated as needle-punched, thermally locked, wovenscrim reinforced GCLs containing sodium bentonite powder. Thespecific GCL descriptions are as follows:

- Type A GCL - Carrier ¼ Woven Geotextile, Cover ¼ Non-wovenGeotextile, Dry Bentonite Mass ¼ approx 3.6 kg/m2

- Type B GCL - Carrier ¼ Woven Geotextile, Cover ¼ Non-wovenGeotextile, Dry Bentonite Mass ¼ approx 4.0 kg/m2

- Type C GCL - Carrier ¼ Woven/Non-woven Composite,Cover ¼ Non-woven Geotextile, Dry Bentonite Mass ¼ approx3.7 kg/m2

The sodium bentonite powder used at all sites was designated asa civil engineering grade, Wyoming-type bentonite with indexproperties shown in Table 2.

2.3. Site final use

A summary of the subgrade and cover soil profiles, GCL types,installation and exhumation dates, and rainfall data for each site ispresented in Fig. 1.

ion Mean annualrainfall (mm)a

Exhumationdate

Approx.age (years)

785 Feb 08 64235 Aug 09 8

ing 447 May 09 3

ing 447 Sep 09 5

Cap 705 Feb 10 91502 Feb 10 9

ion nearest to the site.

Table 2Sodium bentonite specified values.

Property Value Units

Typical physical propertiesMontmorillonite content 70 %Dry Screen (passing 75 mm) 80 %Wet Screen (retained 75 mm) 2 %Bulk Density 0.9 t/m3

Liquid Limit 500 %Moisture Contenta 10 %Cation Exchange Capacity 85 meq/100 gpH 9 e

Typical chemical analysisSiO2 63.8 %Al2O3 13.6 %Fe2O3 2.8 %Na2O 2.3 %MgO 2.0 %CaO 0.2 %K2O 0.2 %TiO2 0.3 %LOI (Loss On Ignition)a 14.8 %

a Moisture content is the water content after oven drying. LOI is the water contentafter melting in flux (w1100 �C).

Table 3Chemical composition of tailings leachate at sites WA-2a & WA-2b.

Property Value Units

TSSa 38.5 mg/LTDSb 22,863 mg/LECc 32.4 mS/cmCa 798 mg/LMg 69.4 mg/LMn 64.2 mg/LNa 1397.8 mg/LFe 0.13 mg/LSO4 416 mg/LpH 6.6 e

NH4 72.2 mg/LCl 10,595 mg/L

a TSS ¼ Total suspended solids.b TDS ¼ Total dissolved solids.c EC ¼ Electrical conductivity.

J. Buckley et al. / Geotextiles and Geomembranes 33 (2012) 7e14 9

2.3.1. WA-1Site WA-1 consisted of a single GCL used as a primary cap for

a municipal solid waste installation in Western Australia. The GCLwas overlaid by approximately 0.5 m of sandy cover soil. A similarmaterial was placed as a blinding layer underneath the GCL. Thelandfill cap drained into a series of valley drains. The primaryfunction of the GCL is to mitigate rainfall infiltration into theunderlying landfill cell, thereby reducing generation of leachateand the costs associated with its management and disposal.

2.3.2. QLD-1The GCL at QLD-1 was also installed as a primary cap for

a regional solid waste installation in North Queensland. The coversoil overlying the GCL consisted of a 0.4m thickness of a sandy loammixed with sugar cane ash. A decomposed granite (mixture ofsand/fine gravel) bedding layer was originally placed under theGCL. The function of the GCL is also to reduce infiltration into thelandfill cell, in a very high rainfall area.

2.3.3. Sites WA-2a & 2bThe application for the GCL at sites WA-2a and WA-2b is

a secondary containment dam liner for mineral processing tailings,installed under a HDPE geomembrane. The subgrade soil consists ofpoorly graded, clayey sand. Confinement of the lining system isfrom a variable thickness of the actual tailings, with the chemicalcomposition of the liquor as shown in Table 3. This liquor with theabove TDS (total dissolved salt), TSS (total soluble salts) and EC(electrical conductivity) levels is best described as saline brine, with

Fig. 1. Summary of exhumation sites.

an ionic strength of w0.2 M, calculated based on the analysesreported.

Chemical compatibility testing with the GCL bentonite and thisliquor was conducted some time prior to installation of the GCL atSite WA-2b. This testing produced favourable results, whichconstituted a pre-approval by the project consultants, for supply ofthe GCL with this bentonite for future installations (GeofabricsAustralasia personal communication, 2009).

2.3.4. SA-1The function of the GCL at site SA-1 is to form a primary cap over

wood processing waste, thereby limiting rainfall infiltration intothe waste body and minimising methane gas generation. The sitelocation in South Australia typically experiences higher rainfallthan the rest of the state. The GCL was immediately overlaid by0.20 m of drainage sand, then 0.40 m of clay and 0.30 m of topsoil,totalling 0.90m of cover material. The subgradematerial consists ofpredominately fine-medium sand.

2.3.5. NSW-1At NSW-1, the GCL is used as a primary cap over a large

municipal waste installation which has been developed intorecreational playing fields. NSW-1 is situated in an environmentallysensitive area being in close proximity to a major watercourse. Thecover material consists of 0.20 m of weathered coal-stone/sandygravel directly above the GCL, overlaid by 0.30 m of topsoil. Thesubgrade is a gravelly sand material, similar in nature to the over-lying coal-stone.

2.4. Durability stresses

2.4.1. RainfallAustralia’s climate is dominated by the dry, sinking air of the

sub tropical high pressure belt which moves north and south withthe seasons. This causes the rainfall pattern over Australia to bestrongly seasonal and helps to define the main climate regionsshown in Fig. 2 (Bureau of Meteorology, 2009).

The GCL exhumation sites are situated in distinctly differentseasonal rainfall zones, as follows:

Sites WA-1, WA-2a & 2b

These sites are situated in sub-tropical Western Australia, wherethe rainfall zone is characterised by Winter Dominant Rainfall,having a marked wet winter and dry summer.

Site QLD-1

Fig. 2. Australian major seasonal rainfall zones Courtesy Australian Bureau of Meteorology website, 2009.

J. Buckley et al. / Geotextiles and Geomembranes 33 (2012) 7e1410

This site is situated in tropical North Queensland, where therainfall zone is characterised by Summer Dominant Rainfall, havinga marked wet summer and dry winter.

Site SA-1

This site is situated in south-east South Australia, where therainfall zone is characterised by Winter Rainfall, having a wetwinter and low summer rainfall.

Site NSW-1

This site is situated in the central coast of New South Wales,where the rainfall zone is characterised by Uniform Rainfall, havingno defined wet or dry season and a median annual rainfall�0.350 m.

A graph of rainfall for each site, from 1998 to 2009, is shown inFig. 3 and the mean annual historical rainfall is shown in Table 1.

2.4.2. Age of sitePrevious literature (Meer and Benson, 2007) has stated that Ca

for Na exchange within cover GCLs can occur within 5 years intemperate North America. Egloffstein (2002) suggested that cationexchange can occur in GCLs over a period of months to a few years.Older studies (Dobras and Elzea, 1993) showed that Ca for Naexchange in base liners is probably common and occurs withina few years. The ages of the exhumation sites varied between 3 and

Fig. 3. Annual rainfall at each exhumation site from 1998 to 2009.

8 years and all exhumation sites were selected on the basis that anypotential cation exchange could occur in a minimum of 3 years andideally at least 5 years, in Australia.

2.5. Exhumation methodology

The sampling of exhumed GCL was generally carried out inaccordance with ASTM D6072 e “Standard Guide for ObtainingSamples of GCL’s” and in general agreement with previous researchperformed by Zanzinger and Touze-Foltz (2009), in terms of GCLdescription, site description, soil description, sampling and testingprocedures. However, specific requirements at each stage of theexhumation are outlined in the following sub-sections:

2.5.1. Pre-exhumationThe size of the exhumed GCL sample was typically 1.0 m by

1.0 m. The overlap for the repair piece is 0.5 mminimum; therefore,the dimensions of the replacement GCL were required to bea minimum of 2.0 m by 2.0 m.

2.5.2. ExhumationThe chosen area for exhumation was a minimum of 2.6 m by

2.6 m, depending on the depth of cover soil at that point. Carefulsiting of the GCL exhumation area was necessary to avoid GCLsamples takenwithin 1m of seam/overlap areas, or the end of a roll.Exhumation of the outer GCL edge, consultation with originalinstallation staff and inspection of installation photos were carriedout in order to ensure that the exhumation area was as central tothe roll width as possible.

Depending on the cover soil, exhumation was commenced byhand or carefully with machinery. If machinery was used, the tophalf of cover only was excavated by machine; the lower half wasthen carefully excavated by hand. In order to minimise contami-nation of the GCL, the GCL was vigorously brushed to clean off coversoil, prior to removal of the GCL.

2.5.3. Sample cuttingA 1.0 m by 1.0 m plywood template was placed directly over the

GCL and the GCL sample was cut around the template with a sharputility knife. The underside of the GCL sample was then thoroughlycleaned. The GCL sample was wrapped with polythene wrap andtaped to prevent any moisture variation. If more than 1 GCL samplewas exhumed, each GCL sample was individually wrapped and

Table 4Chemical properties of the subgrade and cover soil materials exhumed.

Site Cation concentrationof percolating water(meq/100 g)

ECa

(mS/cm)RMDb

(M1/2)Id (M)

Ca Mg Na

WA-1 Subgrade 1.9 0.2 0.2 168 0.14 0.0021Cover 7.7 0.5 0.2 169 0.07 0.0021

QLD-1 Subgrade 1.0 0.1 <0.1 67 <0.09 0.0009Cover 3.1 0.4 <0.1 57 <0.05 0.0007

WA-2a Subgrade 3.7 1.4 2.2 1750 0.97 0.0222WA-2b Subgrade 4.2 3.0 1.0 556 0.37 0.0071SA-1 Subgrade 20.9 1.2 1.4 184 0.30 0.0023

Cover 2.9 0.9 1.4 45 0.72 0.0006NSW-1 Subgrade 11.5 4.5 9.7 N.D.c 2.43 0.0225

Cover 7.9 3.0 12.3 N.D. 3.73 0.0202

a E.C. ¼ Electrical Conductivity (saturated paste).b RMD ¼ Ratio of sum of monovalent to square root of sum of divalent

cations ¼ Mm/Md1/2 (Meer and Benson, 2007).

c N.D. ¼ Not Determined.d I ¼ Ionic strength calculated by E.C. � 0.0127, accept for NSW-1 samples where

it was estimated by assuming all counter anions were Cl, using conversion factors of200 for Ca, 121 for Mg and 230 for Na.

J. Buckley et al. / Geotextiles and Geomembranes 33 (2012) 7e14 11

taped. Next, the wrapped GCL sample was placed between a pair ofplywood templates, taped securely & then wrapped again with PEfilm. Taping of the PE wrapping was sufficient to prevent ingress ofmoisture during transport and the sample was shipped within 24 hafter packaging to minimize moisture content changes inside thepackage.

2.5.4. Soil samplingSamples of the subgrade and cover soils were placed in tightly

sealed, appropriately labelled containers, and then transported onice to avoid significant changes in chemical composition of the soilmatrix due to microbial activity to arrive within the lab analysisholding time, which could be as short as 7 days for certain tests.

2.5.5. Restoration of GCL & cover soilThe restoration of the exhumed GCL sample was generally

carried out in accordance with established practice for GCL repairand patching. After thorough cleaning of the underlying overlaparea, liberal quantities of bentonite paste were applied under thereplacement GCL. The cover soil and vegetation was then carefullyreinstated to a standard acceptable to the site owner.

3. Testing

3.1. Test methods

Independent laboratories provided detailed mineralogical andchemical analyses of both the bentonite within the GCL andsurrounding soils and these analyses were carried out in order toidentify any changes that may have occurred and to better evaluatethe impact that these changes may have on the long-term perfor-mance of the GCL. These analyses included: total metals by ICP-AES(inductively coupled argon plasma atomic absorption spectrom-etry) on acid-digested extracts; soluble cations and anions by ICP-AES on 1:5 soil:water extracts; exchangeable cations by ICP-AESusing ammonium displacement method (Rayment and Higginson,1992); pH, total soluble salts and total alkalinity on 1:5 soil-waterextracts; total Cl using colorimetric methods on acid-digestedextract; total dissolved salts and EC on a saturated paste extract;particle size analysis by sieving and hydrometer methods. Allmethods followed standard internationally recognized methods(eg Australian Standard (AS), American Society for Testing andMaterials (ASTM), American Public Health Association (APHA),United States Environmental Protection Agency (USEPA) ormethods developed in-house by the consulting laboratory in theabsence of documented standards.

Hydraulic and physical testing was carried out by in-houselaboratories on the exhumed GCL ie. Hydraulic conductivity, swellindex and moisture content, as described below.

3.2. Exhumed GCL specimen preparation

Each exhumed GCL bulk sample was cleaned of adjacent soils.From each bulk sample, 10� (0.09 m� 0.15 m) specimens were cutfor mass per unit area (MPUA), thickness and moisture contentanalysis and 10 � 0.1 m diameter specimens were cut for hydraulicconductivity analysis. Bentonite was extracted from remainingspecimens for swell index, specific gravity and the qualitative/quantitative analyses.

3.3. Hydraulic conductivity analysis

Generally, three specimens from each site were tested inaccordance with ASTM D 5887 (2009), Method C (Falling Head,Rising Tailwater) and hydraulic conductivity results determined

using Appendix X2 of this standard. A pressure control panel, withtriaxial cells and bladder accumulators was used. Specimens forhydraulic conductivity measurements were carefully selectedbased on mass and uniformity of bentonite distribution, ensuringspecimens were as close to representative as possible. Wherenecessary, to minimise sidewall leakage during analysis, the topand base geotextile layers around the circumference of each spec-imenwere separated, wetted with de-ionised water and then edge-rolled with dry bentonite powder of similar quality to the bentonitewithin the GCL.

3.4. Swell index and specific gravity analysis

A single test was performed on each sample for swell index and,where possible, specific gravity, using bentonite extracted fromcorresponding specimens. Swell index was performed in accor-dance with ASTM D 5890-06 and specific gravity was performed inaccordance with ASTM D854-06. The exhumed GCL sample washydrated with de-ionised water to improve the efficiency ofremoving the bentonite, which was dried, ground and sieved, priorto swell index testing.

4. Results and discussion

The chemical composition analyses showing the major cationconcentrations of the percolating water from the cover andsubgrade soils for each site are shown in Table 4, along with thecalculated RMD values (Meer and Benson, 2007), electricalconductivity (E.C.) and ionic strengths. Changes to the major basecation composition (as oxides from ignited samples) of thebentonite samples were used to infer changes to the total calcium,magnesium and sodium contents from the exhumed GCL for eachsite (Table 5). The specification values from the bentonite supplier,dated July 2001, as well as results of independent reference testsconducted on the bentonite rawmaterial in July 2006 and February2007, are also presented in Table 5. These specification and refer-ence values are representative of the bentonite used in the GCLoriginally supplied to all the exhumation sites. Table 5 also showsthe percentage of exchangeable sodium retained within thebentonite following exhumation. The results for exhumedGCL properties such as bentonite moisture content, hydraulic

Table 5Chemical composition and exchange properties of exhumed GCL bentonite.

Site Properties of exhumed bentonite

Calciumoxide e CaO (%)

Magnesiumoxide e MgO (%)

Sodiumoxide e Na2O (%)

Exchangeablesodium e Na (%)

Cation exchangecapacity (meq/100 g)

Ref. samp. e Jul 06 0.46 2.15 2.37 84.2 62Ref. samp. e Feb 07 0.56 2.17 2.60 N.D.a N.D.Spec e Jul 01 0.2 2.0 2.3 N.D. 85WA-1 1.61 2.09 0.66 N.D. N.D.QLD-1 1.71 1.98 0.70 0.6 64WA-2a 0.66 2.20 1.69 43.3 65WA-2b 1.33 2.32 1.05 20.8 69SA-1 1.73 2.19 0.37 1.5 64NSW-1 1.21 2.14 1.64 38.3 80

a N.D. ¼ Not determined.

J. Buckley et al. / Geotextiles and Geomembranes 33 (2012) 7e1412

conductivity and swell index compared to cover soil and bentoniteproperties are summarised in Table 6.

4.1. WA-1

As shown in Table 4, the exchangeable divalent cation (ie.calcium and magnesium) levels in the cover soil at WA-1 werehigher than the corresponding properties in the subgrade soil. Theexchangeable monovalent (ie. sodium) levels for both soils wererelatively low compared to the exchangeable calcium and magne-sium levels. Therefore, there is a greater risk of calcium for sodiumcation exchange occurring in the bentonite within the GCL due totransport of soluble divalent cations. This effect is also reflected inthe RMD (ratio of monovalent to the square root of the sum ofdivalent cations) values (Table 4) of the leachate, which indicate thecapacity of percolating waters in equilibriumwith the soil materialto initiate cation exchange within the bentonite in the GCL. RMDvalues �0.07 M1/2 are considered to present a high risk of loss ofsodium saturation in sodium bentonites, whereas less significantchanges are observed for RMD values�0.14M1/2 (Benson andMeer,2009). Therefore, RMD values <0.14, can have a significant impacton GCL hydraulic conductivity and swelling performance, especiallywhere there is potential for wet-dry cycling. The RMD value for thesoluble cations from the cover soil at WA-1 was at the limit of0.07 M1/2, indicating a high potential of calcium for sodiumexchange in the GCL.

The exhumed bentonite at WA-1 had significantly lower totalsodium content than the combined total calcium and magnesiumlevels, as shown in Table 5. Also, the WA-1 total sodium is signifi-cantly lower than the bentonite specification and reference sampletotal sodium values. This, combined with the slightly elevated ECvalues (Table 4), and the increase in total calcium and magnesium

Table 6Exhumed GCL properties compared with cover soil and bentonite characteristics.

Site Site Properties of exhumed GCL

GCL moisturecontent (%)

Swell index(mL/2 g)

Hydraulicconductiv

GCL specvalues

Specvalues

15 27 3 � 10�

WA-1 WA-1 35.0 7 5.6 � 10�

QLD-1 QLD-1 54.5 5 5.3 � 10�

WA-2a WA-2ab 20.5 21 2.1 � 10�

WA-2b WA-2bb 56.1 13 2.9 � 10�

SA-1 SA-1 113.6 6 3.0 � 10�

NSW-1 NSW-1 86.5 18 2.2 � 10�

a N.A. ¼ Not applicable.b WA-2a and WA-2b RMD results are for subgrade soil, GCL covered with geomembrac N.D. ¼ Not determined.

compared to the specification and reference values (Table 5) wouldindicate that sodium is being replaced by calcium and magnesium,within the bentonite exchange complex. Also, this could be influ-enced by high calcium loading in the cover soils.

The WA-1 exhumed GCL results in Table 6 show that, eventhough calcium for sodium exchange has occurred and the swellindex is below specification, the hydraulic conductivity is within anorder of magnitude of the GCL specification value. Such lowerpermeability results have been observed recently for similarly agedexhumed GCLs from the USAwhich had been mostly exchanged byCa and/or Mg during their service life (Benson et al., 2010; Scaliaand Benson, 2011). Even though the rainfall for the year previousto the exhumation (2007 ¼ 735 mm) was close to average, the GCLmoisture content was relatively low at WA-1 and was below thevalue of 85% mentioned by Meer and Benson (2007) as beingindicative of sustained desiccation. To reiterate their findings, GCL’swith field moisture contents below 85% were more susceptible tothe combined detrimental effects of cation exchange followed bydesiccation, which generally resulted in higher GCL hydraulicconductivity. GCL’s with field moisture contents above 100%generally maintained lower hydraulic conductivities, and in someinstances, within specification limits. Gates (2008) estimated thehydraulic conductivity of a GCL with a water content of 50%, incontact with percolating water fromWA-1 cover soil material to beup to 4 � 10�11 m/s, based on an estimated ionic strength of0.002 M, GCL thickness of 5 mm and a steady-state 5 mm head ofpercolating water. This result closely approximates the measuredWA-1 hydraulic conductivity, albeit at a lower GCL moisturecontent. No plant roots were visible in the exhumed GCL samplefrom WA-1. At site WA-1, the combined effect of the calcium forsodium exchange in the bentonite and moderate potential for wet-dry cycling, due to the shallow, sandy cover soil and the low GCL

ity (m/s)Cover soil leachateRMD (M1/2)

Exchangeablesodium e Na (%)

Rootpenetration

11 N/Aa N.Dc N/A

11 0.07 N.D. N8 <0.05 0.6 Y11 N/Aa 43.3 N10 N/Aa 20.8 N10 0.72 1.5 N11 3.73 38.3 N

ne.

J. Buckley et al. / Geotextiles and Geomembranes 33 (2012) 7e14 13

moisture content, does not appear to have significantly affected theGCL hydraulic performance, even after approximately 6 years inservice at time of exhumation.

4.2. QLD-1

The RMD determined on the cover soil leachate at QLD-1 is thelowest of all the exhumation sites, as shown in Table 4. This iscaused by the immeasurably low level of soluble sodiumcombined with the much higher soluble calcium content in thecover (and subgrade) soils. The exchangeable sodium content forthe exhumed bentonite, shown in Table 5, is the lowest of all sitesand these results indicate that bentonite in the GCL’s from theQLD-1 site would have the most severe calcium for sodiumexchange. The GCL moisture content at QLD-1, shown in Table 6,appeared to be low in light of the high average rainfall experi-enced at the site. However, the month of the exhumation (August)is typically one of the drier months in this region. The cover soil atQLD-1 consists of a 50/50 loam/sugar cane ash mixture and isrelatively free draining. The fact that the site is contoured and welldrained could also have limited the infiltration of rainwater intothe GCL.

Compared to all exhumation sites, as shown in Table 6, the swellindex result for QLD-1 is the lowest and the hydraulic conductivityresult is the highest. The hydraulic conductivity value is just over 3orders of magnitude higher than the specified hydraulic conduc-tivity. The low RMD value of the leachate and low exchangeablesodium content of the exhumed bentonite probably accounts forthe significant increase in GCL hydraulic conductivity as well as thevery low swell index, at least in part. However, other issues may becontributing to the poor hydraulic performance. For example, thebentonite in the QLD-1 sample had a distinctive appearance, unlikeany of the other exhumed bentonite samples, consisting of a gran-ular texturewith what appeared to be an organic “waxy” coating onthe individual granules, possibly caused by the high organiccontent of the sugar cane ash used in the cover soil. Therefore,further investigation is warranted at this site, involving a compati-bility test with pristine GCL in contact with percolating watercollected from an elution test with the site cover soil and rainwater.The exhumed GCL samples from QLD-1 were also observed to havesignificant amounts of root penetration through the entire thick-ness of GCL, which could have contributed significantly to theincreased hydraulic conductivity. The cover soil thickness wasapproximately 400 mm and the landfill cap was planted witha variety of running Couch grass. As previously mentioned, the siteis well drained, but is located in the subtropics with an averageannual rainfall over 4 m, a significantly higher rainfall thanoccurred in any of the previously published studies (e.g. James et al.,1997; Benson et al., 2007, 2010; Meer and Benson, 2007; Scalia andBenson, 2011).

Table 7Summary of GCL exhumation results.

Site Threshold properties of exhumed GCL

GCL field moisturecontent (%)

Swell index(mL/2 g)

Hydraulicconductivity

WA-1 Low (<70%) Low (<10) Low (<1 � 10�10

QLD-1 Low Low High (>1 � 10�9)WA-2a Low High (>24) LowWA-2b Low Medium (10-24) Medium

(1 � 10�10e1 � 1SA-1 High (>100) Low MedNSW-1 Medium (70e100) Med Low

a WA-2a and WA-2b RMD results are for subgrade soil, GCL covered with geomembra

4.3. WA-2a

The barrier system at site WA-2a comprised a GCL directlyoverlaid by a geomembrane (to contain mineral processing tail-ings); therefore the effect of cover soil was not applicable at thissite. The RMD value determined from the soluble cations of thesubgrade soil leachate (Table 4) is above Benson and Meer’s (2009)recommendation of 0.14 M1/2 and the results from Table 5 indicatethat there is a high percentage of exchangeable sodium remainingin the exhumed bentonite. Therefore, calcium for sodium exchangeappears not to have occurred within the bentonite at this GCLinstallation. Given the elevated EC (Table 4) this is not unexpected,since Ca for Na exchange on smectite surfaces is repressed atelevated ionic strength (Laudelout, 1987). The exhumed GCL at thissite had the lowest moisture content of all sites, but the swell indexwas the closest to specification of all sites. The hydraulic conduc-tivity of theWA-2a exhumed GCL remained below the specificationvalue and had the best hydraulic performance of the program.However, the age of the installation is only 3 years and since it wasoverlain directly by a geomembrane it may warrant further inves-tigation after more years of service.

4.4. WA-2b

As for WA-2a, the GCL at site WA-2b was directly overlaid bya geomembrane. The subgrade soil leachate RMD value (Table 4) isalso above Benson andMeer’s (2009) limit of 0.14 M1/2 for potentialfor cation exchange. Table 5 indicates that there still is a moderatepercentage of exchangeable sodium remaining in the exhumedbentonite. The total calcium (Table 5) level in the exhumedbentonite is slightly higher than the total sodium level, indicatingsome calcium for sodium exchange did occur, in line with its lowerEC (Laudelout, 1987) compared to WA-2a (Table 4). As shown inTable 6, the GCL moisture content was higher than previous sites,but still lower than Meer and Benson’s (2007) limit of 85% (similarto the WA-1 site). Also, the swell index was low (Table 6) and thehydraulic conductivity of the exhumed GCL was one order ofmagnitude higher than specification values. In light of the aboveresults, it appears that calcium for sodium exchange has initiatedafter 5 years of service, but has thus far only moderately affectedthe GCL hydraulic performance.

4.5. SA-1

The RMD value for SA-1 cover soil leachate is well above0.14 M1/2 (Table 4) indicating a limited potential of calcium forsodium exchange. However, Table 5 shows that the exchangeablesodium content of the bentonite is quite low and the bentonitecalcium and magnesium oxide contents are significantly higher. Inaddition, the sodium oxide content was lower, than those of the

Cover soil leachateRMD (M1/2)

Exchangeablesodium e Na (%)

Rootpenetration

) Low (<0.07) Severe NoneLow Severe SevereHigha (>0.14) None None

0�9)Mediuma (0.07e0.14) Mild None

High Mild NoneHigh None None

ne.

J. Buckley et al. / Geotextiles and Geomembranes 33 (2012) 7e1414

reference samples, indicating a certain level of calcium for sodiumexchange. The cover soil system at SA-1 is comprised of multiplesoil layers and, over the 91/2 year service period, there could havebeen a contribution from the upper layers to the measured calciumand magnesium load in the GCL bentonite. The GCL moisturecontent for SA-1, shown in Table 6, is the highest of all the exhu-mation sites and is above Meer and Benson’s (2007) value of 100%for optimal GCL hydraulic performance, which is surprising sincethe site has the second lowest average rainfall. However, thecombination of total cover thickness, humidity of the wood wastepile and multi-layer nature of the cover soil system could haveassisted in maintaining a high GCL moisture content. The swellindex is relatively low and the GCL hydraulic conductivity is oneorder of magnitude higher than the specification. Therefore, theabove results show that the calcium for sodium exchange occurringin the bentonite has had a moderate effect on the hydraulicperformance of the GCL.

4.6. NSW-1

As shown in Table 4, NSW-1 has the highest RMD value for thecover soil leachate of all the exhumation sites. The exchangeablesodium content in the cover soil and the exchangeable sodiumcontent of the exhumed bentonite is correspondingly high, asshown in Table 5. From these results, it would appear that calciumfor sodium exchange has not occurred in the exhumed GCLbentonite and the potential for it to occur is low. The moisturecontent of the GCL, as shown in Table 6, is above Meer andBenson’s (2007) limit of 85%, but lower than the value foroptimal performance of 100%. The average rainfall at this site ishigher than most of the other sites. The swell index result isapproximately 2/3 of the specification value, but the GCL hydraulicconductivity is lower than the specification. After 9 years ofservice, the hydraulic performance of the GCL at NSW-1 is effectiveand calcium for sodium exchange within the bentonite has notoccurred.

5. Conclusions

A summary of the results of the exhumations and the potentialfor calcium for sodium exchange within the bentonite is shown inTable 7. In partial accordance with previous recent studies (e.g.James et al., 1997; Benson et al., 2007, 2010; Meer and Benson,2007; Scalia and Benson, 2011), it appears that superior fieldhydraulic performance, ie. low hydraulic conductivity, in GCLs arerelated to high cover soil leachate RMD which provides a low riskof bentonite calcium for sodium exchange. For example, thehydraulic performance of the GCL at site NSW-1 remained withinspecifications even after >9 years in a high rainfall (>1.5 mannually) environment where the RMD values for cover andsubgrade leachates were high. However, as observed here for siteWA-1, while low RMD values of subgrade and cover soil leachates,coupled with low moisture content resulted in calcium forsodium exchange, this did not appear to influence the hydraulicperformance after 6 years. This result is in apparent contradictionto the expected behaviour anticipated by Meer and Benson(2007).

Acknowledgements

The authors thank all the councils, stakeholders and GeofabricsAustralasia branch staff that have participated and assisted us withthe exhumation program and W. Hornsey of Geofabrics Australasiafor his support of this project. We wish to thank the three anony-mous reviewers who made insightful and constructive improve-ments to this manuscript.

References

ASTM D 5887, 2009. Standard Test Method for Measurement of Index Flux throughSaturated Geosynthetic Clay Liner Specimens Using a Flexible Wall Per-meameter. American Society for Testing and Materials, West Conshohocken,Pennsylvania, USA.

ASTM D 5890, 2006. Standard Test Method for Swell Index of Clay MineralComponent of Geosynthetic Clay Liners. American Society for Testing andMaterials, West Conshohocken, Pennsylvania, USA.

ASTM D 6072, 2009. Standard Guide for Obtaining Samples of GCL’s. AmericanSociety for Testing and Materials, West Conshohocken, Pennsylvania, USA.

Australian Bureau of Meteorology website, 2009. Commonwealth of Australia.Benson, C.H., Meer, S.R., 2009. Relative abundance of monovalent and divalent

cations and the impact of desiccation on geosynthetic clay liners. Journal ofGeotechnical and Geoenvironmental Engineering, ASCE 135 (3), 349e358.

Benson, C.H., Ören, A.H., Gates, W.P., 2010. Hydraulic conductivity of two geo-synthetic clay liners permeated with a hyperalkaline solution. Geotextiles andGeomembranes 28, 206e218.

Benson, C.H., Thorstad, P.A., Jo, H.-Y., Rock, S.A., 2007. Hydraulic performance ofgeosynthetic clay liners in a landfill final cover. Journal of Geotectnical andGeoenvironmental Engineering 133 (7), 814e827.

Bouazza, A., Gates, W.P., Abuel-Naga, H., 2006. Factors Impacting Liquid and GasFlow through Geosynthetic Clay Liners. Geosynthetics e Recent Developments.Commemorating Two Decades of Geosynthetics in India (Chapter 9). IndianInternational Geosynthetics Society, New Delhi. 119e146.

Dobras, T.N., Elzea, J.M., 1993. In-situ soda ash treatment for contaminated geo-synthetic clay liners. In: Proceedings of Geosynthetics 1993, vol. 3. IndustrialFabrics Association, St Paul, pp. 1145e1160.

Egloffstein, T.A., 2002. Bentonite as sealing material in geosynthetic clay liners einfluence of the electrolytic concentration, the ion exchange and ion exchangewith simultaneous partial desiccation on permeability. In: Proceedings of theInternational Symposium on Clay Geosynthetic Barriers, Nuremberg,pp. 141e153.

Francisca, F.M., Glastein, D.A., 2010. Long term hydraulic conductivity of compactedsoils permeated with landfill leachate. Applied Clay Science 49, 187e193.

Gates, W.P., 2008. Independent Report e WA-1 Cover and Sub-grade Soil Analysis.SmecTech Research Consulting.

Gates, W.P., Bouazza, A., Churchman, G.J., 2009. Bentonite clay keeps pollutants atbay. Elements 5, 105e110.

Hornsey, W.P., Scheirs, J., Gates, W.P., Bouazza, A., 2010. The impact of miningsolutions/liquors on geosynthetics. Geotextiles and Geomembranes 28,191e198.

James, A.N., Fullerton, D., Drake, R., 1997. Field performance of GCL under ionexchange conditions. Journal of Geotechnical and Geoenvironmental Engi-neering 123 (10), 897e901.

Koerner, R.M., 2005. Designing with Geosynthetics, fifth ed. Pearson Prentice Hall,New Jersey, USA. 796 p.

Laudelout, H., 1987. Cation exchange equilibria in clays. In: Newman, A.C.D. (Ed.),Chemistry of Clays Clay Minerals. Mineralogical Society, London,pp. 225e236.

Meer, S.R., Benson, C.H., 2007. Hydraulic conductivity of geosynthetic clay linersexhumed from landfill final covers. Journal of Geotechnical and Geo-environmental Engineering, ASCE 133 (5), 550e563.

Rayment, G.E., Higginson, F.R., 1992. Australian Laboratory Handbook of Soil andWater Chemical Methods. Inkata press, Melbourne, Victoria, Australia. 330 p.

Rowe, R.K., 2001. Geotechnical and Geoenvironmental Engineering Handbook.Kluwer Academic Publishing, Norwell, USA. Liner systems. (Chapter 25) (pp.739e788).

Scalia, J., Benson, C., 2011. Hydraulic conductivity of geosynthetic clay linersexhumed from landfill final covers with composite barriers. Journal ofGeotechnical and Geoenvironmental Engineering, ASCE, 137 (1), 1e13.

Zanzinger, H., Touze-Foltz, N., 2009. Clay geosynthetic barriers performance inlandfill covers. In: Proceedings GeoAfrica, pp. 1e9.