See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/245294832

Relative Abundance of Monovalent and Divalent Cations and the Impact of

Desiccation on Geosynthetic Clay Liners

Article in Journal of Geotechnical and Geoenvironmental Engineering · March 2009

DOI: 10.1061/(ASCE)1090-0241(2009)135:3(349)

CITATIONS READS

77 595

2 authors, including:

Craig H Benson
University of Virginia
415 PUBLICATIONS 15,512 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Containment Systems and Landfill Barriers View project

Dry cover material over tailing facilities View project

All content following this page was uploaded by Craig H Benson on 06 May 2018.

The user has requested enhancement of the downloaded file.

 Relative Abundance of Monovalent and Divalent Cations
and the Impact of Desiccation on Geosynthetic Clay Liners
Craig H. Benson1 and Stephen R. Meer2

Abstract: Laboratory experiments were conducted on a geosynthetic clay liner 共GCL兲 containing Na–bentonite to determine how the
swell index and hydraulic conductivity of GCLs are affected by wet-dry cycling with solutions having different relative abundance of
monovalent and multivalent cations. Relative abundance of monovalent and multivalent cations was characterized by the RMD of the test
solution, which is defined as the ratio of the total molarity of monovalent cations to the square root of the total molarity of multivalent
cations at a given ionic strength. RMD was found to control the final swell index, relative abundance of monovalent and divalent cations
in the final exchange complex, and the final hydraulic conductivity of bentonite exposed to wet-dry cycling. Ionic strength affects the
number of wet-dry cycles required for a change in hydraulic conductivity to occur and the rate of change in swell index. Large increases
in hydraulic conductivity and loss of swelling capacity occurred for solutions having RMD艋 0.07 M1/2. Modest or small changes in
hydraulic conductivity and swell index were obtained when the RMD was 艌0.14 M1/2. These findings suggest that chemical analysis of
the pore water in cover soils may prove useful in evaluating the compatibility of GCLs and cover soils used in applications where wet-dry
cycling may occur.
DOI: 10.1061/共ASCE兲1090-0241共2009兲135:3共349兲
CE Database subject headings: Clay liners; Landfill; Geosynthetics; Hydraulic conductivity.

Introduction exchange of multivalent cations for the native Na results in in-

creased hydraulic conductivity and decreased swell potential
Geosynthetic clay liners 共GCLs兲 are factory-made clay liners that 共Shan and Daniel 1991; Gleason et al. 1997; James et al. 1997;
consist of a layer of bentonite 共3.2– 6.0 kg/ m2兲 bonded to a geo- Ruhl and Daniel 1997; Petrov and Rowe 1997; Shackelford et al.
synthetic material. Most GCLs contain bentonite sandwiched be- 2000; Egloffstein 2001; Jo et al. 2001, 2004, 2005; Kolstad et al.
tween two geotextiles that are bonded using needle punching, 2004, Guyonnet et al. 2005; Lee and Shackelford 2005; Lee et al.
stitching, or adhesives. In some cases, the bentonite is bonded 2005; Katsumi et al. 2007兲. When cation exchange is concomitant
directly to a geomembrane or a geomembrane is laminated to one with wet-dry cycling, increases in hydraulic conductivity can
of the geotextiles. For GCLs that do not include a geomembrane, occur that may render a GCL ineffective as a hydraulic barrier
the effectiveness as a hydraulic barrier is controlled by the hy- 共Melchior 1997; Lin and Benson 2000; Egloffstein 2001; Benson
draulic conductivity of the bentonite. The hydraulic conductivity et al. 2007; Meer and Benson 2007兲. These increases in hydraulic
of the sodium bentonite used in most GCLs is on the order of conductivity have been attributed to macroscopic features 共e.g.,
cracks, intergranule pores兲 formed during drying that do not swell
10−9 cm/ s when permeated with deionized 共DI兲 water at stresses
shut during rehydration 共Lin and Benson 2000; Melchior 2002;
typical of cover applications 共Shan and Daniel 1991; Petrov and
Benson et al. 2007; Meer and Benson 2007兲.
Rowe 1997; Shackelford et al. 2000; Jo et al. 2001兲.
Recent field studies of covers with GCLs have confirmed that
GCLs present an attractive alternative to compacted clay liners
multivalent-for-monovalent cation exchange can occur relatively
as the hydraulic barrier layer in landfill cover systems because of
rapidly and that large increases in hydraulic conductivity of GCLs
their low hydraulic conductivity 共⬇10−9 cm/ s to DI water兲, ease
can occur under some circumstances 共Egloffstein 2001; Melchior
of installation, thinness, and perceived resistance to environmen-
2002; Benson et al. 2004, 2007; Meer and Benson 2007兲. Water
tal stresses 共e.g., freeze-thaw and wet-dry cycling兲 共Bouazza
percolating downward from the overlying cover soils is believed
2002兲. However, the bentonite in GCLs is sensitive to chemical to be the source of the multivalent cations 共Melchior 2002; Ben-
interactions with the hydrating liquid, and ion exchange that oc- son et al. 2007; Meer and Benson 2007兲. For example, Meer and
curs in bentonite can alter its physical properties. In particular, Benson 共2007兲 measured the hydraulic conductivity, swell index,
and cation exchange complex of GCLs exhumed from four land-
1
Wisconsin Distinguished Professor and Chairman, Dept. of Geologi- fill covers that had been in service between 4.1 and 11 years.
cal Engineering, Univ. of Wisconsin-Madison, Madison, WI 53706 They found that divalent cations 共primarily Ca兲 had replaced most
共corresponding author兲. E-mail: [email protected] of the native Na in the exchange complex of the bentonite. Hy-
2
Engineer, Sigma Environmental Services, Inc., 1300 West Canal Str, draulic conductivities of the exhumed GCLs fell in a broad range
Milwaukee, WI 53233. E-mail: [email protected]
共5.2⫻ 10−9 – 1.6⫻ 10−4 cm/ s兲, but most were greater than
Note. Discussion open until August 1, 2009. Separate discussions
must be submitted for individual papers. The manuscript for this paper 10−6 cm/ s. The swell index of bentonite from the exhumed GCLs
was submitted for review and possible publication on October 29, 2007; was typical of Ca–bentonite, and analysis of the exchange com-
approved on May 13, 2008. This paper is part of the Journal of Geotech- plex showed that nearly all of the native Na had been replaced by
nical and Geoenvironmental Engineering, Vol. 135, No. 3, March 1, Ca, and to some extent Mg.
2009. ©ASCE, ISSN 1090-0241/2009/3-349–358/$25.00. Meer and Benson 共2007兲 also reviewed field data for 15 GCLs

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MARCH 2009 / 349

Downloaded 03 Mar 2009 to 69.91.147.111. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
where the in-service hydraulic conductivity could be determined. Table 1. Properties of Cover Soils for Elution Tests
Of these 15 GCLs, 11 had hydraulic conductivities ⬎10−6 cm/ s Percent CaCO3
and nine had hydraulic conductivities ⬎10−5 cm/ s. Only one Soil fines Liquid Plasticity Paste content
GCL had a hydraulic conductivity typical of a new GCL with ID type 共%兲 limit index pH 共%兲
sodium bentonite 共2 ⫻ 10−9 cm/ s兲. GCLs that had undergone in-
1 Sandy clay 74 26 11 7.4 2.6
creases in hydraulic conductivity also exhibited near complete
2 Silty sand 20 36 9 7.2 2.2
cation exchange 共Ca or Mg for Na兲. In contrast, the GCL that had
3 Silty sand 17 39 11 5.6 1.0
a hydraulic conductivity 2 ⫻ 10−9 cm/ s did not exhibit cation ex-
change, despite being buried for 5 years. The cover soil overlying 4 Clayey sand 24 24 12 5.1 1.0
this GCL was sodic 共Na rich兲, and therefore was not a source of 5 Silty sand 31 32 8 5.8 1.0
multivalent cations. 6 Silty sand 45 49 18 5.2 1.4
The findings in Meer and Benson 共2007兲 suggest that large 7 Silty sand 43 43 13 5.0 1.3
increases in hydraulic conductivity can occur in response to wet- 8 Lean clay 99 42 25 7.9 0.9
dry cycling when water percolating through the cover profile is 9 Sandy silt 82 24 2 7.8 1.9
dominated by multivalent cations. In contrast, when the percolat- 10 Silty sand 7 —a —a 8.0 1.9
ing water is dominated by monovalent cations, wet-dry cycling 11 Clayey sand 31 28 13 7.2 0.5
should have much less effect 共and potentially no effect兲 on the a
Not measured.
hydraulic conductivity of GCLs. This hypothesis was evaluated in
this study by subjecting bentonite from a GCL as well as GCL
specimens to wet-dry cycling using solutions having different Hydraulic conductivity of the as-received GCL was deter-
relative abundance of monovalent and divalent cations 共Na and mined in a flexible-wall permeameter following the methods de-
Ca兲. After each wet-dry cycle, the hydraulic conductivity of the scribed in ASTM D 5084 共ASTM 2004兲. DI water was used as the
GCL and the swell index and cation exchange complex of the permeant liquid, the effective confining stress was 20 kPa, and the
bentonite were determined. Changes in the hydraulic conductiv- hydraulic gradient was 75. Two specimens were tested, and had
ity, swell index, and the cation exchange complex were related to hydraulic conductivities of 2.1⫻ 10−9 and 1.0⫻ 10−9 cm/ s.
the relative abundance of monovalent and multivalent cations in
the hydrating solution.
Swell Index Tests
Swell index tests were conducted according to methods described
Materials and Methods in ASTM D 5890 共ASTM 2004兲. Air-dried bentonite was ground
using a mortar and pestle until 100% passed a No. 200 United
Geosynthetic Clay Liner States sieve. Approximately 90 mL of DI water was poured into a
clean 100 mL graduated cylinder. Two grams of dry bentonite
A roll of GCL provided by a manufacturer was used as the source was then placed in the graduated cylinder in 0.1 g increments.
of the GCL specimens and bentonite used in the study. The GCL Additional test solution was used to rinse any particles adhering
contained granular sodium bentonite encased by two geotextiles to the sides of the cylinder and to fill the cylinder to the 100 mL
共a slit-film woven geotextile and a nonwoven geotextile兲 bonded mark. After 24 h of exposure, the swell index was recorded.
by needle punching. The mass per unit area of air-dry bentonite in
the GCL was 4.3 kg/ m2, the initial air-dry thickness ranged from
Solutions
5.8 to 7.0 mm, and the average initial air-dry water content of the
bentonite was 7.0%. The bentonite consisted of sand-size granules Column elution tests were conducted on 11 soils sampled from
共0.075– 2.0 mm兲 composed primarily of clay-size particles 共87% the surface layer of landfills throughout the United States to ob-
finer than 2 ␮m兲 and had a liquid limit of 504 and a plasticity tain an estimate of the ionic strength and relative abundance of
index of 465. X-Ray diffraction showed that the bentonite con- cations in water contacting GCLs in the field. A sample of each
tained 80% montmorillonite, 7% plagioclase, 6% cristobalite, and cover soil was collected in a 20 L bucket, which was sealed and
trace levels 共艋2 % 兲 of illite, mica, heulandite, gypsum, and shipped to the laboratory for testing. Index properties of the soils
quartz. are summarized in Table 1.
The cation exchange complex of the bentonite was determined The column elution tests were conducted in rigid-wall per-
by extraction using the ammonium acetate method 共Thomas meameters similar to those described in ASTM D 5856 共ASTM
1982兲 on 10 g of dry bentonite crushed to pass a No. 20 United 2004兲. Specimens were prepared by compaction in a stainless
States standard sieve. Soluble salts were extracted beforehand steel compaction mold 共diameter= 105 mm, height= 75 mm兲 to a
using the saturation extraction procedure described in Rhoades dry unit weight corresponding to 85% relative compaction per
共1982兲 with a solid-DI water ratio of 1:5. The ammonium acetate standard Proctor to simulate the low compactive effort normally
extraction was conducted with a 1 M ammonium acetate solution applied to surface layer soils. Both ends of the specimen were
at a solid-liquid ratio of 1:5. The mixture was shaken for 24 h, covered with disks of nonwoven geotextile, and a porous stone
after which the solid and liquid were separated by vacuum filtra- was placed on top of the upper geotextile to distribute the influent
tion using Whatman No. 42 filter paper. Concentrations of the water. Effluent from the column was analyzed for concentrations
exchangeable cations Na, K, Ca, and Mg in the extract were mea- of Ca, Mg, Na, and K by AAS, as described previously.
sured using atomic absorption spectroscopy 共AAS兲 following A solution representing synthetic rainwater was allowed to
USEPA Method 200.7. The exchange complex of bentonite in the slowly drip 共2 mL/ h兲 onto the porous stone to simulate the slow
GCL contained Na 共73% mole fraction兲, Ca 共22% mole fraction兲, unsaturated infiltration that might occur in the field. Ionic compo-
Mg 共3% mole fraction兲, and K 共2% mole fraction兲 in the initial sition of the synthetic rainwater was based on an analysis of rain-
condition. water chemistry from 18 locations in North America, Europe, and

350 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MARCH 2009

Downloaded 03 Mar 2009 to 69.91.147.111. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
 Table 2. Ionic Strength, RMD, pH, and Electrical Conductivity of Test
Solutions Used in Laboratory Tests
Ionic Electrical
strength RMD conductivity
共M兲 共M0.5兲 pH 共S/m兲
0 — 7.0 0.05
0.005 0.007 6.5 9.9
0.005 0.07 6.7 7.7
0.005 0.7 6.6 6.0
0.011 0.007 6.2 22.0
0.011 0.07 6.0 18.0
0.011 0.7 6.0 14.9
0.025 0.007 6.5 44.2
0.025 0.07 6.2 36.2
0.025 0.10 6.2 36.0
0.025 0.14 6.3 35.6
0.025 0.7 6.0 32.0

cations. A similar dependence on soil type is not evident in the

RMD of the effluent, with effluent from the clayey and nonclayey
soils having a wide range of RMD.
Eleven solutions were used in the study. Ionic strength and
RMD of the solutions are summarized in Table 2. The RMD of
these solutions spans the range of RMD in the effluent from the
column elution tests 关Fig. 1共b兲兴. Three ionic strengths were used
共0.005, 0.011, and 0.025 M兲. The upper bound on steady concen-
trations observed in the elution tests was represented by an ionic
Fig. 1. Ionic strength 共a兲; RMD 共b兲 from column tests on cover soils
as function of eluted volume
strength of 0.005 M, and the upper bound from the entire set of
elution data was represented by an ionic strength of 0.025 M 关Fig.
1共a兲兴. The intermediate ionic strength 共0.011 M兲 was used to rep-
resent an intermediate condition. Type II DI water 共ASTM D
Asia 共Meer and Benson 2004兲. The synthetic rainwater was pre- 1193, ASTM 2004兲 was used as a control. Jo et al. 共2001, 2004兲
pared by dissolving NaCl and CaCl2 salts in DI water to create a and Kolstad et al. 共2004兲 show that cation valence is the most
solution having an ionic strength of 0.0008 M, RMD of significant factor affecting swell of bentonite and hydraulic con-
0.02 M1/2, and pH 7.1. RMD is defined as M m / M 1/2 ductivity of GCLs at a given molarity, whereas cation species for
d , where
M m⫽total molarity of monovalent cations and M d⫽total molarity a given valence has no measurable impact. Thus, all solutions
of multivalent cations in the solution, and represents the relative were prepared with anhydrous NaCl and CaCl2 dissociated in
abundance of monovalent and multivalent cations at a given ionic Type II DI water 共i.e., Na was the sole monovalent cation and Ca
strength. For inorganic aqueous permeant solutions, ionic strength was the sole divalent cation兲.
and RMD are master variables controlling the hydraulic conduc-
tivity of GCLs for pH between 2 and 12 共Jo et al. 2001, 2004; Batch Tests
Kolstad et al. 2004兲. The denominator of RMD includes divalent
and other multivalent cations because Jo et al. 共2001兲 show that Batch tests were conducted using bentonite from the GCL to
multivalent cations have a similar impact on hydraulic conductiv- evaluate how the combined effects of desiccation and cation ex-
ity at the same molarity, regardless of valence. However, in most change affect swelling of the bentonite and composition of the
pore waters in cover soils including those evaluated in this study, exchange complex. For each batch test series, 80 g of bentonite
divalent cations dominate the multivalent cations. Thus, the mul- from the GCL was ground using a mortar and pestle until it
tivalent nomenclature is retained herein for consistency with past passed the No. 20 United States sieve. A 2 g subsample was set
studies, but the solutions that were used in this study only con- aside for swell index testing, and the remainder was mixed in 2 L
tained monovalent and divalent cations. glass jars with one of the salt solutions summarized in Table 2.
Ionic strength and RMD of the effluent from the cover soil The jars were capped and tumbled end-over-end at 30 rpm.
elution tests is shown in Fig. 1 as a function of eluted volume. After tumbling, the slurry was emptied into polypropylene
Both ionic strength and RMD typically decreased with eluted pans for air drying under a vacuum hood. The mass of each pan
volume and then leveled off, although the magnitude of the drop was weighed daily. Once the mass was stable, the bentonite was
varied considerably. Between cover soils, the steady-state ionic reground until it passed the No. 20 United States standard sieve.
strength varied by approximately a factor of ten and the RMD A 10 g portion of the reground bentonite was set aside to measure
varied by approximately a factor of 20. Effluent from the more the cation composition of the exchange complex and 2 g was set
clayey soils typically had lower ionic strength than effluent from aside for a swell index test. The bentonite was then mixed again
the other soils, which may reflect the greater affinity of clays for with an identical salt solution at the same solid-to-liquid ratio.

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MARCH 2009 / 351

Downloaded 03 Mar 2009 to 69.91.147.111. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
 The GCLs were placed in the permeameter and allowed to
hydrate in the permeant liquid for 48 h under no hydraulic gradi-
ent as recommended in Jo et al. 共2001兲. The solutions shown in
Table 2 were used as permeant liquids. After hydration, a hydrau-
lic conductivity test was conducted following the methods in
ASTM D 5084. Each specimen was permeated for 30 days to
simulate percolation during wet spring conditions, as suggested
by Lin and Benson 共2000兲. During this period, the data were
inspected to determine if the hydraulic conductivity was steady
共⫾25% from the mean and no visible trend兲 and the ratio of
incremental inflow to outflow was between 0.75 and 1.25. In all
cases, both of these criteria were satisfied during the 30 day per-
meation period, and generally in ⬍10 days. Because the intent
was to simulate cation exchange that occurs during a wet spring
condition, establishing chemical equilibrium between the bento-
Fig. 2. Ca, Mg, Na, and K concentrations in batch-test solution as nite and the permeation liquid was not a termination criterion.
function of tumbling time Wet-dry cycling was conducted using the method in Lin and
Benson 共2000兲. After permeation, GCL specimens were air dried
on a bench until the weight of the specimen did not change.
This process was repeated until the 80 g of bentonite was ex- Overburden pressure was not applied during drying. Meer and
hausted, or the swell index and exchange complex data indicated Benson 共2007兲 compared hydraulic conductivities of specimens
that exchange of Ca for Na was complete. dried with and without overburden pressure, and indicate that
The tumbling time required to achieve equilibrium between overburden pressure during drying had no noticeable effect on the
the bentonite and the solution was determined by collecting ali- hydraulic conductivity. Typically 7 – 10 days were required to
quots of the solution over time while tumbling the mixtures for complete the drying cycle, after which the air-dry water content
48 h. Each aliquot was vacuum filtered using a Buchner funnel of the bentonite was on the order of 20–30%. Specimens were
and Whatman No. 42 filter paper. Cation concentrations in the subjected to 5–9 wet-dry cycles using this procedure. After the
filtrate were determined by AAS as cited previously. Results of final wetting, the thickness of the GCL was measured, the water
the analysis showed that the concentrations of Na, K, Ca, and Mg content and swell index of the bentonite were determined, and the
did not change significantly after approximately 14 h 共Fig. 2兲. composition of the exchange complex was measured.
Therefore, all batch tests were conducted with a convenient 24 h The wet-dry procedure that was used is severe, as the GCL is
tumbling time. dried to a water content at the low end of water contents observed
in the field 共Meer and Benson 2007; Benson et al. 2007兲. How-
ever, as described subsequently, the GCLs that experienced
Hydraulic Conductivity Tests changes in hydraulic conductivity due to wet-dry cycling follow-
Falling-head hydraulic conductivity tests were conducted on the ing the aforementioned procedure had hydraulic conductivities at
GCLs in flexible-wall permeameters according to methods de- the end of testing comparable to those measured on specimens
scribed in ASTM D 5084. An average effective stress of 20 kPa exhumed from field sites by Melchior 共2002兲, Meer and Benson
was applied to simulate the effective stress applied to the GCL in 共2007兲, and Benson et al. 共2007兲.
a final cover with approximately 1 m of cover soil. No backpres-
sure was applied so that samples for pH and electrical conductiv-
ity measurements could be collected easily. Results and Discussion
The hydraulic gradient that was applied varied depending on
the hydraulic conductivity of the specimen, with a maximum hy-
Swell Index
draulic gradient of 75 being applied to specimens with lower
hydraulic conductivity. This gradient is higher than the field gra- Swell index of the bentonite subjected to wet-dry cycling using
dient. However, Shackelford et al. 共2000兲 indicate that the hy- the batch procedure is shown in Fig. 3共a兲. A decrease in swell
draulic conductivity of GCLs is relatively insensitive to the index following the first wet-dry cycle was observed for all solu-
hydraulic gradient. Thus, the effect of the elevated gradient is tions. For some solutions 共DI water and salt solutions having
believed to be negligible. Lower gradients were used for the more RMD= 0.7 M0.5兲, the decrease in swell index was followed by an
permeable specimens to permit convenient collection of effluent. increase in swell index with additional wet-dry cycles 共bentonites
GCLs specimens were prepared for hydraulic conductivity that “retained swell”兲, whereas the swell index changed negligibly
testing using the method described in Jo et al. 共2001兲. A razor or decreased modestly 共bentonites that “lost swell”兲 in the other
knife was used to cut the GCL along the outer circumference of a solutions. For most solutions, the swell index ceased to change
stainless steel cutting ring 共152 mm diameter兲. A small volume of significantly after 3 – 4 cycles 共DI water is an exception兲. The
permeant liquid was applied to the GCL along the inner circum- bentonites that retained swell 共solid symbols兲 retain a swell index
ference of the cutting ring during trimming to induce local hydra- of at least 23 mL/ 2 g after repeated wet-dry cycling, whereas the
tion and prevent loss of bentonite. After cutting, the specimen was bentonites that lost swell 共open symbols兲 have a swell index
removed from the cutting ring and excess geotextile fibers along 艋15 mL/ 2 g. These changes in swell behavior reflect the relative
the edge of the GCL were removed. Bentonite paste, prepared abundance of monovalent and divalent cations in the hydrating
with bentonite from the GCL and the permeant liquid, was ap- solution, and their impact on the primary cations in the exchange
plied around the perimeter of the GCL to reduce the potential for complex. This correspondence is illustrated in Fig. 4; swell index
sidewall leakage. increases monotonically as the mole fraction of Na in the ex-

352 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MARCH 2009

Downloaded 03 Mar 2009 to 69.91.147.111. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
 Fig. 3. Swell index 共a兲; mole fraction of Na on exchange complex of the bentonite 共b兲 as function of number of wet-dry cycles

change complex increases. Except for one case, the bentonites Na–salts in the Na–bentonite. In contrast, the bentonites that lost
that have lost swell have a mole fraction of Na less than 0.4. swell were hydrated with solutions having a greater abundance of
The bentonites that retained swell were hydrated with solu- divalent cations 共RMD艋 0.07 M0.5兲, which resulted in the re-
tions having the greatest relative abundance of monovalent cat- placement of Na by Ca as the number of wet-dry cycles in-
ions: DI water or solutions having an RMD= 0.7 M0.5. DI water is creased. This exchange effect is illustrated in Fig. 5 using data
included as a hydration solution that contains predominantly from the batch tests conducted with an ionic strength of 0.005 M
monovalent cations, because the cations present in the batch liq- and RMD= 0.007 M0.5. The rate at which this effect occurs is
uid prepared with DI water are derived primarily from soluble

Fig. 5. Mole fraction of Na and Ca as function of number of wet-dry

Fig. 4. Relationship between swell index of bentonites in batch tests cycles for bentonite in batch tests. Solution had ionic strength of
and mole fraction of Na in exchange complex 0.005 M and RMD= 0.007 M0.5.

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MARCH 2009 / 353

Downloaded 03 Mar 2009 to 69.91.147.111. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
 Fig. 7. Hydraulic conductivity of GCLs permeated with solutions
having ionic strength= 0.005 and 0.025 M and RMD ranging from
0.007 to 0.7 M0.5 as function of number of wet-dry cycles

where the exchange complex consists primarily of divalent cat-

ions 关Fig. 3共b兲兴 共Norrish and Quirk 1954; McBride 1994; Kolstad
et al. 2004兲.
The sensitivity of the final swell index to ionic strength and
RMD of the test solution is shown in Fig. 6. RMD has a greater
effect on the swell index of bentonite than ionic strength after
repeated cycles of wetting and drying. Swell index is modestly
Fig. 6. Swell index of bentonite from final cycle of batch tests as affected by ionic strength, varying by at most 10 mL/ 2 g over the
function of ionic strength 共a兲; RMD 共b兲 of solution range of ionic strengths that were used, and is more strongly
affected by RMD, decreasing by 19– 25 mL/ 2 g as the RMD
decreases.

illustrated in Fig. 3共b兲. More rapid replacement of Na by Ca oc-

Hydraulic Conductivity
curs with wet-dry cycling when the ionic strength is higher or the
RMD is lower. Hydraulic conductivities of GCL specimens subjected to repeated
Replicate tests were conducted with two solutions that pro- wet-dry cycling via permeation followed by air drying are shown
duced bentonites that retained swell 共DI water and a solution with in Fig. 7. Increases in hydraulic conductivity of 4–5 orders of
I = 0.005 M and RMD= 0.7 M0.5兲 to determine if the decrease in magnitude occurred in the GCL specimens permeated with test
swell index observed for these solutions during the first cycle was solutions having RMD= 0.007 M0.5, whereas no discernable
anomalous. Nearly identical swell indices were obtained from change in hydraulic conductivity occurred when the GCLs were
these replicate tests, even for DI water, confirming that the de- permeated with solutions having RMD= 0.7 M0.5, regardless of
crease observed in the first cycle in all tests was a real phenom- whether the ionic strength was high or low. This distinct differ-
enon 关Fig. 3共a兲兴. However, the mechanism responsible for this ence in behavior is consistent with the changes in swell index
decrease in swell index during the first cycle remains unknown. shown in Fig. 3, and can be attributed to the amount of Ca-for-Na
The final swell indices 共i.e., after the last wet-dry cycle兲 of the exchange that occurred during wet-dry cycling. Final mole frac-
bentonites that retained swell vary with the ionic strength of the tions of Na and Ca on the exchange complex of bentonite from
hydrating solutions. The largest final swell index 共40 mL/ 2 g兲 the GCLs permeated with solutions having RMD= 0.7 M0.5,
corresponds to hydration with DI water and the lowest to a solu- 共XNa = 0.65 and 0.66, XCa = 0.28 and 0.26兲 were found to be similar
tion having an ionic strength of 0.025 M 共25 mL/ 2 g兲. The solu- to those for bentonite from a new GCL 共XNa = 0.73, XCa = 0.22兲
tions with low and intermediate ionic strength 共0.005 and 共Table 3兲. In contrast, the mole fractions of Na and Ca for GCLs
0.011 M兲 yielded an intermediate 共and the same兲 swell index permeated with test solutions having RMD= 0.007 M0.5 共XNa
共35 mL/ 2 g兲. All of these solutions also have RMD of at least = 0.03 and 0.02, XCa = 0.84 and 0.94兲 show that nearly complete
0.7 M0.5. This variation in swell index with ionic strength reflects replacement of Na by Ca occurred during wet-dry cycling. Be-
the well-known sensitivity of osmotic interlayer swell to concen- havior between these two extremes was observed for the GCLs
tration in bentonites where the exchange complex is composed permeated with solutions having ionic strength= 0.025 M and in-
predominantly of monovalent cations 关Fig. 3共b兲兴. In contrast, the termediate RMD 共0.07– 0.14 M0.5兲. In particular, the final hydrau-
final swell indices of the bentonites that lost swell are essentially lic conductivity decreased and mole fraction of Na in the
independent of ionic strength, which reflects the insensitivity of exchange complex increased as RMD increased 共Table 3兲.
crystalline interlayer swelling to concentration in bentonites The large increases in hydraulic conductivity evident in Fig. 7

354 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MARCH 2009

Downloaded 03 Mar 2009 to 69.91.147.111. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
Table 3. Hydraulic Conductivity of GCL and Mole Fraction of Na, Ca, Mg, and K in Exchange Complex of Bentonite after Final Cycle of Wetting and
Drying
Ionic Total Hydraulic Mole fraction in exchange complex
strength RMD number of conductivity
共M兲 共M0.5兲 cycles 共cm/s兲 Na Ca Mg K
— — 0 1.0⫻ 10−9 0.73 0.22 0.03 0.02
2.1⫻ 10−9
0.005 0.007 9 2.5⫻ 10−4 0.03 0.84 0.08 0.05
0.005 0.7 9 1.6⫻ 10−9 0.66 0.26 0.05 0.02
0.025 0.007 5 2.4⫻ 10−5 0.02 0.94 0.01 0.03
0.025 0.07 5 7.1⫻ 10−6 0.14 0.81 0.03 0.02
0.025 0.10 5 4.9⫻ 10−7 0.24 0.70 0.04 0.02
0.025 0.14 5 1.3⫻ 10−8 0.25 0.68 0.05 0.02
0.025 0.7 9 2.9⫻ 10−9 0.65 0.28 0.05 0.02

for lower RMD are caused by desiccation cracks in the bentonite cracks swelled shut during rehydration with solutions having
that do not swell shut during rehydration. Cracks formed in the RMD= 0.7 M0.5 关Fig. 8共d兲兴, but did not completely close in GCLs
bentonite of all GCLs during drying, regardless of the test solu- permeated with test solutions having RMD= 0.007 M0.5 关Fig.
tion that was used. Examples of cracks formed during desiccation 8共b兲兴. The effect of Ca-for-Na exchange on swelling of the ben-
are shown in Fig. 8 for the tests conducted with I = 0.005 M and tonite is responsible for these differences in self-healing capacity.
RMD= 0.007 M0.5 关Fig. 8共a兲兴 or 0.7 M0.5 关Fig. 8共c兲兴 共white arrows GCLs exposed to solutions that induce Ca-for-Na exchange
on the photographs in Fig. 8 indicate locations of cracks兲. These 共lower RMD兲 lose their swelling capacity 关Fig. 3共a兲兴 and ability to

Fig. 8. Bentonite in GCL specimen permeated with I = 0.005 M and RMD= 0.007 M0.5 solution after air drying 共a兲; following final wetting cycle
共b兲; exposed bentonite of GCL specimen permeated with I = 0.005 M and RMD= 0.7 M0.5 solution following after air drying 共c兲; and following
final wetting cycle 共d兲. Examples of cracks illustrated with white arrows.

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MARCH 2009 / 355

Downloaded 03 Mar 2009 to 69.91.147.111. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
 than the other GCLs, and may not have reached its final hydraulic
conductivity when sampled. Additionally, the ionic strength of the
pore water from the cover soil at Site D was the lowest of those
evaluated by Meer and Benson 共2007兲. This factor may also have
contributed to a greater number of wet-dry cycles being required
before a large increase in hydraulic conductivity occurred 共see
subsequent discussion兲.
No threshold in RMD is apparent in Fig. 9 above which the
hydraulic conductivity of the GCL is unaffected by wet-dry cy-
cling, although a threshold may have been identified if tests had
been conducted for RMDs between 0.14 and 0.7 M0.5. However,
the trend in Fig. 9 suggests that only modest changes in hydraulic
conductivity should be expected when the permeant liquid has
RMD⬎ 0.14 M0.5. Findings reported in Meer and Benson 共2007兲
suggest that RMD of pore waters can be estimated reliably using
Fig. 9. Hydraulic conductivity of GCL at end of final wet-dry cycle the batch water leach test defined in ASTM D 6141 共ASTM
versus RMD of permeant solution 2004兲. Thus, batch water leach testing may prove to be useful for
evaluating the compatibility between cover soils and GCLs.
Comparison of the hydraulic conductivities in Fig. 9 corre-
self-heal, as reported previously by Lin and Benson 共2000兲, Meer sponding to RMD of 0.007 and 0.7 M0.5 and ionic strengths of
and Benson 共2007兲, and Benson et al. 共2007兲. In contrast, GCLs 0.005 and 0.025 M suggests that RMD has a greater influence on
exposed to solutions that do not induce Ca-for-Na exchange final hydraulic conductivity than ionic strength. The greater im-
共higher RMD兲 retain their swelling capacity 关Fig. 3共a兲兴 and ability portance of RMD is also evident in Fig. 7. For RMD of 0.7 M0.5,
to self-heal during rehydration following desiccation. essentially the same hydraulic conductivity was obtained for both
ionic strengths for all wet-dry cycles. Similarly, for RMD of
0.007 M0.5, comparable hydraulic conductivities were obtained
Practical Implications for ionic strengths of 0.005 and 0.025 M after the hydraulic con-
ductivity increased and leveled off 共although the final hydraulic
The hydraulic conductivities obtained after wet-dry cycling with conductivities are slightly higher for the tests conducted with the
the solutions having RMD艋 0.07 M0.5 are in the range of 7.1 0.005 M solution兲. However, when the ionic strength was higher,
⫻ 10−6 – 2.5⫻ 10−4 cm/ s. These hydraulic conductivities are simi- the hydraulic conductivity increased after fewer wet-dry cycles.
lar to those of the most permeable GCLs exhumed from final These findings suggest that ionic strength controls the number
covers by Melchior 共2002兲, Benson et al. 共2007兲, and Meer and of cycles required for an increase in hydraulic conductivity, and
Benson 共2007兲, which ranged between 4.8⫻ 10−5 and 9.4 RMD controls the final hydraulic conductivity. Similar effects
⫻ 10−4 cm/ s. These hydraulic conductivities are also similar to were observed for index swell of the bentonite exposed to wet-dry
the final hydraulic conductivities reported by Lin and Benson cycling using the batch procedure 共Fig. 3兲, and are apparent in the
共2000兲 for GCLs subjected to wet-dry cycling in the laboratory Na and Ca fractions in the exchange complex 共Table 3兲. As shown
using 0.0125 M Ca solutions. In contrast, the GCLs subjected to in Fig. 10, the mole fraction of Na and Ca at the end of wet-dry
wet-dry cycling with solutions having RMD of 0.7 M0.5 have a cycling is essentially a unique function of the RMD of the per-
final hydraulic conductivity less than 2.9⫻ 10−9 cm/ s, which is meant solution, and is independent of the ionic strength. Given
similar to the average hydraulic conductivity of the exhumed that GCLs in covers generally are anticipated to have a service
GCLs described in Mansour 共2001兲 共1.9⫻ 10−9 cm/ s兲. This cor- life of decades during which they could be exposed to numerous
respondence between field and laboratory conditions for GCLs wet-dry cycles, RMD of the pore water of the adjacent cover soil
having both high and low hydraulic conductivities suggests that should be a primary factor considered when evaluating the com-
the testing method used in this study, while severe, provides a patibility of GCLs and cover soils. Moreover, the pore water in
reasonable estimate of field conditions. adjacent cover soils should be evaluated under realistic elution
Influence of the relative abundance of monovalent and multi- conditions, as the RMD of the pore water will vary as the water
valent cations on the hydraulic conductivity after wet-dry cycling content of the cover soil changes during wetting and drying.
is shown in Fig. 9 in terms of the hydraulic conductivity mea- Changes in water content alter cation concentrations and the ionic
sured at the end of the last cycle versus the RMD of the permeant strength, and result in a nonlinear variation in RMD even when
liquid. The hydraulic conductivity decreases rapidly as the RMD the population of cations in the pore water remains unchanged
increases from 0.007 to 0.14 M0.5. Maximum hydraulic conduc- 共i.e., due to the square root term in the denominator of RMD兲.
tivities of the GCLs exhumed by Meer and Benson 共2007兲 that The final hydraulic conductivity of the GCLs evaluated in this
were not overlain by a geomembrane are also shown in Fig. 9 study is shown in Fig. 11 as a function of the index swell of the
along with the RMD of the pore water in the cover soil. Good bentonite after the final hydraulic conductivity test. The hydraulic
agreement exists between the relationship between hydraulic con- conductivity diminishes as the index swell increases, which re-
ductivities and RMD for the specimens tested in this laboratory flects the effect of cation exchange on the ability of the bentonite
study and for the GCLs exhumed by Meer and Benson 共2007兲. to swell during rehydration and close macroscopic features that
The exception is the GCL exhumed from Site D by Meer and formed during drying. Relatively low hydraulic conductivities
Benson 共2007兲. This GCL was exhumed from a cover over a 共艋10−8 cm/ s兲 are achieved when the index swell exceeds
small landfill in a national forest in northern Wisconsin where wet 15 mL/ 2 g, which is the same threshold observed for bentonites
conditions are common and tree cover is extensive. Thus, the that lost swell in the batch tests 共Fig. 3兲. The trend between hy-
GCL at Site D may have been exposed to fewer wet-dry cycles draulic conductivity and index swell also has a shape similar to

356 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MARCH 2009

Downloaded 03 Mar 2009 to 69.91.147.111. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
 Fig. 12. Hydraulic conductivity versus time for GCL permeated with
0.0125 M CaCl2 solution for 1,600 days, dried to gravimetric water
content of 50%, and permeated again with 0.0125 M CaCl2 solution
共adapted from Benson et al. 2007兲
Fig. 10. Mole fraction of Na or Ca in exchange complex of bentonite
from GCL specimens subjected to wet-dry cycling as function of
RMD of permeant solution. For specimens permeated with DI water, divalent or multivalent cations is known to result in GCLs with
RMD was set at 1.5 M0.5. hydraulic conductivities on the order of 10−8 cm/ s 共Egloffstein
2001; Jo et al. 2004, 2005; Benson et al. 2007兲. This effect is
illustrated in Fig. 12, which shows hydraulic conductivity versus
the relationship between hydraulic conductivity and RMD, which time for a GCL permeated with a 0.0125 M CaCl2 solution by
suggests that RMD of the permeant liquid is the primary factor Benson et al. 共2007兲. The GCL was permeated for 1,599 days
affecting the hydraulic conductivity of GCLs exposed to wet-dry with the CaCl2 solution, which resulted in a gradual increase in
cycling. hydraulic conductivity to 2.3⫻ 10−8 cm/ s that was maintained for
The correspondence between hydraulic conductivity and index the duration of the test period. After 1,599 days, the specimen
swell in Fig 11 suggests that index swell testing may prove to be was removed and dried to a water content of 50%, and then per-
a convenient means to assess the compatibility of GCLs with meated again with the CaCl2 solution for 30 days. A single cycle
cover soils. However, the index swell data in Fig. 11 are from of desiccation caused the hydraulic conductivity of the GCL to
bentonites that have undergone repeated wet-dry cycling. A swell increase to 4.7⫻ 10−6 cm/ s. Thus, large increases in hydraulic
test conducted with pore water and fresh bentonite from a GCL conductivity of GCLs that are initially hydrated require both cat-
would not necessarily show a similar diminished swell index, as ion exchange and desiccation. Protective methods that ensure ini-
reported by Meer and Benson 共2007兲. Rather, an index swell pro- tial hydration of the bentonite and prevent wet-dry cycling 共e.g., a
cedure with pore water and wet-dry cycling would need to be GCL overlain by a geomembrane that is promptly covered with
developed. soil兲 are likely to preclude large increases in hydraulic conductiv-
An additional important consideration is that cation exchange ity even if cation exchange occurs.
alone will not result in large increases in hydraulic conductivity.
Most pore waters tend to be dilute 共e.g., as shown in Fig. 1兲, and
cation exchange with dilute solutions that contain predominantly Summary and Conclusions

A series of batch tests and hydraulic conductivity tests was con-

ducted on a GCL containing Na–bentonite to evaluate how swell
index and hydraulic conductivity are affected by wet-dry cycling
in solutions having different relative abundance of monovalent
and multivalent cations. Relative abundance of monovalent and
multivalent cations was characterized by the RMD of the test
solution, which is defined as the ratio of the total molarity of
monovalent cations to the square root of the total molarity of
multivalent cations. Test solutions having a range of RMD and
two ionic strengths were prepared using NaCl and CaCl2 dis-
solved in DI water.
Results of the batch and hydraulic conductivity tests showed
that RMD controls the final swell index, the relative abundance of
monovalent and divalent cations in the final cation exchange com-
plex, and the final hydraulic conductivity of bentonite exposed to
wet-dry cycling. Ionic strength affects the number of wet-dry
Fig. 11. Hydraulic conductivity of GCL at end of final wet-dry cycle cycles required for a change in hydraulic conductivity to occur
versus swell index for final wet-dry cycle and the rate at which the cation exchange complex and swell

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MARCH 2009 / 357

Downloaded 03 Mar 2009 to 69.91.147.111. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright
index change with wet-dry cycling. Lower swell index and higher ated with weak inorganic salt solutions.” Clays Clay Miner., 52共6兲,
hydraulic conductivity were obtained as the RMD decreased, 661–679.
which reflected greater replacement of the native Na in the ben- Jo, H., Benson, C., Lee, J., Shackelford, C., and Edil, T. 共2005兲. “Long-
tonite exchange complex with Ca. Large increases in hydraulic term hydraulic conductivity of a non-prehydrated geosynthetic clay
liner permeated with inorganic salt solutions.” J. Geotech. Geoenvi-
conductivity and loss of swelling capacity occurred when the so-
ron. Eng., 131共4兲, 405–417.
lution had an RMD艋 0.07 M0.5. Modest or small changes in hy-
Jo, H., Katsumi, T., Benson, C., and Edil, T. 共2001兲. “Hydraulic conduc-
draulic conductivity and swell index were obtained when the tivity and swelling of nonprehydrated GCLs permeated with single-
RMD was 艌0.14 M0.5. These findings suggest that chemical species salt solutions.” J. Geotech. Geoenviron. Eng., 127, 557–567.
analysis of the pore water in cover soils may prove useful in Katsumi, T., Ishimori, H., Ogawa, A., Yoshikawa, K., Hanamoto, K., and
evaluating the compatibility of GCLs and covers soils used in Fukagawa, R. 共2007兲. “Hydraulic conductivity of nonprehydrated
applications where wet-dry cycling may occur. geosynthetic clay liners permeated with inorganic solutions and waste
The findings also illustrate that large increases in hydraulic leachates.” Soils Found., 47共1兲, 79–96.
conductivity of GCLs used in cover applications where the ben- Kolstad, D., Benson, C., and Edil, T. 共2004兲. “Hydraulic conductivity and
tonite is initially hydrated occur when the native Na is replaced swell of nonprehydrated GCLs permeated with multi-species inor-
by multivalent cations and the bentonite is exposed to wet-dry ganic solutions.” J. Geotech. Geoenviron. Eng., 130共12兲, 1236–1249.
cycling. Protective measures that ensure initial hydration and Lee, J., and Shackelford, C. 共2005兲. “Impact of bentonite quality on hy-
draulic conductivity of geosynthetic clay liners.” J. Geotech. Geoen-
prevent wet-dry cycling or cation exchange will likely prevent
viron. Eng., 131共1兲, 64–77.
large increases in hydraulic conductivity of GCLs used in cover Lee, J., Shackelford, C., Benson, C., Jo, H., and Edil, T. 共2005兲. “Corre-
applications. lating index properties and hydraulic conductivity of geosynthetic
clay liners.” J. Geotech. Geoenviron. Eng., 131共11兲, 1319–1329.
Lin, L., and Benson, C. 共2000兲. “Effect of wet-dry cycling on swelling
Acknowledgments and hydraulic conductivity of GCLs.” J. Geotech. Geoenviron. Eng.,
126, 40–49.
Financial support for this study was provided by the United States Mansour, R. 共2001兲. “GCL performance in semi-arid climate conditions.”
Environmental Protection Agency 共USEPA兲 共Contract No. 2C- Proc., 8th Sardinia 2001, Int. Waste Management and Landfill Symp.,
R361-NAEX兲 and the United States National Science Foundation T. Christensen, R. Cossu, and R. Stegmann, eds., CISA, Cagliari,
共NSF兲 under Grant Nos. CMS-9900336 and CMMI-0625850. Italy, 219–226.
McBride, M. 共1994兲. Environmental chemistry of soils, Oxford University
David Carson and Thabet Tolaymat were the project managers for
Press, New York.
the portion funded by USEPA. The findings in this paper are Meer, S., and Benson, C. 共2004兲. “In-service hydraulic conductivity of
solely those of the writers. This paper has not been reviewed by GCLs in landfill covers: Laboratory and field studies.” Rep. No. EPA/
USEPA or NSF. Endorsement by USEPA or NSF is not implied 600/R-05/148, U.S. Environmental Protection Agency, Washington,
and should not be assumed. Tammy Rauen, Sabrina Bradshaw, D.C.
and Joseph Scalia of the University of Wisconsin-Madison as- Meer, S., and Benson, C. 共2007兲. “Hydraulic conductivity of geosynthetic
sisted with the testing program. James Olsta of CETCO provided clay liners exhumed from landfill final covers.” J. Geotech. Geoenvi-
valuable comments during preparation of the manuscript. ron. Eng., 133共5兲, 550–563.
Melchior, S. 共1997兲. “In-situ studies on the performance of landfill caps.”
Proc., Int. Containment Technology Conf. 1997, U.S. Dept. of Energy,
Germantown, Md., 365–373.
References
Melchior, S. 共2002兲. “Field studies and excavations of geosynthetic clay
barriers in landfill covers.” Clay geosynthetic barriers, H. Zanzinger,
ASTM. 共2004兲. Annual book of ASTM standards, Vol. 04.08, 04.09, and
R. M. Koerner, and E. Gartung, eds., Nuremberg, Germany, 321–330.
110.1, West Conshohocken, Pa.
Norrish, K., and Quirk, J. 共1954兲. “Crystalline swelling of montmorillo-
Benson, C., Abichou, T., and Jo, H. 共2004兲. “Forensic analysis of exces-
nite, use of electrolytes to control swelling.” Nature (London), 173,
sive leakage from lagoons lined with a composite GCL.” Geosynthet.
255–257.
Int., 11共3兲, 242–252.
Petrov, R., and Rowe, R. 共1997兲. “Geosynthetic clay liner 共GCL兲—
Benson, C., Thorstad, P., Jo, H., and Rock, S. 共2007兲. “Hydraulic perfor-
Chemical compatibility by hydraulic conductivity testing and factors
mance of geosynthetic clay liners in a landfill final cover.” J. Geotech.
impacting its performance.” Can. Geotech. J., 34, 863–885.
Geoenviron. Eng., 133共7兲, 814–827.
Bouazza, A. 共2002兲. “Geosynthetic clay liners.” Geotext. Geomembr., 20, Rhoades, J. D. 共1982兲. “Soluble salts.” Methods of soil analysis. Part 2.
3–17. Chemical and microbiological properties, 2nd Ed., A. L. Page, R. H.
Egloffstein, T. 共2001兲. “Natural bentonites-influence of the ion exchange Miller, and D. R. Keeney, eds., Soil Science of America, Madison,
and partial desiccation on permeability and self-healing capacity of Wis., 167–179.
bentonites used in GCLs.” Geotext. Geomembr., 19, 427–444. Ruhl, J., and Daniel, D. 共1997兲. “Geosynthetic clay liners permeated with
Gleason, M., Daniel, D., and Eykholt, G. 共1997兲. “Calcium and sodium chemical solutions and leachates.” J. Geotech. Geoenviron. Eng.,
bentonite for hydraulic containment applications.” J. Geotech. Geoen- 123, 369–381.
viron. Eng., 123, 438–445. Shackelford, C., Benson, C., Katsumi, T., Edil, T., and Lin, L. 共2000兲.
Guyonnet, D., Gaucher, E., Gaboriau, H., Pons, C., Clinard, C., Norotte, “Evaluating the hydraulic conductivity of GCLs permeated with non-
V., and Didier, G. 共2005兲. “Geosynthetic clay liner interaction with standard liquids.” Geotext. Geomembr., 18, 133–162.
leachate: Correlation between permeability, microstructure, and sur- Shan, H., and Daniel, D. 共1991兲. “Results of laboratory tests on a
face chemistry.” J. Geotech. Geoenviron. Eng., 131共6兲, 740–749. geotextile/bentonite liner material.” Proc., Geosynthetic ’91, Indus-
James, A. N., Fullerton, D., and Drake, R. 共1997兲. “Field performance of trial Fabrics Association International, St. Paul, Minn., 517–535.
GCL under ion exchange conditions.” J. Geotech. Geoenviron. Eng., Thomas, G. W. 共1982兲. “Exchangeable cations.” Methods of soil analysis.
123, 897–901. Part 2. Chemical and microbiological properties, 2nd Ed., A. L. Page,
Jo, H., Benson, C., and Edil, T. 共2004兲. “Hydraulic conductivity and R. H. Miller, and D. R. Keeney eds., Soil Science of America, Madi-
cation exchange in non-prehydrated and prehydrated bentonite perme- son, Wis., 167–179.

358 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / MARCH 2009

View publication stats Downloaded 03 Mar 2009 to 69.91.147.111. Redistribution subject to ASCE license or copyright; see http://pubs.asce.org/copyright