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

net/publication/238790976

Permeability of Damaged Geosynthetic Clay Liners

Article  in  Geosynthetics International · January 2000

DOI: 10.1680/gein.7.0168

CITATIONS READS
35 393

2 authors, including:

Francesco Mazzieri
Università Politecnica delle Marche
50 PUBLICATIONS   382 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Francesco Mazzieri on 17 February 2015.

The user has requested enhancement of the downloaded file.

Technical Paper by F. Mazzieri and E. Pasqualini

PERMEABILITY OF DAMAGED GEOSYNTHETIC

CLAY LINERS

ABSTRACT: The results of an experimental test program for the permeability of a

damaged, adhesive-bonded, geosynthetic clay liner (GCL) are presented. GCL
specimens were damaged to simulate damage occurring during handling or installation.
Either a portion of the GCL was removed to simulate tearing with bentonite loss, or the
GCL was punctured with a metal object to simulate damage without bentonite loss. In
general, positive results were obtained for the case of damage with bentonite loss: holes
up to 3 cm in diameter self-healed with only a slight increase in hydraulic conductivity
compared to an intact GCL. However, the self-healing process can be compromised if
torn GCLs with bentonite loss are placed on a coarse subgrade with large pore openings,
which allows the bentonite to migrate from the puncture in the GCL. Damage without
bentonite loss was simulated for both hydrated and unhydrated GCL specimens.
Complete self-healing of punctures was generally obtained, but it was also shown that
the conversion of sodium bentonite to calcium bentonite causes a drastic reduction of
the self-healing capacity of the GCL, confirming that ion exchange is particularly
harmful to the overall performance of GCLs.

KEYWORDS: GCL, Bentonite, Permeability, Damage, Self-Healing, Ion exchange.

AUTHORS: F. Mazzieri, Researcher, and E. Pasqualini, Professor, Department of

Earth Sciences, University of Ancona, Via Brecce Bianche 60131 Ancona, Italy,
Telephone: 39/071-2204792, Telefax: 39/071-2204729, E-mail: [email protected]
and [email protected], respectively.

PUBLICATION: Geosynthetics International is published by the Industrial Fabrics

Association International, 1801 County Road B West, Roseville, Minnesota
55113-4061, USA, Telephone: 1/651-222-2508, Telefax: 1/651-631-9334.
Geosynthetics International is registered under ISSN 1072-6349.

DATES: Original manuscript received 20 November 1999, revised version received

20 April 2000 and accepted 27 April 2000. Discussion open until 1 January 2001.

REFERENCE: Mazzieri, F. and Pasqualini, E., 2000, “Permeability of Damaged

Geosynthetic Clay Liners”, Geosynthetics International, Vol. 7, No. 2, pp. 101-118.

GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2 101

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

1 INTRODUCTION

Geosynthetic clay liners (GCLs) first appeared in the early 1980s and, since then,
have been gaining acceptance in a wide range of geotechnical and environmental ap-
plications. GCLs are frequently used as low permeability barriers in landfill liner and
cover systems and in surface facilities for the containment of water (e.g. decorative
ponds, canals, reservoirs, etc.).
GCLs are considered particularly suitable for lining applications over large surfaces
because placement and seaming procedures for GCL panels are relatively simple. At
the same time, however, GCLs are very sensitive to inappropriate installation tech-
niques. Didier and Norotte (1998) illustrated typical errors in GCL handling and instal-
lation that lead to extensive GCL damage, which may potentially compromise the
hydraulic performance.
Although it is generally recognised that GCLs have the property to self-heal if “mod-
erately” damaged, there are relatively few publications regarding the experimental
evaluation of the hydraulic performance of damaged, i.e. torn or punctured, GCLs.
A detailed experimental program was conducted to gain further insight and knowl-
edge and to examine the ramifications of GCL damage on their performance. Tearing
(with bentonite loss) or puncturing (without bentonite loss) of GCL specimens was sim-
ulated. The permeability of damaged specimens was measured and compared with that
of intact specimens. The influence of different factors, such as the extent of damage,
the role of the underlying material, confining stress, hydration state, and ion exchange,
on the hydraulic performance of the GCL was investigated.

2 LITERATURE REVIEW

Previous research has directly addressed the issue of the permeability of damaged
GCLs. Shan (1990) reported that a GCL was capable of “self-healing” holes up to 2.5
cm in diameter with only a slight increase in the hydraulic conductivity (for specimen
diameters of 10 cm). Bouazza et al. (1996) measured the permeability of damaged, 10
cm-diameter GCL specimens; the damage was simulated by making a 1 cm-diameter
hole, or by puncturing the specimen with a thin pin. Bouazza et al. (1996) reported a
minor increase in the hydraulic conductivity of the damaged specimens compared to
intact specimens.
Another group of studies have addressed the behaviour of GCLs in the presence of
dry-wet cycles (Heyer 1995; Boardman and Daniel 1996), freeze-thaw cycles (Kraus
et al. 1997; Hewitt and Daniel 1997), and distortion caused by differential settlement
of landfill covers (La Gatta et al. 1997). In all of the mentioned conditions, the bentonite
layer was subjected to extensive cracking and fissuring, and the hydraulic performance
of the GCLs strongly depended on the capacity of the bentonite to seal, or “self-heal”,
the cracks. In general, it has been concluded that the self-heal potential of GCLs is very
high, owing to the unique swelling property of the sodium bentonite when hydrated with
water. However, recent experimental evidence (Benson 1999) shows a decay in the self-
heal capacity of a GCL subjected to dry-wet cycles when desiccation was coupled with
an ion exchange of sodium for calcium.

102 GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

Fox (1998) investigated the effect of stress concentration, caused by an overlying

layer of coarse material (i.e. gravel with particle sizes ranging from 21 to 50 mm), on
the hydraulic conductivity of an intact adhesive-bonded GCL. Bentonite migration was
observed as a result of the stress concentration, but it was concluded that the effect on
hydraulic conductivity was insignificant even with a high confining stress (300 kPa).
Case histories have been reported where the field performance of the GCLs was de-
fective. Based on the analysis of a failure that occurred in Italy, Mazzieri and Pasqualini
(1997) suggested that puncturing by plant roots might have negatively impacted the
permeability of an adhesive-bonded GCL. James et al. (1997) described a similar fail-
ure that occurred in the UK for a similar GCL, but excluded root damage as being re-
sponsible for the leakage through the GCL. Instead, James et al. (1997) attributed the
failure to extensive ion exchange in the sodium bentonite, as a result of a calcium-en-
riched environment.

3 MATERIALS AND TEST METHODS

3.1 GCL

A geotextile-encased, adhesive-bonded GCL was selected for the experimental pro-

gram. The GCL is specified to contain a minimum of 5.0 kg/m2 of sodium bentonite
sandwiched between a polypropylene woven slit-film geotextile (150 g/m2) and a poly-
ester lightweight open-weave geotextile (40 g/m2). Free swell tests (ASTM D 5890) on
bentonite taken from the GCL typically gave values in excess of 25 mL.
The manufacturer states that joints can be constructed simply by overlapping the ad-
jacent panels: the hydrated bentonite swells out from the openings of the lightweight
geotextile, thus forming a seal in the overlapping area.

3.2 Preparation of Damaged GCL Specimens

Unused GCL specimens were tested in the present investigation. Circular speci-
mens of 10 cm in diameter were accurately cut from a GCL roll and artificially dama-
ged. Two basic patterns of damage were simulated: (i) damage resulting in bentonite
loss from the geotextile casing, i.e. tearing; and (ii) damage not resulting in bentonite
loss, i.e. puncturing.
The first pattern of damage was intended to simulate tears or a localised disturbance
of the GCL, resulting in loss or displacement of bentonite particles from between the
geotextile layers. This type of damage is believed to be more likely during storage, han-
dling, or installation, when the GCL is not yet hydrated. Bentonite loss was simulated
by cutting circular portions from the centre of unhydrated GCL specimens (Figure 1).
Both geotextiles and bentonite were cut and removed creating a hole in the specimen.
The adhesive used to bond the GCL components allowed the bentonite core to be cut
with a minimum amount of additional bentonite loss.
The second pattern of damage was intended to simulate punctures caused by roots
or sharp stones in the subgrade, or an inadequately selected cover layer. Evidence from
exhumed GCL samples showed that roots can fully puncture the geotextiles (Figure 2a)
and grow in the bentonite core, or puncture through the entire thickness of the GCL.

GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2 103

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

Removed portion

Figure 1. Damaged GCL specimen used to simulate tearing and bentonite loss (removal
of a central 3 cm-diameter portion from a 10 cm-diameter specimen).

To simulate the damage caused by plant roots, the GCL was punctured with metal ob-
jects such that the objects protruded from both sides (Figure 2b). Care was taken to pre-
vent any loss of bentonite during specimen preparation. To simulate different levels of
damage, a pin-size wire (0.1 cm in diameter) and a thicker nail (0.7 cm in diameter)
were used to puncture the specimens. Unhydrated and hydrated specimens were tested
to simulate damage that occurs both during and after installation.

3.3 Permeability Tests

Permeability tests on damaged and undamaged specimens were carried out in flex-
ible wall permeameters. Due to the specific nature of the tests, no existing standard pro-
cedure for permeability testing was strictly applicable; hence, a detailed description of
the adopted test procedure follows.
Specimens damaged with a circular hole (i.e. torn with resulting bentonite loss) were
placed in the permeameter between filter paper and porous disks. This experimental set-
up prevented the migration of bentonite from the damaged area and simulated damage
in the bentonite layer but not in the geotextiles, i.e. a damaged GCL underlain by a fine
gravel layer preventing migration of bentonite particles.
Specimens punctured with the pin-like wire (damage without bentonite loss) were
placed between two porous plates, which accommodated the protruding ends of the
wire without deforming them. To install the nail-punctured specimens in the permea-
meter, it was necessary to place two additional layers of fine gravel between the porous
disks and the specimen to contain the protruding ends of the nail.
Permeability tests on unhydrated specimens (damaged by a hole or punctured) were
started by applying the required confining isotropic pressure; after a few minutes the
unhydrated specimen stopped compressing, then the permeation was started immedi-
ately. A head loss of 50 cm was applied across the specimen, usually from the bottom

104 GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

(a)

Exhumed GCL

(b)

GCL

Figure 2. Punctured GCLs without bentonite loss: (a) exhumed GCL punctured by roots;
(b) simulated puncturing of a 10 cm-diameter GCL specimen.

upward to facilitate the expulsion of air bubbles. During this phase, the permeability
of the unhydrated bentonite was very high and the damaged area acted as an orifice;
thus, a high flow rate was observed. Afterward, the flow rate was typically slowed down
by bentonite hydration and swelling. Typically, it was necessary to increase the hydrau-
lic head loss up to 100 cm to obtain reliable permeability readings in a reasonable time
interval. The test was terminated when a steady-state unit flow rate through the speci-
men was achieved.
Specimens to be damaged when hydrated were first permeated intact under the re-
quired isotropic effective stress, and a reference hydraulic conductivity was measured.
To damage the specimen, the test apparatus was momentarily disassembled and then

GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2 105

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

immediately reassembled. The confining stress was restored and the specimen was
again permeated until a new steady permeability value was achieved. Additional swell-
ing of the bentonite may have occurred during unloading; however, considering that the
drainage lines were closed during unloading and that the time necessary to disassemble
and reassemble the apparatus was typically less than half an hour, the additional swell
of bentonite was considered negligible.
Tap water (pH ≈ 7.0 to 7.5; electrical conductivity ≈ 0.5 to 0.6 mS/cm; major dis-
solved ions: Ca2+ (calcium) = 92.5 mg/L, Na+ (sodium) = 16.5 mg/L, SO42- (sulphate)
= 84.3 mg/L, Cl - (chlorine) = 19.7 mg/L) was used as the permeant in all of the tests
to obtain more realistic results regarding hydraulic conductivity (Daniel 1994).

4 RESULTS: DAMAGED GCLs WITH BENTONITE LOSS

4.1 Influence of the Extent of GCL Damage

To investigate the limits of the self-healing potential of unhydrated GCLs, a test se-
ries was performed on specimens damaged by the removal of 1, 1.5, 2.0, 2.5, and 3.0
cm-diameter circular portions. A confining stress of 50 kPa was applied. The specimens
were permeated immediately after the application of the confining stress. During the
permeameter tests, the fluid flow through the specimen occurred mainly through the
damaged area and through the macropores between the unhydrated bentonite particles.
Therefore, the resistance to fluid flow of the porous medium was very low, the flow ve-
locity was relatively high, and the fluid flow was definitely not laminar. Moreover, the
specimen thickness varied with time due to bentonite hydration and swelling. Conse-
quently, the interpretation of permeability data in terms of Darcy’s hydraulic conductiv-
ity may be questionable for this test phase.
The flow rates often substantially decreased during hydration, reaching steady val-
ues comparable to those observed for intact specimens; during this latter steady-flow
phase, it was assumed that the flow occurred through the whole specimen and became
laminar. However, it was impossible to differentiate the exact transition between the
two phases during the tests. Hence, the permeability results were interpreted using
Darcy’s law from the beginning to the end of the tests, but in terms of permittivity, ψ,
as follows:

V
Ψ= (1)
A∆h ∆t

where: V = permeated volume; A = cross-section area of the specimen; ∆h = applied

head loss; and ∆t = elapsed time. If the thickness of the specimen, tGCL , is known (end
of the test), then:

k = Ψ t GCL (2)

where: k = hydraulic conductivity.

Figure 3 illustrates the typical results of a permeability/self-healing test on a dam-
aged specimen. The permittivity was calculated with reference to the water volume

106 GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

flowing out of the specimen, which initially differs from the water volume flowing in,
due to water absorption by the unhydrated bentonite. The permittivity rapidly decreases
with time when the damaged specimen is continuously permeated. In this case, the hy-
draulic performance of the damaged GCL, expressed as permittivity, becomes compa-
rable to that obtained for an intact specimen with an identical confining stress.
The final hydraulic conductivity values (calculated using the final thickness of the
specimens that was corrected for unloading rebound) are plotted versus the hole diame-
ter in Figure 4. It can be noted that for a damaged 10 cm-diameter specimen, the hydrau-
lic conductivity is not affected by hole diameters of up to 1.5 cm. The hydraulic
conductivity is approximately one order of magnitude higher than that of the intact GCL
specimen for the largest diameter hole tested.
Post-test inspections of the healed specimens revealed that the initial hole was com-
pletely filled by the swollen bentonite, i.e. an apparently continuous bentonite layer of
uniform thickness had reformed. For some tests, the swollen bentonite from the dam-
aged specimen area was sampled for water content measurements using a circular cut-
ting shoe. The water content of the self-healed area was found to be much higher
(typically double) than the surrounding area; therefore, the properties of bentonite var-
ied across the specimen. In particular, the permeability was higher in the self-healed
area, due to the higher void ratio of bentonite. Hence, assuming that the water head is
constant across the specimen, the flow rate is higher in the self-healed area than the sur-
rounding area. Consequently, the hydraulic conductivity values, which were calculated
using the total cross-section area of the specimen, depend on the relative size of the in-
tact and removed portions of the specimen.
To verify the above statement, a permeability test was performed on a 7 cm-diameter
GCL specimen that was damaged with a 2 cm-diameter hole. The calculated final hy-

10-4
Damaged with a 2 cm-diameter hole
(effective stress = 50 kPa)
10-5 Undamaged
(effective stress = 50 kPa)
Permittivity, ψ (s-1)

10-6

10-7

10-8

10-9
0.1 1.0 10.0 100.0 1000.0
Permeation time (hours)
Figure 3. Permittivity of damaged and undamaged GCL specimens under an effective
stress of 50 kPa.

GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2 107

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

10-7
Undamaged GCL; specimen diameter = 10 cm

Hydraulic conductivity, k (cm/s)

(effective stress = 50 kPa)
Self-healed GCL; specimen diameter = 10 cm
(effective stress = 50 kPa)
Self-healed GCL; specimen diameter = 7 cm
(effective stress = 50 kPa)

10-8

10-9
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Hole diameter (cm)

Figure 4. Influence of hole diameter on the final hydraulic conductivity of self-healed

GCL specimens.

draulic conductivity of this specimen is higher than the value calculated for the larger,
10 cm-diameter specimen (Figure 4); however, visual observation of the specimen at
the end of the test confirmed that self-healing had still occurred.
The calculated hydraulic conductivity is considered to be a result of the testing
conditions. Had the same test been performed on a specimen with a diameter larger than
10 cm, the final calculated hydraulic conductivity would have been an intermediate va-
lue between that of the intact material and that obtained with a specimen diameter of
10 cm. This implies that, provided self-healing is achieved and the defect is completely
sealed, the overall hydraulic performance of an actual GCL barrier is not significantly
influenced by a defect. On the other hand, the results suggest that self-healing involves
the swelling of the bentonite immediately adjacent to the damaged area and does not
affect the rest of the specimen. Consequently, there is an upper limit value for the hole
diameter, whereby self-healing cannot occur however large the specimen diameter.

4.2 Influence of Continuity of Flow and Underlying Material

For the adopted testing conditions, the damaged GCLs were continuously per-
meated for the duration of the test. To assess the role played by the continuity of hydrau-
lic flow in the self-healing mechanism, damaged specimens were tested allowing the
bentonite to hydrate without any applied hydraulic gradient for at least 48 hours, at an
effective stress of 50 kPa. Then, the specimens were temporarily permeated only for
the time necessary to obtain a reliable permeability measurement. The results show that
the response of damaged specimens is much slower, in the absence of continuous flow
(Figure 5). This can only be attributed to the influence of hydraulic flow on the damaged

108 GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

area, considering that a sharp reduction of permeability occurs as the continuous hy-
draulic flow is restarted.
The results suggest that the presence of flow through the damaged specimen en-
hanced the self-healing process for the current experimental conditions. One possible
interpretation of these findings is that the initially high flow velocity through the dam-
aged area promotes both bentonite hydration and the migration of hydrated bentonite
particles from the adjacent portion of the specimen. The bentonite particles accumulate
at the effluent end and rapidly form a seal by freely swelling. However, it should be
noted that this mechanism is possible because further bentonite migration from the
GCL “sandwich” is prevented by a physical barrier (filter paper and porous stones in
the current experimental program).
To validate this hypothesis, tests were replicated where the filter paper and the po-
rous stones were replaced by a layer of gravel. Preliminary hydration was allowed for
48 hours. The first permeability measurements result in high hydraulic conductivity
values, similar to that observed in specimens underlain by porous filters (Figure 5).
However, when continuous permeation starts, the permittivity does not show any sig-
nificant reduction, but, instead, an increase. Specimen inspection confirmed that ben-
tonite migration into the void spaces of the gravel had occurred. Thus, the damage was
not repaired because the bentonite was dispersed into the gravel layer.
Although damage in the form of geotextile tearing and removal appears to be unlike-
ly, care should be exercised when glued GCLs are placed on soil layers with coarse par-
ticles. In fact, in addition to the increased risk of GCL puncturing by sharp particles
during installation, it must be considered that the self-healing capacity may be compro-
mised in the case of severe damage of the geotextile layers.

10-4
Continuous permeation

10-5 Test with gravel

(effective stress = 50 kPa)
Permittivity, ψ (s-1)

Test with porous filters

10-6 (effective stress = 50 kPa)
Hydration and
temporary
10-7 permeation

10-8

10-9
0 48 96 144 192 240 288 336
Time (hours)
Figure 5. Influence of the continuity of flow and the underlying material on the self-heal
capacity of GCL specimens.

GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2 109

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

10-8

Hydraulic conductivity, k (cm/s)

10-9

10-10 Damaged GCLs

Calculated using Equation 3
Undamaged GCLs
Fitted curve for
undamaged GCLs
10-11
10 100 1000
Effective stress (kPa)
Figure 6. Influence of the effective stress on the final hydraulic conductivity of damaged
and undamaged GCL specimens.

4.3 Influence of Confining Stress

The hydraulic conductivity of intact GCLs decreases with increased confining stress
during hydration (Estornell and Daniel 1992; Petrov et al. 1997). The confining stress
prevents bentonite particles from freely hydrating, which results in a lower swelling
volume, but denser state.
As shown in Section 4.1, the self-healing mechanism is governed by swelling of the
bentonite in the damaged area. The applied confining stress should reduce the volume
of the removed portion by compressing the adjacent unhydrated bentonite. However,
the confining stress also significantly reduces the swelling volume; therefore, the effect
of increased confining stress on the self-healing mechanism is not easily predictable.
A series of self-healing tests was carried out at confining stresses ranging between
15 and 450 kPa on specimens with a 2.0 cm-diameter hole. The lower stress range (15
to 50 kPa) is a typical condition for landfill cover liners. The higher range (200 to 450
kPa) might be representative of a municipal solid waste (MSW) landfill bottom liner,
where the GCL can be coupled with a geomembrane to form a composite barrier. It is
expected that the geomembrane will not leak for the early operational period of the fa-
cility, so that the GCL may remain substantially unhydrated while the landfill is filled;
thus, the confining stress acting on the GCL will gradually increase. In the present study,
the possible chemical content of the leachate is not considered.
Figure 6 shows the final hydraulic conductivity of damaged specimens versus the
confining stress. The hydraulic conductivity values were calculated using the final un-

110 GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

loaded thickness of the specimens, measured at the end of the test and corrected for
rebound due to unloading, which was estimated based on swelling tests performed in
an oedometer using the same vertical stresses (for details refer to Mazzieri (1998)). Fig-
ure 6 also illustrates the hydraulic conductivity of undamaged specimens versus the
confining stress.
Although a difference can be noted between damaged and undamaged specimens,
the hydraulic conductivity values for the damaged GCLs are very low, suggesting com-
plete self-healing. Despite the scatter of the data, it is possible to identify a decreasing
trend up to a confining stress of 200 kPa for the damaged GCLs; for higher stresses, the
hydraulic conductivity of damaged specimens remains generally constant.
To further explain these results, it can be assumed that the self-healed specimen is
composed of two homogeneous sections of different permeability connected in parallel.
Hence, the overall hydraulic conductivity is given by:
k sh A sh + k i A i
k av = (3)
A

where: kav = average hydraulic conductivity of the specimen; ksh = hydraulic conductiv-
ity of the self-healed portion of the specimen; Ash = cross-section area of the self-healed
portion of the specimen (= 3.2 cm2); ki = hydraulic conductivity of the intact portion
of the damaged specimen (cm/s); Ai = area of the intact portion of the specimen (= 78
cm2); A = total cross-section area of the specimen (= 81 cm2).
To estimate the values of ksh and ki , the self-healed portion of the specimen was
sampled and the average void ratio values of the bentonite in the self-healed and the
intact portion were evaluated. It was not considered correct to assume that the ki values
of the undamaged specimens at the same effective stress would be the same for dam-
aged specimens because of the different hydration conditions (i.e. because of lateral
migration of bentonite in the damaged area), resulting in an average void ratio higher
than in undamaged specimens at the same effective stress.
Using the test results for the undamaged GCL at different confining stresses, an em-
pirical relationship between the hydraulic conductivity, k, expressed in cm/s, and the
void ratio, e, of the hydrated bentonite was developed (Mazzieri 1998) for the GCL
tested in the present study:

k = 4.9 × 10−10 1 +
e 
e
3
(4)

At the end of the tests, the bentonite in the self-healed portion was sampled using
a cutting shoe. Water content measurements were used to estimate the void ratio of the
bentonite in self-healed portion, esh , and the initially intact portion, ei .
Although Equation 4 was obtained for void ratios ranging from 0.50 to 3.43, it was
extrapolated to higher void ratios to estimate the hydraulic conductivity of both the self-
healed portion, ksh , and the initially intact portion, ki , by substituting esh and ei , respec-
tively, into Equation 4 .
In Figure 6, the hydraulic conductivity values, kav , calculated according to Equation
3 are compared to the final hydraulic conductivity values. Although the assumption of
a uniform void ratio in the two specimen portions is very simplified, there is a good
agreement between the calculated and measured values. In the low stress range, the

GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2 111

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

term ksh Ash is negligible with respect to ki Ai , which is decreasing due to increasing ef-
fective stresses during hydration (i.e. decreasing ki ); in the high stress range, the hy-
draulic conductivity of the initially intact portion becomes so low that the term ksh Ash
is predominant.

5 RESULTS: DAMAGED GCLs WITHOUT BENTONITE LOSS

The effect of punctures on the performance of a GCL was examined. As mentioned

in Section 3.2, care was taken to prevent any bentonite loss and additional tearing of
the geotextiles while puncturing the GCLs.
In this damaged condition, the permeability of the GCL can be impacted either if
the puncturing body has inherent permeability (e.g. rotten plant roots or wood) or if a
preferential seepage path is created at the interface between the extraneous object and
the bentonite core. Only the second situation is examined in the present study because
metal objects were used to puncture the GCL.

5.1 GCL Damage During Unhydrated Conditions

Tests were carried out at confining stresses of 15, 25, 50, and 300 kPa for specimens
damaged with a pin-like wire and at 50 kPa for a specimen damaged with the thick nail
(Section 3).
The permeability test results of unhydrated punctured specimens show initial out-
flow rates higher than that for intact specimens tested at an identical effective stress and
applied head loss (Figure 7), particularly at lower confining stresses. However, the
permeability decreases rapidly reaching values very close to those of intact specimens.
Therefore, the overall permeability of the GCL appears to be affected by damage only
during the initial phase, probably due to the time required for the hydrating bentonite
to fill the small gap created by the puncture. Although great care was taken to avoid
disturbance of the bentonite layer, it must be considered that an amount of displacement
occurred while puncturing the GCL with the pin-like wire. These considerations are not
possible for the nail-damaged specimen because the measured outflow rates were sig-
nificantly affected by air bubbles initially contained in the gravel layer.
In Figure 8, the final hydraulic conductivity values for damaged specimens are
compared with those of undamaged specimens. The hydraulic conductivity could not
be accurately evaluated for specimens tested between gravel layers due to the uneven
final thickness of the GCL; hence, the results are presented in terms of permittivity. The
slight differences between the calculated permeability values of intact and damaged
specimens are not significant considering the approximations made for the calcula-
tions, i.e. the estimation of the specimen thickness under a load.
The test results suggest that the final GCL performance is not affected by the simu-
lated damage pattern, thus, extending the results reported by Bouazza et al. (1996), who
tested an unhydrated GCL that was punctured only one half of the thickness. The results
obtained also show that complete puncturing is typically sealed by self-healing: hy-
drated bentonite adheres to the puncturing body irrespective of the applied confining
stress. However, the applied confining stress influences the duration of the initial phase
of relatively high permeability that, for the simulated damage, seems to be reduced by

112 GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

25 Punctured with 1 mm wire

(effective stress = 50 kPa)
Intact
20 (effective stress = 50 kPa)
Outflow volumes (mL)

Head loss increased

from 50 to 100 cm
15

10

0
0 24 48 72 96 120 144 168 192 216 240 264 288 312 336
Permeation time (hours)
Figure 7. Difference between the initial outflow rates for intact and punctured GCL
specimens.

Undamaged
(hydraulic conductivity)
Damaged, 1 mm ---diameter wire
(hydraulic conductivity)
10-8
Damaged, 7 mm ---diameter nail
Hydraulic conductivity, k (cm/s)

(permittivity)
Permittivity, ψ (s-1)

10-9

10-10

10-11
10 100 1000
Effective stress (kPa)

Figure 8. Permeability of undamaged and punctured (when unhydrated) GCL

specimens as a function of effective stress.

GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2 113

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

the confining stress. The confining stress presumably squeezes the unhydrated benton-
ite against the surface of the puncturing body, thus enhancing the formation of a seal.
The presence of coarse soil underneath the GCL does not appear to play any role in the
sealing mechanism because the same results were obtained using both the porous filters
and the gravel. However, as shown inSection 4.2, the resultscould be different ifpunctur-
ing causes significant geotextile damage and the GCL is installed on a coarse subgrade.

5.2 GCL Damage During Hydrated Conditions

The primary purpose of this phase of the test program was to simulate the effect of
the growth of roots through the bentonite layer of an installed GCL. It should be consid-
ered that roots can grow only when the GCL is placed under relatively thin cover layers;
consequently, relatively low effective stresses were selected (10, 25, and 50 kPa). It was
also assumed that the GCL remained hydrated.
As explained in Section 3.2, the specimens were hydrated under the selected stress
and then temporarily unloaded before damaging. The GCL was punctured by slowly
forcing the metal object through the geotextile and bentonite layers, preventing any
removal of bentonite or tearing of the geotextile, in an effort to reproduce puncturing
by roots.
After the restoration of hydration stresses on the damaged specimens, little consoli-
dation was observed, confirming that additional swelling due to unloading was negli-
gible. In Figure 9, the final hydraulic conductivity values of damaged and intact
specimens are plotted in relation to the confining stresses. Upon permeation, the perme-
ability of the damaged specimens does not significantly change compared to the un-
damaged phase. These results suggest that, at least under the applied laboratory
conditions, puncturing does not significantly influence the permeability of GCLs.

5.3 GCL Damage Coupled with Ion Exchange

Mazzieri and Pasqualini (1997) observed a dramatic increase in permeability for ex-
humed GCL specimens punctured by roots, when compared with undamaged speci-
mens of the same exhumed material. Subsequent evidence (Mazzieri 1998) revealled
extensive ion exchange in the exhumed GCL due to the replacement of sodium by cal-
cium ions, presumably resulting from leaching of the calcareous gravel used as cover
soil. Within the framework of the experimental program, it was deemed interesting to
investigate the behaviour of a GCL damaged after having undergone ion exchange.
To induce an extensive exchange of sodium ions (Na+) for calcium ions (Ca2+) with-
in a reasonable time period, an intact sodium-bentonite GCL specimen was hydrated
for four days in a 0.5 N solution of calcium chloride (CaCl2), at an effective stress of
50 kPa. The specimen was then permeated by distilled water, to reduce the ionic
strength of the pore fluid. As expected, the permittivity is initially high (Figure 10), but
gradually decreases reaching a value of approximately 8 × 10-9 s-1. Figure 10 also
shows that the permittivity of an identical undamaged specimen hydrated directly in
distilled water is approximately 2 × 10-9 s-1. This difference was expected, and it is
in accordance with permeability values usually described for calcium and sodium ben-
tonite. The ion exchange was confirmed at the end of the test by measuring free swell
(ASTM D 5890) of bentonite taken from the specimen, which gave a value of 5 mL,

114 GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

10-8
Undamaged
(hydraulic conductivity)
Hydraulic conductivity, k (cm/s) Damaged, 1 mm ---diameter wire
(hydraulic conductivity)
Permittivity, ψ (s-1)
Damaged, 7 mm ---diameter nail
(permittivity)

10-9
10 100
Effective stress (kPa)
Figure 9. Permeability of undamaged and punctured (when hydrated) GCL specimens,
as a function of effective stress.

10-4
Sodium bentonite;
undamaged
10-5 Sodium bentonite;
damaged
Permittivity, ψ (s-1)

Calcium bentonite;
10-6 undamaged
Calcium bentonite;
damaged
10-7

10-8

10-9
0 5 10 15 20 25 30 35 40 45
Permeation time (hours)
Figure 10. Permittivity of intact and damaged sodium- and calcium-bentonite GCL
specimens.

GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2 115

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

whereas the sodium bentonite taken from the unused material typically gives values in
excess of 25 mL.
After a steady permeability value was attained, the test was temporarily stopped to
damage the specimen by puncturing it with a 0.7 cm-thick nail. When permeation was
restarted, a dramatic increase in permeability initially occurs (Figure 10). Afterward,
the permeability shows a significant response as a result of continuous permeation. This
is believed to depend on the time necessary for additional swelling of bentonite at the
bentonite-nail interface; however, the seal formation is not complete because the new
steady-state permittivity is approximately one order of magnitude higher than the initial
value. Figure 10 also shows that the permittivity of the identically damaged GCL speci-
men without ion exchange does not change.
Although the authors of the present study admit that conditions simulated in the lab-
oratory cannot fully reproduce the process of root growth in the field, these results seem
to indicate that puncturing may represent a serious concern when coupled with ion ex-
change. Moreover, according to Benson (1999), ion exchange also reduces the resist-
ance of a GCL to dry-wet cycles. Hence, based on the available results, it appears
reasonable to state that puncturing by roots may accelerate the failure of the GCL under
ion exchange conditions.

6 CONCLUSIONS

The permeability of damaged, commercially available GCL specimens was investi-

gated. The authors of the present paper emphasise that the investigation was performed
on unused specimens and, therefore, it was assumed that the swelling capacity of ben-
tonite was completely effective in the self-healing process (with the only exception of
induced ion exchange). Moreover, the results refer to one type of GCL (adhesive-
bonded) and may not be valid for other types of GCLs (e.g. needle-punched).
For GCL damage involving tearing of the geotextiles and loss of bentonite, the GCL
typically has a very good self-healing capacity. Even though the bentonite in the dam-
aged area is significantly swollen (i.e. it has a high void ratio), this area of relatively
high permeability has a negligible influence on the performance of the GCL, provided
a complete seal is achieved.
The self-healing of holes is not reliable if the geotextiles are damaged and the GCL
is placed in contact with a coarse subgrade layer allowing the bentonite to migrate from
the geotextile-bentonite-geotextile “sandwich”. Special attention must be paid during
the installation of GCLs on coarse subgrade layers.
With respect to damage caused by puncturing, the results are encouraging because
the hydraulic performance was found to be virtually insensitive to punctures, provided
that the composition of the bentonite remains as sodium bentonite. Ion exchange, caus-
ing a reduction of swelling properties (e.g. the exchange of calcium for sodium ions),
drastically reduces the capacity of the bentonite to seal punctures. This conclusion can
be reasonably extended to the penetration of roots and pipes, leachate removal wells
in landfill liners, and other situations where a hydraulic seal at the interface with a dif-
ferent material is required.

116 GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

REFERENCES

ASTM D 5890, “Standard Test Method for Swell Index of the Clay Mineral Component
of Geosynthetic Clay Liners”, American Society for Testing and Materials, West
Conshohocken, Pennsylvania, USA.
Benson, C.H., 1999, “Final Covers for Waste Containment”, Control and Management
of Pollutants, Proceedings of the XVII Geotechnical Conference of Turin, Italy, Vol.
11, pp. 1-35.
Boardman, B.T. and Daniel, D.E., 1996, “Hydraulic Conductivity of Desiccated Geo-
synthetic Clay Liners”, Journal of Geotechnical Engineering, Vol. 122, No. 3, pp.
204-209.
Bouazza, A., Van Impe, W.F. and Van Den Broeck, M., 1996, “Hydraulic Conductivity
of a Geosynthetic Clay Liner Under Various Conditions”, Environmental Geotec-
hnics, Kamon, M., Editor, Balkema, Vol. 1, Proceedings of the Second International
Congress on Environmental Geotechnics, Osaka, Japan, November 1996, pp.
453-458.
Daniel, D.E., 1994, “State of the Art: Laboratory Hydraulic Conductivity Tests for Sat-
urated Soils”, Hydraulic Conductivity and Waste Contaminant Transport In Soil,
ASTM Standard Technical Publication No. 1142, Daniel, D.E. and Trautwein, S.J.,
Editors, proceedings of a symposium held in Philadelphia, Pennsylvania, USA, pp.
30-78.
Didier, G. and Norotte, V., 1998, “Mise en oeuvre des geosynthetiques béntonitiques”,
Proceedings of GeoBento 98, INSA-Lyon, Paris, France, February 1998, pp. 45-63.
(in French)
Estornell, P.M. and Daniel, D.E., 1992, “Hydraulic Conductivity of Three Geosynthetic
Clay Liners”, Journal of Geotechnical Engineering, Vol. 118, No. 10, pp. 1592-1605.
Fox, P.J., 1998, “Research on GCLs at Purdue University”, Geotechnical News, March
1998, pp. 35-40.
Hewitt, R.D. and Daniel, D.E., 1997, “Hydraulic Conductivity of Geosynthetic Clay
Liners After Freeze-Thaw”, Journal of Geotechnical and Geonvironmental Engi-
neering, Vol. 123, No. 4, pp. 305-313.
Heyer, D., 1995, “Basic Examination on the Efficiency of GCLs”, Geosynthetic Clay
Liners, Koerner, R.M., Gartung, E. and Zanzinger, H., Editors, Balkema, 1995, pro-
ceedings of an international symposium held in Nurnberg, Germany, April 1994, pp.
101-111.
James, A.N., Fullerton, D. and Drake, R., 1997, “Field Performance of GCL Under Ion
Exchange Conditions”, Journal of Geotechnical and Geoenvironmental Engineer-
ing, Vol. 123, No. 4, pp. 897-901.
Kraus, J.F., Benson, C.H., Erikson, A.E. and Chamberlain, E.J., 1997, “Freeze-Thaw
Cycling and Hydraulic Conductivity of Bentonitic Barriers”, Journal of Geotechni-
cal Engineering and Geoenvironmental Engineering, Vol. 123, No. 5, pp. 229-238.
La Gatta, M., Boardman, B.T., Cooley, B.D. and Daniel, D.E., 1997, “Geosynthetic
Clay Liners Subjected to Differential Settlement”, Journal of Geotechnical and
Geoenvironmental Engineering, Vol. 123, No. 5, pp. 254-262

GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2 117

 MAZZIERI AND PASQUALINI D Permeability of Damaged GCLs

Mazzieri, F., 1998, “Comportamento Idraulico dei Geocompositi Bentontici”, Tesi di

Dottorato di Ricerca, Politecnico di Torino, Italia, 190 p. (in Italian)
Mazzieri, F. and Pasqualini, E., 1997, “Field Performance of GCL: A Case Study”, En-
vironmental Geotechnics, Bouazza, A., Kodikara, J. and Parker, R., Editors, Balke-
ma, Vol. 1, Proceedings of the 1st Australia-New Zealand Conference on
Environmental Geotechnics: GeoEnvironment 97, Melbourne, Australia, November
1997, pp. 289-294.
Petrov, R.J., Rowe, R.K. and Quigley, R.M., 1997, “Selected Factors Influencing GCL
Hydraulic Conductivity”, Journal of Geotechnical and Geoenvironmental Engineer-
ing, Vol. 123, No. 8, pp. 683-695.
Shan, H.Y., 1990, “Laboratory Tests on Bentonitic Blankets”, M.S. Thesis, University
of Texas, Austin, Texas, 84 p.

NOTATIONS

Basic SI units are given in parentheses.

A = cross-section area of specimen (m2)

Ai = cross-section area of intact portion of specimen (m2)
Ash = cross-section area of self-healed portion of specimen (m2)
e = void ratio of bentonite (dimensionless)
esh = void ratio of bentonite in self-healed portion of specimen (dimensionless)
ei = void ratio of bentonite in initially intact portion of specimen
(dimensionless)
k = hydraulic conductivity (m/s)
kav = average hydraulic conductivity (m/s)
ki = hydraulic conductivity of intact portion of damaged specimen (m/s)
ksh = hydraulic conductivity of self-healed portion of specimen (m/s)
tGCL = thickness of GCL specimen (m)
V = permeated volume (m3)
∆h = applied head loss (m)
∆t = elapsed time (s)
ψ = permittivity (s-1)

118 GEOSYNTHETICS INTERNATIONAL S 2000, VOL. 7, NO. 2

View publication stats