The Positive Effects of Silica Fume on the Permeability, Swelling Pressure and Compressive Strength of Natural Clay Liners

www.elsevier.com/locate/enggeo

Engineering Geology 73 (2004) 145–156

The positive effects of silica fume on the permeability, swelling

pressure and compressive strength of natural clay liners

Ekrem Kalkana, Suat Akbulutb,*

aEarthquake Research Center, Ataturk University, 25240 Erzurum, TurkeybDepartment of Civil Engineering, Ataturk Universiy, 25240 Erzurum, Turkey

Received 24 January 2003; accepted 16 January 2004

Abstract

Clays are essential materials to reduce the hydraulic conductivity of natural clay liners in landfill sites. Impermeable

compacted clay liners are needed the landfill to be designed. It is known that clays with high plasticity absorb water several

times as much as their weights. The clay liners subjected to water pressure in landfills generate high permeability in time and

instability problems in their body due to their expansive capacity. Though the compacted clay liners possess many advantages

such as low permeability and large capacity of attenuation, they have high shrinkage and high expansive potential causing

instability problem. The aim of this study is to examine the effects of silica fume on permeability, swelling pressure and

compressive strength of the compacted clay liners as a hydraulic barrier. The test results showed that the compacted clay

samples with silica fume exhibit quite low permeability, swelling pressure and significantly high compressive strength as

compared to raw clay samples. Thus, silica fume appears to be promising for construction material of liners subjected to

leachate in solid waste containment systems.

D 2004 Elsevier B.V. All rights reserved.

Keywords: Permeability; Compressive strength; Swelling pressure; Clay liner; Silica fume; Landfill

1. Introduction utilize natural materials such as compacted clay or

Compacted clay liners are widely used in solid

waste landfills due to their cost effectiveness and large

capacity of attenuation. Traditionally, clay barriers for

the containment of landfill leachate are made up of

compacted clay liners. In the absence of impermeable

natural soils, compacted mixtures of bentonite and

sand have been used to form barriers to fluids (Kennay

et al., 1992; Daniel and Wu, 1993). The liner systems

0013-7952/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.enggeo.2004.01.001

* Corresponding author. Fax: +90-442-236-0957.

E-mail address: [email protected] (S. Akbulut).

shale, bitumen, soil sealant synthetics and membranes

(Prashant et al., 2001). The main requirements of liners

are to ensure the minimization of pollutant migration,

low swelling and shrinkage and resistance to shearing

(Brandl, 1992; Kayabalı, 1997; Cazaux and Didier,

2000). These measures generally involve the applica-

tions of low permeable natural clays and sand–ben-

tonite mixtures or synthetic materials (Van Ree et al.,

1992). There are several criteria that may be used to

evaluate the performance of landfill liners with respect

to chemical migration (Katsumi et al., 2001; Bouazza,

2002; Shackelford et al., 2000). Membranes composed

of soil should have lower hydraulic conductivity and

Fig. 1. The grain-size distribution of clays and silica fume used in

the tests.

E. Kalkan, S. Akbulut / Engineering Geology 73 (2004) 145–156146

diffusion coefficient than those previously demanded,

and they have to look for more efficient soil types for

the future constructions (Foged and Baumann, 1999;

Patrick et al., 2000).

The thickness used for such liners varies from a

few decimeters to more than 1 m. Typically, the

hydraulic conductivity must also be less than or equal

to 1�10� 7 cm/s for soil liners and covers which are

used to contain hazardous waste, industrial waste and

municipal waste (Daniel and Benson, 1990; Das,

1997). The required liner thickness for domestic and

light industrial waste types is to be 0.6 m for com-

pacted clay liners (Bouazza and Van Impe, 1997).

Very little data is available to prove the correlation

between the internal compressive shear of clay CL and

silica fume SF additive material. Therefore, we exam-

ined the suitability of silica fume, which is an abun-

dantly available product, for the construction of

hydraulic barrier in landfill. We have experienced on

a membrane composed of the clay and silica fume with

different contents. The performance of silica fume–

clay composite liners with the optimum moisture

content was evaluated to develop as an alternative

liner material with low permeability, low swelling

pressure and high compressive strength for landfills.

Data for consistency limits, dry unit weight, perme-

ability, swelling pressure and compressive strength

have been obtained from index tests, compaction,

permeability, odometer and unconfined compression

tests under laboratory conditions. From a logical

standpoint, it was found out that a higher bonding

strength between clay and silica fume particles led to

higher internal shear strength in the clay–silica fume

mixture. Therefore, in order to evaluate the interaction

between silica fume and silt or clay particles in

composed samples, the samples were magnified 5000

times by means of a scanning electron microscope

(SEM). Observations of scanning electron microscope

showed that the structure of a material had a profound

influence on its engineering properties such as perme-

ability, strength and stiffness due to silica fume content

in the sample. The chemical analysis, which has

positive effects to improve the properties of composite

samples, also indicated that it could be a chemical

reaction between silica fume and clay particles.

The objective of this paper is to present a novel

liner material compacted of raw clay and silica fume

and to document the structure of clay samples stabi-

lized with different silica fume contents. Clay and

silica fume serve yielding low permeability, low

swelling pressure and high compressive strength.

Various ratios of silica fume to clay were tested to

obtain the most desirable mixture ratio of the ideal

liner material for waste containment systems, and

optimum results were obtained in composite samples

with 25% silica fume content.

2. Materials

2.1. Clays

The natural clay materials (samples 1, 2, 3 and 4)

originated from a clay pit in Oltu deposits in northeast

of Turkey. The clay deposits were discovered during a

researcher working (Akbulut, 1999). This initial in-

vestigation was later used as the base of investigation

for the present study. The physical properties of

natural clays used in the tests such as particle size,

Atterberg limits, maximum dry unit weight and opti-

mum water content were determined in accordance

with ASTM D 422, D 4138, D 1557 and D 2216,

respectively. The hydrometer tests were performed to

find the proportion of the clay and to plot the

granulometry curves of clays. The grain-size distribu-

tion curves and the index properties of natural clays

are given in Fig. 1 and Table 1, respectively. The

results of Atterberg tests show that clays (four differ-

Table 3

Leachate properties used in the tests

pH Conductivity, k (ms/cm) COD (mg/l)

7.71 20.2 6568.9

Table 1

Index properties of the clays used in the tests

Property Sample

1

Sample

2

Sample

3

Sample

4

Density, cs (kN/m3) 26.0 26.6 26.3 26.8

Liquid limit (%) 65 60 65 62

Plastic limit (%) 42 38 35 35

Plasticity index (%) 23 22 30 27

Clay content (%) 20 21 24 22

Activity, A 1.15 1.04 1.25 1.22

E. Kalkan, S. Akbulut / Engineering Geology 73 (2004) 145–156 147

ent clays) are high-plasticity clays as shown in Fig. 3.

The clays are classified as high-plasticity clays CH by

the Unified Soil Classification System. The samples

for the tests at 20 jC were prepared by using the

municipal water under laboratory conditions. The

chemical compositions of clay particles ( < 2-Am frac-

tions) were analyzed in Askale Cement Factory. The

results of the analysis are given in Table 2.

2.2. Silica fume

Silica fume obtained from a Ferro-Chromite Fac-

tory in Antalya was used in different proportions with

clay in the tests. The granulometry of the silica fume

is shown in Fig. 1. The density of silica fume is

c = 20–25 kN/m3, and its bulk density is cv= 3.0–5.0kN/m3. The chemical analysis of silica fume is given

in Table 2.

2.3. Leachate

In order to determine the effect of leachate on

permeability and swelling pressure of composite sam-

ples with 25% silica fume, leachate from municipal

Table 2

Chemical compositions of the clays and silica fume used in the tests

Property Sample

1

Sample

2

Sample

3

Sample

4

Silica

fume

Al2O3 17.82 13.63 13.94 13.24 1–3

CaO3_a 26.09 23.18 27.50 23.09 –

CaO3_b 24.95 17.81 29.79 21.26 –

CaO 9.55 7.98 11.02 8.26 0.5–1

Fe2O3 8.03 5.79 6.21 7.56 1–2

MgO 2.38 4.34 3.48 6.15 0.8–1.2

SiO3 0.15 0.54 0.12 0.44 –

SiO2 44.27 54.67 41.59 44.69 85–95

Loss on

ignition

10.10 11.00 12.45 13.19 –

landfill area in Erzurum city was used in the perme-

ability and odometer tests. The leachate properties,

such as contamination and heavy metal ions, are

summarized in Table 3 and 4, respectively. The

contamination level of leachate was found high,

6568 mg/l, according to Chemical Oxygen Test

(COD). Heavy-metal ion concentrations were deter-

mined using an atomic adsorption spectrophotometer

(Schimadzu). Various metal ion concentrations such

as Cu, Zn, Pb, Cd, Cr and Fe were determined from

the leachate as can be seen in Table 4.

3. Methods

3.1. Clay–silica fume mixture preparation

The clay soils were first dried in heating oven at

approximately 105 jC before using in the mixtures.

The total dry weight of the mixture required to

prepare a specimen, W, is known from the specimen’s

dimensions and dry unit weight, cd. W can be

expressed as, W=Wc +Ws, where Wc and Ws are the

weights of the dry clay soil and silica fume, respec-

tively. To prepare the clay–silica fume mixture, first,

the required amounts of clay and silica fume were

measured by a total dry weight, W, of sample and

mixed together in the dry state. As the silica fume

tended to lump together, it required considerable care

and time to mix them to get an even distribution of the

silica fume in the mixture. The dry clay–silica fume

mixture was then mixed with the required amount of

water that is explained as optimum moisture content.

All mixing was done manually, and proper care was

taken to prepare homogeneous mixtures at each stage

of mixing.

Table 4

The amounts of heavy metal in leachate

Cu Zn Pb Cd Cr Fe

0.046

mg/l

0.249

mg/l

0.077

mg/l

0.005

mg/l

1.279

mg/l

1.757

mg/l

eering Geology 73 (2004) 145–156

3.2. Atterberg tests (consistency limits)

In order to determine the liquid limit, plastic limit,

and plasticity index, various percentages of clay–

silica fume mixtures were carried out. Liquid limit

and plastic limit tests were conducted in accordance

with ASTM D 4318. The results for consistency limits

were obtained immediately from the tests.

3.3. Compaction tests

Standard proctor tests (ASTM D 698) have been

carried out to determine the compaction properties of

the samples in which clay and silica fume were mixed.

The necessary compacting was used to obtain a

homogeneous clay membrane. To ensure uniform

compaction, the entire required quantity of the moist

clay–silica fume mixture was placed inside the

mould-collars assembly and compressed in three steps

alternately from the two ends till the specimen

reached the dimensions of the mould. Fig. 4 shows

the light (standard proctor) compaction curves of the

raw clay and the silica fume mixtures. The results

showed that the raw clay could be compacted to a

significantly greater dry unit weight than the silica

fume–clay mixtures at relatively lower water content.

All the specimens for the compressive tests were

prepared at the optimum water content state of them.

3.4. Clay–silica fume specimen preparation

For unconfined compression tests, cylindrical speci-

mens were prepared at maximum dry unit weight and

optimum moisture content of the raw clay and silica

fume–clay mixtures. A 35-mm-inner diameter and 70-

mm-long mould was used to prepare the samples for

the standard proctor test. The compacted specimen was

extruded from the mould using a hydraulic jack. A

minimum of three specimens was prepared for each

combination of variables for compressive tests.

3.5. Unconfined compression tests

The compressive strength of compacted clay with

silica fume samples was determined from the uncon-

fined compression tests (ASTM 2166). The uncon-

fined compression test is widely used as a quick,

economical method of obtaining the approximate

E. Kalkan, S. Akbulut / Engin148

compressive strength of the cohesive soils. In this

laboratory work, three tube samples with a length/

diameter ratio 2 were prepared (L: 70 mm and D: 35

mm). Samples were placed in a moist container to

prevent from drying while waiting a turn at the

compression machine. At least three specimens were

tested for each combination of variables at a defor-

mation rate of 0.16 mm/min. The specimens were

prepared as explained before in the method of spec-

imen preparation. The silica fume inclusions had a

significant effect on the stress–strain behavior. It was

observed that the increase in the compressive strength

depended on the silica fume content in specimen. The

lower the compressive strength of the raw clay spec-

imen, the higher was the increase in compressive

strength due to the silica fume inclusion.

3.6. Permeability tests

In the falling-head permeability test, the soil spec-

imen was placed inside a tube, and allowed a flow

through the specimen. The initial head difference h1 at

time t = 0 was recorded, and water was allowed to

flow through the soil in order to obtain h2 in the final

head at time t= t. The hydraulic conductivity K was

calculated by the following equation:

K ¼ 2:303aL

Atlog

h1

h2ð1Þ

where K is the hydraulic conductivity, in centimeters

per second, h is the head difference, in centimeters, at

any time t, A is the area of specimen, in square

centimeters, a is the area of standpipe, in square

centimeters and L is the length of specimen, in

centimeters. The permeability apparatus had a plastic

mould 10 cm wide, 20 cm high and 2 mm thick. The

test apparatus consisted of a mould with lids and a

standpipe 10 mm in diameter and 100 cm high.

The permeability of clay samples with silica fume

was calculated from the falling-head permeability tests

done under laboratory conditions. Falling-head perme-

ability tests were performed according to the ASTM D

5084. Composted clay samples with silica fume in

different ratios of 0%, 5%, 10%, 15%, 20%, 25%, 30%

and 50% were subjected to the permeability tests. To

obtain the accurate results from the tests, five different

samples were used for each composite sample. During

E. Kalkan, S. Akbulut / Engineering Geology 73 (2004) 145–156 149

the permeability tests, prepared samples in moulds

were saturated under water pressure for a day or more,

and then permeability values were calculated for 48 h.

At least five specimens were tested for each combina-

tion of permeability values in permeability tests.

3.7. Odometer tests

Swelling pressure tests were conducted on com-

pacted samples of the raw clay and silica fume–clay

mixtures in initial water contents and dry densities for

odometer cells. The swelling pressure of each speci-

men was directly measured from the surcharge, which

loads the sample. The sample was confined in the

consolidation ring, and water was added to the sam-

ple, and it was allowed to swell freely. The samples

used in the tests were 74 mm in diameter and 20 mm

high in standard Odometer apparatus (ASTM D

2435). Samples of modified natural clays were all

initially compacted (in optimum water content) in a

standard Procter mould and extruded using a cutting

ring before odometer tests. Compacted composite

samples were subjected to the neat water and the

leachate to determine effects of silica fume on the

swelling pressure. As the samples were swelling, the

deflection of the dial gauge was set up to zero. As a

result, the samples showed no further tendency to

swell and the maximum surcharge load, Pms, at that

point was used for the calculation of the swelling

pressure. The swelling pressure can be expressed as

Swelling pressure ¼ Pms=A ð2Þ

where swelling pressure is in kilopascals, Pms is the

maximum surcharge load on the specimen in kilo-

newtons, and A is the area of the specimen, in square

centimeters. The time required to reach a maximum

value varies considerably, depending upon the percent-

age of clay (or silica fume). For the lowest percentage

of clay, swell stops in a day, but the swell for the

highest percentage of clay stops in 2 days or more.

Fig. 2. Influence of the silica fume content on the consistency

indexes of clay samples (sample 3).

4. Results and discussion

4.1. Effect of silica fume on consistency limits

The effects of silica fume content on liquid limit,

plastic limit and plasticity index for sample 3 are

presented in Fig. 2. The liquid limit and plasticity

index values slightly decreased with increase in silica

fume content up to 50% in all the clay samples.

However, the plastic limit slightly increased in silica

fume content up to 50% in all the clay samples. The

liquid limit of composed clay samples decreased from

65% to 56%, from 60% to 53%, from 63% to 53%

and from 65% to 53% for the samples 1, 2, 3 and 4,

respectively. The plastic limit of composite samples

slightly increased from 29% to 33%, from 25% to

31%, from 32% to 36% and from 26% to 34% for the

samples 1, 2, 3 and 4, respectively. The reason of this

could be explained depending on the soil type and

associated exchangeable cations (Bell, 1993; Sivapul-

laiah et al., 2000). Due to this change in consistency

limits, some of the composite samples with high silica

fume contents changed the soil groups from high-

plasticity clay group (CH) to low-plasticity clay

groups (OH) as shown in Fig. 3.

4.2. Effect of silica fume on compaction effort

Standard proctor tests were done on both raw clay

and silica fume–clay composite samples to determine

their optimum water content and maximum dry unit

weight relationships. These results, shown in Fig. 4,

were used to determine the optimum water content and

maximum dry unit weight for raw or composite clay

sample preparation. The addition of silica fume to clay

samples increased the optimum moisture content and

Fig. 5. Influence of the silica fume content on the optimum water

content of clay samples.

Fig. 3. The plasticity chart for clays and silica fume–clay samples

used in the tests.

E. Kalkan, S. Akbulut / Engineering Geology 73 (2004) 145–156150

reduced the maximum dry density for the same com-

paction effort (Figs. 5 and 6). The significance of these

changes depends upon the amount of silica fume added

and the chemical composition of the clay minerals.

Silica fume, depending on its content, increased the

total particle surface of the mixture as compared with

that of raw clay sample. Therefore, the optimum water

content increased in the composite samples. Of course,

depending on an increase in optimum water content,

maximum dry unit weight decreased in the composite

samples gradually. The optimum water contents of the

composite samples ranged from 25% to 30%, and the

maximum dry densities ranged from 12.4 to 13.8 kN as

compared all the samples with each other. The low

Fig. 4. Influence of the silica fume content on the compaction

curves of sample 3.

optimum water contents and the high dry unit weights

occurred in raw clay samples, while the high optimum

water contents and the low dry unit weights occurred

in 50% silica fume–clay composite samples.

4.3. Normalization of the experimental results

Practicing engineers who are accustomed to using

conventional experimental results may not prefer the

normalized values. Hence, the results are normalized

to gain an international significance to the experimen-

tal results. The conventional permeability–silica fume

Fig. 6. Influence of the silica fume content on the maximum dry unit

weight of clay samples.

Fig. 8. Relationship between normalized swelling pressure and

silica fume content.

E. Kalkan, S. Akbulut / Engineering Geology 73 (2004) 145–156 151

content, swelling pressure–silica fume content and

compressive strength–silica fume content relation-

ships were modified in terms of normalized perme-

ability, normalized swelling pressure and normalized

compressive strength which accounts for the variation

in maximum dry densities.

Normalized permeability ¼ KexpðGstd=cdmaxÞ ð3Þ

Normalized swelling pressure ¼ Pswðcdmax=GstdÞð4Þ

Normalized compressive strength ¼ Pcsðcdmax=GstdÞð5Þ

where Kexp is the permeability from tests, Psw is the

swelling pressure from tests, Pcs is the compressive

strength from tests, cdmax is the maximum dry density

of the sample and Gstd is the standard value of specific

gravity. A specific gravity value of 2.65 was adopted

as the standard specific gravity in this investigation

since it represents that value for most of the soils (Das,

1983). The permeability, swelling pressure and com-

pressive strength values presented in Figs. 7–10 were

normalized using a specific gravity of 2.65.

4.3.1. The normalized permeability and the normal-

ized swelling pressure of composite samples

The Figs. 7 and 8 represent the silica fume content

against the normalized permeability and the normal-

ized swelling pressure for clay samples, respectively.

Permeability and swelling pressure values steadily

Fig. 7. Relationship between normalized permeability and silica

fume content.

decreased with increasing silica fume content and

the low values were finally reached in the composite

samples with 30% and 50% silica fume contents.

Silica fume contents up to 25% significantly affected

permeability and swelling pressure values. Permeabil-

ity and swelling pressure decreased with more than

25% silica fume content for all the samples. After that,

these values seemed to be less affected by silica fume

contents for all the samples.

As can be seen in Fig. 7, silica fume decreased

the permeability of composite samples, and low-

permeability values were yielded in the composite

samples with 20%, 30% and 50% silica fume as

Fig. 9. Relationship between normalized compressive strength and

silica fume content.

Fig. 10. Effects of maximum dry unit weight on (a) normalized

permeability, (b) normalized swelling pressure and (c) normalized

compressive strength of composite samples.

E. Kalkan, S. Akbulut / Engineering Geology 73 (2004) 145–156152

compared with those of raw clay samples. As compared

with raw clay samples, it was found that the perme-

ability of composite clay samples with 25% silica fume

decreased from 3.8� 10� 7 to 1.46� 10� 7 cm/s, from

3.68� 10� 7 to 1.41�10� 7 cm/s, from 3.51�10� 7 to

9 .03 � 10� 8 cm/s and f rom 2.94 � 10� 7 to

1.22� 10� 7 cm/s for the samples 1, 2, 3 and 4, res-

pectively. The permeability of all the samples was

calculated by averaging five test results shown in

Fig. 7.

In each sample, it was observed that improvement

in swelling pressure was obtained by using silica

fume contents. The swelling pressure in composite

samples was decreased with silica fume contents as

shown in Fig. 8. It was observed that the decrease in

the swelling pressure occurred with rising silica fume

contents, which ranged from 0% to 25%. The swell-

ing pressure of the samples with silica fume content

more than 25% remained almost the same. The

swelling pressure of composite samples with 25%

silica fume contents decreased from 88 to 9 kPa, from

98 to 11 kPa, from 109 to 19 kPa and from 108 to 16

kPa for the samples 1, 2, 3 and 4, respectively. It can

be seen from Figs. 7 and 8 that the regression

coefficient (R) is quite high, and its values are 0.95

for two cases. The given equations are also valid only

for the normalized permeability and normalized

swelling pressure.

4.3.2. The normalized compressive strength of

composite samples

The effects of silica fume contents against the

normalized compressive strength for composite sam-

ples are presented in Fig. 9. Maximum compressive

strength values were obtained in composite samples

with 30% and 50% silica fume contents. Compressive

strength for all samples significantly increased with

silica fume content up to 25%. However, after that,

compressive strength was slightly affected by silica

fume content rise.

Silica fume increased the strength of the composite

samples. It was found that the compressive strength of

natural clays increased due to the rise of 25% silica

fume content from 84 to 116 kPa, from 77 to 113 kPa,

from 91 to 114 kPa and from 85 to 110 kPa for

samples 1, 2, 3 and 4, respectively. Silica fume

reinforcement resulted in an increase in the compres-

sive strength of clay samples (Kalkan and Akbulut,

Fig. 11. The effect of leachate on properties of composite samples

(a) permeability and (b) swelling pressure.

E. Kalkan, S. Akbulut / Engineering Geology 73 (2004) 145–156 153

2002). The increase in compressive strength was

attributed to the internal friction of silica fume par-

ticles and chemical reaction between silica fume and

clay materials. Clay samples treated with silica fume

reached their peak strength of the 30% silica fume

contents. An increase in silica fume content in clay

made the clay sample with silica fume more brittle

than the raw clay sample, though the fracture in the

raw clay samples was ductile as compared to all the

composite samples. The equation given in Fig. 9 is

valid only for the normalized compressive strength,

and the regression coefficient (R) is 0.95 for this case.

4.4. Effect of dry unit weight on the properties of

composite samples

The relationships of maximum dry density–nor-

malized permeability, maximum dry density–normal-

ized swell pressure, and maximum dry density–

normalized compressive strength for all the raw and

silica fume–clay samples are shown in Fig. 10. While

the permeability and the swelling pressure decreased

(Fig. 10a and b), the compressive strength increased

with a decrease in maximum dry density (Fig. 10c).

4.5. Effect of leachate on composite samples

From permeability and swelling pressure tests sub-

jected to the neat water and the leachate, it was seen

that composite samples with 25% silica fume had low

permeability and swelling pressures as compared with

raw clay samples (0% silica fume). Therefore, perme-

ability and swelling pressure tests using leachate were

done on the samples with 25% silica fume. Composite

samples had more resistance to leachate than raw

samples as can be seen in Fig. 11. The permeability

and swelling pressure values of raw clay and compos-

ite samples with 25% silica fume content, which were

tested with leachate, are shown in Fig. 11a and b. The

main trends in Fig. 11a and b are similar to those in

Figs. 7 and 8; the permeability and swelling pressure

of samples with 25% silica fume content decreased as

compared with those of raw clay samples.

4.6. Image of composite samples

In order to evaluate the interaction between silica

fume and silt or clay particles, composed samples

with 0%, 10%, 20% and 30% silica fume were

magnified 5000 times by means of a scanning electron

microscope (SEM) modeled Jeol 6400 SEM. After

curing stabilized specimens, small samples were taken

from the stabilized composite samples for microscopic

analysis to determine the nature of the pores and the

effect of silica fume contents on the pore structure.

Fig. 12 shows SEM micrographs of composite sam-

ples prepared with silica fume contents (0%, 10%,

20% and 30%). Samples with 30% silica fume content

show a denser structure than those of the samples with

0%, 10% and 20% silica fume contents. The clay or

silt particles are coated with a thin layer of silica fume

and form groups. Fig. 12a (0% silica fume) shows that

large continuous pores (up to 30 Am) among the soil

particles provide the large portion of total void ratio,

Fig. 12. Images of composite sample 3 with (a) 0% SF, (b) 10% SF, (c) 20% SF and (d) 30% SF.

E. Kalkan, S. Akbulut / Engineering Geology 73 (2004) 145–156154

while small connected pores (less than 5 Am) exist

among the micro-fine silica fume particles.

It was seen from the images that silica fumes

covered the surrounding of silt and clay particles

and filled the voids in samples. Silt and clay grains

showed angular or subangular shapes (Fig. 12a). In

the samples with 10% silica fume, grain surfaces were

partly covered by silica fumes (silica fume micro

particles), and most of the pore spaces still contained

air (Fig. 12b). In the samples with 20% silica fume,

soil grains were covered by fume in significant ratio,

and micropores were filled by fume (Fig. 12c). In the

samples with 30% silica fume, all grains were covered

by relatively thick silica fume material, which formed

cementing medium (Fig. 12d). Fig. 12 shows the silica

fume particles settle in the pore space among the silt–

clay grains, then the settled silica fume particles react

to form hydration products (flocculation products) in

the surrounding of soil grains. This textural event

caused a significant improvement in permeability,

swelling pressure and compressive strength (reached

to ultimate values in 50% silica fume content). This

effect also can be seen in Figs. 7–9. A detailed

examination of each micrograph reveals that most of

the flocculation products are deposited on the surfaces

of the soil grains or at the contact points. For this

reason, these micrographs and the experimental

results above have also released a possibility that a

chemical reaction with silica fume and clay particles

may occur.

ering Geology 73 (2004) 145–156 155

4.7. Chemical modification of silica fume–clay

mixtures

It is difficult to give a reaction of silica fume with

clay under normal conditions. However, interpretation

of chemical reactions must be explained to understand

silica fume–clay modification. To produce the mod-

ification of clay, the important two effects are the

quality and quantity of silica fume added to clay and

the chemical composition of clay. The clays used in

this research have significant quantity of calcium

compounds (Table 2), which form Ca2 + ions and

hydroxyl ions by reacting water molecules. The active

silica reacts with calcium hydroxide and forms calci-

um silicate hydrate gels (CaSiO3�H2O). The basic

reaction of silica fume–calcium in clay could be

indicated as

Ca2þ þ OH� þ soluble silica

! calcium silicate hydrate ð6Þ

It was found out that the material from this reaction

became stronger and more brittle than previous form

as can be seen in Fig. 9. There are similar researches

to these concepts (Bell, 1993; Sherwood, 1993).

E. Kalkan, S. Akbulut / Engine

5. Conclusions

The following conclusions are derived from this

investigation:

Silica fume decreased the liquid limits and

plasticity index and increased the plastic limits in

all the clay samples. For this reason, the soil types

of composite samples with high silica fume

contents changed from high-plastic clays (CH) to

low-plastic clays (OH). Silica fume slightly increased the optimum water

content and decreased the maximum dry unit

weights of the samples for all the composite

samples in the same compaction effort. A significant improvement on the permeability,

swelling pressure and compressive strength of

composite samples was obtained by using silica

fume. As the permeability and swelling pressure

decreased, the compressive strength of clay

samples proportionally increased with silica fume

contents for all the samples. The low permeability

and swelling pressure and high compressive

strength were calculated in the composite samples

with 25% or more silica fume contents. Observations of scanning electron microscope

showed that the structure of raw clay samples

could be changed through silica fume contents in

the sample. The structure of a material had a

significant influence on its engineering properties

such as permeability, strength and stiffness. The chemical analysis of the silica fume–clay

mixture products showed that a chemical reaction

between silica fume and clay particles could appear

depending on the chemical characteristics of clay

and the properties of silica fume type. The investigation showed that the silica fume is a

valuable material to modify the properties of clay

liners to be used in the landfill sites. In the design

of a liner system, the silica fume content in the

liner should be taken into account due to its

positive effects on the permeability, swelling

pressure and compressive strength of clay liners.

Acknowledgements

This research was conducted by a project, num-

bered 2002/139, supported by the Research Develop-

ment Center of Ataturk University. The authors also

thank Prof. Dr. Sahin Gulaboglu, from the Chemical

Engineering Department, for his valuable contribu-

tions to clarify the chemical reaction in the study.

References

Akbulut, S., 1999. Improvement of geotechnical properties of gran-

ular soil by grouting. PhD thesis, Technical University of Istan-

bul. Istanbul.

Bell, F.G., 1993. Engineering treatment of soils, Published by E and

FN Spon, an imprint of Chapmen and Hall, Boundary Row,

London.

Bouazza, A., 2002. Geosynthetic clay liners. Geotext. Geomembr.

20, 3–17.

Bouazza, A., Van Impe, W.F., 1997. Bottom liner design of some

landfills in Belgium and other countries. In: Marinos, P.G.,

Koukis, G.C., Tsiambous, G.C., Stournaras, G.C. (Eds.), Pro-

ceedings of International Symposium Engineering Geology and

the environment, Athens, Greece, vol. 2. Balkema, Botterdam,

pp. 1629–1933.

E. Kalkan, S. Akbulut / Engineering Geology 73 (2004) 145–156156

Brandl, H., 1992. Mineral liners for hazardous waste containment.

Geotechnique 42, 57–65.

Cazaux, D., Didier, G., 2000. Field evaluation oh hydraulic per-

formances of geosynthetic clay liners by small and large-scale

tests. Geotext. Geomembr. 18, 163–178.

Daniel, D.E., Benson, C.H., 1990. Water content-density criteria

for compacted soil liners. J. Geotech. Eng. ASCE 12 (116),

1811–1830.

Daniel, D.E., Wu, Y.K., 1993. Compacted clay liners and covers for

sites. J. Geotech. Eng. 119 (2), 223–237.

Das, B.M., 1983. Advanced Soil Mechanics. McGraw Hill, New

York.

Das, B.M., 1997. Prenciples of geotechnical engineering. Fourth

Edition California State University, Sacramento.

Foged, N., Baumann, J., 1999. Clay membrane made of natural

high plasticity clay: leachate migration due to advection and

diffusion. Eng. Geol. 54, 129–137.

Kalkan, E., Akbulut, S., 2002. Improving of shear strength of

natural clay liners from clay deposits NE Turkey. 3rd Interna-

tional Conference on Landslides, Slope Stability and the Safety

of Infrastructures. Cl-Premier Ltd., Singapore, pp. 295–300.

Katsumi, T., Benson, C.H., Foose, G.J., Kamon, M., 2001. Perfor-

mance-based design of landfill liners. Eng. Geol. 60, 139–148.

Kayabaly, K., 1997. Engineering aspects of a noval landfill liner

material: bentonite-amended natural zeolite. Eng. Geol. 46,

105–114.

Kennay, T.C., Van Veen, V.A., Swallow, M.A., Sungalia, M.A.,

1992. Hydraulic conductivity of compacted bentonite – sand

mixtures. Can. Geotech. J. 29, 638–640.

Patrick, J.F., Daniel, J.D.B., David, G.M., 2000. Hydraulic perfor-

mance of geosynthetic clay liners under gravel cover soils. Geo-

text. Geomembr. 18, 179–201.

Prashant, J.P., Sivapullaiah, P.V., Sridharan, A., 2001. Pozzolanic

fly ash as a hydraulic barrier in land fills. Eng. Geol. 60,

245–252.

Shackelford, C.D., Benson, C.H., Katsumi, T., Edil, T.B., Lin, L.,

2000. Evaluating the hydraulic conductivity of GCLs permeated

with non-standard liquids. Geotext. Geomembr. 18, 133–161.

Sherwood, P.T., 1993. Soil stabilization with cement and lime.

State-of-the-Art Review, Transportation Research Laboratory.

HMSO, London.

Sivapullaiah, P.V., Sridharan, A., Bhaskar, R.K.V., 2000. Role of

amount and type of clay in the lime stabilization of soils.

Ground Improv. 4, 37–45.

Van Ree, C.C.D.F., Westtrate, F.A., Meskers, C.G., Bremmer, C.N.,

1992. Design aspects and permeability testing of natural clay

and sand-bentonite liners. Geotechnique 1 (42), 49–56.