Clay Minerals, (2018) 53, 721–732

OPEN ACCESS

Effects of clay-mineral type and content on

the hydraulic conductivity of bentonite–sand
mixtures made of Kunigel bentonite
from Japan
M A S A N O R I KO H N O 1 , * , YO S H I TA K A N A R A 2 , M A S A J I K ATO 3 AND
T S U YO S H I N I S H I M U R A 1
1
Tottori University, Tottori 680-8552, Japan
2
Kyoto University, Kyoto 615-8540, Japan
3
Hokkaido University, Hokkaido 060-8628, Japan

(Received 29 November 2017; revised 16 June 2018; Accepted Manuscript published online: 17 December 2018;
Version of Record published online: 1 February 2019; Guest Associate Editor: R. Dohrmann)

A B S T R AC T : Clay-mineral type and content, bulk mineralogical composition and alteration of

bentonite are very important factors for the ultra-long-term stabilization of barriers and backfills in
radioactive waste disposal. This study investigates the effects of clay-mineral type and content on the
swelling characteristics and permeability of bentonite–sand mixtures with clay minerals using one-
dimensional swelling-pressure and constant-pressure permeability tests. The hydraulic conductivity of
bentonite–sand–clay mineral mixtures increased with increasing content of non-swelling alteration
products of montmorillonite. Furthermore, hydraulic conductivity was comparable to that determined
with the Kozeny–Carman equation for a specific surface area, suggesting that hydraulic conductivity
may be estimated based on the abundance of expected alteration products of montmorillonite. This study
provides a basis for evaluation of the hydraulic conductivity of bentonite–sand mixtures with known
quantities of expected alteration products of montmorillonite.

KEYWORDS: clay mineral, swelling pressure, hydraulic conductivity, bentonite–sand mixtures, alteration, specific
surface area.

Engineered barriers and backfill for the geological stabilization. The use of compacted sodium-type
disposal of radioactive waste require ultra-long-term bentonite–sand mixtures for this purpose has received
increased attention. Previous work has studied the
swelling characteristics, permeability and self-sealing
capacity of bentonite and bentonite–sand mixtures
This paper was presented during the ‘7th International (e.g. Komine & Ogata, 1994, 1999a; Mollins et al.,
Conference on Clays in Natural and Engineered Barriers
1996; Cho et al., 1999; Sivapullaiah et al., 2000;
for Radioactive Waste Confinement’, September 2017.
*E-mail: [email protected] Shirazi, 2010; Suzuki et al., 2013). Pusch (1999)
https://doi.org/10.1180/clm.2018.52 studied the swelling pressure and hydraulic conduct-
ivity of various clay materials such as kaolinite,
This is an Open Access article, distributed under the terms
smectite and palygorskite. Sellin & Leupin (2013)
of the Creative Commons Attribution licence (http://
creativecommons. org/licenses/by/4.0/), which permits have reviewed most of the results from over the past 35
unrestricted re-use, distribution, and reproduction in any years on the safety-relevant properties of clays used as
medium, provided the original work is properly cited engineered barriers for the geological disposal of

Copyright © Mineralogical Society of Great Britain and Ireland 2018

722 Masanori Kohno et al.

radioactive waste. However, few quantitative data on and contents on the swelling characteristics and
the effects of clay mineral types, contents and permeability of bentonite–sand mixtures in laboratory
alterations on the swelling characteristics and perme- experiments. Bentonite–sand–clay mineral mixtures
ability of bentonite–sand mixtures have been reported might be affected by the alteration of bentonite via
to date. Furthermore, many researchers (e.g. Nakayama changes in the clay-mineral content. To clarify the
et al., 2004; Yamaguchi et al., 2007; Cuisinier et al., effect of clay-mineral content, the bentonite–sand–clay
2008; Chen et al., 2016) have studied the swelling mineral mixtures were subjected to one-dimensional
characteristics and permeability of altered bentonite swelling-pressure and constant-pressure permeability
using alkaline solutions and various approaches. tests.
Yamaguchi et al. (2013) reported an increase in the
hydraulic conductivity of bentonite–sand mixtures
after alteration by alkaline fluids. However, the METHODS
swelling characteristics and permeability of benton- Materials
ite–sand mixtures containing kaolinite and chlorite as
alteration minerals have yet to be investigated. In the One-dimensional swelling-pressure and constant-
ultra-long-term use of bentonite–sand mixtures for pressure permeability tests with samples containing
geological disposal of radioactive waste, it is assumed one clay mineral were carried out to clarify the
that montmorillonite will be at least partly altered to influence of clay mineralogy on swelling pressure
illite or chlorite, and feldspar to kaolinite (e.g. Inoue and hydraulic conductivity. The clay minerals used
et al., 1987; Japan Nuclear Cycle Development were kaolinite, 10 Å halloysite, talc, mica and chlorite.
Institute, 2000; Nakayama et al., 2004; Yokoyama The clay-mineral samples were collected primarily
et al., 2006; Savage et al., 2007; Yokoyama & from ancient hydrothermal fields. The samples passed
Nakamura, 2010). These alterations may deteriorate through a 75 µm mesh sieve. Next, the kaolinite, mica
the properties of bentonite–sand mixtures. Thus, the and chlorite were mixed with bentonite and sand to
swelling characteristics and permeability of bentonite form bentonite–sand–clay mineral mixtures for one-
appear to be affected by alteration. The rate of dimensional swelling-pressure and constant-pressure
alteration of bentonite–sand mixtures is slow. Several permeability tests. The bentonite and silica sand used
mathematical models have been developed to predict in this study were Kunigel-V1 collected in Tsukinuno
the ultra-long-term effects of alteration on the Mine, Japan, and Mikawa no. 6 silica sand, respect-
properties of bentonite–sand mixtures (e.g. Ishikawa ively. Kunigel-V1 is composed of montmorillonite
et al., 1994; Savage et al., 2002; Ishii et al., 2013). (46–49%), quartz (29–38%), feldspars (2.7–5.5%),
However, very few experimental data on the modifi- calcite (2.1–2.6%), dolomite (2.0–2.8%), zeolites (3.0–
cation of hydraulic conductivity have been reported. 3.5%) and pyrite (0.5–0.7%) (Ito et al., 1993; Japan
Hydrothermal alteration by volcanic–hydrothermal Nuclear Cycle Development Institute, 2000). Mikawa
systems and weathering over geological time and no. 6 silica sand is composed mainly of quartz. In order
historical periods have occurred widely throughout the to avoid confusion with montmorillonite (clay mineral),
Pacific Rim region. As a result of these processes, a wide we use the term ‘clay mineral sample’ for kaolinite,
variety of clay-rich sediments, bentonites and soils occur 10 Å halloysite, talc, mica and chlorite. In addition, the
on the surface and in the crust of the Earth. The mineral kaolinite, mica and chlorite used for the bentonite–
composition of bentonite depends on the parent material sand–clay mineral mixtures are referred to as ‘non-
and weathering environment. Bentonite is composed swelling alteration products’ (NSAPs) (Fig. 1). Table 1
mainly of montmorillonite mixed with other clay and Fig. 1 show the particle densities of the samples and
minerals (e.g. illite, kaolinite, chlorite) and non-clay proportions of bentonite, silica sand and clay mineral
minerals (e.g. quartz, feldspars, carbonates, pyrite). samples. The particle densities (ρS) of bentonite–sand
Furthermore, the physical and mechanical properties of mixtures were calculated from equation 1:
clay minerals vary, as shown by their structures.
m
Thus, clay-mineral type and content, crystal- rs ¼ m mN mS (1)
chemical composition and assemblage and the þB
þ
rB rN rS
expected alteration products of bentonite are important
factors for the ultra-long-term stabilization of barriers where m is the mass, ρ is the particle density and the
and backfill in radioactive waste disposal. This subscripts B, N and S represent bentonite, mixed
study investigated the effects of clay-mineral types NSAPs and silicate sand, respectively.
 Effects of clay mineral type and content on hydraulic conductivity 723

FIG. 1. Proportions of components in bentonite–sand–clay mineral mixtures. Mixed NSAPs were kaolinite, mica and
chlorite.

The ratio of bentonite and mixed NSAPs to sand was determined by N2 adsorption with the Brunauer–
7:3 in dry mass. The materials were dried at 40 ± 3°C in Emmett–Teller (BET) method after degassing at 200°C
an electric oven to achieve a constant mass and to avoid for 1 h. The specific surface areas of the samples are
structural collapse caused by dehydration of 10 Å listed in Table 2.
halloysite and montmorillonite. The mixed NSAP
replacement ratio to bentonite (α), defined as the mass
of the clay-mineral sample divided by the mass of the
Sample preparation and test procedure
bentonite and clay mineral sample, ranged from 0% to The specimens used for one-dimensional swelling-
100% at 25% intervals (Fig. 1). Three materials pressure and constant-pressure permeability tests were
(bentonite, mixed NSAPs and silica sand) were 50 mm in diameter and 10 mm in height. The dried
mixed in a container. Bentonite and the clay mineral clay mineral and bentonite–sand mixtures were
sample were identified by X-ray diffraction (XRD) compacted statically to a dry unit mass of 1.4 mg/m3.
traces of untreated powder samples (Fig. 2). The XRD The compacted specimens were produced using an
was performed with a Rigaku Ultima IV diffractometer apparatus composed of a cylindrical mould, pistons
(40 kV, 20 mA, 0.03 s/step) using Cu-Kα radiation, a and an oil-pressure jack. The tests were performed at
0.15 mm receiving slit, a 0.5° divergence slit and a 2° 22 ± 1°C in a temperature-controlled room.
scattering slit. The kaolinite sample is composed mainly Figure 3 shows the experimental configuration of
of kaolinite with traces of quartz (Fig. 2c). The talc the one-dimensional swelling-pressure test. The testing
sample consists mainly of talc and traces of chlorite system consisted of an elevating-load machine,
(Fig. 2e). The chlorite sample is composed mainly of cylindrical mould (acrylic resin ring), base plate,
chlorite with minor quartz, feldspars calcite and loading plate, load cell, porous stone, water-immersing
sepiolite (Fig. 2g). The specific surface areas of the container and displacement transducer. The test was
clay-mineral powder samples and silicate sand were performed at constant volume. A specimen was placed
in the cylindrical mould with a base plate, the assembly
TABLE 1. Particle densities of samples. was inserted into the water-immersing container and
the container was filled with distilled water. A
0.003 kN load was then applied to ensure that the
Samples Particle density (g/cm3) height of the specimen did not change, and the axial
Bentonite 2.7a swelling pressure was determined from the vertical
Silicate sand 2.663 force applied at that time. A displacement transducer
Clay mineral samples was used to measure the displacement of the specimen
Kaolinite 2.577 throughout the test. The maximum value of displace-
10 Å halloysite 2.736 ment was 0.01 mm. The swelling pressure is given by
Talc 2.786 equation 2:
Mica 2.904 F
Chlorite 2.841 Ps ¼ (2)
A
where Ps is the swelling pressure, F is the vertical force
a
Japan Nuclear Cycle Development Institute (2000). and A is the cross-sectional area of the specimen. A
724 Masanori Kohno et al.

FIG. 2. Random XRD traces of (a) bentonite, (b) silica sand and clay-mineral powders: (c) kaolinite, (d) 10 Å halloysite,
(e) talc, (f ) mica and (g) chlorite. c.p.s. = counts per second.

data logger was used to measure vertical force at 1 s steel. The electronic balance (0.1 mg accuracy) was
intervals. The test was conducted for at least 72 h, after used to measure the amount of water discharge, with
which measurements were taken until the swelling consideration of evaporation. The runoff volume was
pressure had been constant for 24 h. measured at 1 min intervals, and the data were
Figure 4 shows the experimental configuration of transmitted from the balance to a personal computer.
the constant-pressure permeability test. The testing Each specimen was saturated using a water-filled
system consisted of a compressor, pressure water tank, decompression container. The swelling pressure of the
pressure regulator, pressure gauge, water pipe, perme- clay mineral samples attained equilibrium at ∼2 days
ability cylindrical mould (stainless steel), filters ( paper (Fig. 5). Therefore, the water-immersion times were
and porous stone), O rubber ring and electronic determined to be 4 days for the clay-mineral samples
balance. The pipe and cylinder were made of stainless and 7 days for bentonite. The water-immersion time for
 Effects of clay mineral type and content on hydraulic conductivity 725

TABLE 2. Specific surface areas of clay-mineral powder R E S U LT S A N D D I S C U S S I O N

samples and silica sand, measured according to the
BET method. Swelling-pressure and hydraulic conductivity measure-
ments were performed with mixed NSAPs and bentonite
(Figs 5, 6) and admixtures thereof (Fig. 7). The
Samples Specific surface area (m2/g) reproducibility of the one-dimensional swelling-pressure
and constant-pressure permeability tests was assessed in
Kaolinite 40.42
triplicate experiments. Sample data in Fig. 5 represent
10 Å halloysite 61a
Talc 5.39b
the average of at least three measurements under the
Mica 4.85 same conditions. Figure 6 shows the error bars (range of
Chlorite 5.20 maximum and minimum values) of the sample data. The
Silica sand 0.09 standard deviation of the swelling pressure was
0.008 MPa. The differences between the maximum
a and minimum values of the hydraulic conductivity of
The Clay Science Society of Japan (2009).
b samples were less than one order of magnitude.
Holland & Murtagh (2000).
Therefore, the variation in the swelling pressure and
hydraulic conductivity of the samples was very small.
bentonite corresponds to at least the duration of
the one-dimensional swelling-pressure tests. The
minimum degree of saturation of specimens was Swelling pressure and hydraulic conductivity of
98.9%. Therefore, the specimens used were considered the clay-mineral samples and bentonite
to be water-saturated. Distilled water was passed The swelling pressures of five clay-mineral samples
through the specimen at constant water pressure and bentonite are shown in Fig. 5. The average
using a compressor and pressure water tank. swelling pressure was highest for bentonite (Ps =
Measurements were taken for ≥24 h after the first 0.48 MPa), followed by 10 Å halloysite (Ps =
confirmation of water runoff. The water pressure was 0.24 MPa) and kaolinite (Ps = 0.17 MPa). In contrast,
kept lower in the constant-pressure permeability test values for talc and mica were low (Ps = 0.01 and
than in the swelling-pressure test to avoid consolida- 0.02 MPa, respectively), and that for chlorite was
tion or water path (Table 3). As an exception, the water <0.002 MPa. The bentonite required a longer time than
pressure of the chlorite sample and chlorite–sand the clay mineral samples to reach equilibrium swelling
mixtures (α = 100%, Ps <0.002 MPa) was 0.002 MPa. pressure. However, because bentonite has low perme-
Consolidation or water path of these specimens was not ability, a longer time was required for the water to reach
observed after the test. The hydraulic conductivity of all parts of the specimen. The swelling pressures of the
clay mineral samples and bentonite–sand mineral clay mineral samples differed according to the clay-
mixtures was calculated from Darcy’s law (equation 3): mineral type. In particular, the swelling pressures of
10 Å halloysite and kaolinite were greater than those of
QL the other clay-mineral samples. Because the specific
k¼ (3)
hAt surface areas of 10 Å halloysite and kaolinite were also
larger than those of the other clay-mineral samples
where k is the hydraulic conductivity, Q is the runoff (Table 2), the adsorbed water owing to surface tension
volume, L is the height of the specimen, h is the increased. Thus, the pore-water pressure and the
difference in water level, A is the cross-sectional area of swelling pressure of halloysite and kaolinite increased.
the specimen and t is time. h was calculated according The relationships between swelling pressure and
to the pressure in the pressure water tank (e.g. at water hydraulic conductivity for the five clay-mineral
pressure 0.1 MPa, h is 1020.4 cm). The water pressure samples and bentonite are shown in Fig. 6 and
was controlled using the pressure regulator and the Table 4. Hydraulic conductivity was lowest for
pressure gauge. The accuracy was to 0.002 MPa. The bentonite (average of k = 5.3 × 10–13 m/s), followed
hydraulic conductivity of bentonite–sand mixtures by kaolinite and 10 Å halloysite (10–10 m/s) and mica,
depends slightly on the hydraulic gradient i (= h/L) talc and chlorite (10–8 m/s). The hydraulic conductivity
(Nakamura et al., 2011). In this study, the hydraulic of bentonite in this study was comparable to that
gradient (i = 20.4–3571.4) did not affect hydraulic reported in previous studies under similar conditions
conductivity. (4.8 × 10–12 m/s) (Suzuki et al., 1992). The hydraulic
726 Masanori Kohno et al.

FIG. 3. Schematic diagram of the experimental configuration for the one-dimensional swelling-pressure test.

conductivity of the specimens decreased with increas- force and hydraulic conductivity will be clarified in
ing swelling pressure, and analysis revealed a strong future studies. In particular, montmorillonite in ben-
correlation (R = 0.97; Fig. 6). The hydraulic conduct- tonite has a 2:1 layer. Because it is a swelling clay
ivity of the clay-mineral samples varied according mineral containing water and cations between layers,
to the clay-mineral type. Interestingly, the swelling however, a significantly larger electric double layer is
pressure (Ps), hydraulic conductivity (k) and specific formed compared with those for mica or chlorite. It is
surface area (S) of the kaolinite and halloysite (both suggested that the formation of an immobile water film
with 1:1 layer structures) were similar, and the Ps, k in the electric double layer and physicochemical action
and S of talc (2:1 layer), mica (2:1 layer) and chlorite (adsorption and flocculation) by the swelling clay
(2:1:1 layer) were comparable. The large specific mineral contributed significantly to the low hydraulic
surface area resulted in an increase in pore-water conductivity of bentonite (Achari et al., 1999; Mitchell
viscosity, which in turn led to low hydraulic conduct- & Soga, 2005). The hydraulic conductivities of
ivity (Achari et al., 1999). The relationship between kaolinite, mica and chlorite, which might have
the clay-mineral structures or the interlayer bonding formed during the alteration of bentonite under
 Effects of clay mineral type and content on hydraulic conductivity 727

FIG. 4. Schematic diagram of the experimental configuration for the constant-pressure permeability test.

FIG. 5. One-dimensional swelling pressure of clay-mineral samples and bentonite (Kunigel V1).

repository conditions, and those of mica and chlorite bentonite–sand mixtures decreased with increasing
were approximately three to five orders of magnitude swelling pressure, and analysis revealed a strong
greater than the hydraulic conductivity of bentonite. correlation (R = 0.87).
These experimental results suggest that an increase in The relationships between the mixed NSAP replace-
non-swelling clay minerals due to alteration of ment ratio to bentonite and hydraulic conductivity of
bentonite increases hydraulic conductivity. the bentonite–sand–clay mineral mixtures are shown in
Fig. 8. The hydraulic conductivities of the kaolinite–
sand, mica–sand and chlorite–sand mixtures (α =
Swelling pressure and hydraulic conductivity of
100%; k = 6.3 × 10–10 m/s, 3.6 × 10–8 m/s and 9.4 ×
bentonite–sand–clay mineral mixtures
10–8 m/s, respectively) were more than two to four
The relationship between the swelling pressure and orders of magnitude greater than that of the bentonite–
hydraulic conductivity of the bentonite–sand–clay sand mixture (α = 0%, k = 1.4 × 10–12 m/s). The
mineral mixtures is shown in Fig. 7. The swelling hydraulic conductivity of the bentonite–sand mixture
pressure of the bentonite–sand–clay mineral mixtures (α = 0%) was similar to that observed previously under
ranged from 0.00 to 0.25 MPa. The swelling pressure the same test conditions (1.0 × 10–12 m/s) (Suzuki
decreased with increasing replacement of bentonite by et al., 1992). Compared with the hydraulic conduct-
NSAPs, without major differences among the various ivity of the bentonite–sand mixtures (α = 0%), that of
types of NSAPs. The hydraulic conductivity of the the bentonite–sand–clay mineral mixtures did not
728 Masanori Kohno et al.

TABLE 3. Swelling pressure (Ps) of specimens and water

pressure (Pw) for the constant-pressure permeability test.

Ps Pw
Specimen (MPa) (MPa)

Bentonite 0.48 0.35

Clay mineral samples
Kaolinite 0.17 0.16
10 Å halloysite 0.24 0.22
Talc 0.01 0.01
Mica 0.02 0.02
Chlorite <0.002 0.002
Bentonite–sand mixture (α = 0%) 0.25 0.24
Bentonite–sand–clay mineral
mixtures
Kaolinite mixture (α = 25%) 0.20 0.18
(α = 50%) 0.12 0.10 FIG. 7. Relationship between swelling pressure and hydraulic
(α = 75%) 0.05 0.04 conductivity of bentonite–sand–clay mineral mixtures.
(α = 100%) 0.03 0.02
Mica mixture (α = 25%) 0.19 0.16 differ in magnitude for α = 25%. However, a difference
(α = 50%) 0.11 0.10 of one order of magnitude was noted for α = 50%, and
(α = 75%) 0.05 0.04
a difference of one to two orders of magnitude was
(α = 100%) 0.01 0.01
observed for α = 75%.
Chlorite mixture (α = 25%) 0.19 0.18
The hydraulic conductivity of the bentonite–sand–
(α = 50%) 0.11 0.08
clay mineral mixtures increased with the mixed NSAP
(α = 75%) 0.04 0.02
(α = 100%) <0.002 0.002 replacement ratio. For mixtures without bentonite (i.e.
clay mineral–sand mixtures), hydraulic conductivity
differed depending on the mixed NSAP mineral type,
suggesting that an increase in clay minerals due to
alteration leads to increased hydraulic conductivity.
Conversely, the hydraulic conductivity of kaolinite,
mica and chlorite (α = 25%) was comparable, and was
not affected by the mixed NSAP type. When a small

FIG. 6. Relationship between swelling pressure and FIG. 8. Relationship between mixed NSAP replacement
hydraulic conductivity of clay-mineral samples and ratio to bentonite and hydraulic conductivity of bentonite–
Kunigel V1. sand–clay mineral mixtures.
 Effects of clay mineral type and content on hydraulic conductivity 729

TABLE 4. Hydraulic conductivity (k) of samples. mineral content. Among models representing the
hydraulic conductivity of the porous media, the
Kozeny–Carman equation (equation 4) uses the specific
Samples k (m/s)
surface area. As this equation incorporates macropara-
Bentonite (5.3 ± 1.5) × 10–13 meters (i.e. void ratio) and microparameters (i.e. specific
Clay mineral samples surface area), it is often used to evaluate the hydraulic
Kaolinite (1.7 ± 0.3) × 10–10 conductivity of bentonite–sand mixtures (e.g. Ren et al.,
10 Å halloysite (1.2 ± 0.3) × 10–9 2016; Kobayashi et al., 2017). The Kozeny–Carman
Talc (3.1 ± 0.5) × 10–8 equation (Kozeny, 1927; Carman, 1937) is:
Mica (5.3 ± 0.4) × 10–8
Chlorite (5.3 ± 0.7) × 10–8 rw g 1 e3
kKC ¼ C  2 2 (4)
h Sm rs 1 þ e
where kKC (m/s) is hydraulic conductivity, C is a
amount of bentonite was added, the hydraulic dimensionless shape constant with an approximate value
conductivity was not affected considerably by the of 0.2, ρw (= 997.770 kg/m3) is the density of water at
mixed NSAP type. In other words, hydraulic conduct- 22°C, g is gravitational acceleration (= 9.8 m/s2),
ivity appears to be affected significantly by bentonite η (=0.000958 Pa·s) is the coefficient of viscosity at
swelling. This swelling fills most voids when the 22°C, Sm (m2/kg) is the specific surface area per unit mass
bentonite content is large. However, as the mixed of particles, ρs (kg/m3; using equation 1) is the particle
NSAP content increases, swelling pressure decreases density and e (ρs/ρd – 1, ρd = 1400 kg/m3) is the void
and the filling of voids becomes difficult, resulting in ratio. The dimensionless shape constant C = 0.2 gives the
increased hydraulic conductivity. Therefore, hydraulic best fit with experimental results (Carman, 1973).
conductivity may not be affected significantly by clay- In this study, the specific surface areas (Sm) of
mineral type when bentonite is present. bentonite–sand–clay mineral mixtures were calculated
according to mixed NSAP replacement ratio to bentonite
(α) using equation 5 (Komine & Ogata, 1999b):
Hydraulic conductivity of bentonite–sand–clay
mineral mixtures as determined by the Kozeny– SB rB VB þ SN rN VN þ SS rS VS
Sm ¼ (5)
Carman equation rB VB þ rN VN þ rS VS
Various parameters, such as the specific surface area where S is the specific surface area per unit mass, ρ is the
and void ratio, change due to differences in the clay particle density, V is the volume of the component and

TABLE 5. Hydraulic conductivities (k, kKC), specific surface areas (Sm) and void ratios (e) of the bentonite–sand–clay
mineral mixtures.

Mixed NSAP α (%) k (m/s) kKC (m/s), equation 4 Sm (m2/g), equation 5 e

–12 –12
0 1.4 × 10 1.6 × 10 264.99 0.92
Kaolinite 25 5.4 × 10–12 2.6 × 10–12 205.64 0.90
50 1.9 × 10–11 5.1 × 10–12 146.28 0.89
75 6.9 × 10–11 1.4 × 10–11 86.93 0.87
100 6.0 × 10–10 1.4 × 10–10 27.57 0.86
Mica 25 6.2 × 10–12 3.0 × 10–12 199.58 0.94
50 6.4 × 10–11 6.9 × 10–12 134.16 0.97
75 4.8 × 10–10 2.7 × 10–11 68.75 0.99
100 3.6 × 10–8 1.2 × 10–8 3.33 1.02
Chlorite 25 4.0 × 10–12 3.0 × 10–12 199.64 0.94
50 1.3 × 10–11 6.7 × 10–12 134.28 0.95
75 9.2 × 10–11 2.6 × 10–11 68.93 0.97
100 9.4 × 10–8 1.0 × 10–8 3.57 0.99
730 Masanori Kohno et al.

the subscripts m, B, N and S represent bentonite–sand– replacement ratio. The particle density of kaolinite was
clay mineral mixtures, bentonite, mixed NSAPs and less than that of bentonite. Therefore, the void ratios
silicate sand, respectively. This equation shows the Sm excluding kaolinite increased with mixed NSAP
changes due to differences in the volume of the replacement ratio. For α = 75% and 100%, the kKC
component (i.e. α). Therefore, the calculated specific hydraulic conductivity values differed from the
surface area (Sm; Table 5) changes depending on the experimental k values by about one digit. This
difference in mixed NSAP type and α. When using N2 problem should be addressed in the future. On the
gas adsorption, only the outer surface area of montmor- other hand, for α = 0% and 25%, the kKC and k values
illonite may be determined. As water also enters the were similar. These results suggest that the hydraulic
interlayer region, the specific surface area of the conductivity may be estimated according to the mixed
bentonite was calculated using equation 6 (Komine & NSAP replacement ratio (i.e. degree of alteration)
Ogata, 1994, 1999b): when the proportion of clay minerals formed by
  alteration is small. These data provide a basis for
CM C evaluating the hydraulic conductivity of bentonite–
SB ¼ SM þ 1  M Sother (6)
100 100 sand mixtures with their alteration products in clay
radioactive waste-disposal systems.
where SB is the specific surface area per unit mass of
bentonite (= 388.8 m2/g), CM is the montmorillonite
content by percentage of weight (48%; Komine &
CONCLUSIONS
Ogata, 1999b), SM is the specific surface area of
montmorillonite in bentonite (= 810 m2/g; Komine & To clarify the effects of clay mineral type and content,
Ogata, 1999b) and Sother is the specific surface area of bentonite–sand–clay mineral mixtures and clay
component minerals, excluding montmorillonite, in mineral specimens were subjected to swelling-pressure
bentonite. Sother (i.e. ∼52%) is composed mainly of and permeability tests. These experiments produced
quartz, feldspars and carbonates. The specific surface the following results:
areas of these minerals are considered to be very small, as
represented by silica sand in Table 3. The assumed value (1) Swelling pressure and hydraulic conductivity
of <1 m2/g is consistent with Komine & Ogata (1999b). differed according to clay-mineral type. In
The calculated specific surface area was substituted into particular, the hydraulic conductivity of kao-
equation 4 to calculate hydraulic conductivity kKC and to linite, which is considered to have formed by
compare it with the experimental value k (Table 5). the alteration of bentonite, and those of mica
Table 5 and Fig. 9 show the hydraulic conductivities and chlorite were approximately three to five
(k, kKC), specific surface areas and void ratios of the orders of magnitude greater than the hydraulic
bentonite–sand–clay mineral mixtures. The specific conductivity of bentonite. Thus, alteration
surface area decreased with increasing mixed NSAP probably increases hydraulic conductivity.

FIG. 9. Comparison between hydraulic conductivities (k, kKC): (a) kaolinite, (b) mica and (c) chlorite mixtures. The
numerical values (Sm, e) in parts (a), (b) and (c) are the specific surface area (m2/g) based on equation 5 and the void
ratio, respectively.
 Effects of clay mineral type and content on hydraulic conductivity 731

(2) The swelling pressure of bentonite–sand–clay Cho W.J., Lee J.O. & Chun K.S. (1999) The temperature
mineral mixtures decreased with increasing effects on hydraulic conductivity of compacted
mixed NSAP replacement ratio, without major bentonite. Applied Clay Science, 14, 47–58.
difference in the various mixed NSAP types. Cuisinier O., Masrouri F., Pelletier M., Villieras F. &
The hydraulic conductivity of bentonite– Mosser-Ruck R. (2008) Microstructure of a com-
pacted soil submitted to an alkaline PLUME. Applied
sand–clay mineral mixtures increased with
Clay Science, 40, 159–170.
the mixed NSAP replacement ratio. These Holland H.J. & Murtagh M.J. (2000) An XRD morph-
findings suggest that an increase in clay ology index for talcs: the effect of particle size and
minerals due to alteration increases hydraulic morphology on the specific surface area. Advances in
conductivity. X-ray Analysis, 42, 421–428.
(3) The hydraulic conductivity values were Inoue A., Kohyama N., Kitagawa R. & Watanabe T.
similar to those calculated using the (1987) Chemical and morphological evidence for the
Kozeny–Carman equation and specific conversion of smectite to illite. Clays and Clay
surface area. These results suggest that Minerals, 35, 111–120.
hydraulic conductivity may be estimated Ishii T., Yahagi R., Owada H., Kobayashi I., Takazawa M.,
Yamaguchi K., Takayama Y., Tsurumi S. & Iizuka A.
from the clay-mineral content (i.e. degree of
(2013) Coupled chemical–hydraulic–mechanical
alteration). modelling of long-term alteration of bentonite. Clay
Minerals, 48, 331–341.
Our data provide a basis for the evaluation of the Ishikawa H., Shibata H. & Fujita T. (1994) Simulation of
swelling characteristics and permeability of engineered the thermal transformation of smectite to illite as the
barriers and backfills for the geological disposal of buffer material of radioactive waste disposal. Journal
radioactive waste. The relationship between the clay- of the Clay Science Society of Japan, 34, 149–156 (in
mineral structures or the interlayer bonding force and Japanese).
hydraulic conductivity will be clarified in future Ito M., Okamoto M., Shibata M., Sasaki Y., Danhara T.,
studies. Such research requires an accumulation of Suzuki K. & Watanabe T. (1993) Mineral Composition
Analysis of Bentonite. Power Reactor and Nuclear
data on the swelling pressures and hydraulic conduct-
Development Fuel Corporation, PNC TN8430 93-003
ivities of a large number of bentonite–sand–clay (in Japanese).
mineral mixtures and clay mineral samples. Japan Nuclear Cycle Development Institute (2000) H12:
Project to Establish the Scientific and Technical Basis
for HLW Disposal in Japan, Supporting Report 2,
ACKNOWLEDGEMENTS Repository Design and Engineering Technology.
Second Progress Report on Research and
This work was partly supported by Grants-in-Aid for Development for the Geological Disposal of HLW in
Scientific Research ‘KAKENHI’ (grant number Japan, JNC TN1410 2000-003.
16K21174) of the Japanese Society for the Promotion of Kobayashi I., Owada H., Ishii T. & Iizuka A. (2017)
Sciences (JSPS). The support is gratefully acknowledged. Evaluation of specific surface area of bentonite-
The authors thank Mr Sumi K. and Mr Otsuki R. from engineered barriers for Kozeny–Carman law. Soils
Tottori University for their assistance with the experi- and Foundations, 57, 683–697.
ments. The authors also thank two reviewers for their Komine H. & Ogata N. (1994) Experimental study on
helpful comments and advice. swelling characteristics of compacted bentonite.
Canadian Geotechnical Journal, 31, 478–490.
Komine H. & Ogata N. (1999a) Experimental study on
REFERENCES
swelling characteristics of sand–bentonite mixture for
nuclear waste disposal. Soils and Foundations, 39,
Achari G., Joshi R.C., Bentley L.R. & Chatterji S. (1999) 83–97.
Prediction of the hydraulic conductivity of clays using Komine H. & Ogata N. (1999b) Evaluation for Swelling
the electric double layer theory. Canadian Characteristics of Buffer and Backfill Materials for
Geotechnical Journal, 36, 783–792. High-Level Nuclear Waste Disposal – Influence of
Carman P.C. (1937) Fluid flow through a granular bed. Sand–Bentonite Content and Cation Compositions in
Transactions of the Institution of Chemical Engineers, Bentonite. Central Research Institute of Electric Power
15, 150–156. Industry Report, Abiko Research Laboratory Rep. No.
Chen B., Guo J. & Zhang H. (2016) Alteration of U99013 (in Japanese).
compacted GMZ bentonite by infiltration of alkaline Kozeny J. (1927) Ueber kapillare Leitung des Wassers im
solution. Clay Minerals, 51, 237–247. Boden. Sitzungsber. Akad. Wiss. Wien, 136, 271–306.
732 Masanori Kohno et al.

Mitchell J.K. & Soga K. (2005) Fundamentals of Soil Sivapullaiah P.V., Sridharan A. & Stalin V.K. (2000)
Behavior (3rd edition). John Wiley & Sons, Inc., Hydraulic conductivity of bentonite–sand mixtures.
Hoboken, NJ, USA. Canadian Geotechnical Journal, 37, 406–413.
Mollins L.H., Stewart D.I. & Cousens T.W. (1996) Suzuki H., Shibata M., Yamagata J., Hirose I. &
Predicting the properties of bentonite–sand mixtures. Terakado K. (1992) Characteristic Test of Buffer
Clay Minerals, 31, 243–252. Material (I). Power Reactor and Nuclear Development
Nakamura K., Tanaka Y. & Hironaga M. (2011) Survey on Fuel Corporation, PNC TN8410 92-057 (in Japanese).
Current Status of Laboratory Test Method and Suzuki K., Asano H., Yahagi R., Kobayashi I., Sellin P.,
Experimental Consideration for Material Containing Svemar C. & Holmqvist M. (2013) Experimental
Bentonite. Central Research Institute of Electric Power investigations of piping phenomena in bentonite-
Industry Report, Civil Engineering Research based buffer materials for an HLW repository. Clay
Laboratory Rep. No. N10026 (in Japanese). Minerals, 48, 363–382.
Nakayama S., Sakamoto Y., Yamaguchi T., Akai T., The Clay Science Society of Japan (2009) Handbook of
Tanaka T., Sato T. & Iida Y. (2004) Dissolution of Clays and Clay minerals (3rd edition). Gihodo
montmorillonite in compacted bentonite by highly Shuppan, Tokyo, Japan (in Japanese).
alkaline aqueous solutions and diffusivity of hydrox- Yamaguchi T., Sakamoto Y., Akai M., Takazawa M.,
ide ions. Applied Clay Science, 27, 53–65. Iida Y., Tanaka T. & Nakayama S. (2007)
Pusch R. (1999) Is Montmorillonite-Rich Clay of MX-80 Experimental and modeling study on long-term
Type the Ideal Buffer for Isolation of HLW? SKB alteration of compacted bentonite with alkaline
Technical Report, Rep. No. SKB-TR-99-33, Stockholm. groundwater. Physics and Chemistry of the Earth,
Ren X., Zhao Y., Deng Q., Kang J., Li D. & Wang D. 32, 298–310.
(2016) A relation of hydraulic conductivity–void ratio Yamaguchi T., Sawaguchi T., Tsukada M., Kadowaki M.
for soils based on Kozeny–Carman equation. & Tanaka T. (2013) Changes in hydraulic conductivity
Engineering Geology, 213, 89–97. of sand–bentonite mixtures accompanied by alkaline
Savage D., Noy D. & Mihara M. (2002) Modelling the alteration. Clay Minerals, 48, 403–410.
interaction of bentonite with hyperalkaline fluids. Yokoyama S., Nakamura K., Tanaka Y. & Hironaga M.
Applied Geochemistry, 17, 207–223. (2006) Review in Alteration of Bentonite under
Savage D., Walker C., Arthur R., Rochelle C., Oda C. & Alkaline Conditions. Central Research Institute of
Takase H. (2007) Alteration of bentonite by hyperalka- Electric Power Industry Report, Civil Engineering
line fluids: a review of the role of secondary minerals. Research Laboratory Rep. No. N05042 (in Japanese).
Physics and Chemistry of the Earth, 32, 287–297. Yokoyama S. & Nakamura K. (2010) Alteration Behavior
Sellin P. & Leupin O.X. (2013) The use of clay as an of Bentonite Barrier of Radioactive Waste Disposal by
engineered barrier in radioactive-waste management – Alkaline Solutions – Part 1: Permeability Change of
a review. Clays and Clay Minerals, 61, 477–498. Compacted Bentonite Immersed in Alkaline Solutions.
Shirazi S.M., Kazama H., Salman F.A., Othman F. & Central Research Institute of Electric Power Industry
Akib S. (2010) Permeability and swelling characteristics Report, Civil Engineering Research Laboratory Rep.
of bentonite. International Journal of the Physical No. N09015 (in Japanese).
Sciences, 5, 1647–1659.