Matric suction measurement 127

The preparation of the apparatus entails the following steps:

1. Remove the cover of the chamber and put in some de-­aired water that
is several cm above the ceramic disk. Replace the cover and secure it.
2. At the water reservoir outlet, connect a short tube with its free end
submerged in a beaker of water and apply air pressure not exceeding
the AEV of the ceramic disk. The water inside the chamber should flow
through the HAE ceramic disk into the water reservoir, exit through
the outlet tube into the beaker quickly, and eventually stop.
3. Monitor the outlet of the tube in the beaker for air bubbles for at least
10 minutes. It takes 10 minutes or more after the water has exited
from the water reservoir for air bubbles to start appearing. The pres-
ence of minute air bubbles suggests that the HAE ceramic disk is not
fully saturated. Repeat steps 1 to 3 to saturate the ceramic disk.
4. If steps 1 to 3 are not able to saturate the disk, connect the bottom
inlet tube to an external de-­aired water supply reservoir. Apply a low
vacuum pressure of 10–30 kPa at the air pressure inlet valve and
repeat steps 1–3. If step 4 also fails, it is highly likely that there is a
fine crack on the ceramic disk or the epoxy seal around the perim-
eter of the ceramic disk has broken in parts. When this happens, the
ceramic disk must be left to dry and then removed. The ceramic disk
is either replaced or removed, dried and re-­sealed into the water reser-
voir again. Steps 1–3 are again repeated.

When the ceramic disk is fully saturated, the null-­type axis translation appa-
ratus is considered prepared (ready) for matric suction measurement. For
matric suction measurement, the following steps are followed:

1. A representative soil specimen of diameter larger than the diameter

of the ceramic disk is placed onto the ceramic disk, and a 1 kg dead-
weight is placed on the soil specimen to ensure good contact between
soil specimen and ceramic disk.
2. Quickly replace the cover of the chamber and increase the air pressure
to 10 kPa.
3. Note the reading of the water pressure transducer at the bottom of the
water reservoir.
4. If the water pressure becomes increasing negative instead of positive,
increase the air pressure by a further 10 kPa and observe.
5. Repeat Step 4 until the water pressure stabilises at zero or small posi-
tive air pressure.
6. When the water pressure reading stabilises near zero or small positive
pressure, the matric suction of the soil specimen is given by the differ-
ence between air pressure and water pressure.

The operation of the null-­type axis translation apparatus is labour inten-

sive, as the air pressure needs to be adjusted frequently to avoid the water
128 Laboratory tests for unsaturated soils

pressure going too highly negative (> −60 kPa) and causing cavitation in the
water pressure measuring system. It is possible to automate the null-­type
axis translation apparatus to adjust the air pressure automatically when the
water pressure falls below a certain value using a feedback control loop,
as shown in Figure 8.5b. The algorithm for controlling the air pressure is
shown in Figure 8.5c. More details on automating the null-­type axis trans-
lation apparatus can be found in Leong et al. (2009). Typical responses of
the air and water pressure readings of the null-­type axis translation during
matric suction measurement are shown in Figure 8.6.

100 100
Pore-water pressure
Air pressure
80 80

Matric suction (kPa)

60 60
Pressure (kPa)

40 40

20 20

0 0

–20 –20
0 50 100 150 200 0 50 100 150 200
Time (min) Time (min)

(a)

400 400
Pore-water pressure
Air pressure
300 300
Matric suction (kPa)
Pressure (kPa)

200 200

100 100

0 0

–100 –100
0 50 100 150 200 0 50 100 150 200
Time (min) Time (min)

(b)

600 Pore-water pressure

600
Air pressure
500 500
Matric suction (kPa)

400 400
Pressure (kPa)

300 300
200 200
100 100
0 0
–100 –100
0 50 100 150 200 0 50 100 150 200
Time (min) Time (min)

(c)

Figure 8.6 A
 ctive control null-type axis translation apparatus test results (from
Leong et al. 2009). (a) K80S20 – w = 21.7%. (b) K80S20 – w = 14.5%.
(c) K80S20 – w =11.1%.
 Matric suction measurement 129

8.5 HYGROMETER

The direct methods to measure matric suction are always limited in range
due to the AEV of the ceramic disk which at present is limited to about
1,500 kPa. For matric suction beyond the residual suction of the soil, the
water phase is no longer continuous, and moisture transfer is only possible
through vapour. Provided there is negligible salt content in the pore water,
matric suction beyond the residual suction range is taken to be the same
as total suction. Experimental evidence shows that the water characteristic
curves of soils and filter papers based on matric suction and total suction
converge after suction of 1,000–3,000 kPa (Croney and Coleman 1948;
Houston et al. 1994; Leong et al. 2002).
The hygrometer is an instrument that measures relative humidity (see
Chapter 7). Instruments used in agronomy to measure relative humidity
include the psychrometer, dewcell and dewprobe sensors and infrared gas
analyser (Rosenberg et al. 1983). Previously, thermocouple psychrometers
have often been used for the measurement of total suction in soils (Brown
1970; Krahn and Fredlund 1972; Richards 1974; van der Radt et al. 1987;
Hamilton et al. 1979; Lee and Wray 1992; Ridley 1993). However, the psy-
chrometer is known to show fluctuations in relative humidity readings due
to non-­constant air temperature. Currently, the chilled mirror hygrometer
has become more common in the laboratory to measure total suction (Gee
et al. 1992; Bulut et al. 2002; Leong et al. 2003; Schanz et al. 2004).
The chilled mirror hygrometer is included as Method D in ASTM
D6836-­16 for determining the soil-­water characteristic curve from 1,000
kPa to 100,000 kPa. The chilled mirror hygrometer is also available com-
mercially for laboratory use. Hence, the chilled mirror hygrometer is the
only instrument considered here for the measurement of total suction.

8.5.1 Calibration
It is important to verify the calibration of the chilled mirror hygrometer. This
is usually performed with standard solutions. Saturated solutions of magne-
sium chloride (MgCl2.6H20), magnesium nitrate (Mg(NO3)2.6H2O), sodium
chloride (NaCl), potassium chloride (KCl), potassium nitrate (KNO3) and
potassium sulphate (K2SO4) are typically used as the standard solutions. The
relative humidity of the saturated solutions as given in ASTM E104-­20a and
the relative humidity of the standard saturated solutions measured using a
chilled mirror hygrometer, WP4, manufactured by Decagon Devices USA,
which has since merged with Meter Group Inc. USA, are shown in Table 8.4.
If the range of suctions to be measured is known, it is better to use solu-
tions within the range to check the calibration of the chilled mirror hygrom-
eter. Various molal solutions of sodium chloride (NaCl) and potassium
chloride (KCl) can be prepared to generate the required relative humidity.
Procedures for preparing such salt solutions are given in ASTM E104-­20a.
130 Laboratory tests for unsaturated soils

Table 8.4 Relative humidity from ASTM E104-20a of various solutions at 25°C

Relative humidity from Measured relative

Solutions ASTM E104-­20a (%) humidity by WP4 (%)
Magnesium Chloride (MgCl2. 6H2O) 32.8 ± 0.2 32.7
Magnesium Nitrate 52.9 ± 0.2 53.0
[Mg(NO3)2. 6H2O]
Sodium Chloride (NaCl) 75.3 ± 0.1 75.3
Potassium Chloride (KCl) 84.3 ± 0.3 84.3
Potassium Nitrate (KNO3) 93.6 ± 0.6 93.9
Potassium Sulphate (K2SO4) 97.3 ± 0.5 97.7
Pure Distilled Water 100 99.9

Usually, the manufacturer would specify the calibration check needed

before using their instrument. For the WP4, a one-­point calibration check is
recommended, and it would be better for the one-­point to be within the suc-
tion range of the soil specimen to be measured. The WP4 will be described
here, as it is the most common chilled mirror hygrometer used in geotechni-
cal engineering. Similar chilled mirror hygrometers are available from other
manufacturers but not specifically for use with soil samples. The decision to
describe the WP4 here does not represent any endorsement of the product
but is more of a case where there is more cumulative experience gained from
using the WP4. The WP4 has gone through several upgrades. The current
model is the WP4C, which has a measurement range of 33 kPa to 30,000
kPa. However, experiences of the measurement accuracy were mixed for
suctions below 1,000 kPa and sometimes below 3,000 kPa, as the data
points may not merge into the SWCC from the pressure plate test. The accu-
racy of the suction measurement is 50 kPa for suctions less than 5,000 kPa,
and 1% for suctions from 5,000 to 30,000 kPa. The WP4C does not have
an option to provide the relative humidity reading. Hence, if relative humid-
ity reading is needed, Equation 7.5 can be used to convert the total suction
measured to the relative humidity value. The operating temperature range is
from 5°C to 40°C. Temperature during measurement can be set within
5°C–40°C. Temperature resolution is 0.1°C and accuracy is ±0.2°C. The
manufacturer claims that the WP4C can give a reading within 5 minutes in
the quick mode and between 10 and 15 minutes in the precise mode. It has
been found that some soil specimens can take longer than 15 minutes to
reach equilibrium. Readings are updated every 5 minutes until equilibrium.

8.5.2 Test procedures
The WP4 is shown schematically in Figure 8.7. The WP4 is a standalone
unit with a sample drawer, an LCD display panel and buttons. The sample
drawer is slid out to receive the soil specimen container. The soil specimen
 Matric suction measurement 131

OPEN/CLOSE

READ

(a) (b)

Figure 8.7 Schematic drawing of WP4. (a) Schematic drawing. (b) Front view.

container is either plastic or steel. The steel container is supposed to give

better accuracy, as it allows the soil specimen to reach temperature equilib-
rium faster. The container is filled with the soil till half full and then placed
into the slot of the drawer for the soil specimen container. The drawer is
then pushed back into the unit, and the knob is turned to lift the soil speci-
men container to contact a seal at the top of the chamber. The sealed cham-
ber is supposed to be airtight, and the temperature of the chamber can be
controlled within 5°C and 40°C. A small internal fan circulates the air in
the air space above the soil specimen to reduce equilibrium time. The WP4
will continuously record the relative humidity of the air space above the
soil specimen and the surface temperature of the soil specimen displaying
the reading every 5 minutes until equilibrium. The chamber contains a mir-
ror whose temperature is precisely controlled by a thermoelectric (Peltier)
cooler and detection of the instance at which condensation first appears on
the mirror is by a beam of light directed onto the mirror and reflected into
a photodetector cell. The purpose of the photodetector cell is to sense the
change in reflectance when condensation first occurs on the mirror so that
the temperature at which it occurs can be recorded. The temperature is given
by a thermocouple attached to the mirror. Contamination of the mirror by
either spillage of soil specimen or condensation in the sample chamber can
lead to inaccurate readings.

8.5.3 Precaution
The chilled mirror hygrometer when measuring the suction of an aque-
ous solution gives the osmotic suction of the aqueous solution. When it is
measuring the suction of a soil specimen, it is measuring both matric and
osmotic suctions. It is a common practice to infer the third suction compo-
nent when two other suction components are measured (e.g., Tsai and Petry
1995; Tang et al. 2002). The chilled mirror hygrometer should not be used
to infer either matric or osmotic suction when either osmotic or matric suc-
tion is measured using other methods, as the error can be large (Leong et al.
132 Laboratory tests for unsaturated soils

1200

1000
1:1
Total Suction - Osmotic Suction (kPa)

800

Tap Water
600 NaCl 7.6 g/l
NaCl 16.0 g/l
400 NaCl 23.5 g/l
NaCl 32.5 g/l
200

0
0 200 400 600 800 1000 1200

–200

–400

–600
Matric Suction (kPa)

Figure 8.8 C
 omparison of inferred matric suction (total suction – osmotic suc-
tion) and measured matric suction for compacted soil samples using
different solutions [Data from Leong et al. (2009)].

2007). A plot comparing the inferred matric suction given by (total suction
– osmotic suction) and measured matric suction for compacted residual soil
samples using different solutions is shown in Figure 8.8. In Figure 8.8, the
total suction was measured using the chilled mirror hygrometer, the matric
suction was measured using a null-­type axis translation apparatus and the
osmotic suction was determined from the electrical conductivity of the solu-
tion used to compact the soil samples. Figure 8.8 shows that the errors are
large even for soil samples compacted using tap water.

REFERENCES

Blight, G. E. (1966). Strength characteristics of desiccated clays. Journal of the Soil

Mechanics and Foundations Division, 92(6): 18–38.
Bulut, R., Hineidi, S. M., & Bailey, B. (2002). Suction measurements—filter paper
and chilled mirror psychrometer. In Proceedings of the Texas Section American
Society of Civil Engineers, Fall Meeting, Waco, 2–5 October.
 Matric suction measurement 133

Cassel, D. K. & Klute, A. (1986). Water potential: Tensiometry. Methods of soil anal-
ysis, Part 1, Physical and mineralogical methods, 2nd ed., A. Klute, Ed. Madison,
WI: American Society of Agronomy, Soil Science Society of America.
Croney, D. & Coleman, J. D. (1948). Soil thermodynamics applied to the move-
ment of moisture in road foundation. In Proceedings of the 7th International
Conference on Applied Mechanics, Vol. 3, pp. 163–167.
Fredlund, D. G. (1973). Volume change behaviour of unsaturated soils. University of
Alberta, Edmonton, Alta, Canada, vol. Ph. D., p. 490.
Fredlund, D. G. & Rahardjo, H. (1993). Soil mechanics for unsaturated soils. New
York: Wiley.
Gee, G., Campbell, M., Campbell, G., & Campbell, J. (1992). Rapid measurement
of low soil potentials using a water activity meter. Soil Science Society of America
Journal, 56: 1068–1070.
Guan, Y. & Fredlund, D. G. (1997). Use of the tensile strength of water for the
direct measurement of high soil suction. Canadian Geotechnical Journal, 34(4):
604–614.
He, L., Leong, E. C., & Elgamal, A. (2006). A miniature tensiometer for measure-
ment of high matric suction. In G. A. Miller (Ed.), 147 edn. Unsaturated soils
2006. Reston Va, [Great Britain]: American Society of Civil Engineers, vol. 2, pp.
1897–1907.
Houston, S. L., Houston, W. N., & Wagner, A.-M. (1994). Laboratory filter paper
suction measurements. Geotechnical Testing Journal, 17(2): 185–194.
Leong, E. C., He, L., & Rahardjo, H. (2002). Factors affecting the filter paper
method for total and matric suction measurements. Geotechnical Testing Journal,
25(3): 321–332.
Leong, E. C., Lee, C. C., & Low, K. S. (2009). An active control system for matric
suction measurement. Soils and Foundation, 49(5): 807–811.
Leong, E. C., Widiastuti, S., Lee, C. C., & Rahardjo, H. (2007). Accuracy of suction
measurement. Geotechnique, 57(6): 547–556.
Lourenço, S. D. N., Gallipoli, D., Toll, D. G., & Evans, F. D. (2009). On the mea-
surement of water pressure in soils with high suction tensiometers. Geotechnical
Testing Journal, 32(6): 1–7.
Rahardjo, H. (1990). The study of Undrained and Drained Behaviour of Unsaturated
Soil. In Department of Civil Engineering. University of Saskatchewan, Saskatoon,
vol. Ph.D.
Ridley, A. M. & Burland, J. B. (1993). A new instrument for the measurement of soil
moisture suction. Géotechnique, 43(2): 321–324.
Ridley, A. M. & Burland, J. B. (1999). Use of the tensile strength of water for the
direct measurement of high soil suction: Discussion. Canadian Geotechnical
Journal, 36(1): 178–180.
Schanz, T., Agus, S. S., & Tscheschlok, G. (2004). Determination of hydro-­
mechanical properties of Trisoplast. Research Report Bo-­015/03. Laboratory of
Soil Mechanics. Weimar, Germany: Bauhaus-­University Weimar.
Tang, G. X., Graham, J., Blatz, J., Gray, M., & Rajapakse, R. K. N. D. (2002).
Suctions, stresses and strengths in unsaturated sand–bentonite. Engineering
Geology, 64(2–3): 147–156.
Tarantino, A. & Mongiovi, L. (2002). Design and construction of tensiometer
for direct measurement of matric suction. In Proceedings of 3rd International
Conference on Unsaturated Soils, Recife, vol. 1, pp. 319–324.
134 Laboratory tests for unsaturated soils

Tsai, C.-H. & Petry, T. M. (1995). Suction study on compacted clay using three mea-
surement methods. Transportation Research Record No, 1481: 28–34.
Wijaya, M. (2017). Compression, shrinkage and wetting-­induced volume change of
unsaturated soils. In Civil and Environmental Engineering. Singapore: Nanyang
Technological University, vol. Ph.D.
Wijaya, M. & Leong, E. C. (2016). Performance of high-­capacity tensiometer in
constant water content oedometer test. International Journal of Geo-­Engineering,
7(1): 13.
Chapter 9

Indirect suction
measurement methods

9.1 INTRODUCTION

In this chapter, several indirect suction measurement methods are described.

All indirect suction measurement methods use one common measurement
principle, that is, the use of a proxy to provide the suction value. The indi-
rect methods described are the more common and accessible methods that
can be readily adopted by a soil testing laboratory. These methods include
the filter paper method, thermal conductivity sensor, electrical resistance
sensor, capacitive sensor and electrical conductivity measurements. The suc-
tions measured are total suction (filter paper), matric suction (filter paper,
thermal conductivity sensor, electrical resistance sensor, capacitive sensor)
and osmotic suction (electrical conductivity).

9.2 RELATED STANDARDS

1. ASTM D1125 Test Methods for Electrical Conductivity and Resistivity

of Water.
2. ASTM D4542-­ 15 Test Methods for Pore Water Extraction and
Determination of the Soluble Salt Content of Soils by Refractometer.
3. ASTM D5298–16. Standard Test Method for the Measurement of Soil
Potential (Suction) Using Filter Paper.
4. ASTM D6836–16. Standard Test Methods for Determination of
the Soil Water Characteristic Curve for Desorption Using Hanging
Column, Pressure Extractor, Chilled Mirror Hygrometer, or Centrifuge.
5. ASTM E230/E230M Specification and Temperature–Electromotive
Force (emf) Tables for Standardized Thermocouples.
6. ASTM E337 Test Method for Measuring Humidity with a Psychrometer
(the Measurement of Wet-­and Dry-­Bulb Temperatures).
7. ASTM E832-­81 (2019), Standard Specification for Laboratory Filter
Papers.

DOI: 10.1201/b22304-9 135

136 Laboratory tests for unsaturated soils

9.3 FILTER PAPER METHOD

The use of filter paper as a proxy to determine matric and total suction was
developed in soil science and has been standardised for measurement of soil
potential (suction) in ASTM D5298-­16 (2016). The test procedure is simple,
straightforward and does not require any special equipment. It involves the
measurement of the water content of the filter paper that has equilibrated
with a soil specimen in a closed container. It is generally accepted that when
the filter paper is placed in direct contact with the soil, moisture transfer
is by capillary flow and matric suction is measured. When the filter paper
is not in contact with the soil, moisture transfer is by vapour transfer. The
filter paper comes to equilibrium with the partial pressure of water vapour
in the air of the container enclosing the filter paper and the soil, measuring
total suction.
Gardner (1937) was probably the first to use filter paper to measure soil
suction. The filter paper used was Schleicher & Schuell No. 589 (S&S
589) white ribbon filter paper. Calibration of the filter paper was first
done by equilibrating both wet and dry filter papers above sulphuric acid
solutions in closed containers for low water contents and then transfer-
ring the equilibrated filter papers to a centrifuge to obtain the desired
suctions for high water contents. Since then, the filter paper method for
suction measurement has become more widely used due to the accessibil-
ity of filter papers. The two most common filter papers used for suction
measurement are Schleicher & Schuell No. 589 filter paper (McQueen and
Miller 1968b; Al-­Khafaf and Hanks 1974; McKeen 1980; Harrison and
Blight 1998; Leong et al. 2002) and Whatman No. 42 filter paper (Fawcett
and Collis-­George 1967; Hamblin 1981; Chandler and Gutierrez 1986;
Chandler et al. 1992; Harrison and Blight 1998; Deka et al. 1995; Leong
et al. 2002; Power et al. 2008). Other filter papers that have been used
include Whatman No. 1 filter paper (El-­Ehwany and Houston 1990),
Thomas Scientific No. 4705-­F10 filter paper (Miller and Nelson 1992)
and Fisher quantitative coarse (9.54 Å) filter paper (Houston et al. 1994).
In fact, any adsorbent materials besides filter paper can be used. Sibley
and Williams (1990) evaluated five different absorbent materials
(Whatman No. 42, unwashed dialysis tubing, washed dialysis tubing,
Millipore MF 0.025 pm and Millipore 0.05 pm) for suction measurements
and concluded that the Whatman No. 42 filter paper is the most appropri-
ate for suction measurement.
ASTM D5298-­16 (2016) specifies that the filter paper used must be ash-­
free quantitative Type II filter papers (ASTM E832-­81 2019). Examples of
ash-­free quantitative Type II filter papers are Whatman No. 42, Fisherbrand
9-­790A and Schleicher & Schuell No. 589 White Ribbon.
 Indirect suction measurement methods 137

9.3.1 Calibration
It is recommended that the calibration curve of filter paper (for a batch) be
obtained before use (Likos and Lu 2002; Leong et al. 2002) and the filter
paper calibrated according to the method (contact or non-­contact) that it is
used (Ridley and Wray 1996; Power et al. 2008). ASTM D5298-­16 (2016)
suggests that the calibration curve can be obtained by suspending a filter
paper over a salt solution of various concentrations (minimum 50 mL) using
an improvised platform of inert material such as plastic tubing or stainless-­
steel screen. This method of calibration corresponds to the non-­contact filter
paper method that measures total suction. Houston et al. (1994), Bulut et al.
(2002), Leong et al. (2003) and Power et al. (2008) found that the calibration
curves for the contact method which measures matric suction and the non-­
contact method which measures total suction are different. Discrepancies
between suctions measured using contact and non-­contact methods are
more severe in the low suction range, <100 kPa. Bulut and Wray (2005),
and Marinho and Oliveira suggest that the difference is due to insufficient
equilibrium time for the non-­contact filter paper. As the equilibrium time is
increased, the suction measured using the non-­contact method approaches
slowly towards the suction measured using the contact method but never
reaches it even after an excessively long equilibrium time. Such excessive
equilibrium time is not suitable for practical purposes. Hence, it is not pos-
sible to calibrate filter paper for the low suction range (<100 kPa) using
the non-­contact method. In a footnote, ASTM D5298-­16 (2016) mentioned
that the filter paper may be calibrated by using the pressure membrane for
the range 100–1,500 kPa and the ceramic plate for the range 10–100 kPa.
Leong et al. (2020) calibrated the Whatman No. 42 filter paper in the low
suction range (< 16 kPa) using the Tempe cell and capillary rise method.
Fawcett and Collis-­ George (1967) calibrated Whatman No. 42 filter
papers, treated with 0.005% mercury chloride (HgCl2), with soil samples
from a soil column at equilibrium with a water table at suctions of 1 and 10
kPa. Hamblin (1981) calibrated Whatman No. 42 filter papers using a suc-
tion plate apparatus for suction up to 7 kPa. Deka et al. (1995) calibrated
the Whatman No. 42 filter paper using a tension table for suctions from 1
to 65 kPa. The principles of a tension table are similar to those of a suction
plate or the hanging column (Method A) described in ASTM D6836-­16. It
is possible to apply a maximum suction of up to 80 kPa using a tension table
or hanging column with a vacuum pump (Parcevaux 1980). Leong et al.
(2002), Power et al. (2008) and Elgabul (2013) calibrated Whatman No. 42
filter papers in a pressure plate for matric suctions less than 1,000 kPa.
Most of the calibration curves were developed from initially dry filter
paper. Hence, the calibration equation usually corresponds to the wetting
curve of the filter paper. As the filter paper is thin, the amount of hysteresis
138 Laboratory tests for unsaturated soils

Table 9.1 C
 alibration curves for Whatman No. 42 and Schleicher & Schuell No. 589 filter
papers

(a) Whatman No. 42 filter paper

References Calibration curves*
Hamblin (1981) log ψ = 8.022 − 3.683 log wf
Chandler and Gutierrez (1986) log ψ = 4.84 − 0.0622 wf wf < 47
Chandler et al. (1992) log ψ = 6.05 − 2.48 log wf wf ≥ 47
Greacen et al. (1987) log ψ = 5.327 − 0.0779 wf wf < 45.3
ASTM D5298-­16 log ψ = 2.412 − 0.0135 wf wf ≥ 45.3
Leong et al. (2002) log ψ = 4.945 − 0.0673 wf wf < 47
log ψ = 2.909 − 0.0229 wf wf ≥ 47

(b) Schleicher & Schuell No. 589 filter paper

References Calibration curves*
McQueen and Miller (1968) log ψ = 5.238 − 0.0723 wf wf < 54
log ψ = 1.8966 − 0.01025 wf wf ≥ 54
Al-­Khafaf and Hanks (1974) log ψ = 4.136 − 0.0337 wf wf < 85
log ψ = 2.0021 − 0.009 wf wf ≥ 85
McKeen (1980) log ψ = 4.9 − 0.0624 wf wf < 66
log ψ = 1.25 − 0.0069 wf wf ≥ 66
Greacen et al. (1987) log ψ = 5.056 − 0.0688 wf wf < 54
ASTM D5298-­16 log ψ = 1.882 − 0.0102 wf wf ≥ 54
Leong et al. (2002) log ψ = 5.438 − 0.069 wf wf < 54
log ψ = 2.659 − 0.018 wf wf ≥ 54

* Note: ψ = suction in kPa, wf = filter paper water content in %

between the drying and wetting filter paper is small in the range of about
2% of the water content (Leong et al. 2002).
Many calibration equations have been proposed in the literature for the
filter paper. A summary of the calibration equations is listed in Table 9.1. If
the filter paper used is not calibrated, it is recommended to use that given in
ASTM D5298-­16 (Leong et al. 2007).

9.3.2 Test procedures
As the water content of the filter paper is used to determine suction, any
accidental transfer of fluid to the filter paper during water content determi-
nation will cause inaccurate measurement and, hence, wrong suction value.
The filter paper should not be handled with bare hands; vinyl surgical non-­
powdered or similar gloves are recommended to be used when handling tools
or containers during the filter paper water content determination (Bulut and
Leong 2008). A pair of tweezers should be used for picking up, placing or
transferring the filter paper. Hence, a strict protocol is needed when using
the filter paper method to obtain a reliable suction measurement.
 Indirect suction measurement methods 139

Although filter paper size is not rigidly fixed, the most used filter paper
size is 55 mm in diameter. Such a filter paper weighs approximately 0.2 g.
ASTM D5298-­16 (2016) requires the weight of the filter paper to be weighed
to a resolution of 0.0001 g. The filter paper from the box usually has some
initial moisture content of about 3%–6% (Leong et al. 2020). ASTM
D5298-­16 (2016) recommended that the filter papers be dried for at least 16
hours or overnight in a drying oven and placed in a desiccator over a desic-
cant for storage before use. This is not necessary if the filter paper will be
wetted to water content greater than 6% during moisture equilibration with
the soil specimen. If organism growth or biological decomposition of the
filter paper is expected, the filter paper can be pre-­treated by dipping the
filter paper into a 2% concentration of formaldehyde prior to use (ASTM
D5298-­16, 2016). Biological decomposition is possible if the filter paper is
in a warm and moist environment for more than 14 days. Mercury chloride
(HgCl2) (Fawcett and Collis-­George 1967; Hamblin 1981) and pentachlo-
rophenol solutions (McQueen and Miller 1968; Al-­Khafaf and Hanks 1974;
Bulut 1986) have been used to pre-­treat filter papers before use. However,
such treatment may alter the characteristics of the filter paper and hence its
calibration curve. Pre-­treated filter papers should be used with caution.
The soil specimen used should be as large as possible. ASTM D5298-­16
(2016) suggests that the mass of the soil specimen be between 200g and
400g. For total suction measurement, the soil specimen can be fragmented,
and the filter paper placed above the soil specimen with a spacer which can
be a wire screen, O-­ring or other inert material with minimum surface area
between the filter paper and the soil (Figure 9.1a). For matric suction mea-
surement, the whole soil specimen should be used, and the filter paper can
be sandwiched between two halves of the specimen (Figure 9.1b). In both
cases, the soil specimen shall be placed in an airtight container made of
rust-­free metal or glass with minimal air space to reduce equilibration time
and to minimise suction changes in the specimen. ASTM D5298-­16 (2016)
specifies two filter papers to be used in the total suction measurement and
three filter papers to be used in the matric suction measurement. For the
matric suction measurement, a filter paper is sandwiched between two
slightly larger diameter filter papers (about 3–4 mm larger in diameter) to
prevent soil contamination. The use of more than one filter paper will
increase equilibration time and may have the unintended consequence of
the filter paper not reaching suction equilibrium with the soil specimen
(Leong et al. 2002).
The lid of the airtight container can be additionally sealed using tape and
placed in an insulated chamber such as a cooler box or Styrofoam box,
where temperature can be maintained within ±1°C when the variation of
the external temperature is ±3°C (ASTM D5298-­ 16). The equilibration
period is suggested to be seven days and is expected to be less in ASTM
D5298-­16 (2016). The length of equilibration time depends on the degree of
saturation of the soil specimen. Drier soil specimens may take a longer time.
140 Laboratory tests for unsaturated soils

Filter paper
Tape for Spacer (non-porous)
additional
sealing

Soil specimen
Airtight container

(a)

Tape for
additional
sealing Soil specimen

Filterpaper
Airtight container

Soil specimen

(b)

Figure 9.1 F ilter paper method for measuring soil suction. (a) Non-contact filter
paper. (b) Contact filter paper.

Experiments have shown that 14–15 days may be more appropriate (Leong
et al. 2002; Bulut and Wray 2005).
At the end of the equilibration period, the water content of the filter
paper is determined. The equipment needed for the water content determi-
nation should be prepared before removing the filter paper from the con-
tainer. The filter paper is transferred to a metal (aluminium stainless steel)
moisture container with lid using a pair of tweezers and gloved hands
quickly (within 3s–5s) for weight determination to the nearest 0.0001 g.
Plastic bags with ziplock may also be used in place of the metal moisture
container, but greater care needs to be exercised. First, weigh the ziplock
plastic bag to 0.0001 g. Place the filter paper quickly into the ziplock plastic
bag and press the plastic bag flat on a clean and dry surface to remove the
 Indirect suction measurement methods 141

air before zipping it for weighing to 0.0001 g. After weighing, transfer the
filter paper to an evaporation dish or moisture can for drying in the oven.
After drying, repeat the previous steps to obtain the dry weight of the filter
paper using a new ziplock plastic bag. The temperature of the oven should
be the same as for soil moisture container determination – i.e., 105°C–110°C.
If metal moisture container is used, the moisture container can be placed on
a metal block with a flat surface to increase the cooling rate (ASTM
D5298-­16). Metallic moisture container when hot is known to weigh less
than when cold due to adsorbed moisture and should be used with caution,
as it will introduce errors in the water content determination of the filter
paper.

9.3.3 Calculations
The final water content of the filter paper wf is determined using Equation 9.1.

mff  mf 0
wf  (9.1)
mff

where
mff = mass of filter paper after oven drying
mf0 = mass of filter paper before oven drying

Using wf, the suction ψ can be obtained using a calibration equation of

the form shown in Table 9.1 if a calibration of the filter paper was per-
formed or using the ASTM D5298-­16 equation (Leong et al. 2007).

9.3.4 Summary
It is recommended to use only the contact filter paper method to measure
matric suction. When matric suction exceeds 1,000 kPa, contact and non-­
contact filter paper methods will yield similar suction values (Leong et al.
2016). If the non-­contact filter paper method is used, it should only be used
for suction above 1,000 kPa, and the matric suction calibration equation
can be used to determine the suction.

9.4 THERMAL CONDUCTIVITY SENSOR

A thermal conductivity sensor (TCS) relies on the rate of heat dissipation of

a porous material as it equilibrates with the surrounding soil to determine
the matric suction. A typical TCS consists of a heater, a temperature-­sensing
unit and a ceramic block. An example of a TCS is the 229 Water Matric
Potential Sensor by Campbell Scientific shown in Figure 9.2. Comparison
142 Laboratory tests for unsaturated soils

Cable Shield (grey)

Copper wire + (blue)

Epoxy backing Constantan wire – (red)

60 mm
Heater + (black)

Ceramic block Heater – (green)

Stainless steel probe enclosing a heating element

and thermocouple temperature sensor

15 mm

Figure 9.2 S chematic representation of the 229 Water Potential Sensor from
Campbell Scientific.

Table 9.2 Comparison of 229 and FTC thermal conductivity sensors

Suction Resolution
Sensor Company range (kPa) (oC) Accuracy Dimensions
229 Campbell 10–2,500 ±0.1°C − 15 mm (D) ×
Scientific 60 mm (L)
FTC GCTS Testing 1–1,000 − 5% of reading 28 mm (D) ×
Systems (up to 500 kPa 60 mm (L)
in wetting)

of the 229 Water Matric Potential Sensor and the Fredlund thermal con-
ductivity (FTC) the pressure plate at equilibrium sensor from GCTS Testing
Systems is given in Table 9.2. The heater and temperature sensor are placed
inside a needle probe and embedded inside the ceramic block. The cables
extend out from the back of the block through a backing. The principle of
the TCS is that the thermal conductivities of air (= 0.026 Wm−1K−1), water
(≈0.60 Wm−1K−1) and the ceramic are different. The thermal conductivity
of the 229 Water Potential Sensor when dry is about 0.35 Wm−1K−1 (Reece
1996). To do a measurement, the heater is turned on for a fixed period of
time. The heat causes the temperature to rise in the ceramic block, and the
temperature rise ΔT is recorded by the temperature sensor. The amount of
heat generated by the heater is fixed by the duration of heating, but ΔT is
dependent on the amount of water present in the ceramic block as negli-
gible heat is dissipated by the air in the pores. An example of temperature
rise ΔT with a reduction of water content (i.e., increase in matric suction) is
shown in Figure 9.3. Therefore, thermal conductivity of the ceramic block
 Indirect suction measurement methods 143

Figure 9.3 T
 emperature rises in a TCS at different matric suctions. (from Leong
et al. 2011).

will increase with its water content, leading to a lower temperature rise ΔT.
When the TCS is embedded in the soil, the water content of the ceramic
block will equilibrate with the surrounding soil to reach matric suction
equilibrium. Hence, knowing the calibration curve of the ceramic block
enables the matric suction of the soil to be determined.

9.4.1 Calibration
The thermal transfer properties of a TCS depend on the interfaces between
the different components that make up the TCS: the interface between the
stainless-­steel probe and the ceramic, the arrangement of the heating ele-
ments and temperature sensor in the probe, the homogeneity of the ceramic
block. Differences in manufacturing quality of the TCS make it necessary to
calibrate each sensor individually for a more accurate reading (Fredlund and
Wong 1989; Reece 1996; Flint et al. 2002; Hu et al. 2007).
The TCS can be calibrated using a pressure plate apparatus. A soil mix-
ture is prepared and placed in the pressure plate with the TCS embedded
(e.g., Fredlund and Wong 1989; Reece 1996; Feng and Fredlund 1999;
Fredlund et al. 2000; Flint et al. 2002). Hu et al. (2007) calibrated without
a soil mixture using only kaolin paste between the tip of the FTC ceramic
block and a pressure plate. The matric suction is increased, and when the
reading of the TCS becomes stable over time, the reading of the TCS is
deemed to correspond to that of the applied matric suction. The air pressure
is then increased for the next matric suction calibration. This technique
allows several TCSs to be calibrated at the same time. Variation of how the
144 Laboratory tests for unsaturated soils

TCS reading is taken exists. For example, Flint et al. (2002) released the air
pressure of the pressure plate apparatus when the soil mixture came to equi-
librium before making measurements to preclude the effects of air pressure
on the TCS reading. Campbell (2006) thought that the heat transfer in the
ceramic block is highly dependent on the air pressure and hence releasing
the air pressure will allow the TCS to make measurements in the environ-
ment (atmospheric pressure) that it is used. Using this method, special
attachments to the pressure plate need not be made as the cables can be
left in the pressure plate apparatus during suction equilibrium. The cables
are connected to the readout unit for reading after air pressure is released.
However, using the pressure plate does not guarantee that the matric suc-
tion of the soil is the same as the applied matric suction. It has been
observed that the matric suction of soil in the pressure plate at equilibrium
is always lower than the applied matric suction (Gee et al. 2002; Creswall
et al. 2008).
When using the TCS, it is important to know the response time of the
TCS. Using the calibration procedures mentioned earlier will not allow
the response time to be evaluated as the TCS is following a drying curve
when air pressure is increased. An alternative procedure is to compact soil
samples at various water content to calibrate the TCS (Leong et al. 2009).
The TCS is embedded in the soil samples, and the response of the TCS
with time is recorded. The matric suctions of the compacted samples are
then determined by independent measurement, e.g., using a null-­type axis
translation apparatus or tensiometer (see Chapter 8). However, this
method of calibration is highly labour intensive and may not be suitable
in practice.
The following procedures are typically used when calibrating using a
pressure plate apparatus:

1. Saturate the ceramic block of the TCS by immersing the ceramic sen-
sor under water and apply a vacuum, the same way as saturating
porous stones for saturated soil tests.
2. Prepare a saturated soil mixture. Ideally, the soil which the TCS is to
be used with or kaolin may be used.
3. Prepare the pressure plate apparatus. Ensure that the ceramic plate is
fully saturated.
4. Place the soil mixture into a steel or plastic ring (diameter of at least
two times the diameter of the TCS and height of at least two times the
length of the TCS) and place it into the pressure plate.
5. Embed the TCS to be calibrated, ensuring that the ceramic block is
fully embedded into the soil.
6. Connect the wires to the readout unit. Special attachments to the pres-
sure plate are needed to allow the wires to be brought out of the pres-
sure plate.
7. Apply the lowest suction of the calibration range.
 Indirect suction measurement methods 145

8. Take measurements of the TCS periodically to determine if the equi-

librium condition has been reached. Close off the drainage valve of the
pressure plate and release the suction before making a TCS measure-
ment. The reading of the TCS is usually performed at a fixed heating
time, e.g., for the TCS 229, this is at the 30s.
9. When equilibrium at the applied suction is attained, apply the next
suction and repeat steps 7 and 8 until the upper limit of the calibration
range.

Suitable functions to describe the calibration curves are exponential or

power functions as given in Equations 9.2 and 9.3, respectively.

  exp  a.T  b  (9.2)

  a.T b (9.3)

where a and b are curve-­fitting parameters.

The fitting parameters in Equations 9.2 and 9.3 can also be determined
using normalised temperature change to be between 0 and 1 where the nor-
malised temperature change ∆Tnorm is given by Equation 9.4.

Tdry  T
Tnorm  (9.4)
Tdry  Twet

where ∆Tdry is the temperature change when the TCS ceramic block is
dry, and ∆Twet is the temperature change when the TCS ceramic block is
saturated.

9.4.2 Test procedures
The TCS is normally used in the field but it is useful in a model labora-
tory test or as an alternative to measure matric suction of soil samples. The
TCS is first saturated and left with its ceramic block in water. To make a
suction measurement, a pre-­bored hole about the size of the TCS is first
made into the soil sample. The TCS is then inserted and the wires are con-
nected to a readout unit. The TCS and soil sample should be covered or
placed in an airtight container to prevent the soil sample from drying dur-
ing measurement. Take readings periodically until there is no change in the
reading. As the TCS starts from the fully saturated condition, the soil sam-
ple should be large enough such that its suction is not greatly affected by
moisture exchange between the soil sample and TCS during measurement.
Equilibration time depends on the suction of the soil sample, and the con-
tact condition between the TCS and the soil sample (Leong et al. 2011). The
matric suction is then obtained from the calibration equation.
146 Laboratory tests for unsaturated soils

9.5 ELECTRICAL RESISTANCE SENSORS

Electrical resistance sensors (ERSs) rely on measuring the electrical resis-

tance of a porous material as it equilibrates with the surrounding soil to
determine the matric suction. The development of the ERS dates back to
1940 (Bouyoucos and Mick 1940). Materials that have been used as porous
materials include gypsum, fibreglass or nylon (Scalon et al. 2002). Gypsum
is commonly used in early ERSs, and such sensors are still being manufac-
tured commercially e.g., 227-­L Soil Matrix Potential Block manufactured
by Delmhorst. The ERS suffers from problems ranging from sensitivity to
soil salinity and temperature, hysteresis, stability of calibration curve and
durability of the block.
Another ERS is the Watermark Model 200 sensor manufactured by the
Irrometer Company, Inc. A schematic drawing of the Watermark Model 200
sensor is shown in Figure 9.4a. It consists of two concentric stainless-­steel
electrodes embedded in a uniform granular matrix containing a gypsum
wafer encased in a synthetic membrane and covered by a perforated
stainless-­steel shell with plastic end caps (Figure 9.4b). The perforated
stainless-­steel shell helps to ensure that the granular matrix is packed more
uniformly automatically under uniform pressure. Both the synthetic mem-
brane and the perforated stainless-­steel shell make the sensor more robust
for handling. The gypsum wafer buffers against salinity variations less than
0.2 S/m (electrical conductivity of saturated gypsum). The upper matric suc-
tion limit of the sensor is controlled by the air-­entry value (AEV) of the
granular matrix as the matrix remains saturated above the AEV. For
Watermark sensors, the lowest suction limit is about 10 kPa (Armstrong and
Thomson 1985; Thomson and Armstrong 1987; Spaans and Baker 1992).
The highest matric suction limit of the sensor depends on the smaller pore

Sensor matrix

Synthetic membrane PVC Concentric

enclosing granular matrix tube electrodes
Gypsum
Screw
Perforated steel shell Transmission
matrix

(a) (b)

Figure 9.4 S chematic drawing of Watermark Model 200 sensor. (a) Schematic
drawing. (b) Section view.
 Indirect suction measurement methods 147

Table 9.3 Comparison of 227-L and Watermark 200 ERSs

Suction Operating
range Physical quantities temperature
Sensor Company (kPa) (mm, g) Accuracy range
Gypsum Delmhorst 10–1,000 22.5 (D) × 28.6 ±(10% of reading > 0°C
Block (L), 180 + 2 kPa) from
−100 to −5 kPa
Watermark Irrometer 0–200 0 – 80°C
Model 200 Sensors

sizes of the granular matrix which has been specified by the manufacturer as
239 kPa. Comparison of the 227-­L and Watermark 200 sensors is given in
Table 9.3.

9.5.1 Calibration
Scalon et al. (2002) recommend that ERS should be soaked for about 24
hours and then oven-­dried two to three times before calibration and use.
The calibration equation given by the manufacturer should be verified even
though some ERSs are marketed as “identical” with only a single calibra-
tion equation independent of soil types. For Watermark Model 200 sensors,
McCann et al. (1992) and Spaans and Baker (1992) found their calibration
equation to be different from that given by the manufacturer. In addition,
Spaans and Baker (1992) found poor repeatability of the calibration curve.
Hence, for better accuracy, each sensor should be calibrated individually
before use.
The ERSs can be calibrated using a pressure plate apparatus (Thomson
and Armstrong 1987; McCann et al. 1992). Sensors are embedded in a satu-
rated soil in the pressure plate, and the drying curve is measured by increas-
ing the air pressure incrementally similar to the procedures described for
TCS. Alternatively, the sensors can be calibrated using a direct matric suc-
tion measurement device like the tensiometer (Spaans and Baker 1992) by
embedding the ERSs and a tensiometer in the same soil.
Temperature effects are known to affect the accuracy of ERS. Generally,
the correction for temperature effects is about 3% (Campbell and Gee 1986;
Thomson and Armstrong 1987; Wang 1988). Equation 9.5 by Campbell
and Gee (1986) may be used to correct for temperature effects.

Rc  Rm 1  0.03  Tc  Tm   (9.5)

where
Rc = corrected resistance for calibration temperature Tc
Rm = measured resistance at temperature Tm
148 Laboratory tests for unsaturated soils

The Model 200SS Watermark Model 200SS sensor relies on the calibra-
tion equation (Equation 9.6) of the form given by Thomson and Armstrong
(1987).

 20R  11 R  1 k
 ua  uw    3 2
0.00279R  0.190109R  3.71485R  6.73956 R  1 k
(9.6)

9.5.2 Test procedures
The test procedures using the ERSs are similar to that for the TCS. ERSs
should be saturated first before use following the saturation procedures
mentioned earlier. For example, the Watermark Model 200 sensors’ granu-
lar matrix is initially filled with air. The air in the granular matrix must
be replaced by water to obtain correct readings. An alternative saturation
procedure for the Watermark Model 2000 sensors is provided by Barani
(2021):

1. Submerge the sensor to less than halfway for 30 minutes to allow cap-
illary action to draw the water into the inner pores.
2. Let it dry for about 6–8 hours.
3. Repeat Steps 1 and 2 two more times.
4. Finally, submerge the sensor fully in water for about 6–8 hours.
5. The sensor should be fully saturated and is now ready for use.

To make a suction measurement on a soil sample, a pre-­bored hole about the

size of the ERS is first made into the soil sample. The ERS is then inserted,
and the connections to a readout unit are made. The ERS and soil sam-
ple should be covered or placed in an airtight container to prevent the soil
sample from drying during measurement. Take readings periodically until
there is no change in the reading. As the ERS starts from the fully saturated
condition, the soil sample should be large enough such that its suction is not
greatly affected by any moisture exchange between the soil sample and the
ERS during measurement. Equilibration time depends on the suction of the
soil sample and the contact condition between the ERS and the soil sample
(Leong et al. 2011). The matric suction is then obtained from the calibration
equation.

9.6 CAPACITANCE SENSOR

Capacitance sensors (CSs) are similar to the TCS and ERS in that they
rely on a porous matrix to provide a repeatable matric suction reading.
The difference is that CS relies on measuring the dielectric permittivity of
the porous material as it equilibrates with the surrounding soil to deter-
mine the matric suction. The CSs make use of the difference in dielectric
 Indirect suction measurement methods 149

permittivity ε of air (1) and water (80) in the pore space. The ε of dry soil
is about 3 to 4 and that of saturated soil is about 40–50. The ε of ice is 5.
Dielectric permittivity refers to the ability to hold a charge and does not
measure water content directly. There are several different types of sensors
that make use of dielectric permittivity to measure water content in soil.
Examples are time-­domain reflectometry (TDR), standing-­wave technique
and CSs for soil moisture. Early suggestions of using dielectric permittiv-
ity for matric suction measurement use TDR (Noborio et al. 1999; Or and
Wraith 1999).
An example of a CS is the Teros-­21 sensor manufactured by the Meter
Group Inc. that is shown schematically in Figure 9.5. The width, length and
thickness of the sensor are 35 mm, 96 mm and 15 mm, respectively. The
suction measurement range is 5–100,000 kPa with a resolution of 0.1 kPa,
and the operating temperature range is −40°C–60°C. The accuracy of the
sensor is ±(10% of reading +2 kPa) for a suction range of 9–100 kPa. It
consists of a solid-­state circuit board sandwiched between two 8-­mm thick
and 32-­mm in diameter ceramic wafers (porous matrix). Each ceramic wafer
is protected by 0.3-­mm-­thick perforated stainless-­steel plates on the outer
surface that are grounded. The perforated stainless steel serves two pur-
poses: to protect the ceramic wafer and as an electromagnetic shield, as the
sensor uses an alternating current of 70 MHz to charge. The charge-­release
time is measured by the circuit. In principle, the electromagnetic field is con-
fined to within the ceramic wafers and, hence, the sensor is independent of
soil type. Effects of salinity of the soil do not affect the sensor reading, as
long as it is below 10 dS/m. The sensor is factory calibrated from 20 to 80
kPa and, hence, is most sensitive in this suction range. It is less sensitive
beyond 500 kPa due to the water retention characteristic of the ceramic
wafer. According to the manufacturer, the amount of hysteresis between 20
and 100 kPa suction is less than 10 kPa.

Vinyl filled with

polyurethane resin with
temperature sensor below

Printed solid-state circuit board

Perforated steel shell

Figure 9.5 Schematic drawing of Teros-21 capacitance sensor.

150 Laboratory tests for unsaturated soils

9.6.1 Calibration
The capacitance sensor has seen several improvements since it was first
made commercially. For example, the current Teros-­21 Gen 2 has shown
improvements in design from its predecessors, MPS-­1, MPS-­2, MPS-­6 and
Teros-­21 Gen 1. The suction measurement range of MPS-­1 and MPS-­2 is
from 10 to 500 kPa, MPS-­6 is from 9 to 100,000 kPa and Teros-­21 is from 5
to 100,000 kPa. From MPS-­2 onwards, measurement is done using 70 MHz
instead of 5 MHz, as it was found that more stable readings were obtained
at the higher frequency. The MPS-­1 has a one-­point factory calibration, the
MPS-­2 has a two-­point factory calibration and the MPS-­6 has six-­point
factory calibration (Decagon 2015). Since MPS-­6, all later sensors have
six-­point factory calibrations done between 20 and 80 kPa suction. Hence,
theoretically, no calibration of the sensors needs to be performed, but the
sensors still suffer from non-­uniformity, and at least a one-­point verification
for the sensors is needed to ensure accuracy.

9.6.2 Test procedures
The use of CS is the same as for TCS and ERS, the sensor should be saturated
first before use. The response time and accuracy of the CS are dependent on
the contact condition of the porous matrix with the soil. To install the CS into
a soil sample, a slot slightly larger than the sensor needs to be first made into
the soil sample. Moisten the removed soil and use it to pack around the sensor
such that it covers the entire porous matrix before installing. Once installed,
take an initial reading and then at half an hour, doubling the time interval for
subsequent readings. The readings should show an increase in matric suction
with time and reach a stable reading (equilibrium condition). The equilibrium
time of the CS varies from a few hours to several days and does not depend
on the matric suction of the soil sample to be measured (Tripathy et al. 2016).
The CS may not work well for sandy samples in the laboratory, as it is
difficult to make a slot in a dry soil sample without collapse. Sandy soil can-
not be moistened to adhere to the porous matrix. In addition, the matric
suction of sandy soil may change greatly with the introduction of a small
amount of water.

9.7 ELECTRICAL CONDUCTIVITY OF PORE WATER

Osmotic suction depends on the concentration of dissolved salts in the pore

water. The measurement of the electrical conductivity of the pore water can
be performed using an electrical conductivity meter. The amount of pore
water required depends on the electrical conductivity meter. There are com-
pact electrical conductivity meters which can measure electrical conductiv-
ity of small fluid volume. For example, Leong et al. (2007) used the Horiba
B-­173 which can measure the electrical conductivity of a drop of fluid
 Indirect suction measurement methods 151

(~1 ml) but for a limited range of electrical conductivity. The measurement
range of Horiba B-­173 is from 0 to 19.9 mS/cm with an accuracy of ±2%
full scale. Benchtop type of electrical conductivity meters has a wider range.
For example, the Horiba ES-­12 electrical conductivity meter has a measure-
ment range of 0–199.9 mS/cm and an accuracy of 0.5% full scale.
The electrical conductivity can be used to indirectly obtain the osmotic
suction of soil. There are many methods to extract the pore water from a
soil sample (USDA 1954; Iyer 1990; Fares 2009). Examples of such methods
are gas extraction, centrifuging, saturation extraction or dilution, immisci-
ble liquid displacement, leaching and pressurised squeezing. Among the
methods, extracting the pore water by dilute suspension of soil or saturated
soil paste (saturated extract) and mechanical squeezing are more popular.
Hence, these two methods will be described here.

9.7.1 Saturated extract
The saturated extract method assumes that all the salts present in the soil
will dissolve or remain in the solution as distilled water is added. In the
saturation extract method, the soil sample can be air-­dried or used as is.
Drying in the oven at 105°C is not recommended, as it will convert parts of
gypsum to plaster of Paris, which is more soluble (Rhoades 1982). In either
case, the initial dry density and water content of the soil sample need to be
determined first. This can be done with a sub-­sample. The preparation of
a saturated soil paste is given in USDA (1954). Distilled water is added to
the soil sample until water can be extracted with or without application of
a small suction. During the addition of distilled water, stir the soil-­water
mixture using a spatula. Note the volume or weight of the distilled water
added. Soil-­water mixtures at saturation of 1:1, 1:2 and 1:5 by weight have
been used to obtain the saturation extract. Some studies have shown that
the increase in water content leads to a reduction in the electrical conductiv-
ity estimated for the soil pore water (e.g., Hogg and Henry 1984; He et al.
2013). After the addition of distilled water, the soil-­water mixture is left to
stand for 4–6 hours before extraction.
For extraction of the pore water, prepare a clean conical flask, a piece of
filter paper, funnel and connector to a vacuum source as shown in Figure 9.6.
A small hand-­held vacuum pump may suffice as the vacuum source. Pour all
the soil mixture into the funnel. If required, apply a small suction using the
vacuum source. Collect as much of the extract as possible. Measure the elec-
trical conductivity of the extract using an electrical conductivity meter.
Electrical conductivity is affected by temperature, increasing approximately
by 1.9% per °C. The measured electrical conductivity, ECmeasured, can first be
corrected to the reference temperature 25°C, EC25°C, using Equation 9.7
(USDA 1954).

EC25o C  ft  ECmeasured , (9.7)

152 Laboratory tests for unsaturated soils

Funnel
Filter paper
Soil-water mixture Vacuum source

Saturation
Conical flask
extract

Figure 9.6 Set-up for filtering of soil-water mixture.

Figure 9.7 P
 ore fluid squeezer: (a) Schematic drawing (modified from Manheim
1966 and ASTM D4542 − 15). (b) Main components.
 Indirect suction measurement methods 153

where the temperature correction factor ft is given by Equation 9.8 (Sheets

and Hendrickx 1995).

 t 
ft  0.4470  1.4034 exp    (9.8)
 26.815 

The electrical conductivity of the pore water in the original soil EC is then
obtained by scaling EC25°C with the volumetric water content to give the EC
of the original soil using Equation 9.9.


EC  EC25o C (9.9)
 measured

where
θmeasured = volumetric water content of the soil-­water mixture from which
the saturation extract was obtained.
The osmotic suction can be obtained from the electrical conductivity
using Figure 7.1 or Equation 7.4.

9.7.2 Mechanical squeezing
The saturation extract method is widely used because of its simplicity.
However, there are criticisms of the method. Some studies have found that
there is a difference between the electrical conductivity calculated using
Equation 9.9 at different levels of dilutions (Khan 1967), while others show
that the effect is negligible (Argo 1997; Smethurst et al. 1997).
Mechanical squeezing of a soil sample enables the free pore water from a
sample with dissolved salts to be obtained from the soil (Fredlund et al.
2012). This method does not require any treatment of the soil prior to the
pore-­water extraction. The squeezing can be performed using a pore fluid
squeezer. A pore fluid squeezer is shown in Figure 9.6. The pore fluid
squeezer comprises three main parts: a base, a thick-­walled cylinder and a
piston (ram). The soil sample is placed into the cylinder and pressure is
applied on the piston using a compression machine to extract the pore water.
The pore water is collected from the effluent outlet using a syringe. If the
base is made to fit a sampling tube, a sample within the sampling tube can
be directly compressed to extract the pore water.
The procedures for extracting pore water by mechanical squeezing are
given in ASTM D4542 – 15:

1. Clean the components of the mechanical squeezer thoroughly and

rinse with distilled water and dry. Check that no rust or dirt should
remain in any part or the mechanical squeezer.
154 Laboratory tests for unsaturated soils

2. Place a clean piece of filter paper (5–10-­μm pore size) on top of the
mechanical squeezer base.
3. Place the soil sample into the heavy cylinder. Drier soil samples will
require more soil to be used.
4. Place the ram and apply pressure slowly until the first drop of water is
seen coming out of the effluent outlet.
5. Insert a clean, disposable syringe (25 mL) in the effluent outlet to col-
lect the subsequent pore water coming out of the effluent outlet.
6. Continue applying pressure until no more water is expelled or until
the syringe is full. ASTM D4542 – 15 specifies a maximum squeezing
pressure of 80 MPa based on the limits of pressure used by Kriukov
and Komarova (1956) of 59 MPa and Manheim (1966) of 101 MPa.
Krahn and Fredlund (1972) used a maximum pressure of 10 MPa.
7. Remove the syringe with the maximum pressure still being applied.
8. Transfer the extracted pore water to an electrical conductivity meter
for measurement.

The osmotic suction can be obtained from the electrical conductivity using
Figure 7.1 or Equation 7.4.

9.8 SUMMARY

A summary of the indirect suction measurement methods for total, matric

and suction components is given in Table 9.4. The thermal conductivity,
electrical resistance and conductance sensors for matric suction measure-
ment are low-­cost sensors that could be considered when a large number of

Table 9.4 Summary of indirect measurement methods

Suction Approx.
component Technique Suction range (kPa) Equilibration time
Total Non-­contact filter paper All ≥ 14 days
method
Matric Contact filter paper All ≥ 14 days
Thermal conductivity sensor 10–2,500 Hours to days
Electrical resistance sensor 10–200 Hours to days
(granular matrix)
10–1,000
Capacitance sensors 5–100,000 Hours to days
Osmotic Saturation extract technique 0–1,500 Minutes to hours
Squeezing technique 0–1,500 Minutes to hours
 Indirect suction measurement methods 155

measurements are needed at the same time. Although they are of lower accu-
racy than the direct measurement methods, they are better than the filter
paper method, as they are less dependent on the operator, do not require a
high-­resolution weighing balance (0.0001g) and have a shorter equilibrium
time, as little as 2 hours.

REFERENCES

Aitchison, G. D. (1964). Engineering concepts of moisture equilibria and moisture

changes in soils. Statement of review panel, Ed., Moisture equilibria and moisture
changes in soils beneath covered areas. Australia: Butterworths, pp. 7–21.
Argo, W. R., Weesies, N. J., Bergman, E. M., Marshal, M., & Biernbaum, J. A. (1997).
Evaluating Rhizon soil solution samplers as a method for extracting nutrient solu-
tion and analysing media for container-­grown crops. Horticulture Technology, 7:
404–408.
Armstrong, C. F., Ligon, J. T., & Thomson, S. J. (1985). Calibration of Watermark
Model 200 soil moisture sensor. ASAE Paper No. 85-­2077. ASAE, St, Joseph, MI
49085.
Barani, J. (2021 February 01). How to prepare and condition Watermark 200SS
soil water tension sensors. https://www.baranidesign.com/faq-articles/2021/2/1/
how-to-condition-install-watermark-200ss-soil-water-tension-sensors
Bouyoucos, G. J. & Mick, A. H. (1940). An electrical resistance method for the
continuous measurement of soil moisture under field conditions. Michigan
Agricultural Experiment Station Tech. Bull., 172.
Brown, A. J. (1907). On the existence of semipermeable membrane enclosing the
seeds of some of the Gramineae. Annals of Botany, 21: 79–87.
Bulut, R. & Leong, E. C. (2008). Indirect measurement of suction. Geotechnical and
Geological Engineering Journal, 26(6): 633–644.
Bulut, R. & Wray, W. K. (2005). Free energy of water – suction – in filter papers.
Geotechnical Testing Journal, 28(4): 355–364. https://doi.org/10.1520/GTJ12307
Bulut, R., Hineidi, S. M., & Bailey, B. (2002) Suction measurements-­filter paper
and chilled mirror psychrometer. In Proceedings of the Texas Section, American
Society of Civil Engineers, Fall Meeting, Waco, Texas, USA (2002), pp. 2–5.
Campbell, G. S. & Gee, G. W. (1986). Water potential: Miscellaneous methods. In
A. Klute (Ed.), Methods of soil analysis, Part 1, 2nd ed. Madison, WI: American
Society of Agronomy, pp. 619–633.
Chandler, R. J. & Gutierrez, C. I. (1986). The filter-­paper method of suction mea-
surement. Geotechnique, 36(2): 265–268.
Chandler, R. J., Crilly, M. S., & Montgomery-­Smith, G. (1992). A low cost method
of assessing clay desiccation for low-­rise buildings. Proceedings of the Institution
of Civil Engineers, 92(2): 82–89.
Corwin, D. L. & Yemoto, K. (2019). Measurement of soil salinity: Electrical conduc-
tivity and total dissolved solids. Soil Science Society of America Journal, 83: 1–2.
https://doi.org/10.2136/sssaj2018.06.0221
Al-­Khafaf, S. & Hanks, R. J. (1974). Evaluation of the filter paper method for esti-
mating soil water potential. Soil Science, 117(4): 194–199.
156 Laboratory tests for unsaturated soils

Crilly, M. S., Schreiner, H. D., & Gourley, C. S. (1991). A simple field suction mea-
surement probe. In Proceedings of the 10th African Regional Conference on Soil
Mechanics and Foundation Engineering, Lesotho, pp. 291–298.
Croney, D. & Coleman, J. D. (1961). Pore pressure and suction in soil. In Conference
on Pore Pressure and Suction in Soils, London: Butterworths, pp. 31–37.
Decagon (2015). MPS-­2 & MPS-­6—Dielectric water potential sensors. Operator’s
manual. Pullman, WA: Decagon Devices, Inc.
Duran, A. J. G. (1986). Study of effect of contact on the filter paper technique in the
measurement of soil suction. M.Sc. dissertation, Imperial College, London.
El-­Ehwany, M. & Houston, S. L. (1990). Settlement and moisture movement in col-
lapsible soils. Journal of Geotechnical Engineering, 116(10): 1521–1535.
Fares, A., Deb, S. K., & Fares, S. (2009). Review of vadose zone soil solution sam-
pling techniques. Environmental Reviews, 17: 215–234. https://doi.org/10.1139/
A09-­010
Fawcett, R. G. & Collis-­George, N. (1967). A filter-­paper method for determining
the moisture characteristics of soil. Australian Journal of Experimental Agriculture
and Animal Husbandry, 7: 162–167.
Feng, M. & Fredlund, D. G. (2003). Calibration of thermal conductivity sensors
with consideration of hysteresis. Canadian Geotechnical Journal, 40: 1048–1055.
Fredlund, D. G. (1992). How negative can pore-­water pressures get? Geotechnical
News, 9(3): 44–46.
Gardner, R. (1937). A method of measuring the capillary tension of soil moisture
over a wide moisture range. Soil Science, 43: 277–283.
Greacen, E. L., Walker, G. R., & Cook, P. G. (1987). Evaluation of the filter
paper method for measuring soil water suction. International Conference on
Measurement of Soil and Plant Water Status, pp. 137–143.
Greenspan, L. (1977). Humidity fixed points of binary saturated aqueous solutions.
Journal of Research of the National Bureau of Standards Section A: Physics and
Chemistry, 81A(1): 89–96.
Gutierrez, C. I. G. (1985). Assessment of filter paper technique for the measurement
of soil water suction, M.Sc. dissertation, Imperial College, London.
Hamblin, A. P. (1981). Filter-­paper method for routine measurement of field water
potential. Journal of Hydrology, 53: 355–360.
Hansen, H. C. (1926). The water-­retaining power of the soil. Journal of Ecology, 14:
92–111.
Harrison, B. A. & Blight, G. E. (1998). The effect of filter paper and psychrom-
eter calibration techniques on soil suction measurements. In Proceedings of the
Second International Conference on Unsaturated Soils, International Academic
Publishers, Beijing, China, Vol. 1, pp. 362–367.
He, Y., DeSutter, T., Hopkins, D., Jia, X., & Wysocki, D. A. (2013). Predicting ECe of
the saturated paste extract from value of EC1:5. Canadian Journal of Soil Science,
93(5): 585–594. https://doi.org/10.4141/cjss2012-­080
Hillel, D. (1980). Fundamentals of soil physics. San Diego, CA: Academic Press,
p. 403.
Hogg, T. J. 7 Henry, J. L. (1984). Comparison of 1:1 and 1:2 Suspensions and extracts
with the saturation extract in estimating salinity in Saskatchewan soils. Canadian
Journal of Soil Science, 64(4): 699–704. https://doi.org/10.4141/cjss84-­069
Houston, S. L., Houston, W. N., & Wagner, A. N. 1994. Laboratory filter paper
suction measurements. Geotechnical Testing Journal, GTJODJ, 17(2): 185–194.
 Indirect suction measurement methods 157

Hu, Y., Vu, H., & Fredlund, D. G. (2007). Calibration of thermal conductivity sensors
for suction measurement. In 60th Canadian Geotechnical Conference & 8th Joint
CGS/IAH: CNC Groundwater Conference (Ottawa Geo 2007) [Proceedings] (21
October 2007): 1–17.
Iyer, B. (1990). Pore water extraction – Comparison of saturation extract and high-­
pressure squeezing. In K. B. Hoddinott & R. O. Lambs (Eds.), Physico-­chemical
aspects of soils and related materials, ASTM STP 1095. Philadelphia: American
Society for Testing and Materials, pp. 159–170.
Khan, S. U. (1967). Some aspects of equilibrium solutions for a solonetzic and
chernozenie soil, Ph.D. thesis, Department of Soil Science, University of Alberta,
Edmonton.
Krahn, J. & Fredlund, D. G. (1972). On total, matric and somotic suction. Journal of
Soil Science, 114(5): 339–348.
Kriukov, P. A. & Manheim, F. T. (1982). Extraction and investigative techniques
for study of interstitial waters of unconsolidated sediments: A review. In K. A.
Fanning & F. T. Manheim (Eds.), The dynamic environment of the ocean floor.
Toronto: LexingtonBooks, pp. 3–26.
Kryukov, P. A. & Komarova, N. A. (1956). Issledovanie rastvorov pochv, ilov i
gornykh porod (Studies on the interstitial waters of soils, muds and rocks).
Mezhdunarodnomu kongressu pochvovedov, 2nd Komissiya, Doklady. 6: 151–184.
Lee, H. C. & Wray, W. K. 1992, Evaluation of soil suction instruments. In Proceedings
of the 7th International Conference on Expansive Soils, Dallas, USA, pp 307–312.
Leong, E. C., Kizza, R., & Rahardjo, H. (2016). Measurement of soil suction using
moist filter paper. In 3rd European Conference on Unsaturated Soils - E-­UNSAT
2016, Delage, P., Cui, Y. J., Ghabezloo, S., Pereira, J. M., and Tang, A. M. (Eds.),
E3S Web of Conferences. https://doi.org/10.1051/e3sconf/20160910012
Leong, E. C., Tripathy, S., & Rahardjo, H. (2003). Total suction measurement of
unsaturated soils with a device using the chilled-­mirror dew-­point technique.
Geotechnique, 53(2): 173–182.
Leong, E. C., Widiastuti, S., Lee, C. C., & Rahardjo, H. (2007). Accuracy of suction-
measurement. Geotechnique, 57(6): 547–556.
Leong, E. C., Wijaya, M., Tong, W. Y., & Lu, Y. (2020). Examining the contact fil-
ter paper method in the low suction range. Geotechnical Testing Journal, 43(6):
1567–1573. https://doi.org/10.1520/GTJ20190237
Leong, E. C., Zhang, X. -H. and Rahardjo, H. (2011). Calibration of a thermal
conductivity sensor for field measurement of matric suction. Geotechnique, 61(1):
81–85. https://doi.org/ 10.1680/geot.9.P.008.
Likos, W. J. & Lu, N. (2002). Filter paper technique for measuring total soil
suction. Transportation Research Record, 1786(1): 120–128. https://doi.
org/10.3141/1786-­14
Malazian, A., Hartsough, P., Kamai, T., Campbell, G. S., Cobos, D. R., & Hopmans, J.
W. (2011). Evaluation of MPS-­1 soil water potential sensor. Journal of Hydrology,
402(1–2): 126–134, https://doi.org/10.1016/j.jhydrol.2011.03.006
Manheim, F. T. (1966a). A hydraulic squeezer for obtaining interstitial water from
Consolidated and unconsolidated sediments. U. S. Geol. Surv. Profess. Paper 550-­
C, C256.
Manheim, F. T. (1966b). A Hydraulic Squeezer for Obtaining Interstitial Water from
Consolidated and Unconsolidated Sediment. U.S. Geological Survey Professional
Paper 550-­C, pp. 256–261.
158 Laboratory tests for unsaturated soils

Marinho, F. & Oliveira, O. (2006)., The filter paper method revisited. Geotechnical
Testing Journal, 29(3): 1–9.
Marinho, F. A. M. (1994). Shrinkage behaviour of some plastic clays, Ph.D. thesis,
University of London.
McCann, I. R., Kincaid, D. C., & Wang, D. (1992). Operational characteristics of
the Watermark Model 200 soil water potential sensor for irrigation management.
Applied Engineering in Agriculture 8(5): 605–609. ASAE, 2950 Niles Rd., St.
Joseph, MI 49085
McKeen, R. G. (1980). Field studies of airport pavements on expansive soils. In 4th
International Conference on Expansive Soils, pp. 242–261.
McKeen, R. G. (1985). Validation of procedures for pavement design on expansive
soil behaviour. Report No. DOT/FAA/PM-­85/15. Washington, DC, USA.
McKeen, R. G. (1988). Soil characterization using suction measurements. In 25th
Proceedings of Paving and Transp. Conference, University of Mexico, Albuquerque,
N.M.
McQueen, I. S. & Miller, R. F. (1968a). Determination of a soil moisture potential. In
Water in the unsaturated zone: Proceedings of the Wageningen Symposium, P. E.
Rijtema & H. Wassink (Eds.), International Association of Science and Hydrology
Publication No. 82, pp. 147–155.
McQueen, I. S. & Miller, R. F. (1968b). Calibration of a wide-­range gravimetric
method for measuring moisture stress. Soil Science, 106(3): 225–231.
Miller, D. J. & Nelson, D. (1992). Osmotic suction as a valid stress state variable
in unsaturated soils. In Proceedings of the 7th International Conference on
Expansive Soils, Dallas, USA, pp. 179–184.
Moncur, M. C., Blowes, D. W., & Ptacek, C. J. (2013). Pore-­water extraction from
the unsaturated and saturated zones. Canadian Journal of Earth Sciences, 50(10):
1051–1058. https://doi.org/10.1139/cjes-­2012-­0165
Noborio, K., Horton, R., & Tan, C. S. (1999). Time domain reflectometry probe for
simultaneous measurement of soil matric potential and water content. Soil Science
Society of America Journal, 63: 1500–1505.
Or, D. & Wraith, J. M. (1999). A new soil matric potential sensor based on time
domain reflectometry. Water Resources Research, 35: 3399–3407.
Power, K. C., Vanapalli, S. K., & Garga, V. K. (2008). A revised contact filter paper
method. Geotechnical Testing Journal, 31(6): 461–469. https://doi.org/10.1520/
GTJ101099
Rhoades, J. D. (1982). Soluble salts. In Methods of soil analysis, Part 2, Chemical
and Microbiological Properties, 2nd ed., ASA-­SSSA, Agronomy Monograph, 9:
167–179.
Ridley, A. M. & Wray, W. K. (1996). Suction measurement: A review of current
theory and practices. In First International Conference on Unsaturated Soils,
Rotterdam, the Netherlands: A. A. Balkema, pp. 1293–1322.
Ridley, A. M. (1993). The measurement of soil moisture suction, Ph.D. thesis,
University of London.
Scanlon, B. R., Aandraski, B. J., & Bilskie, J. (2002). Miscellaneous methods for
measuring matric or water potential. In J. H. Dane & G. C. Topp (Eds.), Methods
of soil analysis, part 4, physical methods. Madison, WI: American Society of
Agronomy, pp. 643–670.
Schröder, H. (1911). Über die selektiv permeable hülle des weizenkornes. Flora, 102:
186–208.
 Indirect suction measurement methods 159

Sheets, K. R., & Hendrickx, J. M. H. (1995). Noninvasive Soil Water Content

Measurement Using Electromagnetic Induction. Water Resources Research,
31(10): 2401– 2409, https:/doi.org/10.1029/95WR01949.
Sibley, J. W. & Williams, D. J. (1990). A new filter material for measuring soil suc-
tion. Geotechnical Testing Journal, 13(4): 381–384.
Sibley, J. W., Smyth, G. K., & Williams, D. J. (1990). Suction-­moisture content cali-
bration of filter papers from different boxes. Geotechnical Testing Journal, 13(3):
257–262.
Spaans, E. J. A. & Baker, J. M. (1992). Calibration of Watermark soil moisture sen-
sors for soil matric potential and temperature. Plant Soil, 143: 213–217.
Stocker, O. (1930). Über die Messung von Bodensaughkräften und ihren verhältnis
Suden Wurzelsaughräften. Zeitschrift für Botanik., 23: 27–56.
Swarbrick, G. E. (1995). Measurement of soil suction using the filter paper method.
In Unsaturated Soils: Proceedings of the 1st International Conference on Un­saturated
Soils, A. A. Balkema, Rotterdam, Netherlands, pp. 653–658.
Thomson, S. J. & Armstrong, C. F. (1987). Calibration of the Watermark 200 soil
moisture sensor. Applied Engineering Agriculture, 3: 186–189.
Tripathy, S., Al-­Khyat, S., Cleall, P. J. Baille, W., & Schanz, T. (2016). Soil suc-
tion measurement of unsaturated soils with a sensor using fixed-­matrix porous
ceramic discs. Indian Geotechnical Journal, 46: 252–260. https://doi.org/10.1007/
s40098-­016-­0200-­z
van der Raadt, P., Fredlund, D. G., Clifton, A. W., Klassen, M. J., & Jubien, W. E.
(1987). Soil suction measurement at several sites in Western Canada, Trans­
portation Res. Rec. 1137, Soil Mech. Considerations in Arid and Semi-­Arid Areas,
Transportation Res. Board, Washington, DC, pp. 24–35.
Wagenet, R. J. & Jurinak, J. J. (1978). Spatial variability of soluble salt content in a
mancos shale watershed. Soil Science, 126: 342–349.
William, O. B. & Sedgley, R. H. (1965). A simplified filter paper method for deter-
mining the 15-­atmosphere percentage in soils. Australian Journal of Experimental
Agriculture and Animal Husbandry, 5: 201–205.

FURTHER READING

Walshire, L., Berry, W., & Taylor, O.-D. (2020). Redesigned filter paper method:
Protocol and assessment using reconstituted samples. Geotechnical Testing
Journal. https://doi.org/10.1520/GTJ20190420
Chapter 10

Soil-water characteristic curve

10.1 BACKGROUND

The soil-­water characteristic curve (SWCC) relates the water content of a

soil with its suction(s). The first SWCC was plotted in soil physics by Edgar
Buckingham in 1907 for six soils ranging in texture from sand to clay. The
SWCC has since become an important relationship in many other disci-
plines and as such there are many alternative names for the SWCC (Ng and
Menzies 2007; Fredlund et al. 2001):

1. Soil-­water retention curve (SWRC)

2. Soil-­moisture retention curve
3. Moisture retention curve
4. Water retention curve
5. Suction-­water content relationship
6. Retention curve

In this book, the term “SWCC” is preferred, as it is more commonly used

in geotechnical engineering (Fredlund et al. 2001). The SWCC is considered
one of the most important properties of unsaturated soils, as it can be used
to estimate the effect of negative pore-­water pressure on shear strength,
permeability and compressibility of unsaturated soils. In order to guide the
engineer in designing the test programme for SWCC, there are several ques-
tions which need to be considered:

1. Does the analysis require only drying SWCC, wetting SWCC or both
drying and wetting SWCC, along with its scanning curves?
2. What is the soil state (i.e., void ratio, density, water content, etc) which
is needed in the analysis?
3. Does the soil have bimodal pore-­size distribution?
4. Does the soil have significant change in volume during the test?
5. What is the suction range of interest?

DOI: 10.1201/b22304-10 161

162 Laboratory tests for unsaturated soils

6. What is the available equipment to measure matric suction and vol-

ume change?
7. What is the time constraint in obtaining the SWCC?

By answering all of these questions, suitable tests can be performed to obtain

the appropriate SWCC.

10.2 RELATED STANDARDS

1. ASTM-­D6836-­02. 2008. Standard Test Method for Determination

of Soil Water Characteristic Curve for Desorption Using Hanging
Column, Pressure Extractor, Chilled Mirror Hygrometer, or Centrifuge.
ASTM Book of Annual Standards.
2. BS EN ISO 11274. 2014. Soil Quality. Determination of the Water-­
Retention Characteristic. Laboratory Methods.

10.3 SWCC CONVENTION

To date, many SWCCs have been determined. However, the SWCC pre-
sented takes many forms. In determining SWCC, it is not sufficient to
ask, “What is the SWCC of soil?” without giving further information on
“which” SWCC is required. Generally, SWCC is plotted with volumetric
water content on the y-­axis and suction on the x-­axis. The following SWCC
descriptions should be adopted for clarity:

1) Water content – gravimetric water content based SWCC (w-­SWCC),

volumetric water content based SWCC (θ-­SWCC) and degree of satu-
ration based SWCC (S-­SWCC)
2) Volume measurement – SWCC without volume measurement
(w-­SWCC) and SWCC with volume measurement (θ-­SWCC and
S-­SWCC)
3) Path – initial drying, main drying, main wetting and scanning SWCC
4) Shape – unimodal and bimodal (or multimodal) SWCC

Aside from the general description, it is also important to specify the initial
condition of the soil, such as initial dry density, initial water content, initial
soil suction and the soil structure. Incorrect initial soil conditions may give
a misleading SWCC.
Standard parameters that can be obtained from the SWCC are as follows
(Fredlund et al. 2001; Ng and Menzies 2007):

1. Initial/saturated gravimetric water content (ws), volumetric water con-

tent (θs), and degree of saturation (Sr,s)
2. Air-­entry value (AEV) which is only obtainable directly from S-­SWCC
 Soil-water characteristic curve 163

3. Residual matric suction (sr) and residual water content (wr)

4. Suction at the intersection points
5. Slope of each linear segment of the curve
6. Inflection point suction and water content (sinf and winf, respectively)
7. Curve-­fitting equation and parameters

The x-­axis of SWCC represents the soil suction which can be either matric
suction or total suction. At low suction, the x-­axis of the SWCC is matric
suction and at high suctions where the air phase becomes continuous, the
operative suction is total suction. The transition suction where this occurs
depends on the soil type. However, it is quite common to not differenti-
ate the suction into matric or total suction. In soil physics-­and agronomy-­
related disciplines, the unit used for soil suction is pF which is equal to the
logarithm to the base of ten of the absolute value of the column height of
water in centimetres that equilibrates with the suction of the soil (Fredlund
et al. 2001). However, in geotechnical engineering practice, stress unit (i.e.,
kPa) is more common and relevant. The suction range of interest can be as
low as 0.1 kPa up to around 1 GPa. Thus, it is quite common to use log scale
for the suction x-­axis.

10.4 THEORY

The typical shape of the SWCC is sigmoid, as shown in Figure 10.1. Such
an SWCC is now more commonly known as unimodal SWCC as opposed
to bimodal or multimodal SWCC which consists of more than one sigmoid.

Figure 10.1 Typical shape of SWCC.

164 Laboratory tests for unsaturated soils

The unimodal drying SWCC (desorption SWCC) can be discretised into

three linear segments and zones (Wijaya and Leong 2016; Kohgo 2003)
which are the initial drying line (boundary effect zone), virgin drying line
(transition zone) and residual drying line (residual zone). The suction at the
intersection of initial drying and virgin drying lines (s2) is commonly referred
to as AEV for a soil that does not change in volume on drying as s2 is the
same for w-­SWCC, θ-­SWCC and S-­SWCC. However, for a soil that experi-
ences a change in volume on drying (e.g., most fine-­grained soils), s2 is AEV
only for S-­SWCC (Fredlund et al. 2011; Wijaya et al. 2015). The suction at
the intersection of virgin and residual drying lines (s3) is commonly referred
to as residual matric suction (sr) and the corresponding water content as the
residual water content (wr). Similarly, the unimodal wetting SWCC (adsorp-
tion SWCC) can be discretised into three linear segments which are starting
from the high suction, initial wetting line, virgin wetting line and residual
wetting line. The suction at the intersection of the initial and virgin wetting
lines is commonly referred to as water-­entry value (WEV).
To obtain S-­SWCC and θ-­SWCC, the volume of the soil specimen at each
suction equilibrium is required. When volume measurement is difficult to
carry out during the SWCC test, it is possible to use a shrinkage curve to
replace the volume measurement.

10.4.1 SWCC equations
The SWCC test provides discrete data points where each data point repre-
sents a pair of suction and the corresponding equilibrium water content.
However, a continuous relationship between suction and water content is
more desirable for application. Hence, a number of unimodal SWCC equa-
tions have been proposed which can be classified into two categories:

1. Empirical equations
2. Physical-­based equations

10.4.1.1 Empirical equations
Empirical SWCC equations only consider the shape of the SWCC, and the
parameters of the equations are obtained by optimising the curve fit to the
data and may not have any physical meaning. Two commonly used empiri-
cal equations are van Genuchten (1980) and Fredlund and Xing (1994)
SWCC equations. The van Genuchten (1980) equation is given by Equation
10.1.

w s  wr
w s  m
 wr (10.1)
1   a.s n 
 
 Soil-water characteristic curve 165

The Fredlund and Xing (1994) SWCC equation is given by Equation 10.2.

C s
w  w sat mf
(10.2a)
   s 
nf

ln
  exp    
1 
  af  

 s 
ln  1  

C s  1  r 
, (10.2b)
 106 
ln  1  
 r 

where wsat is the saturated gravimetric water content, s is the matric suction,
C(s) is a correction function to force the SWCC to have zero water content
at Ψ = 1 GPa, Ψr is an empirical parameter which represents the residual
matric suction sr, af, nf and mf are the curve-­fitting parameters. More empiri-
cal equations are shown in Table 10.1.

Table 10.1 Unimodal SWCC empirical equation

Reference Equation Input parameter

van Genuchten
w s 
ws  wr   wr
General notation:
(1980) m v
ws = saturated water content
1   av s n 
v
wr = residual water content
  s = soil suction
av, nv, mv = van Genuchten (1980)
Brooks and For s < s2: w  s   w s
Corey (1964) empirical parameters
S 
ab s2 = suction at the intersection
For s ≥ s2 w  s    2  between initial drying line and
 S  virgin drying line
Fredlund and w sC  s  ab = Brooks and Corey (1964)
Xing (1994) w s  mf
empirical parameter
   s  
nf af, nf, mf, ψr= Fredlund and Xing
ln  exp 1     (1994) empirical parameters
  af   bd, c, dd = Feng and Fredlund

(1999) empirical parameters
 s 
ln 1  

C s 1 
r 

 106 
ln 1  
 r 
Feng and w s bd  c  s dd
Fredlund (1999) w s 
bd  s dd
166 Laboratory tests for unsaturated soils

Although empirical equations are convenient for use, there are several
disadvantages (Wijaya and Leong 2016):

1. The parameters are determined by optimising the curve fit to the data.
Sometimes, optimisation does not work without constraining the
range of the parameters.
2. Different parameters in the equation may give the same SWCC which
causes difficulties when relating the non-­unique parameters to other
soil properties.

10.4.1.2 Physical-based equations
Physical-­based equations use parameters obtained directly from the curve.
Thus, the parameters of physical-­based equations are unique. Compared to
empirical equations, physical-­based equations are more complicated in form
and cumbersome to use.
Examples of physical-­ based equations proposed for unimodal and
bimodal SWCCs are given in Table 10.2 and Table 10.3, respectively. More
details of bimodal SWCCs are given later in the chapter.
Wijaya and Leong (2016) proposed a physical-­based equation that can be
applied to both unimodal and bimodal SWCCs. The proposed equation is
shown in Equation 10.3.
n
s
w  s   w sat  m1 log   
 s1 
T  log s, log s , m , m
i2
i i i i 1 , ci  (10.3a)

1 1  cosh k i  x  x i   
Ti  x, x i , mi , mi 1, k i    mi  mi 1   x  x1   ln  
2 k i  cosh k i  x i  x0   
 
(10.3b)

2
ci  (10.3c)
 si  
log  
 si  

where mi is the slope of segment i, si is the intersection between segment i

and segment i-­1, ci is the curvature parameter that controls the curve joining
segment i and segment i-­1, si+ is the point of convergence between the curve
and segment i, si-­is the point of convergence between the curve and segment
i-­1 and s1 represents the minimum x-­axis value and can be simply taken as
0.1 or 1 kPa. All mi and si values can be determined directly from straight
lines drawn as linear approximations of the curve.
Table 10.2 Unimodal SWCC equation that uses the unique parameters approach (Wijaya 2017)

Reference Equation Input parameter

Pham and Fredlund  10  6 General notation:
(2008) w  S2 f2  S3  m3 log   s = matric suction
 s  s1 = matric suction at wsat
scpi For Unimodal SWCC:
fi  s, si , cpi   fi  cpi i cpi ; g i  s, si , c pi   g i  1  f i  s, s, c pi 
si  s s2 = matric suction at the intersection point of
segments 1 and 2
df i  s, si , c pi  ln 10  s3 = matric suction at the intersection point of
f i  s, si , c pi , mi 1, mi     f i  f i  g i  mi 1  mi 
log s 2c pi segments 2 and 3
smin = matric suction at zero water content
 s  ws = saturated water content
Si  fi mi  mi1  log    f i
 s2 
When it is desired force w = 0 at s = 106 kPa m1 = slope of segment 1
w s   m2  m1  log s2  m2 log s3 m2 = slope of segment 2
m3  m3 = slope of segment 3

Soil-water characteristic curve

106
log
s3
m1  m3 s
Gould et al. (2012) w  w sat  S2  S3  log Additional parameter for Pham and Fredlund
2 smin
(2008):
x 1   x  cp2 = parameter that controls the curvature at s 2
2
x 
f  x, y, cGi   cGi log tan1  cGi log   ln 1   cGi log   cp3 = parameter that controls the curvature at s 3
y  y 2   y  Additional parameters for Gould et al. (2012):
 
cG2 = parameter that control the curvature at s2
mi1  mi
Si    f  s, si , cGi   f  smin , si , cGi   cG3 = parameter that control the curvature at s3
cGi 

167
 168
Table 10.3 Bimodal SWCC equations that used unique parameters approach (Wijaya 2017)

Laboratory tests for unsaturated soils

Reference Equation Input parameter
Gitirana n
fi  fi1 General notation:
and
Fredlund
w  fn1   i 2  s 
di AEV = air-­entry value
Ψr = curve-­fitting parameter which represents residual
(2004) 1    matric suction
 s i  s i 1  sm-­i = matric suction at inflection point of segment i

fi 

tani 1  ri2 ln  s / si   
s = matric suction
s1 = Matric suction at wsat
1 r tan   i
2 2
i For Bimodal SWCC:
s2 = matric suction at the intersection point of
 1 
1 tan  
i 1
2
i
ri2 ln2  s / si  

a 2 1  ri2 tan2 i  w segments 1 and 2
1 r tan   s3 = matric suction at the intersection point of
i
i
2 2
i 1  tan i
2
segments 2 and 3
  s4 = matric suction at the intersection point of
1  i1  i  segments 3 and 4
di  2 exp  ; ri  tan  ;
 ln  si1 / si    2  s5 = matric suction at the intersection point of
segments 4 and 5
i1  i   w i  w i1   s6 = smin = s at zero water content
i   i  arctan   w1 = wSat
2  ln  si1 / si   w2 = w at s2
w3 = w at s3
Unimodal equation one bending point; i = 2 w4 = w at s4
For two bending point (n = 2)and bimodal SWCC (n = 4): w5 = w at s5
w6 = 0
Satyanaga ª m1 = slope of segment 1
et al.
­ ª § si  s · º½ § s ·º
n ° « ln ¨ ¸ »° « ln ¨1  ¸» m2 = slope of segment 2
« © si  sm  i «1  © < r ¹ »
(2013) w wr  ¦w i 1  w i 1
®
°erf
¯ «¬ ssi
¹ »¾
»¼ °¿ « § 106 · »
m3 = slope of segment 3
m4 = slope of segment 4
i 2 « ln ¨1  ¸» m5 = slope of segment 5
«¬ © < r ¹ »¼
x
a = additional curve-­fitting parameter of Gitirana and
1 § x2 · Fredlund (2004)
erf 1
³
f
2S
exp ¨ 
© 2
¸ dx
¹
Ssi = additional curve-­fitting parameter of Satyanaga et
al. (2013)
For unimodal SWCC n = 2 while for bimodal SWCC n = 3 ni, li, mLi, λι = curve-­fitting parameters of Li (2009)
equation
Li (2009)
s 2 s3
 
n1/log s3 s
2
or Li et al. w  s    w sat  w 3 
(2014)
s
 
n1/log s3 s
2
s 2 s3
 
n1/log s3 s
2

l1s3 
mL1

  w sat  w 3  / 1
s  l1s3 
mL 1
mL 1

n /log  s s 

Soil-water characteristic curve

5
2
s 4s5 4
  w 4  w5 
n /log  s s 
2 5 n /log  s s  2 5
s  ss
4
4 5
4

l2s5 
mL 2

  w 4  w 5  / 2
smL2  l2s5 
mL 2

Additional constrain:
 w sat  w 3  1 1    w 4  w 5  1 2   w sat

169
170 Laboratory tests for unsaturated soils

10.4.2 Hysteresis
Soil on drying and wetting experiences hysteresis where the water content
of the soil on the wetting path is less than the water content on the dry-
ing path at the same suction (Pham et al. 2005; Haines 1930; Hillel 2003;
Ng and Menzies 2007; Fredlund et al. 2000; Klausner 1991). Figure 10.2b
shows a schematic of the hysteresis phenomenon. Initial drying SWCC is
obtained for a soil drying from the fully saturated condition to the dry con-
dition beyond the residual suction. Main wetting SWCC is obtained for a
soil wetting from the dry condition beyond the residual suction to zero suc-
tion condition. At zero suction, if the soil is dried again, it will give the main
drying SWCC. The gap on the y-­axis between the initial drying SWCC and
the main drying SWCC is the residual air content and represents the voids in
the soil that are occluded. In nature, it may not be possible for the occluded
voids to become saturated. When the soil is wetted (or dried) at any point
on the main drying (wetting) SWCC, it will give the scanning curve. The
main drying and main wetting SWCCs form the bounds enclosing the zone
where scanning curves can occur. Thus, a soil undergoing multiple drying
and wetting cycles may be on the main drying SWCC, main wetting SWCC
or scanning curve. There are a number of hypotheses about the occurrence
of hysteresis in SWCC (Hillel 2003; Klausner 1991):

1. Non-­uniform pore-­size distribution which caused irregularities of the

void passages leading to the ink bottle effect.
2. The contact angle at an advancing interface during the wetting process
is different from that at a receding interface during the drying process.

Figure 10.2 Drying and wetting SWCC shows hysteresis.

 Soil-water characteristic curve 171

3. Entrapped air in the soil.

4. Non-­reversible changes of volume during drying and wetting or aging.

Some hysteresis models for SWCC are presented in Table 10.4. Mualem’s
(1977) hysteresis model can be used to estimate the main wetting SWCC
from the main drying SWCC and can be easily incorporated into any SWCC
equation. However, the shape of the main wetting curve is predetermined
regardless of the availability of the main wetting SWCC.
Feng and Fredlund’s (1999) hysteresis model can be used when there
are at least two data points to calibrate the parameters of the main wet-
ting SWCC equation, while the extended Feng and Fredlund (1999) hys-
teresis model (Pham et al. 2005) can be used to estimate the main drying
and the main wetting SWCC from the initial drying SWCC. Pham et al.
(2005) conducted statistical studies on the hysteresis of SWCC and con-
clude that:

1. Occluded air bubble accounts for about 5%–15% of the degree of

saturation and an average of 10% can be assumed. Thus, the degree
of saturation where main drying and main wetting SWCC converge is
about 90%.
2. The ratio of slopes of the main drying and main wetting SWCCs on a
semilogarithmic plot RSL is between 1 and 2, depending on soil type,
as shown in Table 10.4.
3. The horizontal distance between the main virgin drying line and main
virgin wetting line (DSL) is between 0.2 and 0.5 log suction depending
on soil type as shown in Table 10.4.

The preceding conclusions can be applied to any hysteresis model. However,

most of the soils used by Pham et al. (2005) do not undergo volume change
during drying and wetting. Hence, the conclusions may not be applicable to
clay and highly expansive soils.

10.4.3 The effect of dry density on SWCC

As the water can only fill up the void space in a soil, the SWCC depends on
the dry density of the soil. Wijaya and Leong (2017) show that the effect of
dry density on the SWCC will only shift the initial portion of the w-­SWCC
downwards (Figure 10.3a) but will shift the entire S-­SWCC to the right
(Figure 10.3b).
Thus, the effect of density is easier to account in w-­ SWCC than in
S-­SWCC. For w-­SWCC, the dry density only affects ws of the initial drying
line while the virgin drying line and residual drying line are unaffected. For
soils which do not experience volume change due to a change in matric
 172
Table 10.4 SWCC hysteresis model

Laboratory tests for unsaturated soils

Reference Equation Input parameter
Inverted Mualem  w  w min  General notation:
(1977) equation we    w = water content (can be either gravimetric, volumetric or degree of
(Pham et al.  w u  w min  saturation)
1/ 2
2005) w we  1  1  w de  s   Si = degree of saturation of the initial drying curve
Feng and Main drying curve: Sd = degree of saturation of the main drying curve
Fredlund Sw = degree of saturation of the main wetting curve
w u bd  c  s dd
(1999) wd  s  Ss = degree of saturation of the initial drying curve at the lowest suction
bd  s dd (usually is 100%)
Main wetting curve: Su = degree of saturation that becomes the meeting point between main
drying and main wetting curve
w u bw  c  s dw
ww  s  wd = water content at drying SWCC
bw  s dw ww = water content at wetting SWCC
c and ws are obtained from drying SWCC s = soil suction
2 data points (sw1,ww1) and (sw2,ww2) are required we = normalised water content
to obtain bw and dw which can be defined as wmin = meeting point between drying and wetting SWCC at higher soil
(Pham et al. 2003): suction
wu = meeting point between drying and wetting SWCC at lower soil
  w w1  c   w u  w w 2   suction
log  
  w u  w w1   w w 2  c   wed = normalised water content of the drying SWCC
dw 
 
wew = normalised water content of the wetting SWCC
log sw 2 c = empirical parameter for Feng and Fredlund (1999) for both drying
sw1
and wetting SWCC

bw 
 w w1  c  Swd1 w

w u  w w1
Extended Feng Initial drying curve: bd, dd = empirical parameters for Feng and Fredlund (1999) for drying
and Fredlund S b  c  s dd SWCC
(1999) equation Si  s   s d bw, dw = empirical parameters for Feng and Fredlund (1999) for wetting
(Pham et al. bd  s dd
SWCC
2005) Main drying curve: ww1 = first water content located at the wetting SWCC
Su bd  c  s dd sw1 = suction of the ww1
Sd  s   ww2 = second water content located at the wetting SWCC
bd  s dd
sw2 = suction of the ww2
Main wetting curve:
Su bw  c  s dw Recommended value for RSL and DSL (Pham et al. 2005)
Sw  s  
bw  s dw Soil type RSL DSL
Sand 2 0.2
Additional relationship:
1
Sandy loam 2.5 0.25
  RSL Silt loam and clay loam 1.5 0.5
bd  ; RSL = d d
bw   Compacted silt and 1 0.35

 10 DSL
 
dd
dw compacted sand
 

Soil-water characteristic curve

 b1/ 2d  Su is approximately 0.95Ss to 0.85Ss and 0.9Ss can be taken as reasonable
DSL  log  1d/ dw  value (Pham et al. 2005)
 bw 

173
174 Laboratory tests for unsaturated soils

Figure 10.3 E ffect of density on SWCC. (a) w-SWCC under different density. (b)
S-SWCC under different density.
 Soil-water characteristic curve 175

suction (coarse-­grained soil), the S-­SWCC can be obtained by using w-­SWCC

and initial void ratio e0, as shown in Equation 10.4.

w  Gs
S (10.4)
e0

However, for soils which experience volume change due to a change in mat-
ric suction, the shrinkage curve is needed to provide the equilibrium void
ratio (e) at each suction to construct S-­SWCC. Several equations have been
proposed to account for density in the SWCC, as shown in Table 10.5.
Equation (10.3) can be extended to account for the effect of density on
w-­SWCC (Wijaya and Leong 2017) as shown in Equation 10.5a.

s
w  s   w s,f  m1 log
s1
T2  logs, logs1, logs2,f , m2 , m1, k 2  (10.5a)
T3  logs, logs1, logs3 , m3 , m2 , k3 

where ws,f is the saturated gravimetric water content post-­density change,

s2,f is the matric suction at the intersection between segment 1 and segment
2 post-­density change and s2,f and ws,f can be obtained, respectively, using
Equations 10.5b and 10.5c.

w s ,0  w s , f

s2,f  s2,0  10 m2 m1

(10.5b)

S e
ws, f   rf  (10.5c)
 Gs 

where ws,0 is the initial saturated gravimetric water content pre-­density

change, s2,0 is the matric suction at the intersection between segment 1 and
segment 2 pre-­density change, Srf is usually assumed to be 100% (which
may not be due to the presence of occluded air voids), Gs is the specific grav-
ity and e is the void ratio of the soil.
For coarse-­grained soils such as sand, it is more convenient to express e in
Equation (10.5c) in terms of relative density Dr, maximum void ratio emax
and minimum void ratio emin as shown in Equation (10.5d). Both emax and
emin can be obtained from laboratory tests following ASTM-­D4254-­14
(2014).

S 
w s,f   r  emax 1  Dr   eminDr  (10.5d)
 Gs 
 176
Table 10.5 SWCC equations which account for density

Reference Equation Parameters

Laboratory tests for unsaturated soils

Gallipoli et al. 
mg
 w = gravimetric water content
(2003) – modified  1  wsat = saturated gravimetric water content
van Genuchten Sr   ng  Sr = degree of saturation
(1980) Equation 1   s  e   

n0 = initial porosity
    e = void ratio
ϕ and Ψ require additional SWCCs to be determined e0 = initial void ratio
Tarantino (2009) – B t / ng s = matric suction
  1/B t
ng
  sr = residual matric suction
van Genuchten   e 
(1980) Equation Sr  1    s  sp = matric suction of virgin drying line at 0 water
  A t   
content
m2 = slope of virgin drying line
Salager et al. 
(2010) – modified w sat  ln 1   s / sr    af, nf, mf = Fredlund and Xing (1994) curve-­fitting
w 1   parameters
ln exp 1   s / a    
mf
Fredlund and Xing  ln 1  10 / sr  
6
nf
ng, mg = van Genuchten (1980) curve-­fitting
(1994) equation f  
parameters
As, Bs = Salager et al. (2010) curve-­fitting
   a  parameters
 ln 1   f   
w sat    sr    ϕ, ψ = Gallipoli et al. (2003) curve-­fitting
mf  3.67 ln 1  parameters
 sp     106   
m2 ln   ln 1     At, Bt = Tarantino (2009) curve-­fitting parameters
 af     sr   ζ = Zhou et al. (2012) curve-­fitting parameters
 
 
mf 1
1.31  m2 af 
nf  3.72  
   a    satw   106  
 ln 1   f     1.31mf af  sr  ln 1    
  sr      sr  
mf 1  
 ln 1  10   
6

    
  sr  

af = AS e0BS
As and Bs require additional SWCCs to be determined
Zhou et al. (2012) S
Se    e 1  Se  dei


ei
Se is determined from other SWCC equations. Solution based on
Simpson’s rule is given as:
e S  Se ,0 S S 
ln f  e ,f f  Se ,0    4f  e ,0 e ,f   f  Se ,f 

Soil-water characteristic curve

ei 6  2 
1
f x 
x 1  x 


ζ requires additional SWCCs to be determined

Zhou et al. (2014) – 
modified Fredlund 1  ln 1   s / sr   
Sr  1  
ln exp 1   s  n / a    
n0nf
and Xing (1994)  ln 1  10 / sr  
n0nf 6

equation 0 f 
 ln 1   s / sr   
e 
w 1  
   
n0mf
 ln 1 10 / sr  
6
G s ln exp 1   s  n0 / af  

177
n0nf
  
178 Laboratory tests for unsaturated soils

10.4.4 Bimodal/multimodal SWCC
Some soils exhibit bimodal or multimodal SWCCs due to the presence of
macropores and micropores in the soils. Such a phenomenon is commonly
observed in soils with bimodal grain-­size distribution (Zhang and Chen
2005; Satyanaga et al. 2013). In a soil with bimodal grain-­size distribution,
the arrangement of coarse grains and fine grains can lead to large pores (mac-
ropores) and small pores (micropores), respectively, creating a dual-­porosity
soil (Burger and Shackelford 2001; Zhang and Chen 2005). However, com-
paction or other features such as cracks in the soil may also give rise to
dual porosity as well (Li 2009; Satyanaga et al. 2013; Li et al. 2014). Figure
10.4 shows a bimodal SWCC consisting of five segments where segments
1 and 2 are attributed to the macropores, while segments 3, 4 and 5 are
attributed to the micropores. It is also possible for a SWCC to appear to
be weakly bimodal due to small difference in sizes between micropores and
macropores, inaccuracy of the SWCC test or use of different methods to
obtain the SWCC. Thus, it is important to know a priori if the SWCC of a
soil is unimodal or bimodal, as a bimodal SWCC requires smaller suction
increments to capture the AEVs of the macropores and micropores accu-
rately. Zou and Leong (2019) proposed a classification tree to determine if a
soil with bimodal grain-­size distribution has a unimodal SWCC or bimodal
SWCC. Knowing a priori if the SWCC is unimodal or bimodal enables bet-
ter design of the suction increments for the SWCC test. The classification
tree is given in Figure 10.5a and the parameters used in the classification tree
are defined in Figure 10.5b.

Figure 10.4 Bimodal SWCC.

 Soil-water characteristic curve 179

(a)

50 100
Frequency Histogram Major Peak
45 90
Cumulative Frequency Graph
40 80
Grain-Size Frequency Plot
35 70

Percentage Passing (%)

30 60
Frequency Percentage (%)

25 50

20 Y 40

15 30
Minor Peak
10 20

5 10

0 0
0.0006 0.002 0.006 0.02 0.06 0.2 0.6 2 6
Logarithmic Scale Grain Size (mm)

(b)

Figure 10.5 C
 lassification tree and parameters for determining the type of SWCC
for a soil with bimodal grain-size distribution (from Zou and Leong, 2019).
(a) Classification tree. (b) Parameters Y and MaP (percentage for major peak)
for classification tree.

10.4.5 Using shrinkage curve as an alternative

volume measurement
Measuring the volume of soil specimen during the SWCC test is sometimes
not possible due to several reasons:

1. Volume change of the specimen is non-­ uniform and the specimen

becomes highly distorted making accurate volume measurement difficult.
2. Excessive handling of the specimen may cause disturbance to the spec-
imen and may “damage” delicate specimen.
180 Laboratory tests for unsaturated soils

In such situations, determining the shrinkage curve independently provides

a good alternative to volume measurement during the SWCC test. However,
it is important that a shrinkage curve is developed for the specimen that is
“identical” to that used for determining the SWCC.

10.5 TEST METHODS

10.5.1 Specimen preparation
Soil specimens for SWCC test can be categorised into the following:

1. Natural or undisturbed specimen

2. Reconstituted specimen
3. Compacted specimen

A natural or undisturbed specimen is used when SWCC of the in situ soil is

required. However, such soils may exhibit spatial variability which should
be considered when determining a representative SWCC.
In some cases, the soil needs to be reconstituted because it is not possible
to obtain an undisturbed soil sample, e.g., sands. For such soils, the soil
specimen is reconstituted to its in situ dry density. Different reconstitution
methods can be used, e.g., air, water or vacuum pluviation (Vaid and
Negussey 1988; Lagioia et al. 2006), dry and moist tamping (Frost and Park
2003) and under compaction (Ladd, 1973).
When compacted soil is of interest, static or dynamic compaction can be
used. Static compaction tends to produce more uniform specimens com-
pared to dynamic compaction. To further improve the quality of statically
compacted soils, an assembly of rings is used as the mould (Satija 1978; Goh
2012), where a soil is first compacted into a ring which is then flipped over
for the next ring to be attached and the next layer of soil to be compacted
(see 6.4.3.3 Static compaction test). This multi-­ring mould is to ensure that
none of the soil layers are over-­compacted. For dynamic compaction, it is
difficult to obtain an undisturbed soil specimen by trimming a soil com-
pacted in the standard mould. Hence, it is quite common to dynamically
compact the soil directly into a specimen mould so that the soil specimen
can be obtained without trimming. Different compaction methods, efforts
and water contents will lead to different dry density and soil structure thus
producing different soils and different SWCCs. Even when dry density is
maintained constant, the soil structure/fabric is different and effectively, a
different soil is obtained.

10.5.2 Test procedures
In general, the methods for determining the SWCC test can be broadly
divided into two as follows:
 Soil-water characteristic curve 181

Procedure A: Changing suction of the soil specimen and measuring its

water content
Procedure B: Changing water content of the specimen and measuring its
suction

In procedure A, a method to impose the suction is needed such as by using a

suction table, axis-­translation apparatus (Figure 10.6), vapour equilibrium
technique or centrifuge.

Figure 10.6 Axis-translation apparatuses. (a) Pressure plate apparatus. (b) Tempe cell.
182 Laboratory tests for unsaturated soils

Procedure A usually requires a longer time, as it is required to ensure that

the soil specimen achieves equilibrium at the imposed suction (usually indi-
cated by negligible changes in the weight of the specimen).
In procedure B, the water content of the soil specimen is changed for
instance by evaporation (non-­steady state or steady state via controlled rela-
tive humidity) and measuring its suction using devices such as tensiometer,
null-­
axis-­
translation apparatus or chilled mirror hygrometer (Schindler
1980). It is faster than procedure A as evaporation can change the water
content of the specimen rapidly. As both the water content and suction of
the soil specimen can be monitored continuously, many data points can be
obtained. However, there is no guarantee that the water content of the soil
specimen is homogeneous throughout the soil specimen during the test, and
the suction of the soil specimen may exceed the capacity of the suction mea-
suring device. An example of a commercially available apparatus imple-
menting procedure B is the HYPROP® which has two tensiometers installed
at two depths in the soil specimen and a balance to monitor the weight of
the soil specimen continuously. The averaged suction from the readings of
the two tensiometers and the water content of the soil specimen calculated
from its weight change are used to plot the SWCC.
Volumetric water content and degree of saturation can then be derived
from the gravimetric water content and void ratio. Gravimetric water con-
tent can be calculated using Equation 10.6.

 M w  
wi %   i  ref  1   1 100% (10.6)
 Mref  100  

where Mi is the weight of specimen at a particular suction or point of

interest, Mref is the reference mass, wref is the reference gravimetric water
content and wi is the water content at a particular suction or point of
interest. The reference point can be any point where both the mass and
gravimetric water content of the soil specimen are known. Two suitable
reference points are

1. initial condition prior to the SWCC test (wref is the gravimetric water
content of trimmings or left-­over sample, Mref is the mass of the speci-
men prior to test), and
2. final condition at the end of the SWCC test (wref the final gravimetric
water content and Mref is the final mass of the soil specimen).

It is always a good practice to determine both points of reference. The first

point of reference enables the water content of the specimen to be calculated
(and hence SWCC to be plotted) while the test is still ongoing, but it is less
accurate as the water content was determined from trimmings or left-­over
 Soil-water characteristic curve 183

sample compared to the second point of reference where water content of

the actual soil specimen is determined. However, the second point of refer-
ence should be used to obtain the more accurate SWCC.
If volume measurement is performed during the test, the void ratio of the
specimen can be calculated using Equation 10.7.

 w  G
ei  Vi  1  ref  s  1 (10.7)
 100  M ref

where Vi is the volume of specimen at particular suction or point of interest,

Gs is the specific gravity of the soil and ei is the void ratio at particular suc-
tion or point of interest.
For most sandy and gravelly soils, the instantaneous total volume can be
approximated by the initial total volume V0, as most sandy and gravelly
soils do not show volume change with change in matric suction. ASTM-­
D6836-­02 (20) recommends five methods to determine the SWCC (A, B, C,
D and E):

Method A: Reducing pore-­water pressure (Hanging column)

Method B: Axis-­ translation principle with volumetric measurement
(pressure plate)
Method C: Axis-­ translation principle with gravimetric measurement
(pressure plate)
Method D: Adjusting water content and measuring the total suction
(hygrometer)
Method E: Matric suction is applied by applying a centrifugal field
(centrifuge)

BS EN ISO 11274:2014 recommends four methods to determine the SWCC

(a, b, c, d):

1. Method a: Reducing pore-­water pressure (suction table)

2. Method b: Reducing pore-­water pressure (hanging column)
3. Method c: Axis-­translation principle with gravimetric measurement
(pressure plate)
4. Method d: Axis-­translation principle with gravimetric measurement
(pressure membrane)

In BS EN ISO 11274:2014, the results are expressed on a volume basis

based on the initial volume of the soil specimen used in the test.
Preferably one method should be used to determine the complete SWCC
of a soil. However, ASTM D6836-­02 (2016) allows two or more methods
(Methods A to E) to be used to obtain the SWCC of a soil. The range of suc-
tion applicable for each method is shown in Table 10.7.
184 Laboratory tests for unsaturated soils

10.5.3 Suction intervals
The recommended suction intervals in ASTM D6836-­02 (2016) for Methods
A to E are summarised in Table 10.8. The recommended suction intervals
in ASTM D6836-­02 (2016) may be excessive to determine a unimodal
SWCC and insufficient to determine bimodal or multimodal SWCC. Zou
and Leong (2019) recommend the suction intervals given in Table 10.9 to
determine unimodal and bimodal SWCCs.

10.5.4 SWCC test on coarse-grained soils

SWCC test on coarse-­grained soil is considerably fast, and the ceramic disk
used to test a coarse-­grained soil is usually only up to 100 kPa. The volume
change of coarse-­grained soils due to the change in suction is negligible and
thus can be ignored. Since the volume change is negligible, there will be a
negligible amount of water coming out from the specimen until the applied
suction is higher than the AEV and the shapes of w-SWCC, ­θ-SWCC and
S-SWCC are similar.
The test is commonly performed on a reconstituted specimen which
makes it possible to repeat the test. The saturation of the specimen can be
done by either sprinkling water from the top of the specimen or by using a
burette to allow the water to flow up the specimen from the bottom of the
specimen. The water level in the burette should be maintained equal to or
slightly higher than the specimen height.
Coarse-­grained soils have AEV as low as less than 1 kPa up to around 20
kPa, depending on the grain-­size distribution and the density of the speci-
men. Thus, suction range of interest can be from 0.1 kPa up to 100 kPa. The
suction interval can be very difficult to determine as the differences between
AEV and sr can be very small. An incorrect suction interval may give a mis-
leading AEV. Figure 10.7 shows the same SWCC-­w with different suction
intervals. The AEV is located between 0.7 kPa and 1.3 kPa. Thus, accurate
determination requires the suction interval to be around 0.1 kPa. Using a
higher suction interval will lead to misleading information on the shape of
the SWCC and the location of the AEV and sr.
To accurately pinpoint the location of the AEV, a trial-­and-­error approach
is recommended as follows:

1. Start by applying a higher suction interval (i.e., 1 kPa).

2. Once the AEV is exceeded, an obvious difference can be seen in the
amount of water that flows out during the suction application.
3. The soil is then re-­saturated, and 0.1 kPa suction is reapplied to
remove excess water.
4. Increase suction up to the largest suction prior to AEV that has been
applied from the previous test.
 Soil-water characteristic curve 185

Figure 10.7 SWCC of sandy soils determined by using different suction intervals.

5. The third and following suction is by using a smaller interval (i.e., half
of the previously used suction interval).
6. The test is repeated until a reasonably accurate AEV can be obtained.

10.5.5 SWCC test on fine-grained soils

SWCC test duration for fine-­grained soil is much longer compared to that
for coarse-­grained soils. For fine-­grained soils, AEV may range from around
less than 100 kPa up to more than 1,000 kPa. Thus, suction range of interest
can be from 1 kPa to 1,000,000 kPa.
However, it is also important to consider the fluctuation of suction at the
site. In locations where the groundwater table is relatively shallow, the range
of suction of interest can be quite low. On the other hand, when the ground-
water level is quite deep, the range of suction of interest can be very high.
Countries in temperate zones may have a very deep groundwater level in
summer and a very shallow groundwater level in spring. Thus, a larger range
of suction is expected.

10.5.6 Accuracy
The accuracy of the SWCC is dependent on a number of factors:

a) Condition of the soil.

b) Condition of the apparatus especially where ceramic plates or mem-
branes are used. The ceramic plates may not be fully saturated or
have minor defects not visible to the naked eye that may affect the
test results. Generally, new ceramic plate performs well, but their
186 Laboratory tests for unsaturated soils

performance deteriorates over time due to soil or water impurities

clogging up the pores or shrinkage cracks appearing over time. Pres-
sure membrane is susceptible to deterioration due to chemicals present
in the pore water or algae growth.
c) Size of the specimen. BS EN ISO 11274:2014 recommends that speci-
men height should not exceed 5 cm whereas ASTM-­D6836-­02 (20)
specifies a minimum height of 0.5 cm. A taller specimen will incur a
longer equilibrium time.
d) Temperature change can cause condensation and re-­absorption making
the equilibrium condition of the soil specimen difficult to determine.
e) Humidity of the supplied air pressure. It is well-­known that a soil
specimen under an applied suction using the axis-­translation principle
will reach equilibrium asymptotically or never in the theoretical sense
(Topp et al. 1993; Gee et al. 1992). Hence, equilibrium is assumed to
be reached when the weight change of a specimen is less than the reso-
lution of the weight or volume measuring instrument.
f) Usually, the salt in the pore water of the soil specimen is not consid-
ered. Hence, the osmotic pressure is not normally considered when
determining the SWCC.
g) Good capillary contact must always be maintained between the soil
specimen and the suction table, ceramic plate or membrane. Poor con-
tact results in water contents higher than expected for a given suction.
h) Air may become entrapped when wetting a soil specimen, resulting in
a lower water content at a given suction.

10.6 ESTIMATION OF SWCC

Due to the required testing duration, it is desired to estimate SWCC by using

basic soil properties. Grain-­size distribution is commonly used as the basic
information in estimating SWCC (and is commonly referred to as pedo-­
transfer function).
However, SWCC is not a unique curve, as it is affected by the initial void
ratio of the soil or whether it is for the drying or wetting path. Thus, it is
important to verify whether the estimation method which is used complies
with the behaviour of SWCC.
In the following section, estimation methods such as pedo-­transfer func-
tion and the one-­point estimation method will be described.

10.6.1 Pedo-transfer function
Pedo-­transfer function is used to estimate SWCC by using soil properties
such as grain-­size distribution. Examples of pedo-­transfer function are given
in Table 10.6.
Table 10.6 Pedo-transfer function

Reference Equation Parameters

Aubertin et al. (2003) Sr  Sc  1  Sc  Sa Sr = degree of saturation, unitless
MK Model Sc = saturation associated with the capillary component, unitless

Sa  min 0.5 1  Sa   1  Sa   ,
   Sa = saturation associated with the adhesive component, unitless
hc0 = equivalent capillary height which is related to an equivalent pore
 m
2  diameter and the solid surface area (cm)
 
h 
2
  
Sc  1    c 0   1 exp  m hc 0 h   hs = soil suction represented as a head (cm)
  hs    s
  hr = residual soil suction represented as a head (cm)
  hn = normalisation parameter (1 cm)

   
2
 ln 1  hs  hc 0 3 h0 suction head represented as dry soil condition (cm)
 hr  hn ac = adhesion coefficient, unitless
Sa  ac 1 


ln 1  h0
  

hr  e
1/ 3 hs
hn
1
6 m = pore-­size coefficient, unitless
e = void ratio
D10 = diameter corresponding to 10% passing on grain-­size distribution curve
hn = 1 cm
D60 = diameter corresponding to 60% passing on grain-­size distribution curve
h0 = 107 cm CU = uniformity coefficient

Soil-water characteristic curve

Coarse-­grained soils: LL = Liquid limit (%)
hc 0  cm  
0.75 ρs = density of the soil solids (kg/m3)
1  1.17 log  CU   eD10
1
m= ; ac = 0.01; hr  cm   0.86h1c.02
CU
D
CU = 60
D10
For Fine-­grained soils:
0.15 s 1.45
hc 0  cm   LL
e
m = 0.0003; ac = 0.0007

187
188 Laboratory tests for unsaturated soils

Table 10.7 Suction range for different methods

ASTM-­D6836-­02 (2016) BS EN ISO 11274:2014

Method Suction range (kPa) Method Suction range (kPa)
A 0 – 80 a 0–50
B 0 – 1,500 b 0–20
C 0 – 1,500 c 5–1,500
D 1,000 – 100,000 d 33–1,500
E 0 – 120

Table 10.8 Summary of ASTM D6836-02 (2008) methods for determining SWCC

Method A1 Methods B & C Method D Method E2

Suction level 0.05 10 Not specified 0.5
applied (kPa) 0.2 50 2
0.4 100 8.5
1 300 34
2 500 120
4 1000
6 1500
10
15
20
40
1 Converted from 5, 20, 40, 100, 200, 400, 600, 1000, 1500, 2000 and 4000 mm of water.
2 Estimated from angular velocities of 100, 200, 400, 800 and 1500 rpm and may vary with
centrifuge.

Figure 10.8 shows the ability of the Modified Kovacs (MK) Model
(Aubertin et al. 2003) in estimating w-­SWCC under different density by
using sand specimens data from Wijaya and Leong (2017). Figure 10.8
shows that the MK Model gives a reasonable w-­SWCC regardless of density
despite the estimated w-­SWCCs do not perfectly merge into a single virgin
drying curve. It is important to note that the MK Model estimates the degree
of saturation instead of gravimetric water content. Thus, directly converting
S-­SWCC into w-­SWCC is only possible for coarse-­grained soil or soil which
does not change in volume when suction is changed. When the soil changes
in volume, estimation of the shrinkage curve is required.
Table 10.9 Recommended suction intervals from Zou and Leong (2019)

Method A B&C D E*
SWCC Type Unimodal Bimodal Unimodal Bimodal Unimodal Bimodal Unimodal Bimodal
0.2 0.2 1 1 500 500 0.2 0.2
1 0.5 10 4 1,000 1,000 1 0.5
10 1 100 10 10,000 2,000 10 1
Suction Level 30 2 500 40 50,000 5,000 40 2
Applied (kPa) 80 5 1,500 100 100,000 10,000 120 5
10 200 20,000 10

Soil-water characteristic curve

30 500 50,000 50
80 1,500 100,000 120
* Suction levels are suggestions only and depend on centrifuge.

189
190 Laboratory tests for unsaturated soils

Figure 10.8 C
 omparison between estimated SWCC using MK Model (Aubertin
et al., 2003) and SWCC data from Wijaya and Leong (2017).
(a) w-SWCC of sand specimens under different density. (b) S-SWCC
of sand specimens under different density.

10.6.2 One-point method
One-­point method (e.g., Chin et al. 2010) is a “bridge” between a com-
plete laboratory test and a pedo-­transfer function. The problem with pedo-­
transfer function is its difficulty in estimating the correct SWCC, and it is
possible for the entire estimated SWCC to be completely different from the
actual SWCC. In the one-­point method, only one experimental point of the
 Soil-water characteristic curve 191

SWCC is needed (and thus, there is at least one point that is correct). Chin et
al. (2010) recommend different equations to estimate Fredlund and Xing’s
(1994) equation parameters for fine-­grained and coarse-­grained soils based
on the percentage of soil passing standard sieve No. 200 (P200). Fine-­grained
soils are defined as those soils with P200 ≥ 30%.
For P200 ≥ 30%:

af  2.4  x   722 (10.8a)

nf  0.07  x 
0.4
(10.8b)

mf  0.015  x 
0.7
(10.8c)

 r  914 exp  0.002  x   (10.8d)

For P200 <30%

af  0.53  D50 
0.96
(10.9a)

nf = x (10.9b)

mf  0.23 ln  x   1.13 (10.9c)

 r  100kPa, (10.9d)

where x is a curve-­fitting parameter which ranges from 0 to 300.8 and is

obtained by curve-­fitting the experimental SWCC point. Chin et al. (2010)
recommend the experimental SWCC point to be determined at suction of
100 kPa or 500 kPa (preferably 500 kPa) for fine-­grained soils and 10 kPa
for coarse-­grained soils.

REFERENCES

ASTM-­D4254-­14 (2014). Standard test methods for minimum index density and
unit weight of soils and calculation of relative density. West Conshohocken, PA:
ASTM Book of Annual Standards.
ASTM-­D6836-­02 (2008). Standard test method for determination of soil water char-
acteristic curve for desorption using hanging column, pressure extractor, chilled
mirror hygrometer, or centrifuge. West Conshohocken, PA: ASTM Book of Annual
Standards.
Aubertin, M., Mbonimpa, M., Bussière, B., & Chapuis, R. P. (2003). A model to
predict the water retention curve from basic geotechnical properties. Canadian
Geotechnical Journal, 40(6): 1104–1122.
192 Laboratory tests for unsaturated soils

Brooks, H. & Corey, A. T. (1964). Hydraulic properties of porous media. Fort

Collins, CO: Colorado State University. Hydrology Paper No. 3.
BS EN ISO 11274 2014 Soil quality. Determination of the water-­retention character-
istic. Laboratory methods. UK: BSI.
Chin, K. B., Leong, E. C., & Rahardjo, H. (2010). A simplified method to esti-
mate the soil-­ water characteristic curve. Canadian Geotechnical Journal,
47(12):1382–1400.
Feng, M. & Fredlund, D. G. (1999). Hysteretic influence associated with thermal
conductivity sensor measurements. Proceedings from Theory to the practice of
Unsaturated Soils Mechanics in Association with the 52nd Canadian Geotechnical
Conference and the Unsaturated Soil Group, Regina, Sask 14: 14–20.
Fredlund, D. G., Rahardjo, H., Leong, E. C., & Ng, C. W. W. (2001). Suggestions
and recommendations for the interpretation of soil-­water characteristic curves.
In Proceedings of the Fourteen Southeast Asian Geotechnical Conference, Hong
Kong, vol. 1: pp. 503–508.
Fredlund, D. G., Shuai,F., & Feng, M. (2000). Use of a new thermal conductiv-
ity sensor for laboratory suction measurement. In Proceedings of the Asian
Conference in Unsaturated Soils, UNSAT ASIA 2000. Singapore: A.A. Balkema,
pp. 275–280.
Fredlund, D. G. & Xing, A. (1994). Equations for the soil-­water characteristic curve.
Canadian Geotechnical Journal, 31(4): 521–532.
Frost, J. & Park, J. (2003). A critical assessment of the moist tamping technique.
Geotechnical Testing Journal, 26(1): 57–70. https://doi.org/10.1520/GTJ11108J
Gallipoli, D., Wheeler, S. J., & Karstumen, M. (2003). Modelling the variation
of degree of saturation in a deformable unsaturated soil. Géotechnique, 53(1):
105–11.
Haines, W. B. (1930). Studies in the physical properties of soil – V:The hysteresis
effect in capillary properties and the modes of water distribution associated there-
with. Journal of Agricultural Science, 20(1): 97–116.
Hillel, D. (2003). Introduction to environmental soil physics. San Diego, CA:
Academic Press, Elsevier Inc.
Klausner, Y. (1991). Fundamentals of continuum mechanics of soils. New York:
Springer-­Verlag.
Kohgo, Y. (2003). Review of constitutive models for unsaturated soils and initial-­
boundary value analysis. In Proceedings of the 2nd Asian Conference on
Unsaturated Soils. Karube, D., Iizuka, A., Kato, S., Kawai, K., & Tateyama, K.
(eds.) UNSAT-­ASIA 2003, Osaka-­Japan.
Ladd, R. (1978). Preparing test specimens using undercompaction. Geotechnical
Testing Journal, 1(1): 16–23. https://doi.org/10.1520/GTJ10364J
Lagioia, R., Sanzeni, A., & Colleselli, F. (2006). Air, water and vacuum pluviation of
sand specimens for the triaxial apparatus. Soils Foundation, 46(1): 61–67.
Li, X. (2009). Dual-­porosity structure and Bimodal Hydraulic property functions for
unsaturated coarse granular soils. In Civil engineering. Hong Kong University of
Science and Technology, Hong Kong, vol. Ph.D.
Li, X., Li, J. H., & Zhang, L. M. (2014). Predicting bimodal soil-­water characteristic
curves and permeability functions using physically based parameters. Computers
and Geotechnics, 57: 85–96.
Mualem, Y. (1977). Extension of the similarity hypothesis used for modeling the soil
water characteristics. Water Resources Research, 9: 1324–1331.
 Soil-water characteristic curve 193

Ng, C. W. W. & Menzies, B. (2007). Advance unsaturated soil mechanics and engi-
neering. Abingdon, Oxon: Taylor and Francis.
Pham, H. Q., Fredlund, D. G., & Barbour, S. L. (2003). A practical hysteresis model
for the soil-­water characteristic curve for the soils with negligible volume change.
Géotechnique, 53: 293–298.
Pham, H. Q., Fredlund, D. G., & Barbour, S. L. (2005). A study of hysteresis models for
soil-­water characteristic curves. Canadian Geotechnical Journal, 42(6): 1548–1568.
Salager, S., El Youssoufi, M. S., & Saix, C. (2010). Definition and experimental deter-
mination of a soil-­water retention surface. Canadian Geotechnical Journal, 47(6):
609–622.
Satyanaga, A., Rahardjo, H., & Leong, E. C. (2013). Water characteristic curve of
soil with bimodal grain-size distribution. Computers and Geotechnics, 48: 51–61.
https://doi.org/10.1016/j.compgeo.2012.09.008.
Schindler, U. (1980). Ein Schnellverfahren zur Messung der Wasserleitfähigkeit im teil-
gesättigten Boden an Stechzylinderproben. Archiv für Acker-­und Pflanzenbauund
Bodenkunde, 24(1): 1–7 (In german).
Tarantino, A. (2009). A water retention model for deformable soils. Géotechnique,
59: 751–762.
Topp, G. C., Galganov, Y. T., Ball, B. C., & Carter, M. R. (1993). Soil water desorp-
tion curves. In M. R. Carter (Ed.), Soil sampling and methods of analysis. Ann
Arbor, MI: Lewis Publishers, pp. 569–579.
Vaid, Y. & Negussey, D. (1988). Preparation of reconstituted sand specimens. In R.
Donaghe, R. Chaney, & M. Silver (Eds.), Advanced triaxial testing of soil and
rock. West Conshohocken, PA: ASTM International, pp. 405–417.
Van Genuchten, M. T. (1980). A closed-­form equation for predicting the hydraulic con-
ductivity of unsaturated soils. Soil Science Society of America Journal, 44(5): 892–898.
Wijaya, M. (2017). Compression, shrinkage and wetting-­induced volume change of
unsaturated soils. In Civil and environmental engineering. Singapore: Nanyang
Technological University, vol. Ph.D.
Wijaya, M. & Leong, E. C. (2016). Equation for unimodal and bimodal soil-­water
characteristic curves. Soils and Foundations, 56(2): 291–300.
Wijaya, M. & Leong, E. C. (2017). Modelling the effect of density on the unimodal
soil-­water characteristic curve. Géotechnique, 67(7): 637–645.
Wijaya, M., Leong, E. C., & Rahardjo, H. (2015). Effect of shrinkage on air-­entry
value of soils. Soils and Foundations, 55(1): 166–180.
Zhou, A. N., Sheng, D., & Carter, J. P. (2012). Modelling the effect of initial density
on soil-­water characteristic curves. Géotechnique, 62: 669–680.
Zhou, W.-H., Yuen, K.-V., & Tan, F. (2014). Estimation of soil-­water characteristic
curve and relative permeability for granular soils with different initial dry densi-
ties. Engineering Geology, 179(0): 1–9.
Zou, L. & Leong, E. C. (2019). A classification tree guide to soil-­water characteristic
curve test for soils with bimodal grain -size distribution, Geotechnical Engineering
Journal of the SEAGS & AGSSEA, 50(1): 28–36.

FURTHER READING

Wijaya, M. & Leong, E. C. (2017). Modelling the effect of density on the unimodal
soil-­water characteristic curve. Géotechnique, 67(7): 637–645.
Chapter 11

Permeability
Steady-state methods

11.1 BACKGROUND

Permeability is one of the important hydraulic properties of unsaturated

soils. It governs seepage (transient and steady-­state) and affects the mag-
nitude and distribution of pore-­water pressures in unsaturated soils. Water
flow in unsaturated soils is only through the water phase, and, hence, the
water phase must be continuous in order for flow in the liquid form to
occur. The rate of water flow through saturated and unsaturated soils is
governed by Darcy’s law (Fredlund et al. 2012). In saturated soils, the coef-
ficient of permeability Ks is a function of the void ratio e (Taylor 1948; Lame
and Whitman 1979). However, the coefficient of permeability for unsatu-
rated soil depends on any two of three possible volume-­mass properties:
void ratio e, gravimetric water content w and degree of saturation S (Lloret
and Alonso 1980; Fredlund 1981).
When determining the permeability of unsaturated soils, it is important to
review the various standards for determining permeability for saturated
soils. A summary of the range of saturated coefficient of permeability in the
standards is given in Table 11.1. Currently, ASTM D7664-­10 is the only
standard for measurement of unsaturated permeability of soils. The meth-
ods given in ASTM D7664-­10 are suitable for determining permeability of
unsaturated soils from the saturated coefficient of permeability to a perme-
ability of 10−11 m/s. Hence, all the conditions specified in the other standards
for determination of saturated permeability are equally applicable for unsat-
urated soils.
The methods described in ASTM D7664-­10 are considered non-­steady-­
state or transient methods which will be covered in the next chapter. In this
chapter, the steady-­state method of determining the permeability function of
unsaturated soil is discussed. Such a test can be conducted using a rigid-­wall
permeameter or a flexible-­wall permeameter. Because of the possibility of
specimen shrinkage from the wall of the rigid-­wall permeameter at high suc-
tions, a flexible-­wall permeameter is preferred. The triaxial cell offers a
convenient way to modify it for permeability measurement of unsaturated
soils. Such a system will be described in this chapter.

DOI: 10.1201/b22304-11 195

196 Laboratory tests for unsaturated soils

Table 11.1 V
 alidity range for determination of saturated
coefficient of permeability in standards

Range of saturated coefficient

Standard of permeability (m/s)

ASTM D2434 – 22 > 10−7

ASTMD5084 – 16a 10−11–10−6
ASTM D5856-­15 < 10−5
BS 1377-­5: 1990 10−2–10−5
BS 1377-­6: 1990 Not explicitly mentioned
BS EN ISO 17892-­11:2019 Full range

Table 11.2 Recommended maximum hydraulic gradient in standards

Hydraulic ASTM ASTM ASTM BS BS 1377–6: BS EN ISO

conductivity D5084 D5856 D2434−22 1377–5: 1990 17892-­
(m/s) − 16a − 15 1990 11:2019

> 1e-­5 - - Low 0.2 – 1* Not 1

1e-­5 – 1e−6 2 2 hydraulic explicitly 2
1e-­6 – 1e−7 5 5 gradient mentioned 5
(0.05) except for
1e-­7 − 1e−8 10 10 N.A clays soil, i 10
1e−8 − 1e−9 20 20 > 20+ 20
N.A.
< 1e−9 30 30 ≥ 30

* for ks range of 10−2 − 10−5 m/s

i = hydraulic gradient.

11.2 RELATED STANDARDS

1. ASTM D2434 – 22 Standard Test Methods for Measurement of

Hydraulic Conductivity of Coarse-­Grained Soils.
2. ASTM D5084 – 16a Standard Test Methods for Measurement of
Hydraulic Conductivity of Saturated Porous Materials Using a
Flexible Wall Permeameter.
3. ASTM D5856-­ 15 Standard Test Method for Measurement of
Hydraulic Conductivity of Porous Material Using a Rigid-­ Wall,
Compaction-­Mold Permeameter.
4. ASTM D6035/D6035M Standard Test Methods for Determining the
Effect of Freeze-­Thaw on Hydraulic Conductivity of Compacted or
Intact Soil Specimens Using a Flexible Wall Permeameter.
5. ASTM D6836 Standard Test Methods for Determination of the Soil
Water Characteristic Curve for Desorption Using Hanging Column,
Pressure Extractor, Chilled Mirror Hygrometer, or Centrifuge.
 Permeability 197

6. ASTM D7664 Standard Test Methods for Measurement of Hydraulic

Conductivity of Unsaturated Soils.
7. BS 1377-­ 5: 1990 Methods of Test for Soils for Civil Engineering
Purposes Part 5: Compressibility, Permeability and Durability Tests.
8. BS 1377-­ 6: 1990 Methods of Test for Soils for Civil Engineering
Purposes Part 6: Consolidation and Permeability Tests in Hydraulic
Cells and with Pore Pressure Measurement.
9. BS EN ISO 17892-­11:2019 Geotechnical Investigation and Testing.
Laboratory Testing of Soil Permeability Tests.

11.3 THEORY

During measurement of permeability, it is important to ensure that the

water content or degree of saturation of the soil does not change. This con-
dition can be assured using the axis-­translation technique (Hilf 1956) by
controlling air and water pressures independently such that the matric suc-
tion of the soil specimen is kept constant during the permeability measure-
ment. A summary of permeameters using such a technique to determine
the air and water coefficients of permeability for unsaturated soils is given
in Table 11.3. Air coefficient of permeability is measured sometimes as the
air permeability function is an inverse of the water permeability function
(Fredlund et al. 2012). It is possible to convert the air permeability function
to the water permeability function, but this book concentrates only on the
water phase; the flow of air is outside the scope of this book and will not be
discussed further.
In applying the axis-­translation technique, the porous elements in the tri-
axial cell will need to be replaced with high air-­entry ceramic disk to sepa-
rate the air and water phases. During permeability measurement, the matric
suction (hence, water content or degree of saturation) needs to be main-
tained constant and thus the permeability measurement can only be done
under the constant-­head condition. Under variable head conditions, the air
pressure must also vary in tandem to maintain constant matric suction
which can be very complicated. The principle of the permeability test is to
first equilibrate the soil specimen to the required matric suction at a speci-
fied net normal stress, as illustrated in Figure 11.1a. During measurement
(Figure 11.1b), the water pressures across the soil specimen are adjusted
such that the average pressure remains the same as before. Using Figure 11.1b
for illustration, the water pressure was adjusted to be 190 kPa at the top of
the specimen and 210 kPa at the bottom of the specimen such that the water
pressure difference is 20 kPa, giving a hydraulic gradient i across the soil
specimen as i = 20 kPa/(9.81 kN/m3 x 0.1m) = 20.4 while the average pore-­
water pressure in the specimen remains at 200 kPa (= (190 kPa + 210
kPa)/2). Hence, water will flow from the bottom of the soil specimen to the
top of the soil specimen under a hydraulic gradient of 20.4. When the
 198
Laboratory tests for unsaturated soils
Table 11.3 Summary of permeameter systems used.

Range of kw Range of ka Suction Measurement of volume

References measured (m/s) measured (m/s) range (kPa) change of specimen Soil tested

Klute (1965) Not mentioned Not mentioned 0–90 Not measured Not mentioned
Barden and Pavlakis (1971) 10−12–10−10 10−8–10−3 0–95 Not measured Compacted soils
Undisturbed clay, loam and sand-­
Fleureau and Taibi (1994) up to 10−8 up to 10−8 0–80 Not measured
kaolin mixture
Dane et al. (1998) Not measured up to 10−4 0–4 Not measured Sand
Not Using non-­contacting
Huang et al. (1998) 10−11–10−8 0–90 Silty sand
measured transducers
Not
Gan and Fredlund (2000) 10−10–10−5 0–100 Not measured Residual and saprolitic soils
measured
Using confining fluid
Samingan et al. (2003) 10−12–10−6 10-­7–10-­12 0–300 Residual soils
volume
Goh, et al. (2015). 10− –10−
9 10 Not measured 100–300 Not measured Sand-­kaolin
 Permeability 199

Specimen dimensions: Diameter = 100 mm, Length, L= 100 mm

= 350 kPa = 350 kPa

3 3
3 = 350 kPa

ua = 300 kPa , uw = 200 kPa ua = 300 kPa , uw = 190 kPa ua = 300 kPa , uw = 150 kPa

ua = 300 kPa , uw = 200 kPa ua = 300 kPa , uw = 210 kPa ua = 300 kPa , uw = 150 kPa

i = uw/ wL = 0 i = uw/ wL = 20.4 i = uw/ wL = 0

uw_avg = 200 kPa uw_avg = 200 kPa uw_avg = 150 kPa
ua – uw_avg = 100 kPa
ua – uw_avg = 100 kPa ua – uw_avg = 150 kPa
3 – ua = 50 kPa
3 – ua = 50 kPa 3
– ua = 50 kPa

(a) (b) (c)

Figure 11.1 P
 rinciple of constant-head permeability test for unsaturated soil
using axis-translation technique. (a) Start. (b) During measurement.
(c) Next suction.

volume of water flowing into the soil specimen, Qin, is equal to the volume
of water flowing out of the specimen, Qout, the steady-­state flow condition is
reached, and the water coefficient of permeability can be computed using
Darcy’s law:
Q
k= (11.1)
i ⋅ A⋅t

where
Q = average of Qin and Qout
t = time to obtain Q
A = cross-­sectional area of soil specimen

11.4 TEST METHOD

The soil specimen tested can be intact/undisturbed or reconstituted. For soil

specimen preparation, see Chapter 3. Suitable specimen size to be used for
permeability measurement is usually thin so that the steady-­state condition
200 Laboratory tests for unsaturated soils

can be attained sooner. Most standards specify height to diameter ratio of

1:1 but BS 1377:1990: Part 6 suggests that the height/diameter ratio of 1/2.5
to 1/4 can be used but should not exceed 1. The specimen height is limited
by the largest particle size which should not be more than 1/6 of the speci-
men height. For testing unsaturated soil, the specimen size should allow a
sufficient quantity of water to flow out of or into the soil specimen so that
the water volume change of the specimen is measurable. Hence, the speci-
men size should not be too small. A suitable specimen size is one with a
diameter of at least 50 mm and height to diameter ratio of at least 1:2.
As air and water pressures are controlled at both ends of the soil speci-
men, the high-­air entry (HAE) ceramic disk must allow air pressure to be
distributed across the soil specimen cross-­section. This can be affected by
having part of the cross-­sectional area of the HAE ceramic disk covered
with a porous element that is permeable to air as well, such as a sintered
bronze. Possible arrangements are shown in Figure 11.2. However, any of
these arrangements will reduce the cross-­sectional area for water flow and
increase the duration of the test. A better alternative is to etch shallow
grooves into the HAE ceramic disk, as shown in Figure 11.3. A modified
triaxial cell to conduct a permeability test for unsaturated soils is shown in
Figure 11.4.
A diffused air volume indicator is shown in Figure 11.4 for removing air
that diffused through the HAE ceramic disk and reappearing in the water
line. This is required when the test duration is very long and measuring the
water volume of the specimen is needed. Details of the diffused air volume
indicator can be found in Fredlund et al. (2012).

Figure 11.2 P
 ossible arrangements of HAE ceramic disk with sintered bronze
filter on the platens.
 Permeability 201

Figure 11.3 S uggested HAE ceramic disk for permeability test of unsaturated
soils.
To air compressor

Data acquisition system

A D C B
Triaxial permeameter

Vent

Diffused
air volume
indicator
(DAVI)
Soil specimen
High-air entry
disks
Cell
pressure
line
Drainage

Drainage Digital water

pressure/volume
Top flushing line
controller
Air pressure
transducer

Bottom flushing line Top pore-air

Top pore-water pressure line
pressure line
Water pressure Bottom pore-air
transducer pressure line
Top auto
volume
change Digital air pressure/volume controller
Bottom
indicator
auto
Differential volume
pressure change
transducer indicator

LEGEND:

Plumbing line Bottom pore-water

pressure line
Signal line

To de-aired water supply

supply

Figure 11.4 S chematic drawing of flexible-wall permeameter system for unsatu-

rated soils (from Agus et al. 2003).
202 Laboratory tests for unsaturated soils

The test procedures can be divided into four stages: setting up, saturation,
consolidation and measurement of water permeability.

1. Setting up
The apparatus is first prepared to ensure that the HAE ceramic disks
are fully saturated and the water lines are flushed of air bubbles. The
air valves of the triaxial cell are shut and the triaxial cell without
the soil specimen is filled with water. A cell pressure (at the air-­entry
pressure of the ceramic disk or the maximum air pressure that is ap-
plied during the test) is applied and water flows through the HAE
ceramic disks. Allow the water lines to discharge underwater in a
beaker of water to check for the presence of air bubbles. This is
repeated several times. If air bubbles cannot be effectively removed
in this way, half-­fill the triaxial cell and ensure that the top platen
is submerged in the water as well. Apply a vacuum of say 20 to 30
kPa at the top of the triaxial and leave overnight. Repeat the afore-
mentioned procedures of pressuring the ceramic disks without the
soil specimen to check the saturation condition of the ceramic disks.
Once the ceramic disks are fully saturated, connect the water lines,
ensuring that they are filled with water and there are no air bubbles
in them. If the ceramic disks are exposed for a long period of time,
place wetted cotton wools over them to minimise the possibility of
the ceramic disks desaturating.
Prepare the soil specimen; ensure that it is trimmed to the correct
dimensions. Measure the soil specimen’s dimensions and weight. De-
termine the initial water content of the soil specimen from the trim-
mings. Place the soil specimen into the triaxial cell and place a rub-
ber membrane over it using a membrane stretcher. Secure the rubber
membrane to the bottom pedestal with an O-­ring. Place the top platen
on the soil specimen, lift the membrane over the top platen and secure
with an O-­ring. Fill the triaxial cell with water.
2. Saturation
The soil specimen is then saturated by using a small back pressure
(about 5 kPa) to allow water to flow upwards through the soil speci-
men and drain out from the top. The soil specimen can be left in this
condition overnight. The saturation condition of the soil specimen can
be checked by determining the Skempton’s pore-­water pressure pa-
rameter B. The B value is measured by increasing the cell pressure and
then monitoring the corresponding change in pore-­water pressure. The
pore-­water pressure parameter B is given by the ratio between the
change in pore-­water pressure and the change in applied cell pressure.
The soil specimen is saturated when B is 1. Usually, a B value of 1 can-
not be obtained, especially for stiffer soils, and it is generally accepted
that the soil specimen is fully saturated when B > 0.95 in saturated soil
tests (Black and Lee 1973).
 Permeability 203

3. Consolidation
The soil specimen is then consolidated isotropically to the desired net
normal stress and matric suction. If the change in net normal stress and
matric suction is large, it can be done in several steps (minimum two
steps) to reach the desired values. Double drainage should be adopted
to reduce the consolidation time, as it might take several days at high
matric suction values. The consolidation process can be monitored by
recording the volume of water flowing out from the soil specimen with
time. Consolidation is deemed to be completed when there is negli-
gible water flow either into or out of the soil specimen. At this stage,
the soil specimen is said to be in equilibrium at the applied net normal
stress and matric suction.
4. Measurement of water permeability
The water coefficient of permeability is determined by applying a hy-
draulic gradient across the soil specimen by changing the water pres-
sures applied to the top and bottom of the specimen such that the
average water pressure remains the same as before the change (see
Figure 11.1). The direction of water flow can be upwards or down-
wards but upward water flow is preferred as there is less likelihood
for water to enter the air pressure lines with upward water flow. The
maximum hydraulic gradient depends on the permeability of the soils
(see Table 11.1). High hydraulic gradient is required to reduce the
test duration as the rate of water flow is very slow. It is important to
monitor the water inflow and outflow to establish the steady-­state
flow condition. An example is shown in Figure 11.5.

0.020

0.015
Water flow rate, vw (cm /h)
3

0.010

0.005

0.000
inflow

-0.005 outflow
steady state

-0.010
0 25 50 75 100 125 150
Elapsed time, t (h)
Figure 11.5 E xample of water inflow and outflow during permeability test to
establish steady-state flow condition.
204 Laboratory tests for unsaturated soils

At steady-­state flow condition, the water coefficient of permeability can be

obtained using Darcy’s law (Equation 11.1). The impedance of the ceramic
disks needs to be considered in the determination of the water coefficient of
permeability of the soil when the water coefficient of permeability of the soil
is close to that of the ceramic disks or lower. The impedance of the ceramic
disk can be considered by treating the ceramic disk – soil – ceramic disk as
a three-­layered system. For one-­dimensional flow, the flow velocity v through
each layer must be the same as given by Equation 11.2.

=
v v=
b v s = vt (11.2)

where vb, vs and vt are the flow velocities through the bottom ceramic disk,
soil and top ceramic disk, respectively.
The loss in total head across the three layers H is equal to the sum of the
total head loss in each layer, i.e.:

H = Hb + H s + Ht (11.3)

where Hb, Hs and Ht are the total head losses through the bottom ceramic
plate, soil and top ceramic plate, respectively.
Using Equations 11.2 and 11.3 and applying Darcy’s law, the coefficient
of permeability of the three-­layered system is given by Equation 11.4.

ht + hs + hb
k= (11.4)
ht hs hb
+ +
kt ks kb

where hb, hs and ht are the thicknesses of the bottom ceramic plate, soil and
top ceramic plate, respectively.
The water coefficient of permeability of the soil specimen kw can be
obtained if the coefficients of permeability of the ceramic disks, kt and kb
and the thicknesses of the ceramic disks, hb and ht, are known. By rearrang-
ing Equation 11.4, the water coefficient of permeability is given by Equation
11.5.

hs
kw = (11.5)
 h  ht hb  
 −  + 
 k  kt kb  

Equation 11.5 shows that if the coefficients of permeability of the disks

are much higher than the coefficient of permeability of the soil specimen, kw
is equal to k, which is measured. However, if the coefficients of permeability
of the disks are of the same order or lower than the coefficient of permeabil-
ity of the soil specimen, the measured k is for the disks rather than for the
 Permeability 205

soil specimen. Hence it is important to check when the water coefficient of

permeability of the soil approaches that of the ceramic disks.
A temperature correction factor can be applied to the water coefficient of
permeability of the soil as the viscosity of water changes with temperature.
Usually, the water coefficient of permeability is determined at the standard
temperature of 20°C.

REFERENCES

Barden, L. & Pavlakis, G. (1971). Air and water permeability of compacted unsatu-
rated cohesive soils. Journal of Soil Science, 22: 302–318.
Benson, C. H. & Gribb, M. M. (1997). Measuring unsaturated hydraulic conductiv-
ity in the laboratory and field. In S. Houston & D. G. Fredlund (Eds.), Unsaturated
soil engineering practice, ASCE, geotechnical special publication, No. 68: 113–168.
Black, D. K. & Lee, K. L. (1973). Saturating laboratory samples by back pressure.
Journal of the Soil Mechanics and Foundation Division, ASCE, 99(SMII): 75–93.
Brooks, R. H. & Corey, A. T. (1964). Hydraulic properties of porous medium.
Hydrology Papers. No. 3. Fort Collins, Co: Colorado State University.
Childs, E. C. & Collis-­George, G. N. (1950). The permeability of porous materials.
Proceeding of the Royal Society, 210A: 392–405.
Dane, J. H., Hofstee, C., & Corey, A. T. (1998). Simultaneous measurement of capil-
lary pressure, saturation and effective permeability of immiscible liquids in porous
media. Water Resources Research, 34(12): 3687–3692.
Fleureau, J. M. & Taibi, S. (1994). New apparatus for the measurement of water-­
air permeabilities. In Proceeding of Geo-­environmental Conference, Edmonton,
Canada.
Fredlund, D. G. (1981). Panel discussion: Ground water and seepage problems.
In Proceedings of the 10th International Conference on Soil Mechanics and
Foundation Engineering, Stockholm, Sweden, vol. 4, pp. 629–641.
Fredlund, D. G. & Rahardjo, H. (1993). Soil mechanics for unsaturated soils. New
York: John Wiley & Sons, Inc.
Fredlund, D. G., Rahardjo, H., & Fredlund, M. D. (2012). Unsaturated soil mechan-
ics in engineering practice. New York: John Wiley & Sons, Inc..
Gallage, C., Kodikara, J., & Uchimura, T. (2013). Laboratory measurement of
hydraulic conductivity functions of two unsaturated sandy soils during drying
and wetting processes. Soils and Foundations, 53(3): 417–430.
Gan, J. K. M. & Fredlund, D. G. (2000). A new laboratory method for the measure-
ment of unsaturated coefficients of permeability of soils. In H. Rahardjo, D. Toll,
& E. C. Leong (Eds.), Unsaturated soils for Asia. Rotterdam: Balkema, 381–386.
Goh, S. G., Rahardjo, H., & Leong, E. C. (2015). Modification of triaxial appara-
tus for permeability measurement of unsaturated soils. Soils and Foundations,
Japanese Geotechnical Society, February, 55(1): 63–73.
Huang, S., Fredlund, D. G., & Barbour, S. L. (1998). Measurement of the coefficient
of permeability for a deformable unsaturated soil using a triaxial permeameter.
Canadian Geotechnical Journal, 35: 411–425.
Iwata, S., Tabuchi, T., & Warkentin, B. P. (1995). Soil-­water interactions. New York:
Marcel Dekker., Inc.
206 Laboratory tests for unsaturated soils

Kaye, G. W. C. & Laby, T. H. (1973). Tables of physical and chemical constant and
some mathematical functions. 14th Edition. London: Longman.
Klute, A. (1965). Laboratory measurement of hydraulic conductivity of unsaturated
soils. In C. A. Black, D. D. Evans, J. L. White, L. E. Esminger, & F. E. Clark (Eds.),
Method of soil analysis. Monograph 9. Madison, Wisconsin: American Society of
Agronomy, 1: 253–261.
Lambe, T. W. & Whitman, R. V. (1968). Soil mechanics. New York: John Wiley &
Sons, Inc.
Lloret, A. & Alonso, E. E. (1980). Consolidation of unsaturated soils including
swelling and collapse behaviour. Géotechnique, 30(4): 449–477.
Rahimi, A., Rahardjo, H., & Leong, E. C. (2015). Effect of range of soil-­water char-
acteristic curve measurements on estimation of permeability function. Engineering
Geology, February, 185: 96–104.
Richards, L. A. (1931). Capillary conduction of liquids through porous medium.
Journal of Physics, 1: 318–333.
Samingan, A. S., Leong, E. C., & Rahardjo, H. (2003). A flexible wall permeameter
for measurements of water and air coefficients of permeability of residual soils.
Canadian Geotechnical Journal, June, 40(3): 559–574.
Taylor, D. W. (1948). Fundamentals of soil mechanics. New York: Wiley.

FURTHER READING

Leong, E. C. & Rahardjo, H. (1997). Permeability functions for unsaturated soils.

ASCE Journal of Geotechnical and Geoenvironmental Engineering, 123(12):
1118–1126.
Chapter 12

Permeability
Transient-state methods

12.1 BACKGROUND

Measurement of the coefficient of permeability of unsaturated soils by

the steady-­state methods, though more accurate than the transient-­state
(unsteady-­state) methods, is more difficult and time-­consuming. The condi-
tion for steady state is achieved when the inflow is equal to outflow. In some
cases, it is near impossible to obtain the steady-­state condition. To partially
alleviate the problem of long testing time, several transient-­state methods
have been developed to measure the coefficient of permeability of unsatu-
rated soils (permeability function). In some of these methods, a mixed-­phase
formulation is used to obtain both the soil-­water characteristic curve and
the permeability function of the soil (e.g., Garnier et al. 1997; Fujimaki
and Inoue 2003). As this book concentrates on only the water phase, such
mixed-­phase formulations or methods involving evaporation (Wind 1968;
Schindler et al. 2010; Wendroth et al. 1993; Minasny and Field 2005) are
not included.
A review of transient-­state methods to measure unsaturated permeability
can be found in Benson and Gribb (1997). Transient-­state methods include
Bruce-­Klute adsorption, sorptivity, outflow, instantaneous profile and ther-
mal methods. There are many variations of the outflow method. These
include multistep, one-­step and continuous outflow methods. Due to the
advances in computer modelling, it is possible to combine experimental
results and numerical modelling to better estimate unsaturated permeability
(Kool et al. 1987). Examples are the HYPROP method (Schindler et al. 2010;
Tian et al. 2019), XMSO-­EVA method (Schelle 2011) and TWRI method
(Wayllace and Lu 2012). Such methods are outside the scope of this book.
The transient-­state methods involved two types of main apparatuses: a
soil column which could be placed vertically or horizonally (Bruce-­Klute
adsorption, instantaneous profile, thermal) or a cell containing a soil speci-
men (sorptivity, outflow). The soil column and cell apparatus have been
adopted in ASTM D7664-­10 (reapproved 2018) for the measurement of
permeability of unsaturated soils. In the standard, the soil column for the
instantaneous profile method is described as Method A, and the cell for the

DOI: 10.1201/b22304-12 207

208 Laboratory tests for unsaturated soils

multistep outflow method is described as Method B. In Method A, the addi-

tional instruments required are suction and water content measurement
devices. In Method B, the cell adopted is similar to a pressure plate appara-
tus where matric suction is controlled using the axis-­translation technique
(see Chapter 11). Additional instruments required include flow volume mea-
surement and, optionally, suction measurement.
A Method C is also described in ASTM D7664-­10. Method C involves the
use of a centrifuge to impose a steady-­flow condition. The centrifuge used
can be as big as a geotechnical centrifuge or as small as a benchtop centri-
fuge. In Method C, a special permeameter will have to be constructed, and
the test requires a device to provide an infiltration flow and devices to mea-
sure matric suction and water content of the soil specimen in the centrifuge
at steady-­state flow condition (Zornberg and McCartney 2010; McCartney
and Zornberg 2010). However, it was found that the steady-­state condition
was not fulfilled in the centrifuge permeameter at low G-­ levels, and a
transient-­state analysis should be performed instead (Parks et al. 2011).
Transient analysis using a centrifuge analysis can provide “acceptable” char-
acterisation of the soil-­water characteristic curve and permeability function
over wide suction and water content ranges. Other limitations of measuring
permeability of unsaturated soils in a geotechnical centrifuge include limit of
reading for time, total mass, inflow mass, outflow mass, rate of inflow, fluid
held by surface tension in effluent reservoir and on reservoir base plate,
evaporation effects (which increases with the increasing G-­level) and side-
wall leakage (Timms et al. 2014). In view of the requirements and limitations
of the centrifuge permeameter, this method is omitted from this chapter.

12.2 RELATED STANDARDS

1. ASTM D2434 – 22 Standard Test Methods for Measurement of

Hydraulic Conductivity of Coarse-­Grained Soils.
2. ASTM D5084 − 16a Standard Test Methods for Measurement of
Hydraulic Conductivity of Saturated Porous Materials Using a
Flexible Wall Permeameter.
3. ASTM D5856-­ 15 Standard Test Method for Measurement of
Hydraulic Conductivity of Porous Material Using a Rigid-­ Wall,
Compaction-­Mold Permeameter.
4. ASTM D6035/D6035M Standard Test Methods for Determining the
Effect of Freeze-­Thaw on Hydraulic Conductivity of Compacted or
Intact Soil Specimens Using a Flexible Wall Permeameter.
5. ASTM D6836 Standard Test Methods for Determination of the Soil
Water Characteristic Curve for Desorption Using Hanging Column,
Pressure Extractor, Chilled Mirror Hygrometer, or Centrifuge.
6. ASTM D7664 Standard Test Methods for Measurement of Hydraulic
Conductivity of Unsaturated Soils.
 Permeability 209

12.3 INSTANTANEOUS PROFILE METHOD

The instantaneous profile method using a soil column is described in ASTM

D7664-­10 Method A. The instantaneous profile method makes use of mass
continuity to evaluate the instantaneous water content and suction profiles
of the soil column one-­dimensional (1D) water flow condition. The soil col-
umn can start at different initial conditions, and different boundary con-
ditions can be imposed during the test. Four conditions are described in
ASTM D7664-­10. These are infiltration from the top of a dry soil column
(Method A1), imbibition from the bottom of a dry soil column (Method
A2), drainage from the bottom of a saturated soil column (Method A3) and
evaporation from the top of a saturated soil column (Method A4). During
the test, the suctions and water contents at various levels in the soil column
are monitored. Depending on the conditions imposed and the soil type, the
test duration can range from one week to several weeks.
As a 1D water flow condition is assumed, the void ratio of the soil should
not change during the test, and side wall leakage should not be present.
Hence, the instantaneous profile method using a soil column precludes high
plasticity soils which may experience change in void ratio during the test.
Clays which have very low permeability are also not possible to test using
the instantaneous profile method because of the prohibitively long testing
time and the requirement to use a long soil column. As the soil column gets
taller, it is increasingly more difficult to ensure that the soil column is homo-
geneous. If suction is directly measured during the test, the suction range is
limited to being less than 80 kPa (limitation of tensiometer).
The aforementioned requirements/limitations are specifically listed in
ASTM D7664-­10 under each sub-­method. Methods A1 and A2 are suitable
for fine-­grained sands and for low-­plasticity silts, but the permeability mea-
surement range is restricted within the matric suction range of tensiometers,
i.e., 0 to 80 kPa. Methods A3 and A4 are suitable for fine-­and coarse-­
grained sands. Methods A3 can measure permeability for the matric suction
range from 0 to 200 kPa, while Method A4 can measure permeability for
the matric suction range from 0 to 1000 kPa. For Method A4, desiccation
shrinkage and cracking may occur as evaporative flux is applied on the top
surface of the soil column.

12.3.1 Test set-up
The soil column container can be a cylinder made of non-­reactive metals,
acrylic or PVC (ASTM D7664-­10). The use of acrylic is preferred as the
water movement can be seen (Yang et al. 2004; Krisdani 2009). The soil
column can be prepared inside the container by compaction, wet tamping or
pluviation. Hence, the container should be strong enough to withstand the
stresses imposed during soil column preparation. A typical soil column set-
­up is shown in Figure 12.1. There should be provisions for the installation
210 Laboratory tests for unsaturated soils

Water content sensors (TDR, capacitance)

Suction sensors (tensiometers)

To manometer
or Mariotte bottle
Computer and
data acquisition
unit Water volume outflow
measurement

Figure 12.1 Typical soil column set-up for instantaneous profile method.

of suction and water content sensors along the soil column. Typical sen-
sors for direct suction measurement are small-­tip tensiometers which have
quick response to suction changes (Yang et al. 2004; Krisdani et al. 2009),
but these sensors are limited to suctions less than 80 kPa. Typical sensors
for water content measurement are time-­domain reflectometry (TDR) sen-
sors (Yang et al. 2004; Krisdani et al. 2009) or capacitance sensors (CSs;
Caicedo 2017). The holes are made in the cylinder so that the sensor cable
or non-­sensing part of the sensor can exit the cylinder. For tensiometer, the
diameter of the hole should be bigger than the diameter of the sensing tip
so that the tensiometer can pass through from the outside of the cylinder. A
plug can be provided to make a watertight seal. The use of a cylinder with
a larger diameter will reduce the need for the plug design to consider the
curvature of the cylinder surface or else an additional block will need to
be introduced at the cylinder surface to provide a flat contact surface for
the plug to achieve a watertight seal (see Figure 12.2). For the TDR sensor,
the hole should be large enough to accommodate the connector. The TDR
probe is disconnected and placed on the inside of the cylinder, and then the
connector is fixed from the outside of the cylinder. Usually, an additional
O-­ring at the connector is sufficient to achieve a watertight seal. For the CSs,
a hole large enough for the cable to pass through is sufficient. A plug similar
 Permeability 211

Acrylic block to provide flat surface for plug to seal

Plug
O-ring

Through hole
for tensiometer
or cable

O-ring Plug

Wall of soil column cylinder

(a) (b)

Figure 12.2 P
 lug for watertight seal of small-tip tensiometer or cable. (a) Side view.
(b) Top view.

to that for the tensiometer can be used (Figure 12.2). The minimum internal
diameter of the cylinder recommended by ASTM D7664-­10 is 200 mm, as
it can serve to minimise side wall leakage, as well as detect preferential flow
of the water due to macro features in the soil column. ASTM D7664-­10
recommends that the top sensors should be within 10 mm of the surface of
the soil column and the bottom sensors within 5 mm of the bottom of the
soil column, and at least three sensors (or sets of sensors) evenly spaced in
between. The height of the soil column should be at least 0.5 m for coarse-­
grained soils and 1 m for fine-­grained soils in order to capture representa-
tive distribution of matric suction and water content with the height of the
soil column. A taller soil column is required if the test needs to achieve the
unit hydraulic gradient condition in the upper part of the soil column. The
bottom of the soil column should be free draining. A perforated plate made
from metal or acrylic with a piece of filter paper or high permeability geo-
textile at the bottom of the soil column will usually suffice to prevent loss of
soil particles and the avoidance of creating a capillary break.
At the base of the soil column, provision for measuring water outflow
with time (Methods A1 and A3) and/or imposing a constant head condition
(Methods A2 and A4) are required. For measuring water outflow (methods
A1 and A3), a weighing balance, a graduated cylinder or other means can
be used to record the volume of water outflow from the base with time. For
Method A3, a Mariotte bottle may be used to maintain a constant head
condition and to measure the volume of water outflow at the same time.
212 Laboratory tests for unsaturated soils

H
w
H
0 kPa (atmospheric)

+ wH

Connection end to system to

maintain constant pressure

Figure 12.3 Principle of Mariotte bottle to maintain a constant pressure at inlet.

The principle of the Mariotte bottle is shown in Figure 12.3. For Methods
A2 and A4, a manometer system or a Mariotte bottle can be used to main-
tain a constant head condition at the base of the soil column. The Mariotte
bottle provides a convenient means of determining the volume of water
inflow into the soil column, but if the manometer is used, an inflow volume
measurement system such as the one mentioned in Chapter 11 needs to be
added.
Infiltration at the top of the soil column can be applied by a peristaltic
pump system where the amount of water flowing onto the top surface of the
soil column can be measured using a weighing balance or graduated cylin-
der (Method A1). However, the flow rate into the soil column is usually very
small (less than the saturated permeability of the soil). Provision needs to be
made to distribute the infiltration uniformly across the top surface of the
soil column while ensuring that the evaporation rate is low. A piece of filter
paper can be placed on the top surface of the soil column on to which the
inflow line drips (Yang et al. 2004) or a cup to receive the water from the
inflow line and distribute it through a series of cotton fibre wicks draped
across the soil surface (ASTM D7664-­10). To impose an evaporative flux
boundary on the top surface of the soil column, an infrared lamp to provide
a constant temperature to the soil surface and an electric fan to circulate the
air to the soil surface can be used (ASTM D7664-­10). It is also possible to
provide a constant relative humidity by circulating vapour from a salt solu-
tion through the top surface of the soil column (Figure 12.4). Verification of
the temperature and relative humidity at the top of the soil column can be
 Permeability 213

Peristaltic pump

Parafilm

Constant
relative
humidity

Soil column

Salt solution

Figure 12.4 M
 aintaining constant relative humidity at the top surface of the soil
column using salt solution.

performed by monitoring the temperature and relative humidity using a

capacitive psychrometer with temperature sensor.

12.3.2 Test procedures
The test procedures can be divided into four stages: sample preparation, soil
column preparation, column test and finishing.

12.3.2.1 Sample preparation
Conduct basic soil properties on the soil sample to obtain grain size distri-
bution, maximum and minimum dry densities and specific gravity. Air dry
the sample and prepare about twice the mass of the sample required to fill
the soil column.

12.3.2.2 Soil column preparation

First check that the instruments, suction and moisture content sensors, are
calibrated and working. Usually, small-­tip tensiometers are used to mea-
sure matric suction. These tensiometers should be saturated and tested as
explained in Chapter 8.
214 Laboratory tests for unsaturated soils

Initially, a dry soil column is prepared by dry pluviation. Select a target

dry density to conduct the test. Ideally, the soil column should be prepared
in lifts of equal height until the whole soil column is prepared. A convenient
lift height is the vertical spacing of the sensors. However, if the vertical spac-
ing exceeds 15 cm, a convenient lift height should be selected such that the
maximum lift height does not exceed 15 cm. Weigh out the sand into bags
or containers to meet the target dry density for each lift height. Apply suit-
able dry pluviation method to achieve the target dry density for each lift.
The soil in each lift can be compacted by static compaction, a piston com-
pactor or a drop weight. Ensure that the soil column is secured if piston or
drop weight is used to compact the soil layer.
At each sensor placement level, carefully place and position the sensors.
Tighten the plug to seal. Place some soil from the next lift height to cover/
protect the sensors and continue filling to the next sensor placement level.
Repeat until the entire soil column is prepared. Check the senor readings to
ensure that they reached the equilibrium condition.
For Methods A3 and A4, saturate the soil column by allowing water to
flow into the soil column from the bottom. This can be done using the
manometer to impose a small water pressure. Check that the wetting front
is moving up slowly through the soil column and the soil is not disturbed.
Alternatively, the soil column for Method A3 and A4 can be prepared by
wet pluviation. Regardless of whether a dry or wet pluviation method was
used, it is good practice to allow the soil column to drain and to re-­saturate
again to ensure that there are no trapped or occluded air bubbles. This can
be repeated several times. The other advantage of Methods A3 and A4 over
Methods A1 and A2 is that the working condition of the matric suction and
moisture sensors can be checked during this process.

12.3.2.3 Column test
12.3.2.3.1 Infiltration
In Method A1, the initial condition of the soil column is dry. A constant
infiltration rate is applied to the top surface of the soil column and a free
drainage boundary is applied to the bottom of the soil column. Permeability
is determined at various matric suctions by imposing different infiltration
rates. The number of infiltration rates chosen should be at least three, and
all shall be less than the saturated permeability of the soil. It is convenient to
choose the infiltration rate based on a factor of the saturated permeability
of the soil. The saturated permeability of the soil can first be determined
using test methods for saturated soils. The infiltration rate should start from
the smallest infiltration rates. Convenient rates are 1,000, 100 and 10 times
lower than the saturated permeability of the soil.
Prepare the top surface of the soil column as described in test set-­up to
ensure a uniform distribution of the water. Start the pump and the data
 Permeability 215

recorder. Record inflow volume, matric suctions at the different depths of

the soil column and water content at the different depths of the soil column.
Continue the test until there are no further changes in the matric suctions
and water contents (steady-­state condition).

12.3.2.3.2 Imbibition
In Method A2, the initial condition of the soil column is dry. A zero-­water
pressure condition is applied to the bottom of the soil column while the top
of the soil column is open to the atmosphere. Shut the drainage valve off at
the bottom of the soil column and connect the manometer with inflow vol-
ume measurement or Mariotte bottle to the bottom of the soil column. Set
the water pressure to zero (atmospheric pressure) at the bottom of the soil
column. Start the data recorder and open the drainage valve. Record inflow
volume, matric suctions at the different depths of the soil column and water
content at the different depths of the soil column. Continue the test until
inflow volume becomes zero, i.e., water stops entering the bottom of the soil
column due to capillary rise.

12.3.2.3.3 Drainage
In Method A3, the initial condition of the soil column is saturated and under
hydrostatic water pressure condition with the top of the soil column at zero
pore-­water pressure. Shut the drainage valve off at the bottom of the soil
column and connect the manometer with outflow volume measurement or
Mariotte bottle to the bottom of the soil column. Set the water pressure
to zero at the bottom of the soil column. Start the data recorder and open
the drainage valve. Record outflow volume, matric suctions at the different
depths of the soil column and water content at the different depths of the
soil column. Continue the test until outflow volume becomes zero, i.e., the
change in 30 minutes is less than 1%.

12.3.2.3.4 Evaporation
In Method A4, the initial condition of the soil column is saturated and under
hydrostatic water pressure conditions with the top of the soil column at
zero pore-­water pressure. Turn the drainage valve off at the bottom of the
soil column and connect the manometer or Mariotte bottle to the bottom
of the soil column. Set the water pressure to zero at the top of the soil col-
umn. Open the drainage valve to ensure that no water is flowing out from
the soil column and the pore-­water pressure condition in the soil column is
hydrostatic. Turn the drainage valve off. Apply a constant relative humidity
boundary condition at the top of the soil column and start the data recorder.
Record matric suctions at the different depths of the soil column and water
content at the different depths of the soil column. Continue the test until
216 Laboratory tests for unsaturated soils

there is no change in matric suction and water content with time. The test-­
stopping criterion can be taken as the changes being less than 1% of the
measured values within 24 hours (ASTM D7664-­10).

12.3.2.4 Finishing
The water content distribution of the soil column at the end of the test can
be determined by removing soil samples from different depths of the soil
column and drying in the oven for water content determination. This can
be used to check against the water contents measured by the water content
sensors.

12.4 MULTISTEP OUTFLOW METHOD

Multistep outflow methods can be performed using a pressure plate appa-

ratus. The change in matric suction is performed by applying the axis-­
translation technique. The pressure plate apparatus can be treated as a
rigid-­wall permeameter (Method B1) such as using a Tempe cell or volumet-
ric pressure plate apparatus. Tempe cell and volumetric pressure plate are
mentioned in Chapter 10. However, if a standard pressure plate apparatus
is used, the water pressure that can be applied below the high air entry
(HAE) ceramic disk is small. As for a flexible-­wall permeameter (Method
B2), the standard triaxial apparatus can be modified to become a flexible-­
wall permeameter (Goh et al. 2015) or the triaxial permeameter described in
Chapter 11 can be used. In both test methods, the water outflow with time
needs to be measured.

12.4.1 Test set-up
The addition of water outflow measurement is required to make the stan-
dard pressure plate apparatus suitable for use in Test Method B. The water
outflow volume measurement can be done using a burette or a digital pres-
sure/volume controller with a reading resolution of 1 mL or better. The
connection tubing from the pressure plate apparatus to the water volume
measurement device should be incompressible, as it affects the water volume
measurement.
The HAE ceramic disk and the water reservoir below the ceramic disk
should be fully saturated before the test. The saturation of the pressure plate
apparatus is similar to that in Chapter 10.
As Test Method B has a long test duration, flushing of diffused air from
the water reservoir below the HAE ceramic disk is needed, as it may inter-
fere with the water outflow volume measurement. The flushing of diffused
air can be affected using a diffused air volume indicator (Samingan et al.
2003; Fredlund et al. 2012).
 Permeability 217

12.4.2 Test procedures
The test procedures can be divided into four stages: sample preparation,
preparation of apparatus, permeability test and finishing.

12.4.2.1 Sample preparation
Conduct basic soil properties on the sample to obtain grain size distribu-
tion, maximum and minimum dry densities, and specific gravity. Both intact/
undisturbed and reconstituted soil specimens may be used for Test Method
B. The preparation of the soil specimen is given in Chapters 3 and 10.

12.4.2.2 Preparation of apparatus
The permeability of the HAE ceramic disk should be tested periodically
to ensure that it is not clogged or defective. For the rigid-­wall permeam-
eter, the measurement of permeability of the saturated HAE ceramic disk
can be conducted by first saturating the HAE ceramic disk in a vacuum
desiccator under distilled water at least overnight and then assembling it
into the pressure plate apparatus. Half-­fill the rigid-­wall permeameter and
apply high air pressure and measure the volume of water exiting from the
permeameter with time. The saturated permeability of the HAE disk can be
determined using the constant head permeability equation given in Chapter
11. The pressure gradient is given by the air pressure across the thickness
of the ceramic plate. For the flexible-­wall permeameter, the measurement
of the saturated permeability of the HAE ceramic disks is given in Chapter
11. ASTM D7664-­10 recommends that if there is a decrease in saturated
permeability of the HAE ceramic disks by five times from the saturated per-
meability when new, the HAE ceramic disk should be discarded or cleansed.

12.4.2.3 Permeability test
In ASTM D7664-­10, the HAE ceramic disk is dried in the oven at 110°C
and placed dry in the set-­up. The soil specimen is then placed on top of
the HAE ceramic disk, the pressure chamber is closed and then the whole
system of soil specimen and HAE ceramic disk is saturated. This is a very
difficult and slow process in practice as the permeability of the HAE ceramic
disk is very low and the air in the system will have to be removed through
the small drainage tubing. It is very difficult to ensure that all the air in the
system is fully removed by this process.
Instead, the permeameter system can first be ensured to be fully saturated,
i.e., the HAE ceramic disk, water reservoir below the HAE ceramic disk and
the water drainage lines are fully saturated. For the rigid-­wall permeameter,
the soil specimen in the ring can be placed on top of the HAE ceramic disk
and the bottom drainage line connected to a water burette filled with water
218 Laboratory tests for unsaturated soils

that is allowed to flow into the soil specimen by upward seepage and capil-
lary action. When air bubbles ceased to appear from the surface of the soil
specimen, the soil specimen can be further saturated by closing the rigid-­
wall permeameter and applying a small vacuum (~−20 kPa) at the air pres-
sure inlet. For the flexible-­ wall permeameter, the saturation of the soil
specimen can follow the procedures as mentioned in Chapter 11. The degree
of saturation in the flexible-­wall permeameter can be checked using Skempton
pore-­water pressure parameter B. A B parameter greater than 0.9 is accepted
as full soil saturation in ASTM D7664-­10 and 0.95 in BS 1377-6 (1990).
In ASTM D7664-­10, the test starts with the measurement of the saturated
permeability of the soil specimen. For the rigid-­wall permeameter, this can
only be done if there is water above the soil sample. The saturated permea-
bility test is done as a constant head permeability test and the equations in
Chapter 11 are applicable. For the permeability of the unsaturated soil, the
test starts at equal water and air pressures. The maximum air pressure
applied should not exceed the HAE of the ceramic disk. ASTM D7644-­10
suggests 300 kPa at the start of the test. The air pressure is then maintained
constant while the water pressure is reduced in steps to apply the matric
suctions.

12.4.2.4 Rigid-wall permeameter
The test is started with the soil specimen in the fully saturated condition.
The matric suctions for the whole test should be planned. The maximum
matric suction is governed by the air-­entry value (AEV) of the ceramic disk
and the final water pressure below the ceramic disk. For example, if the
AEV of the ceramic disk is 500 kPa and the final water pressure below the
ceramic disk is 50 kPa, and the target test matric suctions are 100, 200, 300
and 400 kPa, then the initial air and water pressures can be set at 450 kPa.
To achieve the matric suctions of 100, 200, 300 and 400 kPa, the air pres-
sure is kept constant at 450 kPa while the water pressure is reduced to 350,
250, 150 and 50 kPa, respectively, in the test.
The test is started with both air and water pressures at their maximum
value (i.e., 450 kPa). The matric suction is applied by keeping air pressure
constant and decreasing the water pressure. The water starts to flow out
from the soil specimen when the matric suction is applied. At each matric
suction, measure the volume of the water outflow from the soil specimen
with time. The measurements should be done at regular intervals to obtain
the nonlinear shape of the water outflow volume with time. The test at each
matric suction can stop when the water outflow becomes negligibly small. In
ASTM D7644-­10, negligibly small water outflow is defined when the water
outflow volume in one hour is less than 1% of the water outflow volume in
the previous hour. The next matric suction is then applied and the measure-
ment of water outflow with time is then repeated. This process is carried out
until the last matric suction.
 Permeability 219

At the highest applied matric suction, the matric suction can be decreased
in the reverse manner and volume of water inflow can be measured with
time to obtain the unsaturated permeability for the wetting path.
For the rigid-­wall permeameter, it is possible to measure the change in
height of the specimen and/or to apply a net vertical stress during the test by
adding a piston and a load cap on top of the specimen. Examples of tests
conducted in such a manner are given in Wayllace and Lu (2012), and
Fredlund and Houston (2013).

12.4.2.5 Flexible-wall permeameter
The test procedures using the flexible-­wall permeameter are largely similar
to those for the rigid-­wall permeameter except that there is an additional
stress state variable to control which is the net confining stress. The test is
usually conducted at a constant net confining stress. The test is started with
the soil specimen in the fully saturated condition. The matric suctions for
the whole test should be planned as explained for the rigid-­wall permeam-
eter. The maximum matric suction is governed by the AEV of the ceramic
disk and the final water pressure below the ceramic disk. For the flexible-­
wall permeameter, the net confining stress, i.e., the difference between cell
pressure and air pressure must be maintained throughout the test.
The test is started by applying equal air and water pressures that are both
at the maximum values (e.g., 450 kPa). If the air pressure is adjusted, the cell
pressure needs to be adjusted accordingly to maintain constant net confining
pressure. At no time during the test should the confining pressure be smaller
than the air pressure, as the soil specimen may be damaged when this happens.
The matric suction is applied by keeping air pressure constant and decreasing
the water pressure. Keeping the air pressure constant is sensible in the flexible-­
wall permeameter, as it means that the cell pressure does not need to be
adjusted throughout the test to maintain constant net confining pressure.
The water starts to flow out from the soil specimen when the matric suc-
tion is applied. At each matric suction, measure the volume of the water
outflow from the soil specimen with time. The measurements should be
done at regular intervals to obtain the nonlinear shape of the water outflow
volume with time. The test at each matric suction can stop when the water
outflow becomes negligibly small. In ASTM D7644-­10, negligibly small
water outflow is defined when the water outflow volume in 1 hour is less
than 1% of the water outflow volume in the previous hour. The next matric
suction is then applied and the measurement of water outflow with time is
then repeated. This process is carried out until the last matric suction.
At the highest applied matric suction, the matric suction can be decreased
in the reverse manner and volume of water inflow can be measured with
time to obtain the unsaturated permeability for the wetting path. In the wet-
ting test, it is possible to reduce the matric suction to zero to measure the
saturated permeability.
220 Laboratory tests for unsaturated soils

12.4.2.6 Finishing
At the end of the permeability test, the air and water pressures are reduced
to zero in quick steps and then the water pressure line is locked to prevent
the soil specimen from swelling by imbibing the water. For the flexible-­wall
permeameter, the cell pressure will also have to be reduced in tandem with
the air pressure. The final net confining pressure at zero air pressure is only
reduced to zero after the water pressure line is locked.
The soil specimen can then be removed. The dimensions, weight and final
water content of the specimen are then determined.

12.5 CALCULATIONS AND DATA INTERPRETATION

12.5.1 Instantaneous profile method

The instantaneous profile method applies mass continuity to transient pro-
files of suction and water content to determine the permeability of unsatu-
rated soils. It does not matter which initial soil condition and boundary
conditions were applied in the soil column test to obtain the transient pro-
files, the calculation to obtain the unsaturated permeability is the same. The
profiles of suction and water content with time for the four different soil
column tests are shown schematically in Figure 12.5.
The instantaneous profile method can also be applied when only either
the matric suction profile or the volumetric water content is measured.
However, the soil-­water characteristic curve must be obtained separately.
The missing matric suction profile or volumetric water content profile is
then obtained using the soil-­water characteristic curve.
The instantaneous profile method uses Darcy’s law as given in
Equation 12.1.

∂H
v = −k , (12.1)
∂z

where
v = flow velocity
k = permeability
z = vertical coordinate (increasing in gravity direction)
H = total head for vertical flow (= Hp + z)
Hp = pressure head

The hydraulic gradient can be expressed using Equation 12.2.

∂H ∂H p 1 ∂uw
− = +1 = +1 (12.2)
∂z ∂z γ w ∂z
 Permeability 221

Method Suction profiles Water content profiles

A1 Pore-water pressure Volumetric water content
(dry, infiltration) - 0 + 0 +

t1 t1

t2 t2

t3 t3

t4
t4

t0 t0
Depth Depth

A2
Pore-water pressure Volumetric water content
(dry, imbibition)
- 0 + 0 +

t0 t0

t4
t4

t3 t3

t2
t2

t1 t1

Depth Depth

Figure 12.5 S chematic representation of suction and water content profiles in

soil columns.
(Continued)
222 Laboratory tests for unsaturated soils

A3
Pore-water pressure Volumetric water content
(saturated,
- 0 + 0 +
drainage)

t4
t4

t3

t3
t2
t2
t1 t0 t1 t0

Depth Depth

A4
Pore-water pressure Volumetric water content
(saturated,
- 0 + 0 +
evaporation)

t4 t4
t3 t3
t2 t2
t1
t1

t0
t0

Depth Depth

Figure 12.5 (Continued).

 Permeability 223

At any time t, Equation 12.2 can be expressed in finite difference form for
two sensor depths zm and zm+1 giving Equation 12.3.

 ∂H  1  ( uw )m − ( uw )m+1 
−  =   +1 (12.3)
 ∂z m γ w  zm+1 − zm 

The volume of water Vm passing depth m between time t and Δt can be

obtained by integrating the volumetric water content θ downstream of depth
m for the time interval Δt (Equation 12.4).

 n 
Vm = A  ∑(
 i = m
θit − θit + ∆t )(z i +1 − zi ) 

(12.4)

where
i = integration number representing sensor number
n = maximum number of sensors
A = cross-­sectional area of soil column

Hence, the permeability (kw)m at depth m can be obtained using Equation

12.5.

 
Vm  1 
( kw )m = (12.5)

A∆t  − ∂H

( ∂z ) 

m 

The value of (kw)m is associated with the average suction of [(uw)m +

(uw)m+1]/2 and the volumetric water content of θm.
The above calculations are repeated at each of the measurement points
and for each time to define the permeability function. For infiltration
(Method A1), it may be possible to reach a steady-­state (unit hydraulic gra-
dient) condition as indicated by the shaded zone in the suction profile and
volumetric water content profile. If this is observed, the permeability at the
suction and volumetric water content is given by the infiltration rate
(Fredlund et al. 2012).

12.5.2 Multistep outflow method

The original multistep outflow method was developed by Gardner (1956).
Gardner (1956) method is adopted in ASTM D7664-­10 for data interpreta-
tion. The test results from the rigid-­wall and flexible-­wall permeameters are
interpreted in the same way. To analyse the test results, it is assumed that
the specimen is homogeneous and non-­deformable, flow is one-­dimensional,
gravitational flow is negligible, unsaturated permeability kw is constant and
suction is linearly related to the volumetric water content for each suction
224 Laboratory tests for unsaturated soils

increment, and impedance of the ceramic disk is negligible. The equation

used to analyse test results from the multistep outflow method is Fick’s sec-
ond law of diffusion shown as Equation 12.6. There are different approaches
to arriving at Equation 12.6 (Benson and Gribb 1997).

∂ ( ua − uw ) ∂ 2 ( ua − uw )
=D (12.6)
∂t ∂z 2

where D is diffusivity given by Equation 12.7.

∂ ( ua − uw )
D = kw (12.7)
∂θ

The solution to Equation 12.6 is given by Gardner (1956) and is repro-

duced in Equation 12.8.

V −V   8   Dπ 2 
ln  ∞  = ln  π 2  −  4h2  t (12.8)
 V∞     s 

where
V = outflow volume at time t
V∞ = final outflow volume for suction increment
hs = height of soil specimen

The diffusivity D for each suction increment can be obtained by plotting

V −V 
ln  ∞  versus t. A best fit line is drawn through the plot, and the slope
 V∞ 
 Dπ 2 
of the line is given by  2 
. Once D is obtained, kw can be obtained using
 4hs 

For each suction increment (ua–uw)

—

V t
V

t t

Figure 12.6 Illustration of obtaining D from multistep outflow test for one suc-
tion increment.
 Permeability 225

Equation 12.7. The steps are illustrated in Figure 12.6. Using Figure 12.6, kw
is given by Equation 12.9.

∆V
Vt
kw = D (12.9)
∆ ( ua − uw )

where Vt is the total volume of the soil specimen. Repeating the calculation
for each of the suction increments, the permeability function is obtained.

12.6 SUMMARY

The accuracy of the transient-­state methods in obtaining the permeabil-

ity function for unsaturated soils is not as good as the steady-­state meth-
ods. For the instantaneous profile method, the range of suction over which
permeability is determined is lower than in the multistep outflow method.
Preparing a homogeneous soil column may not be easy, and the instruments
(suction and water content sensors) may interfere with the flow process. For
the large-­diameter soil column, preferential flow and fingering processes may
be observed and flow may not be one-­dimensional. Vapour flow processes
may also occur, and there is inaccuracy in measuring suction and water
content as the changes may be small. There is also no control over the stress
state. For the multistep outflow method, there is better control of the quality
of the soil specimen, as it is small. The original multistep outflow method
was designed to be used with the test to obtain the soil-­water characteristic
curve, thus it is an expedient process where both the soil-­water characteris-
tic curve and permeability function can be obtained at the same time. The
accuracy of determining the permeability is obtained with smaller suction
increments. However, this will increase the test duration and increases the
likelihood of diffused air appearing in the water reservoir beneath the HAE
ceramic disk and interferes with the water outflow/inflow volume measure-
ment. In Gardner’s (1956) method of calculating permeability, plate imped-
ance has been neglected. Many have tried to improve the outflow method
(Miller and Elrick 1958; Rijtema 1959; Kunze and Kirkham 1962; Kool et
al. 1985; van Dam et al. 1994; Crescimanno and Iovino 1995; Hwang and
Powers 2001; Figueras et al. 2006; Wayllace and Lu 2012; Shao et al. 2017).

REFERENCES

Benson, C. H. & Gribb, M. M. (1997). Measuring unsaturated hydraulic con-

ductivity in the laboratory and field. In Houston, S. & Fredlund, D. G. (Eds.),
Unsaturated soil engineering practice, Geotechnical Special Publication No. 68.
New York: ASCE, pp 113–168.
226 Laboratory tests for unsaturated soils

Caicedo, B. (2017). Physical modelling of freezing and thawing of unsaturated soils.

Géotechnique, 67(2): 106–126. http://dx.doi.org/10.1680/jgeot.15.P.098.
Crescimanno, G. & Iovino, M. (1995). Parameter estimation by inverse method based
on one-­step and multi-­step outflow experiments. Geoderma, 68(4): 257–277.
Figueras, J., Gribb, M. M., & McNamara, J. P. (2006). Design of a new auto-
mated multi-­step outflow test apparatus, presented at the AGU fall meeting. San
Francisco, CA: American Geophysical Union, www.agu.org
Fredlund, D. G. & Houston, S. L. (2013). Interpretation of soil-­water characteristic
curves when volume change occurs as soil suction is changed. In Proceedings of
the First Pan-­Am Conference on Unsaturated Soils, UNSAT 2013, Advances in
Unsaturated Soils, Editor: Bernardo Caicedo, February 20–22, pp. 15–31.
Fredlund, D. G., Rahardjo, H., & Fredlund, M. D. (2012). Unsaturated soil mechan-
ics in engineering practice. New York: John Wiley & Sons, Inc., 926 pages (ISBN
978-­1-­118-­13359-­0).
Fujimaki, H. & Inoue, M. (2003). A transient evaporation method for determin-
ing soil hydraulic properties at low pressure. Vadose Zone Journal, 2: 400–408.
https://doi.org/10.2136/vzj2003.4000a
Garnier, P., Rieu, M., Boivin, P., Vauclin, M., & Baveye, P. (1997). Determining the
hydraulic properties of a swelling soil from a transient evaporation experiment.
Soil Science Society of America Journal, 61: 1555–1563. https://doi.org/10.2136/
sssaj1997.03615995006100060003x
Goh, S. G., Rahardjo, H., & Leong, E. C. (2015). Modification of triaxial apparatus
for permeability measurement of unsaturated soils. Soils and Foundations, 55(1):
63–73. http://dx.doi.org/10.1016/j.sandf.2014.12.005.
Hopmans, J. W., Šimůnek, J., Romano, N., & Durner, W. (2002). Simultaneous deter-
mination of water transmission and retention properties: Inverse methods. In J.
H. Dane & G. C. Topp (Eds.), Methods of soil analysis. Part 4. SSSA Book Ser. 5.
Madison, WI: SSSA, pp. 963–1008.
Hwang, S. & Powers, S. E. (2001). Using a multi-­step outflow technique to estimate
unsaturated hydraulic conductivity function for quartz sands with variable inter-
facial properties, AGU fall meeting, San Francisco, CA: American Geophysical
Union, www.agu.org
Kool, J. B., Parker, J. C., & Van Genuchten, M. T. (1985). Determining soil hydraulic
properties from one-­step outflow experiments by parameter estimation: I. Theory
and numerical studies. Soil Science Society Of America Journal, 49(6): 1348–1354.
Kool, J. B., Parker, J. C., & Van Genuchten, M. Th. (1987). Parameter estimation
for unsaturated flow and transport models – A review. Journal of Hydrology, 91:
255–293.
Krisdani, H., Rahardjo, H., & Leong, E. C. (2009). Use of instantaneous profile
and statistical methods to determine permeability functions of unsaturated soils.
Canadian Geotechnical Journal, 46(7): 869–874.
Kunze, R. J. & Kirkham, D. (1962). Simplified accounting for membrane impedance
in capillary conductivity determinations. Soil Science Society of America Journal,
26(5): 421–426. https://doi.org/10.2136/sssaj1962.03615995002600050006x
Leij, F. J. & van Genuchten, M. Th. (1999). Characterization and measurement of
the hydraulic properties of unsaturated porous media. In M. Th. van Genuchten et
al. (Eds.), Characterization and measurement of the hydraulic properties of unsat-
urated porous media: Proceedings of the International Workshop. 22–24 October
1997. Riverside, CA: U.S. Salinity Lab., pp. 1–12.
 Permeability 227

McCartney, J. S. & Zornberg, J. G. (2010). Centrifuge permeameter for unsaturated

soils II: Results and analysis. Journal of Geotechnical and Geoenvironmental
Engineering, 136(8): 1064–1076.
Miller, E. E. & Elrick, D. E. (1958). Dynamic determination of capillary conductivity
extended for non-­negligible membrane impedance. Soil Science Society of America
Journal, 22(6): 483–486. https://doi.org/10.2136/sssaj1958.0361599500220006
0002x
Minasny, B. & Field, D. F. J. (2005). Estimating soil hydraulic properties and
their uncertainty: The use of stochastic simulation in the inverse modelling of
the evaporation method. Geoderma, 126: 277–290. https://doi.org/10.1016/j.
geoderma.2004.09.015
Parks, J., Stewart, M., & McCartney, J. S. (2012). Validation of a centrifuge perme-
ameter for investigation of transient infiltration and drainage flow processes in
unsaturated soils. Geotechnical Testing Journal, 35(1): 182–192.
Rijtema, P. E. (1959). Calculation of capillary conductivity from pressure plate
outflow data with nonnegligible membrane impedance. Journal of Agriculture
Science, 4: 209–215.
Samingan, A. S., Leong, E. C., & Rahardjo, H. (2003). A flexible wall permeameter
for measurements of water and air coefficients of permeability of residual soils.
Canadian Geotechnical Journal, June, 40(3): 559–574.
Schelle, H., Iden, S. C., & Durner, W. (2011). Combined transient method for deter-
mining soil hydraulic properties in a wide pressure head range. Soil Science Society
of America Journal, 75: 1681–1693. https://doi.org/10.2136/sssaj2010.0374
Schindler, U., Durner, W., von Unold, G., & Müller, L. (2010). Evaporation method
for measuring unsaturated hydraulic properties of soils: Extending the measure-
ment range. Soil Science Society of America Journal, 74: 1071–1083. https://doi.
org/10.2136/sssaj2008.0358
Shao, L., Wen, T., Guo, X., & Sun, X. (2017). A method for directly measuring the
hydraulic conductivity of unsaturated soil. Geotechnical Testing Journal, 40(6):
907–916. https://doi.org/10.1520/GTJ20160197. ISSN 0149-­6115
Tian, Z., Kool, D., Ren, T., Horton, R., & Heitman, J. L. (2019). Approaches for esti-
mating unsaturated soil hydraulic conductivities at various bulk densities with the
extended Mualem-­van Genuchten model. Journal of Hydrology, 572: 719–731.
https://doi.org/10.1016/j.jhydrol.2019.03.027
van Dam, J. C., Stricker, J. N. M., & Droogers, P. (1994). Inverse method to deter-
mine soil hydraulic functions from multistep outflow experiments. Soil Science
Society of America Journal, 58: 647–652. https://doi.org/10.2136/sssaj1994.036
15995005800030002x
Wayllace, A. & Lu, N. (2012). A transient water release and imbibitions method for
rapidly measuring wetting and drying soil water retention and hydraulic conduc-
tivity functions. Geotechnical Testing Journal, 35(1): 1–15.
Wendroth, O., Ehlers, W., Hopmans, J. W., Kage, H., Halbertsma, J., & Wösten, J.
H. M. (1993). Reevaluation of the evaporation method for determining hydraulic
functions in unsaturated soils. Soil Science Society of America Journal, 57: 1436–
1443. https://doi.org/10.2136/sssaj1993.03615995005700060007x
Wildenschild, D., Hopmans, J. W., & Šimůnek, J. (2001). Flow rate dependence of
soil hydraulic characteristics. Soil Science Society of America Journal, 65: 35–48.
https://doi.org/10.2136/sssaj2001.65135x
228 Laboratory tests for unsaturated soils

Wind, G. P. (1968). Capillary conductivity data estimated by a simple method.

pp. 181–191. In Water in the unsaturated zone: Proc. Symp., Wageningen, the
Netherlands. June 1966. Vol. 1. Gentbrugge, Belgium: International Association
of Scientific Hydrology.
Yang, H., Rahardjo, H., Wibawa, B., & Leong, E. C. (2004). A soil column apparatus
for laboratory infiltration study. Geotechnical Testing Journal, 27(4): 1–8.
Zornberg, J. G. & McCartney, J. S. (2010). Centrifuge permeameter for unsaturated
Soils I: Theoretical basis and experimental developments. Journal of Geotechnical
and Geoenvironmental Engineering, 136(8): 1051–1063.

FURTHER READING

Hopmans, J. W., Šimůnek, J., Romano, N., & Durner, W. (2002). 3.6.2. Inverse meth-
ods. In J. H. Dane & G. Clarke Topp (Eds.), Methods of soil analysis. https://doi.
org/10.2136/sssabookser5.4.c40
Chapter 13

Oedometer test

13.1 BACKGROUND

The oedometer test (ASTM D2435/D2435M 2011; BS EN ISO 17892-­5

2017) is an incremental loading compression test that is commonly carried
out to determine the one-­dimensional consolidation properties of saturated
soils. The test can also determine the secondary compression (creep) of satu-
rated soils by extending the test duration under a constant load and also the
swelling potential of expansive soils or collapse potential of collapsible soil
(ASTM D3877 2008; ASTM D4546 2008; and ASTM D5333 2003). While
the oedometer test is commonly used for saturated soil, it can also be used
for unsaturated soil. Three tests for unsaturated soils that can be performed
using the oedometer apparatus are

1. unsaturated one-­dimensional compression test,

2. wetting test, and
3. swelling/collapse test.

The test procedures for unsaturated soils using the oedometer apparatus
have not been standardised for the first two tests. However, the test pro-
cedures for swelling and collapse of soils have been standardised in ASTM
D3877 (2008), ASTM D4546 (2008) and ASTM D5333 (2003). In this
chapter, the test procedures for the first two tests will be elaborated while the
test procedures for the swelling and collapse of soils will only be described
for unsaturated soils.
Unlike the saturated oedometer test, which is only dependent on effective
stress, the unsaturated oedometer test is dependent on net normal stress and
matric suction/water content. As a result, it requires more effort to deter-
mine the parameters to define the unsaturated compression curve. Thus, it is
important to understand the field condition such as

1. the expected net normal stress that applies at the site, and
2. the initial water content/matric suction and its variation at the site.

DOI: 10.1201/b22304-13 229

230 Laboratory tests for unsaturated soils

13.2 RELATED STANDARDS

1. ASTM D3877. 2008. Standard Test Methods for One-­Dimensional

Expansion, Shrinkage, and Uplift Pressure of Soil-­Lime Mixtures.
2. ASTM D4546. 2008. Standard Test Methods for One-­Dimensional
Swell or Collapse of Cohesive Soils.
3. ASTM D4829. 2011. Standard Test Method for Expansion Index for
Soils.
4. ASTM D5333. 2003. Standard Test Method for Measurement of
Collapse Potential of Soils.
5. ASTM D4186. 2006. Standard Test Method for One-­Dimensional
Consolidation Properties of Saturated Cohesive Soils Using Controlled-­
Strain Loading.
6. ASTM D2435/D2435M. 2011. Standard Test Methods for One-­
Dimensional Consolidation Properties of Soils Using Incremental
Loading.
7. BS EN ISO 17892-­5. 2017. Geotechnical Investigation and Testing.
Laboratory Testing of Soil. Part 5: Incremental Loading Oedometer
Test.
8. BS 1377-­ 5 (1990) Methods of Test for Soil for Civil Engineering
Purposes – Part 5: Compressibility, Permeability and Durability Tests,
BSI.

13.3 THEORY

13.3.1 Settlement and heave

One major consideration in geotechnical engineering design is the esti-
mation of settlement (downward movement of soil relative to the ini-
tial elevation) and heave (upward movement of soil relative to the initial
elevation). There are five sources of settlement or heave that need to be
considered:

1. Soil distortion
2. Primary compression (consolidation)
3. Secondary compression (creep)
4. Wetting-­induced volume change (swelling and collapse)
5. Drying-­induced volume change (shrinkage)

Soil distortion is a condition where the soil particles move due to loading.
Figure 13.1 illustrates settlement-­inducing soil distortion. Consider an ini-
tially horizontal ground level indicated by a straight line subjected to a load.
Once the load has been applied, the soil is displaced sideways without caus-
ing volume change. The difference in elevation between the initial ground
 Oedometer test 231

Applied load
Distorted soil Distorted soil

Initial ground level

si

Figure 13.1 Settlement due to soil distortion.

level and the post-­loading ground level is referred to as settlement/heave due

to soil distortion (si).
Soil distortion occurs instantaneously and thus is considered immediate
settlement. The analysis can be done by using elastic theory, and modulus is
taken as undrained modulus (Eu). Modulus Eu can be obtained by using an
undrained triaxial test, which is explained in Chapter 15.
Primary compression is a reduction in the volume of voids in a soil. For
saturated soil, it is known as consolidation, as it involves the drainage
of pore water with time. When a soil is subjected to a stress increase Δσ,
its pore-­water pressure uw will increase (Δuw). Under one-­ dimensional
loading:

∆uw = B∆σ (13.1)

where B is the Skempton pore pressure parameter. For saturated soils, B = 1,

and B is less than 1 for an unsaturated soil.
When the soil is subjected to an additional load of Δσ, σ is calculated
using Equation 13.2.

σ = σ 0 + ∆σ (13.2)

where σ0 is the initial overburden pressure. Excess pore-­water pressure (ue)

which is equal to Δσ is generated and uw is equal to

uw = u0 + ue (13.3)

where u0 is the initial pore-­water pressure and is usually hydrostatic pres-

sure. As σ equals to σ0 + Δσ all the time, σ’ increases as ue gradually dis-
sipates over time as water flows out from the voids and causes reduction
of the void volume, which is commonly represented as reduction of void
232 Laboratory tests for unsaturated soils

ratio (Δe) and gravimetric water content (Δw). Reduction in gravimetric

water content due to the reduction in void ratio can be calculated using
Equation 13.4.

∆w ⋅ Gs = ∆e ⋅ Sr (13.4)

where Gs is specific gravity and Sr is degree of saturation, which remains at

100% for saturated soils. The process of reduction in soil volume due to the
dissipation of ue is commonly referred to as consolidation.
If the soil is fully saturated, the rate of consolidation is dependent on the
soil’s permeability. Coarse-­grained soils (i.e., sand) have very high permea-
bility, and thus consolidation takes place rapidly (in terms of minutes to
days). Thus, the compression settlement of coarse-­grained soils is considered
immediate settlement. For fine-­grained soils (i.e., silts and clay), consolida-
tion may take months to years, and thus is referred to as time-­dependent
settlement or consolidation settlement. If the soil is unsaturated (Sr < 100%),
compression of the voids filled with air will occur first and no water will
flow out from the soil pores, and thus water content remains constant.
Compression/decompression that does not cause any change in water con-
tent is referred to as the constant water content (CWC) condition. In the
CWC condition, the pore-­water pressure response depends on the initial
degree of saturation, which is described as follows (Wijaya and Leong 2016;
Wijaya 2017):

1. If the final degree of saturation remains less than 100%, no excess

pore-­water pressure will be generated, and no water will enter or
leave the soil element. However, the pore-­water pressure will increase
(although it remains negative) as degree of saturation increases due to
the reduction in void ratio under the CWC condition. Such behaviour
is known as mechanical wetting.
2. If the final degree of saturation is 100%, excess pore-­water pressure
will be generated. Once the degree of saturation reaches 100%, the
soil will behave like a saturated soil undergoing consolidation.

Unsaturated soil loaded under the CWC condition may experience wetting-­
induced volume change. However, it is difficult to separate pure compression
behaviour from the wetting-­induced volume change (Alonso et al. 1990),
especially when there is no measurement of pore-­water pressure during the
test.
Secondary compression or creep is a change in volume under constant σ’
without any change in uw (Holtz and Kovacs 1981; Holtz et al. 2011).
Secondary compression takes place over a very long time. It is unclear if
secondary compression occurs at the same time as primary consolidation or
only after primary consolidation ends. Secondary compression is usually
negligible for most soils and is only significant for peaty or organic soils. The
 Oedometer test 233

secondary compression index can be determined using an oedometer test by

extending the test duration for a constant load.
Some soils may either swell or collapse on wetting. Swelling is defined as
an increase in the soil volume due to wetting and may occur in both satu-
rated and unsaturated soils. Collapse is usually associated with unsaturated
soils on wetting. Soils that experience swelling will also experience shrinkage
on drying (drying-­induced volume change). The shrinkage curve described in
Chapter 5 can be used to estimate the amount of swelling/shrinkage with
respect to the change in water content. If soil movement is constrained for a
swelling soil (i.e., constant volume), swelling pressure will be generated.
Swelling and shrinkage potentials differ according to soil types, which can be
categorised as follows:

1. Highly expansive soils, e.g., soils that contain montmorillonite or illite

2. Moderately expansive soils: fine-­grained soils
3. Non-­expansive soils: coarse-­grained soils

Highly expansive soils require a thorough investigation of their swelling

and shrinkage properties. Swelling and shrinkage can be ignored for non-­
expansive coarse-­grained soils such as gravel and sand. Some soil deposits
that are aeolian, loessal, subaerial, mudflows, alluvial, residual or manmade
fills experience significant collapse when wetted. Collapse is defined as an
irrecoverable settlement due to rearrangement of particles that is caused by
wetting, additional loading or both. Collapse due to wetting is also known
as hydro-­consolidation, Whether a soil experiences swelling or collapse after
being wetted depends on the applied stress and stress history.

13.3.2 Compression, shrinkage and wetting-induced

volume change
Compression, shrinkage and wetting-­and drying-­induced volume change
of soil are stress path dependent (different stress paths may lead to differ-
ent results) as illustrated in Figure 13.2a. Figure 13.2b shows CWC curves
under different initial water contents, while Figure 13.1c shows the wetting
path along the CWC curve under different inundation pressures.

13.3.3 Saturated compression test

In Figure 13.2a, the initial soil specimen condition is located at point 1 where
the net normal stress is zero and the natural or in situ soil suction is equal
to sn. When the soil specimen is subjected to a saturated compression test,
the soil is first inundated with water. Inundation is commonly carried under
constant stress conditions, and the plane that represents constant stress is
referred to as the constant stress (CP) plane. Inundation causes the soil suc-
tion to go to zero, and the soil volume will increase (swelling). As a result,
point 1 will move to point 2s. When the soil becomes saturated, the pore-­air
234 Laboratory tests for unsaturated soils

CWC test

CWC test

Pi,1
Ps
PT P
c,s
sn Pi,1 Pi,2
Pi,2 PT
s0 Pc,d Pi,2

Pc(s)

Pc,d2 Pi,2

(a)

3c Collapse region
1 Swelling region
Shrinkage curve

2s
8s
Swelling

3d 1 3c pc,d,ec
pi 4c
e

4cs
8c Collapse

7s 7s
–u a (kPa) –u a (kPa) pT

(b) (c)

Figure 13.2 S chematic of unsaturated oedometer test in general. (a) 3D Schematic

of compression shrinkage and wetting induced behaviour of unsatu-
rated soil. (b) Constant water content compression curve under
different initial water content. (c) Wetting induced volume change
under different inundation pressure.

pressure (ua) becomes pore-­water pressure (uw) and net normal stress (σ−ua)
is equal to the effective stress (σ−uw). On further loading, the effective stress
will increase and the soil specimen will move along the reloading curve until
it reaches its saturated preconsolidation pressure (points 2s-­3s-­4s-­5s), which
is equal to Pc,s with void ratio (ec) equal to ec,1. Further loading will cause
the volume change to follow the normal compression line (NCL) which is
indicated by points 5s-­6s-­7s. Figure 13.3 shows the saturated compression
curve obtained from a kaolin specimen, KOdSat1 (Wijaya 2017) along with
its typical parameters, such as swelling index (Cs), recompression index (Cr),
compression index (Cc) and saturated preconsolidation pressure (Pc,s).
 Oedometer test 235

1.7

1.6
Inundation
Cr
1.5
Cr
Cs
1.4

1.3
e

1.2

1.1 Cc

0.9

0.8
Pc,s
1 10 100 1000
v' (kPa)

Figure 13.3 Saturated compression test of KOdSat1 from Wijaya (2017).

13.3.4 CWC compression test

If the soil is loaded without inundation (points 1-­2c-­5s), the loading is under
CWC condition, and the test is referred to as the CWC test. Under this load-
ing, the soil suction decreases due to mechanical wetting (the degree of satu-
ration increases due to a decrease in void ratio under the CWC condition),
as shown in Figure 13.4. If the soil suction becomes zero, further loading
will cause the compression curve to merge with the NCL (points 5s-­6s-­7s).
If the soil specimen is dried (point 1 to point 2d), the water content
decreases and soil suction increases. If the water content is low enough,
loading the soil specimen under the CWC condition (points 2d-­3c-­4c-­5c-­6c)
may cause the soil specimen to remain unsaturated beyond its apparent pre-
consolidation pressure or yield pressure (Pc,d) located at point 5c. Drying the
soil specimen will cause the preconsolidation void ratio to decrease (similar
to the soil specimen being subjected to a higher preconsolidation pressure
and then unloaded) and ec becomes ec,2 while the yield pressure increases
from Pc,s to Pc,d.
Drying the soil specimen will not cause further shrinkage if the soil speci-
men has reached the zero-­shrinkage line. The void ratio on the zero-­shrinkage
line is referred to as emin. At this point, soil suction has no effect on soil
structure. If points 2d and 3d are located on the zero-­shrinkage line, any
236 Laboratory tests for unsaturated soils

0.8
0.7
0.6
0.5
0.4
0.3
1 10 100 1000 10000
–uw (kPa) –uw (kPa)

100
0
0
–100
–50
–200
–300 –100
–400
–500 –150
1 10 100 1000 10000 1 10 100 1000 10000
–uw (kPa) –uw (kPa)
(a) (b)

Figure 13.4 C
 hange in pore-water pressure of soil specimen subjected to CWC
test. (a) Kaolin specimen Kw30s468 (Wijaya 2017). (b) Compacted
silty clay 7-10-H (Jotisankasa et al. 2007).

further drying will no longer reduce the prior compression void ratio (e0)
and will no longer reduce ec. If the soil specimen is loaded under CWC con-
dition (points 3d-­7c-­8c) the soil specimen will yield at Pc,d2, which is higher
than Pc,d despite Pc,d and Pc,d2 having the same preconsolidation void ratio
ec,2. The reason is due to the presence of water, which acts to reduce friction
between the soil particles. When there is less water at the interface between
particles, it requires a higher pressure to reach the same amount of particle
rearrangement.

13.3.5 Wetting-induced swelling/swelling pressure

If the soil is inundated at point 2c, the soil suction will reduce under constant
net normal stress conditions. As swelling does not affect the preconsolida-
tion void ratio (similar to unloading, which does not change the preconsoli-
dation pressure), the soil specimen will swell to the saturated compression
curve (point 3s) that is obtained if the soil specimen is inundated at point
1. However, if the soil specimen is inundated under constant volume condi-
tions (i.e., the soil specimen is prevented from swelling), swelling pressure
(Ps) will be generated that can be obtained by first projecting point 2c to the
saturation plane under the constant void ratio (point 3s’) and then drawing
 Oedometer test 237

a horizontal line that intersects the saturated compression curve (point 4s).
The effective stress required to reach point 4s is the swelling pressure. The
value of ec remains at ec,1 regardless of if the soil is inundated at point 1,
point 2c or not inundated, as inundation is equivalent to unloading, which
does not affect the most compact structure that the soil has experienced
before.
If the soil specimen is inundated at point 3c, the soil specimen will swell
to point 8s, which is located on the second saturated reloading line (points
8s-­9s-­6s) that lies below the first saturated reloading line (points 2s-­3s-­
4s-­5s). The second saturated reloading line will merge into the NCL at point
6s with ec equals to ec,2, which is equivalent to CWC loading (points
2d-­3c-­4c-­5c).
Point 4c is the point on the unsaturated compression curve (points 2d-­3c-­
4c-­5c-­6c) which has the same void ratio as the saturated compression curve
(points 8s-­9s-­6s-­7s). At this point, inundation will not cause any swelling or
swelling pressure. Net normal stress located at this point is referred to as the
transition pressure PT (Wijaya and Leong 2015).
Figure 13.5 shows the swelling test result on a kaolin specimen dried to
5% water content (25) and then subjected to stress path A-­B-­C-­D-­E. The
soil was reloaded and then unloaded prior to wetting to remove any distur-
bance on the soil specimen. On inundation, the soil swells from point E and
reaches point F. Any further loading will cause the specimen to follow the
saturated path F-­G-­H-­J compression line to be located at the NCL. The
unsaturated compression line (UCL) is obtained from the specimen which
was air dried to 1% water content (Kw1) and is observed to be parallel to
NCL.

pc,s
1.4
A
1.3
C Cr,d-1 B
1.2 F(P i,e s,i )
Cr,s
1.1 E(Pi,e d,i)
D(P c,d,e c)
H (Pc,s,e c)
1 Cr,d-2
e

0.9 G

0.8 CWC path A-B-C-D-E

Inundation path E-F
0.7 Saturated path F-G-H-J J

0.6
1 10 100 1000 10000
-ua (kPa)

Figure 13.5 W
 etting-induced swelling on specimen w5 (from Wijaya and Leong
2015).
238 Laboratory tests for unsaturated soils

13.3.6 Wetting-induced collapse
If the soil specimen is inundated at net normal stress higher than PT (i.e.,
point 6c) with Pi equal to Pi,2, the soil specimen will undergo irrevers-
ible volumetric compression, which is referred to as collapse or hydro-­
consolidation from point 6c to point 7s (located at the same net normal
stress). Collapse occurs as the unsaturated compression curve (points 2d-­3c-­
4c-­5c-­6c) is located above the saturated compression curve (points 8s-­9s-­
6s-­7s). Table 13.1 shows the relationship between swelling, collapse and
swelling pressure that can be obtained from the CWC test (Wijaya and
Leong 2015). Figure 13.6 shows a kaolin specimen subjected to the CWC
stress path A-­B-­C-­D-­E-­D and then wetted when it is on the UCL. It shows
that the soil volume decreases from point D to F and falls onto the NCL
following the saturated path F-­G.

13.3.7 Constant-suction compression test

For the constant-­suction (CS) test, the soil suction is maintained constant
given by the unsaturated compression curve (points 2d-­3cs-­4cs). As a result,
the yield pressure is a function of soil suction Pc(s). It is very difficult to

Table 13.1 R
 elationship between compressibility wetting-induced volume change/
pressure of soils in CWC test (Wijaya and Leong 2015)

Relationship Equation Parameters

Swelling/ If Pi < Pc,s Input parameters
collapse Pi = inundation pressure
P   Pc , d 
volume ∆e = C r , s log  c , s  − C r , d log  P  Pc,s = saturated preconsolidation
change  Pi   i  pressure
If Pc,s ≤ Pi ≤ Pc,d Pc,d = unsaturated preconsolidation
pressure
∆e = C c log ( Pc , s ) − C r , d log ( Pc , d ) Cr,s = saturated unloading/reloading
index
+ ( C r , d − C c ) log ( Pi ) Cr,d = CWC unloading/reloading
If Pi > Pc,d index
Cc = compression index
P  Output parameters
∆e = C c log  c , s 
 Pc , d  PT = transition pressure
Positive Δe indicates swelling Ps = swelling pressure
Negative Δe indicates collapse Δe = wetting-­induced volume change
represented by a change in the void
Transition  C r ,s log ( Pc ,s ) − C r ,d log ( Pc ,d ) 
ratio
 
pressure  ( C r , s − C r ,d ) 
PT = 10
Swelling If Pi < PT
pressure C r ,d / C r , s
 P 
Ps = Pc ,s  i 
 Pc ,d 
 Oedometer test 239

1.4

1.3 A

1.2 C B

1.1

1 E
e

D(Pc,d,ec)
0.9
F(Pc,s,es,i)
0.8
CWC path A-B-C-D-E-D
0.7 Inundation path D-F
G
Saturated path F-G
0.6
1 10 100 1000 10000
-ua (kPa)

Figure 13.6 W
 etting-induced collapse on specimen w10 (from Wijaya and Leong
2015).

maintain constant suction as compression causes instantaneous mechanical

wetting. Thus, the test must be carried out at a very slow loading rate to
allow the soil to equilibrate at the constant suction.
If the water content of the soil is very low, say at point 3d, the degree of
saturation hardly changes, and thus mechanical wetting is insignificant. At
such low water content (or high suction), the soil-­water characteristic curves
(w-SWCC and S-SWCC) merge into a single line as indicated in Chapter 10.
Thus, the CWC test located at very low water contents is almost equivalent
to the CS test.

13.3.8 Configuration of oedometer tests

Oedometer tests can be differentiated into:

1. Saturated oedometer test (BS EN ISO 17892-­5 2017; ASTM D2435/

D2435M 2011).
2. Suction-­control oedometer test (SC test).
3. Constant water content oedometer test (CWC test).

The saturated oedometer test is a compression test (under incremental load-

ing), where the soil specimen is assumed to remain saturated throughout the
test (Sr = 100%). The saturation is achieved by inundating the soil specimen.
It has been standardised in ASTM D4186 (2006) and BS EN ISO 17892-­5
(2017) and mentioned in many books (e.g., Head 2011). Thus, it will not be
further elaborated in this book.
The SC test requires extensive modification of the oedometer set-­up as it
requires suction control to be incorporated. Suction is controlled usually by
240 Laboratory tests for unsaturated soils

using the axis-­translation technique (Kayadelen 2008; Aversa and Nicotera

2002), applying constant relative humidity (Salager et al. 2011) and apply-
ing constant suction through a semi-­permeable membrane via a salt solu-
tion (Delage et al. 1992; Bulolo and Leong 2021). If the axis-­translation
technique is adopted, the top cap must be modified to allow air pressure to
be applied to the top of the specimen, the porous stone at the bottom of the
specimen must be replaced with a sealed high-­air entry ceramic disk and the
whole specimen must be in an air-­tight chamber to maintain the air pres-
sure. Suction measurement devices such as the high-­capacity tensiometer
can also be installed (although not often done) to monitor the specimen
suction throughout the test. If the relative humidity method is adopted, air
at a certain relative humidity is circulated through the specimen. For suc-
tion control via axis-­ translation technique and relative humidity, the
oedometer needs to be enclosed within a sealed chamber. The oedometer
set-­up where a constant suction is applied to the soil specimen through a
semi-­permeable membrane via a salt solution is commonly referred to as an
osmotic oedometer.
In the SC test, the suction is maintained constant throughout the com-
pression test in order to obtain the constant-­suction compression curve.
As any stress change in unsaturated soil may cause an instantaneous
change in the void ratio, only a small stress increment can be used, and suc-
tion equilibrium must be achieved during each loading/unloading stage. As
a result, the SC test requires a very long test duration to be finished.
The procedure of the SC test is highly dependent on the technology used
to control suction. Considering the cost and testing duration required in the
SC test, the application of the SC test remains very much in the research
laboratory and is difficult to implement in practice.
The CWC test does not require any modification to the oedometer appa-
ratus. However, there is no control of soil suction or pore-­water pressure
measurement (due to latency in pore-­water pressure response to loading). As
soil suction is not controlled, the CWC compression curve is obtained instead
of the constant-­suction compression curve. Thus, the interpretation requires
an alternative soil state, such as water content or degree of saturation.
As noted earlier, when the water content of the soil is very low such that
void ratio, degree of saturation and suction are essentially constant during
the test and the test approximates a CS test, the CWC test on such a soil
specimen is similar to the SC test, and the compression curves can be treated
as similar. Table 13.3 summarises some observations on the compression
behaviour of unsaturated soils under CWC and SC tests.
If the CWC test incorporates suction measurement devices such as a high-­
capacity tensiometer (HCT), it is suggested to refer the test as CWC-­P. One
possible set-­up of an oedometer with a HCT is shown in Figure 13.7. The
advantages and disadvantages of each test are summarised in Table 13.2.
 Oedometer test 241

Figure 13.7 Modified base for CWC-P (from Wijaya and Leong 2016).

13.4 TEST METHODS

When testing unsaturated soil, a saturated soil test is also needed to provide
a reference. In double oedometer test (Jennings and Knight 1957), two iden-
tical soil specimens are tested: one under saturated condition and the other
under unsaturated condition (CWC test or SC test). In this section, general
descriptions of oedometer test procedures for both saturated and unsatu-
rated soils are given.

13.4.1 Saturated oedometer test

The test method for the saturated oedometer test has been standardised in
BS EN ISO 17892-­5 (2017), ASTM D2435/D2435M (2011), BS 1377-­5
(1990), and it is briefly summarised next. More details can be found in Head
(2011). There are three testing stages in the saturated oedometer test:

1. Specimen preparation stage

2. Inundation stage
3. Loading/unloading stage

13.4.1.1 Specimen preparation stage

The specimen preparation stage involves the following:
242 Laboratory tests for unsaturated soils

Table 13.2 Advantages and disadvantages of CWC procedures

Procedures Advantages Disadvantages

Suction-­control 1. Easy interpretation 1. The most expensive type
oedometer test as the compression of oedometer test.
(SC test) curve is assumed 2. Maintaining the
to fall under a constant-­suction
single unique soil condition is extremely
suction. difficult and technology
2. Controlled suction dependent.
wetting and drying 3. The equilibrium time
are possible. required to maintain soil
suction caused the testing
duration to be extremely
long.
4. The soil specimen is
expected to undergo
mechanical wetting
first before the soil
suction is recovered
to the soil suction of
interest. As a result, the
specimen experiences
complicated wetting and
drying path.
Constant water 1. Relatively low cost 1. HCT has to be sufficiently
content oedometer in incorporating pressurised. The
test with pore-­ the HCT. pressurising duration
water pressure 2. Pore-­water depends on the high-­
measurement pressure reading capacity tensiometer and
(CWC-­P) is available the pressure magnitude.
throughout the 2. Correlating the
test. compressibility with soil
3. Test duration suction remains difficult.
is relatively
short as it is
controlled by the
equilibrium time
between HCT and
soil specimen
(in a range of 20
minutes to 1 hour
per test).
Constant water 1. The lowest cost, 1. Correlating soil
content oedometer as there is no compressibility with soil
test without pore-­ need to modify suction is not possible.
water pressure the conventional
measurement oedometer test.
(CWC) 2. Has the shortest
testing duration.
 Oedometer test 243

Table 13.3 Summary of unsaturated compression behaviour modified from Wijaya (2017)

Constant variable
plane Comment
Constant water 1. Difficult to separate the pure compression behaviour
content from the mechanical collapse behaviour (Alonso et al.
1990).
2. Drying the soils will increase preconsolidation pressure
and the preconsolidation void ratio (Leong et al. 2013;
Jotisankasa 2005; Jotisankasa et al. 2007; Wijaya and Leong
2015; Wijaya and Leong 2016; Wijaya 2017).
3. Drying will not cause any further change in the
preconsolidation void ratio once the soil has reached the
zero-­shrinkage line (Wijaya 2017)
4. The UCL can either merge or not merge with the
saturated NCL, but the slope of the UCL seems to
increase (less stiff) rather than decrease (Jotisankasa 2005;
Burton et al. 2014).
5. The UCL is constant under the constant degree of
saturation condition and will merge into the saturated
compression line as the soil becomes saturated due to
mechanical wetting (Wijaya and Leong 2015; Wijaya 2017).
6. Matric suction of soil will decrease when total stress is
applied due to mechanical wetting (Wijaya 2017; Wijaya
and Leong 2016; Jotisankasa et al. 2007)
7. Matric suctions at the top and bottom of the specimen
may differ by around 5%–15% for a thin specimen of
about 20 mm (Jotisankasa 2005).
8. The preconsolidation pressure is not only due to the
past maximum compaction stress but also due to the
inter-­particle bonding which is caused by matric suction
(Munoz et al. 2011).
9. Swelling does not change the preconsolidation void ratio
(Wijaya 2017; Wijaya and Leong 2015)
10. The suction-­degree of saturation relationship in the
unloading-­reloading path is reversible. However, the
matric suction which is recorded prior to unloading is
higher than the matric suction which is recorded at the
end of the unloading-­reloading cycle (Tarantino and De
Col 2008).
11. Change in matric suction due to the change in net normal
stress under the CWC condition becomes less significant
as the soil becomes drier (Rahardjo 1990; Jotisankasa et
al. 2007).
12. The slope of the unloading-­reloading line appears to be
independent of matric suction (Munoz et al. 2011).
13. The slope of the unloading-­reloading line under the CWC
condition is much smaller compared to the slope of the
unloading-­reloading line under the saturated condition.

(Continued)
244 Laboratory tests for unsaturated soils

Table 13.3 (Continued)

Constant variable
plane Comment
Constant matric 1. An increase in constant suction will increase the
suction preconsolidation pressure (Alonso et al. 1990; Lloret et al.
2003).
2. There are a number of uncertainties in the UCL of
the unsaturated soils, and all of them are supported by
experimental observations. The slope of the UCL may
either reduce with suction (Alonso et al. 1990; Pham 2005),
increase with suction and then merge with the saturated
NCL (Toll 1990; Alonso 1993; Honda 2000; Sivakumar and
Wheeler 2000; Futai et al. 2002; Silva et al. 2002; Wheeler
and Sivakumar 1995) or be unaffected by matric suction
(Larson et al. 1980; O’Sullivan 1992; Smith et al. 1997;
Arvidsson and Keller 2004; Imhoff et al. 2004).
3. When the degree of saturation equals 1, the UCL merges
with NCL regardless of the matric suction of the soils
(Mun and McCartney 2015).
4. The slope of the unloading-­reloading line appears to be
independent of matric suction (Alonso et al. 1990; Wheeler
and Sivakumar 1995; Munoz et al. 2011; Hoyos et al. 2012).
5. Soils which are dried and then wetted to the desired
matric suction are more susceptible to yield rather than
soils which are only dried to the desired matric suction
(Sivakumar et al. 2006).

1. Set up the oedometer apparatus.

The oedometer must be set up such that it does not cause a change in
water content or cause the specimen to swell. Thus, the procedures
depend on soil type as follows:
• For dry expansive soil, dry porous disk and dry filter paper must
be used (ASTM D2435/D2435M 2011).
• For initially unsaturated soil, damp porous disk and filter paper
may be used (ASTM D2435/D2435M 2011).
• For saturated soil with low affinity for water, saturated disk
and saturated filter paper may be used (ASTM D2435/D2435M
2011).
2. Place the specimen inside the oedometer apparatus.
If the specimen is placed into the oedometer apparatus and the test is
not started immediately, a loose-­fitting plastic or rubber membrane is
used to cover the oedometer to minimise water content change due to
evaporation (ASTM D2435/D2435M 2011).
3. Apply the seating load on the specimen
ASTM D2435/D2435M (2011) recommends a nominal seating load
of about 5 kPa, but a smaller seating load can be used if it is suspected
that the soil will consolidate significantly under the nominal seating
 Oedometer test 245

load. BS EN ISO 17892-­5 (2017) recommends a seating load not ex-

ceeding 3 kPa, while BS 1377-­5 (1990) recommends a seating load not
exceeding 2 kPa.

13.4.1.2 Inundation stage
Inundation is carried out such that the specimen is not allowed to swell
excessively prior to being loaded beyond its preconsolidation pressure
(ASTM D2435/D2435M 2011). Inundation can be carried out as follows:

1. If the soil is saturated under field condition, ASTM, BS and BS EN ISO

suggest proceeding to the loading stage immediately after the applica-
tion of seating load to prevent the specimen from swelling.
2. If a specific condition is to be achieved, ASTM D2435/D2435M
(2011) suggests applying a sufficiently large load (large enough to pre-
vent the soil from swelling above its initial height) and inundating the
specimen under the applied load. Record the inundation stress (verti-
cal stress at which the soil is inundated) and the axial deformation
during the inundation.

13.4.2 Loading/unloading stage
The test procedures are briefly summarised as follows:

1. Load is applied to the soil specimen under K0 condition (the specimen

is confined in an oedometer ring).
2. In the oedometer test, pore-­water pressure measurement is generally
not carried out (ASTM D2435/D2435M 2011). To determine the end
of consolidation without measuring the pore-­water pressure, several
methods can be used:
a. ASTM D2435/D2435M (2011) Method A and BS 1377-­5
(1990): Use a standard load increment duration of 24 hours.
Take at least two time-­versus-­deformation readings at two load
increments, one of which must be taken after the preconsolida-
tion pressure. The test duration may be extended if the time-­
versus-­deformation reading during the 24 hours is suspected to
be insufficient.
b. ASTM D2435/D2435M (2011) Method B and BS EN ISO
17892-­5 (2017): Use the settlement versus time curve to ensure
that primary consolidation is completed, which can be shorter or
longer than 24 hours.
3. Settlement is plotted against square-­root time or log time to obtain
t50 (time to 50% degree of consolidation) or t90 (time to 90%
degree of consolidation), respectively, to determine the coefficient of
246 Laboratory tests for unsaturated soils

consolidation (cv) and the permeability of the saturated soils (ks) at the
average void ratio before and after load increment.
4. Next, load increment is then applied, and steps 1 to 3 are repeated
until the last load increment. Generally, a load increment ratio of
one is used. Different numbers of load increments are suggested.
BS 1377-­5 (1990) recommends four to six load increments. BS EN
ISO 17892-­5 (2017) recommends at least seven load increments,
while ASTM D2435/D2435M (2011) recommends that the num-
ber of load increments should be such that the following can be
obtained:
a. Three points which define a straight line in a log stress space
b. Three points which define a concave up curve when plotted in log
stress space
c. Stress level which is eight times the estimated preconsolidation
pressure
Convenient load increments to use are 6, 12, 25, 50, 100, 200, 400, 800,
1,600 and 3,200 kPa.

5. An unload-­reload is performed at the end of the last load increment if

required. For unloading, the load is removed in two load decrements
and reloaded back to the maximum load. For the unload-­reload, it
is sufficient to wait until the vertical dial gauge stops moving at each
load decrement or increment which is expected to occur within a short
time.
6. After the oedometer test, a compression curve of void ratio versus
load is plotted. The compression index (Cc), saturated reloading index
(Cr,s), saturated unloading index (Cu,s), saturated unloading-­reloading
index (Cur,s) and saturated preconsolidation pressure (Pc,s) can be
determined from the compression curve.

13.4.3 Unsaturated oedometer test

For the oedometer test on unsaturated soil (either CWC or SC test), there
is an additional stage at the beginning compared to the oedometer test for
saturated soil, which is the suction application stage.
The suction application stage is to bring the soil specimen to the initially
desired soil suction. This stage can be carried out by air drying the sample
or using an axis-­translation apparatus such as the pressure plate. The initial
suction application stage is carried out prior to cookie cutting the sample
into the oedometer ring. In the SC test, the initial suction can also be applied
to the specimen in the oedometer apparatus. However, shrinkage of the
specimen may occur, and the specimen may shrink away from the sides of
the oedometer ring. The advantage of the SC test over the CWC test is that
 Oedometer test 247

the initial suction of the specimen can be ensured prior to testing. The pro-
cedures for the oedometer test on unsaturated soil are similar to the oedom-
eter test for saturated soil.

13.4.4 CWC and CWC-P tests

CWC and CWC-­P tests have similar testing procedures except that suc-
tion is monitored during the CWC-­P test. The advantage of the CWC-­P
test is that it is able to provide suction versus void ratio (e) relationship.
However, it is not possible to carry out the CWC-­P test for soil suction
higher than the maximum capacity of the HCT (typically 1,500 kPa). The
CWC test does not have this limitation. Further information on HCT is
given in Chapter 8.
It is important to note that in the CWC test, the water content of the
soil must be maintained constant throughout the test. Thus, the porous
stone and the filter paper must be dry (ASTM D4546-­08, 2008), as wet
filter paper and the wet porous disk might change the soil’s water con-
tent. However, a dry porous disk is affected by evaporation, especially
when the test takes a very long time. One method that can be used to
reduce evaporation is by using a plastic disk at the top and bottom of the
specimen.

13.4.4.1 CWC loading
Test procedures:

1. Increase/decrease the net normal stress applied to the soil specimen.

2. Wait until pore-­water pressure reaches equilibrium. When pore-­water
pressure remains negative on increase of net normal stress, 20 minutes
is sufficient. It is not recommended to wait for a long duration due to
the possibility of water loss due to evaporation. However, when pore-­
water pressure becomes positive on increase of net normal stress 24
hours wait time is required, following the loading duration for satu-
rated soils to ensure that the excess pore-­water pressure has dissipated.
3. Record the final compression (and hence, void ratio) and the final soil
suction after equilibrium.
4. Repeat steps 1 to 3.

Parameters to be reported: σi, ei, si, wi, Sri

The primary compression of unsaturated soil under the CWC condition is
very short as the permeability of air is much higher than the permeability of
water. Therefore, the change in negative-­pore-­water pressure is due to the
change in void ratio, and this is designated as mechanical wetting/drying.
However, HCT may have a delayed response time, and therefore it is
248 Laboratory tests for unsaturated soils

important to wait until HCT reaches equilibrium in order to obtain the cor-
rect pore-­water pressure.

13.4.4.2 Wetting under constant net normal stress

Test procedures:

1. Inundate the soil specimen by filling the consolidometer chamber with

water. The water must be either distilled water or water with similar
chemical composition as the in situ pore water.
2. Observe the change of soil volume and pore-­water pressure until pore-­
water pressure reaches 0 kPa and there is negligible change in soil
volume. Usually, 48 hours duration is sufficient for the pore-­water
pressure to reach 0 kPa.
3. Record the amount of swelling Δe when pore-­water pressure becomes
zero.

Parameters to be reported: Δe, soil-­swelling characteristic curve, swelling

versus time graph
The advantage of CWC-­P is that it is possible to obtain a continuous
relationship between the void ratio and matric suction under the swelling
condition in a relatively short duration compare to the SC test. Therefore,
additional in situ swelling or shrinkage can be estimated based on the change
in the in situ soil suction.
In CWC-­P, either top or bottom is usually undrained due to the placement
of HCT. Thus, it is not recommended to saturate the soil specimen from the
bottom using a pressure line, as it may introduce additional water pressure
which can lift the soil specimen.

13.4.4.3 Wetting under constant volume

Test procedures:

1. Inundate the soil specimen in the oedometer chamber with water. The
water must be either distilled water or water with similar chemical
composition as the in situ pore water.
2. Whenever the soil starts to swell, add additional net normal stress to
maintain the volume of the soil specimen.
3. Once the soil suction has reached equilibrium, record the change in
net normal stress and soil suction. Usually, 48 hours duration is suf-
ficient for the soil to reach equilibrium.

Parameters to be reported: Swelling pressure, suction versus net normal

stress curve
 Oedometer test 249

Based on this stage, it is possible to obtain a continuous relationship

between the change in net normal stress and soil suction. Thus, in situ swell-
ing pressure can be determined based on the change of the in situ soil suc-
tion. However, as it is impossible to maintain no volume change at all during
this stage, care must be taken when interpolating the test result.

13.4.5 Stages in the CWC oedometer test

The CWC oedometer test is the simplest and only requires a conventional
oedometer test. The loading duration is shorter than the saturated oedom-
eter test, as the permeability of air is much higher than the permeability of
water and compression is almost simultaneous on loading. However, as it
does not give any information regarding matric suction, it is difficult to use
CWC test results in models developed for the SC test or CWC-­P test. The
easiest way to use CWC test results is to duplicate the in situ loading stage
and interpolate the CWC test results to estimate the in situ volume change.
Another way is to use another state variable such as water content or degree
of saturation as the model parameters.
The testing stages in CWC can be summarised as follow:

1. Specimen preparation
2. CWC loading/unloading
3. Wetting under constant net normal stress
4. Wetting under constant volume

13.4.5.1 Specimen preparation
Test procedures:

1. Set up the specimen with air-­dried porous stones and air-­dried filter
papers in the oedometer.
2. Use a plastic sheet, moist paper towel or aluminium foil to cover the
soil specimen to minimise evaporation.
3. Apply seating stress to the unsaturated soil specimen.

13.4.5.2 CWC loading/unloading
Test procedures:

1. Increase/decrease the net normal stress applied to the soil specimen.

2. Record the change in height of the specimen and calculate the change
in degree of saturation assuming that the water content remains con-
stant during the loading stage. If the degree of saturation reaches
100%, recalculate the water content.
250 Laboratory tests for unsaturated soils

3. Use a loading duration of either 5, 10 or 20 minutes. It is important to

observe the change in void ratio with time for the first few loadings in
order to ensure that 20 minutes is sufficient. Subsequent loading can
be carried out when there is negligible change in void ratio. Loading
duration should not be excessive to avoid water loss from the soil
specimen due to evaporation.

Parameters to be reported: σi, ei, wi, Sri

13.4.5.3 Wetting under constant net normal stress

Test procedures:

1. Inundate the soil specimen by filling the oedometer chamber with water
or via a water pressure line connected to the bottom of the specimen.
Use either distilled water or water with the same chemical composition
as the in situ pore water. Saturation using a water pressure line from
the bottom is preferred as air can be displaced from the bottom of the
specimen upwards such that all the porous stone and soil specimens
can become fully saturated. However, care must be taken to ensure
that the water pressure does not uplift the soil specimen.
2. Observe the volume change until there is negligible change in soil vol-
ume. Usually, 48 hours is sufficient for the soil to reach equilibrium.
3. Record the amount of swelling Δe.

Parameters to be reported: Δe, swelling versus time graph

This stage provides the maximum swelling potential of the soil.

13.4.5.4 Wetting under constant volume

Test procedures:

1. Inundate the soil specimen by filling the oedometer chamber with

water or by using a water pressure line connected to the bottom of the
specimen. Either distilled water or water with the same chemical com-
position as the in situ pore water is used. Using a water pressure line
to saturate the soil specimen is preferred, as air can be displaced from
the bottom of the soil specimen upwards to ensure that the porous
stone and soil specimen becomes fully saturated. However, care must
be taken such that the water pressure does not lift the soil specimen.
2. Whenever the soil starts to swell, add additional net normal stress to
maintain the volume of the soil specimen.
3. Record the change in net normal stress once the soil has reached equi-
librium. Usually, 48 is sufficient for the soil to reach equilibrium.

Parameters to be reported: Swelling pressure

 Oedometer test 251

This stage provides the maximum swelling pressure which can be exerted
by the soil.

13.4.5.5 Saturated loading/unloading stage

Test procedures:

1. Apply the load either increase (loading) or decrease (unloading) once

the wetting stage is completed.
2. Maintain each loading stage for up to 24 hours or until there is negli-
gible change in volume and pore-­water pressure.

13.4.6 CWC and CWC-P oedometer test procedures

CWC test has been in use for a very long time due to its simplicity and some
of the testing procedures have been standardised in ASTM:

1. ASTM D3877 (2008) for one-­dimensional expansion, shrinkage and

uplift pressure of soil-­lime mixtures.
2. ASTM D4546 (2008) for one-­dimensional swell or collapse of cohe-
sive soil.
3. ASTM D5333 (2003) for measurement of collapse potential of soils.
4. ASTM D4829 (2011) for obtaining the expansion index of soils.

The ASTM standards mainly focus on swelling and collapse behaviours.

However, the volume change of unsaturated soil is sometimes complicated
as swelling and collapse are also affected by the change in net normal
stress and suction or water content. For more robust testing procedures,
Wijaya (2017) recommended a testing procedure to investigate the volume
change of unsaturated soil due to the change in net normal stress and
change in water content.

13.4.6.1 ASTM D3877-08 for one-dimensional expansion,

shrinkage and uplift pressure of soil-lime mixtures
Lime is known to reduce the swelling potential of soils. However, different
types of soil may require a different amount of lime to give optimum results
in terms of cost and reduction in the amount of swelling. The soil-­lime mix-
ture must be prepared in accordance with ASTM D3551 using a mechanical
mixer. The soil-­lime mixture is made as follows:

1. Sieve the soil particles such that the particle size is less than 4.75 mm
(Sieve no. 4).
2. Air or oven dry the soil for around 24 hours. If air drying or oven dry-
ing is suspected to affect the results, the soil should be cured at or near
its final moisture content.
252 Laboratory tests for unsaturated soils

3. Determine the water content of the air-­dried specimen.

4. Determine the weight of the soil which is required to provide the oven-­
dry weight of soil for the desired number of test specimens.
5. Determine the amount of lime based on the oven-­dry weight of soil
required for the final mixture.
6. Place both the dried soil and the lime inside a bowl.
7. Mix dried soil and lime for either about 1 minute or until the mixture
appears uniform in colour.
8. Weigh or measure the required weight of water to produce the desired
water content. The initial amount of water in the air-­dried soil must be
considered. Considering the possibility of evaporation, it is suggested
to use the amount of water such that the final water content is around
1% + the desired water content.
9. Add the water into the soil mixture in a thin stream or a fine spray
while mixing the soil with a mixer.
10. Mix the soil for two and a half minutes after all the water has been added.
11. Stop the mixer and then scrape all parts to return any caked mixture
into the mixing bowl.
12. Repeat steps 10 and 11.
13. Blend the mixture briefly with a trowel or spatula and form a lightly
compacted mound at the bottom of the bowl.
14. Cover the bowl immediately to minimise the evaporation of moisture
and let the soil-­lime-­water mixture reach equilibrium for around 1 hour.
15. Weigh the oedometer ring.
16. Assembled the oedometer along with the oedometer ring, oedometer
collar and load cap.
17. Measure the position of the load cap r1 by using a dial gauge or lin-
ear variable differential transformer (LVDT). The dial gauge or LVDT
must be positioned at the same place when the dial gauge or LVDT
will be used during the test.
18. Assemble the oedometer ring and the extension collar. An extension
collar is used such that a loose soil mixture can be placed.
19. Compact the specimen in the oedometer ring to the desired wet unit
weight by using a suitable compaction hammer. The minimum height
of the specimen shall be 19 mm and 6 mm greater than the depth of
the ring gauge.
20. Remove the extension collar and trim the excess soil from the top of
the soil specimen. It is preferred to make the specimen height equal to
the oedometer ring height for ease of the initial height measurement.
21. Determine the water content of the soil-­lime mixture using the trimmed
soil.
22. Weigh the specimen with the ring.
23. Cover the exposed surface of the specimen using a glass plate, held
with clamps until the specimen is placed in the oedometer.
24. Compute the initial wet density of the specimen.
 Oedometer test 253

25. If the initial wet density of the specimen is not obtained, discard the
specimen and redo the specimen preparation.
26. If there is a need to cure the soil-­lime specimen, it should be done once
the desired wet density has been obtained.
27. Place the soil specimen inside the oedometer and use a seating load of
approximately 2.5 kPa.
28. Re-­measure the position of the load cap r2 by using the dial gauge or
LVDT. The difference between r2 and r1 should give the height of the
specimen. Alternatively, if the soil specimen has the same height as the
oedometer ring, the height of the specimen can be taken as the height
of the oedometer ring.

Once the soil mixture has been prepared, the soil mixture can then be loaded.
There are three procedures available for estimating the swelling pressure of
soil-­lime mixtures:

1. Loaded and expanded

2. Expanded then loaded
3. Individual load expansion

The testing procedures for loaded and expanded are given as follows:

1. Wetting under constant volume conditions: Record the maximum

load.
2. Unloading stage: Reduce the load to ½, ¼ and 1/8 of the maximum
load and finally to the seating load. Maintain each load for either 24
hours or longer such that there is no change in the void ratio. Record
the change in the void ratio (or height) for each unloading step.
3. Determine the water content of the soil specimen.

The testing procedures for expanded and then loaded are given as follows:

1. Wetting under constant net normal stress: Allow the soil specimen to
expand for at least 48 hours or until there is negligible change in the
void ratio.
2. Loading stage: If the loaded and expanded test procedure has been
conducted, the applied load is equal to 1/8, ¼ and ½ of the maximum
load. The subsequent loading stage can be done using a load incre-
ment ratio (LIR) of 1. If no loaded and expanded test procedure has
been carried, loading follows LIR of 1 as usually carried out in an
oedometer test until the load reaches the capacity of the oedometer
apparatus, usually either 3,200 or 6,400 kPa.
3. Unloading stage: Unload until the seating load. The LIR of 1 may be
used for a normal oedometer test.
4. Determine the water content of the specimen.
254 Laboratory tests for unsaturated soils

The test procedures for individual load expansion are given as follows:

1. CWC loading stage: Increment the applied load until the desired net
normal stress.
2. Wetting under constant net normal stress: Wet the soil for either 48
hours or until there is no more change in soil volume.
3. Unloading stage: Decrement the load using a load decrement ratio
(LDR) of ½ until the seating load.
4. Determine the water content of the specimen.

The “loaded and expanded” procedure will give different results compared
to the “expanded then loaded” procedure. While the “loaded and expanded”
procedure is believed to give a more reliable result, it requires greater profi-
ciency especially in controlling the load to be applied during wetting under
the constant volume condition. Therefore, carrying out both test procedures
is better for comparison. The “individual load expansion” procedure is car-
ried out when the in situ soil will be subjected to load prior to wetting. Thus,
the applied load must simulate the in situ condition.

13.4.6.2 ASTM D4546-08 for one-dimensional swell or collapse

of cohesive soil
ASTM D4546-­ 08 describes the procedure to measure wetting-­ induced
behaviour under different net normal stresses and the post-­wetting behav-
iour of natural or compacted soils. There are three methods:

1. Method A: Wetting-­after-­loading test on multiple specimens

2. Method B: Wetting-­after-­loading test on a single specimen
3. Method C: Loading-­after-­wetting test

The specimen preparation is identical for all three test methods described
as follows:

1. Cookie cut the soil specimen using the oedometer ring.

2. Trim and ensure that the specimen height is less or equal to the height
of the oedometer ring.
3. If the soil specimen’s height is less than the height of the oedometer
ring, measure the height of the specimen at three or four places.
4. Use air-­dried porous stones and dry filter paper (or no filter paper).
5. Apply a seating load of 1 kPa.
6. Set the displacement indicator.

The three test methods are summarised in Table 13.4.

Table 13.4 ASTM D4546-08 procedures

Stage Method A Method B Method C

Specimen Prepare as many specimens as needed Prepare a single specimen to be Prepare as many soil specimens as needed
preparation following the previously mentioned tested. following the previously mentioned
procedure to be wetted under procedures to be wetted under different
different net normal stress. The net normal stresses. The wetting pressure
wetting pressure must be selected must be selected such that it covers the
such that it covers the stress range stress range that may be experienced in
that may be experienced in situ situ starting from before construction
starting from before construction until the future possible load applied after
until future possible load applied after the construction is finished. All of the soil
the construction is finished. All of the specimens can be tested at the same time or
specimens can be tested at the same in sequence.
time or in sequence.
CWC loading Increment the net normal stress at CWC loading stage: Increment CWC loading stage: Increase the net normal
stage LIR of 1 until it reaches the wetting the net normal stress using stress using LIR of 1 until it reaches the
pressure. LIR of 1 until it reaches the in wetting pressure.
situ net normal stress at its
sampling depth.
CWC – Decrement the net normal –
unloading stress using LDR of ½ until it
stage reaches the seating pressure.

Oedometer test
(Continued)

255
Table 13.4 (Continued)

256
Stage Method A Method B Method C

Laboratory tests for unsaturated soils

CWC reloading – Increase the net normal stress –
stage up to the in situ net normal
stress corresponding to the
prior or post-­construction
condition, whichever is of
interest.
Wetting under Inundate the soil specimen with water for either 24 to 72 hours, or until there is negligible change in the volume of the
constant net specimen. Water should be either distilled water or water with similar chemical composition as the primary water source
normal stage of wetting in the field.
Saturated – – Increment the applied load using LIR of 1 up
loading stage to either 3,200 or 6,400 kPa. Each loading
stage is maintained for either 24 hours until
there is no excess pore-­water pressure, or
there is negligible change in volume.
Saturated – – Decrement the applied load using LDR of
unloading ½until its seating pressure. Each unloading
stage stage is maintained for either 24 hours or
until there is negligible change in volume and
pore-­water pressure.
Drying stage Remove all water inside the oedometer using suction devices without –
unloading the specimen, and then use filter paper to soak up all free water
in the oedometer.
CWC Remove all the vertical load applied to the soil specimen at once. –
unloading
stage
Specimen Remove the soil specimen and use dry filter paper to remove any free water on the surface of the soil specimen.
removal
Water content Determine the water content of the soil specimen.
determination
 Oedometer test 257

13.4.6.3 ASTM D5333-03 for measurement of collapse

potential of soils
ASTM D5333-­03 determines the collapse potential of soil under two
conditions:

1. At net normal stress corresponding to the desired in situ net normal

stress (prior to or post-­construction condition), Ic
2. At net normal stress equal to 200 kPa where the collapse potential is
referred to as collapse index, Ie

The specimen which is tested can be either a remoulded, compacted or

undisturbed specimen. The test procedures are given as follows:

1. Specimen preparation stage: Apply seating load of 5 kPa.

2. CWC loading stage: Increment the applied load using LIR of 1 with 1
hour loading duration until it reaches the desired load (either 200 kPa
to obtain Ie, or another load to obtain Ic).
3. Wetting under constant net normal stress stage: Inundate the speci-
men with water for either 24 hours or until there is no more change
in volume. Use distilled-­deionised water to determine Ie and use water
from the in situ wetting source for Ic. Water must flow upwards from
the bottom of the soil specimen to avoid air being trapped inside the
soil specimen.
4. Saturated loading stage: Increment the applied load using LIR of 1
with 24 hours loading duration until the applied load reaches either
3,200 kPa or 6,400 kPa.
5. Saturated unloading stage: Decrement the applied load using LDR of
½ with 24 hours loading duration until the applied load reaches seat-
ing pressure.
6. Take out the soil specimen from the oedometer and determine the final
water content.

13.4.6.4 Wijaya (2017) CWC oedometer test procedure

Wijaya (2017) proposed a method to estimate both swelling and collapse by
using at least two specimens. The first specimen is subjected to a saturated
compression test while the second specimen is subjected to a CWC oedometer
test. The stage procedures for the CWC oedometer test are given as follows:

1. Initial drying stage: Air dry the soil specimen until it reaches almost
zero water content to obtain the shrinkage curve. Measure the dimen-
sions of the soil specimen.
2. Specimen preparation stage: Cookie cut the air-­dried soil specimen
using the oedometer ring. If the soil becomes too brittle, use a plastic
258 Laboratory tests for unsaturated soils

film to wrap the perimeter of the specimen in order to provide a small

confinement during the cutting process. Assemble the soil specimen
into the oedometer apparatus.
3. CWC loading stage: Increase the applied load until it reaches the effec-
tive overburden stress. Maintain each loading for 20 minutes.
4. CWC unloading stage: Decrease the applied load using LDR of ½
until it reaches the seating pressure.
5. CWC loading stage: Increase the applied load using LIR of 1 until it
reaches 6,400 kPa.
6. CWC unloading stage: Decrease the applied load using LDR of 1 until
it reaches the seating pressure.
7. Remove the soil specimen from the oedometer and determine the final
water content of the specimen.

As the soil specimen has been air dried until almost zero water content, the
soil will no longer be affected by evaporation. By combining the param-
eters from the saturated and unsaturated compression curve, it is possible
to estimate the volume change of the unsaturated soil which is subjected to
a complex stress path.

13.4.7 SC Oedometer test

13.4.7.1 Constant-suction loading/unloading stage
Stage procedure:

1. Increase/decrease net normal stress σi.

2. Wait for soil suction to be in equilibrium with the applied suction.
3. Record void ratio after the soil specimen has reached equilibrium ei.
4. Repeat steps 1 and 2 until the end of the constant-­suction loading/
unloading stage.

The constant-­suction loading/unloading stage is carried out to obtain a

constant-­suction compression curve and to investigate the wetting-­induced
volume change under specific suction (Vázquez et al. 2013; Cuisinier and
Masrouri 2005). It is important to note that during the loading stage, there
will be mechanical wetting which induces swelling/collapse of the specimen,
while unloading the specimen may cause mechanical drying which may
cause the soil specimen to be stiffer.
As the collapse phenomenon is irreversible, the compression curve which
is affected by the collapse phenomenon may not represent the compression
curve under the constant-­suction condition. Thus, to reduce the effect of
mechanical wetting and drying, it is recommended to use small stress incre-
ments. Nevertheless, very small stress increments may cause the test dura-
tion to be excessively long (Cuisinier and Masrouri 2005).
 Oedometer test 259

13.4.7.2 Suction decrease (SD) stage

Stage procedure:

1. Record the net normal stress at SD stage σi.

2. Decrease the applied suction from s0 to si.
3. Wait until soil suction and the applied suction are in equilibrium.
4. Record the change in void ratio Δe after equilibrium.
5. Either repeat steps 1 to 3 or end this stage.

SD stage will cause either swelling or collapse. Collapse settlement is depen-

dent on the applied load direction and thus will occur only in the vertical
direction. The parameter Ks,I and Cw,I are given by Equations 13.5 and 13.6,
respectively.

e ( si ) − e ( s0 )
κ s,i = (13.5)
s 
ln  0 
 si 

e ( si ) − e ( s0 )
Cw,i = (13.6)
s 
log  0 
 si 

For the SD stage, since si is smaller than s0, κs,i and Cw,i are positive when
the soil specimen is swelling and negative when the soil specimen is collaps-
ing. The two parameters are equivalent and are a function of σi.
While it is possible to carry out the suction increase (SI) stage in the
oedometer test, it is not recommended, as increasing soil suction may
cause volume change to occur not only in the vertical direction but also in
the radial direction, and may not be homogeneous. Volume change in the
radial direction may cause non-­k0 conditions and cause difficulties in the
interpretation. The SI stage is only recommended for the SC triaxial test.

REFERENCES

Alonso, E. E. (1993). Unsaturated soil: Recent developments and application –

constitutive model of unsaturated soils. Barcelona: Civil Engineering European
Courses Programme of Continuing Education.
Alonso, E. E., Gens, A., & Josa, A. (1990). A constitutive model for partially satu-
rated soil. Géotechnique, 40(3): 405–430.
ASTM D2435/D2435m (2011). Standard test methods for one-­dimensional con-
solidation properties of soils using incremental loading. West Conshohocken, PA:
ASTM International.
260 Laboratory tests for unsaturated soils

ASTM D3877 (2008). Standard test methods for one-­ dimensional expansion,
shrinkage, and uplift pressure of soil-­lime mixtures. West Conshohocken, PA:
ASTM International.
ASTM D4546 (2008). Standard test methods for one-­dimensional swell or collapse
of cohesive soils. West Conshohocken, PA: ASTM International.
ASTM D4829 (2011). Standard test method for expansion index for soils. West
Conshohocken, PA: ASTM Standard.
ASTM D5333 (2003). Standard test method for measurement of collapse potential
of soils. West Conshohocken, PA: ASTM International.
Arvidsson, J. & Keller, T. (2004). Soil precompression stress: I. A survey of Swedish
arable soils. Soil and Tillage Research, 77(1): 85–95.
Aversa, S. & Nicotera, M. V. (2002). A triaxial and oedometer apparatus for testing
unsaturated soils. ASTM Compass, 25(1): 13.
BS 1377-­5 (1990). Methods of test for soil for civil engineering purposes – part 5:
Compressibility, permeability and durability tests. UK: BSI.
BS EN ISO 17892-­5 (2017). Geotechnical investigation and testing. Laboratory test-
ing of soil. part 5: incremental loading oedometer test. UK: BSI.
Burton, G. J., Sheng, D., & Airey, D. (2014). Experimental study on volumetric behav-
iour of Maryland clay and the role of degree of saturation. Canadian Geotechnical
Journal, 51(12): 1449–1455.
Cuisinier, O. & Masrouri, F. (2005). Hydromechanical behaviour of a compacted
swelling soil over a wide suction range. Engineering Geology, 81(3): 204–212.
Delage, P., Suraj De Silva, G. P. R., & Vicol, T. (1992). Suction controlled test-
ing of non saturated soils with an osmotic consolidometer. In 7th International
Conference Expansive Soils, Dallas, pp. 206–211.
Futai, M. M., De Almeida, M. S. S., Conciani, W., & Silva Filho, F. C. (2002).
Experimental and theoretical evaluation of plate load test in collapsible soil. In
Proceedings 3rd International Conference on Unsaturated Soils (UNSAT 2002).
Swets and Zeitlinger, Lisse, Recife, Brazil, pp. 721–726.
Head, K. H. (2011). Manual of soil laboratory testing volume 2: Permeability. Shear
Strength and Compressibility Test.Scotland, UK: Dunbeath Mill.
Holtz, R. D. & Kovacs, W. D. (1981). An introduction to geotechnical engineering.
Englewood Cliffs: Prentice–Hall.
Holtz, R. D., Kovacs, W. D., & Sheahan, T. C. (2011). An introduction to geotechni-
cal engineering, 2nd ed. Upper Saddle River, NJ: Pearson Education, Inc.
Honda, M. (2000). Research on the prediction methods for the mechanical behav-
iours of unsaturated soils. Kobe, Japan: Kobe University,vol. Ph.D.
Hoyos, L., Pérez-­Ruiz, D., & Puppala, A. (2012). Modeling unsaturated soil response
under suction-­ controlled true triaxial stress paths. International Journal of
Geomechanics, 12(3): 292–308.
Imhoff, S., Da Silva, A. P., & Fallow, D. (2004). Susceptibility to compaction, load
support capacity, and soil compressibility of Hapludox. Soil Science Society of
America Journal, 68(1): 17–24.
Jennings, J. E. & Knight, K. 1957. The additional settlement of foundations due to
a collapse of structure of sandy subsoils on wetting. Proceeding of the Fourth
International Conference on Soil Mechanics and Foundation Engineering, Vol. 1,
pp. 316–319.
Jotisankasa, A. (2005). Collapse behaviour of a compacted silty clay. In Civil and
environmental engineering. London: Imperial College, vol. Ph.D, p. 422.
 Oedometer test 261

Jotisankasa, A., Ridley, A., & Coop, M. (2007). Collapse behavior of compacted
silty clay in suction-­monitored oedometer apparatus. Journal of Geotechnical and
Geoenvironmental Engineering, 133(7): 867–877.
Kayadelen, C. (2008). The consolidation characteristics of an unsaturated com-
pacted soil. Environmental Geology, 54: 325–334.
Larson, W. E., Gupta, S. C., & Useche, R. A. (1980). Compression of agricultural soils
from eight soil orders1. Soil Science Society of America Journal, 44(3): 450–457.
Leong, E. C., Widiastuti, S., & Rahardjo, H. (2013). Estimating wetting-­induced
settlement of compacted soils using oedometer test. Geotechnical Engineering
Journal of the SEAGS & AGSSEA, 44(1): 26–33.
Lloret, A., Villar, M. V., Sa´Nchez, M., Gens, A., Pintado, X., & Alonso, E. E. (2003).
Mechanical behaviour of heavily compacted bentonite under high suction changes.
Géotechnique, 53(1): 27–40.
Mun, W. & McCartney, J. S. (2015). Compression mechanisms of unsaturated clay
under high stresses. Canadian Geotechnical Journal, 52(12): 2099–2112.
Munoz, J., Castelblanco Delage, P., Pereira, J. M., & Cui, Y. J. (2011). Some aspect
of the compression and collapse behaviour of an unsaturated natural loess.
Geotechnique Letters, 1: 17–22.
O’Sullivan, M. F. (1992). Uniaxial compaction effects on soil physical properties in
relation to soil type and cultivation. Soil and Tillage Research, 24(3): 257–269.
Pham, H. Q. (2005). A volume-­mass constitutive model for unsaturated soils. In
Civil and Geological Engineering. Saskatoon: University of Saskatchewan, vol.
Doctor of Philosopy, p. 532.
Rahardjo, H. (1990). The study of undrained and drained behaviour of unsaturated
soil. In Department of Civil Engineering. Saskatoon: University of Saskatchewan,
vol. Ph.D.
Salager, S., Rizzi, M., & Laloui, L. 2011. An innovative device for determining the
soil water retention curve under high suction at different temperatures. Acta
Geotechnica, 6(3): 135–142. doi: 10.1007/s11440-­011-­0141-­8
Silva, F. C. F., Almeida, M. S. S., & Futai, M. M. (2002) Critical state modelling
of unstaturated soils. In Proceeding of the Third International Conference on
Unsaturated Soil, Recife, Brazil, pp. 133–137.
Sivakumar, V., Tan, W. C., Murray, E. J., & McKinley, J. D. (2006). Wetting, drying
and compression characteristics of compacted clay. Géotechnique, 56: 57–62.
Sivakumar, V. & Wheeler, S. J. (2000) Influence of compaction procedure on the
mechanical behaviour of an unsaturated compacted clay. Part 1: Wetting and
istropic compression. Géotechnique, 50(4): 359–368.
Smith, C. W., Johnston, M. A., & Lorentz, S. (1997). Assessing the compaction sus-
ceptibility of South African forestry soils. II. Soil properties affecting compactibil-
ity and compressibility. Soil and Tillage Research, 43(3–4): 335–354.
Tarantino, A. & De Col, E. (2008). Compaction behaviour of clay. Géotechnique,
58(3): 199–213.
Toll, D. G. (1990). A framework for unsaturated soil behaviour. Géotechnique,
40(1): 31–44.
Vázquez, M., Justo, J. L. & Durand, P. (2013). A simplified model for collapse using
suction controlled tests. In Proceedings of the 18th International Conference on
Soil Mechanics and Geotechnical Engineering, Paris, pp. 1203–1206.
Wheeler, S. J. & Sivakumar, V. (1995). An elasto-­plastic critical state framework for
unsaturated soil. Géotechnique, 45(1): 35–53.
262 Laboratory tests for unsaturated soils

Wijaya, M. (2017). Compression, shrinkage and wetting-­induced volume change

of unsaturated soils. In Civil and environmental engineering (Ph.D.) Singapore:
Nanyang Technological University.
Wijaya, M. & Leong, E. C. (2015). Swelling and collapse of unsaturated soils due
to inundation under one-­dimensional loading. Indian Geotechnical Journal, 46:
239–251.
Wijaya, M. & Leong, E. C. (2016). Performance of high-­capacity tensiometer in
constant water content oedometer test. International Journal of Geo-­Engineering,
7(1): 13.

FURTHER READING

Jotisankasa, A., Ridley, A., & Coop, M. (2007). Collapse ehavior of compacted silty
clay in suction-­monitored oedometer apparatus. Journal of Geotechnical and
Geoenvironmental Engineering, 133(7): 867–877.
Wijaya, M. & Leong, E. C. (2015). Swelling and collapse of unsaturated soils due to
inundation under one-­dimensional loading. Indian Geotechnical Journal: 1–13.
Leong, E. C., Widiastuti, S., & Rahardjo, H. (2013). Estimating wetting-­induced
settlement of compacted soils using oedometer test. Geotechnical Engineering
Journal of the SEAGS & AGSSEA, 44(1): 26–33.
Chapter 14

Constant rate of strain test

14.1 BACKGROUND

Soil deformation is rate dependent due to phenomena such as stress relax-

ation and creep (Bagheri et al. 2019b). Thus, the rate dependency of soil
deformation starts to gain attention for both saturated and unsaturated soils
(Kim and Leroueil 2001; Yin and Hicher 2008; Sorensen et al. 2010; Tong
and Yin 2013; Yin et al. 2014; Rezania et al. 2017a; Rezania et al. 2017b;
Bagheri et al. 2019b; Tanaka et al. 2014). One method to investigate the
effect of strain rate on soil is by employing the constant rate of strain (CRS)
test. In the CRS test, instead of applying an incremental load to the speci-
men, a constant rate of strain is applied. As the test can be completed faster
than the oedometer test, it also serves as an alternative test method.
The challenge in the CRS test for saturated soils is to ensure that the
excess pore-­water pressure generated throughout the test is kept below a
threshold value. Unlike the CRS test for saturated soils, the CRS test for
unsaturated soils may not have this constraint depending on the initial
degree of saturation of the soil. Many developments have been carried out
on the use of the CRS test on unsaturated soil (with or without pore-­water
pressure measurement). In this chapter, the CRS test for unsaturated soils
will be elaborated.

14.2 RELATED STANDARD

1. ASTM D4186/D4186m-­12 (2012) Standard Test Method for One-­

Dimensional Consolidation Properties of Saturated Cohesive Soils
Using Controlled-­Strain Loading, ASTM International.

DOI: 10.1201/b22304-14 263

264 Laboratory tests for unsaturated soils

14.3 THEORY

14.3.1 Saturated CRS test (ASTM D4186/D4186M-12 2012)

The CRS test is an alternative to the oedometer test for testing saturated
soils. Instead of applying an incremental loading and waiting for the excess
pore-­water pressure to dissipate, a CRS is imposed on the specimen such
that the excess pore-­water pressure does not exceed a certain threshold.
There are three obvious advantages of the CRS test compared to the
oedometer test in testing saturated soils:

1. It provides a continuous compression curve and, hence, a continuous

relationship between the void ratio (e) and permeability (kw), coef-
ficient of consolidation (cv) instead of a discrete relationship based on
the incremental loadings.
2. It is faster than incremental loading.
3. When a confined cell with a pressure system is used, it is possible to
saturate the specimen by using a back pressure, and the saturation of
the specimen can be checked by using the Skempton pore-­pressure
parameter B just like in the triaxial test.

The problem with the CRS test for saturated soil is that it generates excess
pore-­water pressure throughout the test. It is recommended that the pore-­
water ratio (Ru) is between 3% and 15%. Therefore, in order to calculate
the applied effective stress, the pore-­water pressure at the bottom of the soil
specimen is monitored throughout the test.
The CRS apparatus is very similar to the oedometer apparatus except that
the loading is strain controlled, and hence a compression machine or load
actuator similar to that in a triaxial test is used. Some of the CRS appara-
tuses, such as those manufactured by GDS Instruments, UK, have the ability
to do stress control as well and thus can replicate the oedometer test.
Figure 14.1 shows a schematic of a CRS apparatus. As the soil specimen
is placed inside the CRS cell, it is possible to saturate the specimen by filling
the CRS cell with water (instead of inundating the specimen like in the
oedometer test) and applying a back pressure (cell pressure) to the specimen.
The saturation of the specimen can be checked by increasing the cell pres-
sure while observing the change in the pore-­water pressure and calculating
the Skempton B value, which is defined in Equation 14.1.

∆uw
B= (14.1)
∆σ c

where Δuw is the change in the pore-­water pressure (usually measured at the
base), and Δσc is the change in cell pressure.
Table 14.1 shows the equations to calculate the parameters in the CRS
test. Some of the derived soil properties are based on the model used, either
linear or non-­linear. The linear model assumes that the soil has a constant
 Constant rate of strain test
Figure 14.1 Standard CRS apparatus (from Wijaya 2017).

265
266 Laboratory tests for unsaturated soils

Table 14.1 List of equations for CRS test (ASTM D4186/D4186M-12 2012)

Axial deformation, Δdn Parameters

∆d n = d n − d 0 dn = LVTD reading
d0 = initial LVDT reading
n = current data point index
base pressure zero reading, ub,0
ub ,0 = ub , s − σ c , s ub,s = base pressure at the end of the
saturation stage
σc,s = cell pressure at the end of the
saturation stage
Corrected base pressure, um,n
um , n = u b , n − u b , 0 ub,n = base pressure reading
Base excess pressure, Δum,n
∆um, n = um, n − σ c , n
Axial force
f n = Qn − Q0 Qn = axial force reading
Q0 = initial force reading
Net axial force, fa,n
f a , n = f n + ML g − ∆f s + Wp − Apσ c , n ML = mass of loading cap
g = gravity acceleration
Δfs = dynamic piston friction from piston
seal dynamic friction calibration
Wp = effective piston weight from piston
uplift calibration
Ap = effective piston area from piston uplift
calibration
σc,n = cell pressure reading
Change in specimen height, ΔHn
∆Hn = ∆d n − δ af , n − σ ap, n δaf,n = apparatus compressibility due to axial
load
δap,n = apparatus compressibility due to cell
pressure
Axial strain, εa,n
∆Hn H0 = initial height of the specimen
ε a,n = 100
H0
Strain rate, ε
∆Hn +1 − ∆Hn −1 1 t = time
ε = ⋅
H0 tn +1 − tn −1
Excess base pressure Δum,n
∆um,n = um,n − σ c ,n
Axial stress, σa,n
f A = specimen area
σ a ,n = a ,n
A

(Continued)
 Constant rate of strain test 267

Table 14.1 (Continued)

Axial deformation, Δdn Parameters

Steady-­state factor, Fn
Linear equation: l = index of data at the beginning of loading

Fn =
(σ a,n − σ a, l ) − ( ∆um,n − ∆um, l ) or unloading stage
(σ a,n − σ a, l )
Non-­linear equation:

Fn =
{log σ a,n }
− ( ∆um, n − ∆um, l )  − log σ a , l
log σ a , n − log σ a , l

Effective axial stress, σ’a,n

Linear equation:
 2 
σ ′a , n =  σ a , n − ∆um, n 
 3 
Non-­linear equation:
( )
1/ 3
σ a , n = σ a3, n − 2σ a2, n ⋅ ∆um, n + σ a , n ∆um2 , n
Permeability, kn
Linear equation:
εHnH0γ w
kn =
2∆um,n
Non-­linear equation:
0.434εH0 Hnγ w
kn = −
 ∆u 
2σ ′a ,n log 1 − m,n 
 σ a ,n 
Coefficient of compressibility mv,n
ε n+1 − ε n−1
mv ,n =
σ ′a ,n+1 − σ a ,n−1
Coefficient of consolidation, cv,n
Linear equation:
kn
c v ,n =
mv ,nγ w
Non-­linear equation:
σ 
H0 Hn log  a , n +1 
σ
 a , n −1 
cv ,n =
 ∆u 
2 ( tn +1 − tn −1 ) log 1 − m, n 
 σ a,n 
Pore-­pressure ratio, Ru,n
∆um, n
Ru, n =
σ a,n
268 Laboratory tests for unsaturated soils

mv, while the non-­linear model assumes that the soil has a constant com-
pressibility index (Cc) (ASTM D4186/D4186M-­ 12 2012). The selection
choice of linear or non-­linear model depends on the specimen behaviour.

14.3.2 Unsaturated CRS test

Similar to the oedometer test, the CRS test can be performed on unsatu-
rated soil as well through either a constant water content test or a constant
suction test. For the constant water content CRS (CWC-­CRS) test, the test
can be performed either with or without pore-­water pressure measurement
(Wijaya 2017). In the CWC-­CRS test with pore-­water pressure measure-
ment (Figure 14.2), the pore-­water measurement is carried out at the bot-
tom of the specimen (Wijaya 2017; Wijaya et al. 2014) or can be placed at
the perimeter of the specimen (Bagheri et al. 2019b). For the constant suc-
tion CRS test, the modification can be made similar to the triaxial apparatus
where the top cap is connected to the air pressure while the bottom porous
disk is replaced with a ceramic disk. Additionally, a high-­capacity tensiom-
eter can be used to measure the soil suction at the perimeter of the specimen
(Bagheri et al. 2019b).
The compression of unsaturated soil under the constant water content
condition is almost instantaneous, as negligible excess pore-­water pressure
is generated during loading (Wijaya 2017). For a soil specimen that has a
high initial degree of saturation, the soil specimen may become saturated
during the test. Hence, the selected strain rate for such a specimen needs to
follow the strain rate recommended for the CRS test for saturated soil such
that the pore-­water ratio (Ru) is between 3% and 15% and the steady-­state
factor (Fn) is bigger than 0.4 as recommended by ASTM D4186/D4186M-­12
2012. It is therefore recommended that the CRS test be conducted for the
saturated soil before conducting the CWC-­CRS test to determine the appro-
priate strain rate.
The CWC-­CRS test has an inundation stage where the soil specimen is
flooded with water. Wijaya and Leong (2015) and Wijaya (2017) have
shown that there is a transition pressure (PT) in the unsaturated compres-
sion curve which causes the inundation stage and produces different phe-
nomena due to the transition pressure which is located at the intersection
between the saturated and unsaturated compression curves, as illustrated in
Figure 14.3.
If the inundation stage is carried out at Pi (which is referred to as inunda-
tion pressure), when Pi is below PT, the soil will swell and follow the swelling
path (SP) when the soil volume is not constrained. As the soil is constrained
in the CRS apparatus, it will follow the constant volume path (CVP), and
the pressure will increase until the soil reaches the saturated compression
curve (Figure 14.3a). The increase in pressure due to inundation under the
constant volume condition is referred to as swelling pressure. Determining
swelling pressure in the CRS apparatus is more convenient than using the
oedometer apparatus, as it is effortless to maintain the constant volume
 Constant rate of strain test
Figure 14.2 M odified CRS apparatus with pore-water pressure measurement (from Wijaya 2017).

269
270 Laboratory tests for unsaturated soils

Figure 14.3 W
 etting-induced behaviour of unsaturated soil (from Wijaya and
Leong 2015). (a) P i < P c,s . (b) P c,s < P i < P c,d . (c) P i = P c,d .
 Constant rate of strain test 271

condition in the CRS apparatus. An example of an inundation stage which

generates swelling pressure is shown in Figure 14.4.
However, if Pi is bigger than Pt, the soil will collapse. In the CRS appara-
tus, as the soil is constrained, and provided that the collapse is not signifi-
cant enough to cause the load cap to no longer be in contact with the load
cell, the soil state will move from the unsaturated compression curve to the
saturated compression curve by following the CVP and thus reducing the net
normal stress. The reduction in net normal stress associated with the collapse
phenomenon is referred to as pressure relaxation. An example of an inun-
dation stage which causes pressure relaxation is shown in Figure 14.5.

Figure 14.4 S welling pressure due to inundation under transition pressure (from
Wijaya and Leong 2015).

Figure 14.5 P
 ressure relaxation due to inundation above the transition pressure
(from Wijaya and Leong 2015).
272 Laboratory tests for unsaturated soils

The normal compression line (NCL) determined in the two figures is

obtained from a saturated oedometer test. It is shown that post inundation,
the soil will follow the saturated path, and there is an agreement between
the CRS test and the oedometer test results.

14.4 TEST METHODS

14.4.1 CRS apparatus calibration stages

For the proper conduct of the CRS test, the CRS apparatus requires calibra-
tion for

1. apparatus compressibility,
2. chamber pressure,
3. piston uplift, and
4. piston seal dynamic friction.

ASTM D4186/D4186M-­12 (2012) requires that the correction be applied

whenever the apparatus compressibility or chamber pressure may cause
deformation of 0.1% specimen height. If the piston seal dynamic friction
correction is up to 0.5% of the maximum applied load, the correction must
be applied. However, it is recommended to carry out all the calibrations.

14.4.1.1 Calibration for apparatus compressibility

The contribution of the compressibility of the apparatus to the deformation
reading when applying load to the specimen must be considered. The proce-
dure for determining the apparatus compressibility is as follows:

1. Place a calibration disk (steel) of approximately the same size as the

specimen together with the filter elements under the loading piston.
2. Load following the same loading path as the actual test and record
both load and deformation readings.
3. Plot the load-­deformation curve which shall be used to correct the load-­
deformation curves of the soil specimen tested using the apparatus.

14.4.1.2 Calibration for chamber pressure

The applied cell pressure in the chamber may affect the deformation reading
in the CRS test and has to be accounted for in the test results. The procedure
for chamber pressure calibration is as follows:

1. Place a calibration disk (steel) of approximately the same size as the

specimen together with the filter elements.
2. Apply the seating net axial force (Fa,0) with zero cell pressure and
record the initial deformation reading d0.
3. Increase the cell pressure to σ3,i.
 Constant rate of strain test 273

4. Adjust the axial force back to Fa,0.

5. Record the axial deformation di and the cell pressure σ3,i.
6. Repeat steps 2 to 5 for increasing cell pressures.
7. Establish a relationship between cell pressure and apparatus deformation.
This correction is subtracted from all deformation readings in a CRS test.

14.4.1.3 Calibration for piston uplift

Some CRS apparatus designs may cause an uplift of the piston due to the
cell pressure and thus affect the force reading. Therefore, correction to the
force reading due to the piston uplift might need to be carried out together
with the calibration for the pressure chamber. The procedure for the piston
uplift calibration is as follows:

1. Assemble the apparatus without a specimen.

2. Increase the cell pressure from zero to the maximum cell pressure.
3. Record the applied cell pressure and the force reading.
4. Create a plot between cell pressure and force reading.
5. The effective area of the piston (Ap) is given as the slope of the line
between cell pressure and force reading and the effective piston weight
(Wp) as the intercept of the line between cell pressure and force reading.

14.4.1.4 Calibration for piston seal dynamic friction

There is friction in the piston seal where the piston enters the CRS chamber,
and this affects the axial force reading. Therefore, correction to the force
reading needs to be carried out. The procedure for the calibration of piston
seal dynamic friction is as follows:

1. Assemble the apparatus without a specimen.

2. Apply and record the targeted cell pressure σ3.
3. Advance the piston at the rate used in the CRS test.
4. Compute the increment in axial force as the difference between the
measured axial force and the uplift force.
5. Take the average of the increment in axial force as the dynamic seal
friction force Δfs.
6. Steps 2 to 5 are repeated for other cell pressures if required.

14.4.2 Specimen preparation
Specimen preparation for the CRS test is similar to the specimen prepa-
ration for the oedometer test. The specimen preparation method (ASTM
D4186/D4186M-­12 (2012)) is summarised as follows:

1. Record the weight (Mring), inside diameter (D) and height (H0) of the
specimen ring. The diameter and height of the specimen can be taken
as the inside diameter and height of the specimen ring, respectively.
274 Laboratory tests for unsaturated soils

2. The sample is first extruded into a short sample tube which is 2.5 mm
larger than the CRS specimen diameter.
3. The sample is then extruded from the sampling tube. The extruding
direction shall be the same as the direction that the soil enters the
tube. If possible, the sample should be extruded vertically; otherwise,
precaution must be taken to ensure that bending stress is not imposed
on the sample due to gravity.
4. Either of the following can be done to reduce the specimen diameter
into the specimen ring diameter:
a) Trimming the specimen by using a trimming turntable until the
specimen can fit into the specimen ring which has been coated
with low-­friction material (silicone/vacuum grease).
b) Push in the specimen ring to the specimen. The specimen is trimmed
to a gentle taper while the specimen ring is pushed by using an
alignment guide until the specimen protrudes at the top of the ring.
5. Trim the top and bottom of the specimen to be flush with the specimen
ring by using the following:
a) Soft to medium soil: wire saw.
b) Stiff soil: Straightedge with a sharpened cutting surface.
6. Any voids caused by trimming should be patched with the soil from
the trimmings and noted in the test.
7. If the specimen height is below the height of the specimen ring, place
the filter paper (or any filtering element) on top of the soil and do
either of the following:
a) Take at least four evenly spaced measurements of the speci-
men height by using a dial comparator or other measuring de-
vice which can minimise penetration into the soil during the
measurement.
b) Take the height of the confinement ring minus the recess spacer
and the filtering element.
8. Clean the outside of the specimen ring and record the weight of both
the specimen ring and the specimen (Mring+soil) and determine the mass
of soil (Msoil) as follows:
(14.2)
Ms = M ring+soil − Mring

9. Determine the initial water content (w0) of the specimen using the
trimmings.

14.4.3 Saturated CRS test

The saturated CRS test procedure is to follow ASTM D4186/D4186M-­12
(2012). The stages in the saturated CRS test can be divided as follows:
 Constant rate of strain test 275

1. Calibration and set up

2. Saturation
3. Loading
4. Unloading
5. Constant load

14.4.3.1 End calibration and set-up stage

The calibration stage for the CRS test involves the calibrations for appara-
tus compressibility:

1. Chamber pressure, piston uplift, and piston seal dynamic friction cali-
brations as detailed in ASTM D4186/D4186M-­12 (2012). Once the
calibration is completed, record the weight of the load cap and top
porous stone.
2. Ensure that all tubes/systems which are used to apply cell pressure,
back pressure and drainage system are completely filled with water.
Record the zero readings of the cell pressure and base pressure, σ0 and
uw,0, respectively.
Place the porous stone as follows:
a. For non-­expansive saturated soil, use a saturated porous stone and
place it into the CRS apparatus with the porous stone slot/base
filled with water. Use a paper towel to remove the excess water.
b. For non-­expansive unsaturated soil, use a saturated porous stone
and place it into the CRS apparatus with the porous stone slot/
base under dry condition. Use a paper towel to remove any excess
water from the porous stone.
c. For expansive soil, use an air-­dried porous stone and place it into
the CRS apparatus with the porous stone slot/base under dry
conditions.
For all of the above procedures, the connection lines/tubes must
be filled with water.
3. Prepare the specimen inside the specimen ring and place it into the
CRS apparatus with filter elements at the top and bottom of the speci-
men (to avoid intrusion of material into the pores of the porous disk).
4. Record the zero-­reading force F0.
5. Adjust the axial loading device such that the load cap of the specimen
is in contact with the piston or load cell.
6. Set the deformation measuring device and record the zero-­reading d0.
7. Applies a small seating pressure. The seating pressure must be
enough to prevent swelling but not cause significant consolidation.
Recommended seating pressure is about 10% of the in situ effective
stress or 0.2% axial strain.
8. Open the valve to fill the CRS cell with water.
276 Laboratory tests for unsaturated soils

14.4.3.2 Saturation stage
The specimen can be saturated by using back pressure as follows:

1. Apply back pressure (cell pressure) to saturate the specimen while

maintaining a constant seating pressure or constant specimen height.
The back pressure is typically from 400 kPa to 1,000 kPa. There are
two methods for applying back pressure:

a. Stepwise increment. A starting increment is typically equal to the

seating pressure and then doubled (typically in the range of 35
kPa to 140 kPa). A typical first-­time increment is 10 minutes.
b. Constant rate increment. Back pressure is progressively increased
up to the maximum value in several hours.
If constant seating pressure or constant specimen height can-
not be maintained, the back pressure increment or rate must be
reduced.
2. Check the saturation by applying an increment of cell pressure. If
the pore-­water pressure at the base increases almost instantaneously
as the increase in cell pressure, the saturation can be considered
successful.
3. Record the post-­ saturation seating force, cell pressure, pore-­
water
pressure at the base, axial deformation and the time prior to the load-
ing stage.
4. Close the drainage valve and check the pore-­water pressure at the
base. Ensure that the pore-­water pressure at the base remains constant
else there is a leak at the base.

14.4.3.3 Loading stage
Select a suitable strain rate such that the pore-­water ratio Ru is between
3% and 15%. The appropriate strain rate can be achieved by increasing the
strain rate and monitoring Ru until it reaches approximately 5% and then
maintaining the constant strain rate. As the equation shown in Table 14.1
is based on Fn > 0.4, it is also important to select a strain rate such that Fn
is bigger than 0.4; otherwise, the parameters calculated in Table 14.1 are
approximate. Some recommended strain rates are as follows:

1. MH material: 10%/h
2. CL material: 1%/h
3. CH material: 0.1%/h

Once an appropriate strain rate for the typical specimen has been obtained,
a nearly constant strain rate (the variation is not more than a factor of 5)
throughout the test must be used.
 Constant rate of strain test 277

14.4.3.4 Unloading stage
For the unloading stage, use half of the strain rate used for the loading stage.
During the unloading stage, it is possible for the pore-­water pressure at the
base to become negative. Thus, cell pressure must be high enough such that
the pore-­water pressure at the base is greater than the saturation pressure.

14.4.3.5 Constant load stage

In the constant load stage, the stress is maintained until the excess pore-­
water pressure dissipates. Record the time, the elapsed duration, cell pres-
sure, base pressure, axial load and axial displacement reading. The constant
load stage is maintained for a specified time or until the base excess pressure
is nearly 0 (about 1% of the total stress).

14.4.3.6 End-stage
1. At the end of the test, open the bottom drainage valve and decrease
the cell pressure to atmospheric pressure.
2. Unload and record the final zero readings for the axial force, cell pres-
sure and base pressure.
3. Drain the water from the cell and remove the specimen ring from the
CRS apparatus.
4. Extrude the specimen and determine the final water content of the soil.

14.4.4 CWC-CRS test
The CWC-­CRS test procedure developed by Wijaya (2017) can be divided
into the following stages:

1. Suction initialisation
2. Calibration and set up
3. Loading
4. Unloading
5. Inundation
6. End

14.4.4.1 Suction initialisation stage

The CWC-­CRS test can be done on a specimen with natural water content
(and, hence, soil suction is measured) or soil suction can be predetermined
either by using suction control or water content control. Water content
control can be carried out by air drying the specimen to the desired water
content and soil suction is measured (Wijaya 2017; Bagheri et al. 2019a;
Bagheri et al. 2019b).