Ultrasonics 41 (2003) 539–542

www.elsevier.com/locate/ultras

Effect of ultrasound on oil removal from soils

a,* b
Y.U. Kim , M.C. Wang
a
Department of Civil and Environmental Engineering, Myongji University, San 38-2, Nam-Dong Yongin, Gyunggi-do 449-728, South Korea
b
Department of Civil and Environmental Engineering, Pennsylvania State University, University Park, PA 16802, USA
Received 11 April 2003; accepted 21 June 2003

Abstract

The soil-flushing method enhanced by ultrasonic waves is a new technique that potentially can become an effective method for
in situ remediation of the ground contaminated by NAPL hydrocarbons. This study investigated the effectiveness of ultrasound
enhancement in the soil-flushing method for a range of conditions involving soil type, soil density, flushing rate, and sonication
power. The study was conducted in the laboratory using specially designed and fabricated equipment. The test results indicated that
the rate of contaminant extraction increased considerably with increasing sonication power up to the level where cavitation oc-
curred. The effectiveness of sonication-enhanced soil-flushing can be expressed as a function of ðD10 Þ2  i, in which D10 is the effective
grain size, and i is the hydraulic gradient.
Ó 2003 Published by Elsevier B.V.

Keywords: Ultrasound; Ground contamination; Soil-flushing method; In situ remediation; Laboratory testing; NAPL hydrocarbons; Stress waves

1. Introduction 2. Current state of knowledge

Petroleum hydrocarbons are commonly found in the
ground in urban and suburban areas due to possible
leakage of gasoline, motor oils, and diesel fuel from
underground storage tanks. The polluted ground needs
to be cleaned in order to avoid potential hydrocarbon
contamination of ground water aquifer. Currently, there
are different remediation methods, e.g. replacement,
pump-and-treat, vapor extraction, and flushing meth-
ods. However, a method that can be effective and also
economical for a broad range of field conditions is not
yet available.
For development of an effective ground remediation
method, there has been considerable research focusing
on the technique of enhancing soil-flushing method.
Although there are data showing that ultrasonic waves
are capable of removing non-aqueous phase liquid
(NAPL) hydrocarbons from soils, a methodology for
evaluating the effectiveness of sonication is yet to be
developed. It is with this objective in mind that this
study was undertaken.
*
Corresponding author. Tel.: +82-31-330-6417; fax: +82-31-336-
9705.
E-mail address: younguk@mju.ac.kr (Y.U. Kim).
Available information about sonication effects on
extraction of NAPL hydrocarbon from porous media is
limited and piecemeal. Surguchev [8] and Simkin [7]
attributed the increased extraction of oil (hydrocarbons)
to a decrease for water and an increase for oil in the
relative phase permeability due to stress waves. For soils
with low permeability, Cleveland and Garg [1] reported
that ultrasonic excitation can suspend fine particles to
which the contaminants are strongly sorbed. The fine
particles subsequently can then be removed by flushing
water through the soil. Also, Reddi and Challa [5], and
Reddi and Wu [6] presented that ultrasonic waves can
increase not only the mobility of NAPL ganglia but the
porosity of the soil as well, resulting in a decrease in
viscosity and buoyant pressure.
Only a few studies are available on the effectiveness of
using acoustic waves to enhance contaminant extraction
in soil-flushing method. Ellen et al. [2] reported a 30%
increase in contaminant extraction due to acoustic ex-
citation. They hypothesized that acoustic waves can
overcome the capillary force on contaminants in a soil
by alternating over- and under-pressures which produce
pulling and pushing actions to contaminant droplets.
Thus, large contaminant droplets can be broken into
smaller droplets. These smaller droplets can be flushed

0041-624X/$ - see front matter Ó 2003 Published by Elsevier B.V.

doi:10.1016/S0041-624X(03)00168-9
540 Y.U. Kim, M.C. Wang / Ultrasonics 41 (2003) 539–542
out more easily. Another study by Iovenitti et al. [4]
reported a 6–26% improvement in contaminant extrac-
tion.
The available information reviewed above shows that
ultrasonic waves could increase percolation rate and
facilitate removal of entrapped contaminants. The de-
gree of enhancement varies with many factors, such as
soil type, soil density, flow rate, temperature, wave fre-
quency, energy level, and others. Since nearly every
available study focused only on limited specific condi-
tions without a systematic investigation for a broad
spectrum of the various influencing factors, a method-
ology is not yet available for evaluating the effectiveness
of ultrasonic waves. Such a methodology is essential in
the practical application of using stress waves to en-
hance the effectiveness of soil-flushing.
3. Laboratory investigation
3.1. Test equipment
The laboratory testing involved one-dimensional soil-
flushing tests and soil characteristics determinations.
The soil-flushing tests were conducted using specially
designed and fabricated test equipment which is shown
schematically in Fig. 1. As shown, the test setup was
composed of two parts––an ultrasonic processor and a
test chamber, which was connected to a de-aired water
reservoir.
The test chamber was made of a Plexiglas cylinder
having an inside diameter of 7.3 cm with a height of 30
cm. The bottom cap was provided with an inlet con-
necting to a reservoir of de-aired water, which was used
Generator
Converter
Water +
Contaminant Acoustic
Outlet Horn
Contaminated
Soil
Water Inlet
Fig. 1. Schematic of laboratory test setup.
to flush contaminants out of the soil sample. The de-
aired water reservoir was connected to a water tap.
The sonication device was a 500 W-Ultrasonic Pro-
cessor manufactured by Heat Systems Ultrasonics. The
Processor was calibrated by the manufacturer. The en-
tire apparatus was composed of a generator, a con-
verter, an acoustic horn, and a flat tip as shown in Fig.
1. The generator (or power supply) converted the con-
ventional 60 Hz AC at 120 V to a 20 kHz electrical
energy at approximately 1000 V. The high-frequency
electrical energy was fed to the converter to transform
the energy to mechanical vibration. The vibrator was
tuned to vibrate at 20 kHz. The acoustic horn and flat
tip amplified the longitudinal vibration of the converter.
3.2. Test soils and contaminant
The test soils were commercially available Ottawa
sand, a fine aggregate, and a natural soil. Their respec-
tive largest particle sizes were 0.8, 4.75, and 4.75 mm;
uniformity coefficients (Cu ) were 1.85, 3.21, and 12.9;
effective grain sizes (D10 ) were 0.27, 0.54, and 0.25 mm;
median grain sizes (D50 ) were 0.45, 1.40, and 1.10 mm;
and Unified soil classifications were SP, SP, and SW.
The test soil specimens made of Ottawa sand had two
void ratios (e) of 0.55 and 0.67; their corresponding
water contents were 20.8% and 25.3%. The test speci-
mens of both fine aggregate and natural soil had a void
ratio of 0.67 with a water content equal to 24.9%.
The test surrogate contaminant was Crisco Pure
Vegetable Oil. The density of Crisco Pure Vegetable Oil
was 0.9182 g/ml at 20 °C, and the viscosity was 67 cP at
20 °C. These characteristics were provided by the Crisco
Company.
3.3. Test procedures
The soil-flushing tests were conducted for two con-
ditions––without ultrasound and with ultrasonic waves
at 20 kHz frequency. The test specimens were prepared
from the mixture of 332.1 g of air-dried test soils and 60
ml of Crisco Pure Vegetable Oil. Each test specimen had
a diameter of 7.3 cm and a height of 5.0 cm. Both Ot-
Overflow tawa sand and fine aggregate specimens were tested
under three levels of hydraulic gradients (i) equal to 1.6,
5.5, and 12.9; whereas, the natural soil specimens were
tested only for 5.5 and 12.9 hydraulic gradients. The
ultrasonic energy levels (W) investigated were 50, 100,
and 140 W for Ottawa sand, and only 100 W for both
fine aggregate and natural soil specimens. The tests were
performed in the Geotechnical Engineering Laboratory
where the room temperature was between 19 and 22 °C.
The major steps involved in the experiments are as
follows. A pre-weighed soil was mixed thoroughly with
a desired amount of Crisco Pure Vegetable Oil. The
mixture of soil and oil was carefully placed into the test
 Y.U. Kim, M.C. Wang / Ultrasonics 41 (2003) 539–542
chamber. The soil specimen was then saturated with
water; the water level was maintained at the top of the
soil specimen. The soil specimen was then subjected to
ultrasonic waves at 20 kHz frequency. Under the action
of ultrasonic waves, the clean water was allowed to flow
upward through the soil specimen under a specified
hydraulic gradient. The effluent was collected in a 500
ml Polypropylene cylinder. The effluent in the cylinder
was allowed to stand overnight for gravitational segre-
gation of oil from water. The volumes of the separated
water and oil were then measured.
4. Influence factors
Numerous factors may influence the effect of soni-
cation on oil removal. Major factors investigated were
sonication power, hydraulic condition, and soil prop-
erty. The effect of sonication power on oil removal was
investigated only for Ottawa sand specimens that were
prepared at 0.67 and 0.55 void ratios and subjected to
1.60 and 5.50 hydraulic gradients. Fig. 2 presents the
graphical relations between percent contaminant re-
moval and sonication power for 5 PV (pore volume)
water flow volume. It is seen that, for a constant void
ratio and hydraulic gradient, the percent contaminant
removal increases with increasing sonication power to a
maximum around 100 W then decreases. The contami-
nant removal at 140 W is about equal to that around 75–
85 W, corresponding to a loss factor of about 1.8. The
drop in contaminant removal beyond about 100 W can
be attributed to the effect of cavitation. When cavitation
occurs, the sound pressure level at a distance drops,
50
45
Contaminant Removal (%)
40
35
30
e=0.55, i=1.60
e=0.67, i=1.60
25
e=0.67, i=5.5
20
0 20 40 60 80 100 120 140
Ultrasonic Po wer (W )
Fig. 2. Ultrasonic power (W) vs. contaminant removal (%).
541
because cavitation takes power away from the field.
Therefore, cavitation can reduce the effective sonication
power in the soil. Note that it may be a sudden drop
rather than a smooth curve shown. A similar cavitation
effect on oil flow through sandstone was also reported
by Fairbanks and Chen [3]. They observed cavitation at
100 W sonication power. The lower sonication power
for their case was probably because of their smaller test
samples which had a diameter of 7.3 cm and a height of
0.5–2.0 cm.
The influence of hydraulic condition was investigated
under 100 W sonication power using three levels of
hydraulic gradient––i ¼ 1:6, 5.5, and 13.0. The test re-
sults for Ottawa sand are shown in Fig. 3, which pre-
sents the percent contaminant removal as a function of
hydraulic gradient for both with and without sonication
conditions. It is seen that the percent contaminant re-
moval decreases with increasing hydraulic gradient. This
can be explained below with a note that increasing hy-
draulic gradient will increase discharge velocity and flow
rate, other factors being equal.
Before the contaminant can be flushed out of the soil,
the soil/contaminant bond must be broken first. Inside
the soil mass, the contaminant is either trapped within the
pore space formed by the interlocked soil particles or
adsorbed on the surface of individual particles or both.
Regardless of the nature of the bond, the breakdown of
the contaminant/soil bond is a time-dependent process.
For slower flushing, the percolating water has longer
time to interact with the soil/contaminant system. As a
result, the slow flushing water is more efficient to remove
the contaminant than the fast flushing as observed from
the experiment.
65
Ottawa Sand
e=0.67
60
e=0.55
e=0.67
Contaminant Removal (%)
e=0.55
55
With 100W Sonication
50
45
40
35
Without Sonication
30
160 0 2 4 6 8 10 12 14
Hydraulic Gradient
Fig. 3. Percent contaminant removal vs. hydraulic gradient.
542 Y.U. Kim, M.C. Wang / Ultrasonics 41 (2003) 539–542
60
50
Contaminant Removal (%)
40
30
Natural Soil (w/ sonication)
Ottawa Sand (w/ soncation)
20 Fine Aggregate (w/ sonication)
Natural Soil (w/o Sonication)
Ottawa Sand (w/o sonication)
Fine Aggregate (w/o sonication)
10 w/ sonication Best Fit Curve
w/o sonication Best Fit Curve
0
0 1 2 3
(D10)2*i
Fig. 4. Contaminant removal vs. discharge velocity.
Both Figs. 2 and 3 also show that the contaminant in
a soil with a higher void ratio can be removed more
easily than a soil with a lower void ratio. The difference
between the two cases appears to be greater for flow
without sonication as shown in Fig. 3. The test results
indicate, as would be expected, that under the same
hydraulic gradient, contaminants in loose soils can be
extracted more easily than in dense soils. The influence
of soil density on contaminant removal seems to be less
significant for the case with sonication than without
sonication.
The combined influence of soil type and hydraulic
condition on contaminant removal is presented in Fig. 4
which shows percent contaminant removal vs. ðD10 Þ2  i;
in which D10 is the effective grain size of test soil and i is
the hydraulic gradient. It appears that D10 is an effective
soil parameter for use to evaluate the influence of soil
type on contaminant removal at least for the range of
soil type and hydraulic conditions investigated.
Fig. 4 summarizes the test results for all three test
soils with different void ratios. It is interesting to note
that, regardless of soil type and void ratio, all data
points, although scattered, fall in two distinct groups––
one for the condition without and the other with soni-
cation. The vertical distance between the two trend lines
reflects the effectiveness of sonication in contaminant
removal. The figure shows that the effectiveness of so-
nication in contaminant removal is greater at lower
discharge velocity. This can possibly be attributed to the
relatively longer time for interaction between sound
wave and contaminant under slower flushing.
5. Summary and conclusions
This study investigated the effectiveness of sonication
in the soil-flushing method. The study was conducted in
the laboratory using a specially designed and fabricated
test device. The test soils were Ottawa sand, a fine ag-
gregate, and a natural soil; and the surrogate contami-
nant was Crisco Vegetable Oil. The test results indicated
that sonication can enhance pollutant removal consid-
erably, and that the degree of enhancement depends on
a number of factors such as sonication power, water
flow rate, and soil type. Increasing sonication power will
increase pollutant extraction only up to the level where
4 cavitation occurs. The effectiveness of sonication de-
creases with flushing rate but eventually becomes con-
stant under higher flow rates. Within the soil type, soil
density, and hydraulic gradient level investigated, the
effectiveness of sonication is highly related with
ðD10 Þ2  i.
References
[1] T.G. Cleveland, S. Garg, Field demonstration of acoustically
enhanced soil washing system for in-situ treatment of low-perme-
ability soils, in: Waste Management Proceedings of the Gulf Coast
Hazardous Substance Research Centers Symposium on Emerging
Technologies, Metals, Oxidation, and Separation, 13 (5–7), 1993,
pp. 519–520.
[2] T.V. Ellen, C.J.E. Lansink, J.E. Sandker, Acoustic soil remedia-
tion: introduction of a new concept, Contaminated Soil 95, 1995,
pp. 1193–1194.
[3] H.V. Fairbanks, W.I. Chen, Ultrasonic acceleration of liquid flow
through porous media, Chemical Engineering Progressive Sympo-
sium Series 67 (1971) 108–116.
[4] J.L. Iovenitti, T.M., Rynne, J.W. Spencer Jr., Acoustically
enhanced remediation of contaminated soil and ground water, in:
Proceedings of Opportunity 95-Environmental Technology
through Small Business, Morgantown, West Virginia, 1995.
[5] L.N. Reddi, S. Challa, Vibratory mobilization of immiscible liquid
ganglia in sands, Journal of Environmental Engineering, ASCE 120
(5) (1994) 1170–1190.
[6] L.N. Reddi, H. Wu, Mechanisms involved in vibratory destabili-
zation of NAPL ganglia in sands, Journal of Environmental
Engineering, ASCE 122 (12) (1995) 1115–1119.
[7] E.M. Simkin, A possible mechanism of vibroseismic action on an
oil-bearing bed, Journal of Engineering Physics and Thermophysics
64 (4) (1993) 355–359.
[8] E.M. Simkin, M.L. Surguchev, Advanced vibroseismic technique
for water flooded reservoir stimulation, Mechanism and field tests
results, The 6th European IOR Symposium, Stavanger, Norway,
1991, pp. 233–241.