Physical water treatment using RF electric elds for the mitigation of CaCO

3
fouling in cooling water
Leonard D. Tijing
a
, Hang Young Kim
b
, Dong Hwan Lee
a
, Cheol Sang Kim
a,
*
, Young I. Cho
c
a
Division of Mechanical Design Engineering, Chonbuk National University, Jeonju, Jeonbuk 561-756, Republic of Korea
b
Department of Precision Mechanical Engineering, Chonbuk National University, Jeonju, Jeonbuk 561-756, Republic of Korea
c
Department of Mechanical Engineering and Mechanics, Drexel University, Philadelphia, PA 19104, USA
a r t i c l e i n f o
Article history:
Received 16 June 2009
Received in revised form 15 November 2009
Accepted 15 November 2009
Available online 8 January 2010
Keywords:
RF electric eld
CaCO
3
fouling
Physical water treatment
Fouling resistance
a b s t r a c t
The present study investigated the feasibility of RF electrical elds in mitigating CaCO
3
fouling in cooling
water. Three different voltages and two frequencies were used for the RF electric elds produced directly
in water with varying cold-side velocity. Articial hard water was used. Fouling resistances for the PWT-
treated cases decreased by 3488% from the values for no-treatment cases, depending on the cold-side
ow velocity. The results showed blunt crystal structures for the PWT-treated cases, while no-treatment
cases had sharp and pointed crystal structures. The new PWT method using RF electric elds presents a
valid tool to mitigate CaCO
3
fouling in cooling water.
2009 Elsevier Ltd. All rights reserved.
1. Introduction
Scale formation on heat transfer surfaces of heat exchangers
presents a critical concern in industrial operation and mainte-
nance. The scale formation is commonly called as mineral fouling.
Circulating water in a cooling tower contains excessive mineral
ions especially calcium and magnesium ions as the evaporation
of pure water is the primary mode of heat rejection, leaving those
mineral ions behind. In particular, calcium ions reacting with car-
bonate ions in water form layers of CaCO
3
deposit on heat transfer
surfaces, decreasing the efciency of heat exchangers because of
the insulating effect of the deposits. Furthermore, the formed
deposits reduce the ow area, thus requiring more pumping power
to achieve the owrate of water corresponding to the clean state
[17]. A 0.8-mm layer of CaCO
3
scale can increase the energy use
by about 10% [8]. Steinhagen et al. [9] showed in a New Zealand
survey that 90% of heat exchangers had fouling problems. If one
can lessen or completely eradicate fouling on heat transfer sur-
faces, this not only increases heat exchanger efciency, but also
reduces the expenses associated with cleaning of fouled heat
exchangers. Furthermore, as the fouling can be mitigated, the
cycles of concentration (COC) can be increased, resulting in water
savings by reduced make-up and blowdown [7,1015].
Calcium carbonate is one of the most common scales found in
cooling-water applications. It exists in three crystal phases but
the two most common are aragonite and calcite. Aragonite has a
specic gravity of 2.95 while calcite has 2.71. The orthorhombic-
shaped and denser aragonite tends to be more adherent to heat
transfer surfaces than the hexagonal-shaped calcite [6,16]. The
chemical reaction for the precipitation of CaCO
3
is found in the lit-
erature [3,4,16].
There are several methods to help lessen, control, or prevent the
scaling of heat exchangers, which can be primarily divided into
two: chemical and non-chemical methods. In the past, the use of
scale-inhibiting chemicals like chlorine and brominated com-
pounds was the best choice for the control of mineral fouling.
Although it had a high success rate, there were also many disad-
vantages and concerns in their use. Aside from the high cost of
chemicals, more stringent environmental laws increased the costs
associated with their storage, handling and disposal. These chem-
icals or accumulated chemical residues over a long period of time
also pose concern to the environment with accidental spills. Thus,
there is a need for a new approach that is safe and clean from both
environmental and cost points of view in the maintenance of heat
exchangers [3,5,8,11,13].
Physical water treatment (PWT) is a non-chemical method to
mitigate mineral fouling with the use of electric or magnetic elds,
catalytic surfaces, ultrasounds, or sudden pressure changes.
Numerous studies have been reported for the effectiveness of the
use of permanent magnets [3,4,8,11,1720], solenoid coils
[3,4,8,2124], catalytic materials like copper, zinc [25,26], and tita-
nium [16], and ultrasounds [10]. In this paper, RF (radio frequency)
electric elds generated between two graphite electrode plates
0017-9310/$ - see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijheatmasstransfer.2009.12.009
* Corresponding author. Tel.: +82 63 270 4284; fax: +82 63 270 2460.
E-mail address: chskim@chonbuk.ac.kr (C.S. Kim).
International Journal of Heat and Mass Transfer 53 (2010) 14261437
Contents lists available at ScienceDirect
International Journal of Heat and Mass Transfer
j our nal homepage: www. el sevi er . com/ l ocat e/ i j hmt
directly in water were used to mitigate mineral fouling in a double-
pipe heat exchanger. The new method of using RF electric elds
directly in water is a major improvement over the above-men-
tioned PWT methods because these previous PWT methods
produce induced electric elds in water so that the eld strength
is often very small (1 mV/cm) and the maximum frequency of
the eld is limited to approximately 2 kHz due to involved physics
laws such as Faradays law [3,4].
1.1. Objectives
Calcium carbonate fouling is a rampant occurrence for cooling-
tower applications. Thus, the present study simulated CaCO
3
foul-
ing by preparing articial hard water through mixing of an appro-
priate amount of NaHCO
3
and CaCl
2
powders into a 150-L tank to
obtain the desired water hardness. The present study used RF elec-
tric elds between two graphite electrode plates positioned paral-
lel to each other at three different voltages (2, 5, and 13 V) and two
different frequencies (13.56 and 27.12 MHz). The objective of the
present study was to investigate the feasibility of RF electric elds
in mitigating CaCO
3
fouling in heat exchangers. The specic aims of
the present study were to determine the values of voltage and fre-
quency of the RF electric elds that would provide the best results
for CaCO
3
fouling mitigation, and to investigate the effects of water
hardness and ow velocity on the efcacy of the RF electric elds
in mitigating CaCO
3
fouling at the optimum voltage and frequency.
2. Experimental method
The schematic diagram of the experimental system is shown in
Fig. 1. It was composed of a hard water tank, a centrifugal pump, a
rotary ow meter, a two-graphite system to produce RF electric
elds, a heat transfer test section, a chiller, a hot water circulating
bath, LabView device, and a personal computer. The hard water
tank with a cover was made of plastic with a maximum capacity
of 200 L. Inside the tank was a copper coil connected to a chiller
to maintain a constant temperature of the hard water during the
tests. The centrifugal pump was controlled by an inverter to con-
trol the desired ow rate together with the rotary ow meter.
The pump in the hot water re-circulating bath was also controlled
by an inverter. The ow circulation loop was composed of copper
tubes, expandable hoses, and heat-resistant Teon tubes.
2.1. Calibration of RTD
Four RTD temperature sensors were used to measure the tem-
peratures at the inlet and outlet of the heat transfer test section.
The RTD sensors were mounted using a T-shaped compression t-
ting. The RTDs were calibrated using a temperature calibration
block bath and a precision thermometer with an accuracy of
0.01 C as the reference together with the LabView Measurement
and Automation Explorer (MAX Ver. 4.00) calibration system.
2.2. Preparation for fouling test and heat exchanger efciency
Before each test, the heat transfer test section was tested for
leakage. After ensuring that there was no leakage in the test sec-
tion, the efciency of the heat exchanger was tested. Both hot
and cold water were kept at constant inlet temperatures of
85.5 0.5 and 23.5 0.5 C, respectively. The heat transfer test sec-
tion was covered with a Styrofoam insulation to minimize any heat
loss to the surroundings. The RTDs were also covered with Styro-
foam to prevent any heat transfer from the surroundings that
Nomenclature
A
o
outer surface area of copper tube (m
2
)
c
p
specic heat of water (J/kg K)
d diameter of copper tube (m)
D diameter of quartz crystal (m)
f frequency (Hz)
H height (m)
L length (m)
M
w
molecular weight (g/mol)
Q heat transfer rate (W)
R
f
fouling resistance (m
2
K/W)
DT
LMTD
log-mean-temperature difference (C)
U overall heat transfer coefcient (W/m
2
K)
W width (m)
Greek symbol
h diffraction angle ()
Subscripts
c cold side
f fouled state
i initial clean state
i inner
o outer
q quartz crystal
t tube (copper)
Fig. 1. Schematic diagram of the present experimental set-up.
L.D. Tijing et al. / International Journal of Heat and Mass Transfer 53 (2010) 14261437 1427
might affect the temperature readings. Both heat transfer test sec-
tion and the entire fouling set-up were grounded to lessen the pos-
sible effect of electrical noise and static electricity on the
temperature readings and in the process as a whole.
The discrepancy in the heat transfer rates between the hot- and
cold-water sides was found to be 8%. In this study, the heat trans-
fer rate at the cold-side Q
c
was used to calculate the overall heat
transfer coefcient, as the hot water side might have incurred par-
asitic losses to the surroundings. The measurement errors of the
diameters and lengths of the tubes in the heat exchanger are found
in Table 1.
2.3. Test section
Fig. 2 shows the double-pipe heat transfer test section. The inner
tube was made of copper with dimensions of d
i
= 13.85 0.05 mm,
d
o
= 16 0.05 mm, and L
t
= 631 1.0 mm, while the outer tube
was made of quartz crystal with D
o
= 28.1 0.05 mm, D
i
= 24.8
0.05 mm, and L
q
= 499 1.0 mm. The effective heat exchanger
length was 600 1.0 mm. The annulus gap formed between the
two tubes had a cross-sectional area of 2.82 cm
2
. O-rings were used
on grooves at the inlet and outlet sides of Teon head blocks for the
copper tube to prevent any leakage between the hot water and the
cold water. O-rings were also used between the quartz crystal and
the Teon head blocks to prevent leakage in the cold-water side.
Four steel rods locked with nuts xed the test section in place. Table
1 gives the dimensions of the heat transfer test section.
2.4. Physical water treatment
Fig. 3 shows the PWT device using two graphite electrode plates
to produce RF electric elds in the present study. The two graphite
plates (L = 155 0.05 mm, W = 7.65 0.05 mm, and H = 35
0.05 mm) were placed in parallel to each other inside a cylindrical
Teon block, forming a rectangular cross-section. The distance be-
tween the two graphite electrode plates was xed at 20 0.05 mm.
The outer casing was made of a cylindrical stainless steel structure
with specially designed screw ports where metal screws were
placed as the connections for the power supply so that any direct
contact of the connecting wires to water could be prevented. The
present PWT device had detachable diverging inlet and converging
outlet head sections. O-rings were used between the main cylinder
casing and the detachable heads, and locked with screw-type locks.
The power supply included a voltage controller (100 mA, 15 V), a
programmable waveform generator (150 MHz), and an oscillo-
scope. The current study utilized two radio frequencies, 13.56
and 27.12 MHz, which are allowed by the Federal Communications
Commission (FCC) for Industrial, Scientic, and Medical (ISM)
equipment [27]. Test voltages ranging from 2 to 13 V were used
to determine the optimum voltage.
The present PWT method utilized oscillating electric elds pro-
duced directly in water. This allows oscillating electric elds at a
signicantly higher strength and frequency than the previous
Table 1
Heat transfer test section details.
Copper tube Quartz crystal
Inner diameter (d
i
) (mm) 13.85 0.05 24.80 0.05
Outer diameter (d
o
) (mm) 16.00 0.05 28.10 0.05
Length (mm) 631 1.0 499 1.0
Effective heat exchange length (mm) 600 1.0
Inner surface area (cm
2
) 261.07
Outer surface area (cm
2
) 301.59
Mean surface area (cm
2
) 281.33
Area (cross-sectional) (cm
2
) 1.51
Area (annulus) (cm
2
) 2.82
Fig. 2. The present heat transfer test section: (a) side view and (b) cross-sectional view.
1428 L.D. Tijing et al. / International Journal of Heat and Mass Transfer 53 (2010) 14261437
PWT methods, which used indirect methods. There are three ways
mineral ions precipitate inside the heat exchanger: (a) directly on
the heat transfer surface where the temperature is high, (b) on the
surface of solid particles suspended in water, and (c) in bulk water
[8]. It is hypothesized that the oscillating electric elds in the pres-
ent PWT method precipitate the dissolved mineral ions such as cal-
cium to mineral salts in bulk water. As the mineral ions continue to
precipitate and adhere on the surfaces of the suspended particles,
the particles grow in size and adhere to the solid heat transfer sur-
face in the form of soft sludge or particulate fouling. This type of
fouling is believed to easily be removed by shear forces created
by ow than those deposits produced from the precipitation of
mineral ions directly on the solid heat transfer surface, i.e., precip-
itation fouling [8].
2.5. Articial hard water
The hardness of water in the study varied from 350 to 550 ppm.
The desired water hardness was obtained by mixing two chemicals
at right proportions (see Table 2): calcium chloride (CaCl
2
) (Showa
Chemical Co., Ltd., Japan) and sodium bicarbonate (NaHCO
3
) (Sho-
wa Chemical Co., Ltd., Japan). During the tests, there were no addi-
tional chemicals (i.e., CaCl
2
or NaHCO
3
) added into the system so
that we allowed the water hardness level to gradually go down
due to fouling formation. The initial water hardness was at
350 20 and 550 20 ppm for all tests.
The hard water reservoir was rst lled with distilled water at a
volume of approximately 150 L after cleaning. With the use of a by-
pass line, the distilled water in the reservoir was continuously cir-
culated (not through the whole fouling system set-up, but only in
the reservoir). CaCl
2
powder was added to the reservoir with gentle
stirring using a metal stick and was left for 10 min to dissolve. Sub-
sequently, NaHCO
3
powder was also added to the reservoir and
was gently stirred. After 57 min, about 100-ml of water sample
was taken out from the bottom of the reservoir (see Fig. 1) and
used for water chemistry measurements. After 10 min of mixing,
two more 100-ml hard water samples were collected for water
chemistry measurements and used as the initial water chemistry
data (t = 0). Samples were taken three more times at 1218 h inter-
val. Each water sample collected was about 100 ml.
2.6. Fouling test
The hot- and cold-water side temperatures at the heat transfer
test section were maintained constant using a hot water re-circu-
lating bath and a chiller, respectively. Hot water owed inside
the copper tube at a constant velocity, while the cold water made
of articial hard water passed through the annulus gap owing in
the opposite direction at three different velocities (0.3, 0.8, and
1.0 m/s). The corresponding velocity at the graphite electrode
plates was 0.12, 0.32, and 0.40 m/s, respectively. The heat transfer
rate varied from 2.9 to 5.7 kW depending on the velocity at the
cold-water side. RTDs connected to a real-time data acquisition
system (LabView 7.0) were used to measure temperatures every
5 min for a test period of 3640 h. Table 3 shows the summary
of the experimental parameters.
The overall heat transfer coefcient U was calculated using the
following equation [28]:
U
Q
c
A
o
DT
LMTD
1
where A
o
is the outer surface area of the copper tube and DT
LMTD
is
the log-mean-temperature difference.
The fouling resistance, R
f
was calculated using the following
equation [28]:
R
f

1
U
f

1
U
i
2
Fig. 3. Present PWT device using two graphite electrode plates.
Table 2
Preparation of articial hard water.
Distilled water volume: 0.150 m
3
Hardness (ppm)
350 550
Calcium chloride (CaCL
2
: M
w
= 110.98 g/mol) (g) 58 97
Sodium bicarbonate (NaHCO
3
: M
w
= 84.01 g/mol) (g) 88 147
Table 3
Summary of the experimental parameters.
Parameter Hot water Cold (hard) water
Inlet temperature (C) 85.5 0.5 23.5 0.5
LMTD (C) 5356
Flow velocity (m/s) 1.11.3 0.31.0
Mass owrate (kg/s) 0.1620.194 0.0840.281
Heat transfer rate (W) 29005800
Reynolds number 43,00053,000 11003900
Water hardness (ppm) 1050 350550
L.D. Tijing et al. / International Journal of Heat and Mass Transfer 53 (2010) 14261437 1429
where U
f
is the overall heat transfer coefcient for fouled states,
while U
i
is the overall heat transfer coefcient corresponding to
the initial clean state. The latter (U
i
) was determined using distilled
water only (without chemicals) and without the use of PWT device
during the initial calibration run prior to the fouling tests with arti-
cial hard water.
The present study was divided into two parts:
(a) The rst set of tests was done at a high articial water hard-
ness of 550 ppm to investigate the feasibility of PWT method
using RF electric elds to mitigate mineral fouling, and to
further determine the optimum voltage and frequency of
the electric eld.
(b) The second set of tests utilized the optimum values for volt-
age and frequency to generate RF electric elds at a lowered
articial water hardness of 350 ppm and at three different
cold-water side ow velocities at the heat exchanger (0.3,
0.8, and 1.0 m/s).
All experimental procedures, materials, equipment, inlet tem-
peratures for hot- and cold-water sides, and volume of water used
were the same for all tests.
2.7. Cleaning procedure
After each test, the hard water tank was cleaned thoroughly.
After the previously used articial hard water was completely
drained, a tap water jet using a hose was then ushed onto the in-
side wall of tank together with a cleaning brush to remove any
scales formed inside the tank. Traces of CaCO
3
agglomerated
deposits were drained out of the tank. After this, tap water was
introduced into the tank and was circulated in the system without
the test section for at least one hour. When most visible traces of
the hard water previously used were gone, the tap water was com-
pletely drained out and new distilled water up to 150 L was sup-
plied to the tank. No chemicals were added during the cleaning
process. The RTDs were thoroughly cleaned by using a ne grain
sandpaper to scratch off gently the scales formed on the RTD sur-
face, and then wiped with tissue paper to ensure a clean RTD sur-
face for the next test. This was done so as to lessen the effect of
scales on the RTD which might have produced erroneous temper-
ature measurement. After completion of each test, the test section
was disassembled, and the fouled tube was dried on a display
board. The quartz crystal tube was thoroughly cleaned with tap
water and wet tissue to remove the scales in the inside of the
quartz tube. The inside of the Teon heads were also cleaned and
dried. After which, a new copper tube was installed in the test sec-
tion, and the quartz tube was also installed putting Teon seal on
both ends of the two tubes to ensure no leakage, as mentioned
early.
2.8. Sampling of CaCO
3
scales
Scanning electron microscopy images and X-ray diffraction
analyses were obtained from the fouled copper tubes.
2.8.1. Scanning electron microscopy (SEM)
Deposit samples of approximately 0.5 0.5 cm were obtained
from the fouled tubes (both no-treatment and PWT-treated cases)
by manually cutting the tube using a saw blade. SEM images were
taken at a magnication of 5000 to examine the deposit surface
topography and geometry. The samples for SEM were prepared
carefully making sure that there were no scratches on their sur-
faces. Care was given not to contaminate the deposit with any
impurities during the sample preparation. The deposit samples
were coated by gold through low-vacuum sputter coating in order
to prevent the accumulation of static electric charge during the
irradiation of electron.
2.8.2. X-ray diffraction (XRD)
XRD analyses were conducted on 1 1 cm samples taken from
the fouled tubes by manual cutting to determine and characterize
the crystallographic structure of the deposits for both the no-treat-
ment and PWT-treated cases. The CaCO
3
scales on the fouled tube
were expected to exhibit calcite, aragonite, or vaterite crystal pat-
terns of CaCO
3
or their combinations.
3. Results and discussion
3.1. Fouling resistance
Fig. 4 shows the results for the fouling tests obtained using
water hardness of 550 ppm and a constant velocity of 0.3 m/s
for both no-treatment and PWT-treated cases. The no-treatment
case was conducted twice for repeatability check and it was
found that both fouling curves looked very consistent with max-
imum 10% difference in the asymptotic fouling resistance value at
the end of the test. Due to a very high water hardness, there was
no induction period observed in all tests. An induction period is
usually indicated by a straight horizontal line in the beginning
of the fouling curve, and is dened as the lateral spreading of
scale deposits on heat transfer surface until it is fully covered
mainly by stable crystal nucleation [29]. The articial hard water
that contained the articially made calcium and bicarbonate ions
reacted quickly to the hot heat transfer surface, making immedi-
ate deposition on the surface. The scale deposition involved the
cumulative effect of direct diffusion of dissolved calcium ions to
the heat transfer surface and the deposition of precipitated cal-
cium salt particles due to supersaturated conditions and acceler-
ated precipitation of calcium salts by PWT [4]. The no-treatment
case (i.e., without the use of PWT device) had a steep increase in
fouling resistance in the rst 10 h of operation, indicating that
CaCO
3
precipitated and continuously adhered to the heat transfer
surface. At t = 12 h, the fouling resistance reached an asymptotic
level, suggesting that a signicant removal of scale particles
was taking place, maybe due to reduced opening and subse-
quently increased shear stress.
The PWT cases gave very different fouling curves. For exam-
ple, the slope in the fouling curve during the rst 10 h was much
No-treatment
-5.E-05
5.E-05
2.E-04
3.E-04
4.E-04
5.E-04
6.E-04
0 6 12 18 24 30 36
Time (hour)
R
f

(
m
2
K
/
W
)
PWT-treated: 2V, 13.56 MHz
PWT-treated: 2V, 27.12 MHz
PWT-treated: 5V, 13.56 MHz
PWT-treated: 13V, 27.12 MHz
76%
83%
85%
Fig. 4. Tests at different voltages and two different frequencies (550 ppm, 0.3 m/s).
1430 L.D. Tijing et al. / International Journal of Heat and Mass Transfer 53 (2010) 14261437
smaller than that observed for the no-treatment case. If one as-
sumes that the deposition of new scales for the PWT case was
approximately the same as that for the no-treatment case, one
could conclude that the removal rate was signicantly greater
for the PWT case than the no-treatment case. The fouling resis-
tance continued to rise, reaching asymptotic levels at approxi-
mately t = 15 h, after which the fouling resistances remained
almost unchanged. The PWT-treated cases at 2 V, 13.56 MHz
and 5 V, 13.56 MHz gave similar fouling resistance drops of
about 76% from the no-treatment case value from 4.1 10
4
to 9.1 10
5
m
2
K/W. The best result in the fouling resistance
drop was obtained from the case of 2 V and 27.12 MHz where
there was an 85% reduction in the fouling resistance. As the fre-
quency in the RF electric eld increased, the efcacy of the PWT
in the prevention of fouling also increased. One can explain that
at a higher frequency, the calcium ions have a greater chance to
Fig. 5. Photographic images of the clean and fouled tubes at articial CaCO
3
hardness.
L.D. Tijing et al. / International Journal of Heat and Mass Transfer 53 (2010) 14261437 1431
collide with each other, thus enabling precipitation in the bulk
water at a higher rate. The precipitated calcium salt particles
in bulk water adhere to the heat transfer surface, but in the soft
sludge form of particulate fouling via gravitational settling and
particle transport. Particulate fouling is known to be less-adher-
ent which can be relatively easily removed by the shear force
produced by ow velocity [4]. Visual inspection on the fouled
tubes also conrmed thinner deposits for the PWT-treated cases
as compared with the no-treatment case (see Fig. 5). It was also
noticeable that as the velocity of cold water in the heat transfer
test section was increased, the fouling deposit decreased its
thickness as expected for both PWT-treated and no-treatment
cases (see Fig. 5b).
Fig. 6 presents the effect of ow velocity on the fouling resis-
tance for hardness of 350 ppm for no-treatment case at three dif-
ferent velocities (0.3, 0.8, and 1.0 m/s). The results show that as
the velocity increased, the fouling resistance decreased. There were
no induction periods for all three cases. At the lowest ow velocity
of 0.3 m/s, the asymptotic fouling resistance was at its highest as
expected (4.46 10
4
m
2
K/W). This was because the ow shear
forces at 0.3 m/s was the least among the three cases, which might
not be high enough to remove the scales from the heat transfer
surface. As velocity was increased to 0.8 and 1.0 m/s, the asymp-
totic fouling resistances decreased to 2.51 10
4
and 1.01
10
4
m
2
K/W, respectively, signicantly lower than the fouling
resistance obtained at 0.3 m/s. This supports the ndings of previ-
ous studies [29] that at higher velocities, even if there was a higher
mass deposition rate, the high ow shear forces might be large en-
ough to continuously remove the scales.
The comparison between the no-treatment cases and the PWT-
treated cases at 2 V and 13.56 MHz is shown in Fig. 7. The test con-
ditions of 2 V and 13.56 MHz were chosen as they were the best
ones among the test conditions used in the present study (see
Fig. 4). The asymptotic fouling resistances dropped by 3488% for
the PWT cases depending on the velocity of cooling water. The
asymptotic fouling resistance at the 0.3 m/s case decreased by
about 88% from the no-treatment case. The RF electric eld is be-
lieved to give a greater opportunity for ions to collide with each
other compared with the previous PWT methods such as perma-
nent magnets or solenoid coils. Accordingly, the bulk precipitation
of calcium particles in water should take place more efciently
than the previous PWT methods. Subsequently, the calcium salt
particles suspended in water adhere to the heat transfer surface
in the form of particulate fouling. As the particulate fouling depos-
its are known to be soft and not so adherent to the heat transfer
surface thus even at lowowvelocities (such as 0.3 m/s), the scales
could be continuously removed by the shear forces created by the
low ow velocity. Table 4 presents the percentage of fouling resis-
tance reduction by PWT-treatment case from the no-treatment
case at different velocities.
3.2. SEM images
The SEM images (5000) of the CaCO
3
scales obtained for both
the no-treatment and PWT-treated cases are shown in Figs. 8 and
9. Fig. 8 shows the SEM images of the deposits for the case with a
water hardness of 550 ppm and at ow velocity of 0.3 m/s. The no-
treatment case (Fig. 8a) showed sharp and pointed needle-like
crystal structures (less than 1 lm in size) while the PWT-treated
cases (Fig. 8be) showed hexagonal (cubic-like) crystal structures.
The former (no-treatment) depicts the orthorhombic shape of ara-
gonite CaCO
3
crystals which are known to be denser and more
adherent than calcite. The latter (PWT-treated) are more of calcite
crystal structures which are believed to be easily removed by shear
forces created by ow. XRD analyses given in Fig. 11 conrm the
crystal structures. The needle-like structures at the no-treatment
case suggests that the crystals might have grown directly on the
heat transfer surface through crystallization reaction with addi-
tional ions diffusing from the bulk water to the heat transfer sur-
face [4]. On the other hand, for the PWT-treated cases, the CaCO
3
crystals were found to be generally bigger than those for the no-
treatment case and appeared to be an accumulation of small-sized
crystallites. It can be speculated that submicron-size CaCO
3
parti-
cles were initially produced in the bulk water by the RF electric
eld and they eventually grewin size due to continuous circulation
0.3 m/s
0.8 m/s
1.0 m/s
-1.E-04
0.E+00
1.E-04
2.E-04
3.E-04
4.E-04
5.E-04
6.E-04
0 6 12 18 24 30 36
Time (hour)
R
f

(
m
2
K
/
W
)
Fig. 6. No-treatment case at different ow velocities (350-ppm CaCO
3
hardness).
0.3 m/s, NT
0.8 m/s, NT
1.0 m/s, NT
-1.E-04
0.E+00
1.E-04
2.E-04
3.E-04
4.E-04
5.E-04
6.E-04
0 6 12 18 24 30 36
Time (hour)
R
f

(
m
2
K
/
W
)
0.3 m/s, PWT-treated 2V, 13.56 MHz
0.8 m/s, PWT-treated 2V, 13.56 MHz
1.0 m/s, PWT-treated 2V, 13.56 MHz
88%
65%
34%
Fig. 7. Comparison of the no-treatment and treated cases using two graphite
electrode plates at 2 V, 13.56 MHz at different ow velocities (350-ppm CaCO
3
hardness).
Table 4
Percentage reduction of fouling resistance by PWT-treatment from the no-treatment
case at different velocities (350-ppm CaCO
3
hardness).
Velocity at the cold-water
side of the heat transfer
test section (m/s)
Equivalent velocity
at the PWT device (m/s)
PWT-treated case R
f
reduction from the
no-treatment case (%)
0.3 0.12 88
0.8 0.32 65
1.0 0.40 34
1432 L.D. Tijing et al. / International Journal of Heat and Mass Transfer 53 (2010) 14261437
in the cooling loop and deposited unto the surface in the form of
small particulates [4].
Fig. 9 shows the SEM images at three different velocities for
350-ppm hard water case. The SEM images for the no-treatment
cases with V = 0.3 and V = 0.8 m/s had less than 1 lm in crystal
thickness, whereas those obtained with V = 1.0 m/s had less than
3 lm in size. All of the three no-treatment cases showed sharp
and pointed crystal structures. The sharp and pointed crystal tips
could have increased the attachment capability of the deposit to
adhere on the heat transfer surface more strongly than blunt
crystals for the PWT cases. On the other hand, the PWT-treated
cases (2 V, 13.56 MHz treatment) showed larger crystal structures
with blunt and dull crystal surfaces. At V = 0.3 m/s, the crystals
looked like a bigger version of the crystals at the no-treatment
condition but with dull surface and thickness of more than
4 lm. The crystal structure obtained at V = 0.8 m/s (PWT-treated)
was hexagonal in shape with thickness of over 4 lm. The PWT-
treated crystal structures obtained at V = 1.0 m/s showed at
and aky-type agglomerated scales. One could see the lesser
steepness in the structure of the images of those obtained for
the PWT-treated cases as compared with those for no-treatment
cases especially when velocity was increased, presumably due
to a larger ow shear forces acting on the deposit. Table 5 pre-
sents the summary of the test results.
3.3. X-ray diffraction
Calcium carbonate is a crystalline substance that exists in three
polymorphs: calcite, aragonite, and vaterite [30]. Each polymorph
has unique crystallographic structure. Every substance has unique
XRD spectrum that serves as like its ngerprint [31]. Fig. 10 shows
the standard XRD spectra of the three crystal phases of calcium
carbonate. Calcite has a prominent peak of intensity at 2h = 29.5
and Miller index of 104, while aragonite has its peak intensity at
2h = 45.9 and Miller index of 221. Miller index is a notation to
quantify the planes and directions in crystal lattices [31].
Fig. 11 presents the results of the present XRD analyses for both
the no-treatment and PWT-treated cases at water hardness of
550 ppm and a ow velocity of 0.3 m/s. The results here were com-
pared to the standard XRD spectra of CaCO
3
found in Fig. 10. XRD
analysis was conducted to determine the crystallographic phase of
the deposit so that the focus was on the spectrum peaks and not on
the intensity. For the no-treatment case, the peaks depicted that of
an aragonite crystal while the results obtained from all the PWT-
treated cases resembled the calcite phase of calcium carbonate.
The XRD results here support the SEM image results given in
Fig. 8. The supersaturated condition in cooling water must have
spontaneously produced coagulated CaCO
3
particles in bulk water
and with the added effect of RF electric elds, more particles must
Fig. 8. SEM photographs (5000) of the scales for the (a) no-treatment and (be) PWT-treated cases (550-ppm CaCO
3
hardness and V = 0.3 m/s).
L.D. Tijing et al. / International Journal of Heat and Mass Transfer 53 (2010) 14261437 1433
Fig. 9. SEM photographs (5000) of the scales for the no-treatment and PWT-treated cases at 350-ppm CaCO
3
hardness.
Table 5
Summary of the test results for fouling resistance, SEM, and XRD.
Parameter R
f
, m
2
K/W (% decrease from no-treatment case) SEM XRD
550 ppm
No-treatment 4.1 10
4
Sharp, pointed, needle-like Aragonite
2 V, 13.56 MHz 9.1 10
5
(;76%) Hexagonal, thick Calcite
2 V, 27.12 MHz 6.1 10
5
(;85%) Hexagonal, thick Calcite
5 V, 13.56 MHz 9.1 10
5
(;76%) Hexagonal, thick Calcite
13 V, 27.12 MHz 6.9 10
5
(;83%) Hexagonal, thick Calcite
350 ppm
No-treatment (0.3 m/s) 4.5 10
4
Sharp, pointed, needle-like Aragonite
2 V, 13.56 MHz (0.3 m/s) 5.4 10
5
(;88%) Hexagonal, thick Aragonite
No-treatment (0.8 m/s) 2.5 10
4
Sharp, pointed, needle-like Aragonite
2 V, 13.56 MHz (0.8 m/s) 8.8 10
5
(;65%) Hexagonal, thick Calcite
No-treatment (1.0 m/s) 1.0 10
4
Sharp, pointed, needle-like Aragonite
2 V, 13.56 MHz (1.0 m/s) 6.7 10
5
(;34%) Agglomerated, at, thick Aragonite
1434 L.D. Tijing et al. / International Journal of Heat and Mass Transfer 53 (2010) 14261437
have precipitated in the bulk water either by collision between
mineral ions or by attachment of ions to the nucleates. Since they
were not directly exposed to the high-temperature heat transfer
surface, the precipitated particles could be in the form of calcite
crystals. On the other hand, the mineral ions that precipitated di-
rectly on the hot heat transfer surface are believed generally to
form into aragonite crystals, as with the case of no-treatment.
The results for the no-treatment cases at a water hardness of
350 ppm and at different ow velocities are shown in Fig. 12. At
the no-treatment case, most of the dissolved ions were believed
XRD spectra of Calcium carbonate
ARAGONITE
024
312
151 1 1 3 1 4 2
222
023
113
132
202
221
220
211
130
112
031
200
012
002
021
111
VATERITE
006
114
201
200
104
112
110
004 103
102
101
100
002
CALCITE
217
300
214
122
211
116
018
024
202
113
110
104
012
20 25 30 35 40 45 50 55 60 65 70
2-theta (degree)
R
e
l
a
t
i
v
e

i
n
t
e
n
s
i
t
y
Fig. 10. Standard XRD spectra of CaCO
3
.
A
A
A
A
A
A
A
A
A
A
A
No-treatment
C
C C
C
C
C
C
C
C
2 V, 13.56 MHz
C
C
C
C
C
C
C
C
C
C
2 V, 27.12 MHz
C C
C
C
C
C
C
C
C
5 V, 13.56 MHz
C
C
A : Aragonite
C: Calcite
C
C
C
C C C
13 V, 27.12 MHz
C
20 25 30 35 40 45 50 55 60 65 70
2-theta (degree)
R
e
l
a
t
i
v
e

i
n
t
e
n
s
i
t
y
Fig. 11. XRD analyses of the scales for the no-treatment and PWT-treated cases at V = 0.3 m/s and 550-ppm CaCO
3
hardness.
L.D. Tijing et al. / International Journal of Heat and Mass Transfer 53 (2010) 14261437 1435
to directly diffuse and deposit on the heat transfer surface, thus
making them more adherent and afnitive on the surface. The
scales formed directly on the heat transfer surface (i.e., precipita-
tion fouling) are known to be tenacious and difcult to remove.
The no-treatment cases depicted those of aragonite crystal struc-
tures, which are known to be a troublesome type of scales. On
the other hand, Fig. 13 shows the XRD spectra for the PWT-treated
cases at 2 V and 13.56 MHz. The results obtained at 0.3 and 1.0 m/s
velocities revealed those of aragonite crystal phase, and the scale
was calcite for 0.8 m/s. There are many factors that could affect
the morphology of crystals on heat transfer surfaces. Among such
were water temperature, hardness, owvelocity at the test section,
and the residence time and frequency of RF electric elds at the
PWT device. The differences in the residence time and frequency
of the treatment of cooling water inside the PWT device might
have an effect in the coagulation tendency of dissolved mineral
ions and particles in the bulk water. At a ow velocity of 0.3 m/s,
the residence time of water (thus calcium ions) at the PWT device
might be long enough, yet the cooling water returned to the PWT
device for treatment less frequently compared to greater ow
cases. The result at a ow velocity of 0.8 m/s showed calcite form
of calcium carbonate. Its residence time and the frequency of treat-
ment at the PWT device might have been just enough for the given
hardness of water. It is of note that the hardness in this test was at
350 ppm which was lower than the 550 ppm in the rst test yet
still at signicantly supersaturated condition.
4. Conclusion
The present study investigated the effectiveness of a new phys-
ical water treatment method using RF electric elds on the mitiga-
tion of CaCO
3
fouling in a concentric tube heat exchanger. The
study utilized RF electric elds produced between two graphite
electrode plates directly in water. Fouling tests were conducted
No-treatment (350 ppm)
A
A
A
A
A
A
A
A
0.3 m/s
A A
A
A
A
A
A
A
A
A
0.8 m/s
A
A
A
A
A
A
A
A
A
1.0 m/s
20 25 30 35 40 45 50 55 60 65 70
2-theta (degree)
R
e
l
a
t
i
v
e

i
n
t
e
n
s
i
t
y
Fig. 12. XRD analyses of the scales for the no-treatment cases at varying ow velocities and 350-ppm CaCO
3
hardness.
PWT-treated: 2 V, 13.56 MHz (350 ppm)
A
A
A
A
A
A
A
A
A
A
A
0.3 m/s
C
C
C
C
C C
C
C
0.8 m/s
C
A
A
A
A
A
C
A
A
A
C
1.0 m/s
20 25 30 35 40 45 50 55 60 65 70
2-theta (degree)
R
e
l
a
t
i
v
e

i
n
t
e
n
s
i
t
y
Fig. 13. XRD analyses of the scales for the PWT-treated cases at varying ow velocities and 550-ppm CaCO
3
hardness.
1436 L.D. Tijing et al. / International Journal of Heat and Mass Transfer 53 (2010) 14261437
at two different water hardness (350 and 550 ppm) using articial
hard water, and at three different cold-water side velocities of 0.3,
0.8, and 1.0 m/s, with the corresponding ow velocities inside the
graphite plates of 0.12, 0.32, and 0.40 m/s.
The fouling resistances for the PWT-treated cases dropped by
3488% compared with those obtained for the no-treatment cases,
depending on the cold-side ow velocity and the water hardness.
SEM images showed hexagonal and agglomerated crystal-shape
structures for the PWT-treated cases as compared to sharp and
needle-like structure for the no-treatment cases. The XRD analyses
showed aragonite crystals for the no-treatment cases while the
PWT-treated cases showed both aragonite and calcite forms of
scale depending on water hardness and ow velocity. The voltage
and frequency for the best mitigating performance were found to
be 2 V and 13.56 MHz, respectively. Based from the ndings of this
study, it is concluded that the new PWT method using RF electric
elds is a valid tool to mitigate mineral fouling in cooling-water
applications.
Acknowledgment
This research was supported by the Ministry of Education,
Science and Technology through the National Research Foundation
of Korea (Grant No. 2009-0075486).
References
[1] T.R. Bott, Fouling of Heat Exchangers, Elsevier Science B.V., The Netherlands,
1995.
[2] E.F.C. Somerscales, Fouling of heat exchangers: an historical review, Heat
Transfer Eng. 11 (1) (1990) 1936.
[3] Y.I. Cho, S.H. Lee, W. Kim, Physical water treatment for the mitigation of
mineral fouling in cooling-water applications, ASHRAE Transactions 109 (1)
(2003) 346357.
[4] Y.I. Cho, A.F. Fridman, S.H. Lee, W.T. Kim, Physical water treatment for fouling
prevention in heat exchangers, Adv. Heat Transfer 38 (2004) 171.
[5] C.B. Panchal, J.G. Knudsen, Mitigation of water fouling: technology status and
challenges, Adv. Heat Transfer 31 (1998) 431474.
[6] T.R. Bott, Aspects of crystallization fouling, Exp. Therm. Fluid Sci. 14 (1997)
356360.
[7] T. Kuppan, Heat Exchanger Design Handbook, Marcel Dekker, Inc., New York,
2000.
[8] The U.S. Department of Energy, Non-Chemical Technologies for Scale and
Hardness Control, DOE/EE-0162, 1998.
[9] R. Steinhagen, H. Muller-Steinhagen, K. Maani, Problems and costs due to heat
exchanger fouling in New Zealand industries, Heat Transfer Eng. 14 (1) (1993)
1930.
[10] S. Radler, U. Ousko-Oberhoffer, Optimised heat exchanger management
achieving nancial and environmental targets, in: Proceedings of the Sixth
International Conference on Heat Exchanger Fouling and CleaningChallenges
and Opportunities RP2, 2005.
[11] H. Muller-Steinhagen (Ed.), Handbook of Heat Exchanger FoulingMitigation
and Cleaning Technologies, Publico Publications, Germany, 2000.
[12] L.D. Tijing, H.Y. Kim, D.H. Lee, C.S. Kim, Y.I. Cho, Use of an oscillating electric
eld to mitigate mineral fouling in a heat exchanger, Exp. Heat Transfer 22
(2009) 257270.
[13] Y.I. Cho, J. Lane, W. Kim, Pulsed-power treatment for physical water treatment,
Int. Commun. Heat Mass Transfer 32 (2005) 861871.
[14] K.D. Demadis, E. Mavredaki, A. Stathoulopoulou, E. Neofotistou, C. Mantzaridis,
Industrial water systems: problems, challenges and solutions for the process
industries, Desalination 213 (2007) 3846.
[15] J. Lane, Y. Cho, W. Kim, Pulsed-power water treatment as a green scale
inhibitor for HVAC and once-through industrial systems, Corrosion (2004) 1
21, paper no. 04541.
[16] L.D. Tijing, B.C. Pak, D.H. Lee, Y.I. Cho, Heat-treated titanium balls for the
mitigation of mineral fouling, Exp. Heat Transfer 21 (2) (2007) 115132.
[17] J.S. Baker, S.J. Judd, Magnetic amelioration of scale formation, Wat. Res. 30 (2)
(1996) 247260.
[18] S.A. Parsons, S.J. Judd, T. Stephenson, S. Udol, B.L. Wang, Magnetically
augmented water treatment, Process Saf. Environ. Prot. 75 (B2) (1997) 98104.
[19] J. Donaldson, S. Grimes, Lifting the scales from our pipes, New Sci. 18 (1988)
4346.
[20] C. Smith, P.P. Coetzee, J.P. Meyer, The effectiveness of a magnetic physical
water treatment device on scaling in domestic hot-water storage tanks, Water
SA 29 (3) (2003) 231236.
[21] X. Xiaokai, Research on the electromagnetic anti-fouling technology for heat
transfer enhancement, Appl. Therm. Eng. 28 (2008) 889894.
[22] X. Xiaokai, M. Chongfang, C. Yongchang, Investigation on the electromagnetic
anti-fouling technology for scale prevention, Chem. Eng. Technol. 28 (12)
(2005) 15401545.
[23] Z. Quan, Y. Chen, C. Ma, C. Wang, B. Li, Experimental study on anti-fouling
performance in a heat exchanger with low voltage electrolysis treatment, Heat
Transfer Eng. 30 (3) (2009) 181188.
[24] G.V. Ushakov, Antiscaling treatment of water by an electric eld in heat-supply
networks, Therm. Eng. 55 (7) (2008) 570573.
[25] G.J. Lee, L.D. Tijing, B.C. Pak, B.J. Baek, Y.I. Cho, Use of catalytic materials for the
mitigation of mineral fouling, Int. Commun. Heat Mass Transfer 33 (1) (2006)
1423.
[26] P.P. Coetzee, M. Yacoby, S. Howell, S. Mubenga, Scale reduction and scale
modication effects induced by Zn and other metal species in physical water
treatment, Water SA 24 (1) (1998) 7784.
[27] Operating Frequencies, Federal Communications Commission, Code of Federal
Regulations, USA, Title 47, vol. 1, Part 18, 18.301, 2007.
[28] F.P. Incropera, D.P. DeWitt, Fundamentals of Heat and Mass Transfer, fth ed.,
John Wiley and Sons, New York, 2002.
[29] Q. Yang, Y. Liu, A. Gu, J. Ding, Z. Shen, Investigation of induction period and
morphology of CaCO
3
fouling on heated surface, Chem. Eng. Sci. 57 (2002)
921931.
[30] C.G. Kontoyannis, N.V. Vagena, Calcium carbonate phase analysis using XRD
and FT-Raman spectroscopy, Analyst 125 (1999) 251255.
[31] W.F. Smith, Principles of Materials Science and Engineering, McGraw-Hill Book
Co., Singapore, 1986.
L.D. Tijing et al. / International Journal of Heat and Mass Transfer 53 (2010) 14261437 1437