You are on page 1of 13

Chemical Engineering Science 59 (2004) 299–311

www.elsevier.com/locate/ces
Investigations in mass transfer enhancement in textiles with ultrasound
V.S. Moholkar

, M.M.C.G. Warmoeskerken
Textile Technology Group, Department of Chemical Engineering, University of Twente, P.O. Box 217, NL - 7500 AE Enschede, Netherlands
Received 13 November 2002; received in revised form 2 September 2003; accepted 23 September 2003
Abstract
A novel application of ultrasound is for the intensiÿcation of wet textile treatments, in which mass transfer in the inter- and intra-yarn
pores of the textile is the basic physical mechanism. This paper describes a simple methodology for the estimation of mass transfer
enhancement in ultrasonic textile treatments. For this study, washing of EMPA 101 fabric, soiled with carbon soot and olive oil, is selected
as a model process. In the absence of precise knowledge of the convection velocity resulting due to transient cavitation, a semi-empirical
method is used to estimate mass transfer enhancement. The experimental soil removal rate during model process is determined by precise
time-controlled ultrasonic treatment of the textile, with the source of cavitation nuclei located close to the textile surface. The mass
transfer in the textile during the model process is found to occur in two distinct steps, characterized by two dierent convective diusion
coecients. This eect is explained in terms of uneven soil distribution in the inter- and intra-yarn region. The mass transfer enhancement
factor, deÿned as ratio of convective diusion coecient to molecular diusion coecient of soil particles, is found to be in the range
1000–2000. In addition, it is found that the mass transfer enhancement increases with acoustic pressure amplitude during textile treatment.
A qualitative estimate of the convection velocities generated in the vicinity of the bubble is provided using numerical simulations of
bubble dynamics equation.
? 2003 Elsevier Ltd. All rights reserved.
Keywords: Cavitation; Ultrasound; Process intensiÿcation; Bubble dynamics; Acoustics; Convection
1. Introduction
Wet textile ÿnishing is deÿned as the treatment of
textile materials with chemicals in order to manipu-
late the properties of the textiles such as color, design,
hydrophobicity, etc. Wet textile ÿnishing basically involves
transport of solid/liquid substances across the textiles with
water as the medium. More precisely, mass transfer in
the inter- and intra-yarn pores of the textiles is the basic
phenomenon in wet textile ÿnishing. Current wet textile
processes suer from two major drawbacks: large process
times and low energy eciency. Improvement of these pro-
cesses basically involves intensiÿcation of mass transfer in
textile materials. More detailed discussion about mass trans-
fer in textile materials is given in Appendix A. Ultrasound
as a means of intensiÿcation of wet textile processes has
been attempted by several researchers in past few years (for

Corresponding author. Current address: “Karishma”, 2/A, Pramod
Nagar, Vijapur Road, Solapur 413 004, Maharashtra, India.
E-mail address: vmoholkar@hotmail.com (V.S. Moholkar).
example McCall et al., 1998; Thakore et al., 1988;
Yachmenev et al., 1998, 1999; Rathi et al., 1997). In-
spite of encouraging results on laboratory-scale studies, the
ultrasound-assisted wet textile processes have not been im-
plemented on industrial scale yet. Two major factors that
have contributed to this eect are: lack of precise knowl-
edge about the physical mechanism of the ultrasonic mass
transfer enhancement in textiles and inherent drawbacks
of the ultrasonic processors, such as directional sensitivity,
erosion of sonicator surface and non-uniform volumetric
energy dissipation. It was recently shown by us (Moholkar
and Warmoeskerken, 2002; Moholkar, 2002) that transient
cavitation in the medium (i.e., water) in the close vicinity
of the textile surface is the basic physical mechanism of
ultrasonic mass transfer intensiÿcation in textiles.
In this paper, we present results of further research in this
subject, i.e., quantiÿcation of the ultrasonic mass transfer
enhancement in textiles. The phenomena of transient cavita-
tion near a solid surface, such as textile, is extremely com-
plex and, hence, it is very dicult to determine exactly the
micro-convection velocity produced due to bubble motion.
For this reason, we take a semi-empirical approach, combin-
0009-2509/$ - see front matter ? 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ces.2003.09.018
300 V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311
ing the experimental results and theoretical model, for the
quantiÿcation of mass transfer enhancement.
2. Model wet textile process
So as to estimate the degree of intensiÿcation of mass
transfer in the textile with ultrasound, a model wet textile
process with a model fabric and a model diusing sub-
stance, which acts as a monitor for mass transfer, needs to
be selected. A proper study of the ultrasonic mass trans-
fer enhancement will require that the model fabric and
mass transfer monitor possess the following properties and
characteristics:
• The monitor for the transport (in the form of a particle
or a molecule) should be chemically inert.
• The monitor should not form any strong chemical or
physical bond with the ÿbers of the textile that can hinder
its transport across the fabric.
• The concentration per unit area of the monitor in the fab-
ric should be constant, so that the samples used for dif-
ferent experiments should have the same initial concen-
tration of the monitor diusing out.
• The monitor should be completely reversible. A complete
removal of the monitor fromthe textile should be possible
and the original textile should be recovered after total
extraction of the monitor.
• The original diusion rate of the monitor should be slow
enough to be “accelerated” by ultrasound. If the diusion
coecient of the monitor is already very fast, then the
acceleration due to ultrasound may not be discernible.
• The size of the monitor particles should be smaller than
the intra-yarn pores of the textile (typically 5 m or so),
that will basically ensure the presence of the monitor in
these pores. Another advantage of having a very small
size of the monitor is the separation of the secondary
eect of ultrasound that is hypothesized to be responsible
for the intensiÿcation of the textile processes, viz. particle
size reduction (Klutz, 1997). Due to very small initial
size, any further reduction in the size of the monitor will
not occur and, hence, the above-mentioned secondary
eect can be isolated.
• The fabric used in the model process should preferably be
a plain weave (with both inter- and intra-yarn porosity)
that oers a simple geometry for the transport of the
monitor.
In view of these requirements, we have selected EMPA
101 fabric (manufactured by ETH, Zurich) as the model
fabric for the experiments, with the model process being
washing of this fabric with ultrasound. This is a plain weave
cotton fabric, mass per unit area 100 g m
−2
, soiled with
carbon soot and olive oil. Its SEM images are shown in
Fig. 1. Fig. 1D shows the cross-section of the EMPA 101
fabric. The presence of the carbon particles inside the yarn
is clearly visible, although in quite small concentration. This
indicates that the soil is unevenly distributed in the inter-
and intra-yarn space.
EMPA 101 fabric satisÿes all of the requirements of the
model fabric and monitor described above except that the
carbon particles are not free to be transported by convec-
tion, but adhere to the ÿber along with the olive oil. These
particles, therefore, need to be loosened from the surface of
the ÿbers before being transported in the medium. This ne-
cessitates pretreatment of the textiles before the ultrasound
treatment such as soaking in a detergent solution. It could
be inferred from Fig. 1 that the carbon particles are very
small, and hence, an exact determination of their size is dif-
ÿcult. However, an approximate particle size that could be
assumed for the purposes of calculations would be ∼0.1 m
or so (Koster, personal communication).
3. Quantiÿcation of mass transfer enhancement: concept
of convective diusion coecient
As stated earlier, it was shown by us (Moholkar and
Warmoeskerken, 2002) that (high-energy) transient bubble
motion near the textile surface is responsible for the cre-
ation of strong convection in the close vicinity of the textile,
which enhances the mass transfer in the textile. However,
the micro-mechanism of this process is not clear. It is not
known by which mechanism, viz. shock wave, high-velocity
micro-jet or oscillatory spherical velocity ÿeld generated
due to radial bubble motion or by combination of all three
possibilities, the convection responsible for the mass trans-
fer enhancement is caused. A conventional approach to the
modeling of the mass transfer enhancement due to convec-
tion, is to solve the mass balance equation after substitution
of the uid velocity. Most of the bubble dynamics equations
available so far apply for a single bubble oscillating in an
inÿnite medium. However, in practical situations, such as
the current study, the net physical/chemical eect is a result
of multi-bubble phenomena. A comprehensive model that
describes dynamics of thousands of bubbles, with strong in-
teraction between them, is not developed yet. Therefore, an
exact determination of the convection velocity is not possi-
ble. We, therefore, adopt a semi-empirical method to esti-
mate the mass transfer enhancement factor by an order of
magnitude.
The carbon particles in the soil in the model fabric are
not soluble in the medium; however, due to their very
small size (∼0.1 m), they can be approximated as rigid
spheres diusing in the continuum of the washing medium.
For this situation, the most common basis for the deter-
mination of the diusion coecient is the Stokes–Einstein
equation
D =
«1
6¬jd
¸
. (1)
V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311 301
(A)
(B)
(C)
(D)
Fig. 1. SEM images of the model EMPA 101 fabric (SEM machine: Jeol Inc., Model GSM 5800, Accelerating voltage: 5 kV, Working distance: 10 mm):
(A) surface of the fabric; (B) a single yarn; (C) ÿber of the fabric: carbon soot on the surface of the fabric is clearly visible; (D) cross-section of a
single yarn: some carbon particles are present inside the yarn. (Cross-section of the yarn was obtained by ÿrst soaking the textile in a monomer solution
of epoxy resin, while taking care to avoid entrapment of air bubbles. It was then put in an oven, where the monomer is converted into a cross-linked
polymer with the textile entrapped in-between. Thus, in this procedure the textile was basically frozen in the resin matrix. Next, slices of the resin matrix
were obtained with a super-cut machine (Reichert Jung Inc., Model 2050)).
If the liquid, in which the particle is suspended, is moving
due to convection currents set by density or pressure dier-
ences or due to external energy input from sources such as
a mechanical stirrer, the diusion coecient of the particle
is much higher than that predicted by the Stokes–Einstein
equation. We characterize this enhanced mobility of the par-
ticle due to the bulk uid motion as convective diusion co-
ecient. The convective diusion coecient is basically a
synonym for the widely used term eddy diusion in trans-
port phenomena, which represents solute transfer due to tur-
bulent uid motion. However, due to lack of knowledge of
the exact magnitude of the uid velocity created by the tran-
sient motion of the bubbles, which will help determine the
Reynolds number for the uid motion in the close vicinity
of the textile and, hence the ow regime, we have used the
term convective diusion coecient.
3.1. Description of the fabric geometry
As described in Fig. 9 (Appendix A), a textile is com-
prised of yarns, which in turn are made of ÿbers. In order
to propose a mathematical model for the diusion process
in the textile, it is necessary to assign a suitable geometry
to the textile. For this purpose, many possibilities exist. For
example: (1) a porous plate (representing the textile as a
whole); (2) a porous cylinder (representing a single yarn in
the textile). In both the cases, a homogeneous porosity in
the structure is assumed.
In order to make a proper choice for the textile geom-
etry, one needs to take into account the relative sizes of
the source of convection (i.e., the bubbles) and the struc-
ture through which the mass transfer occurs due to the con-
vection (i.e., the textile). The typical diameter of a yarn is
302 V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311
∼200 m, as is evident from Fig. 1B. The radius of bub-
bles that undergo transient motion near the textile surface
(for 25 kHz frequency and acoustic pressure amplitudes
1.2–1.5 bar) oscillates between a few microns, and perhaps
100 m (Moholkar, 2002). An intense spherical velocity
ÿeld is created in the close vicinity of the bubble during
compression. During this phase, the porous structure, i.e., a
single yarn in the textile, tends therefore to be larger than
the dimensions of the source of convection, i.e., the bub-
ble. This circumstance justiÿes considering the textile as
a porous plate. However, while analyzing the eects of
ultrasound treatment on large portion of the textile (or on
a macroscopic scale), the dual porosity of the textile (viz.,
inter- and intra-yarn porosity) needs to be taken into account.
3.2. Approach
A ratio of the convective diusion coecient and molec-
ular diusion coecient (obtained from Stokes–Einstein
equation) gives the mass transfer enhancement factor due
to ultrasound. For the estimation of convective diusion co-
ecient, we combine experimental and theoretical methods
in three steps as follows:
1. Use an approximate method of cavitation nucleation near
the fabric surface for the experimental determination of
the rate of soil removal for dierent periods of ultrasound
irradiation. The soil concentration in the fabric can be
calculated from its reectance using the Kubelka–Munk
theory, which is discussed subsequently.
2. Solve the diusion equation for a plane sheet to determine
the concentration proÿles of the diusing substance for
dierent time intervals, and obtain an expression for the
total amount of diusing substance transported across the
plate per unit time.
3. Using experimentally measured values of soil removal
from the model fabric for dierent time periods of ul-
trasonic treatments and the theoretical expression for the
soil removal (obtained in the previous step) to ÿnd out
the convective diusion coecient for soil under the ef-
fect of ultrasound irradiation.
4. Experimental
4.1. Experimental system
The experimental system had two main components: an
experimental cell and an ultrasound horn along with a signal
generator and ampliÿer. A schematic diagram of the set-up
is shown in Fig. 2.
The experimental cell: The experimental cell was made
of three detachable glass rings and a lid made of Teon.
The height of the two rings was 15 mm, which is equal to
one-quarter of the wavelength of 25 kHz ultrasound (i.e.,
60 mm) in water, while the height of the third ring was
Signal
Generator and
Amplifier
Oscilloscope
Current and Voltage
Monitoring Unit
Ultrasound horn and
experimental cell with textile
λ
Timer
switch
Fig. 2. Schematic diagram of the experimental set-up.
60 mm. The cell was mounted on a stainless-steel bot-
tom (thickness: 51 mm) with four vertical bars that act as
support for the glass rings placed above each other. The
stainless-steel bottom of the experimental cell acts as a
rigid reector for the ultrasound waves. The cell had cali-
brated distance marks on it to measure the distance between
the ultrasound horn tip and the rigid bottom. The textile
sample could be placed in-between the glass rings. Rubber
gaskets were placed between the textile and glass rings in
order to avoid leakage. At the bottom of the cell, a special
arrangement was made to place the hydrophone (Bruel &
Kjer Ltd., Type 8103) for the measurement of the acous-
tic pressure amplitude. The output of the hydrophone was
transformed into proportional voltage by a charge ampliÿer
(Nexus Range, Model 2690). This voltage was monitored
on a digital oscilloscope (Tektronics Ltd., Model 430A).
The ultrasound unit: The ultrasound unit comprised of a
special-made ultrasound horn with a central resonance fre-
quency of 25 kHz, when vibrating in air. The horn was
driven by a signal generator (Hewlett-Packard Inc., Model
3324A) and a radio frequency ampliÿer (ENI Inc., Model
2100L). The ampliÿer could supply a maximum of 200 W
of electrical power for a large frequency range (10 kHz–
1 MHz). The output power of the ampliÿer could be con-
trolled by adjusting the voltage of the input signal to the
ampliÿer. The voltage and current supplied to the ultrasound
horn were monitored using a voltage probe (Tektronics Ltd.,
Model 6138A) and a current clamp (Farnell Inc., Model
PR-20). The ultrasound horn was mounted onto the shaft of
a laboratory jack and the experimental cell was placed on
its base. The base of the jack could be raised or lowered to
adjust the distance between the bottom of the cell and the tip
of the horn. A timer switch was added between the signal
ampliÿer output and ultrasound horn. This switch connects
the ampliÿer output to the ultrasound horn only for a speci-
ÿed time period, thus, controlling the time of the ultrasonic
textile treatment accurately.
4.2. Source of cavitation nucleation
The cavitation-nucleation in the medium is an important
factor for any cavitation-aided physical or chemical pro-
V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311 303
cess, because the physical or chemical eect depends, to a
signiÿcant extent, on the population and the size distribu-
tion of these nuclei. The time of ultrasound exposure is also
an important factor in these processes. During the process,
the initial nuclei population and their size distribution un-
dergo changes due to fragmentation of the bubbles during
radial motion and re-growth (and collapse) of the daughter
bubbles. Therefore, for ultrasound processes of longer du-
ration, as compared to the period of acoustic cycle, the ef-
fect of initial nuclei population and size distribution is likely
to be smoothed out. However, for a short ultrasound expo-
sure, the initial nuclei population and size distribution is of
crucial importance because these parameters are not likely
to change much during the exposure. Thus, for experiments
aimed at determining the kinetics of cavitation-aided phys-
ical or chemical processes, the source of cavitation nuclei,
which determines the initial population and size distribu-
tion of the cavitation nuclei, is a parameter of paramount
importance.
In case of bubble activity near a solid surface, the extent of
convection produced by the bubble depends not only on the
population and the size distribution of the cavitation nuclei
but also on their distance from the surface. Depending on
the stando factor, which is the ratio of the distance of the
bubble center from the textile and the radius of the bubble
at beginning of collapse, the bubble may undergo dierent
kinds of motions (Naude and Ellis, 1961; Tomita and Shima,
1986; Blake et al., 1986; Phillip and Lauterborn, 1998). For
large stando factors (3), the bubble shape during the ra-
dial motion remains spherical without appreciable deforma-
tion and, thus, the velocity ÿeld produced around it is also
spherically symmetric. For smaller stando factors (·1.5),
the bubble undergoes deformation producing a high-velocity
micro-jet directed towards the boundary, when the bound-
ary is suciently rigid. In addition, the distribution of the
cavitation nuclei over the surface of the textile is also an
important factor because the dynamics of a single bubble
is inuenced signiÿcantly by the interaction with adjacent
bubbles.
The nuclei close to the textile surface, before the start of
ultrasound irradiation, can create the convection for soil re-
moval almost immediately after ultrasound exposure. The
nuclei, initially away from the textile, may not help in im-
mediate removal of soil because of the rapid reduction in
the intensity of velocity ÿeld away from the center of the
bubble. However, during ultrasonic irradiation, circulatory
ows are set up in the system due to the absorption of the
momentum of the acoustic waves by the liquid. Due to these
ows, a bubble initially present away from the fabric can
migrate towards the fabric surface (if it does not collapse
before reaching the fabric), and create nucleation near tex-
tile that could assist the soil removal. However, in this case,
the rate of soil removal is determined by the time for bub-
ble migration and the intensity of convection created by the
bubble. This obviously can lead to signiÿcant errors in the
calculations of rate of soil removal, especially for smaller
periods of ultrasound exposure. Thus, the choice of source
of nuclei is of paramount importance in the accurate mea-
surement of rate of soil removal.
Conventionally, a suspension of polystyrene latex parti-
cles is used to provide nuclei in the medium (e.g., Holland
and Apfel, 1990). The air pockets entrapped in this suspen-
sion can provide nuclei for cavitation in the medium. How-
ever, these air pockets are distributed all over the medium
along with the polystyrene particle suspension. Therefore,
for the present study with very short periods of ultrasound
exposure, the nucleation by polystyrene latex suspension
may not give good results, for the reasons explained earlier.
Instead, particles with surface crevices that can entrap air
pockets, placed on the fabric surface itself, can provide nu-
cleation at desired location for production of instantaneous
convection with ultrasound exposure without any time-lag
eects arising out of factors just discussed.
Pumice stones, which are widely used for providing nuclei
for boiling, can be used for the present purpose. Since these
stones are commercially available in large sizes (3 mm
or so), they may give rise to non-uniform nucleation. How-
ever, by reducing their sizes by grinding, restricting the size
range after screening using sieves of certain mesh sizes, and
adding an exact quantity of stones to the medium each time,
this problem could be overcome to some extent. It must be
noted that having uniform particle sizes may not be of much
use as far as the uniformity of nucleation is concerned, be-
cause nuclei for cavitation are mainly contributed by the
gas pockets trapped in surface crevices of the particles, and
not by the particles themselves. In view of the above con-
siderations, we have used pumice stones of size range of
200–400 m as a source of nuclei in the present experi-
ments. If added in large amounts, the pumice stones can al-
ter the power consumption of the system due to reduction in
system impedance as a result of large bubble population in
the system that increases the compressibility of the medium
(Moholkar, 2002). Therefore, we have used a very small
amount (20 mg) of pumice stones for the nucleation in each
experiment. For these conditions, it was found that the power
consumption of the ultrasound horn remained unchanged.
4.3. Experimental procedure
4.3.1. Characterization of the ultrasound ÿeld
According to the analysis of the pressure and velocity
variation in the standing wave ÿeld given by Pierce (1989),
the locations of pressure minimum occur at a distance of
z}4, 3z}4, 5z}4, etc., while the locations of pressure max-
imum occur at a distances of z}2, z, 3z}2, etc., from the
rigid surface, for plane-wave reection from it. In order to
conÿrm this theoretical result, pressure measurements were
done in the experimental cell using a very small (4 mm) un-
calibrated hydrophone at two locations viz., at distances z}4
and z}2 from the rigid bottom of the cell with degassed wa-
ter as the medium. The distance between the tip of the horn
304 V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311
8
4
0
8
4
0
F
o
u
r
i
e
r

C
o
e
f
f
i
c
i
e
n
t
s
F
o
u
r
i
e
r

C
o
e
f
f
i
c
i
e
n
t
s
0 10 20 30 40 50 60 40 80 90 100
Frequency [kHz]
0 10 20 30 40 50 60 40 80 90 100
Frequency [kHz]
(A)
(B)
Fig. 3. Characterization of the standing wave ÿeld generated in the
experimental cell: (A) acoustic spectrum at pressure antinode; (B) acoustic
spectrum at pressure node.
and the rigid bottom of the experimental cell was adjusted
to 60 mm, which is the wavelength of 25 kHz ultrasound
in water. The ultrasound horn was driven at 25 kHz with a
power input of 20 W. For this power input, it produced an
acoustic wave with pressure amplitude ∼1.3 bar, which is
above the transient cavitation threshold for bubbles of sizes
5–10 m at 25 kHz frequency (Moholkar, 2002). Fig. 3
shows the FFT of the pressure signals (duration: 12 ms;
sampling frequency: 2.5 MHz) obtained at distances z}4 and
z}2 from the rigid reector. The amplitude of the funda-
mental peak in the acoustic spectrum corresponding to the
driving frequency (indicative of the acoustic pressure am-
plitude) at location z}2 is much higher than the amplitude
of the corresponding peak at location z}4. In addition, the
subharmonic peak, indicative of transient cavitation, is quite
prominent in the acoustic emission spectrum at location
z}2, while it is completely absent in the acoustic emission
spectrum at location z}4. Thus, the characterization of the
acoustic wave ÿeld in the experimental cell approximately
conÿrms the theoretical prediction of the location of the pres-
sure node and the pressure antinode. Another feature of the
acoustic wave phenomena, which is clear from these results
is that the resultant acoustic pressure amplitude at z}4 is not
zero, but has some positive residual value. A possible reason
for this is that due to presence of small bubble fraction the
wavelength of sound is not exactly 60 mm, but somewhat
less, which causes a shift in the position of pressure node.
4.3.2. Washing experiments
Investigations of Moholkar and Warmoeskerken (2002)
revealed that the ultrasonic washing eciency depends on
the gas content of the fabric and the washing medium.
A maximum washing eciency (indicative of maximum
mass transfer) was obtained with a degassed washing
medium and non-degassed fabric, with the fabric positioned
at a pressure antinode. In this study, we use the same condi-
tions for the experiments. In this case, no cavitation nuclei
are contributed by the medium. The circular grooves in the
tip of the horn, which can entrap small amount of air that
can provide alternate source of nuclei in the medium, were
ÿlled with silicon rubber. Hence, pumice stones are the only
source of cavitation nuclei in the system. Eight speciÿc time
intervals ranging from 0.1 to 2 s were selected for ultra-
sound irradiation. For each interval, three experiments were
conducted with dierent pieces of the model fabric to assess
reproducibility of results. Power input to the ultrasound
horn was 20 W at 25 kHz frequency. At this power input,
the ultrasound horn produces an acoustic wave with a pres-
sure amplitude of ∼1.3 bar. The model fabric was soaked
in the detergent solution (1.75 g l
−1
of sodium dodecyl
benzene sulfonate) for 5 min before ultrasonic treatment.
The model fabric was then ÿxed between the glass rings of
the experimental cell in a water bath to avoid entrapment of
air beneath the fabric. In each experiment, 250 ml degassed
water was used as the washing medium. The dissolved oxy-
gen content of the washing medium was lowered to 2 ppm
using a chemical method (van der Vlist et al., 1994). The
time of ultrasound treatment was 3 min. After ultrasound
treatment, the model fabric was removed from the cell and
was dried in air. In order to assess the eect of the acoustic
pressure amplitude on the convective diusion coecient,
the set of experiments at 20 W power input was repeated
at power input of 15 W, for which the ultrasound horn
produces a wave with pressure amplitude of ∼1.1 bar.
4.4. Data analysis
The quantiÿcation of the washing eect can be done by
measuring the reectance of the ultrasound-treated textile
sample. In order to determine the rate of soil removal, we
need to convert the reectance measurements into soil con-
centrations. For this purpose, Kubelka–Munk theory is of-
ten used, which is the simplest tool for the description of
the optical properties of a turbid medium that absorbs and
scatters light (Kubelka and Munk, 1931). The mathemati-
cal expression of the Kubelka–Munk theory that relates the
absorption coecient (K), scattering coecient (S) and the
reection coecient of the layer (R
[
) is
K
S
=
(R
[
−1)
2
2R
[
. (2)
For particles larger than the wavelength of light, the scat-
tering coecient is inversely proportional to the diameter.
If the distribution of particle diameters is very small, then
one can approximate: S∼1}d
¸
. The Kubelka–Munk theory
can be coupled with the Lambert–Beer law to convert the
remission function into absolute concentration. In case of
negligible scattering coecient, the relationship between
V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311 305
absorption coecient K and concentration of soil C is
K = cC ln(10) = c
n
C, where c
n
= c ln(10). (3)
We make two simplifying assumptions: (1) the diameters of
all carbon particles in the soil over textile are the same; (2)
the scattering coecient of carbon particles is negligible.
In the present study, it is not possible to determine c
n
,
since the manufacturer of EMPA 101 model fabric does
not supply the soil separately. Therefore, the soil concentra-
tions need to be deÿned on a relative basis. The reectance
of untreated EMPA 101 textile is 16%, which gives K}S
value of 2.205, corresponding to 100% soil concentration
in fabric. The reectance of completely washed EMPA 101
fabric is ∼70%, which gives K}S = 0.064, corresponding
to 0% soil concentration in fabric (note that scattering co-
ecient is cancelled out in the calculations of relative soil
concentrations due to assumption 1). The fraction of origi-
nal soil removed from the fabric after ultrasound treatment
for dierent time periods (M
t
) is estimated by assuming a
linear relation between K and M
t
with: K =2.205 at M
t
=0;
K = 0.064 at M
t
= 1, respectively.
5. The mathematical model
As discussed earlier, we approximate the geometry of the
textile as a porous plate. One-dimensional unsteady state dif-
fusion (in absence of velocity) in a porous plate is described
by Fick’s second law
cC
ct
= D
c
2
C
cx
2
. (4)
Let −!·x·! denote the boundaries of the plate with thick-
ness 2!. The solutions of Eq. (4) are listed by Carslaw
and Jaeger (1959) for dierent boundary conditions. For the
conditions
C = C
1
, for x =−!, !,
C = C
0
, for t = 0,
the concentration of the diusing substance in the plate is
(Crank, 1975)
C −C
1
C
0
−C
1
= 1 −
4
¬

n=0
(−1)
n
2n + 1
exp
_
−D(2n + 1)
2
¬
2
t
4!
2
_
×cos
_
(2n + 1)¬x
2!
_
. (5)
The total amount of substance, which has left the plate after
a time t(M
t
), and the corresponding amount after inÿnite
time (M

) are related by
M
t
M

= 1 −

n=0
8
(2n + 1)
2
¬
2
exp
_
−D(2n + 1)
2
¬
2
t
4!
2
_
. (6)
Concentration proÿle for Fourier numbers 0.1: It is
convenient to introduce the Fourier number deÿned by
Fo =
Dt
!
2
. (7)
For times such that Fo0.1, the concentration proÿle
can be approximated by a more convenient expression. The
detailed theoretical analysis of this (which is an extension
of the penetration theory) can be found in several references
such as Beek et al. (1999) and Janssen and Warmoeskerken
(1997). We give herewith only the ÿnal result. For longer
contact times, the average concentration in the porous plate
is
C
1
−C
C
1
−C
0
=
8
¬
2
exp(−¬
2
Fo). (8)
This means that plot of logarithm of the expression on LHS
of Eq. (8) against Fourier number should yield a straight
line with slope ∼−9.86.
5.1. Bubble motion near the fabric
The bubble motion near a solid boundary is a multi-faceted
phenomenon. The kind of radial motion the bubble under-
goes in the vicinity of the solid boundary, whether symmet-
ric or asymmetric, depends on the physical characteristics
of the boundary (such as elastic modulus) and also on the
stando factor.
Recently, Brujan et al. (2001) have reported high-speed
imaging studies of a vapor bubble dynamics near an elastic
boundary. They conclude that for boundaries with very low
elastic modulus, the bubble motion is asymmetric (with for-
mation of a jet) only for stando factors smaller than 1. For
cases where stando factor ¿1, the bubble motion remains
symmetric as if occurring in an inÿnite uid. Cotton ÿbers
have a very low static elastic modulus (Morton and Hearle,
1997). It is likely that, at 25 kHz, the elastic modulus is still
relatively small and, therefore, the cotton fabric acts as a soft
boundary for a bubble oscillating in its vicinity. Thus, we
present simulations of the bubble motion and the spherical
velocity ÿeld generated around it using the Gilmore bubble
dynamics equation (Gilmore, 1954)
R
_
1 −
U
c
_
d
2
R
dt
2
+
3
2
_
1 −
U
3c
_ _
dR
dt
_
2
=
_
1 +
U
c
_
H +
U
c
_
1 +
U
c
_
dH
dR
. (9)
H is the free enthalpy on the surface of the bubble
H =
n
n −1
A
1}n
j
0
_
_
_
_
_
P
0
+
2o
R
0
__
R
0
R
_


2o
R
+ B
_
(n−1)}n
−[P

+ B]
(n−1)}n
_
_
_
. (10)
A, B and n are constants (For water, A = 3001 atm,
B = 3000 atm and n = 7). P

, the pressure in the bulk
liquid driving the bubble motion is
P

= P
0
−P
A
sin(2¬[t). (11)
306 V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311
The velocity in the bulk liquid at a distance r fromthe bubble
center is
u(r, t) =
R
2
r
2
_
dR
dt
_
. (12)
5.2. Adoption of the diusion model for the present study
While adopting the above models to explain the results
of the present experiments, we need to take into consider-
ation the possible discrepancies between the experimental
and theoretical results, which could occur as a result of mis-
matches between the assumptions made in the model and
the practical situation. One such mismatch is the geometrical
dierence between the model porous plate and the textile.
As discussed earlier, the textile is modeled as a porous plate
on the basis of the dierences between sizes of an individ-
ual bubble and that of an individual yarn. However, during
the experiments, we will be determining the rate of soil re-
moval from the textile on a macroscopic scale. We will be
quantifying the soil transfer rates by measurement of the re-
ectance of the textile. The reectance is measured over a
large area of the textile (and not for a single yarn). There-
fore, this measurement technique will quantify the soil trans-
fer not only in the intra-yarn pores but also in the inter-yarn
pores. The reectance of the fabric is a function of the av-
erage soil concentration in the textile. Images of the EMPA
101 model fabric shown in Fig. 1 reveal that the soil is dis-
tributed unevenly over the textile surface, with the major
fraction of the soil present in the inter-yarn pores. As dis-
cussed in Appendix A, due to geometrical constraints, the
intensity of the convection resulting from the radial bub-
ble motion in the inter-yarn region is much higher than in
the intra-yarn region. Thus, the soil removal process is ex-
pected to have two steps, with dierent soil removal rates:
(1) rate for smaller time period of ultrasound irradiation
during which only inter-yarn soil is removed; (2) rate for
longer time period of ultrasound irradiation during which
both inter- and intra-yarn soil is removed. It is obvious that
soil removal rate for step 1 is higher than step 2. This could
give rise to a dierence between the theoretical results pre-
dicted by the models with assumption of constant diusion
coecient with homogeneous porosity in the sheet and the
experimental results.
6. Results and discussion
The results of the experiments done with 20 and 15 W
power input to the ultrasound horn are shown in Figs. 4 and
5, respectively. It can be seen that the washing eect for
both power inputs shows a gradual rise with time of ultra-
sound irradiation. However, for shorter ultrasound irradia-
tion the rate of soil removal from the fabric is higher than for
longer ultrasound treatments. This result may perhaps be in-
terpreted from the discussion given in previous section. The
0.5
0.4
0.3
0.2
0.1
0.0
0 0.5 1 1.5 2
Time (sec)
R
e
f
l
e
c
t
a
n
c
e
0.00
0.25
0.50
0.75
1.00
F
r
a
c
t
i
o
n

o
f

o
r
i
g
i
n
a
l

s
o
i
l

r
e
m
o
v
e
d
0 0.5 1 1.5 2
Time (sec)
(A)
(B)
Fig. 4. Experimental results for fabric treatment at 20 W power input:
(A) reectance of the fabrics treated for dierent time periods; (B) soil
removal from the textile with time.
0.00
0.10
0.20
0.30
0.40
R
e
f
l
e
c
t
a
n
c
e
0.00
0.25
0.50
0.75
1.00
F
r
a
c
t
i
o
n

o
f

o
r
i
g
i
n
a
l

s
o
i
l

r
e
m
o
v
e
d
0 0.5 1 1.5 2
Time (sec)
0 0.5 1 1.5 2
Time (sec)
(A)
(B)
Fig. 5. Experimental results for fabric treatment at 15 W power input:
(A) reectance of the fabrics treated for dierent time periods; (B) soil
removal from the textile with time.
results are consistent with the hypothesis that, for shorter ul-
trasound irradiation, only the soil on the fabric surface and
in the intra-yarn pores is removed, while for longer irradia-
V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311 307
0.00
0.25
0.50
0.75
1.00
F
r
a
c
t
i
o
n

o
f

o
r
i
g
i
n
a
l

s
o
i
l

r
e
m
o
v
e
d
Experimenal D = 7.5e-10 D = 2.0e-9
0.00
0.25
0.50
0.75
1.00
F
r
a
c
t
i
o
n

o
f

o
r
i
g
i
n
a
l

s
o
i
l

r
e
m
o
v
e
d
D = 5e-10 D = 1e-9 Experimental
0 0.5 1 1 .5 2
Time (sec) (A)
(B)
0 0.5 1 1 .5 2
Time (sec)
Fig. 6. Features of mass transfer in fabric under ultrasonic treatment: (A)
20 W power input; (B) 15 W power input. (Values of D indicate diusion
coecient in m
2
s
−1
for theoretical calculation of the soil removal curve
using Eq. (6).)
tion soil in the intra-yarn pores is also removed. However,
since the intra-yarn soil forms only a small fraction of the
total soil in the textile, the reectance of the fabric does not
rise signiÿcantly for longer ultrasound irradiation. A com-
parison of Figs. 4B and 5B reveals that larger power input,
which means cavitation at larger pressure amplitude, results
in faster and higher removal of soil from the fabric.
Some experimental errors, however, need to be taken into
account. In Fig. 5A, the reectance of the fabric does not
show a consistent rise with time of irradiation. Between 0.5
and 1 s, it decreases slightly, indicating lesser soil removal
in 1 s ultrasound irradiation than that for 0.5 s. This is in
contradiction with the theoretical speculation that soil re-
moval should increase with time of ultrasound irradiation.
In addition to duration of ultrasound irradiation, the soil re-
moval from the fabric depends on several other factors such
as size distribution and population of bubbles, the location
of bubble from fabric surface. Since we used a crude method
of providing cavitation nucleation near textile surface, it is
likely that the above-mentioned factors varied signiÿcantly
in consecutive experiments. Thus, we attribute the anomaly
of lesser fabric reectance for 1 s ultrasound treatment than
0.5 s to the error in having exactly same kind of nucleation
produced near the textile surface in consecutive experiments.
A comparison of the theoretical and experimental soil re-
moval rate is shown in Fig. 6A and B for 20 and 15 W
power input, respectively. It can be seen that a single value
of the diusion coecient is inadequate to reproduce the
data. To the extent that an “eective diusion coecient”
y = -0.4999x - 0.4671
-1.60
-1.20
-0.80
-0.40
0.0 0.5 1.0 1.5 2.0 2.5
Time (sec.)
Time (sec.)
l
n

(
C
)
y = -0.1963x - 0.5805
-1.2
-0.8
-0.4
0
0.0 0.5 1.0 1.5 2.0 2.5
l
n

(
C
)
(A)
(B)
Fig. 7. The logarithmic average soil concentration for longer contact time
(Eq. (8)): (A) soil concentration for 20 W power; (B) soil concentration
for 15 W power. The slope of the line (= − ¬
2
D}!
2
) indicates the
convective diusion coecient.
can be deÿned, the data imply that it decreases with time.
The initial portions of the curves are described reasonably
well by Eq. (6) with D = 2 × 10
−9
m
2
s
−1
at 20 W and
D = 1 × 10
−9
m
2
s
−1
at 15 W. Thus, the diusion coef-
ÿcient is found to increase with the acoustic pressure am-
plitude, as expected. If the process can be described by a
superposition of two dierent diusive processes, charac-
terized by dierent diusion coecients, at large times the
solution will be dominated by the slower diusive process
and, therefore, one can estimate this smaller diusion co-
ecients from the large-time behavior of the data. On the
basis of this argument, one deduces from the data the values
D = 7.5 ×10
−10
m
2
s
−1
at 20 W and 5 ×10
−10
m
2
s
−1
at
15 W. The shift between the two mass transfer rates occurs
at approximately 0.4 s (referring to Fig. 6A), which corre-
sponds to a Fourier number of ∼0.03 as calculated with the
smaller diusion coecient. A similar analysis for Fig. 6B,
where the shift between mass transfer rates occurs at ∼0.6 s,
reveals that the Fourier number at which this shift occurs
also is 0.03.
Fig. 7 shows plots of variation in average soil concentra-
tion in the fabric for longer contact time using Eq. (8). The
plots are shown for t0.5 s for 20 W, and for t0.6 s for
15 W power input. The value of convective diusion coe-
cient obtained from the slopes of the plots, assuming fabric
thickness of 100 m, are 5 × 10
−10
m
2
s
−1
for 20 W and
3×10
−10
m
2
s
−1
for 15 W. These values fairly match with
308 V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311
4
2
0
5
2
-5
2.5
1.25
0
1.5
0
-1.5
0 1 2 3 4 5
Time (No. of Acoustic Cycles)
0 1 2 3 4 5
Time (No. of Acoustic Cycles)
0 1 2 3 4 5
Time (No. of Acoustic Cycles)
0 1 2 3 4 5
Time (No. of Acoustic Cycles)
V
e
l
o
c
i
t
y

(
m
/
s
)
V
e
l
o
c
i
t
y

(
m
/
s
)
R
a
t
i
o

(
R
/
R
0
)
R
a
t
i
o

(
R
/
R
0
)
(A)
(B)
Fig. 8. Simulations of radial dynamics and spherical velocity ÿeld for
oscillations of 10 m bubble near textile surface. The distance of bubble
center from textile: 50 m, frequency: 25 kHz: (A) acoustic pressure
amplitude = 1.3 bar; (B) acoustic pressure amplitude = 1.1 bar.
the values of convective diusion coecients given in Fig. 6,
calculated from Eq. (6).
Mass transfer enhancement: Assuming that the carbon
particles in the soil on model fabric are spherical and have
size ∼0.1 m, the molecular diusion coecient obtained
from Stokes–Einstein equation is ∼3 × 10
−12
m
2
s
−1
.
This diusion coecient needs further correction for the
porosity and tortuosity of the fabric. The corrected diu-
sion coecient, per deÿnition given in Appendix A, is:
3 × 10
−13
m
2
s
−1
. The ratio of convective diusion coef-
ÿcient and the molecular diusion coecient indicates that
the enhancement factor is in the range 1000–2000.
As mentioned earlier, a direct estimation of the convec-
tion velocities near the textile surface is not possible in the
present experiments. However, simulations of the bubble dy-
namics equation can give us an order of magnitude estimate
of the convection velocities. Fig. 8 shows simulations of the
radial dynamics and the spherical velocity ÿeld of a 10 m
bubble driven at dierent acoustic pressure amplitudes. The
other parameters for the simulation are given in the ÿgure
caption. It can be seen that the bubble creates an oscillatory
velocity ÿeld around it with dierent magnitudes of the ve-
locities in the two directions: towards and away from the
bubble center. The (absolute) magnitude of the velocity ÿeld
in any direction and, hence, the overall convection gener-
ated due to radial bubble motion increases with the acoustic
pressure amplitude. A typical estimate of the time scale of
particle motion will be t ≈ d
2
¸
}D. With substitution of the
representative values d
¸
∼0.1 m and D∼1×10
−13
m
2
s
−1
,
we ÿnd t = 0.1 s. From Fig. 8, it could be perceived that
the time scale of radial motion of the bubble is same as that
of the acoustic wave period, which is 40 s for 25 kHz fre-
quency. Therefore, the eective motion of the particle re-
sulting due to convection created by a single radial bubble
motion is negligible. The net motion of the soil particle un-
der the convection created by the bubble motion is, thus, a
function of the time-averaged velocity over several acoustic
cycles. The estimation of this velocity is highly complicated
because it will not only be a function of the bubble popula-
tion but also of the spatial distribution of the bubbles from
the textile surface and the strong interaction between adja-
cent bubbles. Since the purpose of the simulations presented
here is to give an approximate estimate of the convection
created by the radial bubble motion, we have not incorpo-
rated such complications in our analysis.
7. Conclusion
This paper describes a semi-empirical methodology for
a quantitative estimate of the mass transfer enhancement in
the ultrasonic textile treatments. In absence of precise quan-
tiÿcation of the convection created due to transient bub-
ble motion near the textile surface, we deÿned a convective
diusion coecient, which is representative of the mass
transfer enhancement due to ultrasound. The geometry of
the textile was modeled as a porous plate. The experimen-
tal results were correlated to the theoretical values of soil
removal with time using solutions of the diusion equa-
tion. The soil removal from the textile showed two distinct
regimes with two dierent convective diusion coecients.
This result was explained in terms of the non-uniform ini-
tial distribution of soil in the textile and the dual porosity of
the textile; inter- and intra-yarn porosity. Textile treatments
at higher acoustic pressure amplitudes were found to give
higher diusion coecients. The order of magnitude of the
mass transfer enhancement for the model wet textile pro-
cess, which was deÿned as the ratio of convective diusion
coecient to molecular diusion coecient, is in the range
1000–2000. This methodology can also be applied for other
ultrasonic wet textile treatments; however, the mass trans-
V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311 309
fer enhancement factor may change from system to system
depending on the properties of the textiles and those of the
ultrasound system. We hope that this study will contribute
towards the design and scale-up of ultrasonic wet textile
processes.
Notation
c velocity of sound, m s
−1
C concentration of diusing substance, kg m
−3
or kmol m
−3
d
[
diameter of a ÿber, m
d
¸
particle radius, m
d
,
diameter of a yarn, m
D diusion coecient, m
2
s
−1
D
e
eective diusion coecient, m
2
s
−1
[ frequency of the acoustic wave, Hz
H free enthalpy on the bubble surface, J kg
−1
k
1
mass transfer coecient for intra-yarn
region, m s
−1
k
2
mass transfer coecient for inter-yarn
region, m s
−1
k
3
mass transfer coecient for the boundary
layer, m s
−1
K absorption coecient, m
−1
K
inter-yarn
permeability between the yarns, m
2
K
intra-yarn
permeability between the ÿbers, m
2
K
0
Kozeny’s constant, dimensionless
K
0
Kozeny’s constant, dimensionless
K
1
overall mass transfer coecient in the textile,
m s
−1
! half-thickness of the porous plate, m
M

amount of diusing material transported in
inÿnite time, kg or kmol
M
t
amount of diusing material transported in
time t, kg or kmol
P
A
pressure amplitude of the acoustic wave, Pa
P

pressure in the medium at inÿnity, Pa
P
0
ambient (atmospheric) pressure, Pa
r radial coordinate, m
R bubble radius, m
R
[
reection coecient, dimensionless
R
0
initial bubble radius, m
S scattering coecient, m
−1
t time, s
1 temperature of the medium, K
u velocity in the medium, m s
−1
U bubble wall velocity, m s
−1
t mean uid velocity, m s
−1
x distance coordinate, m
Greek letters
: tortuosity factor, dimensionless
c
[
porosity between the ÿbers, dimensionless
c
n
Naperian molar absorption coecient,
dimensionless
c
,
porosity between the yarns, dimensionless
¸ polytropic constant of the bubble contents,
dimensionless
« Boltzmann constant, J K
−1
z wavelength of the acoustic wave, m
j viscosity of the medium, Pa s
j
L
density of liquid, kg m
−3
j
0
density of the liquid in undisturbed state,
kg m
−3
o surface tension, N m
−1
Acknowledgements
The authors are grateful to Stork Brabant B.V., Nether-
lands for the funding of the project at the University of
Twente.
Appendix A. Transport phenomena in textile materials
A.1. Fluid ow through the fabric
For a uid owing through the textile, the inter- and
intra-yarn regions form two parallel paths. The relative con-
tribution of the ows in these two regions can be judged by
the ratio of the permeabilities of these regions. The perme-
abilities of the inter- and intra-yarn regions can be written
as (Gooijer, 1998)
K
intra-yarn
=
1
16K
0
c
3
[
(1 −c
[
)
2
d
2
[
, (A.1)
K
inter-yarn
=
1
16K
0
c
3
,
(1 −c
,
)
2
d
2
,
. (A.2)
We make an assumption of constant K
0
for the inter- and
intra-yarn pores on basis of similar porosities of an individ-
ual yarn and the overall textile (Gooijer, 1998). The ratio
of Eqs. (A.1) and (A.2) with representative values of the
parameters: c
[
=0.35 −0.45; c
,
=0.4 −0.5; d
,
=250 m;
d
[
= 15 m gives
K
intra-yarn
K
inter-yarn
=
1
200
to
1
2000
. (A.3)
The ratio of K
intra-yarn
and K
inter-yarn
indicates that the per-
meability between the yarns is far higher than the perme-
ability in the yarns. Therefore, most of the ow will follow
the path of least resistance (between the yarns), and practi-
cally no ow will occur through the yarns. The ow through
the inter-yarn region may, however, penetrate an individ-
ual yarn to a small extent. This can create a thin convective
shell near the periphery of the yarn. Nonetheless, the cen-
tral core of the yarn still remains a stagnant core zone with
no ow inside. Therefore, the mass transfer in this region
occurs only by the slow process of diusion.
310 V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311
TEXTILE
YARNS
FIBRES
INTRA YARN
PORE
INTER YARN PORE
Convective shell
in the yarn
Stagnant
Intra-yarn
region
Boundary layer
YARN BULK
C
r
Inter-yarn region
(A)
(B)
Fig. 9. The mass transfer mechanism in the textiles: (A) schematic
diagram indicating the general structure and dual porosity of the textiles;
(B) the three stages of mass transfer in textile materials and the typical
concentration proÿles of the diusing substance.
A.2. Mass transfer in the fabric
For washing, the mass transfer in these textile materials
can be divided into three distinct steps, which occur in series,
as shown in Fig. 9:
1. mass transfer from the stagnant intra-yarn region to the
inter-yarn region through the convective shell on the pe-
riphery of the yarn;
2. mass transfer from the inter-yarn region to the boundary
layer between the bulk liquid and the textile surface;
3. mass transfer from the boundary layer to the bulk liquid.
In order to assess the relative contribution of each of these
steps to the overall mass transfer process, one needs to have
an insight into the nature of uid ow through the textiles.
In order to ÿnd the overall mass transfer coecient (K
1
) in
the textile, we specify k
1
, k
2
and k
3
as the individual mass
transfer coecients for the ÿrst, second and the third step of
mass transfer mentioned above. K
1
can be determined using
familiar relationship in the mass transfer theory
1
K
1
=
1
k
1
+
1
k
2
+
1
k
3
. (A.4)
We make a simplifying assumption that the mass transfer
in the third step (from the boundary layer to the bulk) is
very fast compared to the other steps. Now to estimate the
relative magnitudes of k
1
and k
2
we use the mass transfer
correlations developed for a packed bed. Several correlations
for the mass transfer coecients are listed by Janssen and
Warmoeskerken (1997). A single yarn can approximately
be described as a cylinder. To estimate the inter-yarn mass
transfer coecient, we choose the following correlation for
the ow past long cylinders perpendicular to the ow:
Sh = 0.42Sc
1}5
+ 0.57Re
1}2
Sc
1}3
, 1·Re·10
4
,
Sc0.7, Pe 1. (A.5)
The non-dimensional numbers are deÿned as
Re =
j
L
td
,
j
, Sh =
kd
,
D
, Pe =
td
,
D
. (A.6)
To estimate the inter-yarn mass transfer coecient, we
substitute the following representative values in the above
correlation: t=2 cm s
−1
; d
,
=250 m; j
L
=1000 kg m
−3
;
D = 1 × 10
−9
m
2
s
−1
; c = 0.5 and Sc ≈ 1000. Substitut-
ing the above values, we ÿnd that Sh ≈ 25 and from the
deÿnition of the Sherwood number it follows that k
2

1.4×10
−2
cm s
−1
. To estimate the mass transfer coecient
inside the yarn, we make use of the analysis given by Matsui
et al. (1978), who have shown that for longer contact times
[Dt}d
2
,
] the Sherwood number ∼5.8. Due to the porous
structure of the yarns, the eective diusion coecient in-
side the yarn is smaller than the actual diusion coecient.
Therefore, we use the correction given by Rietema (1976)
to estimate the eective diusion coecient in a porous
medium
D
e
=
c
,
:
2
D. (A.7)
Substituting representative values of c
,
and : as 0.4 and 2,
respectively, gives D
e
≈ 1×10
−10
m
2
s
−1
. The mass trans-
fer coecient inside the yarn (k
1
) is then calculated from
the Sherwood number as ≈ 4 × 10
−4
cm s
−1
. The ratio of
k
1
and k
2
gives us an idea of the mass transfer process in the
textiles: 97% of the mass transfer resistance is determined
by the diusion process inside the yarns. As such, the diu-
sional mass transfer in the yarn is the rate-controlling step in
the overall mass transfer in the textile. This also implies that
enhancement of the mass transfer in the textile would ne-
cessitate conversion of the intra-yarn diusion process into
the faster convection.
References
Beek, W.J., Muttzall, K.M.K., van Beek, J.W., 1999. Transport
Phenomena. Wiley, Chichester.
Blake, J.R., Taib, B.B., Doherty, G., 1986. Transient cavities near
boundaries. Part 1. Rigid boundary. Journal of Fluid Mechanics 170,
479–497.
V.S. Moholkar, M.M.C.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299–311 311
Brujan, E.-A., Nahen, K., Schmidt, P., Vogel, A., 2001. Dynamics of
laser-induced cavitation bubbles near elastic boundaries: inuence of
elastic modulus. Journal of Fluid Mechanics 433, 283–314.
Carslaw, H.S., Jaeger, J.C., 1959. Conduction of Heat in Solids. Clarendon
Press, Oxford.
Crank, J., 1975. The Mathematics of Diusion. Clarendon Press, Oxford.
Gilmore, F.R., 1954. Hydrodynamic Laboratory Report, 26–4, California
Institute of Technology.
Gooijer, H., 1998. Flow resistance of textile materials. Ph.D. Thesis,
University of Twente.
Holland, C.K., Apfel, R.E., 1990. Thresholds for transient cavitation
produced by pulsed ultrasound in a controlled nuclei environment.
Journal of the Acoustical Society of America 88, 2059–2069.
Janssen, L.P.B.M., Warmoeskerken, M.M.C.G., 1997. Transport
Phenomena Data Companion. Delft University Press, Delft.
Klutz, D.S., 1997. Fundamental interactions of ultrasound with cotton
ÿbers and vat dyes. Ph.D. Thesis, North Carolina State University.
Kubelka, P., Munk, F., 1931. Ein beitrag zur optik der farbanstriche.
Zeitschrift fur Technische Physik 12, 593–601.
Matsui, H., Kobayashi, M., Koji, K., 1978. Washing of fabrics Part 1:
Washing of cotton fabrics and its theoretical analysis. Journal of the
Textile Machinery Society of Japan 24 (2), 41–49.
McCall, R.E., Patel, F.M.A., Mock, G.N., Grady, P.L., 1998. Solvent
and ultrasonic alternatives to perchloroethylene drycleaning of textiles.
Textile Chemist and Colorist 30 (11), 11–18.
Moholkar, V.S., 2002. Intensiÿcation of Textile Treatments: Sonoprocess
Engineering. Twente University Press, Enschede.
Moholkar, V.S., Warmoeskerken, M.M.C.G., 2002. Mechanistic Aspects
and optimization of ultrasonic washing. AATCC Review 2 (2),
34–37.
Morton, W.E., Hearle, J.W.S., 1997. Physical Properties of Textile Fibers.
The Textile Institute, Manchester.
Naude, C.F., Ellis, A.T., 1961. On the mechanism of cavitation damage by
non-hemispherical cavities collapsing in contact with a solid boundary.
Journal of Basic Engineering (Transaction of ASME-D) 83, 648–656.
Phillip, A., Lauterborn, W., 1998. Cavitation erosion by single laser
produced bubbles. Journal of Fluid Mechanics 361, 75–116.
Pierce, A.D., 1989. Acoustics: An Introduction to its Physical Principals
and Applications. Acoustical Society of America, New York.
Rathi, N.H., Mock, G.N., McCall, R.E., Grady, P.L., 1997. Ultrasound
aided open width washing of mercerized 100% cotton twill fabric,
AATCC International Conference and Exhibition 1997, Atlanta, GA,
Book of Papers, pp. 254–262.
Rietema, K., 1976. Fysische Transport en Overdrachtsverschijnselen. Het
Spectrum, Utrecht.
Thakore, K.A., Smith, C.B., Hite, D., Carlough, M., 1988. The eects
of ultrasound and the diusion coecient of C.I. Direct Red 81 in
cellulose. American Dyestu Reporter 77 (10), 15–16.
Tomita, Y., Shima, A., 1986. Mechanisms of impulsive pressure
generation and damage pit formation by bubble collapse. Journal of
Fluid Mechanics 169, 535–564.
van der Vlist, P., Warmoeskerken, M.M.C.G., Willemse, S., 1994.
European Patent No. EP9401241.
Yachmenev, V.G., Blanchard, E.J., Lambert, A.H., 1998. Use of
ultrasound energy in the enzymatic treatment of cotton fabric. Industrial
and Engineering Chemistry Research 37 (10), 3919–3923.
Yachmenev, V.G., Blanchard, E.J., Lambert, A.H., 1999. Study of the
inuence of ultrasound on enzymatic treatment of cotton fabric.
Textile Chemist and Colorist and American Dyestu Reporter 1 (1),
47–51.

If the di usion coe cient of the monitor is already very fast. the net physical/chemical e ect is a result of multi-bubble phenomena. It is not known by which mechanism. However. hence. Another advantage of having a very small size of the monitor is the separation of the secondary e ect of ultrasound that is hypothesized to be responsible for the intensiÿcation of the textile processes. shock wave. A conventional approach to the modeling of the mass transfer enhancement due to convection. This is a plain weave cotton fabric.G. • The fabric used in the model process should preferably be a plain weave (with both inter. therefore. Moholkar. Fig. the convection responsible for the mass transfer enhancement is caused. is to solve the mass balance equation after substitution of the uid velocity. Model wet textile process So as to estimate the degree of intensiÿcation of mass transfer in the textile with ultrasound. In view of these requirements. The carbon particles in the soil in the model fabric are not soluble in the medium. the micro-mechanism of this process is not clear. needs to be selected. a model wet textile process with a model fabric and a model di using substance. but adhere to the ÿber along with the olive oil. • The monitor should be completely reversible. EMPA 101 fabric satisÿes all of the requirements of the model fabric and monitor described above except that the carbon particles are not free to be transported by convection. A comprehensive model that describes dynamics of thousands of bubbles. and hence. A complete removal of the monitor from the textile should be possible and the original textile should be recovered after total extraction of the monitor. an exact determination of their size is difÿcult. M. although in quite small concentration.and intra-yarn porosity) that o ers a simple geometry for the transport of the monitor. For this situation. an approximate particle size that could be assumed for the purposes of calculations would be ∼0:1 m or so (Koster. adopt a semi-empirical method to estimate the mass transfer enhancement factor by an order of magnitude. 2. viz. However. therefore. Its SEM images are shown in Fig. Most of the bubble dynamics equations available so far apply for a single bubble oscillating in an inÿnite medium. Quantiÿcation of mass transfer enhancement: concept of convective di usion coe cient As stated earlier. they can be approximated as rigid spheres di using in the continuum of the washing medium. such as the current study. Therefore. Due to very small initial size. we have selected EMPA 101 fabric (manufactured by ETH. • The concentration per unit area of the monitor in the fabric should be constant.300 V. it was shown by us (Moholkar and Warmoeskerken. Zurich) as the model fabric for the experiments. • The original di usion rate of the monitor should be slow enough to be “accelerated” by ultrasound. personal communication). the most common basis for the determination of the di usion coe cient is the Stokes–Einstein equation D= 6 ÄT : dp (1) . then the acceleration due to ultrasound may not be discernible. 1. due to their very small size (∼0:1 m). It could be inferred from Fig. This indicates that the soil is unevenly distributed in the interand intra-yarn space. These particles. any further reduction in the size of the monitor will not occur and. A proper study of the ultrasonic mass transfer enhancement will require that the model fabric and mass transfer monitor possess the following properties and characteristics: • The monitor for the transport (in the form of a particle or a molecule) should be chemically inert. however. The presence of the carbon particles inside the yarn is clearly visible. soiled with carbon soot and olive oil. with the model process being washing of this fabric with ultrasound. high-velocity micro-jet or oscillatory spherical velocity ÿeld generated due to radial bubble motion or by combination of all three possibilities. need to be loosened from the surface of the ÿbers before being transported in the medium. for the quantiÿcation of mass transfer enhancement. mass per unit area 100 g m−2 . in practical situations. so that the samples used for different experiments should have the same initial concentration of the monitor di using out. is not developed yet. with strong interaction between them. 1D shows the cross-section of the EMPA 101 fabric. Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 ing the experimental results and theoretical model.M. particle size reduction (Klutz. This necessitates pretreatment of the textiles before the ultrasound treatment such as soaking in a detergent solution. 3. which enhances the mass transfer in the textile. 2002) that (high-energy) transient bubble motion near the textile surface is responsible for the creation of strong convection in the close vicinity of the textile. • The size of the monitor particles should be smaller than the intra-yarn pores of the textile (typically 5 m or so). 1 that the carbon particles are very small.S. However. an exact determination of the convection velocity is not possible. • The monitor should not form any strong chemical or physical bond with the ÿbers of the textile that can hinder its transport across the fabric. which acts as a monitor for mass transfer. the above-mentioned secondary e ect can be isolated.C. We. 1997). that will basically ensure the presence of the monitor in these pores. viz.

which will help determine the Reynolds number for the uid motion in the close vicinity of the textile and. is moving due to convection currents set by density or pressure di erences or due to external energy input from sources such as a mechanical stirrer. In order to propose a mathematical model for the di usion process in the textile. Moholkar. It was then put in an oven. 1. (Cross-section of the yarn was obtained by ÿrst soaking the textile in a monomer solution of epoxy resin.e. In both the cases.C. The convective di usion coe cient is basically a synonym for the widely used term eddy di usion in transport phenomena. Thus. (2) a porous cylinder (representing a single yarn in the textile).G. in this procedure the textile was basically frozen in the resin matrix.. For this purpose. Model 2050)). in which the particle is suspended. For example: (1) a porous plate (representing the textile as a whole). it is necessary to assign a suitable geometry to the textile. the bubbles) and the structure through which the mass transfer occurs due to the convection (i.. We characterize this enhanced mobility of the particle due to the bulk uid motion as convective di usion coe cient. slices of the resin matrix were obtained with a super-cut machine (Reichert Jung Inc.M. while taking care to avoid entrapment of air bubbles. we have used the term convective di usion coe cient. SEM images of the model EMPA 101 fabric (SEM machine: Jeol Inc. a homogeneous porosity in the structure is assumed. hence the ow regime.V. In order to make a proper choice for the textile geometry. Next. due to lack of knowledge of the exact magnitude of the uid velocity created by the transient motion of the bubbles. (B) a single yarn.1. the di usion coe cient of the particle is much higher than that predicted by the Stokes–Einstein equation. Description of the fabric geometry As described in Fig. the textile).e.. The typical diameter of a yarn is .S. 9 (Appendix A). 3. (C) ÿber of the fabric: carbon soot on the surface of the fabric is clearly visible. where the monomer is converted into a cross-linked polymer with the textile entrapped in-between. which represents solute transfer due to turbulent uid motion. Accelerating voltage: 5 kV.. which in turn are made of ÿbers. If the liquid. (D) cross-section of a single yarn: some carbon particles are present inside the yarn. Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 301 (A) (C) (B) (D) Fig. Model GSM 5800. a textile is comprised of yarns. one needs to take into account the relative sizes of the source of convection (i. many possibilities exist. However. M. Working distance: 10 mm): (A) surface of the fabric.

we combine experimental and theoretical methods in three steps as follows: 1. 3. which is discussed subsequently. M.C. A timer switch was added between the signal ampliÿer output and ultrasound horn. Use an approximate method of cavitation nucleation near the fabric surface for the experimental determination of the rate of soil removal for di erent periods of ultrasound irradiation. The output power of the ampliÿer could be controlled by adjusting the voltage of the input signal to the ampliÿer.. The stainless-steel bottom of the experimental cell acts as a rigid re ector for the ultrasound waves. Rubber gaskets were placed between the textile and glass rings in order to avoid leakage. Model 2690)..302 V. The output of the hydrophone was transformed into proportional voltage by a charge ampliÿer (Nexus Range. Moholkar. However. The base of the jack could be raised or lowered to adjust the distance between the bottom of the cell and the tip of the horn. Model 430A).S. An intense spherical velocity ÿeld is created in the close vicinity of the bubble during compression. The cell was mounted on a stainless-steel bottom (thickness: 51 mm) with four vertical bars that act as support for the glass rings placed above each other. This voltage was monitored on a digital oscilloscope (Tektronics Ltd. Schematic diagram of the experimental set-up... 3. This circumstance justiÿes considering the textile as a porous plate. 2. 60 mm) in water. The experimental cell: The experimental cell was made of three detachable glass rings and a lid made of Te on. 4. 4. which is equal to one-quarter of the wavelength of 25 kHz ultrasound (i. 2. inter.1.G. a special arrangement was made to place the hydrophone (Bruel & K jer Ltd. when vibrating in air. For the estimation of convective di usion coe cient. Type 8103) for the measurement of the acoustic pressure amplitude. The ultrasound unit: The ultrasound unit comprised of a special-made ultrasound horn with a central resonance frequency of 25 kHz.M. The cell had calibrated distance marks on it to measure the distance between the ultrasound horn tip and the rigid bottom.. Model 6138A) and a current clamp (Farnell Inc. Experimental 4. 60 mm.. tends therefore to be larger than the dimensions of the source of convection. the dual porosity of the textile (viz. The soil concentration in the fabric can be calculated from its re ectance using the Kubelka–Munk theory.. 2. Source of cavitation nucleation The cavitation-nucleation in the medium is an important factor for any cavitation-aided physical or chemical pro- . controlling the time of the ultrasonic textile treatment accurately. Solve the di usion equation for a plane sheet to determine the concentration proÿles of the di using substance for di erent time intervals. The voltage and current supplied to the ultrasound horn were monitored using a voltage probe (Tektronics Ltd.e. 2002).. The horn was driven by a signal generator (Hewlett-Packard Inc. During this phase. The height of the two rings was 15 mm. 1B. i. The ultrasound horn was mounted onto the shaft of a laboratory jack and the experimental cell was placed on its base.2.e.2. the bubble. Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 Signal Generator and Amplifier Timer switch ∼200 m. This switch connects the ampliÿer output to the ultrasound horn only for a speciÿed time period. and perhaps 100 m (Moholkar. Experimental system The experimental system had two main components: an experimental cell and an ultrasound horn along with a signal generator and ampliÿer.2–1:5 bar) oscillates between a few microns. A schematic diagram of the set-up is shown in Fig. i.and intra-yarn porosity) needs to be taken into account. Model 3324A) and a radio frequency ampliÿer (ENI Inc. while the height of the third ring was λ Current and Voltage Monitoring Unit Ultrasound horn and experimental cell with textile Oscilloscope Fig. Model PR-20). while analyzing the e ects of ultrasound treatment on large portion of the textile (or on a macroscopic scale).. Model 2100L).. Using experimentally measured values of soil removal from the model fabric for di erent time periods of ultrasonic treatments and the theoretical expression for the soil removal (obtained in the previous step) to ÿnd out the convective di usion coe cient for soil under the effect of ultrasound irradiation. The ampliÿer could supply a maximum of 200 W of electrical power for a large frequency range (10 kHz– 1 MHz). the porous structure. The textile sample could be placed in-between the glass rings. At the bottom of the cell. thus. as is evident from Fig. and obtain an expression for the total amount of di using substance transported across the plate per unit time. a single yarn in the textile. The radius of bubbles that undergo transient motion near the textile surface (for 25 kHz frequency and acoustic pressure amplitudes 1. Approach A ratio of the convective di usion coe cient and molecular di usion coe cient (obtained from Stokes–Einstein equation) gives the mass transfer enhancement factor due to ultrasound.e.

The nuclei close to the textile surface. However. Therefore. thus. the bubble shape during the radial motion remains spherical without appreciable deformation and. the bubble may undergo di erent kinds of motions (Naude and Ellis. In case of bubble activity near a solid surface. a suspension of polystyrene latex particles is used to provide nuclei in the medium (e. can create the convection for soil removal almost immediately after ultrasound exposure. we have used a very small amount (20 mg) of pumice stones for the nucleation in each experiment. placed on the fabric surface itself. especially for smaller periods of ultrasound exposure. 1998).M. restricting the size range after screening using sieves of certain mesh sizes. the initial nuclei population and their size distribution undergo changes due to fragmentation of the bubbles during radial motion and re-growth (and collapse) of the daughter bubbles. Therefore.1. During the process. For smaller stando factors (¡1:5). and adding an exact quantity of stones to the medium each time.S. The time of ultrasound exposure is also an important factor in these processes. It must be noted that having uniform particle sizes may not be of much use as far as the uniformity of nucleation is concerned. is a parameter of paramount importance.. However. For large stando factors (¿3). In view of the above considerations. the extent of convection produced by the bubble depends not only on the population and the size distribution of the cavitation nuclei but also on their distance from the surface. 1986. Since these stones are commercially available in large sizes (¿3 mm or so). and not by the particles themselves. Characterization of the ultrasound ÿeld According to the analysis of the pressure and velocity variation in the standing wave ÿeld given by Pierce (1989).g. when the boundary is su ciently rigid. For these conditions.. we have used pumice stones of size range of 200 –400 m as a source of nuclei in the present experiments. the initial nuclei population and size distribution is of crucial importance because these parameters are not likely to change much during the exposure. particles with surface crevices that can entrap air pockets. the bubble undergoes deformation producing a high-velocity micro-jet directed towards the boundary.3. In addition. the effect of initial nuclei population and size distribution is likely to be smoothed out. pressure measurements were done in the experimental cell using a very small (4 mm) uncalibrated hydrophone at two locations viz. for the reasons explained earlier. Phillip and Lauterborn. by reducing their sizes by grinding. for a short ultrasound exposure. this problem could be overcome to some extent. to a signiÿcant extent. and create nucleation near textile that could assist the soil removal. at distances =4 and =2 from the rigid bottom of the cell with degassed water as the medium. The air pockets entrapped in this suspension can provide nuclei for cavitation in the medium. the distribution of the cavitation nuclei over the surface of the textile is also an important factor because the dynamics of a single bubble is in uenced signiÿcantly by the interaction with adjacent bubbles. Instead.. Holland and Apfel. the nucleation by polystyrene latex suspension may not give good results. However. which is the ratio of the distance of the bubble center from the textile and the radius of the bubble at beginning of collapse. Therefore.C. may not help in immediate removal of soil because of the rapid reduction in the intensity of velocity ÿeld away from the center of the bubble. for ultrasound processes of longer duration. for experiments aimed at determining the kinetics of cavitation-aided physical or chemical processes. If added in large amounts. The nuclei. the rate of soil removal is determined by the time for bubble migration and the intensity of convection created by the bubble. Depending on the stando factor. etc. etc. 3 =2. for plane-wave re ection from it. Blake et al. as compared to the period of acoustic cycle. during ultrasonic irradiation.V. a bubble initially present away from the fabric can migrate towards the fabric surface (if it does not collapse before reaching the fabric). can be used for the present purpose. initially away from the textile. these air pockets are distributed all over the medium along with the polystyrene particle suspension. Thus. before the start of ultrasound irradiation. from the rigid surface. However.G. Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 303 cess. the choice of source of nuclei is of paramount importance in the accurate measurement of rate of soil removal. the pumice stones can alter the power consumption of the system due to reduction in system impedance as a result of large bubble population in the system that increases the compressibility of the medium (Moholkar. the locations of pressure minimum occur at a distance of =4. 5 =4. 1986. 2002). the velocity ÿeld produced around it is also spherically symmetric.. This obviously can lead to signiÿcant errors in the calculations of rate of soil removal. for the present study with very short periods of ultrasound exposure. Tomita and Shima. 4. can provide nucleation at desired location for production of instantaneous convection with ultrasound exposure without any time-lag e ects arising out of factors just discussed. 1990). Experimental procedure 4. the source of cavitation nuclei. However. which determines the initial population and size distribution of the cavitation nuclei. . Due to these ows. 3 =4. they may give rise to non-uniform nucleation. Thus. which are widely used for providing nuclei for boiling. Conventionally. Pumice stones. circulatory ows are set up in the system due to the absorption of the momentum of the acoustic waves by the liquid. M. because nuclei for cavitation are mainly contributed by the gas pockets trapped in surface crevices of the particles. 1961. In order to conÿrm this theoretical result. while the locations of pressure maximum occur at a distances of =2.. in this case. Moholkar. on the population and the size distribution of these nuclei.3. it was found that the power consumption of the ultrasound horn remained unchanged. The distance between the tip of the horn . because the physical or chemical e ect depends.

In this study. the subharmonic peak. Data analysis The quantiÿcation of the washing e ect can be done by measuring the re ectance of the ultrasound-treated textile sample. 2002). 250 ml degassed water was used as the washing medium. The ultrasound horn was driven at 25 kHz with a power input of 20 W. Kubelka–Munk theory is often used. scattering coe cient (S) and the re ection coe cient of the layer (Rf ) is (Rf − 1)2 K = : S 2Rf (2) For particles larger than the wavelength of light. A possible reason for this is that due to presence of small bubble fraction the wavelength of sound is not exactly 60 mm. (B) acoustic spectrum at pressure node. while it is completely absent in the acoustic emission spectrum at location =4.G. The time of ultrasound treatment was 3 min. and the rigid bottom of the experimental cell was adjusted to 60 mm. which is above the transient cavitation threshold for bubbles of sizes 5 –10 m at 25 kHz frequency (Moholkar. we need to convert the re ectance measurements into soil concentrations. which is the wavelength of 25 kHz ultrasound in water.C. but has some positive residual value. pumice stones are the only source of cavitation nuclei in the system. The dissolved oxygen content of the washing medium was lowered to 2 ppm using a chemical method (van der Vlist et al. Characterization of the standing wave ÿeld generated in the experimental cell: (A) acoustic spectrum at pressure antinode.1 to 2 s were selected for ultrasound irradiation. is quite prominent in the acoustic emission spectrum at location =2. If the distribution of particle diameters is very small. The model fabric was soaked in the detergent solution (1:75 g l−1 of sodium dodecyl benzene sulfonate) for 5 min before ultrasonic treatment. indicative of transient cavitation. Moholkar. Thus. the model fabric was removed from the cell and was dried in air. with the fabric positioned at a pressure antinode. the ultrasound horn produces an acoustic wave with a pressure amplitude of ∼1:3 bar. which is clear from these results is that the resultant acoustic pressure amplitude at =4 is not zero. The Kubelka–Munk theory can be coupled with the Lambert–Beer law to convert the remission function into absolute concentration. the scattering coe cient is inversely proportional to the diameter. Another feature of the acoustic wave phenomena. In each experiment. The mathematical expression of the Kubelka–Munk theory that relates the absorption coe cient (K). were ÿlled with silicon rubber. 1931).3. 3 shows the FFT of the pressure signals (duration: 12 ms. Eight speciÿc time intervals ranging from 0. The model fabric was then ÿxed between the glass rings of the experimental cell in a water bath to avoid entrapment of air beneath the fabric. After ultrasound treatment. we use the same conditions for the experiments. it produced an acoustic wave with pressure amplitude ∼1:3 bar. the characterization of the acoustic wave ÿeld in the experimental cell approximately conÿrms the theoretical prediction of the location of the pressure node and the pressure antinode. the relationship between . 1994). At this power input. 4. 3. Power input to the ultrasound horn was 20 W at 25 kHz frequency. 4. In this case.. which is the simplest tool for the description of the optical properties of a turbid medium that absorbs and scatters light (Kubelka and Munk. For this power input. In order to assess the e ect of the acoustic pressure amplitude on the convective di usion coe cient. For each interval.304 Fourier Coefficients 8 V. M. In addition. A maximum washing e ciency (indicative of maximum mass transfer) was obtained with a degassed washing medium and non-degassed fabric. Washing experiments Investigations of Moholkar and Warmoeskerken (2002) revealed that the ultrasonic washing e ciency depends on the gas content of the fabric and the washing medium. three experiments were conducted with di erent pieces of the model fabric to assess reproducibility of results. In case of negligible scattering coe cient.M. The circular grooves in the tip of the horn. In order to determine the rate of soil removal. sampling frequency: 2:5 MHz) obtained at distances =4 and =2 from the rigid re ector. for which the ultrasound horn produces a wave with pressure amplitude of ∼1:1 bar. which causes a shift in the position of pressure node. Fig. The amplitude of the fundamental peak in the acoustic spectrum corresponding to the driving frequency (indicative of the acoustic pressure amplitude) at location =2 is much higher than the amplitude of the corresponding peak at location =4. then one can approximate: S∼1=dp .S. which can entrap small amount of air that can provide alternate source of nuclei in the medium. Hence. the set of experiments at 20 W power input was repeated at power input of 15 W. For this purpose. no cavitation nuclei are contributed by the medium. but somewhat less.4. Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 4 0 0 10 20 30 (A) 8 Fourier Coefficients 40 50 60 Frequency [kHz] 40 80 90 100 4 0 0 10 20 30 (B) 40 50 60 Frequency [kHz] 40 80 90 100 Fig.2.

the concentration proÿle can be approximated by a more convenient expression. Thus. the bubble motion remains symmetric as if occurring in an inÿnite uid. at 25 kHz. which has left the plate after a time t(Mt ).M. Bubble motion near the fabric The bubble motion near a solid boundary is a multi-faceted phenomenon. One-dimensional unsteady state diffusion (in absence of velocity) in a porous plate is described by Fick’s second law @C @2 C (4) =D 2 : @t @x Let −l¡x¡l denote the boundaries of the plate with thickness 2l. (4) are listed by Carslaw and Jaeger (1959) for di erent boundary conditions. (1999) and Janssen and Warmoeskerken (1997). B = 3000 atm and n = 7). Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 305 absorption coe cient K and concentration of soil C is K = C ln(10) = n C. P∞ .C.V. Moholkar. The kind of radial motion the bubble undergoes in the vicinity of the solid boundary. corresponding to 100% soil concentration in fabric. It is likely that. 1975) C − C1 4 =1− C0 − C1 ∞ n=0 (−1)n exp 2n + 1 −D(2n + 1)2 4l2 2 t = 1+ (9) (2n + 1) x : ×cos 2l (5) The total amount of substance. we approximate the geometry of the textile as a porous plate. C = C0 . The re ectance of untreated EMPA 101 textile is 16%. 5. K = 0:064 at Mt = 1. l. the pressure in the bulk liquid driving the bubble motion is P∞ = P0 − PA sin(2 ft): (11) . In the present study. the cotton fabric acts as a soft boundary for a bubble oscillating in its vicinity. Brujan et al. Recently. 1954) R 1− U c d2 R 3 U 1− + dt 2 2 3c U U U H+ 1+ c c c dR dt dH : dR 2 the concentration of the di using substance in the plate is (Crank. the elastic modulus is still relatively small and. we present simulations of the bubble motion and the spherical velocity ÿeld generated around it using the Gilmore bubble dynamics equation (Gilmore.1: It is convenient to introduce the Fourier number deÿned by Dt Fo = 2 : (7) l A. for for x = −l. Cotton ÿbers have a very low static elastic modulus (Morton and Hearle.205. it is not possible to determine n . For the conditions C = C1 . the average concentration in the porous plate is 8 C1 − C = 2 exp(− 2 Fo): (8) C1 − C 0 This means that plot of logarithm of the expression on LHS of Eq. M. They conclude that for boundaries with very low elastic modulus. since the manufacturer of EMPA 101 model fabric does not supply the soil separately. t = 0. where n = ln(10): (3) We make two simplifying assumptions: (1) the diameters of all carbon particles in the soil over textile are the same. For times such that Fo¿0:1. The detailed theoretical analysis of this (which is an extension of the penetration theory) can be found in several references such as Beek et al. 5. The mathematical model As discussed earlier. We give herewith only the ÿnal result. B and n are constants (For water. The solutions of Eq. which gives K=S = 0:064. the bubble motion is asymmetric (with formation of a jet) only for stando factors smaller than 1. 1997). (2) the scattering coe cient of carbon particles is negligible. corresponding to 0% soil concentration in fabric (note that scattering coe cient is cancelled out in the calculations of relative soil concentrations due to assumption 1).S. The fraction of original soil removed from the fabric after ultrasound treatment for di erent time periods (Mt ) is estimated by assuming a linear relation between K and Mt with: K = 2:205 at Mt = 0. For cases where stando factor ¿1. Therefore.G. and the corresponding amount after inÿnite time (M∞ ) are related by Mt =1− M∞ ∞ n=0 H is the free enthalpy on the surface of the bubble  3 R0 2 2 n A1=n  +B P0 + − H= n−1 0  R0 R R −[P∞ + B](n−1)=n    : (n−1)=n 8 (2n + 1)2 2 exp −D(2n + 1)2 4l2 2 t (10) : (6) Concentration proÿle for Fourier numbers ¿0. whether symmetric or asymmetric. For longer contact times.1. (8) against Fourier number should yield a straight line with slope ∼ − 9:86. (2001) have reported high-speed imaging studies of a vapor bubble dynamics near an elastic boundary. depends on the physical characteristics of the boundary (such as elastic modulus) and also on the stando factor. respectively. A = 3001 atm. the soil concentrations need to be deÿned on a relative basis. The re ectance of completely washed EMPA 101 fabric is ∼70%. therefore. which gives K=S value of 2.

G. 4 and 5. 5. 4. Therefore. One such mismatch is the geometrical di erence between the model porous plate and the textile. 1 reveal that the soil is distributed unevenly over the textile surface.and intra-yarn soil is removed. due to geometrical constraints. Images of the EMPA 101 model fabric shown in Fig.5 1 Time (sec) 1. Adoption of the di usion model for the present study While adopting the above models to explain the results of the present experiments.1 5. (2) rate for longer time period of ultrasound irradiation during which both inter. while for longer irradia- .50 0. Thus.3 R2 u(r. We will be quantifying the soil transfer rates by measurement of the reectance of the textile. the soil removal process is expected to have two steps. the intensity of the convection resulting from the radial bubble motion in the inter-yarn region is much higher than in the intra-yarn region. t) = 2 r dR dt : (12) 0. It is obvious that soil removal rate for step 1 is higher than step 2. However.5 2 (A) 1.20 0.25 0.00 Time (sec) Fraction of original soil removed 0.40 0. 0.50 0. M.00 0 0. Moholkar. this measurement technique will quantify the soil transfer not only in the intra-yarn pores but also in the inter-yarn pores.00 Fraction of original soil 0.30 Reflectance 0.75 removed 0.00 0 0. for shorter ultrasound irradiation the rate of soil removal from the fabric is higher than for longer ultrasound treatments. The re ectance of the fabric is a function of the average soil concentration in the textile.5 1 1. As discussed in Appendix A. It can be seen that the washing e ect for both power inputs shows a gradual rise with time of ultrasound irradiation. This result may perhaps be interpreted from the discussion given in previous section. with the major fraction of the soil present in the inter-yarn pores.75 0. As discussed earlier.5 1 1. (B) soil removal from the textile with time. The Fig. This could give rise to a di erence between the theoretical results predicted by the models with assumption of constant di usion coe cient with homogeneous porosity in the sheet and the experimental results. (B) soil removal from the textile with time.2 0. only the soil on the fabric surface and in the intra-yarn pores is removed.5 2 Time (sec) The results of the experiments done with 20 and 15 W power input to the ultrasound horn are shown in Figs.4 0.5 The velocity in the bulk liquid at a distance r from the bubble center is Reflectance 0.00 0 0. we will be determining the rate of soil removal from the textile on a macroscopic scale. we need to take into consideration the possible discrepancies between the experimental and theoretical results.5 2 (A) 1.25 6.C. which could occur as a result of mismatches between the assumptions made in the model and the practical situation.S. the textile is modeled as a porous plate on the basis of the di erences between sizes of an individual bubble and that of an individual yarn.10 0.5 2 (B) Fig. respectively.5 1 Time (sec) 1. with di erent soil removal rates: (1) rate for smaller time period of ultrasound irradiation during which only inter-yarn soil is removed. results are consistent with the hypothesis that.2. Experimental results for fabric treatment at 20 W power input: (A) re ectance of the fabrics treated for di erent time periods. during the experiments. for shorter ultrasound irradiation. 0. Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 0.0 0 0. Experimental results for fabric treatment at 15 W power input: (A) re ectance of the fabrics treated for di erent time periods. Results and discussion (B) 0. However.306 V.M. The re ectance is measured over a large area of the textile (and not for a single yarn).

7. however.03. it is likely that the above-mentioned factors varied signiÿcantly in consecutive experiments. The initial portions of the curves are described reasonably well by Eq.5 2. (8).0 2. Features of mass transfer in fabric under ultrasonic treatment: (A) 20 W power input. one deduces from the data the values D = 7:5 × 10−10 m2 s−1 at 20 W and 5 × 10−10 m2 s−1 at 15 W. (B) 15 W power input. However.20 y = -0. it decreases slightly. The logarithmic average soil concentration for longer contact time (Eq. assuming fabric thickness of 100 m. (6) with D = 2 × 10−9 m2 s−1 at 20 W and D = 1 × 10−9 m2 s−1 at 15 W. Fig. These values fairly match with . since the intra-yarn soil forms only a small fraction of the total soil in the textile. In Fig. (6).25 0.5 1 1 . 6A).2 (B) Fig. we attribute the anomaly of lesser fabric re ectance for 1 s ultrasound treatment than 0:5 s to the error in having exactly same kind of nucleation produced near the textile surface in consecutive experiments. In addition to duration of ultrasound irradiation.5 -0.S.5e-10 Fraction of original soil removed (A) 0. where the shift between mass transfer rates occurs at ∼0:6 s. (Values of D indicate di usion coe cient in m2 s−1 for theoretical calculation of the soil removal curve using Eq.0 0 0.8 (B) D = 5e-10 Time (sec) D = 1e-9 Experimental Fig.5 Time (sec.5 0.5 D = 2.0 0. the data imply that it decreases with time.50 0. tion soil in the intra-yarn pores is also removed.G. A comparison of Figs.) 1.M.5 Time (sec. This is in contradiction with the theoretical speculation that soil removal should increase with time of ultrasound irradiation. 4B and 5B reveals that larger power input. (B) soil concentration for 15 W power. Thus. A similar analysis for Fig. 6.4999x . It can be seen that a single value of the di usion coe cient is inadequate to reproduce the data. need to be taken into account.C.25 -0. Moholkar.) -1. respectively.4 0 0. Thus. 5A.0. results in faster and higher removal of soil from the fabric. The plots are shown for t¿0:5 s for 20 W. which corresponds to a Fourier number of ∼0:03 as calculated with the smaller di usion coe cient. the re ectance of the fabric does not show a consistent rise with time of irradiation. The value of convective di usion coe cient obtained from the slopes of the plots.60 (A) Time (sec) D = 7. The shift between the two mass transfer rates occurs at approximately 0:4 s (referring to Fig.0 1.0 2.5 Experimenal 1.5 2. as expected. M. Since we used a crude method of providing cavitation nucleation near textile surface. Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 1. one can estimate this smaller di usion coe cients from the large-time behavior of the data. 6A and B for 20 and 15 W power input. A comparison of the theoretical and experimental soil removal rate is shown in Fig. (8)): (A) soil concentration for 20 W power.5805 -0.75 0. the re ectance of the fabric does not rise signiÿcantly for longer ultrasound irradiation.00 307 Fraction of original soil removed 0. which means cavitation at larger pressure amplitude.4671 -1. therefore.) 1. and for t¿0:6 s for 15 W power input. are 5 × 10−10 m2 s−1 for 20 W and 3 × 10−10 m2 s−1 for 15 W. Between 0.00 1 1 . at large times the solution will be dominated by the slower di usive process and.00 0. On the basis of this argument.0e-9 2 -1. 7 shows plots of variation in average soil concentration in the fabric for longer contact time using Eq. the di usion coefÿcient is found to increase with the acoustic pressure amplitude.0.80 ln (C) 0 0.75 0. the location of bubble from fabric surface.00 ln (C) y = -0. If the process can be described by a superposition of two di erent di usive processes.5 and 1 s. characterized by di erent di usion coe cients.V. the soil removal from the fabric depends on several other factors such as size distribution and population of bubbles.40 0.0 1. 6B. indicating lesser soil removal in 1 s ultrasound irradiation than that for 0:5 s. To the extent that an “e ective di usion coe cient” can be deÿned. The slope of the line (= − 2 D=l2 ) indicates the convective di usion coe cient.5 2 0. Some experimental errors.1963x . reveals that the Fourier number at which this shift occurs also is 0.50 -0.

which was deÿned as the ratio of convective di usion coe cient to molecular di usion coe cient. calculated from Eq. Fig. The experimental results were correlated to the theoretical values of soil removal with time using solutions of the di usion equation. 6. This di usion coe cient needs further correction for the porosity and tortuosity of the fabric. the values of convective di usion coe cients given in Fig. Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 2 0 0 5 Velocity (m/s) 1 2 3 Time (No. The order of magnitude of the mass transfer enhancement for the model wet textile process. The other parameters for the simulation are given in the ÿgure caption. With substitution of the p representative values dp ∼0:1 m and D∼1×10−13 m2 s−1 . a function of the time-averaged velocity over several acoustic cycles. The soil removal from the textile showed two distinct regimes with two di erent convective di usion coe cients.M. the mass trans- . it could be perceived that the time scale of radial motion of the bubble is same as that of the acoustic wave period.C. Since the purpose of the simulations presented here is to give an approximate estimate of the convection created by the radial bubble motion. which is representative of the mass transfer enhancement due to ultrasound. 7. of Acoustic Cycles) 4 5 Fig. per deÿnition given in Appendix A.25 0 0 1 2 3 Time (No.308 4 V. of Acoustic Cycles) 4 5 1.S. (6). inter. The (absolute) magnitude of the velocity ÿeld in any direction and. the e ective motion of the particle resulting due to convection created by a single radial bubble motion is negligible. The distance of bubble center from textile: 50 m. we deÿned a convective di usion coe cient. The estimation of this velocity is highly complicated because it will not only be a function of the bubble population but also of the spatial distribution of the bubbles from the textile surface and the strong interaction between adjacent bubbles.and intra-yarn porosity. 8 shows simulations of the radial dynamics and the spherical velocity ÿeld of a 10 m bubble driven at di erent acoustic pressure amplitudes.G. Conclusion Ratio (R/R0) Ratio (R/R0) -1. Therefore. As mentioned earlier. frequency: 25 kHz: (A) acoustic pressure amplitude = 1:3 bar. simulations of the bubble dy- This paper describes a semi-empirical methodology for a quantitative estimate of the mass transfer enhancement in the ultrasonic textile treatments. The ratio of convective di usion coefÿcient and the molecular di usion coe cient indicates that the enhancement factor is in the range 1000 –2000. we have not incorporated such complications in our analysis. is: 3 × 10−13 m2 s−1 . A typical estimate of the time scale of particle motion will be ≈ d2 =D. 8. is in the range 1000 –2000. This result was explained in terms of the non-uniform initial distribution of soil in the textile and the dual porosity of the textile. of Acoustic Cycles) 4 5 1. This methodology can also be applied for other ultrasonic wet textile treatments. In absence of precise quantiÿcation of the convection created due to transient bubble motion near the textile surface. hence. Mass transfer enhancement: Assuming that the carbon particles in the soil on model fabric are spherical and have size ∼0:1 m. It can be seen that the bubble creates an oscillatory velocity ÿeld around it with di erent magnitudes of the velocities in the two directions: towards and away from the bubble center. we ÿnd = 0:1 s. The corrected di usion coe cient. thus. however. a direct estimation of the convection velocities near the textile surface is not possible in the present experiments. Textile treatments at higher acoustic pressure amplitudes were found to give higher di usion coe cients. 8. of Acoustic Cycles) 4 5 2 -5 0 1 (A) 2. Simulations of radial dynamics and spherical velocity ÿeld for oscillations of 10 m bubble near textile surface. which is 40 s for 25 kHz frequency. The geometry of the textile was modeled as a porous plate. From Fig. However.5 0 1 (B) 2 3 Time (No. (B) acoustic pressure amplitude = 1:1 bar. the molecular di usion coe cient obtained from Stokes–Einstein equation is ∼3 × 10−12 m2 s−1 .5 2 3 Time (No. Moholkar. The net motion of the soil particle under the convection created by the bubble motion is. the overall convection generated due to radial bubble motion increases with the acoustic pressure amplitude. M.5 Velocity (m/s) 0 namics equation can give us an order of magnitude estimate of the convection velocities.

m s−1 mass transfer coe cient for inter-yarn region. Netherlands for the funding of the project at the University of Twente. Pa radial coordinate.1. (A. Appendix A. Therefore. m diameter of a yarn. The ratio of Eqs. kg or kmol pressure amplitude of the acoustic wave. m s−1 bubble wall velocity.V. m s−1 mean uid velocity. kg m−3 density of the liquid in undisturbed state. J K −1 wavelength of the acoustic wave. m−1 permeability between the yarns.C. m2 permeability between the ÿbers. This can create a thin convective shell near the periphery of the yarn.1) (A.G. Fluid ow through the fabric For a uid owing through the textile.and intra-yarn regions can be written as (Gooijer. Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 309 fer enhancement factor may change from system to system depending on the properties of the textiles and those of the ultrasound system. kg m−3 surface tension. however. m s−1 half-thickness of the porous plate. df = 15 m gives Kintra-yarn 1 1 : (A.S. dimensionless overall mass transfer coe cient in the textile. m2 Kozeny’s constant. We hope that this study will contribute towards the design and scale-up of ultrasonic wet textile processes.. n y Ä Notation c C df dp dy D De f H k1 k2 k3 K Kinter-yarn Kintra-yarn K0 K0 KT l M∞ Mt PA P∞ P0 r R Rf R0 S t T u U v x Greek letters f L velocity of sound.3) = to Kinter-yarn 200 2000 The ratio of Kintra-yarn and Kinter-yarn indicates that the permeability between the yarns is far higher than the permeability in the yarns. dimensionless polytropic constant of the bubble contents.M. 16K0 (1 − f )2 f 3 1 y d2 : 16K0 (1 − y )2 y 3 (A. m re ection coe cient. m amount of di using material transported in inÿnite time. m bubble radius.V. Transport phenomena in textile materials A.2) with representative values of the parameters: f = 0:35 − 0:45. dimensionless porosity between the yarns. m s−1 mass transfer coe cient for the boundary layer. Hz free enthalpy on the bubble surface. kg m−3 or kmol m−3 diameter of a ÿber. m di usion coe cient. the mass transfer in this region occurs only by the slow process of di usion. kg or kmol amount of di using material transported in time t. m particle radius. dimensionless Kozeny’s constant. penetrate an individual yarn to a small extent. dimensionless . Nonetheless.1) and (A. Pa s density of liquid. K velocity in the medium. dimensionless initial bubble radius. tortuosity factor. most of the ow will follow the path of least resistance (between the yarns). dimensionless Boltzmann constant. Pa pressure in the medium at inÿnity. and practically no ow will occur through the yarns. s temperature of the medium. m s−1 absorption coe cient. The relative contribution of the ows in these two regions can be judged by the ratio of the permeabilities of these regions. m2 s−1 e ective di usion coe cient. N m−1 Acknowledgements The authors are grateful to Stork Brabant B.2) We make an assumption of constant K0 for the inter. m −1 0 Naperian molar absorption coe cient. The ow through the inter-yarn region may. m s concentration of di using substance. m viscosity of the medium. y = 0:4 − 0:5.and intra-yarn pores on basis of similar porosities of an individual yarn and the overall textile (Gooijer. m s−1 distance coordinate. Therefore. m2 s−1 frequency of the acoustic wave. dy = 250 m. The permeabilities of the inter.and intra-yarn regions form two parallel paths. J kg−1 mass transfer coe cient for intra-yarn region. M. the central core of the yarn still remains a stagnant core zone with no ow inside. Pa ambient (atmospheric) pressure. m−1 time. the inter. dimensionless porosity between the ÿbers. m scattering coe cient. Moholkar. 1998) Kintra-yarn = Kinter-yarn = 1 f d2 . 1998).

9: 1.S. Several correlations for the mass transfer coe cients are listed by Janssen and Warmoeskerken (1997).G. Mass transfer in the fabric For washing. mass transfer from the boundary layer to the bulk liquid. In order to ÿnd the overall mass transfer coe cient (KT ) in the textile. 1999. 1986.. one needs to have an insight into the nature of uid ow through the textiles. Therefore. Rigid boundary. we specify k1 .R. second and the third step of mass transfer mentioned above. The mass transfer coe cient inside the yarn (k1 ) is then calculated from the Sherwood number as ≈ 4 × 10−4 cm s−1 . To estimate the inter-yarn mass transfer coe cient. To estimate the inter-yarn mass transfer coe cient. 3.B.7) Substituting representative values of y and as 0. dy = 250 m.M.5) (A. Pe 1: (A. Sh = . To estimate the mass transfer coe cient inside the yarn. Transport Phenomena. B. Muttzall. K. KT can be determined using . G. Sc¿0:7. A single yarn can approximately be described as a cylinder. we ÿnd that Sh ≈ 25 and from the deÿnition of the Sherwood number it follows that k2 ≈ 1:4×10−2 cm s−1 .. J. Moholkar. (B) the three stages of mass transfer in textile materials and the typical concentration proÿles of the di using substance. Blake. Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 TEXTILE INTRA YARN PORE YARNS INTER YARN PORE FIBRES (A) YARN BULK familiar relationship in the mass transfer theory 1 1 1 1 = + + : (A. respectively. Transient cavities near boundaries.K. gives De ≈ 1 × 10−10 m2 s−1 . who have shown that for longer contact times [Dt=d2 ] the Sherwood number ∼5:8. Now to estimate the relative magnitudes of k1 and k2 we use the mass transfer correlations developed for a packed bed. This also implies that enhancement of the mass transfer in the textile would necessitate conversion of the intra-yarn di usion process into the faster convection. Part 1. we choose the following correlation for the ow past long cylinders perpendicular to the ow: Sh = 0:42Sc1=5 + 0:57Re1=2 Sc1=3 .. D = 1 × 10−9 m2 s−1 . Doherty..4 and 2.C. k2 and k3 as the individual mass transfer coe cients for the ÿrst. we substitute the following representative values in the above correlation: v = 2 cm s−1 . we make use of the analysis given by Matsui et al. (1978). The mass transfer mechanism in the textiles: (A) schematic diagram indicating the general structure and dual porosity of the textiles.J. mass transfer from the inter-yarn region to the boundary layer between the bulk liquid and the textile surface. 2. A. the di usional mass transfer in the yarn is the rate-controlling step in the overall mass transfer in the textile. = 0:5 and Sc ≈ 1000. which occur in series.4) KT k1 k2 k3 We make a simplifying assumption that the mass transfer in the third step (from the boundary layer to the bulk) is very fast compared to the other steps. Chichester. 1¡Re¡104 . van Beek. the mass transfer in these textile materials can be divided into three distinct steps. Wiley. Substituting the above values. Taib. L = 1000 kg m−3 . Pe = : D D C r Stagnant Intra-yarn region Boundary layer Inter-yarn region Convective shell in the yarn (B) Fig. as shown in Fig.. 479–497. the e ective di usion coe cient inside the yarn is smaller than the actual di usion coe cient.2.310 V. Due to the porous y structure of the yarns. mass transfer from the stagnant intra-yarn region to the inter-yarn region through the convective shell on the periphery of the yarn. Journal of Fluid Mechanics 170.6) The non-dimensional numbers are deÿned as kdy vdy L vdy Re = . References Beek.. In order to assess the relative contribution of each of these steps to the overall mass transfer process. we use the correction given by Rietema (1976) to estimate the e ective di usion coe cient in a porous medium y De = 2 D: (A. As such. The ratio of k1 and k2 gives us an idea of the mass transfer process in the textiles: 97% of the mass transfer resistance is determined by the di usion process inside the yarns. W. M. J.M. 9.W.

2001. Yachmenev. 1988.. Kobayashi.N.D.S.. H.. Lambert... Moholkar.. McCall. Washing of fabrics Part 1: Washing of cotton fabrics and its theoretical analysis. G.K.. E. American Dyestu Reporter 77 (10). 593–601. Journal of the Textile Machinery Society of Japan 24 (2).D. Ph. M. Thresholds for transient cavitation produced by pulsed ultrasound in a controlled nuclei environment. Thesis.. M.. Lambert.. V. New York. Delft University Press. Moholkar. W. 1954. A. Textile Chemist and Colorist 30 (11). Journal of Basic Engineering (Transaction of ASME-D) 83. Rietema. Crank. Mock. University of Twente. P.. Dynamics of laser-induced cavitation bubbles near elastic boundaries: in uence of elastic modulus. AATCC Review 2 (2).H. Conduction of Heat in Solids.N. Book of Papers. 1998. 254 –262. P. J.. Ellis. Journal of Fluid Mechanics 361. Cavitation erosion by single laser produced bubbles. A.....L. North Carolina State University.. Grady. Direct Red 81 in cellulose. 283–314.M.S.V. A. P.S. V.. M. E.G. Fundamental interactions of ultrasound with cotton ÿbers and vat dyes. Carlough... W.B.C. F.C.E. Koji. Vogel. On the mechanism of cavitation damage by non-hemispherical cavities collapsing in contact with a solid boundary. Acoustics: An Introduction to its Physical Principals and Applications. Journal of Fluid Mechanics 433. Acoustical Society of America. A. 1998.S. R. Zeitschrift fur Technische Physik 12. 1998. Thesis.. Moholkar. Flow resistance of textile materials.G.. Thakore. Hite. Enschede. Rathi. E. F. 1959. J.G. Matsui. Phillip.G. Het Spectrum. Nahen. Intensiÿcation of Textile Treatments: Sonoprocess Engineering. H. Lauterborn. Mock.T. Warmoeskerken. Atlanta. Tomita.C. 2002. Oxford.M. Ein beitrag zur optik der farbanstriche. 1975. 1990.. 1986. 1978. Janssen. V. AATCC International Conference and Exhibition 1997. M. pp.L. GA.A.. D.. Gilmore.. Warmoeskerken.. 34–37. 3919–3923. Carslaw.M. V. Mechanistic Aspects and optimization of ultrasonic washing. Warmoeskerken / Chemical Engineering Science 59 (2004) 299 – 311 Brujan. S.. 535–564. A.G.. K.. Physical Properties of Textile Fibers. N. 1976.S. 1931. Willemse.M. 1997. Pierce. 15–16. M. A. L. Textile Chemist and Colorist and American Dyestu Reporter 1 (1)..M. A. 26 – 4.H. Smith. Oxford. McCall.J.. Schmidt..C....A.F. H. K.E. Ultrasound aided open width washing of mercerized 100% cotton twill fabric. 75–116. C. Naude.. 2002. 1999.S. 648–656. Mechanisms of impulsive pressure generation and damage pit formation by bubble collapse. Yachmenev. 1998. C. Patel. California Institute of Technology... Grady.R. Journal of Fluid Mechanics 169. R.. D. EP9401241. C. 1997. 1989.. Hearle.H.. Clarendon Press.. K. Holland.J.E. Solvent and ultrasonic alternatives to perchloroethylene drycleaning of textiles.-A. Study of the in uence of ultrasound on enzymatic treatment of cotton fabric. Y. M. Munk. Gooijer. Jaeger. . 41–49. 1997. 1997. Journal of the Acoustical Society of America 88. Warmoeskerken. Apfel. J.. The Mathematics of Di usion.. P.D. European Patent No..C. Blanchard. 11–18. F.. 47–51. Delft..M... R.. P. Fysische Transport en Overdrachtsverschijnselen. G. The e ects of ultrasound and the di usion coe cient of C.E. K.B. Blanchard. 1994. Use of ultrasound energy in the enzymatic treatment of cotton fabric. Manchester. Ph. Industrial and Engineering Chemistry Research 37 (10).W... Clarendon Press.P. Twente University Press. van der Vlist. 2059–2069. Kubelka. 311 Morton.I. The Textile Institute. Shima. Transport Phenomena Data Companion. Utrecht. Hydrodynamic Laboratory Report. 1961.G. Klutz.