You are on page 1of 9

Colloids and Surfaces A: Physicochem. Eng.

Aspects 210 (2002) 277 /285


www.elsevier.com/locate/colsurfa

Laundry process intensification by ultrasound


M.M.C.G. Warmoeskerken a,b,*, P. van der Vlist a, V.S. Moholkar b,
V.A. Nierstrasz b
b

a
Unilever Research and Development Vlaardingen, Olivier van Noortlaan 120, 3133 AT Vlaardingen, The Netherlands
Textile Technology Group, Department of Chemical Engineering, University of Twente, P.O. Box 217, 7500 AE Enschede, The
Netherlands

Abstract
In domestic textile laundering processes, mass transfer and mass transport are often rate limiting. Therefore, these
processes require a long processing time, large amounts of water and chemicals, and they are energy consuming. In
most of these processes, diffusion and convection in the inter-yarn and intra-yarn pores of the fabric are limiting mass
transport mechanisms. Intensification of mass transport, preferentially in the intra-yarn pores, is the key in the
improvement of the efficiency of wet textile processes. Conventional methods of intensification of mass transport (e.g.
operation at elevated temperatures) are not always feasible due to the undesired side effects such as fabric damage.
Increasing the flow rate does not deliver the desired effect due to the multi-porous complex structure of textile materials.
Van der Donck et al. [Tenside Surf. Det. 35 (1998) 119; 36 (1999) 222] reported that the deformation of yarns by placing
a fabric in a pulsating flow or repeated mechanical elongation of the yarns improved mass transport. However, the
additional mass transport caused by deformation is limited in practice. Power ultrasound is a promising technique to
accelerate mass transport in textile materials. Several papers appeared in this field, which report an improvement in
energy efficiency and processing time of the wet textile processes in the presence of ultrasound. In this paper, the
different time and length scales are discussed in the intensification of the mass transport in laundry processes in the
presence of ultrasound and compared with more conventional processes. It has been concluded that the characteristic
mass transport rates in textiles can be increased by a factor of 6 applying ultrasound.
# 2002 Elsevier Science B.V. All rights reserved.
Keywords: Process intensification; Power ultrasound; Enhanced mass transport; Textile cleaning; Textile laundering

1. Introduction
An ever recurring subject in textile washing
technology is the application of power ultrasonics

* Corresponding author. Tel.: /31-10-460-5135; fax: /3110-460-5192


E-mail
address:
marijn.warmoeskerken@unilever.com
(M.M.C.G. Warmoeskerken).

in laundry machines to accelerate the mass transfer


in textile materials. Most of the domestic laundry
processes in Western Europe are carried out
nowadays in machines of the horizontal rotating
drum type. Although the results in those machines
are generally fully acceptable, the efficiency of the
washing process must be regarded as far from
optimal. It has since long been speculated that
using ultrasound waves could be a much more

0927-7757/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 7 - 7 7 5 7 ( 0 2 ) 0 0 3 7 2 - 2

278

M.M.C.G. Warmoeskerken et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 210 (2002) 277 /285

effective way of supplying mechanical energy, in


analogy with its numerous successful applications
in the area of hard surface cleaning. In fact, in the
patent literature many examples can be found,
especially from Japanese washing machine manufacturers where claims are made for improved
cleaning processes based on ultrasonics. In general, however, the evidence provided to substantiate such claims is very weak, and it is not
surprising to see that so far none of these innovations has reached the stage of commercialisation.
Recently, the Japanese Sanyo Company introduced a new washing machine that uses ultrasound
waves. However, we were not able to measure any
ultrasonic activity in that machine and a further
inspection taught us that there was not an ultrasonic transducer present. So the ultrasonic claims
made by Sanyo are without any foundation.
Clearly, what is really lacking here is a scientifically based understanding of the acoustic principles involved in a textile /water system, in
particular with regard to their interactions with
the fundamentals of the soil loosening and soil
transport processes. This study was aimed at
achieving a better knowledge of the acceleration
of mass transfer in the intensification of textile
laundry processes by ultrasound compared with
conventional methods.

2. The laundry process


The cleaning obtained in a laundry process is the
result of synergistic actions between the so-called
Sinner parameters, being the mechanical energy,
chemical energy, thermal energy and time. These
factors have to perform a simple separation
process in which soil is removed from a textile
substrate. In that process, two important steps can
be distinguished: the soil loosening step and the
soil transfer step. In the soil loosening step, the
physical binding forces between the soil and the
substrate are broken up. In the subsequent soil
transfer step, the loosened soil is transported from
the substrate to the wash liquor. A comprehensive
review on the physical/chemical aspects of soil
loosening is given by Carroll [3]. Recent results in
this area are on the influence of the dynamics of

the surface tension [4] and on the effect of


surfactants on capillary flow phenomena in textile
washing [5].
In both steps, soil loosening and soil transfer,
some liquid flow is indispensable to deliver the
required kinetic energy. The mechanical energy in
a washing machine, created by the rotation of the
drum is aimed at delivering the kinetic energy for
the soil loosening and soil transfer processes. The
focus in this study was on this aspect of textile
washing.
It is generally thought that the drum rotation in
a washing machine creates enough liquid flow to
assist the loosening and transfer of soil properly.
However, since textiles have a very complex
porous structure of fibres and yarns, the translation of the drum rotation into a flow within the
textile pores is very complicated. A woven fabric is
a bi-porous medium with relatively large interyarn pores, the pores between the yarns and
relatively small intra-yarn pores, the pores between
the fibres. Van den Brekel [6] and later Gooijer [7]
showed that most of the liquid flow is passing the
yarns through the inter-yarn pores without penetrating into the intra-yarn pores. This is because
the flow resistance in the relatively small intrayarn pores is much higher than the resistance in
the relatively large inter-yarn pores as will be
explained in more detail in Section 3. Later
Warmoeskerken and Boom [8] introduced a stagnant core and a convective shell model to describe
the flow through the yarns. The stagnant core in
the yarn is the area in which there is no flow at all.
The convective shell is the outer area of the yarn in
which the flow penetrates to some extent. The
model is schematically given in Fig. 1. The transfer
processes in the stagnant core are based on
molecular diffusion while the transport processes
in the outer convective shell are driven by convective diffusion. Since convective diffusion is
much faster than molecular diffusion, the rate of
soil removal will be determined by the size of the
stagnant core. The smaller this core, the faster the
removal process will be. This means that the task
of the mechanical energy in a wash process can be
defined as making the stagnant cores in the yarns
as small as possible. The mechanism for this is
deforming the porous structure in textiles, creating

M.M.C.G. Warmoeskerken et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 210 (2002) 277 /285

279

yarn pores is of the order of magnitude of 0.1 mm


s1 where the order of magnitude for the liquid
flow in the intra-yarn is of only 0.1 mm s 1. These
differences in liquid flow rate can be explained by
the characteristics of the porous material. The
equation of Darcy is used to describe the laminar
flow through the porous material [7,10,11]:
 
K dP
vs 
;
(1)
h dx

Fig. 1. Liquid flow around and through a textile yarn. The dots
in the figure represent the fibres in a yarn.

a squeezing effect. During the tumbling action in


the rotating drum, the yarns in the fabrics will be
deformed by bending, twist, stretch and pressure
forces, as described by Ganguli and van Eendenburg [9]. Van der Donck et al. [1,2] showed in
model experiments that salt release from porous
yarns is accelerated to a large extent if the yarns
are stretched. Their findings support also the idea
of minimising the stagnant core in yarns. So it can
be concluded that the main task of the mechanical
energy in a wash process is creating deformation of
the textiles resulting in decreasing stagnant yarn
cores, rather than producing at random macroscale hydrodynamic phenomena.
In the following section, it will be shown that the
role of ultrasound in enhancing the cleaning
performance in fabric washing process can also
be explained in terms of intensification of mass
transfer by reducing the stagnant cores in the
yarns.

where vs is the superficial liquid velocity, K the


permeability, h the dynamic viscosity and dP /dx
the pressure gradient. Rewriting Eq. (1) results in
the Kozeny model:
 
1
o3
1 dP
vs 
;
(2)
2a2 (1  o)2 a2 h dx
where a is the tortuosity of the porous medium, a
is 4/d with d being the diameter of the pore and o
the porosity. With the use of Eqs. (1) and (2), the
permeability coefficient K can be written as
K

1
o3
;
K0 (1  o)2 a2

(3)

where K0 is the Kozeny constant. The Kozeny


constant is an empirical parameter that needs to be
determined experimentally and is a measure for the
tortuosity and the orientation of the porous
system. The contribution of both the inter-yarn
permeability (Kinter-yarn) and the intra-yarn permeability (Kintra-yarn) can describe the overall permeability of the textile material:
K Kinter-yarn (1o inter-yarn )Kintra-yarn :

3. Time and length scales in laundry


In textile laundry processes, different time and
length scales can de defined. Three different length
scales can be defined for the textile itself: the plug/
fabric level (dplug /0.1 m), the yarn level (dyarn /
200 mm) and the fibre level (dfibre /10 mm). These
three length scales can be related to the bulk liquid
flow, the flow in the inter-yarn pores and the flow
in the intra-yarn pores, respectively. Van den
Brekel [6] found that the liquid flow in the inter-

(4)

Assuming that K0 is constant for the pores


between the yarn and the fibres, the relative
contribution of the inter-yarn permeability and
the intra-yarn permeability can be estimated as
follows:
Kinter-yarn
Kintra-yarn

o 3inter-yarn (1  o intra-yarn )2 a2intra-yarn


o 3intra-yarn (1  o inter-yarn )2 a2inter-yarn

(5)

Substituting representative values for a textile


material for the different parameters, we find that

280

M.M.C.G. Warmoeskerken et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 210 (2002) 277 /285

Kinter-yarn
:2002000:
Kintra-yarn

(6)

This illustrates that the convection and diffusion in


the intra-yarn pores mainly determine the processing time in a laundry process, since liquid flow
and thus mass transport mainly occurs in the interyarn pores.
To describe mass transport from the yarns by
diffusion, we assume that a textile material consists of infinite porous cylindrical yarns with
radius r. Non-stationary diffusion of a component
C in a porous yarn can then be described using
Ficks second law:


@C Deff @
@C

r
:
(7)
@t
@r
r @r
The effective diffusion coefficient (Deff) for a
porous medium is obtained by correcting the
diffusion coefficient for the porosity and the
tortuosity (a ) of the yarn [10]:
Deff 

o intra-yarn
D:
a

(8)

The diffusion coefficients for particles of different


size are calculated using the Stokes /Einstein
relation. With the appropriate boundary conditions
t0 : 0 5r5R; C C0 ;
@C
t0 : r 0;
 0;
@r
r R;
C C ;
and assuming that there is no interaction between
the diffusing particles and the textile material, the
time needed to remove 90% of the initial amount
C0 of a particle from a porous yarn (dyarn /0.5
mm, o intra-yarn /0.5, a /2) by diffusion (Fig. 2)
can be calculated. If it is assumed, for the sake of
simplicity, an enzyme to be completely spherical
with a characteristic diameter of 1/5 nm the time
needed for removal of 90% by diffusion alone is
15 /700 s. For carbon black particles, as present,
e.g. on the textile test cloth EMPA 101, with a
characteristic diameter of 0.1 mm the time is 1000 s,
and for silica particles with an average diameter of
5 mm the time to accomplish 90% removal by

diffusion only is 70 000 s. These diffusion times


exceed the time needed in true laundry operations.
In laundry processes, diffusion in the intra-yarn
pores is rate limiting and mass transport needs to
be increased due to the mechanical action that
leads to deformation of the plug and thereby the
yarns so that liquid is squeezed from the space
between the fibres.

4. Strategies to increase mass transport


The strategies to increase mass transport in
textile materials discussed in this paper are the
deformation of yarns, creating a squeezing effect
and the application of power ultrasound.
4.1. Deformation of yarns
Van der Donck et al. [1,2] reported that the
deformation of yarns by placing a fabric in a
pulsating flow or repeated deformation through
mechanical elongation of the yarns improved mass
transport compared to diffusion alone. When a
yarn is elongated, an amount of liquid is squeezed
out of the yarn (Fig. 3). Van der Donck et al. [1]
measured the increase of conductivity caused by
the release of magnesium sulphate in time from a
cotton yarn impregnated with magnesium sulphate. The magnesium sulphate was squeezed
out of the yarn due to the repeated elongation of
that yarn. In practice this can be realised through
tumbling of the cloth in domestic laundry machines. To describe the phenomena observed, Van
der Donck et al. [1] used the dimensionless Fourier
number that gives the ratio between the diffusion
and the yarn diameter [12]:
Fo

Deff t
2
dyarn

(9)

Their analysis results in a qualitative logarithmic


relation between the soil release, the Fourier
number and the additional soil removed due to
the elongation of the yarn. However, Etters [13 /
15] proposed a mathematically simple empirical
equation that matches the exact solution to

M.M.C.G. Warmoeskerken et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 210 (2002) 277 /285

281

Fig. 2. Time needed to remove 90% of particles from a yarn by diffusion as a function of the particle diameter.

describe diffusion-controlled mass transport in


yarns.

C t 

1
R

g C dr;
t

(11)

C t  C0
[1exp(a(4Fo)b )]c ;
C  C0

(10)

with C t being the average concentration in the


porous yarn,

and a , b and c are constants depending on the


characteristics of the system and are given in [14].
For example, if the thickness of the diffusion
boundary layer between the bulk and the yarn
approaches zero, then a /5.530, b/1.0279 and
c/0.3341. Eq. (10) has been modified in such a

Fig. 3. Liquid flow in a yarn during elongation and relaxation of the yarn.

282

M.M.C.G. Warmoeskerken et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 210 (2002) 277 /285

way that it is possible to describe the influence of


the deformation of the yarn on mass transport

 
 
4Deff t b c
 1exp a
2
dyarn
C  C0


o intra-yarn  o deformation ft

;
o intra-yarn
C t  C0

(12)

where o deformation is the volume fraction squeezed


out of the yarn during deformation and f the
frequency of the deformation. With this modified
equation of Etters, it was possible to compare the
experiments of other authors with our own experiments to increase mass transport. In domestic
laundry processes, the deformation of the plug and
therefore the yarns will be related to the rotation

Fig. 4. Calculated soil removal as function of time according to


Eq. (12) for different deformation frequencies (Deff /4.8/
10 10 m2 s 1, dyarn /3.5 /10 4 m, eintra-yarn /0.5,
edeformation /0.03, a/2.440, b /1.045, c/0.863). Van der
Donck et al. [1] used a frequency of 0.08 and 2.12 Hz. The
slopes of the lines have been calculated as /0.045 and /0.17.
The upper graph shows the relative soil removal. The lower
graph is a semi-log plot of the relative soil retention.

velocity of the drum. In the case of domestic


laundry processes, a deformation frequency between 0 and 2.5 Hz seems realistic. In Fig. 4, the
results are shown using the modified equation
from Etters (Eq. (12)) with the parameters from
Van der Donck et al. [1]. Different constants for a ,
b and c were used to correct for the low amount of
mixing in their experimental set up. The results
compare very well with the results described by
Van der Donck et al. [1]; however, some deviation
is seen at higher frequencies at the end of the
experiment. This has to do with the fact that in the
experimental set up the medium is not infinite.
4.2. Power ultrasound
Ultrasound is a longitudinal pressure wave in
the frequency range above 25 kHz. As the sound
wave passes through water in the form of compression and rarefaction cycles, the average distance between the water molecules varies. If the
pressure amplitude of the sound is sufficiently
large, then the distance between the adjacent
molecules can exceed the critical molecular distance during the rarefaction cycle. At that moment, a new liquid surface is created in the form of
voids. This phenomenon is called acoustic cavitation. The theoretical pressure amplitude to cause
cavitation in water is approximately 1500 bar.
However, in practice acoustic cavitation occurs at
a far lower pressure amplitude, less then 5 bar.
This is due to the presence of weak spots in the
liquid in the form of tiny micro-bubbles that lower
the tensile strength of the liquid. Once formed, the
bubbles can re-dissolve into the liquid, they may
float away, or, depending on the size of the
bubbles, they may grow and shrink in phase with
the oscillating ultrasonic field. This process of
growing and recompressing of bubbles is called
stable cavitation. If the sound field is sufficiently
intense, bubbles of a specific initial size can grow
so quickly and acquire such momentum that the
compression wave that immediately follows after
the rarefaction phase is not able any more to stop
the bubbles growing. Once out of phase with the
ultrasonic field, however, the bubbles are no
longer stable. The pressure within the bubble is
not high enough to sustain the size of the bubble

M.M.C.G. Warmoeskerken et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 210 (2002) 277 /285

283

and, driven by the next compression wave, the


bubbles implodes. This latter process is called
transient cavitation.
In liquids, the collapsing bubbles remain spherical because the ultrasonic waves are uniform.
However, if a transient acoustic bubble collapses
near a solid boundary, the bubble will implode
asymmetrically generating jets of liquid directed
towards the surface of the solid boundary. The
micro-jets resulting from collapsing bubbles at a
solid boundary are counted for the cleaning effect
of ultrasonic waves. This acoustic cavitation
process has been described by many authors like
Apfel [16], Suslick [17] and Neppiras [18]. Its
application in laundry processes is reported recently by Yachmenev et al. [19,20] and by Moholkar et al. [21 /25].

5. Experimental techniques
The experiments have been performed in a
commercially available ultrasonic bath from
Elma GmbH. The volume of the bath was 1 l
and the operating frequency of the ultrasonic
transducer was 33 kHz. In this bath, several saltrinsing experiments have been done by using
cotton test swatches that contained an amount of
dried-on sodium chloride as a tracer. The release
of the salt from the swatches in time was monitored by conductivity measurements of the bulk
fluid. The swatches were kept in slow motion in
the bath by stirring the liquid very gently by a
laboratory impeller.

Fig. 5. Experimental salt-rinsing results with and without the


application of ultrasonics. The slopes of the lines have been
calculated as /0.11 without ultrasound and /0.66 with
ultrasound. (a) The relative salt removal and (b) semi-log plot
of the relative salt removal.

ultrasound can be expressed by the ratio of the


slopes of the lines in Fig. 5(b). For the present
results, this ratio is approximately 6. This means
that in the present experimental set up ultrasound
accelerates the removal of salt by a factor of 6 or

6. Experimental results and discussion


The results of the rinsing experiments are shown
in Fig. 5. Fig. 5(a) shows the decrease of the
amount of salt in the cotton swatches as a function
of time for two cases. The first case without
ultrasound and the second case with ultrasound.
From those results, it is clear that the presence of
ultrasound has an enormous effect on the release
rate of the salt from the swatches. Fig. 5(b) shows
the same results but now plotted on a semi-log
scale. The intensification of mass transfer by

Fig. 6. Comparison of different slopes obtained by textile


deformation and with and without ultrasound.

284

M.M.C.G. Warmoeskerken et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 210 (2002) 277 /285

recalculated does it mean that the stagnant core in


the yarn is reduced by a factor of 2.5.
It is interesting to compare the results obtained
here with the data obtained by Van der Donck et
al. [1]. This has been done in Fig. 6 in which the
slopes of different release profiles are given. The
first two values refer to the work of Van der
Donck et al. The other two values refer to the
results obtained in this study. From this figure, it is
clear that ultrasound gives the best results of all.
So it can be concluded that ultrasound is a very
promising technique to intensify the mass transfer
in laundry processes. The fact that this technique is
not implemented nowadays is caused by the
complexity of the system. The present experiments
are done in a system in which some swatches were
fully submerged in water, and so the liquid /cloth
ratio was unrealistically high. The water level in
European washing machines has been decreased
during the last 10 years due to the environmental
reasons. This has resulted in a very low liquid/
cloth ratio whereby the wash load is not submerged at all. That makes it very difficult to apply
power ultrasonics in our current washing machines. A complete redesign of the washing process
and machine is necessary to make the application
of ultrasound in a laundry processes successful.
Moreover, it has been found that the presence of
air in the system has a dominant effect on the
ultrasonic washing results. The results in deaerated water were remarkably better than those
obtained in normal tap water. This can be
explained in terms of the acoustic impedance as
recently found by Moholkar [21] and Moholkar
and Warmoeskerken [22]. Although a simple
procedure to de-gas water has been developed,
based on stripping out the air by bubbling carbon
dioxide through the water and converting the
dissolved CO2 gas by increasing the pH [26], it is
not simple to apply this principle in a domestic
washing machine. Besides it is expected that
ultrasonics will have a negative effect in terms of
damage of the textiles. So although the ultrasonic
technique is very promising, a lot of research is still
needed before the technology can be applied in a
washing machine.

7. Conclusions
It has been shown that an improvement in mass
transfer in textile materials can be translated into a
reduction of the diffusional stagnant core of textile
yarns. A reduction of this stagnant core can be
achieved in several ways. Deformation of the
porous matrix of textiles, creating a squeezing
effect is the most obvious way. A more effective
and efficient way is the application of power
ultrasound. It has been found that in a normal
ultrasonic bath, the rinsing of salt from textiles can
be speeded up with a factor of 6. Recalculated this
means that the stagnant core in the yarn is reduced
by a factor of 2.5. Comparison with literature data
showed that ultrasound is far most the best way to
intensify the mass transport in textile materials.

Acknowledgements
The Textile Technology Group at the University
of Twente acknowledges the financial support of
the Foundation Technology of Structured Materials in The Netherlands and of the Dutch Ministry
of Economic Affairs.

References
[1] J.C.J. Van der Donck, A. So, G. Frens, Tenside Surf. Det.
35 (1998) 119 /122.
[2] J.C.J. Van der Donck, Tenside Surf. Det. 36 (1999) 222 /
224.
[3] B.J. Carroll, Colloid. Surf. A 74 (1993) 131 /167.
[4] D.J.M. Bergink-Martens, G. Frens, Tenside Surf. Det. 34
(1997) 263 /266.
[5] A.M.D.E. Timmerman, Ph.D. Thesis, TU Delft, The
Netherlands, 2002.
[6] L.D.M. Van den Brekel, Ph.D. Thesis, TU Delft, The
Netherlands, 1987.
[7] H. Gooijer, Ph.D. Thesis, University of Twente, The
Netherlands, 1998.
[8] M.M.C.G. Warmoeskerken, R.M. Boom, in: Proceedings
of the 90th AOCS Annual Meeting and Expo, Orlando,
FL, May 9 /12, 1999.
[9] K.L. Ganguli, J. van Eendenburg, Textile Res. J. 50 (1980)
428 /432.
[10] K. Rietema, Fysische Transport- en Overdrachtsverschijnselen, Het Spectrum, Utrecht, 1976.

M.M.C.G. Warmoeskerken et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 210 (2002) 277 /285
[11] R.B. Bird, W.E. Stewart, E.N. Lightfoot, Transport
Phenomena, Wiley, New York, 1960.
[12] L.P.B.M. Janssen, M.M.C.G. Warmoeskerken, Transport
Phenomena and Data Companion, Edward Arnold, London, 1987.
[13] J.N. Etters, Text. Chem. Colorist 12 (1980) 140 /145.
[14] J.N. Etters, J. Appl. Polym. Sci. 58 (1995) 2325 /2327.
[15] J.N. Etters, J. Appl. Polym. Sci. 63 (1997) 1237 /1242.
[16] R.E. Apfel, in: P.D. Edmonds (Ed.), Methods in Experimental Physics, vol. 19, Academic Press, New York, 1981,
pp. 355 /413.
[17] K. Suslick (Ed.), Ultrasound: Its Chemical, Physical and
Biological Effects, VCH, New York, 1988.
[18] E.A. Neppiras, Physics reports, Rev. Sect. Phys. Lett. 61
(1980) 159 /251.
[19] V.G. Yachmenev, N.R. Bertoniere, J. Blanchard, Textile
Res. J. 71 (2001) 527 /533.

285

[20] V.G. Yachmenev, E.J. Blanchard, A.H. Lambert, Textile


Colorist Chemist Am. Dyestuff Rep. 1 (1999) 47 /51.
[21] V.S. Moholkar, Ph.D. Thesis, University of Twente, The
Netherlands, 2002.
[22] V.S. Moholkar, M.M.C.G. Warmoeskerken, AATCC
Rev. 2 (2002) 34 /37.
[23] V.S. Moholkar, M. Huitema, S. Rekveld, M.M.C.G.
Warmoeskerken, Chem. Eng. Sci. 57 (2002) 617 /629.
[24] V.S. Moholkar, A.B. Pandit, Chem. Eng. Sci. 56 (2001)
6295 /6302.
[25] V.S. Moholkar, M.M.C.G. Warmoeskerken, in: Proceedings of the First AUTEX Conference (Technitex), Povoa
do Varzim, Portugal, June 26 /29, 2001, pp. 204 /213.
[26] P. van der Vlist, M.M.C.G. Warmoeskerken, S. Willemse,
European Patent EP9401241, 1994.