Full text of DOI:10.1111/j.1478-4408.2002.tb00140.x

Blade coating of fabrics: rheology and fluid
penetration
Coloration
Technology
F Davard and D Dupuis
Laboratoire de Physique et Mécanique Textiles, Universite de Haute-Alsace, 11 rue Alfred
Werner, 68093 Mulhouse Cedex, France
The coating of fabrics generally involves the use of complex fluids as coating compounds. During the
coating process, these compounds are submitted to various stresses. This paper investigates the
influence of the rheological properties of the fluid on its penetration into the fabric. Experiments were
performed with several kinds of model polymer solutions on a cotton fabric. Penetration coefficients
were obtained from reflectance measurements on the face and reverse sides of the fabric. The results
show that there are competitive effects between shear-thinning which promotes penetration and
elasticity which tends to reduce it.
Society of Dyers and Colourists
Different techniques can be used such as blade coating
on air, on a roll, on a table or over a blanket. The choice
of the technique can depend on the substrate, on the
expected final result and on the coating medium.
Introduction
Fabric coating involves a thin layer of coating medium
(generally
a complex fluid, i.e. polymer solution,
suspensions of polymer solutions or filled foams) spread
over the textile surface. This coating medium must have
a good affinity and mechanical adhesion for the textile, the
effectiveness of which is improved when the coating
medium penetrates deep into the pores of the fabric. This
must be a controlled phenomenon, however, as the com-
pound must not be visible on the reverse of the fabric. The
capillarity and the pressure generated under the blade are
the two factors which controls this process. Since the fluid
is submitted to stresses during the process, its rheological
properties do not play a negligible role in the penetration.
Depending on the industry (paper or textile
manufacturing) various methods and geometries of coating
process are used, e.g. roll, kiss, withdrawal, curtain or
blade coating, etc. In the blade coating process, the solid
moves at constant velocity U under the blade according to
Figure 1. The reservoir of fluid is located upstream against
the front wall of the blade. The main geometrical
parameters are defined in Figure 1 for the most simple
shape of blade: h2 is the height, l is the width, θ is the angle
between the blade and the horizontal plane and α is the
angle of the blade.
Several experimental techniques have been developed
for the paper manufacturing industry to study the level of
penetration of compounds into paper. The Warren test is
based on the measurement of electric conductivity which
varies with moisture [1]. Ultrasonic techniques can also be
used to measure the migration of water into paper. Bristow
developed an experimental device, the Bristow wheel, to
determine the penetration into paper due to capillarity over
a short timescale [2]. This device was then modified to also
study the influence of temperature and pressure [3,4]. In
fact, it has been demonstrated that pressure plays an
important role on penetration [5]. Salminen showed that
the influence of the surface tension of the fluid is
important only when no pressure is applied [6].
For textile substrates, previous experiments have been
concerned with capillary penetration. Miller and Tyomkin
used a gravimetric technique to determine the quantity of
liquid absorbed by a fabric in the thickness direction [7].
Optical techniques based on reflectance measurements
have also proved suitable [8]. Experiments developed by
Clark et al. for paper allowed the measurement of both the
capillary and forced penetration of a dye solution, and they
went on to relate the measured reflectance to the thickness
of coloured paper by a formula derived from the Kubelka–
Munk equation [9].
Knowing the level of penetration is also of great
importance for the printing of fabrics. Reflectance
measurements have also been used by Lemmen and
Wassink in the case of printing textiles with foams [8].
From the Kubelka–Munk equation, a penetration degree
related to the reflectances of the coloured and raw material
on both the face and reverse sides was defined. Vu et al.
also used optical methods to evaluate the penetration of a
printing paste into a fabric [10]. They measured two
parameters: the colouristic efficiency (on the face side of
l
α
θ
U
h2
Figure 1 Geometry of the blade coating system
Web ref: 20020205
© Color. Technol., 118 (2002) 69

the fabric) and the transpercement (on the reverse side) (i.e.
the amount of paste which permeates the fabric). It is
shown that the quality of printing depends on the
rheological properties of the pastes. In particular, the
penetration is weaker for more elastic pastes.
The aim of this paper is to present experimental results
obtained by coating a cotton fabric with model fluids
characterised by a large variety of rheological behaviours:
Newtonian, shear-thinning with variable elasticity and
highly elastic with constant viscosity.
pressure. The dynamic measurement was performed on the
Kawabata evaluation system (KES FB3 Compression Tester).
The thickness was controlled and the pressure continuously
measured during the compression of the sample.
It is evident that pressure has a marked effect on the
thickness of the fabric. The decrease in thickness with
increased pressure widely depends on the mode of
application. During the coating process, the time that the
fabric is under the blade, and therefore under pressure, is
extremely short (ca. 0.15 s). So the dynamic determination
of the pressure seems to be more suitable and was the
method of choice for this study.
The penetration of the fluid into the fabric was
evaluated by reflectance measurements. With a spectro-
photometer (Microflash, Datacolor), the reflectance was
measured on the face and reverse sides of the coated
sample, respectively called Rmin and R, while the
reflectance of the raw fabric is Rmax. The penetration
coefficient P is defined as shown in Eqn 2:
Experimental
Coating process
Application
The apparatus used in this study was a blade coater
designed for laboratory use by Roaches International in
which pieces of fabric of rectangular shape are fixed on a
table frame. The width of the machine was 35 cm and the
fluid was applied onto the substrate using a roller bank
concept. The velocity of the substrate was maintained at
a constant speed of 74 mm/s. The height of the blade was
also constant at 0.5 mm and the angle θ was 22°. The other
geometrical parameters were: l, 10 mm and α, 40°. After
Rmax - R
Rmax - Rmin
P =
™ 100%
(2)
Thus, when P is zero there is no penetration and when P
= 100% the face and reverse sides are coloured in the same
manner, i.e. the penetration is maximum.
The reflectances were measured in five places for
wavelengths at maximum absorption (λmax) in the range
400–700 nm in 10 nm steps. The specular component of
the reflected light was eliminated. The measurements were
highly reproducible.
coating, the textile material was dried at 125 °C over 15
.
min. The nominal shear rate (γ) under the blade was equal
to 148 s–1 and can be defined by Eqn 1:
U/h2
(1)
Fabric parameters
The fabric used was plane cotton fabric (Steinheil-Dieterlen)
made of yarns of 34.3 tex for a count of 23/23 yarns/cm. The
diameter of the yarns measured by microscopy was 0.35
mm. The thickness of the fabric, which is an important
parameter, is difficult to evaluate since it depends on the
applied pressure. For this purpose, two techniques were
investigated (Figure 2). For static measurements, the sample
was placed between two metallic plates. By means of an
electromagnetic sensor, the distance between the plates was
determined. The top plate was loaded with a known mass
and the steady value of the thickness was obtained at a given
Coating medium
The experiments were performed with different polymer
solutions as the coating medium:
–
aqueous solutions of three carboxymethylcellulose
(CMC) polymers (Sigma-Aldrich) of three different
average molecular weights (90 000, 250 000 and
700 000 g/mol) at different concentrations; the poly-
mers are respectively called LV, MV and HV (i.e. low,
medium and high viscosity)
–
aqueous solutions of polyacrylamide (PAA) at different
concentrations
–
aqueous solutions of CMC with small amounts of the
high molecular weight PAA Separan AP30 (Dow
Chemical).
Static measurements
0.6
0.4
Dynamic measurements
Measurement of parameters for the polymer solutions
To study the penetration of the fluid in the fabric, i.e.
between the yarns and the fibres, the fluid was coloured
with a dye which did not diffuse (or diffused extremely
slowly) into the fibres. The polymer solutions were
therefore coloured with a blue disperse dye, Marine
Eryonil (Ciba), at a concentration of 0.1% w/v.
The shear viscosity was measured with a Bohlin
constant stress rheometer equipped with a cone and plate
geometry (diameter, 40 mm; angle of gap β, 4°). The
viscosity η is related to the shear stress σ by Eqn 3:
0.2
0
0
2
4
6
8
10
12
Pressure, kPa
/
Figure 2 Thickness of the fabric
70 © Color. Technol., 118 (2002)
(3)
Web ref: 20020205

.
n 1
In this instance, the shear rate γ is given by Eqn 4:
&
= K0
(7)
with K0
=
0an 1
&
= W /
(4)
where Ω is the rotation speed of the cone, the plate being
fixed. When an elastic liquid is sheared between two plates
or between a cone and a plate, it exerts a normal force on
the solid surfaces which tends to separate them. This
normal force F was determined with the Weissenberg R21
rheogoniometer equipped with a cone and plate geometry
(diameter, 50 mm; angle of gap β, 4°); F is related to the first
normal stress difference N1 by Eqn 5 (where R is the radius
of the plate):
100
10
1
0.95%
1.08%
1.15%
1.31%
1.45%
N1 = 2F/ R2
(5)
100
0.1
MV 2.6%
LV 8.2%
0.01
0.1
1
10
100
1000
HV 1.7%
Shear rate, s–1
10
Figure 4 Shear viscosity of the standard PAA solutions
1
1000
0.1
0.01
0.1
1
10
100
1000
Shear rate, s–1
100
0.95%
Figure 3 Shear viscosity of the three standard CMC polymer
solutions
1.08%
1.15%
1.31%
1.45%
.
In Figure 3, the viscosity is plotted vs. the shear rate γ for
10
the CMC solutions. These fluids are shear thinning and the
data can be fitted by the Carreau–Yasuda model (Eqn 6,
where a, n and m are parameters of the model):
10
100
1000
Shear rate, s–1
0
Figure 5 First normal stress difference for the standard PAA
solutions
=
1
n
(6)
m
m
&
1+ (a )
(
)
Table 2 Parameters of the power law model for
the PAA solutions
Table 1 Parameters of the Carreau–Yasuda model for the CMC
polymer solutions
Solutions
K0
n
Solution
η0 (Pa.s)
a (s)
m
n
(1–n)/m
PAA 1.45%
PAA 1.31%
PAA 1.15%
PAA 1.08%
PAA 0.95%
27.1
22.5
18.7
15.3
11.9
0.22
0.23
0.24
0.25
0.25
LV 8.0%
LV 8.2%
LV 9.4%
MV 2.3%
MV 2.6%
1.2
3.5
8.5
5.2
7.0
0.16
0.17
0.52
1.00
0.71
2.00
0.67
1.00
1.80
0.50
0.51
0.55
0.67
0.65
0.64
0.63
0.64
0.69
0.63
0.70
0.68
0.60
0.53
0.48
0.45
0.42
0.38
0.75
0.58
0.58
0.60
0.72
0.81
0.88
0.90
0.90
MV 3.0% 25
HV 1.1%
HV 1.3%
HV 1.7%
8.0
13
32
Figures 4 and 5 show, respectively, the shear viscosity and
.
the first normal stress difference N1 vs. γ for the PAA solu-
tions. In this case, the viscosity obeys a power law in the
range of shear rates which are explored. The parameters
of the power law are given in Table 2. The Weissenberg
number We is a measure of the elasticity of the fluid and
is defined by Eqn 8:
The different parameters are listed in Table 1. The Carreau-
Yasuda model can be approximated by a power law for the
highest shear rates (Eqn 7, where K0 is the consistency
value):
Web ref: 20020205
© Color. Technol., 118 (2002) 71

In their experiments, these two characteristic times were
of the same order of magnitude. So, it was suggested that
an influence of the elasticity on the transient response of
the fluids may explain the observed results on the coating
thickness. In this study, the relaxation times were similar
to the above ones but the residence time was much larger
(30 times greater). It is considered that steady rheological
properties (viscosity and elasticity) were a convenient way
of characterising the test fluids.
We = N1 /2
(8)
where σ is the shear stress. It is plotted on Figure 6 vs.
the shear rate; when We > 1 this results in high elasticity.
In order to determine the influence of the elasticity
independently of the viscosity, it is necessary to prepare
fluids with variable elasticity while keeping the same
viscosity. In a similar manner to the experiments of
Sullivan et al. [11,12] small amounts of PAA (0.025–0.2%)
were added to HV polymer solutions at 1.5%. The results
.
showed that when γ > 10 s–1, the viscosity curves were
Results and Discussion
superimposed while N1 significantly increased with the
PAA concentration. Thus, the associated Weissenberg
Figure 8 represents the reflectivity of the face and reverse
sides (Rmin and R, respectively) for three concentrations of
each of the CMC solutions. Except for the polymer LV at
the highest concentration, Rmin has low values independent
of the concentration, implying that the surface readily
becomes saturated by dye-containing fluid. On the other
hand, the reflectivity of the reverse side is sensitive, for
any given polymer, to the polymer concentration; the
higher the concentration, the higher the value of R. Thus,
a smaller amount of fluid penetrates through the fabric
which is therefore less coloured on the reverse side.
number (calculated from Eqn 8) plotted against shear rate
.
γ results in Figure 7.
It has been already underlined that, although the blade
coating is a steady state process, the fluid particles
experience a rapid Lagrangian acceleration under the blade
[12]. Sullivan et al. compared the residence time of a
particle under the blade to a relaxation time trelax of the
fluid evaluated by Eqn 9:
&
trelax = We /
(9)
50
6
40
30
20
10
0
5
4
3
2
0.95%
1.08%
1.15%
1.31%
1.45%
1
0
0
2
4
6
8
10
10
100
Shear rate, s–1
1000
Concentration, %
Rmin (LV)
R (MV)
Figure 6 Weissenberg number for the standard PAA solutions
R (LV)
min (MV)
R
min (HV)
R
R (HV)
Figure 8 Reflectance on the face (Rmin) and reverse (R) sides of
the fabric for the three CMC solutions
2.0
0%
0.025%
0.05%
0.075%
0.1%
1.5
Indeed, the penetration coefficient P decreases when the
polymer concentration increases (e.g. for concentrations of
2.3, 2.6 and 3.0% polymer MV, P was 84, 78 and 23%,
respectively) which is a consequence of the increased
viscosity of the fluids. Since it is reasonable to consider
that the viscosity can be described by a power law for the
shear rates occurring in the process, P is plotted vs. the
parameter K0 previously defined in Eqn 7 (Figure 9). The
experimental data are regrouped on the same master curve.
It seems that K0 is a pertinent parameter to evaluate the
penetration coefficient whatever the type of CMC polymer
that is used. A similar value of K0 may be associated with
0.2%
1.0
0.5
0.0
10
100
Shear rate, s–1
1000
Figure 7 Weissenberg number for the standard HV–PAA
solutions
72 © Color. Technol., 118 (2002)
Web ref: 20020205

50
40
30
20
10
0
100
80
R
min (PAA)
R (PAA)
60
40
20
0
MV
HV
LV
PAA
1
10
K0
100
0
0.5
1.0
1.5
2.0
Concentration, %
Figure 10 Reflectance on the face (Rmin) and reverse (R) sides of
Figure 9 Penetration coefficient vs. the parameter K0 for the
the fabric for the PAA solutions
three CMC solutions and the PAA solutions
100
either a CMC LV solution that is weakly shear-thinning
with a low zero shear viscosity or a CMC HV solution that
is highly shear-thinning with a high zero shear viscosity.
For the PAA solutions, the power law index n is similar
for all the concentrations and K0 allows a direct comparison
of the viscosities at any shear rate. The results in Figure 9
show a decrease in P when K0 increases. There is a slight
difference between the CMC and the PAA solutions, which
is due to the latter experiencing easier penetration into the
fabric. At equivalent values of K0 there is a corresponding
higher penetration for PAA than for CMC solutions. This
can be explained by considering the shear viscosities.
Comparing the rheological behaviours of these fluids, one
observes that, at low shear rates, the viscosities of the PAA
solutions are much larger than those of the CMC solutions.
However, the PAA solutions are highly shear-thinning (0.22
< n < 0.25) so that at high shear rates their viscosities
rapidly decrease and become less than those of the CMC
solutions (n > 0.38).
PAA
HV–PAA
80
60
40
20
0
0
1
2
3
4
5
Weissenberg number
Figure 11 Penetration coefficient vs. the Weissenberg number
for the PAA and HV–PAA solutions
The reflectivity on the face and reverse sides of PAA
solutions are given on Figure 10 for the different concen-
trations. The low values of Rmin, which again appear to be
independent of the polymer concentration, mean, as
previously, that the colour of the face side of the fabric is
saturated. On the other hand, as would be expected, R
increases and the penetration coefficient decreases with
increasing polymer concentration.
The PAA solutions are highly elastic. The plot of the
penetration coefficient vs. the Weissenberg number (Figure
11) shows a decrease in P when We increases. For these
solutions, larger values of We correspond to higher polymer
concentrations and thus to higher viscosities. Conse-
quently, it is not possible to separate the coupled effects
of viscosity and elasticity. The results for the mixed CMC–
PAA solutions (Figure 11) show a similar trend. Since these
six solutions are characterised by the same shear viscosity,
it is possible to conclude that the elasticity reduces the
level of penetration. Similar results have been obtained by
Vu et al. on compounds for textile printing which were
modified by the addition of surfactants [10]. This effect is
less marked than for PAA solutions because, in this case,
it is amplified by the conjugated influence of the viscosity.
The competition between shear-thinning and elastic
behaviours can be compared to results previously obtained
on the study of the flow upstream and under the blade in
the case of coating a plastic film [13]. With flow
visualisations, it has been demonstrated that, for an elastic
fluid, the flow generates large normal stresses. It results
in a swelling of the streamlines (and, in some cases,
instabilities) at the entrance of the device. The normal
stresses act against the flow under the blade and this is
more marked as the angle θ is increased. Indeed, the flow
is not only a shear flow but it has an elongational com-
ponent due to the contraction of the geometry. This effect
can be partly compensated by shear-thinning properties of
the fluid. Since the flow of fluid between the yarns of the
fabrics is comparable to the flow through an orifice, a
similar effect of normal stresses tends to reduce the
penetration of the fluid.
Conclusions
The results of the penetration of coating compounds into
a fabric during the coating process are reported herein. The
coating compounds used were model polymer solutions
Web ref: 20020205
© Color. Technol., 118 (2002) 73

with controlled rheological parameters and the results
show that the penetration depends on both the viscosity
and the elasticity of the fluid. While the elasticity tends
to reduce the penetration, it can be partly compensated by
a shear-thinning behaviour. These results are consistent
with the conclusions of flow visualisation experiments
performed on the same experimental device [13]. The
problem is complicated by the fact that the flow under the
blade – especially when θ is not small (> 10°) – and
between the yarns is not a simple shear flow at a given
shear rate. It has an elongational component and the effect
of large elongational viscosities characterising highly
elastic liquids is to resist the flow, preventing penetration
of the fluid. For rapid processes, when the residence time
of the fluid under the blade is of the same order of
magnitude as the relaxation time of the fluid, transient
effects have to be evaluated. On the other hand, results on
a physical quantity are observed at a macroscopic level
which leads to global information on penetration. In this
work, the influence of the rheology of fluids was
investigated and fluids with approximately the same
surface tension [14] (always spread on the same solid
sample characterised by the same surface energy) were
compared. Obviously, if one looks at a mesoscopic or a
microscopic scale, capillary phenomena are important, due
to the complex structure of the fabric. It is therefore
important to investigate the influence of the interfacial
properties on the penetration of a complex fluid in a non-
homogeneous porous medium.
*
*
*
We are grateful to Professor Pierre Viallier for helpful
discussions.
References
1. E Bohmer and J Lute, Svensk Paperstidning, 18 (1966) 610.
2. J A Bristow, Svensk Paperstidning, 19 (1967) 623.
3. M B Lyne and J S Aspler, Tappi J., 65 (1982) 98.
4. D E Eklund and P J Salminen, Tappi J., 69 (1986) 116.
5. P E Sandas and P J Salminen, Tappi Proc., Coating Conference
(1987) 97.
6. P J Salminen, Tappi J., 70 (1988) 195.
7. B Miller and I Tyomkin, Text. Res. J., 54 (1984) 706.
8. J T E Lemmen and J G Wassink, J.S.D.C., 106 (1990) 142.
9. N O Clark, W Windle and K M Beazley, Tappi J., 52 (1969)
2191.
10. C Vu, P Lombardin, J L Grossiord, I Fietier and A Blanc,
Cahiers de Rhéologie, 14 (1995) 73.
11. T M Sullivan and S Middleman, J. Non-Newt. Fluid Mech.,
21 (1986) 13.
12. T M Sullivan, S Middleman and R Keunings, AIChE J., 33
(1987) 2047.
13. F Davard and D Dupuis, J. Non-Newt. Fluid Mech., 93 (2000)
17.
14. T
M Sullivan, An experimental and computational
investigation of rheological effects in blade coating, PhD
Thesis, University of California (1986).
74 © Color. Technol., 118 (2002)
Web ref: 20020205