Colloids and Surfaces A: Physicochem. Eng.

Aspects 280 (2006) 203215

Liquidpaper interactions during liquid drop impact
and recoil on paper surfaces
Dushmantha Kannangara, Hailong Zhang, Wei Shen

Australian Pulp and Paper Institute, Department of Chemical Engineering, Monash University, Clayton Campus, Vic. 3800, Australia
Received 22 August 2005; received in revised form 27 January 2006; accepted 6 February 2006
Available online 20 March 2006
Abstract
The spreading and recoiling of water drops on several at and macroscopically smooth model surfaces and on sized paper surfaces were studied
over a range of drop impaction velocities using a high-speed CCD camera. The water drop spreading and recoiling results on several model
hydrophobic and hydrophilic surfaces were found to be in agreement with observations reported in the literature. The maximum drop spreading
diameter for those model surfaces at impact was found to be dependent upon the initial drop kinetic energy and the degree of hydrophobic-
ity/hydrophilicity of the surface. The extent of the maximum drop recoiling was found to be much weaker for hydrophilic substrates than for
hydrophobic substrates. Sized papers, however, showed an interesting switch of behaviour in the process of water drop impaction. They behave
like a hydrophobic substrate when a water drop impacts on it, but like a hydrophilic substrate when water drop recoils. Although the contact angle
between water and hydrophilic or hydrophobic non-porous surfaces changes from advancing to receding as reported in literature, the change of
contact angle during water impact on paper surface is unique in that the level of sizing was found to have a smaller than expected inuence on the
degree of recoil. Atomic force microscopy (AFM) was used to probe bres on a sized lter paper surface under water. The AFM data showed that
water interacted strongly with the bre even though the paper was heavily sized. Implications of this phenomenon were discussed in the context
of inkjet print quality and of the surface conditions of sized papers. Results of this study are very useful in the understanding of inkjet ink droplet
impaction on paper surfaces which sets the initial condition for ink penetration into paper after impaction.
2006 Elsevier B.V. All rights reserved.
Keywords: Liquid drop impact; Sized paper surface; Liquidpaper interactions; Inkjet printing; Spreading and recoiling
1. Introduction
Phenomena associated with a liquid drop impaction on a
solid surface and the spreading and recoiling thereafter can be
seen in many industrial applications. Spreading of herbicides on
leaves, spray painting and inkjet printing are just a few exam-
ples to mention. Liquid drop impaction on macroscopically
smooth hydrophilic and hydrophobic solid surfaces has been
studied by many researchers [15]. Some mathematical mod-
eling approaches were explored in these studies to predict the
maximum and minimum diameters of the liquidsolid contact
area at the drop impaction and recoil. The basic consideration
common to all these modeling studies was the conservation of
energy. The impaction of water drops on smooth hydrophilic
and hydrophobic surfaces was summarized by Park et al. [1].

Corresponding author. Tel.: +61 3 99053447; fax: +61 3 99053413.

E-mail address: wei.shen@eng.monash.edu.au (W. Shen).
In mathematical modeling, water drop impact on hydrophilic
surfaces is generally considered to have three stages: the drop
immediatelybefore impaction; the maximumdiameter byforced
spreading and the nal equilibrium state after the dissipation of
the kinetic energy. On the other hand, the water drop impact on
hydrophobic surfaces has four stages: drop immediately before
impaction; the maximumdiameter byforcedspreading; the max-
imum drop recoil (with the liquidsolid contact area reduced to
the smallest value) and the nal equilibrium state.
Unlike smooth and homogeneous substrates such as Teon
and glass, there are much fewer literature reports on liquid
drop impaction on chemically and topographically heteroge-
neous substrates such as paper. Asai et al. [6] and Heilmann [7]
reported their studies of impaction and recoil of inkjet ink drops
on the surface of commercial papers. Asai et al. also modeled
the drop impaction and recoil processes using the energy conser-
vation principle, but they treated the paper surface as a smooth
and homogeneous surface [6]. These authors reported that the
maximum forced spreading of inkjet ink drops on paper was
0927-7757/$ see front matter 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.colsurfa.2006.02.008
204 D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215
independent of the characteristics of paper surface, but depen-
dent only on the ink drop impact velocity [6]. They correlated the
ink drop spreading data with the Reynolds (Re) and the Weber
(We) numbers of the ink drop, without considering paper surface
parameters.
Ink drop impact on a paper surface and the subsequent redis-
tribution of the ink on the paper surface are important factors
affecting inkjet print quality. There is a large amount of litera-
ture addressing aspects of liquid spreading on nonporous solids
[811] and liquidpaper interactions after the liquid and paper
contact has been established [1218]. To address the impact of
a liquid drop on a paper surface in the context of inkjet printing,
it is helpful to rst review the research in the liquid spreading
on and penetration into the paper surface after impact, since ink
spreading and penetration take place immediately after ink drop
impaction and have a strong inuence on print quality.
Several studies revealed that the kinetics of liquid wetting of
smooth solid surfaces could be well characterized by a simple
power law [9,19]. Other studies reported that wetting kinetics
on rough surfaces could also be described by simple power law
models [8,12]. However, Apel-Paz and Marmur [8] showed that
using a simple power law model to describe liquid spreading on
a rough surface was empirical.
The Washburn equation has been used traditionally as a
model to describe the penetration of liquid into a paper [15].
More recent studies showed that the Washburn model gives the
rst order description of the liquid penetration into paper [11].
Roberts et al. [14] conducted a microscopic study and showed
that the penetration of water into an unsized paper was via lm
ow [14] but this lm ow could not be observed from strongly
sized papers [16]. Modaressi and Garnier [12] observed that the
interaction of a water drop and a paper could be described by
two sequential phenomena, i.e., the wetting of the paper surface
by water until a pseudo-equilibrium contact angle was reached,
followed by the water drop being absorbed into the bulk of paper.
They found that the rst process does not occur instantaneously,
if the paper is sized [12]. Lyne and Aspler [13] also published
some empirical data suggesting that a wetting delay might be
appropriate to describe the delay between the onset of liquid
penetration and the moment of liquidpaper contact was made.
The primary interest of this investigation is to study the
behaviour of the liquidpaper interaction in the event of ink drop
impaction on a sized paper surfaces. Empirical observations of
inkjet printing trials showed that inkjet printing quality deterio-
rated badly if ink drops amalgamated after landing on the paper
surface [20]. The ink drop amalgamation could be further pro-
moted by the forced spreading during impaction and the recoil
after the impact. Park et al. [1,21] showed, using data obtained
from impervious and macroscopically smooth solid surfaces,
that the liquid drop contact area with the solid substrates reduced
to a minimum value when the drop reached its maximum recoil
after impaction (expressed by the minimum diameter of the cir-
cular contact area, D
min
). These authors also showed that D
min
could be correlated with the work of adhesion between the liquid
and the solid [1]. The larger the value of work of adhesion, the
weaker the liquid drop recoil, and thus, the larger the D
min
. On
the other hand, Asai et al. [6] suggested that the ink drop recoil
fromthe paper surface could be correlated with Re (the ink prop-
erty) and We (the inertial energy of the ink drop), and showed no
correlation to paper surface properties. Although these authors
did not further comment on their results, their results suggested
that the paper surface may not be regarded as a simple surface.
Characterization of the work of adhesion between liquids
and paper using contact angle methods is very difcult [22,23].
As a result, the large apparent contact angles for water on
sized cellulose bre and paper surfaces were interpreted as that
these surfaces are hydrophobized [12,24]. In fact, the interaction
between an aqueous solution and cellulose bres (even strongly
sized) is rather extraordinary [11]. It may not be appropriate
to describe the sized cellulose bres simply as hydrophobic,
instead, it may be more appropriate to describe the sized bre
surface as a hydrophilic surface covered by domains of sizing
chemicals. Although these domains slow down, the wetting of
the bres by an aqueous solution by pinning the three-phase
contact line, the partial coverage of the bre surface by these
domains could not totally cut off the interactions between liquid
molecules and the hydrophilic areas of the bre surfaces. This
fact may strongly inuence the forced liquid drop spreading and
recoiling.
In this study, behaviour of water drop impaction and recoil
on hydrophilic and hydrophobic model surfaces, as well as on
sized papers is investigated. Atomic force microscope (AFM)
was also used to study the interaction of water and heavily sized
paper surfaces. The results obtained are discussed in the context
of the chemical conditions of the sized paper surface and inkjet
printing quality.
2. Experimental
2.1. Materials and reagents
Water (Millipore, 18 M) and diiodomethane (99%, Aldrich)
were used as liquid probes to determine the surface energy of
the solid surfaces. Surface tension data of these probes obtained
from the literature [25] are listed in Table 1.
Cellulose acetate (99%, Aldrich) and sodium methoxide
(25%in methanol, Aldrich) were used for preparing smooth cel-
lulose lms on glass slides. Analytical grade methanol (BDH)
and acetone (BDH) were used as solvents for preparing the cel-
lulose lms.
Alkyl ketene dimer (AKD) wax was supplied by Hercules
Chemicals Australia Pty Ltd. The wax was re-crystallized in
AR grade acetone three times before use. AR grade n-heptane
(99.5%, Ajax) was used to make up AKDheptane solutions,
which were used for solvent sizing of cellulose lms.
Table 1
Surface tension (mN/m) values of liquid probes [25]
Surface tension
(
lv
)
Polar component
(
p
lv
)
Dispersive
component
(
d
lv
)
Water 72.8 51.0 21.8
Diidomethane 50.8 0.0 50.8
D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215 205
Commercial AKDemulsion (Keydimer E) was obtained from
Eka Chemicals, Australia. An anionic trash catcher (ATC) and
a retention acid (NLP 780) were also obtained from Eka Chem-
icals. Fully bleached Eucalypt Kraft pulp was obtained from the
Maryvale Mill of Australian Paper Pty Ltd. These materials were
used to make handsheets.
Teon sheet (thickness =0.5 mm) was obtained from Good-
fellow (UK). Microscopic glass slides were supplied by Bio
Lab, Australia. A commercial copy paper (Reex, produced by
Australian Paper Ltd.) was used as a paper sample. Reex paper
was made under alkaline conditionusingpredominantlythe fully
bleached eucalypt Kraft pulp. The glass slides were soaked in
a 2% solution of laboratory detergent (RBS 35) overnight and
were then rinsed with ample amount of Millipore water, fol-
lowed by rinsing with methanol and acetone. Teon sheet was
cut into convenient sizes and washed with Millipore water and
then rinsed with methanol and acetone. The Teon and glass
samples were dried in an oven at 60

C, allowing solvents to
evaporate. The dried substrates were used for tests and sample
preparation described below.
2.2. Cellulose lm generation on glass slide
A cellulose acetate lm was rst prepared by slowly dip-
ping a clean microscope slide into a solution of 10% cel-
lulose acetate in acetone (w/v). Upon the withdrawal of the
slide from the solution, acetone evaporated, leaving a thin and
smooth lm of cellulose acetate on the glass slide. Samples
were raised from the cellulose acetate solution at a xed veloc-
ity using an in-house built device. Cellulose acetate lm was
then hydrolyzed to form cellulose in 0.5% sodium methox-
ide (w/v) in methanol for at least 12 h. The slides were then
washed with Millipore water followed by methanol and dried
at 60

C. This method was similar to the one employed by

Toussaint and Luner [26]. An FTIR spectrum of the cellu-
lose lm was taken and it agrees well with the spectrum taken
fromthe micro-granular cellulose sample obtained fromAldrich
(Fig. 1).
2.3. Handsheet forming and AKD sizing of cellulose lms,
lter paper and handsheets
The pulp was disintegrated following the TAPPI test method
(T200 sp-96) and rened using a Valley beater following the
TAPPI test method (T262 sp-96) to the Canadian Standard
Freeness of 390. Sized handsheets (220 mm220 mm) were
made using a Moving Belt Sheet Former [27] for liquid drop
impact study. AKD dosing levels for handsheet sizing were 0,
2.5, 7.5 and 10 kg (AKD)/t pulp (oven-dried at 105

C). All
handsheets were also lled with around 20% ground calcium
carbonate.
Cellulose lms were sized to two different sizing levels using
AKDheptane solutions of two concentrations (0.2 and 0.6 g/L).
Sizing was performed by slowly dipping cellulose lms into
an AKDheptane solution and slowly pulling them out of the
solution using the device for cellulose lm generation. These
slides were then heat treated in an oven at 105

C for 20 min.
Fig. 1. FTIR spectra of (a) cellulose lm generated in this study, and (b) micro-
granular cellulose obtained from Aldrich.
The highly sized lter paper sample was prepared under the
same conditions, but with 0.8 g/L.
2.4. Plasma treatment of commercial copy paper
Reex copy paper (static contact angle with water =119.8

)
was plasma treated using an atmospheric plasma treater (Sigma
Technologies International, Tucson, Arizona, USA) in order to
obtain a wide range of contact angles (100.3, 89.4 and 84.7

,
Table 5). The atmospheric plasma unit was powered by a 450 W
high voltage DC power supply. Ultrahigh purity Ar gas was
fed through the electrode tube at 3.3 L/min for generating the
discharge. The paper samples were xed to the drum and the
distance fromthe drumto the electrode was set to approximately
2 mm. The drumspeedwas 25 m/minandthe paper samples were
passed through the plasma zone three, ve and seven times to
achieve required level of hydrophilicity.
2.5. Drop generation and image capturing for the drop
impact analysis
An OCAH-230 apparatus (Dataphysics, Germany) and
Photron Fastcam Super 10KC were used to conduct the liq-
uid drop impact study. The OCAH-230 apparatus consists of a
computer controlled liquid drop dispensing syringe pump and
a high-speed camera capable of 360 frames per second (fps).
For capturing the drop impact process, another faster camera
(Fastcam Super 10KC, USA) was used. The camera was set at
500 fps, with a shutter speed of 5 s. Subsequent image anal-
ysis was performed using Image-pro Plus and Photron Motion
Tools software. The water drop base diameter change during the
impaction was calculated using a macro programs written for
these software.
A syringe with a at-tipped stainless steel needle (outer
diameter =0.31 mm) was tted to the syringe pump which was
programmed to a dispensing speed of 0.11 L/s. Equivalent
206 D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215
spherical drop diameter (d) was determined by weighing 10
drops of water and using Eq. (1).
d =
3

6m
n
(1)
where m is the total mass of n drops and is the density of
water. It was observed that the drop shape at impact (state 1)
is not perfectly spherical, hence better accuracy was obtained
with this method rather than measuring the drop diameter on the
image taken at impact state 1.
The drop impact velocity was controlled by varying the
height of the needle tip above the substrate surface. It was real-
ized that, since the distance between the tip and substrate was
used for the drop velocity calculation, a systematic error was
unavoidable. This is because it is difcult to precisely deter-
mine the centre of mass of the drop immediately before its
release using the present experimental set up. However, this error
was expected to be constant, as the drop volume had been well
controlled.
2.6. AFM measurements
A ThermoMicroscope AFM was used for measuring bre
swelling in water. Aliquid scanner (TopoMetrix 5180-00) was
used to obtain the images of the cotton lint bre on the surface of
a highly sized lter paper before and after water was introduced
on the lter paper surface. Contact scanning mode was used in
the data acquisition.
3. Surface energy calculations and the liquid drop
impact modeling considerations
3.1. Surface energy calculation from contact angle data
Although the liquid drop impact process is signicantly
affected by liquidsolid interactions, the inuence of the work
of adhesion between the liquid drop and solid substrate in the
impact and recoil processes has not been considered in most
previous studies [2,28]. Some authors used only the liquid prop-
erties and the drop kinetic energies to predict the impact results
[2,6]. However, we realized that the solid surface properties and
the liquidsolid interactions play important roles in liquid drop
impact and recoil on solid surface. This is particularly true with
inkjet printing on sized paper surface.
Contact angle methods are frequently used for measuring
surface energetics of a solid. A common approach adopted in
various mathematical models for calculating the components of
surface energy is to rst calculate the work of adhesion (W
a
)
between the solid surface and liquid probes using Eq. (2), which
assures that the spreading pressure due to vapour adsorption on
the solid surface is negligible [23].
W
a
=
lv
(1 +cos ) (2)
where
lv
is the surface tension of the liquid in the presence of
its vapour and is the equilibrium contact angle between the
liquid and the solid.
Fowkes proposed that the surface free energy (or surface
tension) can be expressed as the sumof two components of inter-
molecular interactions, i.e. the dispersion (
d
) and the polar (
p
)
components:
=
d
+
p
(3)
By incorporating the rule of the geometric mean, which
was initially proposed by Fowkes and Mostafa [29] and later
expanded by Owens and Wendt [30], into Eq. (2), we have:
W
a
=
lv
(1 +cos ) = 2

d
sv

d
lv
+

p
sv

p
lv

(4)
where sv and lv represent the solidvapour and liquidvapour
interfaces, respectively. The dispersion and polar components of
surface tension (
d
sv
and
p
sv
) for many liquids are known. Since
there are two unknowns (
d
sv
and
p
sv
) in Eq. (4), one can calculate

d
sv
and
p
sv
by adopting the two-liquid approach [23].
3.2. Further theoretical and practical considerations on
work of adhesion
It is difcult to measure the work of adhesion between an
aqueous liquid and a paper surface, even if the paper was sized
and a stable apparent contact angle could be determined. Some
researchers reported the surface energy data for sized papers
using the contact angle methods [11,31,32], their results sug-
gested that the heavily sized paper surfaces were similar to a
hydrophobic polymer surface such as Telfon or polyethylene in
that the polar component of the heavily sized paper was zero
[11,31,32]. Some authors [31] used their results to justify the
possible total coverage of the cellulose surface by the sizing
chemical (AKD).
Shen et al. [22] showed that even though contact angle calcu-
lationpredictedthat there was nopolar interactionbetweenheav-
ily sized paper samples and water, strong Lewis acid and base
characteristics of paper surfaces could still be detected using
other methods (e.g. inverse gas chromatography). Although siz-
ing was found to have reduced the Lewis acid and base com-
ponents of the paper or cellulose bre, these components were
reduced only slightly and were still able to interact with a con-
tacting liquid [22].
It has been either suggested or implied [17,22] that water,
or any other polar liquid, is able to have strong polar interac-
tions with sized paper, even though water does not appear to
spread on strongly sized paper surfaces. Fowkes and Mostafa
[29] proposed that the work of adhesion contribution by the
Lewis acidbase interactions could be written as:
W
AB
a
= fN(H
AB
) (5)
where H
AB
is the molar enthalpy of acidbase adduct forma-
tion, N is the number of moles of interacting functional groups
per unit area on the solid surface, and f a conversion factor to
correct enthalpy values to free energy values [23]. Eq. (5) ts
the situation of liquidpaper interaction well, as some literature
results support the view that the surface of a heavily sized paper
was not fully covered by a monolayer of sizing chemical and
D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215 207
Fig. 2. The schematic of the four stages of an aqueous liquid drop impact on a
hydrophobic solid surface.
those uncovered hydroxyl groups of cellulose should be able to
interact with water [17,22,33,34].
3.3. Modeling of maximum spreading
Fig. 2 shows schematically the four stages of an aqueous
liquid drop impact on a hydrophobic solid surface, as was sum-
marized by Park et al. [1]. Most mathematical models generally
consider the conservation of energy of states 1 and 2 (kinetic,
interfacial and viscous dissipation during drop impact [1,2,6]).
The Re and We numbers were used as parameters to empirically
correlate the maximumforced drop spreading diameter obtained
experimentally with the theoretical models. The Re and We num-
bers are dimensionless parameters that provide comparisons of
the kinetic and the viscosity dissipation energies, and the kinetic
and the surface energy of the drop, respectively. They are dened
as follows:
Re =
ud

(6)
We =
du
2

lv
. (7)
where is the liquid density, u is the impact velocity, d is
the liquid drop diameter and is the viscosity of the liquid.
When kinetic energy is large enough to overpower the vis-
cous force (i.e. large Re), the drop will be forced to spread
extensively upon impaction. When We is small, the surface ten-
sion dominates the behaviour of the drop and the drop tends
to keep its spherical shape provided that no wetting takes
place.
Early models for predicting the maximum drop spreading
ratio, D
m
*
(the ratio of drop diameter at state 2 (D
m
) and the
spherical drop diameter (d) at state 1, Fig. 2), are based on
the energy conservation [1,4,6]. It was assumed that the kinetic
energy at maximum spreading is zero and the drop volume is
constant. The energy conservation equation assumed for these
models [1] was as follows:
KE
1
+SE
1
lv
+SE
1
sv
= KE
2
+SE
2
lv
+SE
2
sl
+VE
2
(8)
where KE and SE denote kinetic and surface free energies,
respectively. Superscripts 1 and 2 represents the impact stages 1
and 2, respectively, as shown in Fig. 2, and VE is the viscously
dissipated energy as the drop deforms from state 1 to state 2.
For low viscosity liquids (Re >81(D
m
/d)
4
), VE
2
was estimated
using the relationship proposed by Pasandideh-Fard et al. [35],
Mao et al. [2] and Park et al. [1] and given by Eq. (9).
VE
2
= 0.33
u
2
d

Re
D
2
m
(9)
By assuming the drop shape at maximum spreading (state 2)
being cylindrical [2] and spherical (for near-zero impact veloc-
ity drops) [1], Mao et al. [2] and Park et al. [1] proposed their
equations for the prediction of maximum drop spreading at the
drop impaction. Of course, the assumption about drop shape
was not the only difference between the models proposed by
these authors. Park et al. [1] introduced more mathematical
considerations toaccommodate the decrease insurface andinter-
face energies for spontaneous spreading from state 1 to state 2
(SE
p
). This energy dissipation was estimated by considering
spontaneous dissipation of surface and interfacial energies into
heat at zero impact velocity and assuming it to be the same at
other velocities. Therefore, the energy balance equation intro-
duced by Park et al. [1] is as follows.
KE
1
+SE
1
lv
+SE
1
sv
= KE
2
+SE
2
lv
+SE
2
sl
+VE
2
+SE
p
(10)
Park et al. [1] demonstrated that their model predict maxi-
mum spreading of water drop impaction on both hydrophilic
and hydrophobic surfaces with much better accuracy than Mao
et al. [2] and Fukai et al. [5]. Furthermore, the model by
Park et al. [1] showed good agreement (within 4%) with the
experimental results published by Asai et al. [6] for their
inkjet printing study at low Re and We. The model by Park et
al. [1] was therefore selected to predict maximum spreading
for this work. With given contact angle and calculated val-
ues of Re and We, D
m
can be calculated based on the model
equations.
To predict the drop recoil from the solid surface (i.e. state
3), Park et al. [1] further explored the correlation between the
minimum diameter of the solid/water drop contact area, D
min
,
at the maximum drop recoil and the apparent work of adhesion
(W
a
) calculated fromthe equilibriumcontact angle values. Their
correlation suggested that D
min
is related to the work of adhesion
between the liquid and the solid. Since water was used, D
min
was much smaller for hydrophobic surfaces than for hydrophilic
surfaces.
Agood knowledge on the drop recoil is essential to the under-
standing of inkjet ink drop amalgamation after being printed on
the substrate surface. Ink drop amalgamation is a factor that
seriously compromises the print quality [20]. It is realized that
when a paper surface is used for drop impact study, both surface
energy and surface roughness will affect the forced spreading
of drops on the paper surface. In this study, however, we focus
only on the surface energy aspects. The inuence of paper sur-
face roughness to the drop impact and recoil will be reported
separately.
208 D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215
4. Results and discussion
4.1. The maximum forced spreading of water drops
impacting on various macroscopically smooth and a
commercial paper surfaces
Apparent equilibriumcontact angle values for diiodomethane
and water on various solid surfaces are listed in Table 2. As is
expected, on the known high surface energy substrates such as
glass and cellulose, the contact angle values are substantially
lower than those on the low energy surfaces such as Teon. The
contact angle for diiodomethane on Teon surface was higher
than on other surfaces. Sizing of the cellulose lms with AKD
caused substantial increases in contact angles for both liquid
probes. The lightly sized cellulose lm was still partially wet-
table by water (equilibrium contact angle <90

). In contract
to this, the heavily sized cellulose lm (0.6 g/L AKD) shows
a non-wetting character. Water contact angle on this surface
(101.2

) was close to the value reported for water on AKD

wax, which was 109

[11]. This suggests that at this sizing

level the AKD coverage of the surface was close to the full
coverage. The apparent contact angles for water on handsheets
sized to three different levels were 115.3, 125.2 and 132.8

,
respectively. Contact angles on paper surfaces were very sta-
ble and there was no obvious water absorption by the sheet
over the time of measurements, which is typically in the range
of 10 s.
Fig. 3 illustrates the shapes of water drops at the maximum
forced spreading upon impact on various solid surfaces. The
most obvious feature inFig. 3is that the impactingcontact angles
for water on hydrophilic surfaces are very small, whereas for the
less hydrophilic surface (such as the lightly sized cellulose lm
surface) and hydrophobic surfaces the impacting contact angles
are greater than 90

. FromFig. 3, it can be seen that, it is reason-

able to assume that the shape of the water drop at the maximum
forced spreading on the lightly sized cellulose surface and other
hydrophobic surfaces is cylindrical, as suggested by Mao et al.
[2] and Asai et al. [6]. However, for hydrophilic surfaces such
as glass and unsized cellulose lm, the cylindrical assumption
appears to be less accurate. Fig. 3(a) and (b) showthat the centre
region of the spread drops on hydrophilic surfaces is at whereas
the edges are tapered; this is expected to cause some error in the
modeling, no matter which model is used.
Fig. 3. The shapes of water drops on various substrates at the maximum forced
spreading upon impact (impact velocity =1 m/s).
Table 3 lists the maximum forced spreading ratios for water
drops on various substrates obtained experimentally and using
Parks model [1]. For all macroscopically smooth and impervi-
ous samples, this model exhibited a discrepancy ranging from
0 to 10%. The following trends are also worth mentioning.
First, the model by Park et al. [1] overestimated the D
m
*
for
both hydrophilic and hydrophobic smooth surfaces at all impact
velocities. Second, this model better predicts drop impaction
on smooth and impervious surfaces (i.e. glass and cellulose
lm) than on paper surfaces, to which larger discrepancies were
observed and in some cases Parks model underestimated the
D
m
*
for sized handsheet samples. The major source of the dis-
crepancy between the experimental results and the prediction
obtained using the model is likely to be associated with the
assumption that the solid surface is physically and chemically
ideal. Since the model did not consider heterogeneity of the
Table 2
Solid surface free energies (mJ/m
2
) and work of adhesion with water (mJ/m
2
)
Surface Static contact angle (

) Surface energy (
sv
) Polar component (
p
s
) Work of adhesion (W
a
)
Water DIM
a
Glass 14.2 40.8 74.5 35.3 143.4
Cellulose 25.0 47.5 69.4 33.8 138.8
PE 99.5 47.8 35.6 0.1 60.8
Teon 118.6 83.2 15.9 0.0 38.0
0.2 g/L AKD-cellulose 57.9 58.7 47.5 18.2 111.5
0.6 g/L AKD-cellulose 101.2 69.9 23.9 1.0 48.9
5 g/kg AKD-handsheet 125.2 0.0 30.8
10 g/kg AKD-handsheet 132.8 67.7 26.7 2.5 23.3
a
Diidomethane.
D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215 209
Table 3
Experimental and calculated (using model by Park et al. [1]) values of the maximum spreading ratios of water drops impacting on various substrates
Substrate Dimensionless numbers Impact velocity (m/s) Experimental D
m
*
=D
m
/d D
m
*
by Park et al.s model
Glass We =5 0.44 2.7 2.8
Cellulose Re =887 2.4 2.6
PE 1.7 1.8
Teon 1.6 1.7
0.2 g/L AKD-cellulose 1.7 2.3
2.5 g/kg AKD-handsheet 1.4 1.4
5 g/kg AKD-handsheet 1.3 1.3
10 g/kg AKD-handsheet 1.3 1.2
Plasma 7X-copy paper 1.4 1.9
Plasma 5X-copy paper 1.5 1.9
Plasma 3X-copy paper 1.5 1.8
Glass We =16 0.77 3.0 3.1
Cellulose Re =1552 3.1 3.1
PE 2.2 2.2
Teon 1.9 2.0
0.2 g/L AKD-cellulose 2.3 2.7
2.5 g/kg AKD-handsheet 1.9 1.8
5 g/kg AKD-handsheet 2.0 1.7
10 g/kg AKD-handsheet 1.9 1.6
Plasma 7X-copy paper 2.2 2.4
Plasma 5X-copy paper 2.0 2.3
Plasma 3X-copy paper 1.8 2.2
Glass We =28 1.00 3.2 3.4
Cellulose Re =2016 3.1 3.3
PE 2.3 2.5
Teon 2.2 2.4
0.2 g/L AKD-cellulose 2.5 3.0
2.5 g/kg AKD-handsheet 2.2 2.4
5 g/kg AKD-handsheet 2.2 2.3
10 g/kg AKD-handsheet 2.0 2.3
Plasma 7X-copy paper 2.4 2.7
Plasma 5X-copy paper 2.2 2.6
Plasma 3X-copy paper 2.3 2.5
substrate surface, it is difcult to determine the cause of the
discrepancy between model and the results. However, the fact
that this model more poorly predicts water drop impaction on
paper surfaces, which are known to be physically and chemically
heterogeneous, supports our reasoning.
4.2. The maximum water drop recoil from hydrophobic
surfaces after forced spreading
Two macroscopically smooth hydrophobic surfaces, Teon
and polyethylene (PE), were used to demonstrate the recoil
of water drops from hydrophobic surfaces. Water drop impact
experiments were conducted at drop velocities of 0.44, 0.77 and
1.00 m/s. Figs. 4 and 5 show the state of the drops (impacted
at the speed of 0.77 m/s) at (a) maximum forced spreading, (b)
maximum recoil and (c) after reached equilibrium.
The water drop recoil from Teon and PE surfaces shows
similar behaviour. At the moment of maximumforcedspreading,
both surfaces were not wetted by water and had apparent contact
angles greater than 90

. Due to poor illumination, the image in

Fig. 4(a) is less clear. However, it still shows that the contact
angle is greater than 90

. Images in Figs. 4(a) and 5(a) are in

good agreement with the images reported by Park et al. [1] of a
water drop impacted on a Teon surface.
At the maximumforced spreading, some energy is lost due to
viscous dissipation [1,2,6]. The rest of the energy is converted
into surface energy and interfacial energy, as described in Eq.
(8). The pressure at the rim of the spreading drop is the highest
and points inward to the centre of the drop. It may be possible
to estimate this pressure if clear images (like that in Fig. 3(e))
can be obtained. On a highly hydrophobic surface, the Laplace
pressure may then be estimated as:
P 2

1
D
m
+
1
t

(11)
where P is the inward-pointing pressure exerted by the curved
liquid surface, is the water surface tension, D
m
is the diameter
of the drop at the maximum forced spreading, and 2(1/t) is the
curvature of the rim of the drop shown in the image (Fig. 3(e)).
In the extreme case where the contact angle at the maximum
spreading is 180

, t becomes the thickness of the attened drop

and the pressure (P) can be estimated using Eq. (11). The
subsequent recoiling of the drop is driven by P.
210 D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215
Fig. 4. The water drop impaction on a Teon surface at the velocity of 0.77 m/s:
(a) maximum forced spreading, (b) maximum recoil and (c) after reached equi-
librium.
Since the work of cohesion of water (W
c
=2
water
=
145.6 mJ/m
2
) is much higher than either the work of adhesion for
water/Teon (38.0 mJ/m
2
, Table 2) or for water/PE(60.8 mJ/m
2
,
Table 2), the cohesion of water overcomes the adhesion between
water and these surfaces when water drop recoils. As a result,
the water drop recoils strongly from all hydrophobic surfaces.
It has also been observed that water drop recoil back strongly
from the surfaces of parafn [2] and Teon [1], respectively. It
Fig. 5. The water drop impaction on a PE surface at the velocity of 0.77 m/s: (a)
maximum forced spreading, (b) maximum recoil and (c) after reached equilib-
rium.
Table 4
Static and equilibrium contact angle data, and water drop diameter data
a
at
different stages of drop impaction on smooth model surfaces
SCA ECA D
m
D
min
D
e
Glass
0.44 m/s 14.2 12.2 5.5 5.3 6.3
0.77 m/s 14.2 10.1 6.0 6.0 6.6
1.0 m/s 14.2 9.0 6.4 5.9 6.3
Cellulose
0.44 m/s 25.0 18.1 4.9 4.4 5.4
0.77 m/s 25.0 12.8 6.3 6.2 6.6
1.0 m/s 25.0 11.5 6.3 6.2 6.4
PE
0.44 m/s 99.2 94.1 3.5 3.2 3.0
0.77 m/s 99.2 94.9 4.4 1.7 2.8
1.0 m/s 99.2 92.1 5.1 1.6 2.9
Teon
0.44 m/s 109.2 104.8 3.1 2.2 2.5
0.77 m/s 109.2 103.2 3.8 1.3 2.5
1.0 m/s 109.2 103.7 4.4 1.7 2.5
0.2 g/L AKD-cellulose
0.44 m/s 57.9 74.5 3.4 3.3 3.3
0.77 m/s 57.9 54.9 4.7 4.5 4.2
1.0 m/s 57.9 44.4 5.1 4.5 4.3
0.6 g/L AKD-cellulose
0.44 m/s 101.2 102.8 3.4 2.2 2.7
0.77 m/s 101.2 107.7 Droplet divided into two parts at
the maximum recoil
1.0 m/s 101.2 106.2
a
SCA: static contact angle; ECA: equilibrium contact angle; D
m
: maximum
spreading diameter; D
min
: drop base diameter at the maximum recoiling; D
e
:
equilibrium drop base diameter.
is worth noting that the heavily sized cellulose lms (0.6 g/L
AKD) behave in the same manner as a strongly hydrophobic
surface in the maximum recoil phase (Table 4).
4.3. The maximum water drop recoil from hydrophilic
surfaces after the forced spreading
Two macroscopically smooth hydrophilic surfaces, glass and
cellulose lmon glass slide, were used to demonstrate the recoil
of water drops from hydrophilic surfaces. Water drop impact
experiments were conducted at drop velocities of 0.44, 0.77 and
1.00 m/s. Figs. 6 and 7 show the state of the drops at (a) maxi-
mumforced spreading, (b) maximumrecoil and (c) after reached
equilibrium on glass and cellulose surfaces, respectively.
The water drop recoiled from glass and cellulose surfaces in
a similar manner. At the maximum forced spreading, both sur-
faces were wetted by water and apparent contact angles at this
state were much smaller than 90

. Because the rim curvatures

of the water drop on glass and cellulose surfaces were much
smaller than those on the hydrophobic surfaces, the inward-
pointing pressure was much weaker. As a result, water drops
only showed very weak recoil from the hydrophilic surfaces
after impaction.
Another reason for water drop to have weak recoil fromthese
hydrophilic surfaces was that strong water-substrate adhesion
D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215 211
Fig. 6. The water drop impaction on a glass surface at the velocity of 0.77 m/s:
(a) maximum forced spreading, (b) maximum recoil and (c) after reached equi-
librium.
was established. The work of adhesion between water-glass and
water-cellulose calculated using Eq. (2) (see Table 2) show that
they are of the same magnitude as the work of cohesion of
water. The immediate wetting of these substrates by water upon
impactionsuggests that the adhesionbetweenwater andthe glass
and cellulose surfaces has established in the time scale of the
impaction. The comparison of the works of adhesion between
water and these hydrophilic substrates and the work of cohesion
of water can also explain the observed phenomenon that the
interface areas between these substrates and water established
at maximum forced spreading were not reduced at maximum
recoil.
Fig. 7. The water dropimpactiononthe surface of a cellulose lmcoatedonglass
slide at the velocity of 0.77 m/s: (a) maximum forced spreading, (b) maximum
recoil and (c) after reached equilibrium.
Fig. 8. The water drop impaction on the surface of an untreated copy paper
(Reex) at the velocity of 0.77 m/s: (a) maximum forced spreading, (b) maxi-
mum recoil and (c) after reached equilibrium.
4.4. The maximum water drop recoil from a sized paper
surface after the forced spreading
A sized commercial copy paper was used to demonstrate the
recoil of a water drop from a sized paper surface. Water drop
impact experiments were conducted at drop velocities of 0.44,
0.77 and 1.00 m/s. Fig. 8 shows the state of the drop at (a) maxi-
mumforced spreading, (b) maximumrecoil and (c) after reached
equilibrium.
Compared with macroscopically hydrophobic and
hydrophilic surfaces, sized paper surfaces showed inter-
esting behaviour when a water drop impacts on them. When
a water drop hit the sized paper surface and is forced to
spread, the contact angle between water and the paper at the
maximum forced spreading is greater than 90

(Fig. 8(a)).
The rim of the water drop has a large curvature, which
generates a strong inward-pointing pressure. At this stage, the
sized paper showed typical characteristics of a hydrophobic
surface. The inward-pointing pressure caused the water drop
to strongly recoil after the impaction (Fig. 8(b)). When the
drop recoiled, since the waterpaper contact line was pinned,
the drop behaved differently to a hydrophobic surface (see
Figs. 5(b), 6(b) and 8(b) for comparison). Since the contact line
of water and most paper surfaces did not move backward during
the recoil, the apparent waterpaper interface did not change
after reached to its maximumsize in the forced spreading phase.
In this respect, in the water drop recoil phase, the sized paper
surface behaved somewhat similar to a hydrophilic surface. As
shown in Figs. 6 and 7, a hydrophilic surface only causes very
weak water drop recoil. However, since sized paper behaved
as if it were hydrophobic in the forced spreading phase, water
drop recoiled from the paper surface was much stronger than
from a hydrophilic surface, although soon after the drop recoil
phase started, the sized paper behaved like a hydrophilic
surface.
212 D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215
Table 5
Static and equilibrium contact angle data, and water drop diameter data
a
at
different stages of drop impaction on model paper surfaces
SCA ECA D
m
D
min
D
e
Plasma 7X-copy paper
b
0.44 m/s 84.7 73.4 2.89 2.89 3.00
0.77 m/s 84.7 50.3 4.35 4.33 4.30
1.0 m/s 84.7 35.0 4.86 4.82 4.86
Plasma 5X-copy paper
0.44 m/s 89.4 78.7 3.02 3.00 3.10
0.77 m/s 89.4 51.1 4.09 4.01 3.98
1.0 m/s 89.4 38.0 4.45 4.44 4.44
Plasma 3X-copy paper
0.44 m/s 100.3 82.0 3.09 3.04 3.15
0.77 m/s 100.3 56.0 3.72 3.70 3.72
1.0 m/s 100.3 35.0 4.54 4.30 4.56
2.5 g/kg AKD-handsheet
0.44 m/s 115.3 77.6 2.86 2.86 2.98
0.77 m/s 115.3 47.3 3.91 3.68 3.71
1.0 m/s 115.3 43.6 4.54 4.25 4.58
5 g/kg AKD-handsheet
0.44 m/s 125.2 88.8 2.71 2.56 2.78
0.77 m/s 125.2 55.3 4.00 3.67 3.90
1.0 m/s 125.2 38.4 4.35 3.89 4.23
10 g/kg AKD-handsheet
0.44 m/s 132.8 87.7 2.64 2.40 2.70
0.77 m/s 132.8 63.1 3.76 3.51 3.54
1.0 m/s 132.8 46.9 3.99 3.58 3.83
a
SCA: static contact angle; ECA: equilibrium contact angle; D
m
: maximum
spreading diameter; D
min
: drop base diameter at the maximum recoiling; D
e
:
equilibrium drop base diameter.
b
Passed through plasma seven times.
Such a switch of properties of sized paper during the water
drop impact suggests that paper, even sized, should not be sim-
ply regarded as a hydrophobic surface. Whilst the pinning of
the waterpaper contact line can offer an explanation to this
phenomenon, what can be said in addition is that the adhesion
between water and the sized paper was established after the
impact of the water drop. Such an adhesion is likely to be rather
strong, even though the apparent static contact angle between
water and the sized paper was large (Table 5) and the apparent
work of adhesion calculated using those contact angle data was
very low (Table 2). More discussion on the work of adhesion
will be given in later sections.
Surface roughness of the paper is also likely to have certain
inuence on the water drop recoil, since the roughness of a sur-
face contributes to the pinning of the liquidsolid contact line
[36]. However, surface chemical heterogeneity is known to con-
tribute to the pinning of the contact line as well [37]. Fig. 9 shows
the water drop impaction and recoil process on a lightly sized
cellulose lm surface at the impact velocity of 0.77 m/s. Fig. 9
also shows that the drop recoils in a very similar fashion as it does
on a sized paper surface in that there was little or no loss of the
dropsubstrate interface area at the maximumdrop recoil. How-
ever, in the process of reaching the equilibrium contact angle,
the apparent interface area between the water drop and the lm
reduced slightly. The water drop impaction behaviour at other
Fig. 9. The water drop impaction on the surface of the lightly sized cellulose
lm surface (0.2 g/L AKD) at the velocity of 0.77 m/s: (a) maximum forced
spreading, (b) maximum recoil and (c) after reached equilibrium.
velocities is presented in Tables 4 and 5. This slight reduction
of the contact area between the lightly sized cellulose lm and
the water drop upon recoiling may be attributable to the fact that
cellulose lm is much smoother than the paper surface. Despite
this, the behaviour of water drop recoil from these surfaces is
similar. This suggests that forced wetting and the immediate
establishment of adhesion also contribute to switch of the water
drop behaviour at the impact and recoil process.
It is necessary to mention that some highly sized paper sam-
ples (e.g. 10 g/kg AKD-handsheets, Table 5) showed smaller
D
min
values than D
m
. However, the fact that equilibrium drop
base diameter (D
e
) is similar to D
m
suggests that when the drop
recoiled from the paper surface, the water drop could have left
a thin water lm on the paper surface. Since the area covered
by the thin water lm had a diameter of D
m
, when the water
drop settled after the impaction, it could easily wet the area
that covered the thin water lm, leading to D
e
D
m
(Table 5).
This behaviour could not be observed on PE or Teon sur-
faces, suggesting that a thin water lm could not form on those
surfaces.
4.5. The adhesion of water with sized paper and bre
surface
Park et al. [1] predicted drop recoil (state 3) using the work of
adhesion of the liquid and the solid surface. These authors used a
correlation between D
min
/d (d is the diameter of the liquid drop
before impact, as shown in Fig. 2) and the normalized work
of adhesion (W
a
/
lv
=(1 +cos )) to demonstrate a relationship
between the liquid/substrate adhesion and the liquid drop recoil.
Since the work of adhesion is a thermodynamic quantity, the
contact angle they used for the correlation was the static appar-
ent contact angle. Their correlation showed that, for smooth and
non-porous surfaces (Teon, glass and HMDS coated silicon
D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215 213
Fig. 10. The trend between the normalized work of adhesion (W
a
/
lv
) and the
minimum spreading ratio (D
min
/d). (a) Model surfaces: glass, cellulose lm,
0.2 g/L AKD-sized cellulose lm and Teon, (b) model surfaces and paper sur-
faces.
wafer), as the normalized work of adhesion (1 +cos ) increased,
the D
min
/d also increased. Their results therefore suggested that
a strong work of adhesion between the liquid and the sub-
strate reduces the liquid drop recoil. This method of correla-
tion was used to t the data obtained from the present work
(Fig. 10).
The relationship between the minimum spreading ratio
(D
min
/d) and normalized W
a
is shown in Fig. 10. Fig. 10(a)
shows that the minimum spreading ratio (D
min
/d) on glass,
cellulose lm, lightly sized (0.2 g/L AKD) cellulose lm and
Teon decreases with the decreasing normalized W
a
at all impact
velocities. Similar trend was reported for Teon, hydrophobized
silicon wafer (contact angle =73

), silicon wafer and glass sur-

faces by Park et al. [1]. However, the recoil of water drop from
paper surfaces did not follow similar trend (Fig. 10(b)).
This rather interesting behaviour of waterpaper interaction
suggests that care must be taken if one attempts to use an existing
model for ideal theory surfaces to predict the spreading and
recoil behaviour of a water drop on a paper surface. There are
probably two major reasons why paper behaves so differently to
other substrates.
Firstly, it is not easy (if not impossible) to obtain a thermody-
namically meaningful contact angle value from a paper surface.
The measured value of the apparent contact angle for a liquid on
a solid surface can be strongly inuenced by the paper surface
roughness. Shen et al. [22] showed that the work of adhesion
calculated using apparent static contact angle data on a sized
paper surface was not reliable. Most likely, the contact angle
method will underestimate the work of adhesion.
Secondly, sized papers usually show a substantial contact
angle hysteresis. Early studies by Hodgson and Berg [38]
showed that, while the advancing contact angles of heavy sized
cellulose bres with water ranged from 120 to 140

, the reced-
ing contact angles were always 0

. von Bahr et al. [9] also noted

that the receding contact angle for water on a heavy sized paper
was very low. These observations suggested that, although a
sized paper shows a high degree of resistance to water wetting
and penetration, it adheres well with water when it is in contact
with water. Shen et al. [22] showed that the work of adhesion
between water and sized paper calculated using data obtained
from inverse gas chromatography (IGC) was much higher than
that calculated using the apparent static contact angle data.
Whilst the contact angle results failed to reveal any hydrogen
bonding interactions between water and strongly sized papers
[11,22,31], IGCresults revealedthat there was a stronghydrogen
bonding capability for sized paper surfaces [22]. Their results
suggested that molecules of sizing chemical do not fully cover
the surface of the paper to completely cut off all hydrogen bond-
ing interactions between water and the paper surface. Instead,
the coverage of sizing chemical on paper surface is likely to be
only partial. However, the degree and the evenness of such distri-
bution on bre surface are unknown. It is likely, though, that the
distribution may not be uniformat a molecular level. This allows
the interaction (and therefore adhesion) to be established when
water is in contact with the paper surface. In fact, waterpaper
interactions under this condition can be better described by Eq.
(5). This analysis agrees with the research reported by Str om et
al. [34] that the coverage of AKD on a sized paper surface is far
less than a complete monolayer.
The above analysis allows the following picture to be formed:
the wettingconditionat the waterpaper contact line is verycom-
plicated, as there are areas which are not covered by the sizing
chemical which have very lowcontact angles, whereas areas that
are covered by the sizing chemical have higher contact angles.
The strong chemical heterogeneity of paper surface contributes
to the pinning of the waterpaper contact line, although paper
surface roughness is expected to also contribute to the pinning
[12]. The strong adhesion of water to areas of cellulose bre sur-
face that are not covered by AKD is established as they contact.
This will also allow water to penetrate into the bre wall.
To further discuss the waterpaper interaction, an AFM was
used to acquire images of a cellulose bre on a heavily sized and
calendered lter paper surface before and after it was brought
in contact with water (Fig. 11(a) and (b)). A consideration in
choosing the sized lter paper was that the inuence of ller
particles can be avoided. Since the water drop can be held sta-
bly on the sized paper surface, an AFM image was obtained
under the water environment (Fig. 11(b)). However, the degree
214 D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215
Fig. 11. (a) AFM image of a cotton lint bre on the surface of a strongly sized
and calendered lter paper. (b) AFM image obtained under water of the same
bre.
of difculty in performing the in situ AFM imaging on the same
bre after introducing water hampered the chance of collecting
high quality AFM image. Nevertheless, we believe that these
images contain some information that supports the strong inter-
action between water and a strongly sized paper surface.
The major feature of Fig. 11(a) is a thick and long cotton
lint bre laying diagonally across the gure. On the left side of
the gure are other bres at about the same height as the major
bre in the gure. Fig. 11(b) shows that after water was intro-
duced, the major bre in the gure appeared to be raised up.
The most likely cause for such a change to occur was that water
strongly interacts with the bre, causing the bre to swell. This
interaction therefore leads to a relaxation of the bre matrix.
Unfortunately, attempts of recording scans of smaller scales
were unsuccessful because of the continuous swelling of the
bre.
Such a strong interaction between water and the heavily sized
paper, however, cannot be correctlyreectedbythe contact angle
and the work of adhesion calculated from the contact angle (see
Table 2) for the reasons given previously. The present AFM
results once again showed that water can adhere strongly to the
sized paper surface possibly through areas that are not covered
by the domains of the sizing chemicals. These results are in good
agreement with the observation made by Roberts et al. [14] using
confocal microscopy where they showed water penetrated into
the wall of strongly sized bres after contact.
4.6. Implication of water impact behaviour on paper
surfaces to inkjet printing
The advancement of inkjet printing technology allows much
smaller ink droplets to be delivered on to paper surface. At the
same time, the ink droplet delivery rate has increased signi-
cantly [39]. The development of specialty inkjet papers empha-
sizes the use of a ink receptive coating. The coating layer must
not only absorb the ink rapidly, but also be able to chemically
immobilize the dye molecules. On the other hand, however, nor-
mal uncoated but sized ofce papers are still being used in large
quantities for inkjet printing. A factor which has been demon-
strated to be potentially detrimental to ink print quality is the ink
droplet amalgamation after being printed on the paper surface.
The ink droplet amalgamation obviously depends on the short-
est distance between the printed dots (droplets) on the paper
surface. Since the shortest distance between the printed dots on
a paper surface is determined by the forced spreading of the ink
droplet in the impaction phase and not affected by the process
of drop recoil, the droplet amalgamation is most likely be only
determined by the droplet impaction.
Print quality analysis of inkjet printed solid tone images on
some uncoated but strongly sized and supercalender papers often
show speckles of areas of 1 mm
2
. It is possible that the sur-
face conditions of the speckles are different from that of their
surrounding areas. These areas might either have a higher sizing
level (due to uneven size distribution) or have a greater smooth-
ness or both. In this case, less sized and less calendered areas
around the areas of speckles would absorb ink more strongly.
The speckle areas, which does not absorb ink at the rate as areas
surround it, would have a lower print density.
It is a well observed phenomenon that applying a strong
sizing only, without also applying other appropriate surface
treatments to paper, can signicantly increase the level of inkjet
print mottle [18]. Observations from the present study lend
some support to this phenomenon. On a sized paper surface,
ink droplets are likely to recoil quite strongly in the recoil
phase, since in the impact phase the sized paper behaves like
a hydrophobic surface. Such recoil may provide a chance for
the ink droplets to redistribute before being absorbed into paper.
Since the paper surface is microscopically rather rough, ink
droplets do not necessarily recoil perpendicularly to the paper
surface all the time. This makes the redistribution of the droplets
(before penetration) practically likely. The redistribution of the
droplets may encourage the formation of small reservoirs on
the paper surface and the subsequent penetration of ink in these
D. Kannangara et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 280 (2006) 203215 215
reservoirs may be an enhancing factor to the formation of print
mottle.
5. Conclusions
The impact of water drops ona solidsubstrate is highlydepen-
dent on the drop kinetic energy and the liquidsubstrate interac-
tions. Experiments conducted on hydrophilic and hydrophobic
model surfaces conrmed the following results from the liter-
ature. On a hydrophobic model surface, water drop spreads to
a maximum diameter (D
m
) under the force of impact and then
retracts from the surface to a minimum diameter (D
min
) which
is smaller than D
m
. On a hydrophilic model surface, water drop
spreads to a maximum diameter, but it does not strongly retract
from the surface. Therefore, for a hydrophilic model surface,
D
min
D
m
. This study found that the water drop impaction
on sized paper surfaces exhibits a switch of behaviour. At the
moment of drop impact, a sized paper behaves like a hydropho-
bic surface. During the drop recoil, sized paper behaves like a
hydrophilic surface. When the impacting drop reaches equilib-
rium, the sized paper again behaves like a hydrophobic surface,
allowing no water penetration. When water is in contact with the
sized paper surface, good adhesion between the water and the
sized paper surface can be established in the time scale of the
drop impaction. This adhesion prevents water fromretracting on
the sized paper surface. The results from this study are relevant
and useful to the understanding of inkjet ink droplets impaction
on paper surfaces.
Acknowledgements
This work was carried out as part of the research program for
the CRCSmartprint. Funding received fromthe Federal Govern-
ment and PaperlinXvia the CRCprogramis gratefully acknowl-
edged. One of us, H.Z. would like to gratefully acknowledge the
BASF for the nancial support through the joint research project
(No. 243-7249).
References
[1] H. Park, W.W. Carr, J.Y. Zhu, J.F. Morris, AIChE J. 49 (2003) 2461.
[2] T. Mao, D.C.S. Kuhn, H. Tran, AIChE J. 43 (1997) 2169.
[3] J. Fukai, Y. Shiiba, T. Yamamoto, O. Miyatake, D. Poulikatos, C.M.
Megaridis, Z. Zhao, Phys. Fluids. 7 (1995) 236.
[4] J. Fukai, M. Tanaka, O. Miyatake, J. Chem. Eng. Jpn. 31 (1998)
456.
[5] J. Fukai, Z. Zhao, D. Poulikakos, C.M. Megaridis, O. Miyatake, Phys.
Fluids A 5 (1993) 2588.
[6] A. Asai, M. Shioya, S. Hirasawa, T. Okazaki, J. Imaging Sci. Technol.
37 (1993) 205.
[7] J. Heilmann, Proceedings of NIP17: International Conference on Dig-
ital Printing Technologies 2001, VTT Information Technology, Espoo,
Finland, 2001, p. 735.
[8] M. Apel-Paz, A. Marmur, Colloids Surf. A 146 (1999) 273.
[9] M. Von Bahr, F. Tiberg, V. Yaminsky, Colloids Surf. A 193 (2001) 85.
[10] T. Onda, S. Shibuichi, N. Satoh, K. Tsujii, Langmuir 12 (1996) 2125.
[11] M. Von Bahr, Wetting and capillary ow of surfactant solutions and
inks, Ph.D. thesis, Lund University, Sweden, 2003.
[12] H. Modaressi, G. Garnier, Langmuir 18 (2002) 642.
[13] M.B. Lyne, J.S. Aspler, ACS Symposium Series 200 (1982) 385.
[14] R.J. Roberts, T.J. Senden, M.A. Knackstedt, M.B. Lyne, J. Pulp Paper
Sci. 29 (2003) 123.
[15] H. Yamazaki, U. Munakata, Proceedings of International Conference of
Paper Physics, 1994, p. 913.
[16] R.J. Roberts, T.J. Senden, M.A. Knackstedt, Proceedings of 5th Interna-
tional Paper and Coating Chemistry Symposium, Montreal, June 1619,
2003, p. 303.
[17] J.A. Bristow, in: J.A. Bristow, P. Kolseth (Eds.), Paper Structure and
Properties, Marcel Dekker, Stockholm, 1986, p. 183.
[18] D.F. Varnell, Pulp Paper Mag. Can. 99 (1998) 37.
[19] M. Von Bahr, F. Tiberg, B.V. Zhmud, Langmuir 15 (1999) 7069.
[20] R. Urquhart, Aust. Printer (2004) 63.
[21] H. Park, W.W. Carr, J. Zhu, Proceedings of NIP17: International Con-
ference on Digital Printing Technologies, Florida October, 2001, p. 438.
[22] W. Shen, Y. Filonanko, Y. Truong, I. Parker, N. Brack, P. Pigram, J.
Liesegang, Colloids Surf. A 173 (2000) 117.
[23] J.C. Berg, in: J.C. Berg (Ed.), Wettability, Marcel Dekker, New York,
1993, p. 75.
[24] L. Neimo, in: L. Neimo (Ed.), Papermaking Chemistry, Fapet Oy,
Helsinki, 1999, p. 151.
[25] R.J. Good, in: K.L. Mittal (Ed.), Contact Angle, Wettability and Adhe-
sion: Festschrift in honor of Professor Robert J. Good, Utrecht, The
Netherlands, 1993, p. 3.
[26] A.F. Toussaint, P. Luner, J. Adhesion Sci. Technol. 7 (1993) 635.
[27] L. Xu, I. Parker, Appita J. 53 (2000) 282.
[28] L. Cheng, Ind. Eng. Chem. Proc. D.D. 16 (1977) 192.
[29] F.M. Fowkes, M.A. Mostafa, Ind. Eng. Chem. Proc. D.D. 17 (1978) 3.
[30] D.K. Owens, R.C. Wendt, J. Appl. Polym. Sci. 13 (1969) 1741.
[31] D.T. Quillin, D.F. Cauleld, J.A. Koutsky, Int. J. Polym. Mater. 17
(1992) 215.
[32] Y. Huang, D.J. Gardner, M. Chen, C.J. Biermann, J. Adhesion Sci.
Technol. 9 (1995) 1403.
[33] W. Shen, I. Parker, Cellulose 6 (1999) 41.
[34] G. Str om, G. Carlsson, M. Kiar, Wochenblatt fur papierfabrikation 15
(1992) 606.
[35] M. Pasandideh-Fard, Y.M. Qiao, S. Chandra, J. Mostaghimi, Phys. Fluids
8 (1996) 650.
[36] R. Dettre, R. Johnson, in: F.M. Fowkes (Ed.), Contact Angle, Wettability
and Adhesion, American Chemical Society, Washington, DC, 1964, p.
p. 136.
[37] P.G. De Gennes, Rev. Mod. Phys. 57 (1985) 827.
[38] K.T. Hodgson, J.C. Berg, Wood Fibre Sci. 20 (1988) 3.
[39] H. Kipphan, Handbook of Print Media: Technologies and Manufacturing
Processes, Springer, 2001.