Journal of Colloid and Interface Science 274 (2004) 637–644

www.elsevier.com/locate/jcis

Contact angle measurement on rough surfaces

Tammar S. Meiron,a Abraham Marmur,b and I. Sam Saguy a,∗
a Institute of Biochemistry, Food Science and Nutrition, Faculty of Agricultural, Food and Environmental Quality Sciences,
The Hebrew University of Jerusalem, Israel
b Department of Chemical Engineering, Technion—Israel Institute of Technology, Haifa 32000, Israel

Received 18 August 2003; accepted 6 February 2004

Available online 25 March 2004

Abstract
A new method for the measurement of apparent contact angles at the global energy minimum on real surfaces has been developed. The
method consists of vibrating the surface, taking top-view pictures of the drop, monitoring the drop roundness, and calculating the contact
angle from the drop diameter and weight. The use of the new method has been demonstrated for various rough surfaces, all having the same
surface chemistry. In order to establish the optimal vibration conditions, the proper ranges for the system parameters (i.e., drop volume,
vibration time, frequency of vibration, and amplitude of vibration) were determined. The reliability of the method has been demonstrated
by the fact that the ideal contact angles of all surfaces, as calculated from the Wenzel equation using the measured apparent contact angles,
came out to be practically identical. This ideal contact angle has been compared with three methods of calculation from values of advancing
and receding contact angles.
 2004 Elsevier Inc. All rights reserved.

Keywords: Contact angle; Wetting; Roughness; Wenzel equation; Global minimum free energy; Vibration

1. Introduction could be calculated using appropriate correlations for inter-

facial tensions. However, most practical surfaces are rough
Wetting of solid substrates by liquids is a fundamental and heterogeneous to some extent. On such surfaces the only
phenomenon related to many applications, including lubrica- measurable value is the apparent contact angle, θap , which
tion, coating, printing, waterproofing, and detergency [1–3]. may be very different from the ideal contact angle. There-
Understanding and characterizing the wettability of solid fore, to calculate the surface energy of the solid, θap by itself
surfaces is thus of significant importance. Wettability is of- is insufficient, and additional information regarding the na-
ten characterized by measuring the contact angle formed ture of the surface is required. Moreover, on real surfaces
between a liquid drop and a solid surface. This measurement θap may vary from point to point along the contact line.
is considered to be a relatively simple, useful, and sensitive Therefore, contact angle measurement is practically mean-
tool for assessing surface energies of solids [4], although this ingful only in situations when the drop is axisymmetric in
is not straightforward [5–7]. In addition, surface roughness shape [9].
may be estimated from contact angle measurements [8]. Another characteristic phenomenon of nonideal surfaces
Were the characterized solid surface ideal (i.e., rigid, is contact angle hysteresis [4,10,11]. This implies that the
flat, smooth, chemically homogeneous, insoluble, and non- apparent contact angle may assume a range of values rather
reactive), contact angle measurement and its interpretation than a single, unique one. A necessary condition for con-
would be straightforward. In this case, the measured con- tact angle hysteresis to occur is the existence of multiple,
tact angle would corresponded to the ideal contact angle that metastable states that differ in energy and are separated from
is given by the well-known Young equation. From the ideal each other by energy barriers [4]. A sufficient condition for
contact angle hysteresis is the dependence of the metastable
contact angle and the Young equation, the surface energy
states on the drop volume [10,12]. The existence of a range
of apparent contact angles is the main difficulty in wettabil-
* Corresponding author. ity characterization, since it is not clear how to identify the
E-mail address: [email protected] (I.S. Saguy). ideal contact angle within the hysteresis range.
0021-9797/$ – see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcis.2004.02.036
638 T.S. Meiron et al. / Journal of Colloid and Interface Science 274 (2004) 637–644

The commonly used approach is to characterize a solid (c) calculating the contact angle from the drop diameter and
surface by the advancing (maximum) contact angle or the volume.
receding (minimum) one. This is, in essence, an empiri-
cal approach, since the theory of hysteresis on real, three-
dimensional solid surfaces is not sufficiently developed yet 2. Experimental
to enable the calculation of the ideal contact angle from
hysteresis measurements. Thus, it is as yet difficult to quan- 2.1. Rough surfaces
titatively assess the surface energy of a solid from advancing
or receding contact angle measurements. The other possible
approach is to search for the apparent contact angle corre- The rough model surfaces were prepared as follows: abra-
sponding to the global energy minimum (GEM, the lowest sive papers (WRN, Gamal Sarid, Holon, Israel, grit # 240,
energy out of all possible metastable states) [13]. This con- 320, and 1000) were attached to glass slides (50 × 50 ×
tact angle can, in principle, be correlated with the ideal con- 2 mm) using a double-sided adhesive tape. The rough sur-
tact angle. On a rough surface, this apparent contact angle face side was not touched. In addition, plain glass slides
is related to the ideal contact angle by the Wenzel equa- were used as model substrates of smooth surfaces. The glass
tion [14], if the drop is sufficiently large compared with the slides (either plain or with an attached abrasive paper) were
roughness scale [9,15], and if the liquid completely pene- vertically dipped into melted beeswax (Beeswax, bleached,
trates into the roughness grooves [16]: white, Aldrich, #24322-1, MP 63.5 ◦ C) at 65 ◦ C for about
10 s. Wax residues were removed by about 20 s of hot
cos θW = r cos θi . (1) air flow (∼80 ◦ C) from a fan vertically placed about 3 cm
In this equation, θW is the Wenzel contact angle, i.e., the ap- above the sample. The slides were then moved to covered
parent contact angle at the GEM, θi is the ideal contact angle, plastic Petri dishes and placed on a leveled plate for cool-
and r is the roughness ratio, defined as the ratio between the ing to ambient temperature (24 ± 2 ◦ C) for at least 2 h.
true and the apparent surface area of the solid. On a hetero- Each sample was used only once for contact angle measure-
geneous surface, the contact angle at the GEM is related to ments.
the ideal contact angle through the Cassie equation [17].
Johnson and Dettre [4] suggested that the advancing and 2.2. Scanning electron microscopy
receding contact angles might converge to a common value if
sufficient energy were supplied to overcome the energy bar- One cm2 of a model surface was coated with gold to a
riers between the metastable states. This indeed was demon- thickness of 250 Å and studied by SEM (JSM 35C, Jeol,
strated by Stokes et al. [18], by Andrieu et al. [19] for drops Kyoto, Japan). Images were taken while samples were ro-
on heterogeneous surfaces, and by Decker et al. [20] for tated by 60◦ to allow a 3D perspective. Two samples of each
vertical heterogeneous plates penetrating into a liquid. The roughness were scanned.
latter two groups applied vibrations to supply the neces-
sary energy. It seems reasonable to assume that this common 2.3. Liquids
value may be the contact angle corresponding to the global
minimum. To assess the validity of this assumption, an indi-
The liquids used for contact angle measurements were
cation that results from a mathematical analysis of the Wen-
distilled water (Riedel–de Haen, G. Chromasolv, for gradi-
zel equation may be used [9]. It was proven that when the
ent elution, #34877, Seelze, Germany), and ethylene glycol
drop is sufficiently large compared with the roughness scale
(J.T. Baker, #5387-05, minimum purity 99.0%, Phillipsburg,
(which is a necessary condition for the validity of the Wenzel
NJ).
equation), it becomes axisymmetric as it reaches the GEM.
The opposite statement has not been proven, but a practical
working hypothesis may be made that when a large drop on 2.4. Physical properties of test liquids
a rough surface becomes round following vibrations, it is at
the global minimum in energy. Surface tension of each liquid was measured by a com-
Thus, the objective of the current study is to present a puterized tensiometer (Sigma70, KSV, Helsinki, Finland)
methodology for measuring contact angles on rough sur- equipped with a platinum/iridium microroughened Wil-
faces, based on the above-mentioned considerations. This helmy plate. The probe was dipped at 10 mm/min, and
approach consists of stopped at 4 mm immersion. Each liquid was tested 10 times
with three repetitions at 24 ± 2 ◦ C and ambient relative hu-
(a) vertically vibrating the surface in order to allow the drop midity of about 55%.
to reach the GEM; Liquid density was measured with a calibrated 50-ml pic-
(b) measuring drop circularity by image analysis of a top nometer, with an accuracy of ±1% at 24 ± 2 ◦ C, and viscos-
view of the drop to confirm the required axisymmetry; ity was taken from the literature [21]. Table 1 summarizes
and the physical properties of the liquids.
 T.S. Meiron et al. / Journal of Colloid and Interface Science 274 (2004) 637–644 639

Table 1
Physical properties of test liquids
Liquid Surface tension Density Viscositya
mN/m g/cm3 mPas
(24 ± 2 ◦ C) (24 ± 2 ◦ C) (25 ◦ C)
Water 72.0 ± 0.1 0.998 0.89
Ethylene glycol 48.8 ± 0.0 1.110 16.10
a From Ref. [21].

2.5. Roughness measurement

Surface roughness was determined using an optical pro-

filometer (Model UB16, UBM Messtechnik GmbH, Ettlin-
gen, Germany) equipped with a microfocus measurement
system with a laser sensor. Two roughness parameters were
recorded: Ra , the arithmetical mean deviation of the vertical
profile, and λ, the mean horizontal wavelength [22].
The roughness parameter of interest in contact angle cal-
culations is the roughness ratio, r, which is used in the Wen-
zel equation (1). As explained above, this roughness ratio is
defined as the ratio of the true area of the solid surface to
the apparent area. The experimental measurement of rough-
ness by the above profilometer is done, however, by scanning
of lines, so a method for estimating r from linear measure- Fig. 1. Vibration system for contact angle measurements: (1) CCD camera
ments needed to be developed. It is very difficult to develop a and lens; (2) light source; (3) loudspeaker; (4) sample; (5) vibration input.
general method for this purpose; hence a simplistic model of
regular roughness was assumed in order to get an indication
of the nature of the correlation between r and the linear mea- tact angle corresponding to the GEM: (a) the drop should
surement. The simplistic model assumed the rough surface be vibrated to enable overcoming of energy barriers, and
to consist of square pyramids with a basis a and a height c. (b) the drop size should be sufficiently large, compared with
For this model, it turns out that the roughness ratio and the the scale of surface roughness. As the drop approaches the
linear roughness ratios, rl , in the x and y directions are all GEM, it becomes axisymmetric [9]. This characteristic can
equal and given by be measured by analyzing the top view of the drop and may

serve as a means to follow the approach to the GEM.
 2
2 a Based on this approach, a top-view vibration system was
r = rlx = rly = + c2 . (2)
a 2 built. The system (Fig. 1) consists of a CCD video cam-
era (Cohu Incorporation, 4913-5010 monochrome camera
Thus, this simplistic model demonstrates that if rl is similar
CCIR, San Diego, CA), a lens (Nikon Company, AF Micro-
in the x and y directions, it may serve as a good approxima-
Nikkor 60mmf/2.8D, Tokyo, Japan), a light source (Leica
tion for r. This conclusion was adopted for the interpretation
of the results in the present paper. Company, Model CLS100X, Heerbrugg, Switzerland), and a
The linear roughness ratios, rlx and rly , were calculated fully computerized, controlled loudspeaker (Megavox Com-
from the measured surface profiles by dividing the lengths pany, MGX-600 6 ½ 200 W, Taiwan). The rough model
of the roughness profiles in the x and y directions by the surfaces were horizontally placed on the center of the leveled
lengths traversed in these directions. The experimental re- loudspeaker, fastened with three clips. Liquid drops were de-
sults showed no significant changes in r (or in Ra and λ) posited on the tested surface by a micrometer syringe. The
values measured in the x and y directions. This is realistic, camera axis and the light source were vertical to the studied
since surface roughness in the model surfaces used for this surface. Images of the drops on the surfaces were captured
study appears to be random. Five to ten lines were scanned on a G4-400 Macintosh computer with a CG-7 PCI frame
in the x and y directions for each surface. The roughness grabber (Scion Corporation, Frederick, MD) and analyzed
ratio was taken as the average value of all measurements. (Scion Image 1.62c, NIH-Image Software, NIH, Bethesda,
Maryland). Image analysis of the top view of the drop was
2.6. Contact angle measurements used to calculate the area of the drop. To determine the
drop volume, the tested surfaces were weighed with a mi-
2.6.1. Top-view vibration system crobalance (Precisa Company, Model XT 220A, Dietikon,
As explained in Section 1, two fundamental requirements Switzerland) before and after the drop was applied, and the
have to be satisfied in order to identify and measure the con- volume was calculated using the liquid density.
640 T.S. Meiron et al. / Journal of Colloid and Interface Science 274 (2004) 637–644

(±0.1 mN/m), drop volume (±1.0 µl), and drop radius

(±0.05 mm), the errors in calculating the contact angle were
simulated to be ±1 ◦ . Assuming normal distribution of the
measured data, this error should be regarded as equal to
±2σ , where σ is the standard deviation.

2.6.2. Advancing and receding contact angles—side view

measurements
Advancing and receding contact angles of the liquids on
the test surfaces were measured by a computerized con-
tact angle analyzer (CAM200, KSV, Helsinki, Finland), with
a reported accuracy of ±1◦ . A micro syringe (Hamilton
#81341, 0–1000 µl, Merck Eurolab, Niittyrinne, Finland)
Fig. 2. Effect of input power on the vibration amplitude, at various frequen- was used to create the drop on the surface. Liquid was
cies. (F) 30 Hz; (2) 40 Hz; (Q) 50 Hz; (") 80 Hz. continuously added to measure the advancing contact angle
(drop maximum volume was 0.4 ml). To measure the reced-
The loudspeaker vibrations were controlled with the soft- ing contact angle the liquid was continuously removed. The
ware Cool Edit (Syntrillium Software Company, CoolEdit needle was kept in the drop throughout the procedure. Side
2000, Phoenix, AZ). Software setups were sample rate— view images were captured at a rate of 10 frames/s. Con-
6000 points/s; channel—mono; resolution—8-bit; base tact angles were automatically calculated by the instrument
frequency—30 to 80 Hz; duration 0, 100, and 300 s; wave (by fitting the captured drop shape to the shape calculated
shape—sinus; linear fade-in—30 s; linear fade-out—30 s. from the Young–Laplace equation). Measurements were pre-
The vibration amplitude was controlled with a rheostat (al- formed at ambient temperature (24 ± 2 ◦ C) and repeated at
lowing 10 full turns) connected to an amplifier (0–5 V). least 10 times on fresh surfaces at an advancing or receding
Input voltage and loudspeaker frequency were measured rate of 0.05 ml/min. The measured contact angle was iden-
by a datalogging multimeter (Extech Instruments, ML720, tified as the advancing contact angle when the angle did not
Waltham, MA). Input power was calculated by dividing the increase significantly with drop volume increase. Similarly,
square of the voltage by the loudspeaker resistance. Maxi- the receding contact angle was identified when the angle did
mal working input power was 6.3 W. not decrease considerably with drop volume decrease.
The vibration amplitude of the loudspeaker was quanti-
fied from side-view pictures of the loudspeaker movement.
The amplitude measurement was calibrated against a milli- 3. Results and discussion
metric ruler. The experimental error was taken as the width
of the ruler line. Each measurement was repeated at least
3.1. Surface roughness
five times. The effect of input power on loudspeaker vi-
bration amplitude at different frequencies is represented in
Fig. 2. As can be expected, the amplitude increases with in- In order to demonstrate the feasibility of the proposed
put power, and this increase is larger the lower the frequency. method for measuring contact angles, model surfaces were
The loudspeaker performance is limited at frequencies lower used. All these surfaces have the same surface energy, but
than 30 Hz; therefore, this was the lowest frequency applied vary in their roughness over a wide range. The surface en-
in this study. ergy is the same, because all surfaces were coated by the
The criterion for drop axisymmetry was the roundness, same wax. The roughness is determined by the grit num-
Rd , defined as ber of the abrasive paper. The measured roughness parame-
ters are shown in Table 2. In addition, SEM images of the
4S
Rd = 2
, (3) smoothest and roughest surfaces prior to and after coating
πDmax are shown in Fig. 3.
where S is the area of the drop base and Dmax is the maxi-
mal base diameter as derived from the image analysis. Only
drops exhibiting Rd
0.95 were considered axisymmetric Table 2
and used for the analysis of the results. Roughness parameters of model surfaces

Contact angles were calculated by a computer program Surface Ra (µm) λ (µm) r

that fits a solution of the Young–Laplace equation [23] to Smooth 1.3 ± 0.1 94 ± 10 1.01
the measured radius of the base of the drop, given its vol- Abrasive paper #1000 5.3 ± 0.3 134 ± 15 1.03
Abrasive paper #320 9.3 ± 1.1 239 ± 21 1.09
ume, surface tension, and density. Taking into account the
Abrasive paper #240 11.2 ± 1.3 234 ± 20 1.25
following errors in the measured variables, surface tension
 T.S. Meiron et al. / Journal of Colloid and Interface Science 274 (2004) 637–644 641

Fig. 4. Ethylene glycol drops (520 µl) vibrated at 30 Hz on a model surface

(Ra = 9.27 µm; λ = 239 µm; r = 1.09): (a) 0.00 mm (before vibration);
(b) 0.5 mm during vibration; (c) 0.7 mm during vibration.

for a 520-µl ethylene glycol drop, vibrated at 30 Hz on one of

the rough surfaces (Ra = 9.27 µm; λ = 239 µm; r = 1.09).
Figure 4a shows the initially deposited drop. Figure 4b il-
lustrates the drop under intermediate vibration conditions
(amplitude 0.5 mm). The drop is more circular than at the
Fig. 3. SEM images (X200, 60◦ ) of the surface prior to and after coating initial stage, and its shape is kept when the vibrations are
with beeswax: (a and b) the smoothest surface, respectively; (c and d) the stopped. However, when the loudspeaker amplitude was in-
roughest surface, respectively. creased above a certain value (e.g., 0.7 mm), the drop shape
became irregular, as shown in Fig. 4c.
3.2. Apparent contact angles measured by the top-view The transition to irregular drop shapes depends also on
vibration system the drop volume and time of vibration. As the drop volume
decreases, the transition to irregular shapes occurs under
To initiate a routine procedure for contact angle mea- more extreme vibration conditions. Thus, vibration condi-
surement by the vibration system, the proper ranges for the tions that are suitable for large drops will necessarily be
system parameters needed first to be determined. These in- appropriate also for smaller drops. In addition, the transi-
cluded the drop volume, V , vibration time, t, frequency of tion to irregular shapes develops with time; therefore, too
vibration, f , and amplitude of vibration, A. Since many pa- long vibration times are disadvantageous. Taking into con-
rameters are involved, initial values had to be assumed for sideration all the above arguments, 520-µl ethylene glycol
the drop volume and vibration time, as explained in the fol- drops and 830-µl water drops were vibrated for 100 s at dif-
lowing. ferent frequencies (30, 40, 50, and 80 Hz) and amplitudes
As stated in Section 1, measurements of apparent contact (0.0–1.0 mm).
angles on rough surfaces are meaningful and can be inter- In order to minimize the number of independent variables
preted only when the drop is axisymmetric and sufficiently it may be useful to combine the amplitude and frequency
large (compared with the scale of roughness) [9]. The maxi- of vibrations into a single parameter that makes physical
mal diameter of drop that could be employed in our system sense. Since the input wave into the loudspeaker was sinu-
was about 20 mm. Accordingly, the maximal ethylene glycol soidal, it is reasonable to consider the vibration motion as
and water drop volumes were about 520 and 830 µl, respec- approximately harmonic. Thus, the product f A represents
tively. This implies that the drop diameter was about three the maximum velocity and f 2 A stands for the maximum ac-
orders of magnitude larger than the average vertical irregu- celeration. Figures 5 and 6 show the dependence of the drop
larities of the roughest surface (Ra = 11.23 µm) and almost roundness and measured contact angle on f A and f 2 A, re-
two orders of magnitude larger than its average wavelength spectively, for one of the model surfaces (Ra = 9.27 µm;
(λ = 234 µm). It is assumed that these ratios between drop λ = 239 µm; r = 1.09). Figures 5a and 6a are for ethyl-
base diameter and surface roughness parameters are suffi- ene glycol, and Figs. 5b and 6b are for water. These figures
ciently large for the Wenzel equation to hold. The validity of demonstrate that the correlation between roundness and f A
this assumption will be demonstrated below. appears to be more unique in comparison with the correla-
Regarding vibration time, it was observed in preliminary tion with f 2 A. The same conclusion holds for the correla-
experiments that most changes in drop roundness and con- tion between the contact angle and f A. Consequently, f A
tact angle values occurred during the initial 60 s of vibration. was chosen as the appropriate independent variable for fur-
On the other hand, too long vibration times are inconvenient, ther analysis of Rd and θ . The same pattern was observed
may lead to significant evaporation, and may also induce ir- with all studied surfaces and for smaller drop volumes. The
regular drop shapes. Therefore, the vibration time was set at physical interpretation of this conclusion is that the kinetic
100 s. energy of the drop (proportional to f 2 A2 ) is the determining
The vibration conditions need to be such that the drop factor.
may overcome energy barriers on its way to the GEM; how- Vibrations are applied to the drop in order to bring it to
ever, they should not be too extreme, in order to keep the its GEM [4,19,20,25]. As is well known [4,12], the energy
drop shape symmetric. Typical shapes are shown in Fig. 4 barriers to changes in the contact angle increase as the drop
642 T.S. Meiron et al. / Journal of Colloid and Interface Science 274 (2004) 637–644

(a) (a)

(b) (b)

Fig. 5. (a) Effect of (f A) on θ and Rd (liquid—ethylene glycol; V —520 µl; Fig. 6. Effect of a (f 2 A) on θ and Rd (liquid—ethylene glycol; V —520 µl;
surface—#320; t—100 s): !—Rd ; Q—θ . (b) Effect of v (f A) on θ and Rd surface—#320; t—100 s); !—Rd ; Q—θ . (b) Effect of a (f 2 A) on θ and
(liquid—water; V —830 µl; surface—#320; t—100 s); !—Rd ; Q—θ . Rd (liquid—water; V —830 µl; surface—#320; t—100 s). !—Rd ; Q—θ .

approaches the GEM. Therefore, it is reasonable to assume

that when the contact angle stops varying with an increase in
the vibration energy, the drop is approximately at its GEM.
In addition, as explained in Section 1, the drop has to be
sufficiently round (Rd
0.95). As can be seen in Figs. 5a
and 5b, Rd increases with f A up to a maximum and then
decreases; θ decreases with f A and then becomes constant.
It is important to notice that the maximum in roundness and
the leveling of the contact angle occur approximately at the
same value of f A. Therefore, the values of θ that are ex-
pected to represent the GEM are 71.0 ± 0.7◦ and 98.0 ± 0.8 ◦
for ethylene glycol and water, respectively.
In order to demonstrate that smaller drop volumes do not
enable proper measurement of the contact angle, smaller Fig. 7. Effect of v (f A) on θ and Rd (liquid—ethylene glycol; V —430 µl;
surface—#320; t—100 s). !—Rd ; Q—θ .
drops of ethylene glycol were used. Figure 7 demonstrates
that for 430-µl drops, it is rather difficult to decide whether
the value of the contact angle becomes level before the ommendation that contact radius of a measured drop should
roundness decreases. Figure 8 shows that for 320-µl drops be 0.4–0.5 cm if the solid surface is not very smooth [24].
the contact angle clearly does not become level before the The conditions for measuring contact angles by the vibra-
roundness decreases. These results evidently demonstrate tion method having been established, the effect of surface
the necessity of using large drops. Unfortunately, volumes of roughness on the contact angle can now be studied. The
tested drops are not often reported in the literature; therefore, measured apparent contact angles for the model rough sur-
it is difficult to compare our experimental results with pre- faces are summarized in Table 3. This table also shows the
vious data. However, our results are consistent with the rec- values of the ideal contact angles, as calculated from the
 T.S. Meiron et al. / Journal of Colloid and Interface Science 274 (2004) 637–644 643

Table 4
Advancing and receding contact angles of water and ethylene glycol (eg)
drops on the rough beeswax surfaces
r θad (water) (deg) θre (water) (deg) θad (eg) (deg) θre (eg) (deg)
1.01 111 ± 1 62 ± 1 83 ± 1 51 ± 1
1.03 116 ± 1 48 ± 2 88 ± 1 41 ± 2
1.09 120 ± 1 42 ± 1 93 ± 1 36 ± 1
1.25 127 ± 1 40 ± 1 99 ± 1 28 ± 1

Table 5
Ideal contact angles, θi , calculated from the Wenzel equation, hysteresis
average, averaged cos θad and cos θre and linear regression for water and
ethylene glycol (eg) drops on rough beeswax surfaces

Fig. 8. Effect of v (f A) on θ and Rd (liquid—ethylene glycol; V —320 µl; θi (water) (deg) θi (eg) (deg)
surface—#320; t—100 s). !—Rd ; Q—θ . Wenzel equation (present study) 97.6 ± 0.4 73.3 ± 0.4
θad and θre averagea 83.3 ± 2.4 65.0 ± 1.5
θad and θre cosine averageb 87.3 ± 0.9 79.2 ± 0.4
Table 3
Linear regressionc 90.3 69.4
Apparent contact angles, θ , and ideal contact angles, θi , calculated from the
a From Ref. [20].
Wenzel equation for water and ethylene glycol (eg) drops on rough beeswax
b From Ref. [19].
surfaces
c From Ref. [25].
r θap (water) (deg) θi (water) (deg) θap (eg) (deg) θi (eg) (deg)
1.01 97.3 ± 0.3 97.3 ± 0.3 73.1 ± 0.7 73.2 ± 0.7
1.03 97.9 ± 0.2 97.7 ± 0.2 72.4 ± 0.4 72.9 ± 0.4 from 32◦ to 71◦ with water and ethylene glycol, respectively.
1.09 98.0 ± 0.8 97.3 ± 0.7 71.6 ± 0.4 73.2 ± 0.4 It is worth noting that all currently studied surfaces exhib-
1.25 100.1 ± 0.9 98.1 ± 0.7 69.7 ± 0.6 73.9 ± 0.5 ited hysteresis, even the smoothest one. It will be important
in the future to perform similar studies with a hysteresis-
Wenzel equation (1). It is important to notice that the cal- free reference surfaces. Table 5 compares the present results
culated ideal contact angles on the various rough surfaces for θi that are shown in Table 3 (based on measurements
turn out to be practically identical. This is to be expected for at the GEM and the Wenzel equation) with values calcu-
surfaces that consist of the same wax; however, the fact that lated by the above three methods. It appears that predictions
this is indeed experimentally confirmed serves as a strong by the method of Kamusewitz et al. [25] are closest to the
indication that the measured apparent contact angles indeed present experimental data; however, there are meaningful
relate to the GEM. differences between them.

3.3. Advancing and receding contact angles—side view

measurement 4. Summary and conclusions

A few different methods were suggested in the literature The following points summarize the main achievements
for calculating the contact angle at the GEM from contact of the present study:
angle hysteresis [19,20,25]. One approach proposed that the
ideal contact angle could be approximated by the average (1) A new method for the measurement of apparent contact
of the advancing and receding angles, θad and θre , respec- angles at the global energy minimum on real surfaces
tively, [20]. Alternatively, the average of cos θad and cos θre has been developed. The method consists of vibrating
was employed to estimate the average spreading parameter the surface, taking top-view pictures of the drop, mon-
of heterogeneous surfaces [19]. Another approach suggested itoring the drop roundness, and calculating the contact
using linear regression between θad or θre and the contact an- angle from the drop diameter and weight.
gle hysteresis; the extrapolated value of the contact angle at (2) The use of the new method has been demonstrated
zero hysteresis was assumed to be θi [25]. It is of interest for various rough surfaces, all having the same surface
to compare the contact angles as assessed by these methods chemistry.
with the direct measurement by the top-view vibration sys- (3) The effects of the various system parameters on the mea-
tem. For this purpose, advancing and receding contact angles surement process have been analyzed and optimized.
of water and ethylene glycol were measured. (4) The reliability of the method has been demonstrated by
Table 4 summarizes the advancing and receding contact the fact that the ideal contact angles of all surfaces, as
angles measured on the rough beeswax surfaces. The hys- calculated from the Wenzel equation, using the mea-
teresis range (i.e., the difference between θad and θre ) in- sured apparent contact angles, came out to be practically
creased with surface roughness, ranging from 49◦ to 87◦ and identical.
644 T.S. Meiron et al. / Journal of Colloid and Interface Science 274 (2004) 637–644

Acknowledgments [8] R.J. Good, M.K. Chaudhury, C. Yeung, in: First International Congress
on Adhesion Science and Technology—Invited Papers, 1998.
[9] G. Wolansky, A. Marmur, Colloids Surf. A Physiochem. Eng. As-
This research was partially supported by Nestec Ltd., pects 156 (1999) 381.
Nestle Research Center, Switzerland. We also express our [10] A. Marmur, Adv. Colloid Interface Sci. 50 (1994) 121.
gratitude to Dr. Martin Leser, Dr. Guy Mayor, and Dr. Herib- [11] E. Chibowski, Adv. Colloid Interface Sci. 103 (2003) 149.
ert Watzke, Nestlé Research Center, Lausanne, Switzer- [12] A. Marmur, J. Colloid Interface Sci. 168 (1994) 40.
[13] A. Marmur, J. Imaging Sci. Technol. 44 (5) (2000) 406.
land. [14] R.N. Wenzel, Ind. Eng. Chem. 28 (1936) 998.
[15] S. Brandon, N. Haimovich, E. Yeger, A. Marmur, J. Colloid Interface
Sci. 263 (1) (2003) 237.
References [16] A. Marmur, Langmuir 19 (2003) 8343.
[17] A.B.D. Cassie, Discuss. Faraday Soc. 3 (1948) 11.
[18] J.P. Stokes, M.J. Higgins, A.P. Kushnick, S. Bhattacharya, Phys. Rev.
[1] A.W. Adamson, Physical Chemistry of Surfaces, Wiley, New York, Lett. 65 (15) (1990) 1885.
1990, p. 493. [19] C. Andrieu, C. Sykes, F. Brochard, Langmuir 10 (1994) 2077.
[2] P. Hiemenz, R. Rajagopalan, Principles of Colloid and Surface Chem- [20] E.L. Decker, B. Frank, Y. Suo, S. Garoff, Colloids Surf. A Phys-
istry, Dekker, New York, 1997, p. 248. iochem. Eng. Aspects 156 (1999) 177.
[3] A. Marmur, Colloids Surf. A Physiochem. Eng. Aspects 116 (1996) [21] D.R. Lide, CRC Handbook of Chemistry and Physics, eighty second
55. ed., CRC Press, Boca Raton, FL, 2001.
[4] R.E. Johnson, R.H. Dettre, in: E. Matijevic (Ed.), Surface and Colloid [22] T.R. Thomas, Rough Surfaces, first ed., Longman, New York, 1982.
Science, Wiley–Interscience, New York, 1969, p. 85. [23] H.M. Princen, in: E. Matijevic (Ed.), Surface and Colloid Science,
[5] C. Della Volpe, S. Siboni, M. Morra, Langmuir 18 (4) (2002) 1441. Wiley–Interscience, New York, 1969, p. 1.
[6] S. Shalel-Levanon, A. Marmur, J. Colloid Interface Sci. 262 (2003) [24] D.Y. Kwok, A. Leung, C.N.C. Lam, A. Li, R. Wu, A.W. Neumann,
489. J. Colloid Interface Sci. 206 (1998) 44.
[7] S. Shalel-Levanon, A. Marmur, J. Colloid Interface Sci. 268 (2003) [25] H. Kamusewitz, W. Possart, D. Paul, Colloids Surf. A Physiochem.
272. Eng. Aspects 156 (1999) 271.