Superhydrophilicity of Regular
Elena Martines,†Kris Seunarine,‡Hywel Morgan,§Nikolaj Gadegaard,‡
Chris D. W. Wilkinson,‡and Mathis O. Riehle*,†
Centre for Cell Engineering, IBLS, UniVersity of Glasgow, Glasgow, G12 8QQ, U.K.,
Department of Electronics and Electrical Engineering, UniVersity of Glasgow,
Glasgow, G12 8QQ, U.K., and School of Electronics and Computer Science,
UniVersity of Southampton, S017 1BJ, U.K.
Received July 25, 2005
The hydrophilicity, hydrophobicity, and sliding behavior of water droplets on nanoasperities of controlled dimensions were investigated
experimentally. We show that the “hemi-wicking” theory for hydrophilic SiO2samples successfully predicts the experimental advancing angles
and that the same patterns, after silanization, become superhydrophobic in agreement with the Cassie−Baxter and Wenzel theories. Our
model topographies have the same dimensional scale of some naturally occurring structures that exhibit similar wetting properties. Our
results confirm that a forest of hydrophilic/hydrophobic slender pillars is the most effective superwettable/water-repellent configuration. It is
shown that the shape and curvature of the edges of the asperities play an important role in determining the advancing angles.
Introduction. The wettability of solid surfaces is a subject
that has raised great interest in the past few decades. The
surface energy of a sample will determine if a given liquid
drop will roll up or spread when deposited on it. Roughening
the surface enhances its repellent or wetting properties,1
resulting in “superhydrophobic” or “superhydrophilic” tex-
tures (this nomenclature applies if the liquid considered is
water, as is the case in this work).
Many authors have contributed to the fabrication and
understanding of superhydrophobic surfaces.2-14O ¨ner and
McCarthy10describe a superhydrophobic surface as one
where the advancing angle, θadv, is very high (generally
>150°), and the receding angle, θrec, is such that the drop
exhibits low hysteresis ∆θ (∆θ ) θadv- θrec). Water drops
form beads and roll off this kind of surface, cleaning it in
the process. This phenomenon has been termed the “Lotus
effect” because it is very pronounced on the leaves of the
lotus plant (Nelumbo nucifera).14,15These leaves exhibit a
double-structured roughness, where submicrometric wax
crystals cover a larger micrometric structure; even though
double-scale roughness has been proven to enhance water
repellency,16it has been suggested that the small scale
roughness plays an important role.12The wings of some
insects (e.g., Pflatoda claripennis and in the family Rhino-
termitidae) are also covered with nanometric structures that
are thought to ensure water repellency as well as other
properties;17unlike the irregular topography of wax crystals,
these surfaces are covered with ordered arrays of rounded
The effect of surface roughness on hydrophobicity has
been explained by two different theories. According to
the model developed by Wenzel, it is assumed that the
space between the protrusions on the surface is filled by
the liquid3(Figure 1); this model predicts that both hydro-
phobicity and hydrophilicity are reinforced by the roughness,
* Corresponding author. E-mail: email@example.com; tel: 0044-141-
3302931; fax: 0044-141-3303730.
†Centre for Cell Engineering.
‡Department of Electronics and Electrical Engineering.
§University of Southampton.
Figure 1. Schematic of roughness-filling by water according to
the Wenzel and Cassie-Baxter models.
Vol. 5, No. 10
10.1021/nl051435t CCC: $30.25
Published on Web 10/12/2005
© 2005 American Chemical Society
according to the following relation
where θWis the apparent contact angle on a rough surface,
θYis the ideal contact angle (Young’s angle) of water on a
smooth surface of identical chemistry, and r is the roughness
factor, which is defined as the ratio of actual surface area
over the projected area.
The approach developed by Cassie and Baxter, however,
assumes that air is trapped by the asperities2so that the drop
sits on a composite surface made of air and solid (Figure
1); the relation between the apparent contact angle θCBand
the ideal angle θYis in this case described as
where rfis the roughness factor of the wetted area, and f is
the area fraction of the projected wet area. The product rff
is often called the solid fraction φS.
Both Wenzel and Cassie-Baxter relations were originally
formulated for static drops at equilibrium; yet because it was
shown that low-rate advancing angles and static angles are
essentially identical,18eqs 1 and 2 can be applied to
advancing angles. Fewer attempts have been made to model
the receding angles.11,19It has been shown4that a droplet
can be in either a Cassie-Baxter or a Wenzel state on a
rough hydrophobic surface depending on how it is formed.
Because the advancing angles predicted by both the Cassie-
Baxter and Wenzel theories can be very close to the
experimental values, the receding angles can be used as a
qualitative indication of the state of the drop: if θrecis high
(i.e., the hysteresis is low), then the drop will be in a Cassie-
Baxter (“slippy”) mode; if θrecis low, then the drop will be
in a Wenzel (“sticky”) state. Therefore, obtaining a stable
Cassie-Baxter drop is the ultimate goal for achieving
superhydrophobicity by tailoring surface topography. The
adhesive behavior of water on rough surfaces can also be
assessed by sliding-angle measurements.5,6,9The sliding angle
is defined as the critical angle where a water droplet begins
to slide down an inclined plate: a high sliding angle indicates
a sticky Wenzel state, whereas a low sliding angle suggests
a Cassie-Baxter regime (i.e., the drop will easily roll off a
slightly tilted substrate).4
Much less work has been dedicated to the study of
superhydrophilic (or more generally, superwetting) sur-
faces.1,20,21Assuming that no air is trapped in the roughness
of the hydrophilic surface (for the opposite case, see
Abdelsalam et al.22), the Wenzel model still applies, along
with the hemi-wicking or “composite-drop” model,23where
the drop is assumed to be sitting on a composite surface made
of solid and water.
All of the above-mentioned theories for both hydrophobic
and hydrophilic rough surfaces have been tested by preparing
surfaces structured on the micrometer scale;5-7,9,10,14,16,20only
a few studies have been done recently on nanotopogra-
phies,8,22,24,25mainly because of the difficulty of fabrication.
In the present study, we have fabricated ordered arrays of
nanopits and nanopillars and investigated their dynamic
wettability before and after chemical hydrophobization. Our
model structures resemble the natural submicrometric fea-
tures that ensure the water repellency of biological surfaces
such as the lotus leaf and some insect wings.
Because of the accurate geometrical characterization of
our nanopatterns, the validity of the analytical models
currently available for predicting the wettability of rough
surfaces can be verified, for both the hydrophilic and the
hydrophobic case. To the best of our knowledge, this is the
most extensive quantitative study of the wetting properties
of small-scale topography reported so far.
Experimental Details. (A) Fabrication. Nanopatterns with
increasing solid fraction (two samples with pits and four with
pillars) were fabricated in silicon wafers (4 in., 〈100〉,
p-doped, 525 ( 50 µm thick) across an area of 1 × 1 cm2.
(1) Nanopillars. A 2-nm titanium layer was evaporated on
the silicon (Plassys evaporator), and the samples were spin-
coated immediately with a 60% NEB31A3-40% EC solvent
(Sumitomo Chemical Co Ltd) at 3 krpm for 60 s (150-nm-
thick coating). After a preexposure bake at 90 °C for 2 min,
the wafers were exposed in the e-beam writer (Leica
Microsystems EBPG 5) with the desired pattern. After a
postexposure bake at 85 °C for 2 min, we developed the
samples in Microposit MF CD-26 for 20 s and rinsed them
with reverse osmosis (RO) water; the titanium was etched
(1 part HF: 26 parts RO water) for approximately 2-3 s,
and the samples were rinsed in RO water. The silicon was
then dry-etched using STS-ICP (Surface Technology Systems-
Inductively Coupled Plasma) with C4F8and SF6(unswitched
gases) at an etch rate of 100 nm/min, and finally piranha-
cleaned for 5 min. (2) Nanopits. The silicon wafers were
spin-coated with 40% ZEP520A at 5 krpm (100-nm-thick
coating) and baked at 180 °C for 1 h. The samples were
then exposed in the e-beam writer, developed in o-xylene
for 60 s, and dry-etched as for the pillars. Finally, the surfaces
were piranha-cleaned for 5 min.
(B) Surface Modification. Prior to the measurements of
contact angles on the hydrophilic nanopatterns, we cleaned
all of the samples with an O2plasma for 15 min (BP80 RIE,
flow rate 20 sccm, pressure 30 mT, RF power 100 W); the
contact angles on these surfaces were measured within 24
Subsequently, the same patterns were coated with octa-
decyltricholorosilane (OTS) by modifying the procedure of
Rosloznik et al.:26all of the samples were sonicated for 10
s in 1:1 water-ethanol and 10 s in IPA, then rinsed in
chloroform and blow-dried. After a 15 min O2cleaning, we
sonicated them for 10 s in chloroform and 10 s in IPA, rinsed
them in 1:1 water-ethanol, rinsed them in RO water, and
blow-dried them. The clean samples were then placed in a
ceramic slide holder, which was gently tilted on the side at
90° so that the patterns were facing down. This technique
ensured that if any OTS clusters formed in the solution, then
they had less of a chance of being deposited on the patterned
surfaces. The tilted holder was placed in a glass beaker filled
with a 0.001% solution of OTS (Sigma) in heptane (Sigma).
cos θW) r cos θY
cos θCB) rff cos θY+ f - 1(2)
Nano Lett., Vol. 5, No. 10, 2005
After 3.5 h, we sonicated the holder three times for 1 min in
copious amounts of heptane, then rinsed it in IPA, 1:1 water-
ethanol, and RO water and finally blow-dried it.
(C) Sample Characterization. (1) Scanning Electron
Microscopy. The surfaces were imaged with a Hitachi S4700
prior to and following hydrophobization. After contact angle
measurements, we cleaved them and their SEM profiles were
used to measure the dimensions of the asperities with
ImageJ.27(2) Dynamic Contact Angle Measurements. Images
of the advancing and receding contact angles of filtered
Milli-Q water were captured at a rate of 2 images/s with a
long-distance objective connected to a CCD camera and
analyzed with the FTÅ200 software (First Ten Ångstroms,
v2.0). Water drops were deposited and taken up through a
30-gauge flat-tipped needle, at a rate of 0.25 µL/s; the
maximum volume of the drops was 5 µL on hydrophobic
substrates and 4 µL on hydrophilic ones. The values reported
are averages of at least five measurements made on different
areas of the sample. All of the measurements were performed
at room temperature on a vibration-free platform. (3) Sliding-
Angle Measurements. For sliding-angle measurements on the
hydrophobic samples, water drops of weight ranging from
5 to 40 mg were deposited gently on a horizontal plate fixed
on a goniometer by means of a calibrated micropipet. The
goniometer was rotated slowly until the drops started to slide.
The sliding angle was determined on at least four different
Theoretical Models. (A) Hydrophobic Surfaces. For
vertical structures with a flat top, assuming that the water
does not invade the roughness, then rf) 1 and φS) f. In
this work, φSalways refers to the solid fraction of cylindrical
pillars φS ) πd2/4l2, where d is the base diameter of the
cylinders, and l is their center-to-center pitch. We applied
the Cassie-Baxter and Wenzel formulas to different geom-
etries. (1) The Cassie-Baxter relation was calculated for two
where φBis the ratio of the area of the asperity bases over
the total area. In our case, φB) φS. (2) Wenzel’s relation
was calculated for two cases:
We also calculated the receding angle of composite
drops by assuming that a receding drop leaves a film of
water behind.19In this way, Patankar11derived eq 7 to
predict the receding angles on asperities with a flat top
(rf) 1). We applied eq 7 to cylindrical pillars
In the case of hemispherical asperities, eq 7 becomes
(B) Hydrophilic Surfaces. We applied the composite-drop
(or hemi-wicking) and the Wenzel formulas to different
geometries. (1) The composite-drop relation was calculated
for two cases:
(2) Wenzel’s relation was applied to the case of cylindrical
pillars, as in eq 5.
cos θCB-c) -1 + φS(cos θY+ 1) (3)
cos θCB-h) -1 + φB(cos θY+ 1)2
cos θW-c) r cos θY
cos θW-h) 1 + 4φS(
d- 0.25)cos θY
Figure 2. SEM images of (a) P22 before hydrophobization; (b)
P22 after hydrophobization; (c) profile of P22; (d) sample H83 after
hydrophobization, the insertion shows its profile; (e) P12 after
hydrophobization; (f) profile of P21. The profiles were imaged with
a 90° tilt, the other images were taken at 45°, scale bar ) 500 nm
(a-f) and 200 nm (inset in d), respectively.
Dimensions of the Nanopatternsa
aAll dimensions were measured with ImageJ ((5 nm). In case of hollow
asperities (H), h indicates the depth. In case of pillars (P), h indicates the
maximum distance from the base to the top.
cos θrec-c) 2φS- 1(7)
cos θrec-h) φS(2 + 2 cos θY+ sin2θY) - 1(8)
cos θcomp-c) 1 + φS(cos θY- 1)(9)
cos θcomp-h) φS(2 cos θY+ 3 cos2θY- 1) + 1(10)
Nano Lett., Vol. 5, No. 10, 20052099
Results and Discussion. (A) Sample Characteristics. SEM
images of the patterns immediately after fabrication showed
perfectly cylindrical nanopillars and nanopits, except for
sample P12, where the tall pillars had a cusped top. The
edges of the pillars were rounded after hydrophobization
because of sonication. This is summarized in Figure 2a-b.
In particular, sample P22 has pillars with a hemispherical
top (Figure 2c), and the others look like cylinders with
smooth edges (Figure 2f), except P12 whose shape was
unchanged after coating (Figure 2e). The pitted samples
(Figure 2d) had a cylindrical profile. The different patterns
were named after their solid fraction percentage (H designates
a hollow pattern, i.e., pits, P a protruding pattern, i.e., pillars;
P22 indicates a pillared sample where 22% of the apparent
area is wet by the drop, assuming a cylindrical top). Table
1 shows the dimensions of the patterns (base diameter d,
height h, center-to-center pitch l). The contact angle measured
on the OTS-coated flat silicon was 114 ( 1°, which indicates
that a monolayer was formed (a contact angle for total
coverage was reported to be 115°).26
(B) Hydrophilic Patterns: Experiments and Predictions.
The experimental advancing and receding angles (θadvand
θrec) on hydrophilic silicon are shown in Table 2. The
experimental advancing angles along with the theoretical
predictions are plotted in Figure 3. Because all of the
structures (except P12) were perfectly cylindrical, the Wenzel
and composite-drop curves were plotted for this geometry;
the composite-drop curve for hemispherical-top pillars is also
shown for comparison with the experimental advancing angle
on P12. Young’s angle for these substrates was taken as the
angle measured on the flat control (θY ) 35 ( 1°). The
receding drops on these substrates were pinned (θrec) 0°),
as predicted by Quere.23
Figure 3 shows that the composite-drop formula for
cylindrical asperities is in excellent agreement with the
experimental values, except for sample P12 where θadv )
0°. It should be noted that both the Wenzel formula for
cylinders and the composite-drop formula for hemispherical
tops (eq 10) always predict an apparent angle of 0°, which
is theoretically unattainable because cos θ > 1. Therefore,
the topography-induced superhydrophilicity of sample P12
was due to the high aspect ratio, h/d, of the protrusions,
which acted as a reservoir for the fast spreading of the liquid
front,23a spreading much faster than that on the flat surface.21
Because the composite-drop relation does not depend on the
aspect ratio of the features, this formula predicts that
hemispherical-top pillars will always be superhydrophilic.
In this context, it would be interesting to investigate the effect
of height variation on the reservoir-effect of such structures.
(C) Hydrophobic Patterns: Experiments and Predictions.
The experimental advancing and receding angles (θadvand
θrec) on hydrophobic silicon are shown in Table 2. Young’s
angle for these substrates was taken as the advancing angle
measured on the flat control (θY ) 114° ( 1). The
experimental advancing angles along with the theoretical
predictions are plotted in Figure 4.
Because the end-geometries of the hydrophobized pillars
were somewhat between cylindrical and hemispherical, the
Cassie-Baxter and Wenzel curves were plotted for both
geometries (eqs 3 and 5 in Figure 4a, and eqs 4 and 6 in
Figure 4b), whereas the pits were always modeled as perfect
cylinders (eqs 3 and 5). The sliding angles in Figure 5 were
used as a means to test the predictions derived from Figure
4 concerning the “state” of the drops (i.e., Cassie-Baxter
In the following paragraphs, we discuss our results within
the framework of Patankar’s criterion for designing super-
hydrophobic surfaces.11Briefly, this method consists of
obtaining a stable Cassie-Baxter drop by fabricating struc-
tures, which, given the highest possible aspect ratio, h/d, have
a dimensionless spacing, l/d, such that cos θCB> cos θW, or
Experimental Angles of Water on Hydrophobic (5 µL Drop) and Hydrophilic (4 µL Drop) Nanopatternsa
flat H90 H83P22P21P13 P12
114 ( 1
100 ( 3
125 ( 2
92 ( 2
129 ( 3
89 ( 2
155 ( 2
159 ( 2
140 ( 2
161 ( 2
150 ( 2
164 ( 2
163 ( 2
35 ( 3
36 ( 3
35 ( 3
12 ( 3
19 ( 3
11 ( 3
aAdvancing and receding angles are shown, along with the best theoretical agreement (model): (W) Wenzel, (CB) Cassie-Baxter and (comp) composite-
drop, with (-c) cylindrical and (-h) hemispherical top.
Figure 3. Plot of the apparent advancing angles of a 4-µL drop of
water on hydrophilic patterns as a function of structure geometry.
composite drop (comp-) and Wenzel (W-) curves for cylindrical
(-c-) and hemispherical (-h-) asperities are shown. The Wenzel
curves were plotted for cylindrical asperities only. The composite
drop curves for pillars, P, and pits, H, are different because of a
different dependence on l/d. Individual points indicate experimental
data: (4) H90; (3) H83; (O) P22; (left-facing triangle) P21; (right-
facing triangle) P13; (0) P12.
Nano Lett., Vol. 5, No. 10, 2005
a value of |cos θW- cos θCB| as high as possible; the first
condition ensures that a Cassie-Baxter drop will have lower
energy than a Wenzel drop for the given geometry; the
second condition means that even if cos θCB< cos θW, then
the energy barrier between the two states should be as high
as possible to avoid transitions from a metastable Cassie-
Baxter drop to a lower-energy Wenzel state.
The experimental values of the advancing angles on the
pitted samples (H90 and H83) were in agreement with their
respective Wenzel curves for cylindrical pits (same W-c-H
in both Figure 4a-b). This was supported by sliding-angle
measurements, where 5-10 mg drops were completely
pinned (Figure 5). This result goes against theoretical
expectations because the Cassie-Baxter curve for pits (CB-
c-H) is always at a lower energy than the Wenzel one (W-
c-H), that is, cos θCB-c-H > cos θW-c-H. However, the
Cassie-Baxter regime for pits is stabilized only for l/d <
1.54 (H90) and l/d < 1.48 (H83), which are the critical values
at which the corresponding Wenzel angles become unattain-
able (cos θW< -1); it is therefore possible to obtain a higher
energy configuration with our topography. However, this
predicts the occurrence of stable Cassie drops on pitted
surfaces with a high density of pits of large diameters.
The best agreement with the experimental data for pillars
was found with the models using a hemispherical top (Figure
4b), even when the tops were not perfectly hemispherical
(e.g., as in Figure 2f). This underlines the importance of the
curvature of the edges in determining the advancing angles.
From now on, we will refer to the results plotted in Figure
Samples P22 and P21 had practically the same φSvalues,
but different roughness values, r. Their experimental advanc-
ing angles were close to the intersection between the Cassie-
Baxter curve for pillars and their Wenzel curves. In this case,
the receding angles should allow us to distinguish if the drops
were in either a Wenzel or a Cassie-Baxter state. On
receding, the water droplets were pinned on P22 (θrec) 0°)
because the receding angle never attained a steady state. The
sliding-angle measurements showed that drops of up to 20
mg were pinned on this substrate (upper 90° values in Figure
5). This result suggests that P22 is in a Wenzel state; this
could be explained by the fact that P22 is the shallowest of
the four protruding patterns and it has truly hemispherical
edges: both factors make the Cassie-Baxter regime more
unlikely to happen because sharp edges and a high aspect
ratio are important conditions to ensure air trapping.7O ¨ner
and McCarthy10showed that the receding angles depended
on the three-phase contact line structure, whereas the
advancing angles were unaffected by it. We suggest that the
unusual pinning on sample P22 might be due to the curvature
of the hemispherical tops (see Figure 2c) because the contact
line could be pinned by greater solid-liquid contact and
indeed a hemispherical top will have more solid-liquid
contact than a flat top with the same height (∆φS) 4% for
P22). However, a receding angle of 0° was unexpected, and
we have no explanation for this; between the possible reasons
we cannot exclude a differential hydrophobization of the
Figure 4. Plot of the apparent advancing angles of a 5-µL drop of water on hydrophobic patterns as a function of structure geometry. (a)
Cassie-Baxter (CB-) and Wenzel (W-) curves for cylindrical asperities (-c-). (b) Cassie-Baxter (CB-) and Wenzel (W-) curves for cylindrical
asperities (-c-) and cylinders with hemispherical tops (-h-). The CB and W curves for pits H90 and H83 are the same (W-c-H) in a and b,
whereas the CB and W curves for pillars are different: in a the pillars (P) are modeled as cylinders; in b they are modeled as cylinders with
hemispherical tops. Note that the dependence of the CB curve on l/d for pillars and pits (CB-c-H and CB-c-P) is different. Individual points
indicate experimental data: (4) H90; (3) H83; (O) P22; (left-facing triangle) P21; (right-facing triangle) P13; (0) P12.
Figure 5. Sliding angles of water droplet on hydrophobic patterns
as a function of drop weight. The points at 90° indicate pinned
Nano Lett., Vol. 5, No. 10, 20052101
bottom surface compared to the top of the asperities, which
would have affected the receding angles on the Wenzel
samples; still, only P22 showed this behavior.
Unlike P22, sample P21 had a high receding angle,
showing that it was in a Cassie state. Its closeness to the
critical point would imply that a transition to the Wenzel
state would need very little energy; this hypothesis was
confirmed by the tilting angle measurements, which showed
a sudden jump from 0 to 21° when the drop weight was
increased from 5 to 10 mg. The increased weight forced the
water inside the texture, and the 10-mg drop transitioned to
the Wenzel state.
For sample P13, a Wenzel state would have a lower
energy, but the energy barrier to overcome is bigger than
that in P21. In this case, an increase in drop weight from 5
to 10 mg caused a decrease in sliding angle, confirming the
Cassie-Baxter state of the 5-µL droplet.
Sample P12 had the best water-repellent configuration not
only because the Cassie-Baxter state was at lower energy
than the Wenzel state and the energy barrier between the
two was very high but also because the Wenzel angle
corresponding to this geometry is unattainable (cos θW <
-1). Tilting angles on P12 were 0 ( 2° throughout the whole
range of drop weights, making it impossible to deposit a drop
smaller than 30 µL. This finding is coherent with Patankar’s
prediction that a forest of nanopillars would be the most
effective water-repellent structure,12confirming that for a
given spacing, l, increasing the aspect ratio, h/d, will stabilize
the Cassie-Baxter regime, as was also illustrated by
Yoshimitsu et al.5Sample P12 was superwettable when its
surface was hydrophilic: this similarity not only shows that
roughness with the highest possible aspect ratio, h/d, will
dramatically enhance the wettability of a surface, whether
hydrophilic or hydrophobic, but it also entails the ambiva-
lence of topography-enhanced wetting/dewetting, as high-
lighted already by McHale et al.21
Having verified the validity of the Cassie-Baxter
model for our submicrometric structures, we investigated
Patankar’s12suggestion that the epicuticular wax crystalloids
on the Lotus leaf (200 nm-1 µm) play a significant role in
repelling water droplets. We tested this hypothesis by
applying the Cassie-Baxter formula for hemispherical tops
to our superhydrophobic surfaces (P21, P13, and P12),
assuming an ideal contact angle of cuticular wax θY) 100°:
28the Cassie-Baxter formula predicts apparent contact
angles of 150, 155, and 157°, respectively. This result is only
indicative because the morphology of the wax crystalloids
is very different from those of our model structures.
However, we can confirm Patankar’s hypothesis that rough-
ness at this scale might stabilize or even be the primary cause
of the hydrophobicity of dual-scale topographies. It would
be even more interesting for us to compare our results with
experimental contact angles of insect wings because our
models closely resemble the ordered arrays of nanometric
structures shown by Watson et al.17Unfortunately, this data
is not available at present.
Finally, we verified that the current theoretical predictions
for receding angles agree with our experimental values. The
experimental receding angles of our Cassie-Baxter drops
(P21, P13, and P12) were plotted with eqs 7 and 8 (Figure
6). Our values show a trend similar to the theoretical curves
for cylindrical and hemispherical-ended pillars, but more
work will have to be done in order to obtain better
Conclusions. Regular nanopatterns were fabricated in
silicon wafers, and the behavior of water droplets on these
surfaces was evaluated before and after chemical hydro-
phobization. From our results, we conclude that the composite-
drop (hemi-wicking) model successfully predicts the advanc-
ing angles on hydrophilic patterns. Analogously, the Cassie-
Baxter and Wenzel models gave accurate estimates of the
advancing angles on hydrophobic patterns. These models are
very sensitive to even small variations in the asperity profile,
and we show that if the edges of the cylindrical pillars are
not sharp, a geometry considering a hemispherical-top rather
than a flat one will predict the advancing experimental angles
more accurately; we suggest that the same might be true for
receding angles. In accordance with Patankar’s criterion for
designing a superhydrophobic surface, we confirmed that a
forest of slender pillars is the most stable water-repellent
texture; this same topography exhibited superhydrophilicity,
confirming the ambivalence of topography-enhanced wetting/
dewetting. Finally, our findings support the suggestion that
the epicuticular wax crystalloids of the lotus leaf play a main
role in its water-repellent behavior.
Acknowledgment. Many thanks to Adam S. G. Curtis,
F. Madani, V. Koutsos, L. Csaderova, N. Blondiaux, M.
Robertson, and S. Borzı ´. The partial support by the EC-
funded project NaPa (contract no. NMP4-CT-2003-500120)
is gratefully acknowledged. The content of this work is the
sole responsibility of the authors.
(1) Bico, J.; Thiele, U.; Quere, D. Colloids Surf., A 2002, 206, 41-46.
(2) Cassie, A. B. D.; Baxter, S. Trans. Faraday Soc. 1944, 40, 546.
(3) Wenzel, R. N. Ind. Eng. Chem. 1936, 28, 988-994.
(4) Quere, D.; Lafuma, A.; Bico, J. Nanotechnology 2003, 14, 1109-
(5) Yoshimitsu, Z.; Nakajima, A.; Watanabe, T.; Hashimoto, K. Langmuir
2002, 18, 5818-5822.
Figure 6. Receding angles of water on hydrophobic patterns as a
function of the solid fraction, φS. The experimental values ((f)
P21; (9) P13; (b) P12) are plotted with eq 7 (cyl) and eq 8 (hem).
Nano Lett., Vol. 5, No. 10, 2005
(6) Shibuichi, S.; Onda, T.; Satoh, N.; Tsujii, K. J. Phys. Chem. 1996, Download full-text
(7) Bico, J.; Marzolin, C.; Quere, D. Europhys. Lett. 1999, 47, 220-
(8) Lau, K. K. S.; Bico, J.; Teo, K. B. K.; Chhowalla, M.; Amaratunga,
G. A. J.; Milne, W. I.; McKinley, G. H.; Gleason, K. K. Nano Lett.
2003, 3, 1701-1705.
(9) Miwa, M.; Nakajima, A.; Fujishima, A.; Hashimoto, K.; Watanabe,
T. Langmuir 2000, 16, 5754-5760.
(10) Oner, D.; McCarthy, T. J. Langmuir 2000, 7777-7782.
(11) Patankar, N. A. Langmuir 2003, 19, 1249-1253.
(12) Patankar, N. A. Langmuir 2004, 20, 8209-8213.
(13) Marmur, A. Langmuir 2004, 20, 3517-3519.
(14) Furstner, R.; Barthlott, W.; Neinhuis, C.; Walzel, P. Langmuir 2005,
(15) Barthloot, W.; Neinhuis, C. Planta 1997, 202, 1-8.
(16) Shirtcliffe, N. J.; McHale, G.; Newton, M. I.; Perry, C. C. Langmuir
2005, 21, 937-943.
(17) Watson, G. S.; Watson, J. A. Appl. Surf. Sci. 2004, 235, 139-144.
(18) Kwok, D. Y.; Lin, R.; Neumann, A. W. Colloids Surf., A 1996, 116,
(19) Roura, P.; Fort, J. Langmuir 2002, 18, 566-569.
(20) Bico, J.; Tordeux, C.; Quere, D. Europhys. Lett. 2001, 55, 214-
(21) McHale, G.; Shirtcliffe, N. J.; Aqil, S.; Perry, C. C.; Newton, M. I.
Phys. ReV. Lett. 2004, 93, 036102-1.
(22) Abdelsalam, M. E.; Bartlett, P. N.; Kelf, T.; Baumberg, J. Langmuir
2005, 21, 1753-1757.
(23) Quere, D. Physica A 2002, 313, 32-46.
(24) Burton, Z.; Bhushan, B. Nano Lett. 2005, 5, 1607-1613.
(25) Suh, K. Y.; Jon, S. Langmuir 2005, 21, 6836-6841.
(26) Rozlosnik, N.; Gerstenberg, M. C.; Larsen, N. B. Langmuir 2003,
(27) Rasband, W. S. U.S. National Institutes of Health: Betheda, MD,
(28) Holloway, P. J. Ann. Appl. Biol. 1969, 63, 145-153.
Nano Lett., Vol. 5, No. 10, 2005 2103