ARTICLE

pubs.acs.org/Langmuir

Exceptional Superhydrophobicity and Low Velocity Impact

Icephobicity of Acetone-Functionalized Carbon Nanotube Films
Liqiu Zheng,*,† Zhongrui Li,† Shawn Bourdo,† Khedir R. Khedir,† Madhu P. Asar,† Charles C. Ryerson,‡ and
Alexandru S. Biris*,†
†
Nanotechnology Center, Applied Science Department, Chemistry Department, University of Arkansas at Little Rock, Arkansas 72204,
United States
‡
Terrestrial and Cryospheric Sciences Branch, Cold Regions Research & Engineering Laboratory, Engineer Research and
Development Center, U.S. Army Corps of Engineers, Hanover, New Hampshire 03755-1290, United States
bS Supporting Information
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Downloaded via UNIV COLLEGE LONDON on January 17, 2022 at 15:20:51 (UTC).

ABSTRACT: We present a simple method to produce carbon nanotube-based

films with exceptional superhydrophobicity and impact icephobicity by deposit-
ing acetone-treated single-walled carbon nanotubes on glass substrates. This
method is scalable and can be adopted for any substrate, both flexible and rigid.
These films have indicated a high contact angle, in the vicinity of 170°, proved
both by static and dynamic analysis processes. The dynamic evaporation studies
indicated that a droplet deposited on the treated films evaporated in the constant
contact angle mode for more than 80% of the total evaporation time, which is
definitely a characteristic of superhydrophobic surfaces. Furthermore, the
acetone-functionalized films showed a strong ability to mitigate ice accretion from supercooled water droplets (
8 °C), when
the droplets were found to bounce off the films tilted at 30°. The untreated nanotube films did not indicate similar behavior, and the
supercooled water droplets remained attached to the films’ surfaces. Such studies could be the foundation of highly versatile
technologies for both water and ice mitigation.

’ INTRODUCTION with difficulty.10 Second, the need to optimally surface treat a

One of the most economically damaging solid-water interac-
tions at low temperatures is ice formation, which can have an
extremely negative impact on aeronautics, energy generation,
and power distribution networks.1
3 As a result, lately there have
been reported a number of approaches4 to mitigate ice formation
on surfaces that include surface roughness and morphology
control in parallel with altering the surface energy.5 The mor-
phologic control down to the nanoscale holds great promise for
the development of multifunctional surfaces that have both water
and ice phobicity.
It has been reported6 that superhydrophobicity is a primary
prerequisite of icephobicity and that optimizing the surface
conditions of materials in order to reach extremely high contact
angles (CA) close to 170° could result in the nonwetting/nonicing
of surfaces. Very few studies have presented durable hydropho-
bic surfaces with contact angles even close to 170° coupled with
icephobicity because of two major needs that inevitably arise
when designing materials with such requirements: (1) a mor-
phologically rough surface (that has the ability to trap air) and
(2) a top surface coated with low surface energy species that are
chemically stable. Most of the current strategies for fabricating a
textured surface with high roughness characteristics are primarily
based on template synthesis,7 lithography,8 or electrohydro-
dynamics9 and require either very complicated techniques or costly,
sophisticated devices. Even so, hydrophobicity can be maintained
starting material with low surface energy materials in an efficient
and scalable manner is still an issue. Therefore, successfully over-
coming these limitations has become a major challenge. We have
previously presented the development of tungsten nanorods with
pyramidal-shaped tips coated with highly hydrophobic polymers
(silane and Teflon) for water mitigation.11
13 Here, it was clearly
shown that the accurate control of the spacing between these
metallic nanorods can result in a CA varying from 120° to over
160°, holding great promise for possible icephobic applications.
During all of these studies, the surface chemistry remained
similar, but the roughness and the solid fraction were modified
based on the accurate control of the tungsten nanorods’ growth
and morphology.
Lately, the development of various methods for carbon
nanotube (CNT) synthesis, processing, and deposition has made
these unique materials superb 1D structures with major applica-
tions in a number of areas that include: energy generation,14,15
conductive coatings,16,17 sensors,18 and bionanomedicine.19 In
this work, exceptional surface superhydrophobicity (CA ∼ 169°
( 2°) coupled with resistance to impact ice formation was
achieved, with a remarkably straightforward and simple method:
Received: April 27, 2011
Revised: July 1, 2011
Published: July 11, 2011

r 2011 American Chemical Society 9936 dx.doi.org/10.1021/la201548k | Langmuir 2011, 27, 9936–9943
Langmuir
single-walled carbon nanotubes (SWNTs) were treated in acet-
one (99%) under ultrasonication for about 3 h. The resulting
films of these nanotubes were tested for CA and dynamic water
impaction studies at both room and in freezing conditions
(substrate at
8 °C and the water supercooled at
8 °C). It
was found that the superhydrophobic films presented an enhanced
impact icephobicity that was not observed for the nanotube
films that were not acetone treated (unfunctionalized) or for an
aluminum surface, used as a control. A systematic experimental
investigation—including scanning electron microscopy (SEM),
wettability and icephobicity measurement, Fourier transform
infrared (FTIR) spectroscopy, and Raman scattering spectros-
copy—was subsequently conducted to validate the effectiveness
of this approach. Although this study was intended only as a
proof of concept, this approach could be used to create water and
icephobic surfaces. Nevertheless, a number of issues remain to be
addressed, such as adhesion of the films to the subsurface and
their stability in various environments.
’ MATERIALS AND METHODS
CNT Films Fabrication and Characterization. For the control
film samples, 1 g of slightly oxidized SWNT powder (SWeNT Inc.,
SG65) was dispersed in 50 mL of dimethylformamide (DMF). The
appropriate amount of as-obtained solution was air-brushed onto the top
surface of glass slides, which were heated to 150 °C in order to remove
the DMF solvent from the coatings to obtain the untreated control
samples. For the acetone functionalization, we followed the simple
procedure: 0.1 g of SWNT powder was dispersed in 50 mL of acetone
(99%) and kept under sonication for about 3 h. The resulting dispersion
was then air-brushed onto glass slides and allowed to dry under ambient
conditions. The nanotubes formed a network coating onto the top
surface of the glass substrate attached mainly through van der Waals
forces between the two interfaces.20 The acetone treatment was
performed in order to lower the surface energy of the nanotube samples
and to surface functionalize them with
CHx functional groups, which
have one of the lowest surface energy reduction capabilities reported.21
The CH3 groups have been reported to have lower surface energy
compared to the CH2 ones.21 Raman scattering studies of the nanotube
films were performed using a Horiba Jobin Yvon LabRam HR800
equipped with a charge-coupled detector and a He
Ne laser (632.8 nm)
as an excitation source. The laser beam intensity measured at the sample
was kept at 0.5 mW. Raman shifts were calibrated with a silicon wafer at
the peak of 521 cm
1. The Fourier transform infrared (FTIR) spectra
were collected on a Nicolet MAGNA-IR 550 Series 2 spectrometer with
resolution of 8 cm
1. The reported spectra were an average of 32 scans.
The water contact angles of the nanotube films were measured on a
VCA-Optima system. The morphology of the nanotubes was monitored
with a JEOL 7000F high-resolution scanning electron microscope. For
the dynamic impaction behavior a simple water repellence test was
performed. Water droplets were allowed to fall over a 30o tilted substrate
under test, as produced by a stainless steel needle fixed at a constant
height above the surface. The dropping process was video recorded
continuously at high speed. The camera used for high-speed recording
was the model HiSpec-1 from Fastech. This camera provides a good
compromise between form factor, light sensitivity (monochrome ver-
sion, ISO = 3200), and high-speed capabilities. Although it is capable of
very high speeds at lower resolutions, we chose to use the highest frame
rate of 506 frames/s. That still provides the full resolution of 1280 
1024 pixels. The lens chosen for this application was the Nikon
Macro Zoom Nikkor AF (24
85 mm f/2.8
4.0 D) lens.
Icephobicity Analysis. A temperature-controlled environmental
chamber was custom-designed for the hydrophobic impaction studies.
9937
ARTICLE
Figure 1. The morphologies of (a) the untreated and (b) the acetone-
treated SWNT films as analyzed by SEM. The scale bars are 100 nm
in size.
The water droplets were delivered from a 5 cm3 syringe positioned
above the substrate. A special substrate holder was designed that can be
tilted at any angle ranging from 0° and 90° (Figure S1, Supporting
Information). The temperatures of both the substrate and the syringe
were controlled independently through Peltier cooling modules in order
to reach temperatures as low as
10 °C for the water droplets and

20 °C for the substrate. The cooling plates, including the syringe
cooler, were acoustically isolated from vibrations of the heat sink cooling
fans in order to prevent premature nucleation of the supercooled water.
The adjustable impaction height of the droplets was set at 25 mm for
these experiments. The chamber volume can be adjusted by adding or
removing foam insulation as needed. In these experiments, the empty
volume of the chamber was reduced to 5080 cm3 to minimize the
cooling load of the ambient air in the chamber. The relative humidity
(RH) was reduced to less than 20% with the use of a desiccant placed
inside the chamber. It was imperative that the humidity be reduced to
prevent frost formation on the substrate.
’ RESULTS AND DISCUSSIONS
Superhydrophobic Properties of Acetone-Coated Carbon
Nanotube Films. The surface morphology and chemical com-
position essentially govern the wettability of a solid. First, we
investigated the morphology of both samples (control and
acetone-treated films). The acetone ultrasonication treatment
scarcely changed the morphology of the random network of
SWNTs, as shown in the SEM images (Figure 1a,b) of the
untreated and treated SWNT films. The networks were formed
by laying the SWNTs on the substrate with plenty of empty
interspaces. Meanwhile, many bundles of SWNTs were as-
sembled in both cases, typically attributed to van der Waals
forces.22 The diameters of bundles were found to range from
approximately 1 to 70 nm, which provided a large degree of
roughness to the films’ top surfaces (see Figure S2 in Support
Information).
It is worth pointing out that the roughness of the SWNT
network is higher by several orders of magnitude than that of
graphene films, which simply depend on the interconnection
between individual sheets and a few wrinkles at the edge of some
flakes.23,24 Although the treated and untreated samples share
similar morphologies, water droplets behave very differently on
them, suggesting that the chemical composition (controlling the
surface energy) of the surfaces is quite different. Superhydro-
phobicity can be attained only when a textured structure
combines with the proper chemical composition on a surface,
since increased surface energy would result in a more hydrophilic
behavior.25
The wettability properties of the films made from as-produced
(acid-treated and therefore slightly oxidized) and acetone-treated
dx.doi.org/10.1021/la201548k |Langmuir 2011, 27, 9936–9943
Langmuir
Figure 2. The water droplet CAs on (a) the untreated and (b) the
acetone-treated carbon nanotube films.
Figure 3. Dynamic behavior during the evaporation of water droplet
CAs over the acetone-treated carbon nanotube film.
nanotubes were evaluated by water CA measurements. Un-
treated carbon nanotube films demonstrated a relatively hydro-
philic behavior, as seen in Figure 2a. When a water droplet is
deposited on the surface, it instantly disappears, leaving the trace
of a liquid droplet behind. The acetone-treated SWNT film
exhibits superhydrophobic behavior with an extremely high CA
of 169° ( 2° (Figure 2b). The water droplet maintains an almost
spherical shape due to the strong water repellency of the surface.
Ultralow contact-angle hysteresis (defined as the difference between
the advancing and receding CAs) was observed, which means
that a water droplet would easily bounce off.
Further evidence for the low contact angle hysteresis is
provided in Figure 3, which shows the dynamic evaporation beha-
vior of the droplets deposited onto the acetone-functionalized
nanotube films. Both high static water contact angle and low
contact angle hysteresis are the prerequisites for any surface to be
identified as water repellent. Therefore, monitoring the water
droplet dispensed on the substrate as a function of time, diffusion
of water molecules into the ambient environment, and conse-
quent decrease in the size of water droplet would give valuable
information regarding the kinetics of interfacial interaction
between the solid surface and water droplet. Studies have shown
that the kinetics of water droplet evaporation over the solid
surface can be described as a function of both the contact angle
(CA) and the baseline diameter (BLD).26 The decrease in the
size of the water droplet dispensed on a superhydrophobic
surface, due to the evaporation process and the consequent
variation of both CA and BLD, has shown two important stages
of evaporation. The first stage is called the constant baseline
diameter mode (CBLD) with simultaneous decrease in the CA as
evaporation occurs, as the result of the pinning of the baseline
on the patterned surfaces. The second mode is known as the
9938
ARTICLE
constant contact angle (CCA) mode during which the shrinkage
in BLD from both sides allows the water droplet to maintain its
spherical shape.27 The CA (stable CA) at which the transition
from CBLD to the CCA mode occurs is considered the receding
CA. A direct relationship has been proposed between initial CA
and stable CA in the case of evaporation and advancing and
receding CAs in the case of contact angle hysteresis (CAH)
measurements. Therefore, the difference between initial and
stable CAs is equivalent to the difference between advancing
and receding CAs.28 At the very small size of a water droplet
before vanishing, both CA and BLD decrease dramatically until
the droplet vanishes; this is described as the mixed mode.29
After monitoring the evolution of water droplet evaporation
over the treated samples, it was determined that CCA was the
dominant mode of evaporation. As shown in Figure 3, the
evaporation process started with the CCA mode with a decrease
in the BLD accompanying a quasiconstant CA, which is the sign
of low CAH, and the subsequent easy motion of water droplet
BLD over the superhydrophobic surface of CNTs.30 Moreover,
the droplet retained its spherical shape with evaporation evolu-
tion to keep the CA unchanged at the expense of a decrease in
BLD. The mixed mode occurred after the droplet evaporated in
the CCA mode for more than 80% of the total evaporation time.
Interestingly, no transition from CCA mode to CBLD mode was
observed during the entire evaporation process. Therefore, the
analysis of water droplet evaporation and the dynamic BLD
motion over the functionalized carbon nanotube surfaces indi-
cated a very low CAH of less than 3°, whereas the evaporation
started with CCA mode with both initial and stable CAs of around
169° (the observed CA). As mentioned above, this is equivalent
to an advancing CA/receding CA of almost 169°/169°. The high
advancing CA of more than 150° and low CAH of less than 10°
combined with the bouncing phenomenon of water droplets are
intrinsic properties of superhydrophobic surfaces.31 The only
regime (nonsticky regime) with these properties is the Cassie
regime, in which the water bridges over the surface texture, sitting
mostly on air pockets. All of these findings clearly show that the
water droplet has adopted the Cassie regime of wetting for the
acetone-treated nanotube films. The collapse of the water droplet
at the final stage of the evaporation process, as shown in Figure 3,
resulted in a dramatic increase in the BLD. This provides
additional concrete evidence that the droplet had been in the
Cassie regime before switching to the Wenzel regime at the very
end of the evaporation process. At the very end of evaporation
process and due to the very small size of water droplet compar-
able to the size of surface features, the transition would occur
from Cassie regime to Wenzel regime and consequently wetting
the texture by the BLD expansion.
These strong hydrophobic properties can be explained by the
possible chemical surface alterations of the nanotube films since,
as indicated by the microscopy analysis, there were no significant
morphological and/or roughness surface changes between the
acetone-treated and untreated control nanotube films. As pre-
viously shown,24 acetone is strongly chemisorbed onto the
surface of the nanotubes via C(acetone)
O
C(nanotube)
bonds and mostly attaches to the surface defects present in the
graphitic structure of the nanotubes. To explore the chem-
ical composition change that occurs at the nanotube sur-
faces after acetone treatment, the samples were scrutinized by
Raman-scattering spectra. As shown in Figure 4 (left), both
samples demonstrate a peak (G band) at about 1590 cm
1,
which is the characteristic peak of any graphitic-based structure
dx.doi.org/10.1021/la201548k |Langmuir 2011, 27, 9936–9943
Langmuir
Figure 4. The Raman scattering spectra (left) and the FTIR spectra (right) from the untreated and acetone-treated SWNT samples.
due to the in-plane E2g vibrational mode. The multicomponent
spectral feature of the G band is associated with the tangential
displacements of the C
C bond stretching vibrations.23,32 After the
acetone treatment, the intensity of the G band considerably
decreased because the intrabundle interactions, determining the
relative intensity,33 were drastically decreased by the adsorption
of acetone. The D peak positioned at around 1305 cm
1
represents the disorder in the untreated nanotubes, which is
mostly attributable to structural defects. As reported,34 acetone
functionalization causes a large number of (
CHx) groups to be
adsorbed onto the exposed SWNT surfaces and grooves, as well
as possibly accessible interstitial channels of the bundled SWNT
system, thus significantly increasing the intensity of the D band
(1337 cm
1), as shown in the spectra of the treated SWNT.
Notably, the blue shift (32 cm
1) of the D band after the
treatment further suggests that the acetone treatment induced
new types of surface alterations like attachment of
CH2/
CH3
groups rather than morphological defects. This can be explained
because the dangling bonds could be saturated by hydrogen
bonds and structural defects diminished by taking in new
functional groups.35 These changes of the Raman modes sub-
stantiate that chemical modifications of the nanotube surface
take place during the acetone treatment of the nanotubes.
Therefore, Raman spectra analysis revealed that the surface
chemical composition of the treated sample was altered and
could be responsible for the observed superhydrophobic beha-
vior of the nanotube films.
To further corroborate the different surface chemical compo-
sition of samples, FTIR spectroscopy was used to characterize
the species introduced onto the nanotubes’ surface due to
acetone treatment and functionalization. Sonication-induced
changes in temperature and pressure result in chemical interac-
tions of acetone with the SWNTs.36 As seen in Figure 4 (right),
after the acetone treatment, the spectrum manifests absorption
peaks at 2964 and 2922 cm
1 corresponding to unhindered
asymmetric stretching vibration modes of
CH3 and
CH2,
respectively. Symmetric stretching vibration modes of
CH3
and
CH2 are present at 2875 and 2850 cm
1. In addition, four
more absorption peaks were observed at 1368, 1218, 1081, and
838 cm
1, which are twisting
rocking and wagging progression
CH2 modes, consistent with the results of multiwalled carbon
nanotubes treated in acetone.37 All of these (
CHx) groups are
generated through acetone adsorption and functionalization and
can significantly lower the surface energy of the SWNTs films.23
Nevertheless, such absorption peaks are not present in the FTIR
9939
ARTICLE
spectra of graphene films treated with acetone,34 which illustrates
that the SWNTs can be more efficiently functionalized than flat
graphene flakes. The first reason is that the reactivity of carbon
materials with curved geometry is primarily driven by the
significant strain induced by the pyramidalization given by the
curvature of the graphitic surface and the corresponding mis-
alignment in carbon atoms’ π-orbitals. These effects are expected
to make carbon nanotubes more reactive than flat graphene
sheets.38 Second, the relatively large surface areas of SWNTs are
studded with various defects, such as oxygen functionalities,
Stone Wales defects, and vacancies that are essential sites for
acetone to chemically bind to the nanotube surfaces.39
41 Even
the bundled structure of SWNTs induces multiple sites for
acetone adsorption, such as endohedral sites, interstitial sites,
grooves, and external walls.42 Moreover, sonication in organic
solvents induces dangling bonds onto the nanotube surfaces that
provide additional chemical reactivity.43 As the FTIR spectra
indicates, there is an additional
CO absorption peak at
1709 cm
1 appearing in the treated sample, other than the one
at about 1633 cm
1 present in both spectra. Therefore, we can
deduce that this group comes from the fragments of adsorbed
acetone. Basically, there are no
CHx peaks in the spectrum of
the untreated sample other than two extremely low intensity
peaks corresponding to hydrophilic groups
CO (1633 cm
1)
and
OH (not shown) that stem from the oxidative purification
treatment.44 These hydrophilic groups account for the strong
tendency of water to get absorbed on the untreated SWNT film.
Accordingly, the FTIR spectra tenably support the fact that the
different chemical composition of the films resulted in quite
different surface energy values, which induces the different
wettability despite their similar morphologies. Furthermore,
the highest adsorption energy estimated for acetone on SWNTs
is ∼110 kJ mol
1, which implies that adsorption is very strong
and the chemical modification remains stable until high temper-
atures.43 The nanotubes’ stable graphitic walls’ acetone modi-
fication is also reflected by the lack of aging effects of the corre-
sponding hydrophobic characteristic for the treated samples.
After about 6 months, there has been no major degradation,
and the CA has not been found to decrease significantly
(data not shown here).
The hydrophobicity results presented in Figure 2 can be
explained by the Cassie
Baxter model,45 which indicates that
surfaces with structured nanomorphologies and a low surface
energy are needed for enhanced superhydrophobicity. The
effective contact angle (θ*) for a droplet on such a surface is given
dx.doi.org/10.1021/la201548k |Langmuir 2011, 27, 9936–9943
Langmuir
by the following equation:
cos θ ¼ fA cos θA þ fB cos θB ð1Þ
where fA,B are the surface area fractions of components and θA,B
are the Young contact angles of the components [cos θ = (γSV

γSL)/γLV, in which γSV, γSL, γLV are the interface tension of
solid/vapor, solid/liquid, and liquid/vapor, respectively]. If
textured structures on a surface can entrap air pockets, the
effective contact angle is considerably increased since the contact
angle θBof the air component with water is 180°. Then, the
Cassie
Baxter equation takes a simplified form:
cos θ ¼ fA ðcos θA þ 1Þ
1 ð2Þ
Thus, the ability to capture air bubbles in its structures is the first
optimal condition. The perfect configuration of the random
nanotube networks allows the entrapment of air in its inter-
spaces, which exactly meets the first requirement. To further
optimize an effective contact angle θ*, a smaller fA and a larger θA
of a solid component are required according to the simplified
Cassie
Baxter eq 2. The high aspect ratio of carbon nanotubes in
the network drastically decreases the contact area of a droplet
with the nanoprotrusions on the surface and results in a nearly
spherical droplet. In other words, a smaller area fraction, fA, was
obtained, satisfying the second optimal requirement. The third
optimal condition is a larger θA. Since θA is governed by its
surface energy,25 a larger value can be achieved merely by
functionalizing the solid with low surface energy groups. The
high reactivity of carbon nanotubes makes it possible to effi-
ciently functionalize them with low surface energy groups.
Hence, the significantly low surface energy of the network after
coating greatly increases the θA, which fulfills the third optimiza-
tion condition. Superhydrophobicity can therefore be ensured
due to the unique geometry configuration and the chemical com-
position of the treated sample, which provides a precondition for
icephobicity.
From an analysis of the micrographs presented in Figure 1, it is
possible to estimate the surface area fraction, fA to be used in eq 2.
Consequently, from the measured contact angles, θA in eq 2 can
also be estimated. The apparent contact angle on the prepared
surface can be predicted using the Cassie model. To predict the
observed CA (θ*) for the functionalized CNTs with both CH2
and CH3 groups using the Cassie model, the values of both solid
fraction fA of CNTs at the interface and intrinsic CA (θA) of the
smooth surface covered with both CH2 and CH3 groups are
required. Image analysis was used to estimate the solid fraction by
considering the white areas as solid areas and black areas as
voids.13 The determined fA was found to be around 0.1, and the
θA for the smooth surface of CH2 and CH3 groups was estimated
to be around ∼100°.46 After substituting these values in the eq 2,
the predicted value of θ* was found to be on the order of 157°.
The predicted value of θ* is relatively lower than the observed
one, which was around 169°. The primary cause could be
errors in the estimation of the solid fraction, where the overlap
of CNTs caused the image analysis technique to add some
white areas of CNTs from the underlayer areas that have no
contribution at the interface.
Low Velocity Impact Icephobicity of the CNT Surfaces. To
examine the icephobic properties, the samples were tested under
icing conditions for impact adhesion. The environmental condi-
tions included maintaining both the water droplets and the
substrates at below freezing temperatures, in our case
8 °C.
9940
ARTICLE
Figure 5. Superimposed water droplets’ interaction with the acetone-
treated carbon nanotube films at room temperature (A) and while under
icing conditions (B). Icing conditions include maintaining both the
substrate and the droplets at
8 °C. The substrates were tilted at 30°.
The images were obtained from high-speed-camera frames and by
superimposing the same drop at various time frames. The actual high-
speed movies used for this analysis are provided in the Supporting
Information.
The substrate was tilted at 30°, and single droplets of 10 ( 1 μL
sizes were allowed to impact the substrate from a height of
25 mm. The release of the water droplet from this height as a free
fall object will produce an impact velocity of about 0.7 m/s.
The nondimensional Weber number,47 which is the measure of
impact strength (the ratio of the water droplet’s inertia to its
surface energy), for the water droplet of 1.3 mm radius is about
8.8. At this relatively small Weber number, the water droplet
can retain its integrity after impact without fragmentation. The
humidity inside the chamber was kept at a relative humidity of
20%.
Figure 5 shows that the superimposed droplet sequences
indicate that the droplet bounced off from the acetone-treated
carbon nanotube substrate both at room temperature (A) and
under icing conditions (B). The change in the trajectory and kinetic
energy of the droplet between room temperature and icing condi-
tions could suggest an approach for a possible and very accurate
measurement of the adhesion energy at low temperatures for
any surface. Further research will include such studies for films
of nanotubes treated under various conditions.
On the other hand, similar studies performed for untreated
nanotube samples (both room temperature and icing con-
ditions) and aluminum substrates (icing conditions) indicate
that the droplets remained attached to the substrates and froze
shortly after the impact; as a result, they did not bounce off, as
shown in Figure 6A
C. In this work, the icephobicity of the
prepared surfaces was measured only when both the substrate
and the water droplet were at room temperature and at
8 °C.
More studies regarding the icephobicity of superhydrophobic
surfaces with various conditions are still under investigation and
yet to be published. In the current paper, we only wanted to
present the fact that acetone-treated nanotubes have super-
hydrophobic and icephobic properties.
These studies clearly indicate that superhydrophobicity could
be a premise for icephobicity, but more studies need to be
performed in order to fully understand the ice adhesion over
substrates of various surface energy and roughness values.48,49
Since the samples share similar surface morphology, the conclu-
sion can be drawn that the morphology of surfaces, alone, is of
dx.doi.org/10.1021/la201548k |Langmuir 2011, 27, 9936–9943
Langmuir ARTICLE

Figure 6. Superimposed water droplets’ interaction with the untreated (oxidized) carbon nanotube films at room temperature (A), while under icing
conditions (B), and with an aluminum surface used as a control under icing conditions (C). Icing conditions include maintaining both the substrate and
the droplets at
8 °C. The substrates were tilted at 30°. The images were obtained from high-speed-camera frames and by superimposing the same drop
at various time frames. The actual high-speed movies are provided in the Supporting Information. The scale bars are 3 mm in size.

limited use in ice-prevention technologies. The combination of
surface morphology and proper surface chemistry could effi-
ciently prevent droplets from freezing even at supercooled
substrate temperatures. It has been proposed that the ice
solid
surface contact areas are directly associated with the water solid
contact areas at the sample interfaces, which, in turn, determines
the strength of ice adhesion.50 In essence, the low wetting
hysteresis truly reflects the ice repellent property of the surface.
The presence of entrapped air underneath a water surface could
be responsible for slowing down the droplet solidification.5
Furthermore, if the triple phase line (TPL) is unstable, the water
droplet can freeze easily because the heat transfer rate at the TPL
region is much higher than for other conditions, and solidifica-
tion of a droplet is initiated at the TPL instead of the inner part of
the droplet. The different TPL of water droplets on the surfaces is
responsible for the different wetting properties and even icepho-
bicity. The TPL on a rough surface is metastable, which favors
freeze prevention and the solidification of the liquid water at the
TPL region, thus inhibiting ice formation. Therefore, the gen-
eration of rough surfaces with low surface energies is the most
desirable requirement to reduce ice formation and water icing on
impact. As a result, superhydrophobic materials can be designed
to remain entirely ice free down to
8 °C due to their ability to
repel impacting water before ice nucleation occurs.
Recent studies have further shown a strong correlation
between icephobicity and the wettability of the surfaces.49 The
key parameter involved in this relationship is the receding
contact angle of the water droplet over the various surfaces. A
linear relationship has been reported between ice adhesion
strength on superhydrophobic surfaces and their practical work
of adhesion. The systematic study of icephobicity in terms of the
wetting properties of smooth and textured surfaces (features with
various geometries and sizes) has been previously reported
by Mishchenko et al.6,50,51 It has been shown that a continuous,
9941
10-min stream of supercooled water droplets was entirely repelled
by superhydrophobic surfaces at temperatures ranging between
20 and
25 °C, while the water freezing and ice formation
was quite significant over hydrophilic surfaces and relatively
lower over hydrophobic surfaces. However, freezing occurred
for the supercooled water droplets even on superhydrophobic
surfaces when the surface temperature was lowered to less
than
25 °C. In another work on the fabrication of fluoro-
polymer surfaces with various contact angle hysteresis, Kulinich
and Farzaneh6 confirmed the correlation between ice adhesion
strength and CAH. They also demonstrated that the correlation
between ice repellency and CA of the prepared surfaces was only
true for surfaces with low CAH. Therefore, superhydrophobic
surfaces with low CAH would be the most promising candidates
for minimizing or eliminating ice accumulation by causing
supercooled water droplets to bounce off before they can freeze
on the surface.51 However, there has been some controversy on
the limited applicability of superhydrophobic surfaces for ice-
phobic usage in the presence of high humidity and frost.52
55
Our results also suggest that the correlation between cooled
water droplets/ice repellency and CA of the prepared surfaces is
only true for surfaces with low CAH. Therefore, superhydro-
phobic surfaces with low CAH would be the most promising
candidate to minimize or eliminate the ice accumulation by
bouncing off the impinged supercooled water droplets before
they can freeze on the surface.
’ CONCLUSIONS
The functionalization by acetone of single-walled carbon
nanotubes with low surface-energy groups and the generation
of continuous nanotube films by air-spraying has proven to be a
promising method to mass produce long-lasting superhydropho-
bic coatings (CA close to 170°) at relatively low cost. Reduced
dx.doi.org/10.1021/la201548k |Langmuir 2011, 27, 9936–9943
Langmuir
surface contact time and contact area of the impinging water
droplets on properly designed surfaces can leave the surface ice-
free, as the droplets were found to bounce off during ice condi-
tions before ice nucleation could occur. This simple method
owes its effectiveness to the unique configuration of the nano-
tube random networks, as well as the stable and efficient chemical
modification of the graphitic nanotubes in the presence of
acetone and under sonication. Acetone-treated carbon nanotube
films have both the required roughness and low surface area to
present water repellency at both room temperature and in icing
conditions. Such approaches could become the foundation for
ice-fighting technologies with important implications for the
economic aspects of various industrial fields, including aeronau-
tics, power generation and distribution, and defense.
’ ASSOCIATED CONTENT
bS Supporting Information. Additional information as noted
in the text. This material is available free of charge via the Internet at
http://pubs.acs.org/.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: lxzheng@ualr.edu; asbiris@ualr.edu.
’ ACKNOWLEDGMENT
This research was partially supported by the DOD (Grant No.
W912HZ-09-2-0008). Also the financial support of the Arkansas
Science and Technology Authority (ASTA) grant # 08-CAT-03
is highly appreciated. The editorial assistance of Dr. Marinelle
Ringer is also acknowledged. We also acknowledge the work
done by Mr. Harry L. Maddox in creating the composite drop
sequence images along with image enhancements from the
individual JPEG frames that were derived from the high-speed
videos.
’ REFERENCES
(1) Jafari, R.; Menini, R.; Farzaneh, M. Appl. Surf. Sci. 2010, 257,
1540–1543.
(2) Menini, R.; Farzaneh, M. Surf. Coat. Technol. 2009, 203, 1941–1946.
(3) Ryerson, C. C. Cold Reg. Sci. Technol. 2011, 65, 97–110.
(4) Farzaneh, M.; Ryerson, C. C. Cold Reg. Sci. Technol. 2011,
65, 1–4.
(5) He, M.; Wang, J.; Li, H.; Jin, X.; Wang, J.; Liu, B.; Song, Y. Soft
Matt. 2010, 6, 2396–2399.
(6) Kulinich, S. A.; Farzaneh, M. Langmuir 2009, 25, 8854–8856.
(7) Zhang, G.; Wang, D.; Gu, Z.; Mchwald, H. Langmuir 2005,
21, 9143–9148.
(8) Liu, B.; He, Y.; Fan, Y.; Wang, X. Macromol. Rapid Commun.
2006, 27, 1859–1864.
(9) Jiang, L.; Zhao, Y.; Zhai, J. Angew. Chem., Int. Ed. 2004, 43,
4338–4341.
(10) Onda, T.; Shbuichi, S.; Satoh, N.; Tsujii, K. Langmuir 1996,
12, 2125–2127.
(11) Khedir, K. R.; Kannarpady, G. K.; Ishihara, H.; Woo, J.;
Ryerson, C.; Biris, A. S. Phys. Lett. A 2010, 374, 4430–4437.
(12) Kannarpady, G.; Sharmaa, R.; Liu, B.; Trigwell, S.; Ryerson, C.;
Biris, A. S. Appl. Surf. Sci. 2010, 256, 1679–1682.
(13) Khedir, K. R.; Kannarpady, G. K.; Ishihara, H.; Woo, J.;
Ryerson, C.; Biris, A. S. Langmuir 2011, 27, 4661–4668.
(14) Wang, S.; Khafizov, M.; Tu, X.; Zheng, M.; Krauss, T. D. Nano
Lett. 2010, 10, 2381–2386.
9942
ARTICLE
(15) Li, Z.; Kunets, V. P.; Saini, V.; Xu, Y.; Dervishi, E.; Salamo, G. J.;
Biris, A. R.; Biris, A. S. ACS Nano 2009, 3, 1407–1414.
(16) Sethi, S.; Dhinojwala, A. Langmuir 2009, 25, 4311–4313.
(17) Li, Z.; Kandel, H. R.; Dervishi, E.; Saini, V.; Biris, A. S.; Biris,
A. R.; Lupu, D. Appl. Phys. Lett. 2007, 91, 053115.
(18) Sinha, M.; Ma, J.; Yeow, J. T. W. Carbon Nanotube-Based
Sensors. J. Nanosci. Nanotechnol. 2006, 6, 573–90.
(19) Biris, A. S.; Galanzha, E. I.; Li, Z.; Mahmood, M.; Xu, Y.;
Zharov, V. P. J. Biomed. Opt. 2009, 14, 021007–021011.
(20) Zhao, Y.; Tong, T.; Delzeit, L.; Kashani, A.; Meyyappan, M.;
Majumdar, A. J. Vac. Sci. Technol. B 2006, 24, 331.
(21) Hare, E. F.; Shafrin, E. G.; Zisman, W. A. J. Phys. Chem. 1954,
58, 236–239.
(22) Thess, A.; Lee, R.; Nikolaev, P.; Dai, H.; Petit, P.; Robert, J.; Xu,
C.; Young, H. L.; Seong, G. K.; Rinzler, A. G.; Colbert, D. T.; Scuseria,
G. E.; Tombnek, D.; Fischer, J. E.; Smalley, R. E. Science 1996, 273,
483–487.
(23) Zheng, L. Q.; Li, Z. R.; Bourdo, S.; Watanabe, F.; Ryerson,
C. C.; Biris, A. S. Chem. Commun. 2011, 47, 1213–1215.
(24) Rafiee, J.; Rafiee, M. A.; Yu, Z.; Koratkar, N. Adv. Mater. 2010,
22, 2151–2154.
(25) Quere, D. Physica A 2002, 313, 32–46.
(26) McHale, G.; Aqil, S.; Shirtcliffe, N. J.; Newton, M. I.; Erbil, H. Y.
Langmuir 2005, 21, 11053–11060.
(27) Anantharaju, N.; Panchagnula, M.; Neti, S. J. Colloid Interface
Sci. 2009, 337, 176–182.
(28) Soolaman, D. M.; Yu, H. Z. J. Phys. Chem. B 2005, 109,
17967–17973.
(29) Kim, J.; Ahn, S. I.; Kim, J. H.; Zin, W. Langmuir 2007, 23,
6163–6169.
(30) Kulinich, S. A.; Farzaneh, M. Appl. Surf. Sci. 2009, 255,
4056–4060.
(31) Nosonovsky, M.; Bhushan, B. J. Phys.: Condens. Matt. 2008, 20,
395005–395011.
(32) Kukovecz, A.; Kramberger, C.; Georgakilas, V.; Prato, M.;
Kuzmany, H. Eur. Phys. J. B 2002, 28, 223–230.
(33) Jiang, C.; Kempa, K.; Zhao, J.; Schlecht, U.; Kolb, U.; Basche,
T.; Burghard, M.; Mews Phys. Rev. B 2002, 66, 161404–161407.
(34) Ellis, A. V.; Vijayamohanan, K.; Goswami, R.; Chakrapani, N.;
Ramanathan, L. S.; Ajayan, P. M.; Ramanath, G. Nano Lett. 2003, 3,
279–282.
(35) Zhang, G.; Qi, P.; Wang, X.; Lu, Y.; Mann, D.; Li, X.; Dai, H.
J. Am. Chem. Soc. 2006, 128, 6026–6027.
(36) Yang, C. M.; Kaneko, K.; Yudasaka, M.; Iijima, S. Phys. B:
Condens. Matt. 2002, 323, 140–142.
(37) Ayala, D. A.; Romero, J. M.; Jorge, N. L.; Gomez-Vara, M.;
Jubert, A. H.; Castro, E. A. Spectrochim. Acta, Part A 2006, 64,
717–721.
(38) Niyogi, S.; Hamon, M. A.; Hu, H.; Zhao, B.; Bhowmik, P.; Sen,
R.; Itkis, M. E.; Haddon, R. C. Acc. Chem. Res. 2002, 35, 1105–1113.
(39) Pak, C.; Lee, H.; Kim, J.; Kim, D.; Kim, K. Struct. Chem. 2005,
16, 187–202.
(40) Robinson, J. A.; Snow, E. S.; Badescu, S. C.; Reinecke, T. L.;
Perkins, F. K. Nano Lett. 2006, 6, 1747–1751.
(41) Chakrapani, N.; Zhang, Y. M.; Nayak, S. K.; Moore, J. A.;
Carroll, D. L.; Choi, Y. Y.; Ajayan, P. M. J. Phys. Chem. B 2003, 107,
9308–9311.
(42) Kazachkin, D.; Nishimura, Y.; Irle, S.; Morokuma, K.; Vidic,
R. D.; Borguet, E. Langmuir 2008, 24, 7848–7856.
(43) Koshio, A.; Yudasaka, M.; Zhang, M.; Iijima, S. Nano Lett. 2001,
1, 361–363.
(44) Yudasaja, M.; Zhang, M.; Jabs, C.; Iijima, S. Appl. Phys. A: Mater.
Sci. Process. 2000, 71, 449–451.
(45) Cassie, A. B. D.; Baxter, S. Trans. Faraday Soc. 1944, 40, 546–551.
(46) Kutasi, K.; Bibinov, N.; Keudell, A. V.; Wiesemann, K. J.
Optoelectron. Adv. Mater. 2005, 7, 2549–2556.
(47) Tsai, P.; Pacheco, S.; Pirat, C.; Lefferts, L.; Lohse, D. Langmuir
2009, 25, 12293–12298.
dx.doi.org/10.1021/la201548k |Langmuir 2011, 27, 9936–9943
Langmuir ARTICLE

(48) Meuler, A. J.; Smith, J. D.; Varanasi, K. K.; Mabry, J. M.;

McKinley, G. H.; Cohen, R. E. ACS Appl. Mater. Interfaces 2010, 2,
3100–3110.
(49) Dotan, A.; Dodiuk, H.; Laforte, C.; Kenig, S. J. Adhes. Sci.
Technol. 2009, 23, 1907–1915.
(50) Meuler, A. J.; McKinley, G. H.; Cohen, R. E. ACS Nano 2010, 4,
7048–7052.
(51) Mishchenko, L.; Hatton, B.; Bahadur, V.; Taylor, J. A.;
Krupenkin, T.; Aizenberg, J. ACS Nano 2010, 4, 7699–7707.
(52) Varanasi, K. K.; Deng, T.; Smith, J. D.; Hsu, M.; Bhate, N. Appl.
Phys. Lett. 2010, 97, 234102–234105.
(53) Farhadi, S.; Farzaneh, M.; Kulinich, S. A. Appl. Surf. Sci. 2011,
257, 6264–6269.
(54) Kulinich, S. A.; Farhadi, S.; Nose, K.; Du, X. W. Langmuir 2011,
27, 25–29.
(55) Jung, S.; Dorrestijn, M.; Raps, D.; Das, A.; Megaridis, C. M.;
Poulikakos, D. Langmuir 2011, 27, 3059–3066.

9943 dx.doi.org/10.1021/la201548k |Langmuir 2011, 27, 9936–9943