Journal of Colloid and Interface Science 372 (2012) 231–238

Contents lists available at SciVerse ScienceDirect

Journal of Colloid and Interface Science

www.elsevier .com/locate / jc is

Fabrication and characterization of hierarchical nanostructured smartadhesion surfaces

Hyungoo Lee, Bharat Bhushan ⇑Nanoprobe Laboratory for Bio- & Nanotechnology and Biomimetics, The Ohio State University, Columbus, OH 43210, United States

a r t i c l e i n f o

Article history:Received 13 December 2011Accepted 7 January 2012Available online 16 January 2012

Keywords:AdhesionSuperhydrophobicLotusGeckoPolypropyleneHierarchical structure

0021-9797/$ - see front matter � 2012 Elsevier Inc. Adoi:10.1016/j.jcis.2012.01.020

⇑ Corresponding author.E-mail address: Bhushan.2@osu.edu (B. Bhushan).

a b s t r a c t

The mechanics of fibrillar adhesive surfaces of biological systems such as a Lotus leaf and a gecko arewidely studied due to their unique surface properties. The Lotus leaf is a model for superhydrophobic sur-faces, self-cleaning properties, and low adhesion. Gecko feet have high adhesion due to the high micro/nanofibrillar hierarchical structures. A nanostructured surface may exhibit low adhesion or high adhesiondepending upon fibrillar density, and it presents the possibility of realizing eco-friendly surface struc-tures with desirable adhesion. The current research, for the first time uses a patterning technique to fab-ricate smart adhesion surfaces: single- and two-level hierarchical synthetic adhesive structure surfaceswith various fibrillar densities and diameters that allows the observation of either the Lotus or geckoadhesion effects. Contact angles of the fabricated structured samples were measured to characterize theirwettability, and contamination experiments were performed to study for self-cleaning ability. A conven-tional and a glass ball attached to an atomic force microscope (AFM) tip were used to obtain the adhesiveforces via force-distance curves to study scale effect. A further increase of the adhesive forces on the sam-ples was achieved by applying an adhesive to the surfaces.

� 2012 Elsevier Inc. All rights reserved.

1. Introduction

Superhydrophobic surfaces exhibit extreme water repellentproperties. Certain plant leaves, notably Lotus leaves, are knownto be superhydrophobic and self-cleaning with low adhesion,known as the Lotus effect [1–5]. These properties are achieved byhaving a hydrophobic surface and a hierarchical structure withboth micro- and nanoscale dimensions. These surfaces are of inter-est in various applications, such as self-cleaning windows anddisplay screens for electronic devices, exterior paints for building,navigation ships, textiles, and solar panels. The leg attachmentpads of several creatures, including many insects, spiders, and liz-ards, are capable of attaching to a variety of surfaces and are usedfor locomotion, known as the gecko effect [5–9]. Geckos, in partic-ular, have the largest mass and have developed the most complexhairy attachment structures capable of smart adhesion, which isthe ability to cling to different smooth and rough surfaces anddetach at will. The high-adhesion mechanism of geckos is basedon so-called division of contacts [6,10]. Cumulative van der Waalsattraction results in strong adhesion. In addition to strong adhe-sion, because of a three level hierarchical structure, they havethe ability to adapt to a variety of surfaces. They exhibit wear resis-tance and self-cleaning. They can also detach via peeling action to

ll rights reserved.

provide reversible adhesion. Attempts are being made to developclimbing robots using gecko inspired structures [11].

Due to the highly optimized and efficient properties of livingnature organism surfaces and their potential applications,researchers have studied their mechanisms and exploited themfor commercial applications [5,9,12,13]. In order to get the Lotuseffect, hierarchical structures using wax structures, nanotubes,and nanoparticles have been fabricated by a large number ofinvestigators; for a review, see Bhushan and Jung [4]. In order toobtain the gecko effect, a high density of nanofibers is required.Hard materials such as carbon nanotubes have been used to fabri-cate gecko-like structures to get high fiber density [7,14]. The hardmaterials provide high resistance to wear and surface contamina-tion. To provide higher adhesion and adaptability to mating sur-faces, soft materials such as polymers have been used tofabricate a one-level fibril structure [15–19]. The fabrication pro-cesses used are complicated and generally provide a low aspectratio of length (height) to diameter of fibers, and the diameter ofthe fibers is generally on the microscale. Murphy et al. [20] hadsuggested a fabrication technique of two-level polymer geckostructures, but the fabricated fibers are on the microscale.

In the present study, two-level polymer fibril structures withhigh aspect ratio are fabricated using stacked porous membranesas a template with nanoscale pores. For the first time, by changingthe densities and diameter of nanofibers, superhydrophobic sur-faces with either Lotus effect or gecko effect have been fabricated.

232 H. Lee, B. Bhushan / Journal of Colloid and Interface Science 372 (2012) 231–238

Contact angle and AFM adhesion measurements have been made.By applying an adhesive to the fiber structure surfaces, an increaseof adhesion on the fiber structures was achieved by combining thegecko effect and a classical adhesive approach.

2. Mechanisms of Lotus and gecko

2.1. Contact angle

The primary parameter that characterizes wetting is the staticcontact angle, which is defined as the angle that a liquid makeswith a solid. The contact angle depends on several factors, suchas surface energy, surface roughness, and its cleanliness [3,21–24]. Surfaces with contact angles in the 0� 6 h 6 90� and90� 6 h 6 180� ranges are hydrophilic and hydrophobic, respec-tively. In particular, surfaces with contact angles between 150�and 180� are called superhydrophobic. Water contact angle hyster-esis (CAH) is another property of interest to reduce drag in fluidflow. CAH occurs due to surface roughness and heterogeneity. Con-tact angle hysteresis is a measure of energy dissipation during theflow of a droplet along a solid surface. A liquid droplet on a solidsurface removes contaminant particles by rolling in addition tosliding at low CAH. Surfaces with low CAH (<10�) are generally re-ferred to as self-cleaning [3]. The contact angle can be determinedby minimizing the net surface free energy of the system of a liquiddroplet on a solid surface. The well-known Young equation for thecontact angle provides an expression for the static contact angle forgiven surface energies. On rough surfaces such as hierarchicalstructures, the contact angle is dependent upon the heterogeneous(composite) interface present. For the heterogeneous interface, thecontact angle is given by the Cassie–Baxter equation:

cos h ¼ Rf cos h0 � fLAðRf cos h0 þ 1Þ ð1Þ

where h is the contact angle on a rough surface, h0 is the contactangle on a smooth surface, fLA is the fractional contact area of theliquid–air, and Rf is the surface roughness factor (>1) equal to theratio of the real interface surface area (ASL) to its geometric interfacearea (AF), Rf = ASL/AF.

An expression for CAH as a function of roughness has beendeveloped. The difference of cosines of the advancing (hadv) andreceding (hrec) angles is related to the difference of those for a nom-inally smooth surface (hadv0 and hrec0) and is given as [3]:

cos hadv � cos hrec ¼ Rf ð1� fLAÞ ðcos hadv0 � cos hrec0Þ þ Hr ð2Þ

where Hr is the effect of surface roughness, which is equal to the to-tal perimeter of the asperity per unit area. Thus, CAH is proportionalto the fraction of the solid–liquid contact area (fSL = 1 � fLA).

2.2. Adhesion

The explanation for the adhesion properties of the Lotus surfaceand gecko feet can be found in the morphology of the Lotus surfacestructure and the skin on the toes of the gecko. On the Lotus sur-face, the papillose epidermal cells form asperities and providemicroscale roughness. A range of waxes made from a mixture oflong chain hydrocarbon compounds which are not easily wettedare usually present on the Lotus surface. A microscale roughnesssurface is covered by sub-microscale asperities of three-dimen-sional epicuticular waxes, creating a hierarchical structure. Thehierarchical structure of the Lotus surface has low adhesion dueto the low density of the sub-microscale asperities. Nanoscaleroughness allows water droplets to sit easily on the apex of thenanostructures, because air pockets occur in the valleys of thestructure under the droplet, resulting in high contact angle and

low CAH. Therefore, the Lotus leaves have low adhesion, superhy-drophobicity, and self-cleaning.

The toe skin of the gecko is also comprised of a complex hierar-chical structure of lamellae, setae, branches, and spatulae [25]. Thedivision of contacts serves as a means for increasing adhesion [10].The surface energy approach can be used to calculate adhesionforce in a dry environment in order to calculate the effect of divisionof contacts. The adhesion force of a single contact Fad, based on theso-called Johnson-Kendall-Roberts (JKR) theory, is given by [26]:

Fad ¼32pWadR ð3Þ

where R is the radius of a spatula hemisphere tip, and Wad is thework of adhesion (unit of energy per unit area). It shows that theadhesion force of a single contact is proportional to the lineardimension of the contact. For a constant area divided into a largenumber of contacts or setae n the radius of a divided contact, R1,is given by [6]:

R1 ¼Rffiffiffinp ð4Þ

Therefore, the total adhesion force (F 0ad) for multiple contacts can begiven by:

F 0ad ¼32pWad

Rffiffiffinp� �

n ¼ffiffiffinp

Fad ð5Þ

Thus, the total adhesive force increases linearly with the square rootof the number of contacts. Based on this analysis, one needs to de-velop structures with a high density of nanofibers. A hierarchicalstructure is needed to provide adaptability to a variety of rough sur-faces [8].

3. Experimental

3.1. Materials and sample preparation

Polycarbonate, (PC) (MILLIPORE, MA) a porous membrane(30 lm thickness) with various pore sizes (50 nm, 100 nm,600 nm, and 5 lm diameter) was used to create structures withdifferent diameters and density of fibers. All samples were fabri-cated using polypropylene (PP) (OC01, Laird Plastics, OH), a ther-moplastic polymer. Its molecular chain is (C3H6)n with a meltingtemperature of 130–171 �C and a Young’s modulus of 1.5–2 GPa[27]. PP has low surface energy (about 30 mJ/m2). Fig. 1 showsthe stacks used to fabricate one- or two-layered structures. Forthe samples with a one-layer structure (Fig. 1a), a 1-mm thick PPfilm and a PC membrane were sandwiched between two poly-dimethylsiloxane (PDMS) disks, and the whole stack was againsandwiched with two aluminum sheets to provide support [28].For fabricating the two-layer structure (Fig. 1b), PP film was placedon top of two PC membranes with different sizes corresponding totwo layers. The stacked layers were placed in an oven at 200 �C for40–50 min with a weight of 1 kg in order to melt the PP and fill thepores in the membrane. For reference, polycarbonate melting tem-perature is 267 �C. After heating in an oven, the sandwiched sam-ples were dipped into methylene chloride for 1 h to etch themembranes to realize the polypropylene fibers.

To fabricate large diameter fiber structures, a micropatterned sil-icon mold with 14 lm diameter, 21 lm pitch, and 30 lm height wasused. A negative replica was created using dental wax [29]. PP wasmelted into the negative mold in the oven at 200 �C for 40–50 min.The melted samples were dipped into methylene chloride or a mix-ture of methyl chloride and chloroform for 1 h as in other cases.

The fiber size and geometry of the fabricated arrays aredescribed in Table 1. Due to the high aspect ratio (length/diame-

Fig. 1. Sample fabrication processes for (a) one-layer structure and (b) two-layer(hierarchical) structure using one and two membranes in the stack, respectively.After heating the stacks in an oven, the membranes are etched using methylenechloride to obtain the nanostructured samples.

H. Lee, B. Bhushan / Journal of Colloid and Interface Science 372 (2012) 231–238 233

ter = 30 lm/50 nm) of the 50 nm diameter random fibers, the50 nm diameter fibers collapse under their own weight and liescattered. To obtain the vertically oriented fibers, larger fiber diam-eters were chosen �100 nm, 600 nm, 5 lm, and 14 lm. The 100and 600 nm diameter fibers were needed to provide high fiber den-sity. Using the structures with the different diameter fibers anddifferent density, the gecko effect and Lotus effect can be realized.The SEM images of each fabricated sample are shown in Fig. 2. Thecontrolled sample (flat surface) of PP was fabricated with no PCmembrane by undergoing all steps used in processing the struc-

tured samples. Fiber density in Table 1 was obtained by measuringthe number of fibers in a unit square area in the SEM images.

3.2. Contact angle measurement

The wetting properties of the samples were characterized bycontact angle measurement. The static contact angle (CA) wasmeasured by placing a DI water droplet of about 5 lL on the sam-ples using a microsyringe. After placement, the droplet was imagedand analyzed using ImageJ (drop analysis, LB-ADSA) software. Con-tact angle hysteresis (CAH) was obtained by measuring the contactangle using a tilting stage with an automated goniometer (290-F4,Rame-Hart Instrument, Succasunna, NJ). The tilting stage was tiltedand the pictures of the DI water droplet on the samples were takenuntil the droplet on the samples rolled out. The advancing andreceding angles of the droplet were measured using ImageJ soft-ware to obtain the CAH. The experiments were carried out at roomtemperature (21 �C) and in 45–55% relative humidity.

3.3. Self-cleaning efficiency measurement

To investigate the self-cleaning ability of the fabricated samplesurfaces, the samples were contaminated with silicon carbide pow-der (37 lm particle size, Aldrich, MO). The samples were placed ona tilted stage. The tilting angle of the stage was 50� and 20� for theflat surface sample and the oriented fiber structured samples(600 nm and 5 lm diameter), respectively. The water droplet sizefor the tests was 10 lL using a microsyringe. While the dropletstraveled down, videos were captured to analyze the movementof the droplets and record the sample cleanliness.

3.4. Adhesion measurement

A commercial AFM (Nanoscope IIIa, Bruker, Santa Barbara, CA)was used for this study [8,30]. Silicon nitride tips of nominal30 nm radius attached to the end of a cantilever beam (DNP, springconstant k of 3 N/m) were used for surface height measurements.The tip of 30 nm nominal diameter was also used to obtain theforce-distance curves for each sample. To investigate the scale ef-fect on adhesion force, a glass ball of 30 lm diameter (dry sodalime glass microspheres, Duke Scientific, Palo Alto, CA) was at-tached to the AFM tip and used to obtain the force distance curves.The experiments were performed at room temperature (21 �C) and45–55% relative humidity. For each force distance curve, therewere 512 sampling points. The force distance curves were collectedand calculated to obtain adhesive forces [9,30].

3.5. Application of adhesive to fibers

To further increase the adhesive forces of the surfaces in casehigh adhesion is needed, an adhesive (Elmer’s repositionable gluesticks) was applied to the sample surfaces. The adhesive materialwas heated on a heating stage for 5 min at 30 �C. The structuredsamples were lightly touched to the heated adhesive, and thendried for 1 h. The adhesive forces for the samples were measuredusing the glass ball tip.

4. Results and discussion

4.1. Contact angle

The wetting properties of the samples were investigated bymeasuring the static contact angle and CAH of a water droplet oneach sample. The contact angle and CAH for each sample is shownin Fig. 3. As a control sample, the CA of the flat surface of PP is 99�,

Table 1Fabricated samples used in the experiments.

Sample Layer Fiber size (app. diameter) Fiber orientation Fiber density (No./lm2) Pitch (lm)

Flat sample None – – – –Random 50 nm One layer 50 nm Random 60 –Oriented 100 nm One layer 100 nm Oriented 20 0.22Oriented 600 nm One layer 600 nm Oriented 0.4 1.41Oriented 5 lm One layer 5 lm Oriented 0.016 7.91Oriented 14 lm One layer 14 lm Oriented 0.0022 21Hierarchical structure (0.6/5 lm) Two layer 0.6 nm/5 lm Oriented 0.13 (600 nm) 2.77 (600 nm)

The height of the fibers for all samples is about 30 lm.

Fig. 2. SEM images for samples. The one layer structure of 50 nm diameter random fibers, 100 and 600 nm oriented fibers, 5 and 20 lm diameter oriented fibers, and the twolayer structure of 600 nm diameter with 5 lm diameter oriented fibers are shown.

Fig. 3. The measured CAs and CAHs on all the one- and two-layered samples.

234 H. Lee, B. Bhushan / Journal of Colloid and Interface Science 372 (2012) 231–238

indicating that PP is hydrophobic itself due to its low surface freeenergy. The CA of a random fiber (50 nm diameter) surface is147�, which is higher than that of the flat sample surface. TheCAH of the flat surface is higher than that of the 50 nm diameterfibers because of the formation of air-pockets under the dropleton the 50 nm diameter fibers.

All oriented fiber samples have a high CA with a range from165� to 167�, and the oriented fibers of 100 nm diameter havethe highest values of 175�. The CAs of the oriented fibers wereaffected by the amount of air pockets (the fractional contact areaof the liquid–air, fLA) on sample surfaces [3,4]. The fractional con-tact area of the liquid–air fLA and Rf for one-layer samples can becalculated using the diameter of the fibers and the fiber densityon each sample, as follows:

fLA ¼ 1� pR21N1 and Rf ¼ 1þ 2pR1hN1 ð6Þ

where R1 is the radius of fibers, N1 is the density of the fibers as gi-ven in Table 1, and h is the height of the fibers. For a two-layer sam-

Table 2Calculated fLA and Rf and the measured contact angles, contact angle hysteresis and adhesion forces (measured using the 30 lm diameter tip) of the samples.

Sample fLA Rf Contact angle (�) Contact angle hysteresis (�) Adhesion force (nN)

Flat sample – – 99 40 259Random 50 nm – – 147 21 160Oriented 100 nm 0.84 189 175 29 348Oriented 600 nm 0.86 29 165 24 291Oriented 5 lm 0.69 9 167 4 157Oriented 14 lm 0.65 4 166 28 98Hierarchical structure (0.6/5 lm) 0.96 16 173 1 326

Fig. 4. Optical images of contaminated surfaces after cleaning with a water droplet.

H. Lee, B. Bhushan / Journal of Colloid and Interface Science 372 (2012) 231–238 235

ple, the calculation for fLA was made using the values of nanofibersin the top level in Eq. (6), and the expression for Rf is given as

Rf ¼ 1þ 2phðR1N1 þ R2N2Þ ð7Þ

where R2 and N2 are radius and density of the fibers on the top level.Table 2 shows the calculated values of fLA and Rf. For the ori-

ented fiber samples of 100 nm, 600 nm, 5 lm, and 14 lm diameter,and the two-layered fibers, fLA is 0.84, 0.86, 0.69, 0.65, and 0.96,respectively. Based on Eq. (1), the CA (h) decreases with an increaseof fLA. Although fLA changes much (from 0.85 to 0.65), CAs of theoriented fiber samples change little (165–167�) due to the changeof Rf on the samples as well. The CAHs of the 100 and 600 nmdiameter fibers are higher than those of the 5 lm diameter fibersdue to the number of fibers contacting the water droplet which re-sults in pinning [4]. In the case of the 14 lm diameter fibers, thedecrease of fLA and Rf results in an increase of CAH.

CA of the two-layer fiber sample is 173�, although the diameterof the fibers contacting the water droplet is 600 nm. It is higher thanthat of the CA of the one-layer 600 nm diameter fiber (165�). This isattributed to the smaller number of fibers contacting the waterdroplet than that of the one-layer fiber sample. In other words, itis due to both an increase of fLA and an increase of Rf, resulting inan increase in CA. Due to the increased fLA and Rf, CAH of the two-layer fibers also decreased, inducing superhydrophobicity.

4.2. Self-cleaning ability

The results of self-cleaning measurement on flat surface and ori-ented fiber (600 nm and 5 lm diameter) surfaces are shown inFig. 4. The dark areas on the images correspond to particles andthe bright areas to droplet tracks. The water droplets cleaned thesurfaces of the flat, the 600 nm diameter fiber, and the 5 lm diam-eter fiber samples. The droplet track on the flat sample is larger thanthat on the other samples, although the droplet size was the samefor all samples. It is attributed to the fact the CA of the flat sampleis 99� but the CAs of the other samples are larger than 165�. Thedroplet tracks were larger than the actual contact area betweenthe droplets and the samples. The particles under the water dropletwere lifted up into the water droplet due to adhesion.

The sliding speed of the droplet on the flat surface was muchslower than that on the fiber structured samples. During slidingof the droplet on flat surface, the particles in the water dropletaccumulated and were present on the contact area between thedroplet and the surface. It means that the particles provided anobstacle in the droplet sliding down. The droplets on the flat sam-ples stopped after sliding about two centimeters due to lots ofaccumulated particles in the droplet, so that the droplet cleanedonly some parts of the surface. However, the contact areas of thedroplet with the fiber structured samples were small. The dirtyparticles in the water droplet did not accumulate and removedon the sample surface. The movement of the droplets was hencenot hindered so that the droplet speed on the fiber structured sam-ples was fast, and the droplets rolled down with dirty particles,resulting in the particles’ removal from the surface.

4.3. Adhesion

Adhesive forces (Fad) of all samples were measured using theAFM. Representative force-distance curves are shown in Fig. 5.The left column is for data obtained using the 30 lm diameter

236 H. Lee, B. Bhushan / Journal of Colloid and Interface Science 372 (2012) 231–238

tip, and the right column is for the 30 nm diameter tip. The bend-ing or buckling behaviors of the fibers were not observed in theforce distance curves on the left column.

In the right column in Fig. 5, bending and buckling behaviors ofa single fiber and two level hierarchical structure are shown whilethe 30 nm diameter tip travels downward. Fig. 6 shows the behav-

Fig. 5. Representative force-distance curves for the samples with the one-layer 100 nm, 6(left column) and the 30 nm diameter tip (right column). Both scales of x and y axes in

ior of the 100 nm diameter single fiber with the AFM tip. The tipapproaches the fiber (point A to B) and the tip touches the fiberat point B. From point B to C, the tip presses the fiber and is de-flected so that the fiber is gradually bent and remains pressed.From point C to D, the fiber buckles. However, the tip keeps press-ing the fiber from D to E, inducing more compression of the fiber.

00 nm, 5 lm diameter fibers, and the two-layer fibers using the 30 lm diameter tipthe right hand column are different.

Fig. 6. The force distance curve for the 100 nm diameter single fiber obtained usingthe AFM tip of 30 nm diameter. Schematics of the fiber behavior at each point as theAFM tip travels are shown.

Fig. 7. Adhesive force on all the one- and two-layered samples measured using the30 lm diameter tip. The values above each bar are the number of fibers in unit area(lm2). The measurements were performed at room temperature (21 �C) and 45–55% relative humidity.

Fig. 8. Adhesive force of the fibers with and without the adhesive applied to thesamples.

H. Lee, B. Bhushan / Journal of Colloid and Interface Science 372 (2012) 231–238 237

From point E to F, even though the tip keeps pressing, the fiber isnot bent or buckled anymore so the tip is deflected. From pointF, the tip retracts upward. Even though the tip retracts, the fiberdoes not stand back up (from point F to G). At point G, the fiberstands. From point H to I, the tip retracts, but the fiber does notmove. From point I to J, the fiber stands up again. This phenomenonrepeats at point K to L, M to N, and O to P.

The measured adhesion forces for each sample using the 30 lmdiameter tip are shown in Fig. 7 and Table 2. The force on the flatsurface is 260 nN, but those on the oriented fibers of 100 and600 nm diameter are 350 and 290 nN, respectively. This is attrib-uted to the number of fibers. The flat surface can be consideredas a fiber with a diameter larger than the real contact area betweenthe AFM ball tip and the flat surface. As shown in Eq. (5), with anincrease in the number of fibers (n) the adhesion forces increasein proportion to the square root of n. In the cases of the 100 and600 nm diameter fibers, the number of fibers is 20 and 0.4 per unitarea, respectively. The 100 and 600 nm diameter fibers hence havehigher adhesion than the flat surface, resulting in the gecko effect.However, the 50 nm diameter fibers are not directional (but ran-domly scattered). This results in a decrease in the adhesive forcedue to the decreased real contact area between the fibers and theAFM ball tip, even though the fiber diameter is smaller than the100 and 600 nm diameter fibers. The oriented 5 and 14 lm diam-eter fibers have low adhesion, obtaining the Lotus effect, due to thesmall number of fibers contacting the AFM ball tip. In addition, asthe AFM ball tip goes deeper, the number of the fibers in contactwith the AFM ball tip increases, resulting in an increase of adhesion

force. Therefore, it is possible that the smaller the fiber diameters,the larger the adhesion forces.

Hierarchical structure (two-layered 600 nm diameter fibers)has an adhesive force slightly larger than the one-layered 600 nmdiameter fibers. The density of the two-layered fibers is lower thanthat of the one-layered fibers (Table 1). The two-layered structureis more compliant which leads to a larger number of fibers comingin contact with the tip.

4.3.1. Application of adhesiveApplication of the adhesive to the samples was performed in or-

der to further increase the adhesive forces by combining the geckoeffect and a classical adhesive approach. The 600 nm and 5 lmdiameter fiber samples and the flat surface sample were chosenfor this experiment to compare the adhesive effect on gecko andLotus structures. The adhesive forces on the samples were mea-sured using the glass ball tip. The results for the adhesive forces be-fore and after application of the adhesive are shown in Fig. 8. Asexpected, the adhesive forces for all the samples increased withadhesive application. However, the trend of the adhesive forces isdifferent before and after the adhesive application. Before theapplication of the adhesive, the oriented 600 nm fiber had thehighest value of adhesive force, but after adhesive application theflat surface had the highest adhesive force, and the oriented5 lm diameter fibers are the second. Before adhesive application,the main force between the ball tip and the sample surfaces were

238 H. Lee, B. Bhushan / Journal of Colloid and Interface Science 372 (2012) 231–238

van der Waals. With the adhesive material, the adhesive forceswere governed by the forces of the adhesive material itself. Itshows that the larger fibers and the flat surface have larger adhe-sive forces because they have a larger amount of adhesive.

5. Conclusions

In this study, one- and two-level polymer fibril structures havebeen fabricated using stacked porous membranes as a templatewith nanoscale pores. The wetting properties, self-cleaning ability,and the adhesive forces on one and two-level fibril structures havebeen investigated. By measuring the CA of the flat sample surface,it was shown that PP structural material is a hydrophobic surface.All one-layered structures were superhydrophobic. The CAs of thesamples were governed by the formation of air pockets on the sur-face (fLA) and high values of roughness factor (Rf). The hierarchicalstructure is also superhydrophobic due to the high fLA and Rf. TheCA of the hierarchical structure was higher, and the CAH was lowerthan those of the one-layer 600 nm diameter fiber structure. All thesamples have self-cleaning ability, but the droplets on the flat sur-faces stopped sliding at a certain point due to the large accumula-tion of particles. The droplets on the fiber structured surfacesrolled down all the way, removing the particles from the surfaces.

The oriented fibers of 100 and 600 nm diameter exhibited thegecko effect (high adhesion) due to their high fiber densities andlarge contact areas. Whereas, the oriented fibers of 5 and 14 lmdiameter exhibited the Lotus effect due to their smaller density(N). From the force distance curves measured using the 30 nmdiameter AFM tip, the bending and buckling behaviors of the fiberswere observed for a single fiber of each sample and two-levelhierarchical structure. An adhesive was applied to the samples,and a further increase of the adhesive forces on the samples wasdemonstrated.

Acknowledgments

We would like to thank Yunlu Pan and Dr. Manuel Palacio forscientific discussions on experiments and the data analysis.

References

[1] W. Barthlott, C. Neinhuis, Planta 202 (1997) 1.[2] C. Neinhuis, W. Barthlott, Ann. Bot. 79 (1997) 667.[3] M. Nosonovsky, B. Bhushan, Multiscale Dissipative Mechanisms and

Hierarchical Surfaces: Friction, Superhydrophobicity, and Biomimetics,Springer, Heidelberg, Germany, 2008.

[4] B. Bhushan, Y.C. Jung, Prog. Mater. Sci. 56 (2011) 1.[5] B. Bhushan, Biomimetics; Bioinspired Hierarchical-Structured Surfaces for

Green Science and Technology, Springer-Verlag, Heidelburg, Germany, 2012.[6] E. Arzt, S. Gorb, R. Spolenak, Proc. Natl. Acad. Sci. USA 100 (2003) 10603.[7] K. Autumn, MRS Bull. 32 (2007) 473.[8] B. Bhushan, J. Adhes. Sci. Technol. 21 (2007) 1213.[9] B. Bhushan (Ed.), Springer Handbook of Nanotechnology, third ed., Springer,

Heidelberg, Germany, 2010.[10] K. Autumn, M. Sitti, Y.A. Liang, A.M. Peattie, W.R. Hansen, S. Sponberg, T.W.

Kenny, R. Fearing, J.N. Israelachvili, R.J. Full, Proc. Natl, Acad. Sci. USA 99 (2002)12252.

[11] M. Cutkosky, S. Kim, Phil. Trans. R. Soc. A 367 (2009) 1799.[12] B. Bhushan, Beilstein J. Nanotechnol. 2 (2011) 66.[13] M. Nosonovsky, B. Bhushan, Philos. Trans. R. Soc. A 368 (2010) 4677.[14] Y. Zhao, T. Tong, L. Delzeit, A. Kashani, M. Meyyappan, A. Majumdar, J. Vac. Sci.

Technol. B 24 (2006) 331.[15] M. Sitti, R.S. Fearing, J. Adhes. Sci. Technol. 18 (2003) 1055.[16] S. Kim, M. Sitti, Appl. Phys. Lett. 89 (2006) 261911.[17] A. del Campo, C. Greiner, E. Arzt, Langmuir 23 (2007) 10235.[18] T.S. Kustandi, V.D. Samper, W.S. Ng, A.S. Chong, H. Gao, J. Micromech.

Microeng. 17 (2007) N75.[19] H.E. Jeong, J.-K. Lee, H.N. Kim, S.H. Moon, K.Y. Suh, Proc. Natl. Acad. Sci. USA

106 (2009) 5639.[20] M.P. Murphy, S. Kim, M. Sitti, ACS Appl. Mater. Interfaces 1 (2009) 849.[21] A.W. Adamson, Physical Chemistry of Surfaces, Wiley, NY, 1990.[22] J.N. Israelachvili, Intermolecular and Surface Forces, second ed., Academic,

London, 1992.[23] B. Bhushan, Principles and Applications of Tribology, Wiley, NY, 1999.[24] B. Bhushan, Introduction to Tribology, Wiley, NY, 2002.[25] R. Ruibal, V. Ernst, J. Morphol. 117 (1965) 271.[26] K.L. Johnson, K. Kendall, A.D. Roberts, Proc. R. Soc. A 324 (1971) 301.[27] C. Maier, T. Calafut, Polypropylene: The Definitive User’s Guide and Databook,

William Andrew, NY, 1998.[28] J. Lee, R.S. Fearing, K. Komvopoulos, Appl. Phys. Lett. 93 (2008) 191910.[29] B. Bhushan, M. Nosonovsky, Y.C. Jung, J.R. Soc, Interface 4 (2007) 643.[30] B. Bhushan (Ed.), Nanotribology and Nanomechanics I – Measurement

Techniques and Nanomechanics, II – Nanotribology, Biomimetics, andIndustrial Applications, third ed., Springer-Verlag, Heidelberg, Germany, 2011.