Journal of Colloid and Interface Science 326 (2008) 465–470

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier.com/locate/jcis

UV and thermally stable superhydrophobic coatings from sol–gel processing

Yonghao Xiu a,b, Dennis W. Hess a,∗, C.P. Wong b,∗a School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive, Atlanta, GA 30332-0100, USAb School of Materials Science and Engineering, Georgia Institute of Technology, 771 Ferst Drive, Atlanta, GA 30332-0245, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 11 April 2008Accepted 15 June 2008Available online 28 June 2008

Keywords:SilicaThin filmSuperhydrophobicityFluoroalkyl silaneUV stabilityThermal stabilityAccelerated UV weathering test

A method for the preparation of inorganic superhydrophobic silica coatings using sol–gel processingwith tetramethoxysilane and isobutyltrimethoxysilane as precursors is described. Incorporation ofisobutyltrimethoxysilane into silica layers resulted in the existence of hydrophobic isobutyl surfacegroups, thereby generating surface hydrophobicity. When combined with the surface roughness thatresulted from sol–gel processing, a superhydrophobic surface was achieved. This surface showed im-proved UV and thermal stability compared to superhydrophobic surfaces generated from polybutadieneby plasma etching. Under prolonged UV tests (ASTM D 4329), these surfaces gradually lost superhydro-phobic character. However, when the as-prepared superhydrophobic surface was treated at 500 ◦C toremove the organic moieties and covered with a fluoroalkyl layer by a perfluorooctylsilane treatment,the surface regained superhydrophobicity. The UV and thermal stability of these surfaces was maintainedupon exposure to temperatures up to 400 ◦C and UV testing times of 5500 h. Contact angles remained>160◦ with contact angle hysteresis ∼2◦.

© 2008 Elsevier Inc. All rights reserved.

1. Introduction

The water repellent properties of many biological surfaces, es-pecially plant leaves, have prompted great research interest since,when water drops fall on certain leaf surfaces, they can easily pickup contamination, roll off the surface, and thereby clean the leaves.Because the first observation of this self-cleaning phenomenon wason lotus leaves, the effect is generally termed “lotus effect.” De-tailed inspection of the leaves showed that micron and submicronstructures as well as hydrophobic materials exist on the leaf sur-face [1]. The combination of these factors leads to superhydropho-bicity. Wenzel [2,3] showed that surface roughness produces aphysical enhancement of hydrophobicity which was described by

cos θA = r cos θY . (1)

In Eq. (1), θA is apparent contact angle, θY is the contact angleon a flat surface and r is the ratio of the actual to the projectedarea. This expression predicts that the wetting behavior of a sur-face is enhanced by roughness; thus, creation of roughness on aflat surface with an equilibrium contact angle θY (flat) > 90◦ in-creases the contact angle, while the same roughness on a surfacewith θY (flat) < 90◦ decreases the contact angle. In practice, inti-mate contact is not usually maintained between liquid and solid on

* Corresponding authors. Faxes: +1 404 894 2866 (D.W. Hess), +1 404 894 9140(C.P. Wong).

E-mail addresses: dennis.hess@chbe.gatech.edu (D.W. Hess),cp.wong@mse.gatech.edu (C.P. Wong).

0021-9797/$ – see front matter © 2008 Elsevier Inc. All rights reserved.doi:10.1016/j.jcis.2008.06.042

very rough surfaces with θY (flat) > 90◦; rather, the droplet effec-tively resides on a composite surface. In this case Cassie’s equationapplies [4].

cos θA = f1 cos θY − f2, (2)

where f1 and f2 are the fractions of the surface occupied by solid/liquid and air/liquid, and f1 + f2 = 1.

Self-cleaning surfaces have many potential applications, forexample, surface decontamination on microelectronic equipmentor devices, water repellent coatings, biocompatible surfaces, andfriction reduction. Such surfaces have been created on a vari-ety of materials such as silicones [5], alumina [6–8], silicon [9],silica [10–14], polyelectrolytes [15,16], carbon nanotubes [17],polystyrene [18], PTFE (polytetrafluoroethylene) [19] and fluorocar-bons [20], and polybutadiene [21]. Several criteria have been usedto characterize these surfaces including contact angle (>150◦),contact angle hysteresis (<10◦) and roll off angle. However, limita-tions exist when polymeric materials are exposed to atmosphericconditions where degradation in superhydrophobic properties oc-curs due to UV irradiation, and to exposure to impurities, O2, andmoisture present in the environment. Also, because organic poly-mers are easily deformed when a force is applied, the surfacestructure is not stable when the surface undergoes abrasion fromfriction, handling, or dust particles.

Sol–gel processes offer a method to fabricate porous glass filmsunder ambient conditions. The mild preparation conditions offerthe ability to incorporate a wide range of labile organic speciesinto a glass composite. In addition to the mild processing condi-

466 Y. Xiu et al. / Journal of Colloid and Interface Science 326 (2008) 465–470

tions, sol–gel derived materials exhibit tunable porosity [14,22–24],transparency [25], hardness, and good thermal stability. However,UV stability has not been addressed extensively, especially for longtimes, which is a requirement for outdoor applications. Such appli-cations are especially problematic, since the surface hydrophobicityimparted is usually due to the presence of hydrocarbons [10,14].In the current paper, porous silica generated by sol–gel processingfollowed by surface fluoroalkyl substitution is invoked to improvethe thermal and long term UV stability of superhydrophobic films.This approach to the fabrication of inorganic superhydrophobic sil-ica films uses tetramethoxysilane and isobutyltrimethoxysilane asprecursors. When isobutyltrimethoxysilane is incorporated into sil-ica, the surface contains hydrophobic isobutyl groups, resulting inthe ability to fine tune hydrophobicity and adhesion. Furthermore,this method does not use supercritical drying which is frequentlyemployed for the preparation of superhydrophobic materials [10].The porosity and thus surface structure can be controlled by pa-rameters such as reaction conditions, solvent evaporation rate andprecursor ratio. In addition, due to the fact that C–F bonds arestronger than C–H bonds, it is expected that the incorporation offluoroalkyl groups will impart improved UV stability to the coat-ing and thereby make possible a variety of outdoor applications.Therefore, by designing both surface hydrophobicity and surfacestructure into the film, it should be possible to prepare an inor-ganic coating with stable superhydrophobicity.

2. Experimental details

2.1. Bulk silica formation

The initial solution was prepared using a two-step (acid/basecatalyzed) procedure. Tetramethoxysilane (TMOS), isobutyltrimeth-oxysilane (IBTMOS), ethanol, and water were mixed in the volu-metric ratio of 1:1:2:1 in a 30 ml vial. The pH was adjusted to1.5–1.8 by adding hydrochloric acid, and the mixture stirred for120 min at 60 ◦C to form silica sol. Ammonium hydroxide (NH4OH)was subsequently added to the sol, followed by mixing for 5 minat room temperature. The resulting gel was allowed to age for1 week. Slow evaporation of ethanol and water was permittedthrough a small slit (6 mm × 3 mm) in the vial cap, thereby form-ing bulk silica within the vial.

2.2. Silica film formation by casting

Glass slides were used as substrates for the casting of silicafilms. The slides were cleaned in Piranha solution (70:30 (vol/vol)mixture of 96% sulfuric acid and 30% aqueous hydrogen peroxide)and subsequently rinsed extensively with de-ionized water andethanol. Film casting was performed by invoking two glass slides.One slide served as the substrate, and the other, which was treatedwith trimethylchlorosilane (TMCS) to prevent gel from bonding toit, served as the cover. A chromium film of ∼300 nm thicknesswas sputtered onto the surface of the substrate glass slide priorto gel formation. Bulk silica sol was prepared according the pro-cedure described above; after NH4OH was added to the sol, thesolution was stirred for 3 min. Several drops of the sol were thendispensed onto the substrate glass slide and the sol covered withthe TMCS treated glass slide. After gelation, aging and evaporationof the solvent, a silica film resulted.

2.3. Preparation of superhydrophobic polybutadiene films

A polybutadiene/toluene solution (5 wt% with 2% UV stabilizerTINUVIN 1577FF from Ciba and 1.5% antioxidant IRGANOX 1010FFfrom Ciba) was dispensed onto a silicon substrate at a spin speed

Fig. 1. Morphology of a superhydrophobic polybutadiene surface formed by an SF6

plasma treatment.

of 2000 rpm to cast polybutadiene films (∼10 μm). The film sur-face was subsequently exposed to an SF6 plasma at 150 mTorr and150 W for 10 min to form a superhydrophobic fluorinated polybu-tadiene surface as shown in Fig. 1.

2.4. Characterization

Contact angle measurements, scanning electron microscopy(SEM), atomic force microscopy (AFM), thermogravimetric analysis(TGA), and Fourier transform infrared (FTIR) were used to char-acterize the superhydrophobic surfaces. Water droplets of 1 mmdiameter were placed onto surfaces using a microsyringe and staticimages were recorded from which the contact angle of the dropleton the solid surface was determined. For SEM inspection, sampleswere fixed to aluminum stubs with conductive tape prior to coat-ing with ∼20 nm of gold in an Ernest Fullam sputter coater. Thesamples were then imaged using a LEO 1530 SEM. An ExplorerAFM from Veeco was invoked for roughness determination; mea-surements were performed in contact mode. A cantilever with anominal spring constant of 0.1 N/m was used with minimum force.For TGA analysis, a TGA 2050 thermogravimetric analyzer from TAInstruments was used. An approximately 10 mg silica sample wasplaced in a ceramic pan, and the pan positioned in the TGA fur-nace; a temperature ramp was applied from room temperature to800 ◦C with an air flow of 100 SCCM. FTIR spectra were obtainedusing a Nicolet Magna IR 560 with an ATR accessory. Samples werestored in a drybox to prevent moisture absorption. The spectrom-eter was flushed with dry air continuously before and during FTIRmeasurements.

2.5. Accelerated UV weathering test

UV tests were performed by adhering to standard tests for thefluorescent UV exposure of plastics (ASTM D 4329). A UVA-340fluorescent lamp was used to simulate the short and middle UVwavelength region corresponding to daylight exposure. The testcycle was 8 h UV with the uninsulated black panel temperaturecontrolled at 60 ± 3 ◦C, followed by 4 h condensation with theuninsulated black panel temperature controlled at 50 ± 3 ◦C. A ra-diometer was used to monitor and control the amount of radiantenergy that impinged on the sample. A thermometer was used tomonitor the temperature in the test chamber.

3. Results and discussion

Solvent evaporation of the solution-cast coating results in filmsurface roughness. The presence of hydrophobic chemical groupson the silica surface as illustrated schematically in Fig. 2, alongwith the surface roughness generated from solvent evaporation,

Y. Xiu et al. / Journal of Colloid and Interface Science 326 (2008) 465–470 467

Fig. 2. Illustration of hydrophobic groups on a treated silica surface.

Table 1Physical characteristics of solutions resulting from different sol formulations

TMOS/IBTMOS/ethanol/H2O(volume ratio)

Phase separationduring hydrolysis

Transparencyduring gelation

Precipitationafter gelation

3:1:4:2 No Yes No1:1:2:1 No No No1:3:4:2 Yes No Yes

yields an enhanced surface contact angle. As described below, a su-perhydrophobic silica film can be prepared from sol–gel processingusing TMOS and IBTMOS.

3.1. Acid–base catalyzed sol–gel process

Table 1 indicates that chemical reactions can occur that re-sult in changes in the physical solution characteristics during thebulk silica formation process. When the concentration of IBTMOSwas increased, e.g., TMOS/IBTMOS ratio of 1:3, phase separationoccurred during hydrolysis as evidenced by the fact that the so-lution became cloudy (white) upon stirring. After the addition ofammonium hydroxide and equilibration, a separate viscous phaseappeared at the bottom of the vial. After removal of the upper liq-uid phase, the viscous solution remaining in the vial underwentslow gelation and bulk silica formed. For a TMOS/IBTMOS ratio(v/v) of 1:1, no phase separation occurred; even after 120 minof reaction, the sol was still transparent. After gelation by addi-tion of the ammonium hydroxide solution (1.1 M), the sol became

Table 2Contact angles on surfaces resulting from various formulations of bulk silica

TMOS(ml)

IBTMOS(ml)

Ethanol(ml)

H2O(ml)

1 M HCl(ml)

NH4OH(ml)

Contactangle (◦)

1 3 4 2 0.03 0.5 160.82 2 4 1.9 0.1a 1 156.13 1 4 2 0.03 1 75.3

a HCl concentration is 0.1392 M.

turbid/white and the solution gelled to form bulk silica. The timeneeded for this to occur depended on the amount of ammoniumsolution added; the more ammonium, the shorter the time to de-velop a turbid sol. During the controlled evaporation of solvent,the volume of silica was gradually reduced. Table 2 indicates thesol–gel process conditions and the final contact angles on the as-prepared silica surfaces.

The creation of superhydrophobic surfaces requires controlof the surface roughness. For surfaces formed from mixturesof TMOS/IBTMOS with a ratio of 1:3, superhydrophobicity wasachieved with large pore sizes of ∼10 μm as shown in Fig. 3. Inaddition to the pores generated by evaporation of solvent, the par-ticle size also contributes to the surface roughness as shown inFig. 3d. Surfaces from mixtures of TMOS/IBTMOS with a ratio of 1:1still show superhydrophobicity although the isobutyl trimethoxysi-lane content is reduced. Furthermore, the pores are not sphericaldue to the increase of solvent solubility in the silica and the poresize is much reduced. Surfaces from mixtures of TMOS/IBTMOSwith a ratio of 3:1 did not display superhydrophobicity because ofthe lower ratio of IBTMOS used; that is, the surface is generallyflat with insufficient voids, and the contact angle is only 75.3◦ .

After fracture of bulk silica formed from a 1:1 mixture ofTMOS/IBTMOS, the fractured surface still displayed superhydropho-bicity because the surface remained rough; the water contact angleon this fractured surface was also above 150◦ .

3.2. Creation of superhydrophobic thin films by sol–gel processing

In addition to bulk silica, cast silica thin films were also pre-pared. Water droplet contact angles on different thin film surfaces

Fig. 3. SEM images of surfaces of as-synthesized bulk silica. (a) TMOS/IBTMOS = 1:3, 1 M hydrochloric acid: 0.03 ml, 1.1 M ammonia: 0.5 ml; (b) TMOS/IBTMOS = 1:1,0.1392 M hydrochloric acid: 0.1 ml, 1.1 M ammonia: 1 ml; (c) TMOS/IBTMOS = 3:1, 1 M hydrochloric acid: 0.03 ml, 1.1 M ammonia: 1 ml; (d) high resolution SEM image ofthe structures on a superhydrophobic silica surface from a mixture of TMOS/IBTMOS = 1:3. Inset: images of the shapes of water droplets in contact with the prepared silicasurfaces.

468 Y. Xiu et al. / Journal of Colloid and Interface Science 326 (2008) 465–470

Table 3Thin silica films cast from different ratios of TMOS/IBTMOS

TMOS/IBTMOS(v/v)

Contact angle(◦)

Contact anglehysteresis (◦)

1:3 101.0 Not measurable1:1 167.8 2–33:1 73.6 Not measurable

Fig. 4. High resolution SEM image of a silica film cast from TMOS/IBTMOS = 1:1.

are shown in Table 3. For a TMOS/IBTMOS ratio of 1:3, althoughsuperhydrophobicity was observed on bulk silica, ∼300 nm filmscast from this mixture were not superhydrophobic. A possible rea-son for this result is that large pores such as those observed inthe bulk materials cannot form in thin films. However, with aTMOS/IBTMOS ratio of 1:1, a superhydrophobic thin film with acontact angle of 165–170◦ was achieved. Fig. 4 shows the surfacemicrostructure of a 300 nm silica film cast onto a glass slide froma TMOS/IBTMOS ratio of 1:1; the surface consists of globules (20–50 nm in diameter) surrounding a network of pores (submicronsize) formed by evaporation of solvent from the sol. These obser-vations also indicate that surface roughness exists in two scales:the pores contribute to the submicron scale roughness, and thesilica particles contribute to the nanoscale roughness. This two-scale surface roughness combined with the hydrophobic surfacethat resulted from the isobutyl groups in IBTMOS, generated a su-perhydrophobic film surface.

3.3. Stability of superhydrophobic surfaces

The hydrocarbon side group contained in the superhydrophobicsilica can decompose (or be oxidized in air) at elevated temper-ature and thus cause the loss of superhydrophobicity. TGA stud-ies of samples synthesized with TMOS/IBTMOS ratios of 1:1 and1:3, demonstrate that the onset of weight loss is ∼200 ◦C and∼207 ◦C, respectively, as shown in Fig. 5. With continued heat-ing above the onset temperature, the organic groups in the silicastructure begin to degrade due to oxidation or bond cleavage ofthe isobutyl groups. TGA results showed that 27.7% loss is observedfor a TMOS/IBTMOS ratio of 1:1. EDS results in Table 4 and Fig. 6also indicated that the isobutyl group content was 28.5 wt% inthe 1:1 silica (H atoms were included in the calculation) whichagrees well with TGA data. For a TMOS/IBTMOS ratio of 1:3, sim-ilar results were observed to those with a TMOS/IBTMOS ratio of1:1. Table 5 shows the contact angle data on the heat-treated sur-faces at different temperatures. Contact angle decreased with in-creasing heat treatment temperature and contact angle hysteresisincreased. When the heat treated surface was again treated withIBTMOS/hexane solution, the surface superhydrophobicity was re-covered (CA ∼ 165◦ and hysteresis <5◦). This suggests that thermal

Fig. 5. TGA results for silica from different TMOS/IBTMOS ratios. Temperature ramp-ing rate: 4 ◦C/min.

Table 4EDS analysis results of the thin film from TMOS/IBTMOS = 1:1

Element Wt% At%

C 25.13 35.13O 44.56 46.76Si 30.31 18.11Totals 100.00 100.00

Fig. 6. EDS spectrum of the thin film from TMOS/IBTMOS = 1:1.

Table 5Thermal stability of the as-prepared superhydrophobic surfaces after heating for 2 hat different temperatures

Heat treatmenttemperature (◦C)

Contact angle(◦)

Contact anglehysteresis (◦)

200 165.3 12.6300 91.7 >60400 15.2 Not measurable

degradation of the surface is primarily due to a change in surfacechemistry rather than to a change in the surface roughness. FTIRanalysis of the silica film after a heat treatment at 400 ◦C indicatedthat hydrocarbon absorptions were absent as shown in Fig. 7; CH3

and CH2 stretching vibration (2975–2840 cm−1) and deformationvibration (1232 cm−1, 1173 cm−1, 894 cm−1 and 839 cm−1) inten-sities were below the detectability limit after the heat treatment,confirming decomposition of the organic groups.

Organic polymers with reactive (e.g., double bonds) or low en-ergy (e.g., tertiary hydrogen) structures in the polymer main chain,are vulnerable to UV irradiation. When impurities (i.e., catalyst)and UV absorbing groups are present, the polymer can undergoa photo-oxidation process to form carbonyl or hydroxyl groups onthe surface and thus degrade (unzip) into smaller chains [26]. Theintroduction of these hydrophilic groups leads to the loss of super-

Y. Xiu et al. / Journal of Colloid and Interface Science 326 (2008) 465–470 469

Fig. 7. FTIR spectra of the silica superhydrophobic silica thin film before and afterheat treatment at 400 ◦C for 1 h.

Fig. 8. UV moisture weathering of superhydrophobic PB and silica thin film surfaces.Bottom curve: SF6 RIE etched PB + UV absorber TINUVIN 1577FF (2%) + antioxidantIRGANOX 1010FF (1.5%); top curve: silica film from TMOS/IBTMOS (1:1) using a twostep acid–base catalyzed sol–gel process.

hydrophobicity. As described previously, when the surface is hy-drophilic, surface roughness enhances the hydrophilicity to super-hydrophilicity. For the silica film, only small side groups (isobutylgroups) exist on the surface; these moieties are more UV stablethan are polymers because negligible impurities and less UV-fragiledefects exist on the silica surface; thus the silica film exhibitsbetter UV stability than do organic polymers (polybutadiene flu-orinated in SF6 plasma) as shown in Fig. 8. In addition, the silicamain chain is comprised of Si–O bonds, which have higher bondstrength and thus better UV stability than organic polymer materi-als. Other commercially available polymers such as ethylene–vinylacetate copolymer, polydimethyl siloxane and Teflon, all show sur-face degradation after long time UV aging tests.

When a superhydrophobic surface begins to degrade, the hys-teresis changes more rapidly than does the contact angle. Thisobservation is shown in Table 6, where contact angles changedvery little, while hysteresis changed significantly. Therefore, con-tact angle hysteresis can be used to evaluate the surface chemistrychange when the contact angles remain almost unaltered. The sur-face fraction that is air beneath the water drop becomes smallerand smaller, until it ultimately disappears, and the Cassie model nolonger accurately describes the surface/water contact. Instead, thecontact interface is better described by the Wenzel state. The FTIRspectrum in Fig. 9 shows that after UV irradiation, the SiO–CH3 de-formation vibration band at 1444 cm−1 disappeared due to contin-ued hydrolysis. The ≡CH deformation vibration band at 1352 cm−1

Table 6Effect of UV irradiation on contact angle and hysteresis for a thin film formed fromTMOS/IBTMOS = 1:1

UV weathering testof silica film

Contactangle

Hysteresis

8 days 159–162◦ 3–5◦12 days 158◦ 15–20◦

Fig. 9. FTIR comparison of silica thin films before and after prolonged UV irradiation.Top curve: silica film from TMOS/IBTMOS = 1:1 before UV test; bottom curve: thesame silica film after UV testing for 12 days.

Fig. 10. Superhydrophobic silica thin film AFM surface image after PFOS treatment;inset, water droplet image on the surface; contact angle, ∼168◦; hysteresis, 2◦ .

also disappeared while the OCH2 vibration band at 1330 cm−1 ap-peared after UV irradiation because of oxidative degradation of theisobutyl groups specifically on tertiary carbons.

In order to mitigate the thermal and UV stability degradationof the superhydrophobic TMOS/IBTMOS surfaces, a heat treatmentwas performed on the silica thin films immediately after for-mation, to decompose/oxidize the organic moieties; subsequently,a fluoroalkyl silane treatment was carried out. AFM was usedto probe the surface morphology of PFOS-treated superhydropho-bic silica films. Fig. 10 shows the surface morphology of a silicafilm produced from a 1:1 TMOS/IBTMOS mixture, that was heattreated at 500 ◦C and then PFOS treated; the root mean square(RMS) surface roughness from AFM was 330 nm. On the rough sur-face, air can still be effectively trapped between the water dropletand the substrate. Thus, the Cassie state can be maintained toachieve a superhydrophobic surface with both high contact an-gle and low hysteresis as shown in the inset in Fig. 10. Acceler-

470 Y. Xiu et al. / Journal of Colloid and Interface Science 326 (2008) 465–470

Fig. 11. UV stability of a PFOS-treated rough silica thin film surface.

Table 7Thermal stability of a PFOS-treated superhydrophobic thin film surface after heatingfor 2 h at different temperatures

Temperature(◦C)

Contact angle(◦)

Contact anglehysteresis (◦)

400 166.3 2.1450 160.4 >30500 ∼0 Not measurable

ated UV and thermal stability test results are shown in Fig. 11and Table 7. Because of the attachment of fluoroalkyl groups onthe surface, improved thermal and UV stability were achieved dueto the presence of a linear fluorocarbon chain on the silica sur-face. UV tests showed a stable superhydrophobic surface after a5500 h testing period without degradation of either contact angleor contact angle hysteresis. A thermal treatment at 400 ◦C demon-strated that the surface is stable to this temperature, which isa significant improvement relative to the results obtained witha superhydrophobic surface that contained isobutyl groups. Fur-ther temperature increases resulted in a decrease of the contactangle and an increase of contact angle hysteresis. When the treat-ment temperature was 500 ◦C, the surface was no longer super-hydrophobic (contact angle = ∼0◦) as a result of the decompo-sition/oxidation of the surface fluoroalkyl groups. To our knowl-edge, this is one of the most stable superhydrophobic surfacesreported to date. This finding demonstrates that it may be possibleto use superhydrophobic/self-cleaning surfaces in harsh environ-ments such as high UV irradiation, high temperature, and outdoorapplications.

4. Conclusions

Superhydrophobic bulk silica materials and analogous silica thinfilms were prepared by TMOS/IBTMOS two-step acid base cat-alyzed sol–gel processing. Thermal analysis shows that these su-perhydrophobic materials begin to lose superhydrophobicity whenheated above 200 ◦C. UV weathering tests (ASTM D 4329) indicatethat the silica materials demonstrate better stability compared to

SF6 plasma etched polybutadiene films but lost superhydrophobic-ity due to the oxidative degradation of the isobutyl groups. Whenthe superhydrophobic surfaces from TMOS/IBTOMS mixtures wereheated to 500 ◦C, the surface became superhydrophilic. When aPFOS treatment was performed on the surface, the attachment offluoroalkyl groups resulted in superhydrophobicity. This surface ismuch more stable than a surface with isobutyl groups; after a UVtesting time of 5500 h, no degradation of either contact angle orcontact angle hysteresis was observed for the PFOS-treated surface.Thermal stability of the PFOS-treated rough silica surfaces was alsomuch improved at temperatures up to 400 ◦C. The PFOS-treatedsuperhydrophobic surface is one of the most stable surfaces re-ported to date. This approach to the generation of superhydropho-bic surfaces may have much significance for the successful appli-cation of self-cleaning surfaces in a variety of environments.

Acknowledgments

The authors acknowledge financial support from the NationalScience Foundation (NSF CMMI #0422553) and the National Elec-tric Energy Testing Research and Applications Center (NEETRAC) atGeorgia Institute of Technology.

References

[1] W. Barthlott, C. Neinhuis, Planta 202 (1997) 1.[2] R.N. Wenzel, Ind. Eng. Chem. 28 (1936) 988.[3] R.N. Wenzel, J. Phys. Colloid Chem. 53 (1949) 1466.[4] A.B.D. Cassie, Discuss. Faraday Soc. 3 (1948) 11.[5] J. Genzer, K. Efimenko, Science 290 (2000) 2130.[6] K. Tadanaga, J. Morinaga, A. Matsuda, T. Minami, Chem. Mater. 12 (2000) 590.[7] A. Nakajima, A. Fujishima, K. Hashimoto, T. Watanabe, Adv. Mater. 11 (1999)

1365.[8] D. Schondelmaier, S. Cramm, R. Klingeler, J. Morenzin, C. Zilkens, W. Eberhardt,

Langmuir 18 (2002) 6242.[9] Y. Xiu, L. Zhu, D.W. Hess, C.P. Wong, Nano Lett. 7 (2007) 3388.

[10] A.V. Rao, S.D. Bhagat, H. Hirashima, G.M. Pajonk, J. Colloid Interface Sci. 300(2006) 279.

[11] A. Roig, E. Molins, E. Rodriguez, S. Martinez, M. Moreno-Manas, A. Vallribera,Chem. Commun. (2004) 2316.

[12] Y. Xiu, L. Zhu, D.W. Hess, C.P. Wong, Langmuir 22 (2006) 9676.[13] H. Wang, J. Fang, T. Cheng, J. Ding, L. Qu, L. Dai, X. Wang, T. Lin, Chem. Com-

mun. (2008) 877.[14] N.J. Shirtcliffe, G. McHale, M.I. Newton, C.C. Perry, Langmuir 19 (2003) 5626.[15] L. Zhai, F.C. Cebeci, R.E. Cohen, M.F. Rubner, Nano Lett. 4 (2004) 1349.[16] R.M. Jisr, H.H. Rmaile, J.B. Schlenoff, Angew. Chem. Int. Ed. 44 (2005) 782.[17] L. Feng, Z.L. Yang, J. Zhai, Y.L. Song, B.Q. Lin, Y.M. Ma, Z.Z. Yang, L. Jiang, D.B.

Zhu, Angew. Chem. Int. Ed. 42 (2003) 4217.[18] J.Y. Shiu, C.W. Kuo, P.L. Chen, C.Y. Mou, Chem. Mater. 16 (2004) 561.[19] M. Morra, E. Occhiello, F. Garbassi, Langmuir 5 (1989) 872.[20] P. Favia, G. Cicala, A. Milella, F. Palumbo, R. Rossini, R. d’Agostino, Surf. Coat.

Technol. 169 (2003) 609.[21] Y. Ren, T.P. Lodge, M.A. Hillmyer, Macromolecules 34 (2001) 4780.[22] H. Zhang, G.C. Hardy, Y.Z. Khimyak, M.J. Rosseinsky, A.I. Cooper, Chem.

Mater. 16 (2004) 4245.[23] J.L. Gu, J.L. Shi, G.J. You, L.M. Xiong, S.X. Qian, Z.L. Hua, H.R. Chen, Adv. Mater. 17

(2005) 557.[24] R. Richer, L. Mercier, Chem. Mater. 13 (2001) 2999.[25] H.M. Shang, Y. Wang, S.J. Limmer, T.P. Chou, K. Takahashi, G.Z. Cao, Thin Solid

Films 472 (2005) 37.[26] N.J. Turro, Modern Molecular Photochemistry, Univ. Sci. Books, Sausalito, CA,

1991.