Superoleophobic Cotton Textiles

Transcription

1 2456 Langmuir 2009, 25, Superoleophobic Cotton Textiles Boxun Leng,, Zhengzhong Shao, Gijsbertus de With, and Weihua Ming*,, Key Laboratory of Molecular Engineering of Polymers of Ministry of Education, AdVanced Materials Laboratory, Department of Macromolecular Science, Fudan UniVersity, Shanghai , China, Laboratory of Materials and Interface Chemistry, EindhoVen UniVersity of Technology, P.O. Box 513, 5600 MB EindhoVen, The Netherlands, and Nanostructured Polymers Research Center, Materials Science Program, UniVersity of New Hampshire, Durham, New Hampshire ReceiVed September 22, ReVised Manuscript ReceiVed December 4, 2008 Common cotton textiles are hydrophilic and oleophilic in nature. Superhydrophobic cotton textiles have the potential to be used as self-cleaning fabrics, but they typically are not super oil-repellent. Poor oil repellency may easily compromise the self-cleaning property of these fabrics. Here, we report on the preparation of superoleophobic cotton textiles based on a multilength-scale structure, as demonstrated by a high hexadecane contact angle (153 for 5 µl droplets) and low roll-off angle (9 for 20 µl droplets). The multilength-scale roughness was based on the woven structure, with additional two layers of silica particles (microparticles and nanoparticles, respectively) covalently bonded to the fiber. Superoleophobicity was successfully obtained by incorporating perfluoroalkyl groups onto the surface of the modified cotton. It proved to be essential to add the nanoparticle layer in achieving superoleophobicity, especially in terms of low roll-off angles for hexadecane. Introduction Through millions of years of evolution, nature has developed many interesting superhydrophobic surfaces, such as various plant leaves (a typical example being the lotus leaf), 1,2 legs of the water strider, 3 troughs of the Namib desert beetle, 4 and a gecko s feet. 5 Surface roughness at a dual or multilength scale has shown to be the key in generating the surprising nonwetting behavior. Inspired by this finding, biomimetic artificial superhydrophobic surfaces have been produced via various methods, 6-17 targeting a broad range of potential applications including self-cleaning and anti-biofouling coatings. 9 On the other hand, it is highly desirable for superhydrophobic surfaces to be also oil-repellent to maintain their superhydrophobicity. For instance, in an industrial or household environment, a superhydrophobic surface with poor oil repellency can be easily contaminated by * Towhomcorrespondenceshouldbeaddressed. W.Ming@unh.edu. Fudan University. Eindhoven University of Technology. University of New Hampshire. (1) Herminghaus, S. Europhys. Lett. 2000, 52, (2) Barthlott, W.; Neinhuis, C. Planta 1997, 202, 1 8. (3) Gao, X. F.; Jiang, L. Nature 2004, 432, (4) Parker, A. R.; Lawrence, C. R. Nature 2001, 414, (5) Autumn, K.; Liang, Y. A.; Hsieh, S. T.; Zesch, W.; Chan, W. P.; Kenny, T. W.; Fearing, R.; Full, R. J. Nature 2000, 405, (6) Han, W.; Wu, D.; Ming, W.; Niemantsverdriet, H.; Thune, P. C. Langmuir 2006, 22, (7) Ming, W.; Wu, D.; van Benthem, R.; de With, G. Nano Lett. 2005, 5, (8) Roach, P.; Shirtcliffe, N. J.; Newton, M. I. Soft Matter 2008, 4, (9) Zhang, X.; Shi, F.; Niu, J.; Jiang, Y. G.; Wang, Z. Q. J. Mater. Chem. 2008, 18, (10) Li, X. M.; Reinhoudt, D.; Crego-Calama, M. Chem. Soc. ReV. 2007, 36, (11) Feng, X. J.; Jiang, L. AdV. Mater. 2006, 18, (12) Liu, H.; Zhai, J.; Jiang, L. Soft Matter 2006, 2, (13) Zhai, L.; Berg, M. C.; Cebeci, F. C.; Kim, Y.; Milwid, J. M.; Rubner, M. F.; Cohen, R. E. Nano Lett. 2006, 6, (14) Sun, T. L.; Feng, L.; Gao, X. F.; Jiang, L. Acc. Chem. Res. 2005, 38, (15) Shirtcliffe, N. J.; McHale, G.; Newton, M. I.; Chabrol, G.; Perry, C. C. AdV. Mater. 2004, 16, (16) Zhai, L.; Cebeci, F. C.; Cohen, R. E.; Rubner, M. F. Nano Lett. 2004, 4, (17) Erbil, H. Y.; Demirel, A. L.; Avci, Y.; Mert, O. Science 2003, 299, oily substances, which will in turn compromise the surface superhydrophobicity. Therefore, superlyophobic (or superhygrophobic 18 ) surfaces combining superhydrophobic and superoleophobic properties are desirable for practical applications. 19 Despite extensive investigations on superhydrophobic surfaces, studies on superoleophobic surfaces with high repellency against liquids with low surface tensions (<35 mn/m) have been rather limited so far. Oil repellency has been examined on various perfluorinated, superhydrophobic surfaces, with reported static contact angles (CAs) for benzene, hexadecane, or rapeseed oil in the range of No receding contact angles or roll-off angles for oil droplets were reported on most of the reported surfaces, so it is difficult to judge whether these surfaces are truly superoleophobic. For instance, despite a high static CA (160 ) for rapeseed oil on a superhydrophobic surface made of surface-fluorinated carbon nanotubes, 27 the rapeseed oil droplet would remain pinned to the surface (indicating very low receding CA) when the sample was tilted, underlying the importance of achieving high receding CAs (and thus low contact angle hysteresis and low roll-off angles). Only those surfaces with high CAs (>150 ) and low roll-off angles for oil droplets can be regarded as truly superoleophobic surfaces. Very recently, (18) Marmur, A. Langmuir 2008, 24, (19) Tuteja, A.; Choi, W.; Ma, M. L.; Mabry, J. M.; Mazzella, S. A.; Rutledge, G. C.; McKinley, G. H.; Cohen, R. E. Science 2007, 318, (20) Zimmermann, J.; Rabe, M.; Artus, G. R. J.; Seeger, S. Soft Matter 2008, 4, (21) Cao, L. L.; Price, T. P.; Weiss, M.; Gao, D. Langmuir 2008, 24, (22) Yan, H.; Kurogi, K.; Tsujii, K. Colloid Surf., A 2007, 292, (23) Tian, Y.; Liu, H. Q.; Deng, Z. F. Chem. Mater. 2006, 18, (24) Nicolas, M.; Guittard, F.; Geribaldi, S. Angew. Chem., Int. Ed. 2006, 45, (25) Nakajima, A.; Hoshino, M.; Song, J. H.; Kameshima, Y.; Okada, K. Chem. Lett. 2005, 34, (26) Yabu, H.; Takebayashi, M.; Tanaka, M.; Shimomura, M. Langmuir 2005, 21, (27) Li, H. J.; Wang, X. B.; Song, Y. L.; Liu, Y. Q.; Li, Q. S.; Jiang, L.; Zhu, D. B. Angew. Chem., Int. Ed. 2001, 40, (28) Shibuichi, S.; Yamamoto, T.; Onda, T.; Tsujii, K. J. Colloid Interface Sci. 1998, 208, (29) Tsujii, K.; Yamamoto, T.; Onda, T.; Shibuichi, S. Angew. Chem., Int. Ed. Engl. 1997, 36, (30) Wu, D.; Vranchen, R. J.; van Loenen, B. G. H.; van Benthem, R. A. T. M.; De With, G.; Ming, W. Polym. Mater.: Sci. Eng. 2007, 97, /la CCC: $ American Chemical Society Published on Web 01/22/2009

2 Superoleophobic Cotton Textiles Langmuir, Vol. 25, No. 4, Scheme 1. Schematic Illustration of the Procedure for the Preparation of Dual-Size Structure onto the Surface of Woven Cotton Fibers, Combining an In Situ Stöber Reaction with the Subsequent Adsorption of Silica Nanoparticles truly superoelophobic surfaces have been achieved on the basis of microhoodoo 19 and nanonail 31 structures, as exemplified by low contact angle hysteresis for probe liquids of low surface tension (<30 mn/m) such as octane and ethanol. In both cases, the key to obtaining true superoleophobicity is the re-entrant or overhang surface structure, thus ensuring the entrapment of air beneath the top solid surface and preventing the transition from the Cassie-Baxter state to the Wenzel state. 19,21,31 However, the fabrication of these superoleophobic surfaces involves lithography and etching steps, which may limit their practical applications. We have recently prepared superhydrophobic cotton textiles by covalently introducing a layer of silica particles onto the fiber surface and by subsequent chemical modification. 32 The intrinsic roughness of the woven textiles can be considered as local surface curvature (analogous to a re-entrant structure). Combining this local curvature with a layer of silica particles, it may be possible to produce a superoleophobic surface; it was indeed shown that, with surface perfluorination, the modified cotton was turned highly oleophobic, as demonstrated by a static contact angle of 140 and a roll-off angle of 24 fora15µl sunflower oil droplet. However, the sample cannot be deemed superoleophobic, since a 10 µl hexadecane droplet cannot easily roll off the surface despite a static CA of Previously, we also reported the preparation of superhydrophobic surfaces by using well-defined raspberry-like particles, in a way of mimicking the dual-size surface structure of the lotus leaf. 7 In this contribution, we will introduce the raspberry-like, dual-size structure onto the woven cotton fibers, leading to a triple-size surface structure. As shown in Scheme 1, relatively big silica particles (diameter: 800 nm; we call them silica microparticles thereafter) were in situ generated and covalently bonded to the cotton fibers. After treatment with 3-aminopropyl-triethoxysiloxane (APS) and hydrochloric acid, the surface charge was turned positive due to the protonation of amine groups. Negatively charged silica nanoparticles were then electrostatically adsorbed onto the fiber surface. The obtained roughened structure was stabilized by SiCl 4 cross-linking, followed by surface modification with a perfluoroalkyl silane. The wettability of these samples was evaluated by both hexadecane and water. Experimental Section Materials. Tetraethylorthosilicate (TEOS) was purchased from Fluka, 3-aminopropyl-triethoxysiloxane (APS) and silicon tetrachloride (SiCl 4 ) were obtained from Aldrich, and 1H,1H,2H,2Hperfluorodecyl trichlorosilane (Rf-Si) was obtained from ABCR. A cotton textile was purchased from a local fabric store. Cotton textiles of 4 4cm 2 pieces were cleaned with water, ethanol, and toluene to remove possible impurities. Other chemicals were purchased from Merck and used as received. In Situ Introduction of Silica Microparticles on Cotton Fibers. The process followed the traditional Stöber method as described elsewhere. 32 Briefly, a piece of clean cotton textile was added to a solution of 25 ml of methanol, 75 ml of 2-propanol, and 21 ml of ammonia solution (25%). TEOS (6 ml) was added dropwise to the mixture, and the mixture reacted at room temperature for 6 h under magnetic stirring (120 rpm). After the reaction, the cotton textile was washed with methanol in an ultrasonic bath five times (5 min for each wash) to remove physically adsorbed particles and then dried under N 2 flow. The obtained textile was named as C TEOS. Adsorption of Silica Nanoparticles. The adsorption process was driven by electrostatic attraction. C TEOS was immersed in APS toluene solution (1 vol %) for 30 min, leading to amine groups at the particle and fiber surface, which were then protonated with hydrochloric acid (0.1 mol/l, 30 s). The positively charged textile was dipped into a 2-propanol dispersion of negatively charged silica nanoparticles (1 mg/ml) for 20 min. The adsorption process of silica nanoparticles was repeated up to three times (no further APS or HCl treatment was used) to increase the particle areal density. Three different silica nanoparticles, synthesized with the Stöber method, 33 were used for the adsorption experiments with the following sample IDs for the final modified cotton samples: C a, 160 nm (diameter); C b, 220 nm; C c, 100 nm. The number after sample IDs (below) stands for the number of adsorption cycles for the corresponding sample. Surface Modification by Perfluoroalkyl Silane. The samples after every adsorption cycle were treated with silicon tetrachloride toluene solution (0.5 vol %) for 5 min to enhance the mechanical robustness. These samples were then immersed in a precooled Rf-Si solution in toluene (0.4 vol %) and reacted at 0 C for 3 h. Subsequently, they were flushed with toluene and dried in air. The samples after perfluorination are marked with Rf after each sample ID. For example, C b -1-Rf stands for the sample containing silica microparticles, followed by one cycle of adsorption of 220 nm silica nanoparticles and surface perfluorination. Measurements. A Dataphysics OCA 30 instrument was used to measure the contact angle and roll-off angle at room temperature (about 20 C). Water (surface tension γ ) 72.2 mn/m at 20 C) and hexadecane (γ ) 27.4 mn/m at 20 C) were used as probe liquids. The contact angles and roll-off angles were averaged values by measuring at 3-4 different points on each sample surface. The tilting rate for the roll-off angle measurements was 1 /s. Scanning electron microscopy (SEM) was performed on a Philips XL-30 ESEM FEG instrument in high-vacuum mode and operated at a 2 kv acceleration voltage. Thermogravimetric analysis (TGA) was carried out on a TGA Q500 apparatus under oxygen atmosphere at a flow rate of 20 cm 3 /min. The heating rate was 20 C/min, ranging from room temperature to 700 C. (31) Ahuja, A.; Taylor, J. A.; Lifton, V.; Sidorenko, A. A.; Salamon, T. R.; Lobaton, E. J.; Kolodner, P.; Krupenkin, T. N. Langmuir 2008, 24, (32) Hoefnagels, H. F.; Wu, D.; de With, G.; Ming, W. Langmuir 2007, 23, (33) Stober, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26,

3 2458 Langmuir, Vol. 25, No. 4, 2009 Leng et al. Figure 1. Morphology of samples (a) C TEOS, (b) C b -1, (c) C b -2, and (d) C b -3. Shown in the insets are the images of static water droplets (5 µl) and C 16 H 34 droplets (5 µl) on the respective textiles modified by Rf-Si, namely, (a) C TEOS -Rf, (b) C b -1-Rf, (c) C b -2-Rf, and (d) C b -3-Rf. Results and Discussion As shown in Scheme 1, the multilength-scale roughness based on the cotton fibers was obtained from a two-step procedure. The first step was to introduce silica microparticles ( 800 nm) to cotton fibers via the Stöber reaction. After this treatment, the fiber surface was covered with a layer of silica microparticles (Figure 1a). The silica microparticles were covalently bonded to the cotton fiber due to the abundant hydroxyl groups in cellulose. The second step was to adsorb silica nanoparticles onto the microparticle-covered fibers through electrostatic attraction. This procedure is similar to the layer-by-layer (LbL) self-assembly technique, which employs complementary interactions in solutions, such as the interaction between positively and negatively charged species After reaction with APS, the surface of the silica microparticles was turned positively charged via the protonation of the amine groups by HCl. In addition, the remaining part of the cotton fiber that was not covered by silica microparticles was also turned positively charged, since APS also reacted with the surface hydroxyl groups from the cellulose. The sample was then dipped in a 2-propanol dispersion of silica nanoparticles, which had a zeta potential of about -50 mv, leading to the adsorption of silica nanoparticles onto both the silica microparticles and the cotton fiber surface (Figure 1b-d). After one adsorption cycle (Figure 1b), the silica nanoparticles formed a monolayer on the surface although the layer did not fully cover the silica microparticles and the fiber surface. The adsorption process was repeated up to three times to increase the amount of the silica nanoparticles covering both the silica microparticles and fiber surface (Figure 1c,d). The adsorption of nanoparticles on the fiber surface may be considered as beneficial for the roughened surface, similar to the lotus leaf on which nanostructured protrusions cover both micropapillae and the lower part of the leaf. 37 TGA was used to determine the percentage of inorganic component (silica) in C b samples after each treatment (Figure 2). (34) Tsai, H. J.; Lee, Y. L. Langmuir 2007, 23, (35) Decher, G. Science 1997, 277, (36) Ma, M. L.; Gupta, M.; Li, Z.; Zhai, L.; Gleason, K. K.; Cohen, R. E.; Rubner, M. F.; Rutledge, G. C. AdV. Mater. 2007, 19, (37) Feng, L.; Li, S. H.; Li, Y. S.; Li, H. J.; Zhang, L. J.; Zhai, J.; Song, Y. L.; Liu, B. Q.; Jiang, L.; Zhu, D. B. AdV. Mater. 2002, 14, Figure 2. Weight loss for the cotton textile samples (C b series) before and after each step of the modification by TGA. For the clean cotton prior to the modification, the remaining weight was 0.2% after being heated to 700 C in an oxygen atmosphere. After the Stöber reaction (C TEOS ), the remaining weight increased to 8.4% due to the formation of silica microparticles. This value was higher than that reported previously, probably due to the formation of a silica film. 38 The remaining weight percentage further increased (including the APS contribution) to 10.4%, 14.5%, and 16.4% after one, two, and three cycles of deposition of silica nanoparticles, respectively. Obviously, the surface coverage of the cotton fiber and microparticles by silica nanoparticles can be increased by repeating the electrostatic adsorption process. The morphology (SEM) and TGA data for C a samples are shown in the Supporting Information (Figure S1). After each adsorption cycle, the samples were treated with a SiCl 4 solution in toluene (0.5 vol %) for 5 min to increase the mechanical robustness for the modified cotton fabrics, since the SiCl 4 treatment leads to the formation of cross-links among silica particles (both between nano- and microparticles and among nanoparticles) as well as between silica particles and the cotton fiber surface. Furthermore, the presence of moisture (a trace amount of water) could hydrolyze SiCl 4, allowing silica particles (38) Brzoska, J. B.; Benazouz, I.; Rondelez, F. Langmuir 1994, 10,

4 Superoleophobic Cotton Textiles Langmuir, Vol. 25, No. 4, Table 1. Water Contact Angles and Roll-Off Angles on Cotton Samples Modified with Rf-Si static contact angle [ ] roll-off angle [ ] sample ID 5 µl 10 µl 20 µl 5 µl 10 µl 20 µl C TEOS -Rf 157 ( ( ( 1 34( ( 7 12( 4 C a -1-Rf 158 ( ( ( 1 14( 1 11( 2 7( 2 C a -2-Rf 159 ( ( ( 1 15( 1 10( 1 8( 1 C a -3-Rf 160 ( ( ( 1 15( 3 7 ( 3 5( 1 C b -1-Rf 158 ( ( ( 1 14( 1 12( 3 5( 1 C b -2-Rf 159 ( ( ( 1 14( 1 12( 3 7( 2 C b -3-Rf 159 ( ( ( 1 17( 2 11( 5 7( 1 C c -Rf 159 ( ( ( 1 16( 2 11( 1 8( 1 Table 2. Hexadecane Contact Angles and Roll-Off Angles on Cotton Samples Modified with Rf-Si static contact angle [ ] roll-off angle [ ] sample ID 5 µl 10 µl 20 µl 5 µl 10 µl 20 µl C TEOS -Rf 142 ( ( ( 3 76( ( 5 22( 3 C a -1-Rf 152 ( ( ( 2 32( ( 1 11( 2 C a -2-Rf 151 ( ( ( 2 32( 9 15( 2 9( 2 C a -3-Rf 151 ( ( ( 1 30( 4 15( 2 10( 1 C b -1-Rf 152 ( ( ( 1 30( ( 2 10( 1 C b -2-Rf 152 ( ( ( 1 31( 4 14( 1 9( 2 C b -3-Rf 152 ( ( ( 1 25( 7 14( 3 9( 1 C c -Rf 151 ( ( ( 3 30( 6 17( 5 12( 3 to fuse together to some extent due to the formation of extra silica. This is useful from the standpoint of increasing the film robustness. Of course, the SiCl 4 treatment should be limited to a short period of time, otherwise too much silica could form, which would compromise the hierarchical structure we have aimed for. In the last step, the particle-containing cotton samples were modified by Rf-Si, resulting in the formation of a monolayer of low surface tension perfluoroalkyl tails covering the silica surface. It should be noted that the grafting conditions should be optimized to ensure the monolayer formation, 39 such as the Rf-Si concentration, a low reaction temperature (0 C), and a trace amount of water. The surface wettability was evaluated by static CAs and rolloff angles by using water and hexadecane (C 16 H 34 ) as probe liquids. Those fibers that protrude out from the cotton surface (shown in insets of Figure 1) made it difficult to collect accurate advancing/receding CAs, 32,40 so only static CAs were reported here. The insets in Figure 1 are the optical images of droplets of water and C 16 H 34 (5 µl) on the corresponding samples. These surfaces demonstrated very high repellency against both water and C 16 H 34. Samples C a showed similar high water and oil repellency (Tables 1 and 2). The CAs and roll-off angles of all samples are listed in Table 1 for water and Table 2 for C 16 H 34. For the sample with only silica microparticles (C TEOS -Rf), the static CAs for 10 µl water and C 16 H 34 droplets were 153 and 137, respectively, similar to the values we reported previously. 32 When silica nanoparticles were introduced to the cotton samples, triple-size surface structures were obtained; as a result, the antiwetting property of the cotton samples was improved significantly. Although the addition of nanoparticles did not change much in the static CAs (all above 150 ) for water droplets, the roll-off angles for the droplets of different volumes (5, 10, and 20 µl) decreased substantially (Table 1). The roll-off angles were dependent on the volume of the water droplets. For instance, for the C b -1-Rf sample, the roll-off angles ranged from 14 for a 5 µl droplet (39) Li, S. H.; Zhang, S. B.; Wang, X. H. Langmuir 2008, 24, (40) Wang, T.; Hu, X. G.; Dong, S. J. Chem. Commun. 2007, (41) Zimmermann, J.; Artus, G. R. J.; Seeger, S. Appl. Surf. Sci. 2007, 253, to 5 for a 20 µl droplet. Our results clearly indicate that the samples with additional silica nanoparticles have demonstrated better water repellency than those with only microparticles. 32,40 For C 16 H 34 droplets, the incorporation of silica nanoparticles to the cotton samples led to an increase of about 10 for the static CAs (Table 2), independent of the size of nanoparticles and the number of the adsorption cycles. The static CAs were greater than 150 for 5 µl droplets. More significantly, the roll-off angles decreased dramatically when nanoparticles were incorporated onto the fiber surface (Table 2): for 20 µl droplets, roll-off angles of as low as 9 were observed. Again, the roll-off angles strongly depended on the droplet volume. With the high hexadecane and water repellency (the video clips in the Supporting Information clearly demonstrate that both water and hexadecane hate our modified cotton samples: when the sample is lifted out of these two liquids, the liquid contact line retreats very rapidly from the top edge of the modified cotton), we can conclude that these samples are both superhydrophobic and superoleophobic. A sonication experiment was carried out to examine the mechanical robustness of these samples. C b -2-Rf was subjected to sonication in an ethanol bath (50 ml) for 30 min. The sample was dried and then examined by CA measurements. The water CA values remained essentially unchanged: the static CA and roll-off angle of a 10 µl droplet were 154 ( 1 and 13 ( 3, respectively. On the other hand, despite that the static CA for a10µl C 16 H 34 droplet remained high (144 ( 3 ), the roll-off angle increased from 14 ( 1 to 41 ( 6. We tried to find out whether there were significant changes, both chemically and morphologically, for the sample after sonication. SEM imaging showed no obvious change of the morphology of the fiber surface after sonication (data not shown). X-ray photoelectron spectroscopy (XPS) was used to examine possible changes in the surface chemical composition of the sample after the sonication test; the F/C atomic ratio (1.73) at a takeoff angle of 90 was similar to that before sonication. Therefore, at present, it is not clear to us why the roll-off angle for C 16 H 34 increased. A recent report 41 suggests that, for a superhydrophobic film after being immersed in water, the increase in the roll-off angle of water may be related to the pinning of the contact line due to surface inhomogeneity. In our case, it appears that the sonication may have introduced some sort of surface inhomogeneity, which nonetheless did not show significant effect on the contact line between a water droplet and the surface, but did somehow alter the contact line between a C 16 H 34 droplet and the surface in a way that we do not exactly know, leading to the increase of the roll-off angle. Two distinct models, the Wenzel 42 and Cassie-Baxter 43 models, have been extensively used to explain the roughness effect on the contact angles of liquids on a roughened surface. The Wenzel model describes the wetting regime in which the liquid penetrates into the roughened surface (there is always intimate contact between the liquid and the solid surface), usually leading to high contact angle hysteresis. In contrast, in the Cassie wetting regime, there is air trapped between the solid surface and liquid, which results in much smaller hysteresis. When a liquid droplet sits on a liquid-repellent cotton textile surface, the wetting behavior can be described by the Casssie-Baxter equation, 43 cos θ CB ) f ls cos θ 0 - f lv, where θ CB is the observed CA on a rough surface, θ 0 is the intrinsic CA on the corresponding smooth surface (θ 0 is 120 for water and 80 for C 16 H 34 on a perfluoroalkyl (42) Wenzel, R. N. Ind. Eng. Chem. 1936, 28, 988. (43) Cassie, A. B. D.; Baxter, S. Trans. Faraday Soc. 1944, 40, (44) Michielsen, S.; Lee, H. J. Langmuir 2007, 23,

5 2460 Langmuir, Vol. 25, No. 4, 2009 Leng et al. Figure 3. Morphology of sample C c observed by SEM. Insets are profiles of water and hexadecane (C 16 H 34 ) on the corresponding sample C c -Rf. surface), f ls is the liquid/solid contact area divided by the projected area, and f lv is the liquid/vapor contact area divided by the projected area. This equation has been recently modified to account for the local surface roughness on the wetted area as follows: 32,44,45,46 cos θ CB ) r f f cos θ 0 + f - 1, where f is the fraction of the projected area of the solid surface wetted by water (f lv ) 1 - f ) and r f is the surface roughness factor of the wetted area (r f g 1). By introducing a layer of microparticles to the fiber surface, we showed previously that indeed r f can be increased for a PDMSmodified cotton fiber, leading to superhydrophobic cottons. 32 With the surface perfluoroalkyl modification, the cotton sample was turned highly oleophobic, 32 as demonstrated by a C 16 H 34 CA of 135 (also shown in this study), but not superoleophobic. It has become obvious that, by incorporating a second layer of nanoparticles to the microparticle-covered fiber in the current investigation, r f can be further increased to an even higher level, which allows the wetting by C 16 H 34 to be in the Cassie regime, as clearly manifested by the high static CA and the small roll-off angle for C 16 H 34. The creation of a triple-length-scale roughness (the woven fiber, microparticle, and nanoparticle) has proven to be essential to achieving superoleophobicity. The data in Tables 1 and 2 indicate that the number of adsorption cycles for silica nanoparticles, in spite of increasing the surface coverage of the microparticle-covered fiber by the nanoparticles (Figure 1 and Supporting Information Figure S1), does not have major effect on the CAs and roll-off angles for both water and C 16 H 34. This suggests that the triple-size surface roughness from only one cycle of nanoparticle adsorption is sufficient to allow both water and C 16 H 34 to reside in the Cassie wetting regime. In addition to the silica nanoparticles (diameter: 220 and 160 nm) used for samples C b and C a, even smaller nanoparticles (100 nm) were used to cover the microparticle-covered fiber surface (sample C c ). As shown in Figure 3, both large particles and the fiber surface were fully covered by the nanoparticles after one adsorption cycle. The surface wettability of sample C c was comparable to samples C a and C b (Tables 1 and 2). The size ratio between nanoparticle and microparticle ranges from 1/8 to 1/4 for samples C c,c a, and C b. A free-energy-based modeling in our laboratory on a dual-size structured surface (raspberry-like surface 7 ) suggests that as the size ratio between the small and big particles is less than 1/5, the impact of the size ratio on the surface wettability becomes negligible. The possibility of using different nanoparticles to achieve superhydrophobicity and superoleophobicity offers flexibility and ease from a fabrication standpoint; there is no necessity to use highly uniform nanoparticles or microparticles. Conclusions In summary, we have successfully obtained superlyophobic surfaces on the basis of cotton textiles by introducing a nano/ microparticle dual-size structure to the woven fiber network followed by surface perfluorination. The modified textiles were completely nonwettable by both water and hexadecane, which both showed high contact angles and low roll-off angles. To achieve superoleophobicity, the presence of a triple-length-scale surface roughness has proven to be essential. This approach can be easily extended to using functional polymeric microparticles and nanoparticles, which can be readily synthesized by dispersion and emulsion polymerization, respectively, as building blocks to prepare triple-length-scale structured surfaces. Acknowledgment. We thank the China Scholarship Council for supporting B.L. s stay at Eindhoven. Z.S. acknowledges the financial support from the National Natural Science Foundation of China (NSFC ) and Chinese Ministry of Science and Technology (973 Project No. 2009CB930000). B.L. thanks Chunxia Sun for her kind help with TGA measurements and Ming Yuan for valuable discussions during the manuscript preparation. Supporting Information Available: Figure S1, morphology, and TGA curves of C a samples, as well as images of water (5 µl) and C 16 H 34 (5 µl) droplets on the corresponding samples modified by Rf-Si. Two video clips (H2O.wmv and C16.wmv) showing that, when the sample is lifted out of two liquids (water and hexadecane), the liquid contact line retreats very rapidly from the top edge of the modified cotton. This material is available free of charge via the Internet at LA (45) Marmur, A. Langmuir 2003, 19, (46) Shirtcliffe, N. J.; McHale, G.; Newton, M. I.; Perry, C. C. Langmuir 2005, 21,