Contact Self-Cleaning of Synthetic Gecko Adhesive from Polymer microfibers

Contact Self-Cleaning of Synthetic Gecko Adhesive from Polymer microfibers

(Parte 1 de 2)

Contact Self-Cleaning of Synthetic Gecko Adhesive from Polymer Microfibers

Jongho Lee*,† and Ronald S. Fearing‡

Department of Mechanical Engineering and Department of Electrical Engineering and Computer Sciences, UniVersity of California, Berkeley, California 94720

ReceiVed July 7, 2008. ReVised Manuscript ReceiVed August 21, 2008

Natural gecko toes covered by nanomicro structures can repeatedly adhere to surfaces without collecting dirt.

Inspired by geckos, we fabricated a high-aspect-ratio fibrillar adhesive from a stiff polymer and demonstrated selfcleaning of the adhesive during contact with a surface. In contrast to a conventional pressure-sensitive adhesive (PSA), the contaminated synthetic fibrillar adhesive recovered about 3% of the shear adhesion of clean samples after multiple contacts with a clean, dry surface.

Conventional pressure-sensitive adhesives (PSA) use soft viscoelastic polymers (Young’s modulus <100 kPa measured at 1H z1-3) to make intimate contact with surfaces to achieve high adhesion. However, soft polymers tend to collect dirt and lose adhesion with repeated use. In contrast, a gecko uses millions of keratinous nano and microhairs (Young’s modulus E ≈ 1.5 GPa3,4) to cling to and walk on virtually any surface. These hairs shed dirt particles during contact with a surface, keeping its natural adhesive sufficiently clean to support the gecko’s body weight.5

A key factor in the self-cleaning ability of gecko structures is the nonadhesive default state exhibited by the gecko fibers.6 To adhere, the fibers need to be dragged to expose the spatula tips, increasing the contact fraction by approximately 7.5-fold.6 In contrast to the well-known lotus effect7 in which particles are removedfromanonadhesiveandhighlyhydrophobicsurfaceby water droplets, gecko setae self-clean particles during use, even on dry surfaces. We restrict our discussion here to the selfcleaning of adhesives on dry surfaces during use. Natural gecko setae are the only previously reported self-cleaning adhesive on dry surfaces.

Recently, gecko-inspired synthetic adhesives (GSAs)8 have been fabricated using soft polymers (Young’s modulus e10 MPa)9-14 or hard polymers15-18 (Young’s modulus g1.5 GPa). Also,arraysofcarbonnanotubes(CNT)havebeenusedtoachieve adhesion.19-21Fibrillaradhesivecleaninghasbeendemonstrated usingwater16,22andmechanicalvibration.22Superhydrophobicity mayleadtothecleaningoffibrillaradhesivebywater.23However, no synthetic adhesive has demonstrated self-cleaning on dry surfaces during use, one of the important advantages of a geckoinspiredadhesiveoverconventionalpressure-sensitiveadhesives.

Autumn24 has identified seven benchmark properties that are characteristic of geckolike adhesives, which are (1) anisotropic attachment,(2)ahighadhesioncoefficient,(3)alowdetachment force, (4) material-independent adhesion, (5) self-cleaning, (6) anti-self-adhesion, and (7) a nonsticky default state. Although properties 1-4 and 7 have been previously demonstrated25,26 in a single material, in this letter we report the first geckolike microfibrillarmaterialthatalsodemonstratesself-cleaningduring contact.

Tocreateaself-cleanableadhesive,wefabricatedhigh-aspectratio fibrillar arrays from polypropylene (Young’s modulus E ≈ 1.5GPa,measuredwithSintechtensiletester2/S,MTSSystems). In previous work, these hard-polymer-based fibrillar materials haveshownuniqueadhesionproperties,similartothoseofgecko setae, including sliding enhanced shear adhesion25 with low peelingforceandfrictionaladhesion27withasphericalindenter.28 In this letter, we use a contact “step” protocol similar to that used for natural gecko setal arrays5 to demonstrate self-cleaning of the synthetic fibrillar adhesive. The self-cleaning synthetic

* To whom correspondence should be addressed. E-mail: jongho@ eecs.berkeley.edu.

† Department of Mechanical Engineering.

‡ Department of Electrical Engineering and Computer Sciences. (1) Dahlquist, C. A. In Treatise on Adhesion and AdhesiVes; Patrick, R. L. Ed.;

Dekker: New York, 1969. (2) Pocius, A. V. Adhesion and AdhesiVe Technology, Hanser: Munich, 2002;

Chapter 3. (3) Autumn, K.; Majidi, C.; Groff, R. E.; Dittmore, A.; Fearing, R. J. Exp.

Biol. 2006, 209, 3555. (4) Peattie, A. M.; Majidi, C.; Corder, A.; Full, R. J. J. R. Soc. Interface 2007, 4, 1071. (5) Hansen, W. R.; Autumn, K. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 385. (6) Autumn, K.; Hansen, W. J. Comp. Physiol., A 2006, 192, 1205. (7) Barthlott, W.; Neinhuis, C. Planta 1997, 202,1 . (8) Autumn, K.; Gravish, N. Philos. Tran. R. Soc. A 2008, 366, 1575. (9) Sitti, M.; Fearing, R. S. J. Adhes. Sci. Technol. 2003, 18, 1055. (10) Gorb, S.; Varenberg, M.; Peressadko, A.; Tuma, J. J. R. Soc. Interface 2006, 4, 271. (1) Kim, S.; Sitti, M. Appl. Phys. Lett. 2006, 89, 26191. (12) Murphy, M. P.; Aksak, B.; Sitti, M. J. Adhes. Sci. Technol. 2007, 21, 1281. (13) Crosby, A. J.; Hageman, M.; Duncan, A. Langmuir 2005, 21, 11738. (14) Glassmaker, N. J.; Jagota, A.; Hui, C.-Y.; Noderer, W. L.; Chaudhury, M. K. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 10786.

(15) Geim,A.K.;Dubonos,S.V.;Grigorieva,I.V.;Novoselov,K.S.;Zhukov,

A. A.; Shapoval, S. Yu. Nat. Mater. 2003, 2, 461. (16) Kustandi, T. S.; Samper, V. D.; Yi, D. K.; Ng, W. S.; Neuzil, P.; Sun,

W. AdV. Funct. Mater. 2007, 17, 21. (17) Northen, M. T.; Turner, K. L. Sens. Actuators, A 2006, 130, 583. (18) Schubert, B.; Majidi, C.; Groff, R. E.; Baek, S.; Bush, B.; Maboudian,

R.; Fearing, R. S. J. Adhes. Sci. Technol. 2007, 21, 1297. (19) Zhao, Y.; Tong, T.; Delzeit, L.; Kashani, A.; Meyyappan, M.; Majumdar,

A. J. Vac. Sci. Technol., B 2006, 24, 331. (20) Ge, L.; Sethi, S.; Ci, L.; Ajayan, P. M.; Dhinojwala, A. Proc. Natl. Acad.

Sci. U.S.A. 2007, 104, 10792. (21) Qu, L.; Dai, L. AdV. Mater. 2007, 19, 3844. (2) Sethi, S.; Ge, L.; Ci, L.; Ajayan, P. M.; Dhinojwala, A. Nano Lett. 2008, 8, 822. (23) Bhushan, B.; Sayer, R. A. In Microsys. Technol. 2007, 13,7 1. (24) Autumn, K. In Biological AdhesiVes; Smith, A. M., Callow, J. A., Eds.;

Springer Verlag: Berlin, 2006; p 225-255. (25) Lee, J.; Majidi, C.; Schubert, B.; Fearing, R. S. J. R. Soc. Interface 2008, 5, 835. (26) Kim, S.; Spenko, M.; Trujillo, S.; Heyneman, B.; Santos, D.; Cutkosky,

M. R. IEEE Trans. Robot. 2008, 24,6 5. (27) Autumn, K.; Dittmore, A.; Santos, D.; Spenko, M.; Cutkosky, M. J. Exp.

Biol. 2006, 209, 3569. (28) Schubert, B.; Lee, J.; Majidi, C.; Fearing, R. S. J. R. Soc. Interface. 2008, 5, 845.

10.1021/la8021485 C: $40.75 2008 American Chemical Society Published on Web 09/10/2008

adhesive should be useful in a variety of applications where conventional adhesives can be easily contaminated.

The fibrillar adhesives were fabricated by casting a single layer of a 25-µm-thick polypropylene (P) film (TF-225-4, Premier Laboratory Supply Inc.) in a vacuum oven at 200 °C into a 20-µm-thick polycarbonate (PC) track-etched membrane filter(ISOPORE,MilliporeInc.)containing300-nm-radiuspores as illustrated in Figure 1A. Using a fixed fiber length, this fiber radius was selected to provide bending compliance while preventing fibers from clumping. The polycarbonate filter was etched completely for 10 min in a first bath and 5 min in a second bath of methylene chloride (Figure 1C) to release the polypropylene fibrillar surface and film. The resulting samples were rinsed in isopropyl alcohol and air dried (Figure 1D). The polypropylene film contains approximately 42 million fibers per squarecentimeterwiththeaveragelengthandradiusofthefibers being 18 µm and 300 nm, respectively. The microstructured polypropylene film was cut into 2 cm × 2.5 cm rectangles using a razor blade, anda2c m × 0.5 cm × 0.05 cm load bar with a small hole in which a string goes through was attached to distribute the pulling force uniformly.

To simulate contamination with dirt particles, microspheres with a mean radius of 1.15 µm (gold powder, spherical, radius e2.5 µm, Alfa Aesar) were applied to cover the whole area of fibrillaradhesivesandconventionalpressure-sensitiveadhesives by freely dropping microspheres from about 5 cm above the adhesives. (Au microspheres were supplied in dry powder form with only weak clumping. Au microspheres were applied uniformly with similar density on the PSA and fibrillar surfaces bygravity,withoutapplyinganycontactforce.)Afterapplication, the adhesives were gently shaken to remove excess microsphere particles.AsshowninFigure2A,C,microspheresinitiallycovered most of the area.

The samples were tested using a “simulated step” protocol showninFigure3similartoagecko’swalkingstep.Thesamples were first compressively loaded (<1 N/cm2) onto a clean glass substrate manually with a gloved finger (Figure 3A). (It has been shown previously that the shear strength is independent of the initial normal preload.25) The samples were next loaded parallel to the glass substrate by a weight attached to the load bar through a string (Figure 3B), and then the normal load was removed while maintaining the parallel load (Figure 3C). If the sample could hold the weight, then we removed the sample from the substrate manually (Figure 3D) and increased the weight for the next step. If the sample could not hold the weight, then the sample fell and was caught by a gloved hand just below the weight.Incaseoffailuretosupporttheweight,weusedthesame weight for the next simulated step. Before each simulated step, the glass substrate was cleaned with isopropyl alcohol to remove residual particles. After 30 simulated steps, the fibrillar adhesive shed about 60% of the microspheres onto the glass substrate as shown in Figure 2B. Some microspheres remained embedded between fibers and were not self-cleaned. As a control, we used a 0.2 cm × 0.5 cm conventional pressure-sensitive adhesive (PSA) (Scotch Magic Tape, 3M). After the simulated steps, the soft polymer of the conventional PSA was almost completely coveredbymicrospheres,asshowninFigure2D.Thisispossibly because microspheres not in direct contact with the soft polymer are taken off and recaptured in the exposed area of the soft polymer during simulated steps.

Figure 1. Schematic illustration of the fabrication process of polypropylene fibrillar adhesives. (A) A polypropylene (P) film was cast into a polycarbonate (PC) template for 28 min in a vacuum oven. (B) The casted P film and PC template cooled down to room temperature for 30 min. (C) The PC template was etched completely for 10 min in a first bath and 5 min in a second bath of methylene chloride (MC). (D) The resulting sample was rinsed in isopropyl alcohol and air dried. A string-connected load bar was attached to distribute the pulling force evenly.

Figure 2. Scanning electron micrograph images of the polypropylene fibrillar adhesive and conventional pressure-sensitive adhesives (PSA). (A) Fibrillar adhesive contaminated by gold microspheres. (B) Fibrillar adhesive after 30 contacts (simulated steps) on clean glass substrate. (C) Conventional PSA contaminated by microspheres. (D) Conventional PSA after contact on a clean glass substrate. All scale bars correspond to 10 µm. Microspheres on fibrillar adhesive are removed by simulated steps, but microspheres on PSA cover more area after the steps.

Figure 3. One cycle of simulated steps, with contact with an initially clean glass slide. (A) Applying normal compressive force. (B) Shear load added to the compressive load by a hanging weight. (C) Removing the compressive load (pure shear loading). (D) Detaching the sample.

10588 Langmuir, Vol. 24, No. 19, 2008 Letters

To quantify the self-cleaning capability of the adhesives, the shearadhesionstrengthwasmeasuredbyapplyingaloadparallel to the glass substrate during every simulated step as shown in Figure 3C. (The normal compressive load is zero during this phase of the testing cycle; this is not a friction test.) With no contamination, both fibrillar adhesives and PSAs could sustain a 4 N load parallel to a glass substrate (precleaned microscope slides, Fisher Scientific). (We limited the shear force to 4 N to prevent plastic deformation or tearing of the samples’ thin backing.) After the samples were contaminated, the initial shear loadtriedwas0.2N.Thisshearloadwastestedateverysimulated step until the sample could sustain it (eight simulated steps for fibrillar adhesive sample 1). Once the sample could sustain this load,theshearloadwasincreasedby0.1Nforthenextsimulated step. Following 30 successive simulated steps, fibrillar adhesive sample 1 could sustain a shear load of 1.0 N but did not show further improvement with five more simulated steps. This saturationisconsistentwiththequantityofmicrospheresdeposited on the glass substrate after each step, as shown in the bottom three images in Figure 4. Initial contact steps left many microspheres on the clean glass substrate, with diminishing particle removal after further steps. As expected, the PSA contaminatedbymicrospheresdidnotrecoveranyshearadhesion and could not sustain 0.05 N, even after 35 steps.

The synthetic fibrillar adhesives did not self-clean larger particlesduringcontact.Toobservetheself-cleaningdependence on particle size, four differently sized polystyrene microspheres, 1.5, 2, 3 (also containing 12% 5 µm), and 5 µm in radius (Corpuscular Inc.), were used as dirt particles. Unlike gold microspheres, dry polystyrene microspheres contained lumps of microspheres. To obtain single-sized microspheres, we mixed dry polystyrene microspheres in isopropanol in about a 1:30 ratio. The mixture was ultrasonically agitated (2510 Branson) for10mintoseparatelumps.Then,severaldropsofmicrospheres in an isopropanol suspension were deposited on a clean glass slide and air dried. Air-dried polystyrene microspheres became approximatelysingle-layered.Thepolystyrenemicrosphereswere transferred to fibrillar adhesives by dragging adhesive samples on the glass slide covered with the single-layer microspheres. The shear adhesive strength of clean samples before being contaminated with polystyrene microspheres was 4 N. Prior to reuse, fibrillar samples were cleaned by removing clogged microspheres in an isopropanol bath with an ultrasonic cleaner (2510 Branson) for 2 min. After ultrasonic cleaning, samples could again hold4No f shear.

Contaminated samples with uniformly sized polystyrene microspheresweretestedwiththesamemethodsforgoldparticles as described in Figure 4. After typically 20-25 simulated steps, samplescontaminatedwith1.5-µm-radiusmicrospheresrecovered about 34% (SD ) 13%, three arrays, six measurements) of the shear force of uncontaminated samples, as shown in Figure 5. Samplescontaminatedwith2-µm-radiusmicrospheresrecovered about 29% of the shear force of uncontaminated samples (SD ) 9%, three arrays, six measurements) after 20-25 steps. However,samplescontaminatedwith3-and5-µm-radiusparticles could not sustain 0.2 N in shear (5% of the shear force of uncontaminatedsamples,threesamples,sixmeasurements)even after 25 steps.

Thecontactself-cleaningofthefibrillaradhesivesdemonstrated above is consistent with a greater affinity of the microspheres for the glass substrate than for the fibers as described for natural gecko setae.5 The self-cleaning of natural gecko setae5 was explained by comparing attraction forces and energies acting on a microsphere in contact with spatulae and a glass substrate. Hansen and Autumn5 argue that the small number of spatulae

Figure 4. Steps on clean glass and recovered shear adhesion. Clean samples could hold 4 N shear force. Samples contaminated by gold particles(meanradius1.15µm)recoveredupto33%oftheshearadhesion of clean samples. (×) Indicates a shear force that could not be sustained bytheadhesive;(O)indicatestheshearforcethatwassustained.Fibrillar samples 1-3 are individual samples fabricated by the same methods. Bottom images: optical images showing the whole contact area after each simulated step (1 cm scale bars). (MS) Microspheres deposited on a glass substrate by fibrillar sample 1 at each step. The quantity of microspheres deposited on the glass decreases with increasing step number.

Figure5.Cleaningperformancebymicrospheresize.Whenthesamples were contaminated with 1.5- and 2-µm-radius microspheres, the shear adhesive force recovered to 3 and 29%, respectively, of the clean value after 20-25 steps. For larger particles (3 and 5 µm), the adhesive force did not recover.

Letters Langmuir, Vol. 24, No. 19, 2008 10589 contacting a spherical particle have less net adhesive force than particle adhesion to a flat substrate. We use a similar argument as with Hansen and Autumn5 for self-cleaning using the Johnson-Kendall-Roberts(JKR)contactmodel29andreported surface energies.30,31 Neglecting surface roughness, we can estimate the adhesion forces from the JKR model.29 The sphere-glass pull-off force is

Fsg) 32πRsWsg and the sphere-fiber pull-off force is

Fsf) 32π Rf Rs Rf+Rs

Wsf with mean radius Rs ) 1.15 µm for gold microspheres (Alfa

Aesar) and fiber radius Rs ) 0.3 µm. The work of adhesion is estimated with

Wsg≈2√γsγg and

(ref 32) where the surface energy is γg ) 115-200 mJ/m2 for

SiO230 and γf ) 30 mJ/m2 for polypropylene.31 The ratio of pull-off forces is

N) Fsg

(Parte 1 de 2)

Comentários