bioinspired functionally graded adhesive materials: synergetic interplay of top viscous–elastic...

Bioinspired Functionally Graded Adhesive Materials: SynergeticInterplay of Top Viscous−Elastic Layers with Base MicropillarsHamed Shahsavan and Boxin Zhao*

Department of Chemical Engineering and Waterloo Institute for Nanotechnology, University of Waterloo, 200 University AvenueWest, Waterloo, Ontario, Canada N2L 3G1

*S Supporting Information

ABSTRACT: Inspired by the amazing adhesion abilities of the toe pads of geckos and tree frogs, we report an experimentalstudy on the integration of a dissipative material (resembling the dissipative and wet nature of the tree frog toe pads) to an elasticfibrillar interface (resembling the dry and fibrillar nature of the gecko foot pads). Accordingly, a new type of functionally gradedadhesive is introduced, which is composed of an array of elastic micropillars at the base, a thin elastic intermediate layer and aviscoelastic top layer. A systematic investigation of this bioinspired graded adhesive structure was performed in comparison withthree control adhesive materials: a viscoelastic film, a viscoelastic film coated on a soft elastomer, and elastic film-terminatedmicropillars. The results showed that this graded structure bestows remarkable adhesive properties in terms of pull-off force,work of adhesion, and structural integrity (i.e., inhibited cohesive failure). Moreover, an extraordinary compliance was observed,which is attributed to the polymer slippage at the top layer. Overall, we attribute the improved adhesive properties to thesynergetic interplay of top viscous−elastic layers with the base biomimetic micropillars.

1. INTRODUCTION

Adhesion is an interesting phenomenon in nature, and polymeradhesives are widely used in many manufacturing processes.The importance of adhesion and adhesives is continuouslyincreasing for the development and manufacturing of advancedmaterials at ever-smaller scales. Other than the permanentstructurally adhesive bonding, effective soft and temporaryadhesive materials are in high demand for both consumerproducts (post-it notes and medical bandage) and the assemblyand packaging of functional elements in electronic devices. Thelocomotive organs of the geckos and tree frogs are outstandingexamples of the temporary adhesive materials. The amazingaptitude of these creatures to stick readily and rapidly tosurfaces has attracted extensive research interests on thedevelopment and application of biomimetic structures.The adhesive functionalities of the gecko toe pads are rooted

mainly in their sophisticated micro/nano structures, whichutilize the universal van der Waals intermolecular interactionsand/or capillary forces to generate sufficient adhesive forces.1−4

The hierarchical micro- to nanostructures on the gecko toepads maximize the compliance, contact area, and accordinglyeffective van der Waals forces with the mating surface.5 As aresult of more than a decade of extensive studies, syntheticbiomimetic fibrillar adhesives with a variety of geometries andmaterials have been fabricated. The first class of these structuresis based on the simple surface protrusions such as pillars orposts,6−8 some with different tip shapes, e.g., mushroom-shapedpillars.9,10 These have been fabricated in either single or multilevels from nano to micrometer scales.11,12 Another class ofbiomimetic structures is based on incisions underneath the freesurface of a material. Film-terminated microfibrils,13,14 bridgedmicrofibrils,15 and capped microfluidic channels have beenfabricated in single or multi levels of hierarchy.16 Thus far,mushroom-shaped, film-terminated, and more recently bridged

Received: September 9, 2013Revised: December 13, 2013Published: December 24, 2013

Article

pubs.acs.org/Macromolecules

© 2013 American Chemical Society 353 dx.doi.org/10.1021/ma4018718 | Macromolecules 2014, 47, 353−364

pubs.acs.org/Macromolecules

pillars are introduced as the most effective prototypes. Thesesynthetic structures have shown to increase adhesive strengthand toughness of a flat control by combination of mechanismssuch as contact splitting,17 enhanced compliance,18 and cracktrapping.13 In contrast to geckos, the toe pads of the tree froghave relatively simpler structure, consisting of hexagonalmicroposts separated by grooves or channels which are coveredby viscous fluids. The exact mechanism of adhesion of treefrogs is not fully understood but several mechanisms have beenproposed such as capillary, friction, and viscous forcescontributing to the total adhesion. For instance, it is believedthat microscale hexagonal grooves and channels regulate themeniscus forces associated with the liquids secreted.19 Waterdrainage in contact has been offered as of one mechanism ofthe adhesion in wet condition. Thus, the adhesive ability ofsuch systems may be rooted in a combination of the mentionedmechanisms.20

Inspired by the amazing adhesion abilities of the toe pads ofboth gecko and tree frog and their surface structures, a fewrecent publications reported the integration of a dissipativematerial (like the “wet” nature of the tree frog toe pads) to bothelastic flat and fibrillar interfaces (like the fibrillar nature ofgecko foot pads). Although not explicitly stated, the effect ofchemical, mechanical, and geometrical gradient on the adhesionproperties of such structures has been studied. For instance,Cheung et al. demonstrated a simple method to enhanceadhesion by deposition of a thin layer of silicon oil on an arrayof a biomimetic fibrillar interface. The enhanced adhesion wasattributed mainly to the viscous and induced capillary forcesduring the debonding.21 In a similar attempt, Patil and co-workers coated a layer of liquid PDMS solution with a lowcontent of cross-linker on an array of microposts and foundenhanced adhesion with limited magnitude of deformation andless cohesive separation.22 In both of these works, the fibrillarinterface is immersed in the viscous counterpart. In addition,Carelli et al studied the effect of composition gradient in aviscoelastic bilayer during a debonding process. They foundthat the effect of gradient in composition of viscoelastic layerson adhesion and the mode of failure is complex, depending onthe nature of probe.23 Similarly to this work, Patil and co-

workers used an elastic PDMS skin to coat a viscoelastic layerof the same material to improve reusability and retain adhesivefracture.24

To the best of our knowledge, the possible synergetic effectof the bioinspired fibrillar interfaces and their viscoelasticcounterpart on the adhesive properties in a singular gradedmaterial have not been studied. We recently reported thefabrication of a hybrid adhesive structure consisting of an arrayof elastic biomimetic micropillars terminated with a thinpressure sensitive adhesive film. Localized separation insta-bilities along with more dissipative nature of the top layer inthat structure enhanced the adhesion remarkably.14,25 Theobserved adhesion enhancement of the adhesive film on top ofthe micropillars suggested a significant synergetic interactionbetween the top layer and subsurface micropillars.The objective of this paper is to systematically investigate the

observed synergetic effects in a hybrid adhesive systemconsisting of a viscoelastic top layer and micro fibrillar base.The main difference between our structure and ones previouslyreported is the nature of the backing material and its interactionwith a more dissipative top layer in a nonimmersed geometry.Systematic studies of three comparative adhesive structures, i.e.,viscoelastic film, viscoelastic film coated on a soft elastomer,elastic film-terminated micropillars were performed in parallelto the newly developed graded adhesive structure to acquirefundamental insights in the context of current understandingsof viscoelastic polymer adhesion and biomimetic adhesion. Theresearch findings in this work depict a new outlook to thequestion of “the combined role of shear and normal forces onthe compliance, adhesion and friction of fibrillar structures”,26

which is an unsolved fundamental question in the field ofbiomimetic adhesion.

2. EXPERIMENTAL SECTIONPDMS (Sylgard 184, Dow Corning) resin and curing agent were usedin all adhesive structures. Schematic pathway of fabrication of thegraded adhesive structure is shown in Figure 1. Hexagonal arrays ofPDMS micropillars were fabricated through the soft-lithographytechnique. A 10:1 by weight mixture of PDMS (Sylgard 184, DowCorning) resin and curing agent were poured on a master-mold ofmicroholes fabricated by DRIE process. The PDMS coated master-

Figure 1. Schematic view of the fabrication pathway including the three key steps of soft-lithography and microinking, which are followed by coatingof a viscoelastic layer on top of the elastic film-terminated micropillars.

Macromolecules Article

dx.doi.org/10.1021/ma4018718 | Macromolecules 2014, 47, 353−364354

mold was cured at 90 °C for 1 h and the PDMS was peeled off themaster-mold after curing. The resultant positive structures were pillarswith 50 μm diameter, 150 μm height and 115 μm center-to-centerspacing (surface density of 17%). The final thickness of the curedPDMS backing film was controlled to be 1000 ± 100 μm.The dipping method developed in ref 13 was used to fabricate thin

film-terminated pillars. A thin layer of PDMS was spun on a lowsurface energy microscope glass slide. The reduction of the surface freeenergy was achieved by coating a self-assembled monolayer ofheptadecafluoro-1,1,2,2,tetrahydrodecyltrichlorosilane on a glass slideas described in ref 27. The spin speed and time determine the finalthickness of the terminal film. Fabricated micropillar arrays wereplaced upside-down on the thin PDMS layer and kept in roomtemperature overnight to ensure the contact between all pillars’ tipsand the thin film. The entire system was placed in the oven at 90 °Cfor 1hr and the cured sample was peeled off gently. Thickness of theterminal films, which was varied from 8 to 24 μm, measured byweighting the samples and confirmed by the optical interferometry.Coating of the elastic film-terminated fibrillar structures with theviscoelastic top layer was performed using the spin coating technique.Again, the thickness of the terminal viscoelastic layer was adjustedusing spin speed. Weighting of the samples and optical interferometrywere used to estimate the thickness of the terminal viscoelastic filmwhich varied from 18 to 50 μm.Oscillatory frequency sweep and shear relaxation tests were

performed using a RA2000 model rheometer (TA Instruments) tocharacterize the thermo mechanical properties of 100:10, 100:4, and100:1.6 w/w PDMS mixtures. PDMS mixtures were poured on thebottom plate of a plate−plate geometry and cured at 90 °C for 1 h.The instrument environment was then cooled down to roomtemperature to perform the oscillation. The maximum oscillationamplitude was set to 3% to avoid any nonlinear response. The timedependent shear modulus (G(t)) of the sample was measured in ashear relaxation test for 30 min. The obtained storage (G′) and loss(G″) moduli of the samples were recorded in order to estimate of thelinear rheological properties of the terminal film.Adhesive properties of all samples were characterized by the

indentation of a 6 mm diameter hemispherical fused silica probe(Ispoptics Co., New York) using a microindenter assembled on top ofan inverted optical microscope. The load−displacement data weremeasured with a 0−10 g load-cell (GSO-10, Transducer Techniques).The contact area and deformations during loading and unloadingprocesses were visualized from the bottom-view images. The setup wasequipped with a side-view camera to monitor the possible neckingduring separation and the mode of separation. The loading velocity inindentation tests was maintained 1 μm/s for all of the samples. Uponreaching the preload force, a slight decrease in the force and an

expansion of the contact area were observed due to the stressrelaxation in the viscoelastic layer. The contact time was set to 30 s,where the expansion of the contact area leveled off. The holding timewas also much greater than the effective shear relaxation time of thesample obtained from rheology tests.

3. RESULTS AND DISCUSSION3.1. Fabrication of the Biomimetic Functionally

Graded Adhesives. We use PDMS materials to make thebiomimetic structures because PDMS has low surface energyand adhesion and is commonly used as a releasing agent. Theviscoelasticity of the PDMS can be readily changed by varyingthe curing agent to resin ratio, making it a good model systemto investigate the fundamentals of the complex gradedmaterials. Previous works showed that a flat film of 100:10w/w PDMS can be deemed purely elastic having very lowadhesion. Introduction of the biomimetic micropillars withdifferent tip shapes, e.g., mushroom tipped, bridged, or film-terminated pillars, bestowed nonadhesive elastic PDMSremarkable adhesive properties.9,10,13−15 Recently, we reportedan enhancement of pressure-sensitive adhesives using a thinPDMS film-terminated PDMS fibrillar interface. In that work,the curing agent to resin ratio of the terminal film was set to100:4 w/w. Although more dissipative than 100:10 w/wPDMS, 100:4 w/w PDMS was elastic rather than viscoelas-tic.14,28 We speculated that using more dissipative PDMS as theterminal layer would result in better adhesive properties interms of higher adhesion energy and pull-off force. To make amore viscous layer, we varied the concentration of PDMSmixtures from 100:4 w/w to 100:1.2 w/w. We limited our studyto solid-like state of the matter to avoid excessive complicationsin the study of liquids in contact.Figure 2a shows a schematic view of the new hybrid graded

adhesive structure. It consists of three components: aviscoelastic top layer, an elastic intermediate layer and elasticbiomimetic micropillars. h′ is the thickness of the elastic and hrepresents the thickness of the viscoelastic terminal layer.Figure 2b shows a viscoelastic layer coated on the glasssubstrate (deemed as a simple nongraded structure); Figure 2cshows a viscoelastic low cross-linked PDMS layer coated on afully cross-linked PDMS film (deemed as a chemically gradedstructure); and, Figure 2d shows an elastic film-terminatedfibrillar adhesive (deemed as a geometrically graded structure).

Figure 2. Schematic view of the (a) bioinspired functionally graded adhesive B-E12-VE50, control samples, (b) viscoelastic layer on glass VE50, (c)viscoelastic layer on the polymer P-VE50, and (d) the elastic-film terminated biomimetic fibrillar adhesive B-E12.



For clarity and brevity, we designate the adhesive samples bythe immediate backing material to the terminal layers (i.e., P forthe flat PDMS film, and B for the biomimetic fibrillar adhesive),elasticity (E) or viscoelasticity (VE), and the thickness of eachterminal layer (h or h′). For instance, B-E12-VE50 represents agraded adhesive structure consisting of a terminal viscoelasticlayer of 50 μm in thickness backed with a biomimetic fibrillarinterface with intermediate elastic layer of 12 μm. Note that aglass slide is employed as the supporting substrate for all theadhesive samples and has not been considered as a part of theadhesives.The viscoelastic properties of PDMS with varied degree of

cross-linking were characterized. Figure 3a demonstrates thevariation of both storage and loss components of the modulus(G′ and G″) against content of cross-linking agent for threemixtures of 100:10, 100:4 , and 100:1.6 w/w during a frequencysweep rheology test. The storage modulus for both 100:10 and100:4 w/w is almost independent of the frequency, while itsvariation against frequency is more pronounced for the 100:1.6w/w sample. Also, the loss modulus of the 100:10 w/w PDMSis almost 3 orders of magnitude greater than that of 100:1.6 w/w mixture. The damping factor tan δ of the 100:1.6 w/wmixture is higher than that of both 100:10 and 100:4 w/wmixtures, indicating its more viscoelastic or dissipative nature.Figure 3b shows the variation of the shear modulus against timefor 100:1.6 w/w PDMS mixture. The relaxation test wasperformed under 3% shear strain for 1800 s. The shear modulusregime change took place at t* ≪ 30 s. In fact, 30 s holding

time is quite larger than the stress relaxation time and thus thepolymer can be assumed to be fully relaxed.The 100:1.6 w/w PDMS mixture was chosen to make the

top layer because it is close to the transient gel point of thePDMS and it can be assumed fairly as a viscoelastic material.However, the micropillars could not be topped easily throughthe dipping method due to the cohesive peeling of the lowcross-linked PDMS from the glass substrate. The solutiondeveloped here was to coat an elastic thin film-terminatedfibrillar structure with the liquid PDMS. Figure 4a is obtainedfrom optical microscopy of the sample B-E12, in which theterminal film and micropillars are easily distinguishable. Surfacedefects in the form of cavities were sporadically located on theelastic terminal film. Figure 4b represents the SEM micrographof the sample B-E12. It seems the presence of the slightlyrippled surface on the terminal layer is inevitable as a result ofcapillary rise of the liquid PDMS within the micropillars. Suchsurface instability must be minimized or covered well to avoidany undesirable effect on the adhesion test. Undesired effects ofsuch instabilities on the adhesion have been reported byNadermann et al.29 These surface instabilities can be replicatedon the viscoelastic layer if the thickness of the terminalviscoelastic layer is below a critical amount. A detailed study onthe effects of such surface instabilities on the contact behaviorhas been reported in ref 30. Herein, we minimized the extent ofthis surface instability by making a smooth terminal film. Forthis purpose, the viscoelastic terminal layer was chosen to be

Figure 3. (a) Rheological properties of the 100:10, 100:4, and 100:1.6 w/w PDMS mixtures and (b) variation of shear modulus against time during ashear relaxation test.

Figure 4. (a) Optical micrograph of the sample B-E12 with a surface defect on top right corner, and (b) SEM micrograph of the sample B-E12.



thick enough so that an approximately smooth surface can beachieved.An efficient viscoelastic adhesive must show strong adhesion

to a substrate surface and retain its cohesive strength during theseparation. This requires a careful design of the adhesive interms of both thermomechanical and geometrical properties,such as terminal layers thickness, the fibrils’ arrangement,spacing, and aspect ratio. The effect of geometrical featuressuch as spacing, fibrils’ length, and arrangement have beenstudied thoroughly in recent literature.31,32 It is shown, that theenergy release rate for an elastic film-terminated structure isstrongly proportional to w4, where w denotes the interfibrillarspacing and weakly scales with fibrils length, i.e., L−1/4, where Lis the fibrils length. Moreover, it has been shown that thehexagonal arrangement of the fibrils with the same surfacedensity as of the square arrangement results in strongeradhesion energy and pull-off force.31 For the hybrid gradedadhesive structures reported in this work, the thermomechan-ical and geometrical properties of the micropillars remainconstant; the varied parameters are the thickness and elasticityof each of terminal layers.In order to achieve the most efficient structure, we first varied

the thickness of the elastic terminal layer to obtain an optimumthickness. Figure 5 shows the variation of the pull-off force vs

preload for micropillars topped with different thicknesses of theelastic terminal layer. The pull-off force of elastic film-terminated structures increased with decrease of the thicknessfrom 24 to 12 μm over the entire range of preload. The sametrend was observed for sample B-E8 having the thinnestterminal layer but only at high preload. At low preload, the B-E8 has the lowest pull-off force and the separation instabilitiessuch as cavitation and fingering phenomena were observed ontop of the fibrils. In contrast, for other samples with thickerterminal films, separation instabilities were constrained to onlycrack trapping on the crack line. Thus, we chose sample B-E12as the elastic foundation for the terminal viscoelastic layer inorder to have the highest possible pull-off force while having nopronounced debonding instabilities on top of the fibrils. Ingeneral, we speculated that, the thinnest possible elasticterminal film topped with the thickest possible viscoelastic

layer will create the most efficient adhesive structure. A similarassumption has led to fabrication of efficient reusable PSAs withan elastic skin in the study by Patil et al.24 In that work, theadverse effect of the elastic skin shielding the PSA’s adhesionstrength and viscoelastic loss is attenuated by decreasing theelastic skin thickness. In our study, the elastic film is betweenthe micropillars and the top viscoelastic layer. Thus, it mayshield the desirable effect of the micropillars in enhancement oftotal compliance of the new hybrid structure.Figure 6 shows the variation of pull-off force vs preload for

viscoelastic layers of varied thickness coated on the sample B-

E12. The thicker the viscoelastic layer, the higher the adhesivepull-off force is achieved. The influence of the thickness of aviscoelastic layer on the adhesion properties in an axisymmetricprobe test experiments has been reported in only a fewreferences.14,33−35 Generally, it is believed that the thicker theviscoelastic layer the greater adhesion energy and the pull-offforce. However, to the best knowledge of the authors, notheoretical study on the effect of thickness was reported in theliterature thus far, perhaps because of the complexity ofdebonding instabilities. Note that the effect of thickness on theadhesive properties of the viscoelastic adhesives have been wellstudied for other testing geometries such as peeling geometry.34

Results show that there is an optimum thickness point forviscoelastic adhesives at which the adhesion energy and pull-offmaximizes. From the experimental evaluation of the thicknesseffects in Figures 5 and 6, the sample B-E12-VE50 was chosenas the hybrid graded adhesive structures to investigate theinterplay between the top viscous layer and the micropillarbase. Furthermore, the viscoelastic PDMS having the samethickness was coated on the glass and a thick (≈ 2 mm) flatPDMS elastomer film to obtain flat control samples, VE50 andP-VE50. The initial elastic film-terminated structure, B-E12,was chosen as the third control sample.

3.2. Indentation Behaviors of the Graded Adhesives.The load−displacement curves of the four types of adhesivestructures (B-E12-VE50, VE50, P-VE50, B-E12) are shown inFigure 7. There is no noticeable snap-in force for the elasticsample B-E12. For all other samples, bonding process startswith a pronounced snap-in force when probe approaches to thesurface because of the intermolecular surface forces. Following

Figure 5. Effect of thickness of the elastic terminal layer on the pull-offforce of film-terminated fibrillar adhesives.

Figure 6. Effect of thickness of the viscoelastic top layer on the pull-offforce of the new hybrid structure.



that, a normal compressive load (i.e., preload) of 0.5 mN wasapplied to enlarge the contact area. During unloading, theload−displacement curves were linear for a remarkable range ofretraction distance before reaching the pull-off point where thetensile adhesive force is the highest. The debonding for theelastic sample B-E12 is rapid and has several small zigzag steps,suggesting the presence of crack trapping mechanism duringthe separation. The debonding process of the viscoelasticsamples is gradual and smooth but undergoes a slope changebetween the pull-off point and the separation. This slopechange is indicative of the bulk deformation processes such asfibrillation, cavitation and other instabilities as discussed in latersection. The slop change is most pronounced for the viscousfilm on polymer (P-VE50), becomes less for the viscous film onglass (VE50), and is the least pronounced for graded structure(B-E12-VE50). This observation suggests the possible changeof crack mechanisms during the separation.To quantitatively characterize the crack mechanisms and

energy dissipation processes, we divided the unloading curvesinto two phases, as shown in Figure 8: (phase 1) crack initiationbefore the pull-off point and (phase 2) crack propagation afterthe pull-off point. The consumed energy for each phase gives acoarse idea about the sink of energy and the mode of separationin each system. Inset table in Figure 8 shows the ratio of theenergy required for each phase with respect to the energyrequired for the entire debonding process. The required energyfor the phase 2 of the sample P-VE50 is about 88%; the energydissipation processes include instabilities such as verticalfibrillation, fingering, and cavitation. As a result, this system isprone to experience cohesive failure and the extent of plasticdeformation is also higher than that of other samples. Therequired energy for the phase 2 of the sample B-E12-VE50 isabout 54%, which is much less than that of the sample P-VE50(88%) and even lower than that of sample VE50 (58%). Thecomparison between the relative amounts of energy dissipatedin the crack propagation revealed that less energy is consumedin the separation of the graded adhesive structure than that of asimple viscoelastic film. This situation renders the detachmentprocess to take place at the interface or at the adhesive mode ofseparation. Thus, the base micropillars inhibit the cohesivefailure of the viscous film, leading to an adhesive material with

higher structural integrity. To verify this analysis, we carefullyexamined the debonding process recorded in videos, accessiblein the Supporting Information, from the side-view and bottom-view. As can be seen in movie 1 in Supporting Information, themode of failure for the sample B-E12-VE50 is adhesive sincethere is no residue left on the probe. Although having a largeelastic deformation, the adhesive layer had experienced nodiscernible plastic/permanent deformation. In contrast, intensebulk deformation and vertical fibrillation were developed duringthe separation process for the sample P-VE50. In addition, weobserved cohesive failure for some of the samples. The extentof vertical fibrillation and residual plastic deformation left afterdetachment are shown in movie 2 in the SupportingInformation. The following section shows the detailedevolution of the contact area and its deformation andinstabilities during loading and unloading.

3.3. Evolution of Adhesive Contact Area andSeparation Instabilities. The evolution of the adhesivecontact and separation was observed and recorded by thebottom view camera showing a full indentation cycle on thesamples B-E12 (movie 3 in Supporting Information), P-VE50(movie 4 in Supporting Information), and B-E12-VE50 (movie5 in Supporting Information). The contact area during theloading was measured from the bottom view image as shown inthe first column of Figure 9. In our experiments, the preloadvaried from 0.5 to 10 mN. Dashed circles indicate the contactline at the preload point. Apparently, the sample B-E12-VE50always has the greatest contact area, suggesting this sample hasthe highest compliance, while the sample B-E12 has thesmallest contact area. The contact area for samples VE50 andP-VE50 are roughly comparable to the one for the sample B-E12-VE50 under the preload of 0.5 mN, but the differencebetween them starts to grow for higher preloads. Note that thecontact area increased with preload and became even largerthan the imaging frame for the viscous/elastic-pillar sample B-E12-VE50 under preloads higher than 4 mN. The similar trendswere observed for displacements δ at the preload. Displacementat the preload is the largest for the sample B-E12-VE50 (≈14

Figure 7. Typical load vs displacement curves for indentation test onfour different samples under 0.5 mN preload.

Figure 8. Division of two phases during the debonding of the sampleB-E12-VE50 under 0.5 mN preload. Phase 1: strain energy storage andcrack initiation. Phase 2: crack propagation processes. The inset tablelists the contributions of each phase in total energy dissipation duringthe debonding of the four adhesive structures.



μm for preload 0.5 mN). Samples VE50 and P-VE50 havealmost equal displacement at preload values (≈10 μm forpreload 0.5 mN), and the sample B-E12 under the samepreload has considerably smaller value than that of othersamples (≈6.8 μm). To a first approximation, the contact radius(a) is geometrically related to the depth of penetration or thedisplacement (δ) by a = (R2 − (R − δ)2)1/2, where R is theprobe radius. It is assumed that there is no meniscus raise of thePDMS around the contact with the probe. We compared theestimated values with the experimental values determined fromthe bottom view image. The measured contact area is alwayslarger than the calculated one by 1.2−1.3 times because of theviscoelastic nature of the materials. This method gives anestimate of the contact area for the graded structure B-E12-VE50 for preloads higher than 4 mN. This method has beenused by Paiva et al as well.36

Second column of Figure 9 shows the bottom view images ofthe contact at the pull-off point; third column shows thebottom view images of the contact at an arbitrary point afterpull-off. As expected, during the unloading, significant amountof fingering instabilities and undulations were developed forsamples VE50, P-VE-50, and B-E12-VE50 but not for the elasticsample B-E12. Comparing the three viscoelastic samples, theextent of undulation for the sample B-E12-VE50 is morepronounced than the other two, as can be seen in movie 5. Theeffective contact zone during the separation is indicated bydashed circles and defined as the area in which (1) the probeand adhesive are fully in contact and (2) there is no trace ofundulation. The effective contact zone is smaller for the sampleB-E12-VE50 and the finger instabilities are more developed. Inaddition, the contact line has been located between the fibrilsand not on top of them. The effective contact zone for thesample P-VE50 is remarkably larger than that of the sample B-E12-VE50. It means that the crack propagation is facilitated ontop of the sample B-E12-VE50. As the retraction continues, the

fingering instabilities of the sample B-E12-VE50 resemble thatof the sample P-VE50. However, separation event for thesample P-VE50 takes much longer time than that of the sampleB-E12-VE50. The fingers, and also vertical fibrils, on the flatcontrol sample appear to be pinned while they tend to moveeasily for the new hybrid structure. This can be seen in theforce-displacement curves in Figure 7 for the mentionedsamples. While the retraction trail for the sample B-E12-VE50shows only a slight slope change, it undergoes a huge strainhardening for the sample P-VE50. The fourth column showsthe surface deformation left after the complete separation. Theprobe retraction from the sample VE50 induces a moderatetrace of plastic deformation. Likewise, separation from thesample P-VE50 is occurring with a greater level of plasticdeformation which is easily visible after breakage; in contrast,there is no discernible plastic deformation on the sample B-E12-VE50.

3.4. Effect of Preload on the Adhesive Pull-Off Force.It has been previously reported that the adhesion andmechanical properties of homogeneous flat samples, eitherelastic or viscoelastic, have a limited dependence on preload.37

Preload dependence becomes more important if the materialhas a graded nature. Functionally graded materials have beenshown to have variation of the mechanical properties along thedepth from their surface. For instance, modulus of elasticity ofthe power law graded materials are known to vary with distancefrom the surface.38 Biomimetic fibrillar adhesives have beentreated as the graded materials by Yao and Gao.39,40 Thepreload dependence of mechanical and adhesive properties ofthese adhesives has been well studied in the literature.7,41

Figure 10 shows the variation of the pull-off force with preloadfor the four adhesive structures. The pull-off forces for allsamples increased as the preload increased. The functionallygraded structure B-E12-VE50 has a much higher pull-off forceand greater dependence on the preload than that of the other

Figure 9. Bottom view images of the contact at (i) preload point, (ii) pull-off point, (iii) an arbitrary point after the pull-off, and (iv) afterindentation test for (a) sample VE50, (b) sample P-VE50, (c) sample B-E12, and (d) sample B-E12-VE50. The experiments have been performedunder 0.5 mN preload.



three. This significant preload dependency can be attributed tothe presence of both micropillars and the top viscoelastic layer.3.5. Adhesion Energy Analysis. While pull-off force is

one of the most commonly used indicators of the surfaceadhesion, preload dependence of the pull-off force forbiomimetic fibrillar structures makes it difficult to use.41 Forthis reason, the work of adhesion can be calculated as arelatively universal value to indicate the adhesive properties of amaterial. In an indentation test, the work of adhesion is definedas the hysteresis or energy dissipation per change in the area ofthe contact. The hysteresis is defined as the difference betweenthe stored strain energy due to intersurface attraction andloading, and the work required separating the contactingsurfaces during the unloading.

∫ ∫ ∮δ δ δ= + = + =U U U F F Fd d dHys

d

d

d

1 20

max

max

(1)

U1 is the strain energy stored in the system. U2 is the requiredenergy to separate the probe from the surface, or open a crackat the interface. dmax is the displacement or depth of penetrationof the probe at preload; d is the displacement of the failurepoint. For an entire indentation process, the contact area at thestarting and the failure points are equal to zero and the totalwork of adhesion can be shown as:

∫ ∫ ∮δ δ δ=

Δ=

−+

−=W

U

A

F

A A

F

A A

F

A

d d dadh

hysd

d

d

d

0

max 0 max max

max

max

(2)

Since we have performed the indentation tests with differentpreload values, both the contact area at preload and hysteresisvary. Thus, the slope of the linear fit to the UHys vs ΔA dataprovides the universal work of adhesion value. A similarreasoning has been used in ref 42 to obtain the work ofadhesion for fully elastic systems. Figure 11 shows the variationof the hysteresis against preload and consequently contact areafor the tested samples. It reveals an approximately linearrelationship. Thus, the work of adhesion can be determinedfrom the slop of the linear fitting lines.It is well-known that the critical release energy rate is a

multiplicative function of thermodynamic work of adhesion, or

the quasi-equilibrium energy release rate at v ≈ 0 (G0), anddissipation factors which are dependent on temperature anddebonding velocity(φ(T,v)), i.e., G = G0(1 + φ(T,v)) . It isbelieved that the thermodynamic work of adhesion does notsignificantly depend on the amount of cross-linking agent usedin a polymer. Thus, in our experiments, the only origin ofalteration in adhesion energy is related to the extent ofirreversible energy dissipations. The energy release rate of afully cured and purely elastic PDMS film estimated throughJKR fitting was found about 0.07 J/m2 which is greater than itsthermodynamic work of adhesion (≈0.044 J/m2).42 Accordingto Figure 11, the work of adhesion for the elastic thin-filmterminated micropillars (sample B-E12) is about 0.35 J/m2

which is higher than that of a flat nonpatterned elastic PDMSfilm. This value is very close to the value obtained for a similarstructure in ref 42. This increase is attributed to thecombination of mechanisms such as crack trapping andenhanced compliance. The work of adhesion for the viscoelasticfilm on the glass (sample VE50) is 0.77 J/m2 as a result of themore dissipative nature of the viscoelastic sample. The adhesionenergy for sample P-VE50 is greater than VE50. Larger contactarea and higher compliance of the backing material are themain reasons for this difference. Also, this leads to increment ofthe confinement ratio, rendering the separation process towardan unstable more dissipative event. Finally, the adhesion energyfor the sample B-E12-VE50 is the highest. Even the valuesobtained by superimposing the work of adhesion for both pairsof (B-E12, VE50) and (B-E12, P-VE50) is lower than the workof adhesion obtained for the sample B-E12-VE50. This can beattributed to the synergistic effect of compliance of the backingmaterial and the elasticity of the top layer on the total adhesiveproperties of the system.

3.6. Compliance Analysis of the Graded Adhesives.Enhancement of compliance has been reported as one of themain advantages of the biological fibrillar adhesives.18,42−44

Enhanced compliance in such adhesive systems leads toincrement of effective contact area and accordingly increasein effective surface forces during the interaction with matingsurfaces. It has also been shown that the increase of compliancecan amplify the adhesion hysteresis and energy.42 Nadermannet al investigated the effect of contact compliance on theperformance and strength of biomimetic film-terminatedstructures.29 They showed that the use of arrays of microfibrilsunderneath an elastic PDMS thin film increases the compliance

Figure 10. Variation of pull-off force vs preload for the four differentsamples indicating their preload dependence behaviors.

Figure 11. Plot of the hysteresis versus contact area for tested samplesunderwent different preloads during the indentation tests; slope ofeach line indicates the work of adhesion for each tested sample.



of the interface. Following this work, we hypothesize that thecompliance of the biomimetic film-terminated structure can befurther enhanced using a viscoelastic top layer. To verify thisidea, we calculated the contact compliance and Young’smodulus since they are the determining parameters forpredicting the efficacy of an adhesive material.First, the effective Young’s modulus was calculated based on

the Hertz model. The fitting/numerical results deviated fromthe experimental results, particularly for the samples withviscoelastic top layer at small preloads. In fact, surfaceinteractions at low preloads caused a remarkable deformationof the surface and accordingly snap-in force, hindering the useof Hertz model to predict the correct effective Young’smodulus. The JKR theory includes the surface interactionsbut it cannot be directly used for measurement of the elasticmodulus of viscoelastic materials; modified JKR models havebeen developed for viscoelastic materials which involvedcomplicated mathematical treatments.42,45−47

Given the limitations of the Hertz and JKR models for theloading portion, we opted to use the unloading portion of theindentation cycle. In detail, to calculate the Young’s modulusand compliance, the slope of the load−displacement curve andthe Boussinesq definition of the compliance at the start of theunloading point were used, where the contact area is almostpinned and constant. This approach is essentially the same asthe one which Oliver−Pharr theory used. That is, the effectiveYoung’s modulus can be calculated from the unloading portionof the indentation cycle by relating the stiffness of the material(S = 1/C) to the pinned contact area at the vicinity of theunloading point and reduced Young’s modulus.48−50 One mainaspect of the Oliver−Pharr theory is the approximation of theunknown contact area based on the indentation depth and totaldisplacement. Furthermore, veracity of this theory for bothnonpolymeric materials undergoing plastic deformation andviscoelastic polymeric materials has been reported in ref 49.Since the contact areas for most of our samples are known, theuse of the Oliver−Pharr’s approximation is not necessary.However, the similarity of our approach to theirs makes itreasonable to calculate the compliance at the starting point ofthe unloading indentation curve and the use of the Boussinesqdefinition.Herein, the compliance (C) for a fixed contact area is defined

as the change in displacement per unit force (C = dδ/dF). Thisvalue is obtained by finding the slope of the unloading curve atthe point where the unloading starts. Figure 12 shows theload−displacement curves for indentation of sample B-E12with different preloads. The similar graph can be obtained forother samples tested in this study. The unloading cycles in ourexperiments were linear for a remarkable range of retractiondistance. This situation makes unnecessary the use of Oliver−Pharr theory which is based on power law fitting to theunloading curve.48,49

To quantify the effective Young’s modulus of our samples,we used the Boussinesq definition of the compliance with aconstant contact area upon loading as follow:42

=*

=*

π

CrE E

12

1

2B A

(3)

In this equation r is the contact radius; E* is the combinedeffective Young’s modulus of substrate and probe that can berelated to the elastic modulus of each side by the followingequation:

*=

−+

−E

vE

vE

1 (1 ) (1 )12

1

22

2 (4)

where v is the Poisson ratio of each side of the interface. As ourprobe is fused silica, which is a stiff material, the eq 4 reduces tothe form of (1/E*) = ((1 − v1

2)/(E1)) . We also assume v =0.5.Variation of the compliance and Young’s modulus with

preload is shown in Figure 13. The compliance of the sample B-

E12-VE50 is the highest and dramatically decreases as thepreload increases at low preloads (<3 mN). The compliance ofthe samples B-E12 is less and also dramatically decreases as thepreload increases at low preloads (<3 mN). In contrast, thecompliance of P-VE50 and VE50 are much lower and graduallydecrease with preloads. At higher preloads (>3 mN), thecontact compliance of the samples VE50 and P-VE50 and B-E12 continued to gradually decrease. But for the sample B-E12-VE50, the compliance only slightly decreased for the preloadbetween 3 and 7 mN and started to increase at higher preloads(10 mN). The Young modulus shows the opposite trend to thecompliance as predicted by eq 3. It is interesting to note thatthe effective Young’s modulus of the sample B-E12-VE50 is inthe same order of magnitude of a typical pressure sensitive

Figure 12. Force vs displacement curves of the sample B-E12 underdifferent preloads ranging from 0.5 to 10 mN. The slope of theunloading curve at the preload has been used to estimate thecompliance.

Figure 13. Variation of the compliance and elastic modulus of thetested samples against preload.



adhesive as described by the Dahlquist criterion.51 It is alsoinformative to notice that adding the micropillar to thesubsurface layer of the viscoelastic film reduced the effectivemodulus by 6 times to have a value of 1.5 × 105 Pa. Thisanalysis confirmed that the biomimetic fibrils increase thecontact compliance and verified that the viscoelastic top layerfurther increased the contact compliance. Note that thecompliance and Young’s modulus of the elastic film-terminatedfibrillar interfaces and arrays of simple microfibrils have beentheoretically modeled by Shen and Noderer et al.31,42 Assumingthe additive contributions of each layer to the global structureof the graded adhesive, it might be interesting to apply theseexisting models to predict the contributions of each layer andthe influence of structural parameters. However, we haveobserved a significant interplay between the top viscous layerand the base micropillars as discussed in the following sections,which hinders the analysis of the whole structure using theexisting analytical models and requires more detailed analyticaland numerical studies.3.7. Synergetic Interactions between Biomimetic

Fibrils and Viscoelastic Top Layers. It is known thatbuckling of fibrillar structures can increase the compliance.42

Thus, we further investigated the possible bending and bucklingof biomimetic fibrils at increased preloads. Figure 14 showstypical bottom-view images for the sample B-E12-VE50 atincreased preloads. Examination of the recorded videos of thecontact deformations revealed that the minimum preload forbending the pillars in the sample B-E12-VE50 is around 3 mN.In contrast, a minimum preload of about 25 mN is required toinduce visible bending of the pillars in sample B-E12. Weapproximated the weight of a 50 μm thin PDMS layer for a unitof area (≈ 0.5 mN/m2). This result shows a negligible extraweight on top of the fibrils. Thus, there should be anothermechanism causing the bending and buckling of the pillars thanthe extra weight induced by the viscoelastic top layer. This

suggests the possible synergetic interactions between the fibrilsand viscoelastic top layers.The enhanced compliance for the new hybrid structure can

be attributed to two mechanisms. First, because of theviscoelastic nature of the top layer, it deforms laterally on thesurface under the preload, causing interfacial shear stress to theunderlying elastic film. If the viscoelastic top layer is completelydiffused into the elastic backing layer, it forms an interphase.Thus, the no-slip condition governs the interfacial shear stress.In such a case, the only possible reason for the facilitatedbending of the pillars can be related to the deformation of theintermediate elastic film induced by Poissonian deformation ofthe viscoelastic top layer.The existence of a second scenario is possible as the elastic

layer is fully cross-linked and the diffusion at the interface mightbe limited. It is known that there exists a slippage of theviscoelastic polymers confined between two substrates. This isthe case even for compression of a polymer melt in a confinedgeometry that is chemically modified with the identicalpolymer. Although Brochard and de Gennes reported thatthe identical grafted chains on confining substrates can hinderthe polymer slippage remarkably, in certain stress rates theconformation of the grafted chains shift from “coiled” to“stretched”, causing polymer slippage.52 In our experiments, asignificant amount of shear slippage is expected at the interfacebetween the viscoelastic and elastic layers. This shear slippagecan result in deformation of the elastic film laterally andconsequently bending of the fibrils as illustrated in Figure 15.Moreover, the extent of the polymer slippage at the interface

is affected by both thermomechanical properties of the polymerand surface properties of the substrate. The static and dynamicfriction between the polymer and substrate hinder the slippageand flow of the viscoelastic material.53−55 But the softness ofthe polymer facilitates its slippage on the substrate. It has beenshown by Brown that a surface with greater local mobilityfacilitates the slippage of the polymer.56 The slippage results in

Figure 14. Bottom view images of the pillars in the sample B-E12-VE50 during an indentation test (a) no bending, (b) slight bending at preload of 3mN, and (c) full buckling at preload of 10 mN.

Figure 15. Interplay of the top viscous layer and base micropillar during the compression: the slippage of the viscoelastic polymer stretched theintermediate layer which subsequently induced the bending deformation of micropillars.



the less adhesion force of a viscoelastic material laid on asubstrate with higher local mobility. The lower pull-off forceand higher contact area for the sample P-VE50 comparing tothat of the sample VE50 can be explained by this fact. For thegraded adhesive structure B-E12-VE50, the presence ofmicropillars at the base of the sample may have increasedsurface mobility and the slippage, which in turn enhance thebending and buckling of the micropillars at low preloads. Thissynergetic interaction resulted in higher pull-off force andadhesive energy comparing to that of P-VE50.Finally, it is worthwhile to look at the gecko and tree frog-

inspired hybrid adhesive structure from the perspective offunctionally graded materials and discuss the technicalimplications. There exists a myriad of natural substances havingoutstanding bulk resistance to cracking, deformation, anddamage, thanking to their microstructured or porous-basedgradations. Bamboos, bones, and plant stems are the commonexamples of such structures with the strongest elements locatedwhere the stress is the maximum.38 Particularly, the Humboldtsquid’s body is composed of an extremely stiff beak embeddedin a soft buccal envelop. When hydrated, the chemical gradientin the beak renders a gradient in stiffness ranging 2 orders ofmagnitude from the top to the base.57 It has been shown suchgradients in chemical and mechanical properties of a structureare able to hamper the cracking, deformation, and damagethrough concentrating the stress to the joints betweenmechanically dissimilar materials.38,57 In the other words, theextent and mode of crack growth can be intensified or reducedacross an interface by alteration of the elastic propertiesunderneath the free surface of the material. Therefore, gradedmaterials have attractive potentials as materials resistant tocontact damage.38 Contact mechanics study of both elastic andplastic graded materials has been facilitated by depth-sensingindentation.39,40,58−61 Gradient in the chemical, geometrical,and mechanical properties along the depth axis is also believedto regulate the adhesion properties of materials.59 Thus, far,there is only a few theoretical studies on the mechanics ofelastic graded materials considering adhesive contacts.40,59

Interestingly, bioinspired fibrillar structures can be treated asgraded materials. For instance, Yao and Gao developed aninterfacial crack model and showed propensity of the elasticbioinspired fibrillar adhesives, as graded materials, to increaseadhesion robustness and flaw tolerance.62 While the domain ofthese investigations is limited to only elastic graded materials,there is an appeal to elaborate more about plastic or viscoelasticgraded materials.38,60

The reported bioinspired graded adhesive structures in thiswork composed of three components: an array of elasticmicropillars at the base, a thin elastic intermediate layer and aviscoelastic top layer. The biomimetic fibrillar interfacefunctions as a spring foundation storing the elastic energyduring the bonding. It dramatically increases the compliance ofthe system both before and after bending and buckling of thepillars. The stored energy in the pillars can be retrieved duringthe separation which facilitates the crack propagation atinterface instead of cohesive failure. The intermediate elasticlayer facilitates integration of the viscous layer on top of thebiomimetic fibrillar foundation and transfer the shear stressfrom the top to the base micropillars. The viscoelastic top layerdissipates a large amount of energy during the separationbecause of the bulk deformation and instabilities, whichinduced a shear stress at the interface and enhanced thebending and buckling of the fibril. This synergetic interaction

among the three components resulted in higher pull-off forceand adhesive energy, higher compliance and resistance tocohesive failure. Other than the scientific insights obtained inthis complex system, the introduction of the concept offunctionally graded materials to the biomimetic adhesives mayprovide effective ways for the development of “soft” temporaryadhesive materials for effective adhesion and bonding processesused in biological and mechanical systems’ applications, forinstance, the emerging “soft” electronic devices that are foldableand able to stick to biological tissues.

4. CONCLUSIONS

We made a bioinspired graded adhesive structure composed ofan array of elastic film-terminated micropillars at the base and athin viscoelastic film of the same material on top. A systematicinvestigation of this bioinspired graded adhesive structure wasperformed in comparison with three control adhesive materials:viscoelastic film coated on a glass substrate, viscoelastic filmcoated on a soft elastomer, and elastic film-terminatedmicropillars. Indentation measurements show that more energyis consumed in crack initiation and less energy is consumed incrack propagation during the separation of the functionallygraded adhesive structure than that of a simple viscoelastic film.This facilitates the adhesive mode of separation (or inhibits thecohesive failure of the viscoelastic film) and leads to an adhesivematerial with higher structural integrity. The evolutions ofadhesive contact area and separation instabilities during theindentation were examined, revealing a large amount offingering, cavitation and crack trapping phenomena. Thepreload effects on the adhesive pull-off force and work ofadhesion were investigated along with energy and complianceanalysis. These studies showed a remarkable increment of pull-off force, work of adhesion, and compliance. The significantcompliance of the new structure was attributed to the effect ofinterfacial slippage on the bending and buckling of the fibrilsunderneath the viscoelastic layer. Accordingly, higher amountof energy dissipation was observed for such structure.Retrieving of the elastic energy stored in the fibrils underneatha viscoelastic layer facilitated the crack propagation resulting inthe less bulk deformation during the separation for the gradedstructure when it is compared to that of the control samples.Overall, this work demonstrated that the synergetic combina-tion of gecko-inspired micropillar structure with the viscoelastictop layer resulted in a functionally graded adhesive materialdelivering high adhesion, high compliance, and resistance tocohesive failure.

■ ASSOCIATED CONTENT

*S Supporting InformationMovies taken by a side-view camera for the sample B-E12-VE50(movie 1) and the sample P-VE50 (movie 2) to show thevertical fibrillation and surface deformation during theindentation, and movies 3−5 taken by a bottom-view camerato show the contact formation and separation instabilitiestaking place during the indentation for the samples E12, VE50,and B-E12-VE50, respectively. This material is available free ofcharge via the Internet at http://pubs.acs.org/.

■ AUTHOR INFORMATION

Corresponding Author*E-mail: (B.Z.) [email protected].



http://pubs.acs.org/

mailto:[email protected]

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis work was supported by the Natural Sciences andEngineering Research Council of Canada (NSERC) and theMinistry of Research and Innovation of Ontario. The authorsappreciate the help of Professor Costas Tzoganakis and Dr.Shui Han Zhu with the rheological characterizations.

■ REFERENCES(1) Gorb, S. N.; Sinha, M.; Peressadko, A.; Daltorio, K. a; Quinn, R.D. Bioinspir. Biomim. 2007, 2, S117−25.(2) Autumn, K.; Sitti, M.; Liang, Y. a; Peattie, A. M.; Hansen, W. R.;Sponberg, S.; Kenny, T. W.; Fearing, R.; Israelachvili, J. N.; Full, R. J.Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 12252−6.(3) Persson, B. N. J. J. Adhes. Sci. Technol. 2007, 21, 1145−1173.(4) Shahsavan, H.; Arunbabu, D.; Zhao, B. Macromol. Mater. Eng.2012, 297, 743−760.(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, 681−5.(6) Geim, A. K.; Dubonos, S. V.; Grigorieva, I. . V.; Novoselov, K. S.;Zhukov, A. A.; Shapoval, S. Y. Nat. Mater. 2003, 2, 461−3.(7) Greiner, C.; Campo, A.; Del Arzt, E. Langmuir 2007, 23, 3495−502.(8) Shahsavan, H.; Zhao, B. Langmuir 2011, 27, 7732−42.(9) Murphy, M. P.; Aksak, B.; Sitti, M. Small 2009, 5, 170−5.(10) Del Campo, A.; Greiner, C.; Alvarez, I.; Arzt, E. Adv. Mater.2007, 19, 1973−1977.(11) Murphy, M. P.; Kim, S.; Sitti, M. ACS Appl. Mater. Interfaces2009, 1, 849−55.(12) Jeong, H. E.; Lee, J.-K.; Kim, H. N.; Moon, S. H.; Suh, K. Y.Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 5639−44.(13) Glassmaker, N. J.; Jagota, A.; Hui, C.-Y.; Noderer, W. L.;Chaudhury, M. K. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 10786−91.(14) Shahsavan, H.; Zhao, B. Soft Matter 2012, 8, 8281−8284.(15) Bae, W.; Kim, D.; Suh, K. Nanoscale 2013, 5, 11786−11884.(16) Majumder, A.; Ghatak, A.; Sharma, A. Science 2007, 318, 258−61.(17) Arzt, E.; Gorb, S.; Spolenak, R. Proc. Natl. Acad. Sci. U.S.A. 2003,100, 10603−6.(18) Hui, C.-Y.; Glassmaker, N. J.; Jagota, a. J. Adhes. 2005, 81, 699−721.(19) Persson, B. N. J. J. Chem. Phys. 2003, 118, 7614.(20) Drotlef, D.-M.; Stepien, L.; Kappl, M.; Barnes, W. J. P.; Butt, H.-J.; del Campo, A. Adv. Funct. Mater. 2013, 23, 1137−1146.(21) Cheung, E.; Sitti, M. J. Adhes. 2011, 87, 547−557.(22) Patil, S.; Mangal, R.; Malasi, A.; Sharma, A. Langmuir 2012, 28,14784−91.(23) Carelli, C.; Deplace, F.; Boissonnet, L.; Creton, C. J. Adhes.2007, 83, 491−505.(24) Patil, S.; Malasi, A.; Majumder, A.; Ghatak, A.; Sharma, A.Langmuir 2012, 28, 42−6.(25) Shahsavan, H.; Zhao, B. In 33rd Annual Meeting of AdhesionSociety; Adhesion Society: Daytona Beach, FL, 2013.(26) Jagota, A.; Hui, C.-Y. Mater. Sci. Eng. R Rep. 2011, 72, 253−292.(27) Ashurst, W.; Carraro, C.; Maboudian, R.; Frey, W. SensorsActuators A: Phys. 2003, 104, 213−221.(28) Nase, J.; Lindner, A.; Creton, C. Phys. Rev. Lett. 2008, 101, 1−4.(29) Nadermann, N.; Ning, J.; Jagota, A.; Hui, C.-Y. Langmuir 2010,26, 15464−71.(30) Mukherjee, R.; Pangule, R. C.; Sharma, A.; Banerjee, I. J. Chem.Phys. 2007, 127, 064703.(31) Shen, L.; Hui, C.-Y.; Jagota, A. J. Appl. Phys. 2008, 104, 123506.(32) Liu, J.; Hui, C. Y.; Jagota, A. J. Polym. Sci., Part B: Polym. Phys.2009, 47, 2368−2384.(33) Crosby, A.; Shull, K. J. Polym. Sci., Part B 1999, 37, 3455−3472.(34) Gent, A. N.; Hamed, G. R. Polym. Eng. Sci. 1977, 17, 462−466.

(35) Nase, J.; Ramos, O.; Creton, C.; Lindner, A. Eur. Phys. J. E. SoftMatter 2013, 36, 103.(36) Paiva, A.; Sheller, N.; Foster, M. Macromolecules 2001, 2269−2276.(37) Creton, C.; Leibler, L. J. Polym. Sci., Part B: Polym. Phys. 1996,34, 545−554.(38) Suresh, S. Science 2001, 292, 2447−51.(39) Guo, X.; Jin, F.; Gao, H. Int. J. Solids Struct. 2011, 48, 2565−2575.(40) Yao, H.; Gao, H. J. Adhes. Sci. Technol. 2007, 21, 1185−1212.(41) Long, R.; Hui, C.-Y. Proc. R. Soc. A 2009, 465, 961−981.(42) Noderer, W. L.; Shen, L.; Vajpayee, S.; Glassmaker, N. J.; Jagota,a.; Hui, C.-Y. Proc. R. Soc. A Math. Phys. Eng. Sci. 2007, 463, 2631−2654.(43) Liu, J.; Hui, C.-Y.; Shen, L.; Jagota, A. J. R. Soc. Interface 2008, 5,1087−97.(44) Glassmaker, N. J.; Himeno, T.; Hui, C.-Y.; Kim, J. J. R. Soc.Interface 2004, 1, 23−33.(45) Greenwood, J.; Johnson, K. Philos. Mag. A 1981, 37−41.(46) Falsafi, A. J. Rheol. (N.Y.). 1997, 41, 1349.(47) Hui, C.; Baney, J. M.; Kramer, E. J. Langmuir 1998, 14, 6570−6578.(48) Oliver, W. C.; Pharr, G. M. J. Mater. Res. 1992, 7, 1564−1583.(49) Herbert, E.; Oliver, W. J. Mater. Res. 2009, 24, 626−637.(50) Tawfick, S.; De Volder, M.; Copic, D.; Park, S. J.; Oliver, C. R.;Polsen, E. S.; Roberts, M. J.; Hart, a J. Adv. Mater. 2012, 24, 1628−1674.(51) Dahlquist, C. A. Adhes Age 1959, 2, 25−29.(52) Brochard, F.; De Gennes, P. G. Langmuir 1992, 8, 3033−3037.(53) Newby, B. M.; Chaudhury, M. K.; Brown, H. R. Science 1995,269, 1407−9.(54) Zhang Newby, B.; Chaudhury, M. K. Langmuir 1997, 13, 1805−1809.(55) Zhang Newby, B.; Chaudhury, M. K. Langmuir 1998, 14, 4865−4872.(56) Brown, H. Science 1994, 263, 1411−1413.(57) Miserez, A.; Schneberk, T.; Sun, C.; Zok, F. W.; Waite, J. H.Science 2008, 319, 1816−9.(58) Guler, M. a.; Erdogan, F. Mech. Mater. 2006, 38, 633−647.(59) Chen, S.; Yan, C.; Zhang, P.; Gao, H. J. Mech. Phys. Solids 2009,57, 1437−1448.(60) Gu, Y.; Nakamura, T.; Prchlik, L. Mater. Sci. Eng., A 2003, A345,223−233.(61) Nakamura, T.; Wang, T.; Sampath, S. Acta Mater. 2000, 48,4293−4306.(62) Gao, H.; Yao, H. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 7851−7856.



bioinspired functionally graded adhesive materials: synergetic interplay of top viscous–elastic...

Documents