Effect of Viscoelasticity on Adhesion of Bioinspired Micropatterned

ARTICLE pubs.acs.org/Langmuir

Effect of Viscoelasticity on Adhesion of Bioinspired Micropatterned Epoxy Surfaces Graciela Castellanos,† Eduard Arzt,† and Marleen Kamperman*,‡ † ‡

Functional Surfaces Group, INM - Leibniz Institute for New Materials, Campus D2 2, 66123 Saarbr€ucken, Germany Physical Chemistry and Colloid Science, Wageningen University, Dreijenplein 6, 6703 HB, Wageningen, Netherlands ABSTRACT: The effect of viscoelasticity on adhesion was investigated for micropatterned epoxy surfaces and compared to nonpatterned surfaces. A two-component epoxy system was used to produce epoxy compositions with different viscoelastic properties. Pillar arrays with flat punch tip geometries were fabricated with a two-step soft lithography process. Adhesion properties were measured with a homebuilt adhesion tester using a spherical sapphire probe as a countersurface. Compared to flat controls, micropatterned epoxy samples with low viscoelasticity (i.e., low damping factors) showed at least a 20-fold reduction in pull-off force per actual contact area for both low (E0 = 2.3 MPa) and high (E0 = 2.3 GPa) storage moduli. This antiadhesive behavior may result from poor contact formation and indicates that the adhesion performance of commonly used elastomers for dry adhesives (e.g., polydimethylsiloxane) is governed by the interfacial viscoelasticity. Adhesion significantly increased with increasing viscoelasticity. Micropatterned samples with high viscoelasticity showed a 4-fold reduction in adhesion for aspect ratio (AR) 1.1 patterns but a 2-fold enhancement in adhesion for AR 2.2 patterns. These results indicate that viscoelasticity can enhance the effect of surface patterning on adhesion and should be considered as a significant parameter in the design of artificial patterned adhesives.

’ INTRODUCTION The remarkable attachment system of some animals enables locomotion under different environmental conditions. As a result of these remarkable properties, investigation of artificial, bioinspired adhesives has attracted the attention of many investigators.1
6 The Tokay gecko can firmly adhere and easily detach from almost any kind of surface, due to the microtopography of its toe pads combined with optimized biomechanics.2 These pads are covered by β-keratin fibrils with hierarchical structuring from long micrometer-sized setae to nanosized spatulae. β-Keratin is a stiff and hydrophobic protein material with a bulk elastic modulus of around 2 GPa. Therefore, it would be unexpected for a β-keratin structure to efficiently and reversibly deform and attach to a surface. However, due to the hierarchical construction, and in particular due to the nanosized platelike spatula structures, close (atomic) contact with the surface is formed, which enables strong and reversible adhesion.5 It is generally assumed that the attachment system of geckos deform elastically. Recently, however, it was shown that viscoelasticity might also play an important role in gecko adhesion.7 The development of engineering materials with similar reversible adhesion capabilities is motivated by many potential applications such as in robotics, transport, biomedical devices, or sports equipment. Recently, synthetic patterned surfaces have been developed to mimic the geometry of natural attachment systems.8
13 The stiffness of the materials used in the designs varies greatly from compliant elastomers to extremely stiff carbon r 2011 American Chemical Society

nanotubes. Spolenak et al. developed “adhesion design maps” to predict the optimum pillar geometry for a certain material stiffness by balancing several design parameters (such as fiber strength, fiber condensation, and ideal contact strength).14,15 Whereas many design parameters have been systematically studied,13 the role of stiffness and viscoelasticity on patterned adhesives has not been investigated in detail. By contrast, viscoelastic effect of nonpatterned polymer surfaces has been investigated in great detail since the 1960s.16 It is well-established that the amount of energy dissipation in bulk viscoelastic materials can far exceed the intrinsic work of adhesion between two surfaces. Many theoretical and experimental investigations have been aimed at the deconvolution of these interfacial and bulk contributions.17
20 It was shown that the adhesive strength of viscoelastic adherends is qualitatively linked to the ratio of the interfacial adhesion energy and the bulk elastic modulus of the polymer.21,22 To maximize adhesive energy, this ratio must lie in an optimal range to ensure that the interfacial contribution is strong enough to support the dissipative processes in the bulk. Further investigations with the inclusion of cavitation and fibril formation in the debonding mechanism provided a more complete picture to characterize adhesion of viscoelastic polymer surfaces.23 The knowledge of this well-developed field Received: March 12, 2011 Revised: May 9, 2011 Published: May 23, 2011 7752

dx.doi.org/10.1021/la2009336 | Langmuir 2011, 27, 7752–7759

Langmuir

ARTICLE

Figure 1. Molecular structure of epoxy resin DER 331 (a) and DER 732p (b).

Figure 2. SEM micrographs showing arrays of SU-8 pillars with 9 μm diameter and height of 10 μm (a) and 20 μm (b) and of micropatterned epoxy samples EP-30-p1 (c), EP-30-p2 (d), EP-0-p2 (e), and EP-40-p2 (f).

could be very useful in the investigation of artificial, bioinspired adhesives. Here, we experimentally study the effect of viscoelasticity and stiffness of polymeric materials on their adhesive performance for nonpatterned surfaces and for patterned arrays of micropillars with defined dimensions. To realize a controllable variation in viscoelastic properties of one material, we selected an epoxy system consisting of two epoxy resin components (see Figure 1). By changing the mixing ratio of the two components, the elastic modulus of the material can be tuned over 3 orders of magnitude (from MPa to GPa).24 We fabricated the composites by photolithographic techniques and molding. Dimensions of the micropatterned surfaces were measured with light microscopy and scanning electron microscopy (SEM). Thermomechanical properties and surface free energies were determined with dynamic mechanical thermal analyses (DMTA) and contact angle measurements, respectively. Load
indentation depth curves were obtained with a home-built indentation equipment used to study and determine the adhesive properties of the structures. Our results suggest that viscoelasticity significantly enhances adhesion of micropatterned surfaces.

’ EXPERIMENTAL SECTION Epoxy-based pillar surfaces were fabricated through a two-step soft lithography procedure. Arrays of holes of polydimethylsiloxane (PDMS) were obtained by molding on lithographically patterned SU-8 films containing pillars and subsequently used as template to get epoxy patterned surfaces. A two-step soft lithography process was used because

the high chemical and mechanical stability of SU-8 is optimum for molding soft polymers and SU-8 is not suitable for molding stiff materials.25 Materials. Epoxy-based liquid resins DER 331 and DER 732p were provided by Nordmann, Rassmann (Hamburg, Germany), Amicure PACM curing agent was provided by Air Products (Utrecht, Netherlands), and PDMS Sylgard 184 was purchased from Dow Corning through Arrow Europe (Dreieich, Germany). Silicon wafers (100 orientation, 50.8 mm in diameter and 275 μm thick) were obtained from Crystec (Berlin, Germany), SU-8 type 2010 and 2015 and developer mr-Dev 600 were from Micro Resist Technology (Berlin, Germany), and hexadecafluoro-1,1,2,2-tetrahydrooctyltrichlorosilane was from Alfa Aesar (Karlsruhe, Germany). The lithography mask was obtained from ML&C (Jena, Germany) in quartz, with 0.8  0.8 cm2 chrome patterned fields. Fabrication of the Specimens. Silicon wafers were cleaned with an O2-plasma Tepla Plasma Processor, type 300 Autoload provided by TePla GmbH (Wettenberg, Germany) at 600 W and 0.30 mbar of pressure for 5 min, rinsed with acetone, and blown dry with nitrogen before spin-coating and lithographic processing with SU-8 2010 and 2015 resists. A mask aligner MA 1006 from S€uss MicroTec (Garching, Germany) was used for the irradiation step with a 1000 W lamp. A filter LC-PL-360LP supplied by Laser Components (Olching, Germany) was used for truncating the irradiation wavelengths below 320 nm according to the sensitivity of the SU-8. The fabricated lithographic templates contained arrays of cylindrical pillars with a diameter of 9 μm and aspect ratio (AR) of 1.1 and 2.2 (see SEM images in Figure 2a,b). Prior to molding and to promote the separation of the template and molding material, the photolithographic patterns were silanized. The silanization was carried out in the gas phase using a vacuum desiccator for 30 min and using 20
50 μL of hexadecafluoro-1,1,2,2-tetrahydrooctyltrichlorosilane to obtain a layer of 1
2 nm. Then it was left in the 7753

dx.doi.org/10.1021/la2009336 |Langmuir 2011, 27, 7752–7759

Langmuir

ARTICLE

Table 1. Chemical Composition Range Used To Fabricate the Epoxy-Based Samplesa sample

wt % DER 331

wt % DER 732p

wt % curing agent

EP-0

0.0

83.4

16.6

EP-10 EP-20

8.3 16.4

74.4 65.5

17.3 18.1

EP-30

24.3

56.8

18.9

EP-40

32.0

48.1

19.9

EP-50

39.5

39.5

21.0

a

From EP-0 (0 wt % DER 331) to EP-50 (nominally 50 wt % DER 331 of the total amount of epoxy).

oven for 30 min at 95 °C in vacuum to increase lateral cross-linking and react free OH-groups of the resist surface.12 For the first step of the molding process a 10:1 ratio of Sylgard 184 prepolymer and cross-linker was mixed, degassed, and poured on the silanized SU-8 patterned wafer. Then it was cured at 75 °C for 24 h at 600 mbar. PDMS samples containing holes were then carefully peeled off from the template and silanized overnight under vacuum.26 The fluorinated PDMS templates were used for the second step of the molding process. Epoxy resin (DER 331 and DER 732p) and curing agent (4,40 -methylenebis(cyclohexylamine)) were thoroughly mixed and poured on the PDMS templates, then cured in an oven at 80 °C for 2 h, postcured at 160 °C for 2 h, and finally demolded. Micropatterned epoxy-based surfaces with different epoxy resin ratios were fabricated and are listed in Table 1. The amount of curing agent is determined by its amine hydrogen equivalent weight (AHEW = 53) and the epoxy equivalent weight of the epoxy portion (EEWDER 331 = 182
192; EEWDER 732p = 310
330; average values used). The patterned area for each sample was 0.8  0.8 cm2. Careful demolding and silanization are crucial in order to obtain structures without mold failure. The stiffer the material, the more difficult demolding becomes. The profiles of the patterned surfaces were characterized by light microscopy (Olympus BX51) and by SEM (FEI Quanta 400 ESEM operating at energy between 1 and 15 kV). The nonpatterned samples and PDMS control surfaces were obtained from a PDMS template using a similar procedure. The thickness of the polymeric samples was 2 mm in all cases. Dynamic Mechanical Thermal Analysis. The thermal and mechanical properties of the epoxy materials were studied by DMTA (Q800 V20.9 Build 27). Epoxy samples with dimensions 30  10  2 mm3 were fabricated from a silanized PDMS mold and tested in the temperature range of
100 to 100 °C at a heating rate of 3 °C/min. The tests were performed under nitrogen at an oscillatory frequency of 1.0 Hz and amplitude of 15 μm in tensile mode. Contact Angle Measurements. To characterize the surface free energy of the nonpatterned polymeric surfaces, contact angle measurements were performed with a contact angle measuring system (G2/G40 V3, Kr€uss GmbH, Hamburg). Contact angle values were measured for all epoxy surfaces and for the PDMS control sample using water and 1-octanol. The Owens and Wendt model was used to estimate the surface free energy of the solid surfaces.27 Adhesion Measurements. The adhesion performance of the patterned and nonpatterned epoxy surfaces was tested by recording load
indentation depth curves obtained with a home-built indentation apparatus.28 It consists of a laser interferometer (SP 120 from SIOS, Ilmenau, Germany) and a hexapod positioning stage (F-206 from Physik Instrumente, Karlsruhe, Germany). A sapphire sphere with a diameter of 4 mm, purchased from Goodfellow (Huntingdon, U.K.), was glued to the middle point of a glass spring fixed on both ends. A spherical probe was chosen to prevent possible misalignment between probe and surface. The sample was placed on the positioning stage and moved up at constant velocity against the fixed sphere to a predetermined

Figure 3. Storage modulus (a) and tan δ (b) of different epoxy compositions as a function of temperature at 1.0 Hz. The dotted line shows room temperature. The dimensions of all samples were 30  10  2 mm3. preload, Pp. Subsequently, the sample was retracted at constant velocity until the final detachment event occurred giving the pull-off force (Pc), which is a measure of the adhesion performance. Deflection of the spring was monitored via a laser interferometer and converted into a force. The spring constant used in all experiments was 147 N/m. The indentation depth is obtained as the difference between the stage position and the deflection of the cantilever during contact. The maximum vertical displacement of the stage after contact was 100 μm, and the positioning accuracy was 1 nm. The substrate was cleaned with ethanol and brought in contact with the sample several times before each experiment to avoid changes due to material transfer between the surfaces.28 The temperature and relative humidity (RH) were controlled during experiments and set at ∼25 °C and ∼50% RH. A minimum of three measurements was performed for each data point. To study the effect of loading rate on the adhesion, experiments were carried out at four experimental loading rates (1, 5, 10, and 50 μm/s). For the patterned surfaces we considered the effective area of contact available on the epoxy surfaces, dividing the pull-off force values by the packing density, 100% for the nonpatterned and 18.4% in the case of micropatterned surfaces.

’ RESULTS Arrays of hexagonally distributed epoxy pillars were successfully fabricated. The dimensions of the fabricated micropatterned surfaces were d = 9.0 μm in diameter, 11 μm in interpillar 7754

dx.doi.org/10.1021/la2009336 |Langmuir 2011, 27, 7752–7759

Langmuir

ARTICLE

Table 2. Material Properties of Epoxy Compositionsa DMTA results at room temperature (25 °C) and surface free energy sample

Tg [°C]

storage modulus E0 [MPa]

loss modulus E00 [MPa]

tan δ

γs [mN/m]


11

2.3

0.079

0.034

20.34

EP-10

3

2.6

0.94

0.36

20.98

EP-20

13

5.9

0.70

24.08

EP-30

30

144

114

0.79

29.65

EP-40

44

1006

249

0.25

31.59

EP-50

65

2322

113

0.049

32.92

EP-0

8.4

Loss moduli are obtained using tan δ = E00 /E0 . The glass transition temperature (Tg) is defined as the peak in damping factor (tan δ). γs is the surface free energy. The mechanical properties correspond to values at room temperature. a

Figure 4. Dependence of the pull-off force, Pc, on the preload, Pp, in adhesion experiments on flat (a) and patterned (b) surfaces with pillars of 9 μm diameter and AR of 1.1. The spherical sapphire probe of 4 mm diameter was approached to and retracted from the surface at constant velocity of 5 μm/s. The effective area of contact available on the epoxy surfaces was calculated by dividing the pull-off force values by the packing density. The calculated contact area fraction was 100% for the nonpatterned and 18.4% for micropatterned surfaces.

distance, and h = 10 and 20 μm in height with AR 1.1 and 2.2, respectively. SEM images of samples with AR 1.1 and 2.2 are shown in Figure 2c
f. The dimensions of the resulting epoxy structures are roughly the same as the original SU-8 fibrillar

pattern. In the remainder of the text we will denote all samples with a label, stating the composition (EP-0
EP-50) and the pattern dimensions (np = nonpatterned; p1 = AR 1.1 pillars; p2 = AR 2.2 pillars); e.g., EP-0-p1 denotes the sample with 0% DER 331 and aspect ratio 1.1. The results obtained from DMTA analysis at 1.0 Hz are presented in Figure 3 and summarized in Table 2. The storage modulus (E0 ) curves in Figure 3a reveal that the value of E0 could be varied from a few to thousands of MPa by increasing the ratio of DER 331 to DER 732p epoxy resins for a given temperature. Figure 3b depicts the variation of the damping factor or tan δ with temperature. The glass transition temperature, Tg, was taken as the peak value of tan δ. The glass transition temperature increases with increasing DER 331 epoxy content, from
11 °C (EP-0) to 65 °C (EP-50). At room temperature, EP-30 is close to its Tg and will show significant viscoelastic behavior at room temperature. Surface free energies (γs) of the epoxy composites determined by contact angle measurements are listed in Table 2, and a value of 20.68 mN/m was obtained for PDMS. γs increases moderately with increasing DER 331 content. Figure 4 shows the effective pull-off forces, Pc, plotted versus the applied preload, Pp, for flat (Figure 4a) and patterned samples with AR 1.1 (Figure 4b) for an loading rate of 5 μm/s. Both flat and patterned epoxy composites showed a dependence of pull-off force on preload, with the exception of PDMS surfaces and samples that showed almost no adhesion (EP-50-np, EP-0-p1, EP-10-p1, EP-50-p1, and PDMS-np). Other epoxy samples showed an increase in pull-off force with preload, without reaching a plateau, aside from EP-20-p1. This particular sample reached a plateau after a fast increase in pull-off force. For EP-40-np and EP-30-p1, Pp increased up to 10 times with increasing preload. The pull-off force values of the nonpatterned epoxy surfaces at high preloads first increased with DER331 content until sample EP-40-np and then decreased to low pull-off forces with even higher DER331 content. The patterned epoxy surfaces followed the same trend except that the highest adhesion was found for EP-30-p1 and pull-off forces were always lower than for the flat samples. For EP-50-p1 no measurable adhesion was observed. The effect of aspect ratio on pull-off force is shown in parts a, b, and c of Figure 5 for samples with 0%, 30% and 40% content of DER 331 epoxy, respectively. Micropatterning of EP-0 resulted in negligible adhesion for both AR 1.1 and 2.2, with at least a 20fold reduction in adhesion compared to flat controls. Different behavior was obtained for EP-30 and EP-40. Whereas the micropatterned surfaces with AR 1.1 showed up to 4-fold lower pull-off forces than flat surfaces, increasing the AR of the pillars to 7755

dx.doi.org/10.1021/la2009336 |Langmuir 2011, 27, 7752–7759

Langmuir

ARTICLE

load
indentation depth curves at different rates are depicted in Figure 6 for micropatterned EP-0 and EP-30 epoxy surfaces with AR 1.1 and the corresponding unpatterned controls. The compressive parts of the EP-0-np and EP-0-p1 curves are more or less identical for all velocities, indicating that bulk viscoelastic effects are small. The effects of velocity on adhesion are more pronounced for EP-30. The compressive parts of EP-30-p1 show increased hysteresis and steeper slopes with higher velocity, which are manifestations of bulk viscoelasticity. Additionally, both EP-30-np and EP-30-p1 show an increase in pull-off force with higher velocities (see also Figure 7). Figure 7 shows a logarithmic plot of the pull-off force at a preload of 14 mN as a function of the loading rate for nonpatterned and micropatterned epoxy surfaces. Increasing the rate from 0.1 to 10 μm/s resulted in constant or moderately increased pull-off forces. The largest relative increases in pull-off force for nonpatterned epoxy surfaces were found for EP-20-np and EP30-np. A similar trend is observed for micropatterned surfaces (Figure 7b); EP-20-p1 and EP-30-p2 showed the largest relative increase in pull-off force values with increasing loading rate. We note for samples showing low adhesion that the pull-off force values are too low to observe any trend with loading rate.

’ DISCUSSION

Figure 5. Effect of aspect ratio on pull-off force at constant velocity of 5 μm/s for different epoxy compositions: (a) EP-0, (b) EP-30, and (c) EP40. The effective area of contact available on the epoxy surfaces was calculated by dividing the pull-off force values by the packing density. The calculated surface area packing density fraction was 100% for the nonpatterned and 18.4% for micropatterned surfaces.

2.2 resulted in 2-fold enhanced adhesion compared to the values obtained for flat samples. To study the influence of loading rate on adhesion, experiments were performed at velocities of 0.1, 1, 5, and 10 μm/s. In this set of experiments, the preload remained constant. Representative

Effect of Stiffness on Adhesion. DMTA analysis and contact angle measurements showed that by increasing the amount of DER 331 in the epoxy composites the storage and loss moduli both increase 3 orders of magnitude, whereas the surface free energy only varied between 20 and 32 mN/m. Therefore, the adhesion performance appears to be predominantly influenced by changes in the thermomechanical properties rather than in surface free energy. With respect to the elastic behavior from Table 2 and Figure 4, pull-off forces of both patterned and nonpatterned samples are observed to first increase with increasing storage modulus and then decrease for subsequent higher values of the storage modulus. For a micropatterned surface with flat-punch-shaped tips separating from a flat substrate,32 the intrinsic strength of the pillar/ substrate interface may be higher for stiffer pillars.29 However, an increase in adhesion with higher stiffness is not easy to obtain, while the stiffer the material, the smaller the strain that can be applied before the deformation energy equals the adhesion energy and pull-off will occur. Only for a perfect uniform contact and the use of patterns consisting of long thin pillars that allow significant stretching because of the small individual contact radius can maximized pull-off forces for very stiff materials be realized.29,30 To obtain strong adhesion with the present system with relatively low aspect ratio pillars, the material has to be sufficiently compliant (1) to make close contact with the substrate and (2) for the interface to be able to support significant deformation of the bulk. For elastically deforming flat and micropatterned samples EP-0 (E0 = 2.3 MPa; EP-0-np Pc = 4 mN; EP-0-p1 Pc = 200 μN) and EP-50 (E0 = 2.3 GPa; EP-50-np Pc = 1.5 mN; EP-50-p1 Pc = 65 μN) a lower pull-off force is indeed obtained for the stiffer sample. To explain the increase and subsequent decrease of the intermediate compositions, the viscoelastic contributions have to be taken into account. Viscoelastic Contributions to Adhesion. We note that the pull-off forces of both patterned and flat epoxy surfaces are higher for compositions with a higher damping factor, i.e., a more viscoelastic behavior (see Figure 8). Viscoelasticity enhances 7756

dx.doi.org/10.1021/la2009336 |Langmuir 2011, 27, 7752–7759

Langmuir

ARTICLE

Figure 6. Experimental load vs indentation depth data for (a) EP-0 nonpatterned and (b) AR 1.1 patterned surfaces and for (c) EP-30 nonpatterned and (d) micropatterned surfaces with AR 1.1 at different loading rates and constant preload of 14 mN. Positive loads are compressive and negative loads tensile.

contact formation and invokes dissipative processes, thereby improving the adhesion performance. For both patterned and flat surfaces, the enhancing viscoelastic effect dominates the reducing effect of increasing storage modulus. The preload dependence of the different compositions is also related to the viscoelastic behavior (see Figure 4). Epoxy compositions with higher damping factors show a larger increase in pull-off force with Pp because additional load and time allows a viscoelastic system to improve the contact and thereby the adhesion performance. Viscoelastic effects are also observed in load
indentation depth curves of EP-30 obtained at different velocities (Figure 6); the slope of the compressive part decreases with decreasing velocity. Moreover, reduced hysteresis at compressive loads in the load
indentation depth curves with decreasing velocity is observed. Smaller slopes indicate viscoelastic flow of the material relaxing some elastic strain energy with time, and reduced hysteresis indicates that at lower velocities the stored energy can be more efficiently used to break bonds at the interface rather than being dissipated into the system. Pull-off force values of viscoelastic materials are influenced by the loading rate through two opposing effects: lower rates lead to (1) less energy dissipation in the system, thus decreasing the pull-off forces, and to (2) improved contact formation, resulting in increased pull-off forces. For EP-20 and EP-30, it was observed that the energy dissipation effect dominates and the pull-off force values increase with rate. For EP-40 the effects seems to be balanced, and no significant increase in pull-off force is observed.

This observation is reasonable because EP-40 is much stiffer and contact formation will be more difficult. Flat Punch Tip Geometries. For AR 1.1, the adhesion performance of micropatterned surfaces is reduced compared to the flat controls for all compositions (Figure 4). This antiadhesive behavior cannot be explained by the reduction in surface area for patterned surfaces because pull-off forces per actual contact area were plotted. Actual contact areas were obtained by dividing the pull-off force by the packing density (0.184). As the effective modulus of a micropatterned surface is somewhat lower than flat controls, one should see improved contact formation. However, it is experimentally very difficult to ensure that both the glass probe and the pillars are perfectly flat, resulting in a typical measurement where only a fraction of the pillars is in close contact and will support the load. The combined effects of surface roughness and imperfections will be more severe for stiffer materials. Moreover, the edges of the pillars are somewhat rounded compared to an ideal cylinder and act as preexisting edge cracks. This deviation from the ideal cylindrical shape will facilitate crack propagation and failure of the interface. Interestingly, increasing the AR to 2.2 improved the adhesion performance of the more viscoelastic compositions (EP-30 and EP40), but not of the less viscoelastic and most compliant composition (EP-0) (see Figure 5). One of the key mechanisms of fibrillar adhesion is that the elastic strain energy stored in the pillars is dissipated at pull-off.3 Whereas the amount of elastic strain energy that can be stored in a pillar is directly proportional to the aspect 7757

dx.doi.org/10.1021/la2009336 |Langmuir 2011, 27, 7752–7759

Langmuir

ARTICLE

Figure 8. Variation of pull-off force with damping factor at 14 mN preload for patterned epoxy surfaces with pillars of 9 μm diameter and AR 1.1 (EP-0-p1
EP-50-p1). The spherical sapphire probe of 4 mm diameter was approached to and retracted from the surface at constant velocity of 5 μm/s.

Figure 7. Variation of pull-off force with increasing loading rate at 14 mN preload for nonpatterned (a) and patterned (b) surfaces with AR 1.1 and 2.2. The effective area of contact available on the epoxy surfaces was calculated by dividing the pull-off force values by the packing density. The area packing density ratio was determined to be 100% for the nonpatterned and 18.4% for micropatterned surfaces.

ratio, it is inversely proportional to the elastic modulus of the pillar (at a given stress). Therefore, it is unlikely that the difference in behavior between EP-0 and EP-30 can be attributed to this mechanism. More likely, the increased aspect ratio reduces the effective modulus and enhances bending of the pillars to conform to the spherical probe, which together with the viscoelastic behavior improves contact formation. The stronger interface is capable of supporting forces large enough to invoke dissipative processes. In other words, the viscoelastic properties are more efficiently used for higher AR structures. These results indicate that viscoelasticity can enhance the effect of micropatterning on adhesion; by changing the aspect ratio from 1.1 to 2.2, the adhesion can be changed from 25 to 200% the value of flat controls. It is not immediately clear why micropatterns of the most compliant composition (EP-0) show almost no adhesion. Micropatterned surfaces of the commonly used elastomer, PDMS, with the same geometrical features (diameter 9 μm; AR 1.1) and elastic modulus (2.3 MPa) showed significantly higher adhesion than its flat control (see Figure 4). The difference is most likely caused by the interfacial viscoelasticity of PDMS. PDMS is known to behave in a nearly ideally elastic manner in the bulk but to exhibit significant interfacial viscoelasticity. This effect

greatly enhances contact formation and may reduce the edge stress concentration. Therefore, results from PDMS cannot be generalized to other materials with a similar elastic modulus. For example, elastomeric polyurethane flat-punch-shaped pillars showed significantly reduced adhesion,31 similar to the present epoxy data. Therefore, we conclude that pillar arrays with flat punch tip geometries made of nearly ideally elastic materials with an elastic modulus in the MPa range may generally not be able to make good contact, and these micropatterned surfaces will not reach considerable adhesive strengths. Contact formation, and thus adhesion performance, can be improved by the incorporation of platelike terminal structures,12 by using shear to attain good contact31 or by using more viscoelastic materials.

’ CONCLUSIONS Micropatterned epoxy surfaces with flat punch tip geometries were fabricated with different storage moduli (2.3 MPa
2.3 GPa) and damping factors (0.034
0.79). The adhesion performance of these patterned surfaces was enhanced by viscoelastic behavior of the material. This effect dominated the reducing effect on adhesion of increasing storage modulus. Our results suggest that without bulk or interfacial viscoelasticity pillar arrays with flat punch tip geometries are unable to make close contact with the counter-surface, resulting in poor adhesion. However, viscoelastic contributions can enhance the effect of micropatterning on adhesion, which is important for future development of bioinspired artificial analogues. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]; Tel: þ31(0)317 482358; Fax: þ31(0)317 483777.

’ ACKNOWLEDGMENT We thank Yuri Egorov and Birgit Heiland for the DMTA analysis, Christian Cavelius for help with contact angle measurements 7758

dx.doi.org/10.1021/la2009336 |Langmuir 2011, 27, 7752–7759

Langmuir

ARTICLE

and data analysis, Aude Hettich and Andreas Schneider for SEM images, Griselda Guidoni and Baptiste Girault for useful discussions, and Philip Egberts and Dadhichi Paretkar for useful comments on the manuscript. This work, as part of the European Science Foundation EUROCORES Program FANAS, was supported by the German Science Foundation (DFG) grant AR201/9-1.

’ REFERENCES (1) Creton, C.; Gorb, S. MRS Bull. 2007, 32, 466–468. (2) Autumn, K.; Liang, Y. A.; Hsieh, S. T.; Zesch, W.; Chan, W. P.; Kenny, T. W.; Fearing, R.; Full, R. J. Nature 2000, 405 (6787), 681–685. (3) Jagota, A.; Bennison, S. J. Integr. Comp. Biol. 2002, 42 (6), 1140–1145. (4) Arzt, E.; Gorb, S.; Spolenak, R. Proc. Natl. Acad. Sci. U.S.A. 2003, 100 (19), 10603–10606. (5) Persson, B. N. J.; Gorb, S. J. Chem. Phys. 2003, 119 (21), 11437–11444. (6) Kamperman, M.; Kroner, E.; del Campo, A.; McMeeking, R. M.; Arzt, E. Adv. Eng. Mater. 2010, 12 (5), 335–348. (7) Prowse, M. S.; Wilkinson, M.; Puthoff, J. B.; Mayer, G.; Autumn, K. Acta Biomater. 2011, 7 (2), 733–738. (8) Crosby, A. J.; Hageman, M.; Duncan, A. Langmuir 2005, 21 (25), 11738–11743. (9) del Campo, A.; Greiner, C.; Arzt, E. Langmuir 2007, 23 (20), 10235–10243. (10) Murphy, M. P.; Aksak, B.; Sitti, M. Small 2009, 5 (2), 170–175. (11) Jeong, H. E.; Lee, J. K.; Kim, H. N.; Moon, S. H.; Suh, K. Y. Proc. Natl. Acad. Sci. U.S.A. 2009, 106 (14), 5639–5644. (12) Greiner, C.; del Campo, A.; Arzt, E. Langmuir 2007, 23 (7), 3495–3502. (13) Boesel, L. F.; Greiner, C.; Arzt, E.; Del Campo, A. Adv. Mater. 2010, 22 (19), 2125–2137. (14) Spolenak, R.; Gorb, S.; Arzt, E. Acta Biomater. 2005, 1 (1), 5–13. (15) Greiner, C.; Spolenak, R.; Arzt, E. Acta Biomater. 2009, 5 (2), 597–606. (16) Ting, T. C. T. J. Appl. Mech. 1966, 33 (4), 845–851. (17) Gent, A. N.; Schultz, J. J. Adhes. 1972, 3 (4), 281–294. (18) Andrews, E. H.; Kinloch, A. J. Proc. R. Soc. London, Ser. A 1973, 332 (1590), 385–399. (19) Barquins, M.; Maugis, D. J. Adhes. 1981, 13 (1), 53–65. (20) Maugis, D.; Barquins, M. J. Phys. D: Appl. Phys. 1983, 16 (10), 1843–1874. (21) Crosby, A. J.; Shull, K. R. J. Polym. Sci., Part B: Polym. Phys. 1999, 37 (24), 3455–3472. (22) Shull, K. R. Mater. Sci. Eng., R 2002, 36 (1), 1–45. (23) Lakrout, H.; Sergot, P.; Creton, C. J. Adhes. 1999, 69 (3
4), 307–359. (24) Pokroy, B.; Epstein, A. K.; Persson-Gulda, M. C. M.; Aizenberg, J. Adv. Mater. 2009, 21 (4), 463. (25) del Campo, A.; Greiner, C. J. Micromech. Microeng. 2007, 17, R81–R95. (26) Tan, J. L.; Tien, J.; Pirone, D. M.; Gray, D. S.; Bhadriraju, K.; Chen, C. S. Proc. Natl. Acad. Sci. U.S.A. 2003, 100 (4), 1484–1489. (27) Owens, D. K.; Wendt, R. C. J. Appl. Polym. Sci. 1969, 13 (8), 1741–1749. (28) Kroner, E.; Maboudian, R.; Arzt, E. Adv. Eng. Mater. 2010, 12 (5), 398–404. (29) Tang, T.; Hui, C. Y.; Glassmaker, N. J. J. R. Soc., Interface 2005, 2 (5), 505–516. (30) Persson, B. N. J. J. Chem. Phys. 2003, 118 (16), 7614–7621. (31) Murphy, M. P.; Aksak, B.; Sitti, M. J. Adhes. Sci. Technol. 2007, 21 (12
13), 1281–1296. (32) The radius of curvature of the probe is much larger than the pillar diameter, and the pillar can be considered as a “flat punch” contact contacting a flat substrate. 7759

dx.doi.org/10.1021/la2009336 |Langmuir 2011, 27, 7752–7759