Enhanced Adhesion of Elastic Materials to Small-Scale Wrinkles

Article pubs.acs.org/Langmuir

Enhanced Adhesion of Elastic Materials to Small-Scale Wrinkles Chelsea S. Davis,†,‡ David Martina,§ Costantino Creton,§ Anke Lindner,*,‡ and Alfred J. Crosby*,† †

Polymer Science and Engineering, University of MassachusettsAmherst, 120 Governors Drive, Amherst, Massachusetts 01003, United States ‡ Laboratoire de Physique et Mécanique des Milieux Hétérogènes, UMR 7636 CNRS/ESPCI and §Laboratoire de Sciences et Ingénierie de la Matière Molle, UMR 7615 CNRS/ESPCI, Université Pierre et Marie Curie, Université Paris Diderot, 10 rue Vauquelin, 75005 Paris, France S Supporting Information *

ABSTRACT: The adhesive properties of a material can be greatly affected simply by wrinkling its surface. We show the importance of selecting the wrinkle feature sizes (amplitude, b; and wavelength, λ) that complement the material-defined length scale related to the adhesion energy and modulus. A rigid circular cylindrical punch patterned with aligned wrinkles ranging in amplitude from 0.5 to 5.0 μm with a fixed aspect ratio of 0.1 is used to characterize the adhesion of elastic films of smooth poly(dimethyl siloxane) (PDMS). The cross-linker concentration used to form the PDMS layers is varied to determine the impact of material properties on wrinkled surface adhesion. The elastic films have an average thickness of 240 μm and the average probe radius is 1 mm, leading to a confined contact scenario. The separation stress and work of debonding are presented for each cross-linker concentration with testing rates ranging over 3 orders of magnitude. For stiffer films (10 wt % cross-linker, E' ≈ 3.00 MPa), small wrinkles (b ≈ 0.5 μm) increase the separation stress by nearly 200% relative to a smooth interface whereas large wrinkles (b ≈ 5.0 μm) are shown to reduce adhesion significantly. A substantial increase in the debonding energy is also observed for these small-amplitude wrinkles contacting stiff materials. No discernible impact of wrinkled surface topography on the adhesion of softer (2 and 4 wt % cross-linker, 0.05 MPa < E' < 0.30 MPa) films is measured.

■

INTRODUCTION For reusable adhesives, surface roughness plays an important role in dictating the force and energy required to separate an interface bound by intermolecular forces.1−4 To develop materials with reusable, reliable adhesive properties, wrinkles offer an attractive option for producing well-defined surface roughness. This concept has been demonstrated for bulk wrinkled elastic materials, allowing the role of periodic surface roughness to be determined.5 However, the wrinkle topography effects on the adhesion of a rigid material to a thin, adhesive layer have yet to be determined and are further investigated here. Surface roughness has typically been shown to decrease normal adhesive properties relative to smooth interfaces in elastic systems. This adhesion reduction is often attributed to the decrease in the true contact area at the point of separation because only the tops of the surface features remain in contact.6 The few instances of adhesion enhancement by random surface roughness or wrinkled surfaces on elastic materials have been reported on relatively soft materials with moduli on the order of 10−100 kPa.7−10 In these works, modest adhesion enhancements of roughly 10 to 40% were shown. Briggs and Briscoe utilized randomly rough surfaces to contact smooth elastomers and observed slight increases in the relative separation force of small-scale roughness.7 The materials that Briggs and Briscoe © 2012 American Chemical Society

utilized were soft (0.06 MPa < E < 0.49 MPa), fairly viscoelastic materials, and the slight increase in adhesion energy that they observed can most likely be attributed to a viscous dissipation separation mechanism. They also predicted theoretically that for very low amplitude surface roughness a small increase in normal adhesion should occur. Wrinkles have been used successfully to control the adhesion of several elastic systems.5,11,12 Wrinkles provide a spontaneously forming periodic surface roughness that allows the lateral and height dimensions of the wrinkled structures (peakto-peak wavelength, λ and peak-to-trough amplitude, b) to be predictably controlled. The mechanics governing the formation of wrinkles are well understood.13−15 Several wrinkling techniques allow independently tuned wavelengths and aspect ratios as well as the ability to pattern large areas rapidly.8,16−18 Taking advantage of this elastic surface buckling instability, we can apply surface roughness to various materials through wrinkling techniques and subsequent molding replication steps. In their pioneering wrinkle adhesion work, Chan and coworkers used randomly oriented wrinkles to control the adhesion of butyl acrylate networks and showed increases of Received: June 7, 2012 Revised: August 14, 2012 Published: October 1, 2012 14899

dx.doi.org/10.1021/la302314z | Langmuir 2012, 28, 14899−14908

Langmuir

Article

nearly 50% for small wrinkles.8 More recently, Martina and coworkers have employed periodically buckled surfaces to investigate the adhesive properties of viscoelastic materials.9 This study provided insight into the impact of surface roughness on viscoelastic systems. Here, we present significantly enhanced adhesion in a fairly stiff (E′ ≈ 3.00 MPa) material. We measure the separation force as well as the debonding energy of patterned surfaces relative to the adhesive response of a smooth interface. By employing a wrinkling technique that allows the formation of wrinkles with carefully controlled aspect ratios, we build upon the work of Martina et al.9 An enhancement of nearly 200% relative to a smooth interface is presented. The viscoelastic properties and surface roughness length scales that contribute to enhanced and decreased adhesion of a rigid, wrinkled surface contacting an elastic adhesive layer are identified. To our knowledge, this is the first instance of significant surface-roughness-initiated adhesion enhancement on a stiff, elastic material. These results are significant for industrial applications where fouling and contamination tend to decrease the adhesive performance of soft and/or viscoelastic materials most commonly used for adhesion. Note that throughout this report the terms adhesive and rigid are used to refer to the compliant PDMS layer and resin probe, respectively.

■

EXPERIMENTAL APPROACH

■

FABRICATION OF PROBES

independently of the wrinkle aspect ratio.5,17 To eliminate concerns that residual stresses stored in the buckled surfaces could affect the mechanics of separation, master wrinkle surfaces are replicated with a rigid material. This molding step also allows multiple rigid probes to be fabricated with identical surface features. The topography of the wrinkled surfaces is measured by optical profilometry (Zygo NewView 7300 and Veeco Wyko NT3300). Rigid circular cylindrical probes with a wrinkled surface on one end are formed by utilizing a novel capillary bridge technique that allows the height and radius of each probe to be controlled. Probes are fabricated from a UV-curable resin (Norland Optical Adhesive 63, NOA). Figure 1 contains a schematic of the process used to fabricate rigid circular cylindrical probes. An uncured drop of liquid NOA with an approximate volume of 0.3 mL is placed on a clean 1 cm2 glass slide and suspended dropside down over the master wrinkle surface to be replicated (Figure 1a). Then, using a custom-built sample holder, the drop is lowered toward the master until contact between the drop and the wrinkled surface occurs and a cylindrical capillary bridge of liquid NOA is formed (Figure 1b).19 The height of the glass slide is adjusted until the desired probe radius, c, is achieved (c ≈ 1 mm). The NOA is then exposed to ultraviolet (UV) light (Jelight, 30 mW/mm2) for 4 min to cure the resin. The interface between the cross-linked NOA probe and the wrinkled master is separated at a velocity of approximately 1 mm/s at a 30° peel angle, and the probe is exposed to UV for an additional 2−4 min to ensure the completion of the resin curing process (Figure 1c). This amount of UV exposure far exceeds the exposure time recommended by the manufacturer but is necessary to ensure the perfect replication of the micrometer and submicrometer surface features and eliminate undesired nanoscale surface roughness. The fidelity of the replicated rigid wrinkled probes to the wrinkled master is verified with optical profilometry (Veeco Wyko NT3300), and the profile of the wrinkled probe surface is measured directly as shown in Figure 2c. This wrinkle replication technique can also be utilized to form molded wrinkles on elastic materials. (More details on this process are provided in the Supporting Information.)

We investigate the impact of wrinkle feature size and film modulus on the adhesion of a rigid material to a compliant layer. Specifically, the adhesive impact of rigid surface wrinkles with amplitudes ranging from 0.5 (small) to 5.0 μm (large) and a constant aspect ratio of approximately 0.1 contacting a smooth, thin adhesive layer is measured. The modulus of each thin adhesive substrate is varied by changing the degree of cross-linking through different cross-linker to prepolymer ratios (2 to 10 wt % cross-linker). Controlling the crosslinker percentage adjusts the film storage modulus, E′, from 0.05 (soft) to 3.00 MPa (stiff). For all adhesion tests, the normal force, P, vertical displacement, δ, and contact area, A, are recorded simultaneously as the surfaces are brought into contact and subsequently separated.

Master wrinkle surfaces are formed by utilizing previously reported methods that allow the wavelength to be tuned

■

FABRICATION OF PDMS FILMS The viscoelasticity and modulus of the adhesive film are controlled by varying the concentration of cross-linker added to a two-part elastomer system. When the amount of cross-linking agent available during the curing of the elastomer is changed, the molecular weight between cross-links is altered, affecting the viscoelasticity and modulus of each film. Previous works have thoroughly characterized the rheological properties20 of the elastomers used in these adhesion experiments (Supporting Information 4). Smooth films of poly(dimethyl siloxane) (PDMS) (Dow Corning Sylgard 184) are formed on glass slides. Three compositions of PDMS (2, 4, and 10 wt % cross-linker in prepolymer) are mixed and degassed in a reduced-pressure environment for 15 min. Films are cast by applying the uncured PDMS mixture to a clean glass slide using a doctor blade and a 400 μm spacer to control the film thickness. Cast films are immediately placed in an oven under vacuum for 5 h at 70 °C. Upon removal from the oven, the films cool for several hours before the film thickness is characterized with a pneumatic thickness-measurement device. The final cured films range in

Figure 1. Schematic of rigid wrinkled probe fabrication process. 14900


Langmuir

Article

Figure 2. Wrinkle pattern topography. (a, b) Optical micrographs of rigid wrinkle surfaces for (a) b = 0.5 μm, λ = 5.0 μm and (b) b = 5.0 μm, λ = 48.0 μm. The scale bar applies to a and b. (c) Optical profilometer scan of the surface in image a illustrating the aspect ratio (b/λ ≈ 0.1) of the wrinkles. The lateral dimensions are 53 × 71 μm2, and the peak-to-trough amplitude is 0.48 μm.

thickness from 170 to 357 μm with an average thickness of 240 μm. To provide a smooth, clean surface for adhesion experiments, the PDMS films are removed from the slide on which they were cured and inverted onto a freshly cleaned glass slide just prior to testing.

Table 1. Range of Experimental Variables

■

ADHESION EXPERIMENTS The normal adhesion of rigid wrinkled surfaces is characterized using a custom-built contact adhesion testing apparatus (Figure

characteristic

minimum

maximum

wavelength (λ, μm) amplitude (b, μm) aspect ratio (b/λ) probe radius (c, mm) substrate thickness (h, μm) confinement (c/h) substrate modulus (E′, MPa) cross-linker percentages (x, %) testing velocity (V, μm/s)

5.0 0.47 0.08 0.96 170 3.1 0.05 2 0.1

48.0 4.20 0.13 1.18 357 5.9 3.00 10 100.0

probe and film and is held for 1 s. The contacting surfaces are then pulled away from one another until complete interfacial separation occurs. A fiber optic displacement sensor (Philtec D63 LPT) monitors the vertical displacement of the probe whereas a load cell attached to the probe records normal forces at the interface. Throughout the test, the contact area is imaged with an optical microscope (Zeiss Axio Vario) and a CCD camera (AVT Manta, HIRIS imaging software). To address concerns of chain transfer arising from the use of unextracted, lightly cross-linked PDMS films, 12 adhesion experiments are conducted sequentially on various locations of each PDMS film without cleaning the NOA probe between tests. Following these 12 tests, the probe is rinsed for 30 s with ethanol and dried with compressed air for 2 min, and then 12 additional adhesion tests are conducted. For each cross-linker concentration, no discernible differences are observed in the adhesive behavior among tests 1, 12, 13, and 24, indicating that no change in chain transfer significantly influences the adhesion results presented here.

Figure 3. Schematic of the contact adhesion testing device.

3).21,22 Over the course of a test, a rigid cylindrical probe is brought into contact with an adhesive film at a fixed testing velocity, V, until full contact of the wrinkled peaks and troughs is achieved with the flat film. The resulting maximum compressive force, Pm, is varied as a function of the crosslinking density to achieve full interfacial contact between the 14901


Langmuir

Article

Figure 4. Contact adhesion results. (a) Force versus displacement curves for the three cross-linker concentrations studied. Tests are conducted on the same wrinkled surface (b = 0.5 μm) at a controlled testing velocity of 1 μm/s. (b) Tensile portion of plot a showing the adhesive response of various cross-linker ratios. (c) Representative force versus displacement curves for three wrinkle amplitudes (b = 0.5, 3.0, 5.0 μm). Tests are conducted on a 10% cross-linker PDMS film with a controlled testing velocity of 1 μm/s. (d) Tensile portion of plot c showing the adhesive response of various wrinkle amplitudes.

Figure 5. Adhesion results as a function of testing velocity. Separation stress, σs (a−c), and work of debonding, wdeb (d−f), as a function of velocity for three wrinkle amplitudes (legend in plot a). The green and red boxes contain the results for the two lightly cross-linked (2 and 4%) adhesives. The blue box encompasses the plots for the films with the highest cross-linker concentration (10%).

The compliance (displacement per unit force, C = ∂δ/∂P) of the experimental setup impacts normal contact adhesion experimental results.23 The compliance of the experimental setup is measured independently and accounted for in the

analysis of the data obtained for each adhesion test using a custom macro (Igor Pro 5.0). Additionally, adhesion tests of a smooth, rigid probe in contact with each smooth PDMS film (2, 4, and 10 wt % cross-linker) are completed to establish the 14902


Langmuir

Article

work of debonding is dependent on the wrinkle amplitude and is greater for the smallest wrinkles. wdeb increases dramatically with increasing testing rate, especially for the more lightly crosslinked layers. However, this phenomenon is independent of surface roughness and is due to the dissipative viscoelastic response of the material.25,26

adhesive properties of a smooth interface for each cross-linking concentration. The values measured on the smooth probes are used to analyze the wrinkle adhesion results and determine the effect of wrinkles on adhesion. The adhesive response of these smooth interfaces is provided in the Supporting Information as a function of testing velocity for each cross-linking density. For our experiments, the wrinkle feature size, cross-linker concentration of the adhesive layer, and testing velocity are varied systematically. A summary of the experimental parameters and their ranges is provided in Table 1.

■

DISCUSSION The two main findings of this work are that (1) for stiffer materials (10% cross-linker), small wrinkles enhance adhesion

■

ADHESION RESULTS Normal force as a function of vertical displacement is presented in Figure 4a for varying cross-linker concentrations. Negative force values are compressive, while positive force values are tensile or adhesive. Zero displacement (δ = 0) corresponds to zero force upon retraction. The same data is provided in Figure 4b with the compressive portion of the curve omitted to magnify the region of interest, specifically, the adhesive regime of each test. All three of these tests are conducted at a fixed testing velocity of V = 1 μm/s using the same rigid wrinkled probe (b = 0.5 μm). For the most highly cross-linked adhesive layer (10%, stiff), the peak separation force, Pc, is greater than for the two more lightly cross-linked materials (2 and 4%, soft). Similarly, Figure 4c,d shows the relationship between the normal force and the displacement measured on highly crosslinked (10%) PDMS films with three different wrinkle amplitudes. This allows the stark contrast in the adhesive response of the various wrinkle amplitudes to be more easily observed. The peak separation force for the smaller-amplitude wrinkles (b = 0.5 μm) is much greater than that for the largest wrinkles (b = 0.5 μm). Figure 5 contains a comprehensive summary of the results obtained in this study. As a function of the testing velocity, V, Figure 5a−c shows the impact of wrinkle amplitude on the separation stress (σs = Pc/πc2) for the three cross-linker compositions tested. The wrinkle amplitude does not affect the adhesion of the two softer layers (2 and 4%) but has a dramatic impact on the separation stress of the most highly cross-linked material (10%). As shown in previous wrinkle adhesion studies,5,11,12,24 for the 10% films, wrinkles with small amplitudes require more force to separate the interface than larger amplitude wrinkles. Through integration of the area under the tensile or adhesive portion of the force versus displacement curves, a measure of the energy required to separate the interface (i.e., the debonding energy, Udeb) is obtained: Udeb =

∫0

Figure 6. Effective separation stress normalized by the effective separation stress of a smooth, nonpatterned interface and the adhesive film modulus as a function of (a) wrinkle amplitude and (b) testing velocity. In plot a, results are for tests run at a testing velocity of 1 μm/ s. The dashed lines are a guide to the eye and do not represent numerical fits. The solid black lines are y = 1.0. Values above this line represent adhesion scenarios where the separation stress of the wrinkled interface is greater than that of a smooth interface whereas values falling below unity indicate a decrease in wrinkle adhesion relative to the smooth interface. The error bars represent 10% of an average value taken over three tests.

δ

P ∂δ

(1)

whereas larger wrinkles decrease adhesion as previously reported5,11 and (2) wrinkles or surface roughness on the length scales tested here do not significantly impact the adhesive properties for the 2 and 4% cross-linker compositions. These two results are readily observed in Figure 6.

Although the separation force is higher for the most highly cross-linked adhesive film (10%), the softer films (2 and 4%) are more dissipative and require more energy to separate the interface though less total force is required as shown in Figure 5d−f as a function of the testing velocity where wdeb = Udeb/πc2. Both the separation force and debonding energy descriptors are normalized by the projected area of the probe to account for small variations in the size of the rigid probes as a result of the probe fabrication method described in Figure 1. As for the separation stress data, wdeb is independent of the size of the wrinkled surface features for the two lightly crosslinked films. For the more highly cross-linked PDMS film, the

■

MATERIAL LENGTH SCALE RELATIONSHIPS A material-defined length scale incorporating both the critical strain energy release rate, Gc, and modulus, E, can be used to describe the distance over which adhesive forces act between two contacting surfaces.27 Here, Gc values are determined by averaging the results of five adhesion tests performed by contacting a flat, rigid cylindrical punch with a smooth adhesive 14903


Langmuir

■

Table 2. Material-Property-Defined Length Scalesa cross-linker (wt %)

Gc (J/m )

E′ (MPa)

h (μm)

δc (μm)

2

0.419

0.048

4

0.370

0.291

10

0.142

3.000

240 ∞ 240 ∞ 240 ∞

45.77 8.73 17.47 1.27 3.37 0.05

2

Article

ADHESION OF LIGHTLY CROSS-LINKED FILMS For the lightly cross-linked (2 and 4%) soft PDMS films, δc is larger than the wrinkle amplitude (δc > b). For these two

δc values indicating the material length scale over which adhesion dominates the contact for various PDMS cross-linker concentrations. The tensile storage modulus, E′, values are obtained from rheological experiments.20 The average film thickness is used to calculate the critical displacement in confinement. a

film. These tests are performed at V = 1.0 μm/s to approximate thermodynamic equilibrium at which Gc = wdeb27 For our viscoelastic materials, the most appropriate modulus value to consider is the storage modulus at a testing frequency that corresponds to the normal adhesion testing velocities employed.28,29 The shear storage modulus, G′, is obtained from previously reported cone and plate rheology (2 and 4%) and dynamic mechanical testing experiments (10%) (Supporting Information 4).20 To determine the angular frequency, ω, that corresponds to the contact adhesion testing velocities employed in this study, the relationship between the frequency and strain rate, ε̇, is used30

ω ≈ ε̇ =

V lo

(2)

where lo is the original length over which the strain is applied. In our case, we have used a vertical displacement in the compressive regime of our test that averaged lo ≈ 20 μm for most tests. G′ values at an angular frequency of ω = 0.05 are used that approximate an adhesion testing velocity of V = 1.0 μm/s. Finally, the simple relationship between the shear and tensile moduli, E′ = 3G′, is used to obtain values for the tensile storage modulus of each PDMS formulation. Over the range of testing velocities employed in this study, the frequency dependence of E' is extremely weak allowing the simple approximation of eq 2. The material-defined length scale for the adhesion of bulk materials can be considered to be the critical displacement for infinitely thick systems needed to separate the interface, δ∞ c :

Figure 7. Contact images at Pc for 4 wt % PDMS films tested with rigid probes with (a) smooth and (b) small (b = 0.5 μm) wrinkles and (c) large (b = 5.0 μm) wrinkles. Tests are performed at V = 1 μm/s, and the light areas in the center of each image are the portion of the probe still in contact with the film surface. The scale bar in image a applies to all three images. Figures 8f and 9f contain equivalent contact images at separation for 10 wt % PDMS contacting small (Figure 8f) and large (Figure 9f) wrinkles.

Gc (3) E′ However, in the case of confined contact geometries where c/h ≥ 0.45, finite size effects must be considered. The influence of confinement on the adhesion of viscoelastic films in contact with flat, cylindrical probes has been treated previously.31 δc∞ ≈

⎛ G h ⎞1/2 δct ≈ ⎜ c ⎟ ⎝ E′ ⎠

materials, adhesive forces act over longer distances, leading to a large deformation of the material and masking the effects of the small wrinkle features. The wrinkle topography at the interface is effectively smoothed out so that the separation is similar to that of a smooth interface (Figure 5). The entire surface is brought into contact with very little elastic penalty so that the topography does not play a role in the separation mechanism. For 4% PDMS, the larger wrinkle amplitudes (b = 3.0 and 5.0 μm) are greater than δ∞ c , suggesting that a slight influence of wrinkle topography should be expected. In fact, Figure 5b does show a very slight difference between the separation stress of

(4)

Table 2 contains a summary of the material properties and the critical displacement in confinement, δtc for the three PDMS cross-linker ratios tested in this study. Allowing for the viscoelastic nature of the PDMS films utilized in this study, we find that the actual value of the critical displacement for each cross-linker ratio is most likely between t the confined displacement and that of the bulk (δ∞ c ≤ δc ≤ δc). 14904


Langmuir

Article

Figure 8. Sequence of contact images obtained in adhesion testing of small wrinkles (b = 0.5 μm) on a 10 wt % PDMS film at a testing velocity of 1 μm/s. Images are obtained by reflective microscopy imaging through the PDMS film and focusing on the interface as the probe is brought into contact and subsequently separated. (a−f) Images of the full probe with (a′−f′) magnified views of the contact area. (The white rectangle in image f indicates the location of insets.) Lighter areas (b−f) represent interfacial contact. Time sequence: (a) surfaces are out of contact, (b) the interface is forming along wrinkle peaks (the black line illustrates the contact line front that is moving radially inward (center of probe is not in contact)), (c) maximum compression, (d) initiation of cracks along wrinkle troughs as surfaces begin to separate (cracks are observed as thin, dark lines running parallel to wrinkles), (e) cracks continue to nucleate and propagate, and (f) the peak separation force. The scale bar in image a applies to images a−f and is 500 μm whereas the scale bar in image a′ applies to images a′−f′ and is 100 μm.

small and large wrinkles for the 4% PDMS. The δc values presented here are not meant to be taken as absolute values but rather as general criteria to determine the size scale of surface roughness required to impact adhesion. Figure 6 further illustrates the independence of separation stress on wrinkle amplitude for lightly cross-linked films. This plot shows effective separation stresses (σ̅s = σs/E′) for large and small wrinkles normalized by the effective separation stress of a smooth probe contacting the same smooth film. Regardless of the testing velocity or feature size, the normalized adhesion values do not deviate substantially from unity for the 2 and 4% PDMS films, indicating that there is little to no change in adhesive properties due to the wrinkle topography. To gain a better understanding of the separation process, observations of the contact area over the course of each test are utilized. Figure 7 provides contact images that correspond to the maximum separation force for the 4% cross-linker PDMS. The separation mechanisms are all qualitatively similar for smooth (Figure 7a) and small wrinkles (Figure 7b) and large wrinkles (Figure 7c). It should be noted that the contact line or

perimeter of the contact on the large wrinkles is jagged, indicating that the contact area is conforming to the peaks and troughs of the wrinkles. However, this effective increase in the length of the contact line is negligible relative to the bulk separation mechanism of the interface.

■

ADHESION OF HIGHLY CROSS-LINKED FILMS For the most highly cross-linked (10%) stiff PDMS film, δc < b. Thus, the wrinkle topography effects are not screened or smoothed out as for the more lightly cross-linked interfaces, altering the separation of the interface. Our results, most clearly observed in Figure 6a, lend credence to this assertion and illustrate the inverse dependence of separation stress on wrinkle amplitude. In direct contrast to the wrinkle-independent adhesive response of lightly cross-linked films described above, the adhesion of the highly cross-linked film is greatly affected by the wrinkled surface topography as shown in Figure 6. The smallest wrinkles (b = 0.5 μm) show a marked increase in the separation stress relative to that of the smooth interface. Alternatively, for 14905


Langmuir

Article

Figure 9. Image sequence of contact images obtained in adhesion testing of large wrinkles (b = 5.0 μm) on a 10 wt % PDMS film run at a testing velocity of 1 μm/s. (a−f) Full probe and (a′−f′) magnified views of contact areas. (The white rectangle in image f indicates the location of insets for all images.) Lighter areas show interfacial contact. Time sequence: (a) surfaces are out of contact, (b) the interface is forming along wrinkle peaks, (c) maximum compression, (d) wrinkle troughs begin to separate, (e) separation continues in wrinkle troughs with cracks propagating along troughs parallel to aligned wrinkles, and (f) the peak separation force. The scale bar in image a applies to images a−f and is 500 μm whereas the scale bar in a′ applies to images a′−f′ and is 100 μm.

the largest wrinkles tested (b = 5.0 μm), the separation stress is reduced relative to a smooth surface More insight into these wrinkle amplitude effects can be obtained by again examining the contact images captured during adhesion testing. The separation mechanism of the rigid probes with small wrinkle features (b = 0.5 μm) from fully cross-linked films differs significantly from that of the more lightly cross-linked films. Figure 8 is a time sequence of the contact images for small wrinkles on stiff films. The separation of these interfaces occurs in two steps. Multiple small cracks occur at the interface along small portions of the wrinkle troughs across the entire probe face (Figure 8d,e). These small crack openings require a significant amount of energy to form, leading to higher debonding energy results relative to smooth surfaces. Once a critical portion of the interface separates along the troughs, full or bulk interfacial failure occurs in a manner similar to the separation of a smooth interface (similar to Figure 7a). Namely, the entire interface fails in a single, unstable crack propagating from the probe edges radially inward toward the center of the probe face, independent of wrinkle troughs and peaks (Figure 8f). Videos of the two tests illustrated in Figures 8 and 9 can be found in the Supporting Information, allowing the contact area as well as the dynamics of the separation to be more easily observed.

Alternatively, a time sequence of the contact images for the large-amplitude wrinkles (Figure 9) illustrates a separation mechanism from highly cross-linked films that differs greatly from that of the small-amplitude wrinkles. After the application of the maximum compressive load, the entire wrinkled surface of the probe is in contact with the adhesive film (Figure 9c). The displacement is then reversed, and while still in macroscopic compression (P < 0), the troughs of the large wrinkles separate from the film (Figure 9d). This trough separation typically initiates at the edges of the probe and runs along the wrinkle direction inward toward the center of the contact. Subsequently, as the surfaces are pulled further apart, the peaks of the wrinkles separate from the film surface (Figure 9e,f). Only the peaks of the wrinkles exert a tensile or adhesive force on the interface, greatly reducing the adhesion of the rough surface relative to the smooth separation stress. This two-step separation mechanism is the driving force dictating the reduced adhesive properties of large wrinkles on elastic materials and has been predicted theoretically25 and recently observed experimentally.9 14906


Langmuir

■

TESTING RATE EFFECTS The effect of the wrinkle amplitude on the highly cross-linked films is more pronounced at very slow testing velocities as shown in Figure 6b. The more quickly the interface is separated, the closer the separation stress is to that of a smooth interface.32 Thus far, the critical strain energy release rate has been taken as a constant material property. However, it is actually a velocity-dependent parameter affected by the viscoelasticity and testing velocity. A more accurate expression for the effective critical strain energy release rate is29,33−35 ⎛ ⎛ V ⎞n ⎞ ⎟ ⎟ Gc = Go⎜1 + ⎜ ⎝ V* ⎠ ⎠ ⎝

(5)

(6)

■

CONCLUSIONS The impact of rigid wrinkles on the adhesion of smooth PDMS films with varying cross-linker ratios is presented here. For lightly cross-linked materials, adhesive forces dominate the behavior of the interface over length scales larger than the wrinkle amplitudes tested and wrinkles are subsequently shown to have little to no effect on adhesion. For highly cross-linked, stiffer materials, adhesive forces affect distances shorter than the wrinkle amplitudes, thus wrinkles greatly impact the adhesion behavior of these surfaces. Small wrinkles increase the separation stress and debonding energy nearly 200% over smooth surface adhesion values whereas large wrinkles decrease these adhesion metrics relative to those of smooth surfaces. ASSOCIATED CONTENT

S Supporting Information *

Adhesion results (both separation stress and the work of debonding as a function of testing velocity) for smooth interfaces between a rigid cylindrical indenter and PDMS films with varying cross-linker concentrations. Experimental details and adhesion results for adhesive wrinkles indented by rigid, flat surfaces (the inverse of the contact scenarios described here). Contact videos for the most highly cross-linked film contacting both small and large wrinkle surfaces. Rheological data for varying cross-linker concentrations of PDMS. This material is available free of charge via the Internet at http://pubs.acs.org.

■

■

REFERENCES

(1) Briggs, G.; Briscoe, B. Nature 1976, 260, 313−315. (2) Fuller, K. N. G.; Tabor, D. Proc. R. Soc. London, Ser. A 1975, 345, 327−342. (3) Persson, B. N. J.; Tosatti, E. J. Chem. Phys. 2001, 115, 5597− 5610. (4) Hui, C.-Y.; Lin, Y. Y.; Baney, J. J. Polym. Sci., Part B: Polym. Phys. 2000, 38, 1485−1495. (5) Davis, C. S.; Crosby, A. J. Soft Matter 2011, 7, 5373−5381. (6) Persson, B. N. J.; Albohr, O.; Tartaglino, U.; Volokitin, A. I.; Tosatti, E. J. Phys.: Condens. Matter 2005, 17, R1−R62. (7) Briggs, G.; Briscoe, B. J. Phys. D: Appl. Phys. 1977, 10, 2543− 2466. (8) Chan, E. P.; Smith, E. J.; Hayward, R. C.; Crosby, A. J. Adv. Mater. 2008, 20, 711−716. (9) Martina, D.; Creton, C.; Damman, P.; Jeusette, M.; Lindner, A. Soft Matter 2012, 8, 5350−5357. (10) Guduru, P.; Bull, C. J. Mech. Phys. Solids 2007, 55, 473−488. (11) Kundu, S.; Davis, C. S.; Long, T.; Sharma, R.; Crosby, A. J. J. Polym. Sci., Part B: Polym. Phys. 2011, 49, 179−185. (12) Lin, P.-C.; Vajpayee, S.; Jagota, A.; Hui, C.-Y.; Yang, S. Soft Matter 2008, 4, 1830−1835. (13) Bowden, N.; Whitesides, G. M.; Huck, W. T. S.; Paul, K. E. Appl. Phys. Lett. 2001, 75, 2557−2559. (14) Harrison, C.; Stafford, C. M.; Zhang, W.; Karim, A. Appl. Phys. Lett. 2004, 85, 4016−4018. (15) Allen, H. G. Analysis and Design of Structural Sandwich Panels; Pergamon Press: Oxford, England, 1969; Vol. 51. (16) Yang, S.; Khare, K.; Lin, P.-C. Adv. Funct. Mater. 2010, 20, 2550−2564. (17) Miquelard-Garnier, G.; Croll, A. B.; Davis, C. S.; Crosby, A. J. Soft Matter 2010, 6, 5789−5794. (18) Stafford, C. M.; Vogt, B. D.; Harrison, C.; Julthongpiput, D.; Huang, R. Macromolecules 2006, 39, 5095−5099. (19) De Souza, E. J.; Gao, L.; McCarthy, T. J.; Arzt, E.; Crosby, A. J. Langmuir 2008, 24, 1391−1396. (20) Nase, J. Debonding of Viscoelastic Materials: From a Viscous Liquid to a Soft Elastic Solid. Ph.D. Thesis, Universitat des Saarlandes, Universite Pierre et Marie Curie, 2009. (21) Crosby, A. J.; Shull, K. R.; Lakrout, H.; Creton, C. J. Appl. Phys. 2000, 88, 2956−2966. (22) Chiche, A.; Pareige, P.; Creton, C. C. R. Acad. Sci., Ser. IV: Phys., Astrophys. 2000, 1, 1197−1204. (23) Shull, K. R.; Creton, C. J. Polym. Sci., Part B: Polym. Phys. 2004, 42, 4023−4043. (24) Jin, C.; Khare, K.; Vajpayee, S.; Yang, S.; Jagota, A.; Hui, C.-Y. Soft Matter 2011, 7, 10728−10736. (25) Hui, C.-Y.; Lin, Y. Y.; Creton, C. J. Polym. Sci., Part B: Polym. Phys. 2002, 40, 545−561. (26) Nase, J.; Creton, C.; Ramos, O.; Sonnenberg, L.; Yamaguchi, T.; Lindner, A. Soft Matter 2010, 6, 2685−2691. (27) Shull, K. R. Mater. Sci. Eng., R 2002, 36, 1−45. (28) Shull, K. R.; Flanigan, C. M.; Crosby, A. J. Phys. Rev. Lett. 2000, 84, 3057−3060. (29) Josse, G.; Sergot, P.; Creton, C.; Dorget, M. J. Adhes. 2004, 80, 87−118.

highlights the relationship between Gc and V. As the testing velocity increases, the effective critical energy release rate increases as well. Thus, the critical displacement, δc increases with the testing rate so that the wrinkles have a diminished effect.

■

ACKNOWLEDGMENTS

We thank F. Monti of the ESPCI ParisTech for his help with optical profilometry measurements. We also thank Dr. J. Nase for graciously authorizing the use of her rheological results. C.S.D. acknowledges M. Bartlett of the University of Massachusetts for fruitful discussions and a careful reading of the manuscript. This work was supported by Michelin, the NSF-PIRE Korean Exchange (NSF-0730243), and the National Science Foundation MRSEC at the University of MassachusettsAmherst (NSF-DMR-0820506).

where Go is the critical strain energy release rate at zero velocity corresponding to the thermodynamic work of adhesion, V* is a parameter related to the dissipative properties of the soft viscoelastic material, and n is a fitting parameter (typically 0.4 < n < 0.6 for soft, fairly elastic materials).18,30,34,35 We do not discuss the particulars of eq 5 in detail here, but a simplification of this expression34,35 Gc ≈ Go + Go(V )

■

Article

AUTHOR INFORMATION

Corresponding Author

*[email protected], [email protected]. Notes

The authors declare no competing financial interest. 14907


Langmuir

Article

(30) Fabbroni, E. F.; Shull, K. R.; Hersam, M. C. J. Polym. Sci., Part B: Polym. Phys. 2001, 39, 3090−3102. (31) Webber, R. E.; Shull, K. R.; Roos, A.; Creton, C. Phys. Rev. E 2003, 68, 021805−11. (32) Poulard, C.; Restagno, F.; Weil, R.; Léger, L. Soft Matter 2011, 7, 2543−2551. (33) Maugis, D.; Barquins, M. J. Phys. D: Appl. Phys. 1978, 11, 1989− 2023. (34) Shull, K. R.; Ahn, D.; Chen, W.; Flanigan, C. M.; Crosby, A. J. Macromol. Chem. Phys. 1998, 199, 489−512. (35) Carelli, C.; Deplace, F.; Boissonnet, L.; Creton, C. J. Adhes. 2007, 83, 491−505.

14908


Enhanced Adhesion of Elastic Materials to Small-Scale Wrinkles

Recommend Documents