Tailored polyurethane acrylate blend for large-scale and high-performance micropatterned dry adhesives

For the micropatterned dry adhesives, pre-polymer blends based on aliphatic urethane diacrylate oligomers, UA16 (Miramer UA5216, Miwon Specialty Chemical Co. Ltd., Gwanggyo, South Korea) were used. The photoinitiator Omnirad 500 (IGM Resins B.V., Waalwijk, The Netherlands) was added to all UA16-based blends in a concentration of 5 wt%. For different formulations, modulators were added: U6L7 contained 10 wt% Laromer LR 8887 (BASF AG, Ludwigshafen, Germany); U6S9 contained 1 wt% Miramer SIP900 (Miwon Co.); U6E4 contained 20 wt% Miramer ES4420 (Miwon Co.). All compositions for each pre-polymer blend are summarized in Table 1. All blends were mixed at 2350 rpm and degassed at 1 mbar for 3 min using a SpeedMixer (DAC600.2 VAC-P, Hauschild & Co. KG, Hamm, Germany).

For the flexible templates, two types of silicones were used: Elastosil M4601 (nPDMS, Wacker Chemie AG, München, Germany) and Elastosil M4644 (tPDMS, Wacker AG). The two-component polydimethylsiloxanes were mixed at 2350 rpm and degassed at 1 mbar for 3 min using a SpeedMixer. The pre-polymers were cast into a circular mold (diameter 130 mm; depth 5 mm), in which the master structure was mounted. A micropatterned Ni-shim (Temicon GmbH, Dortmund, Germany) was used as a master structure that comprised mushroom-shape micropillars with diameters and heights of 50 µm. The diameters of the mushroom tips were in the range of 66–76 µm. The pillar array was hexagonally arranged with a center-to-center distance of 100 µm. The mushroom tips hence cover an area fraction of about 42.9%. The PDMS pre-polymer was then cured at 70 °C for 1 h, and the negative microstructure was received after gentle demolding. Several PDMS stamps were prepared by repeating this procedure and then trimmed into square shape, and eventually glued onto the imprinting roll.

Viscosities of the pre-polymer blends were measured with a rheometer (Physica MCR300, Anton Paar GmbH, Graz, Austria) in the temperature range between 25 and 80 °C at a constant shear rate of 1 s⁻¹. For all materials, dynamic mechanic thermal analyses (DMTA) were performed in the temperature range from − 100 to 120 °C in tensile mode at an oscillatory frequency of 1 Hz under N₂ atmosphere. The sample dimensions were 30 mm × 5.5 mm × 2 mm. Elongations at break were measured by (quasi-static) tensile tests (Material Testing Machine Xforce P, Zwick & Roell GmbH & Co. KG, Ulm, Germany) with a crosshead moving rate of 20 mm min⁻¹. Dog-bone-shaped specimens with a thickness of 3 mm exhibited a gage length of 60 and a width of 8 mm. The width of the grips was 17 mm. For all mechanical tests, UA16 blends were cured by UV exposure (365 nm) for 5 min; the silicones were thermally cured at 70 °C for 1 h.

Surface free energies were examined by contact angle goniometry (OCA 35, analysis software: SCA20; DataPhysics Instruments GmbH, Filderstadt, Germany) using water and hexadecane as testing liquids. Smooth films with 120 µm thickness for static contact angle measurements were prepared by coating UA16 polymer blends or PDMS pre-polymers onto flat PET substrates, following by 5-min UV exposure for UA16 polymer blends and thermal curing at 70 °C for PDMS. The surface free energy was deducted from the obtained contact angles by using Wu’s harmonic method [45, 46].

UV-curing rates of the polymer blends were measured by UV-DSC (Mettler Toledo GmbH, Gießen, Germany) under N₂ atmosphere. For each blend, 8–10 mg was tested in an aluminum pan without a lid at 25 °C. The heat flow was recorded over a time frame of 6 min (2 min before UV exposure, 3-min UV exposure, 1 min after exposure).

For scanning electron microscopy (SEM), the samples were sputter-coated with a conductive gold layer at 30 mA for 40 s using a Jeol JFC-1300 auto fine coater (Jeol Ltd., Akishima, Japan). Secondary electron images were acquired at 5 kV accelerating voltage in high vacuum conditions using an FEI Quanta 400 FEG (Thermo Fisher Scientific, Waltham, USA).

For continuous fabrication, a roll-to-roll system (Jakob Weiß & Söhne GmbH, Sinsheim, Germany) with a total web length of 24 m was used. The system consists of several units including coating, thermal drying, imprinting and UV curing and delamination. The imprinting unit is the core of the machine, which includes an imprinting roll and two backing rolls placed on each side; a pressure roll pressed the pre-polymer (i.e., wet coating) against the imprinting roll, and a deflection roll guided the delamination (Fig. 1a). The flexible transparent (tPDMS) or non-transparent (nPDMS) silicone templates were wrapped onto the imprinting roll. The UV-LED radiator with a maximum power of 6 W cm⁻² at 365 nm was placed under the imprinting roll. 30% or 50% of UV power was used for different trials for comparison purposes. Detailed processing parameters for each experiment are summarized in Table 2.

In the present work, the roll-to-roll manufacturing followed a two-step process as illustrated in Fig. 1b: First a non-patterned film (UA16 or U6L7) with a 30 µm wet coating thickness was formed onto the PET substrate (75 µm thickness, Melinex 506, DuPont Teijin Films Europe, Luxemburg) by covering the wet film with a release foil (polyethylene film, 23 µm Hostaphan MP 23, Mitsubishi Polyester Film GmbH, Wiesbaden, Germany) to protect the wet film against direct contact with the micro-patterning molds. After UV curing of this non-patterned ground layer, the release foil was removed. In a second step, the microstructures were imprinted based on coatings of UA16, U6S9 or U6E4 (90 or 120 µm wet coating thickness) as top coating on the ground layer. During imprinting, the distance between the ground layer and the flexible template, d, was adjusted by moving the pressure roll to control the filling of the microcavities by the pre-polymer. Positive values of d mean that there was a gap between the ground layer (without considering the top layer) and the flexible template, whereas negative values represent the flexible template being compressed. Upon UV curing, the microstructures were demolded from the flexible template.

Adhesion tests were carried out on a custom-built testing device as illustrated in Fig. 2a. The device is composed of an interferometer, a pivotable stage and a smooth glass sphere with curvature radius of 15 mm mounted on a glass beam with a spring constant of 2241 N m⁻¹. Forces were calculated based on the beam deflection and the spring constant [24]. Samples with a size of 15 mm × 15 mm were cut from the micropatterned films and subsequently glued to a glass slide with a UV adhesive (BO MV76002, Bohle AG, Haan, Germany). A typical force–displacement curve for a normal pull-off experiment using a spherical probe is shown in Fig. 2b. The glass sphere was brought in contact with the samples at a velocity of 5 µm s⁻¹ until a compressive preload of 30 mN was reached. The apparent contact area, A, was estimated from the radius of the glass sphere and the indentation depth (which is the difference between the displacement at preload and that achieved at first contact) from the equation: \( A = \pi \left[ {R^{2} - (R^{2} - \delta^{2} )} \right] \), where R is the radius of the probe and \( \delta \) is the indentation depth at preload. After preloading, the sphere was retracted in normal direction at a constant velocity of 5 µm s⁻¹. The pull-off force was defined as the maximum tensile force before detachment. Pull-off stresses were determined by dividing pull-off forces by the apparent contact area, A, as described above. The work of separation was calculated by integrating the enclosed area of the stress–displacement curve in the tensile regime (see Fig. 2b). Six positions on each sample were measured, and the mean values were recorded.

Peel tests on smooth, flat glass substrates with a 90° peel angle were performed with 24-mm-wide and 240-mm-long films. The peeling speed was 5 mm s⁻¹. For comparison with our tapes, magic tape (Scotch^® Magic™ Tape, 3 M Deutschland GmbH, Neuss, Germany) was measured under the same condition. The normalized peel force was determined by dividing the measured forces by the width of the tapes.

Shear strength was examined by lap shear test in a tensile tester (Material Testing Machine Xforce P, Zwick & Roell GmbH & Co. KG, Ulm, Germany). The adhesive strip was attached on a glass plate with a contact area of 20 mm × 50 mm. The crosshead speed was 5 mm min⁻¹. Shear strength was calculated by dividing the maximum shear force by the contact area.

The two-step replication, as shown in Fig. 1b, was successfully implemented into the continuous roll-to-roll process: First, the plain ground layer (thickness 30 µm) was coated onto the PET carrier film. Second, the micropatterned adhesive layer was added. To study how the morphology as well as the adhesive properties of the composite film was affected by the thickness of the second layer, three different conditions were investigated: “90_(− 245 µm),” “120_(− 245 µm)” and “120_(− 303 µm).” The first number in the notation represents the layer thickness of the second wet layer (in µm), whereas the second number represents the set distance between the imprinting and the pressure roll during imprinting (see Fig. 1a). Negative distance implies compression between the template and the wet coating during the imprint. In this test, pure UA16 (without additives) was used for the ground and the second layer. The power of the UV curing lamp was set to 50%. Scanning electron micrographs of the microstructures obtained from the three different imprint conditions are presented in Fig. 3a. Pillar dimensions such as pillar height and diameter of the pillar top are shown in Fig. 3b. The figure includes three regimes, highlighting the pillar top diameter: “standard mushroom” with diameters between 67 and 72 µm, corresponding to the dimensions obtained from the Ni-shim master structure; “smaller mushroom” with diameters ranging between 56 and 67 µm; and “no mushroom” with diameters smaller than 56 µm associated with the absence of mushroom tips. Clearly, all three conditions tested in the first tests led to “small mushrooms” with diameters ranging from 57 µm (“90_(− 245 µm)” and “120_(− 245 µm)”) to 60 µm (“120_(− 303 µm)”). Thus, the diameter slightly increased with more negative distances, i.e., higher compressive loads between the imprinting and the pressure roll. The pillar heights ranged between 53 and 56 µm. The specimens made from 120-µm-thick wet coatings and the highest compressive loads (\( d = - 303 \upmu{\text{m}} \)) showed the best adhesion performance with a pull-off stress about 70 kPa and a work of separation about 1.25 J m⁻² (Fig. 3c), which most probably is related to the largest mushroom caps obtained under these conditions [47].

In addition to compression during the imprint, UV-curing conditions are crucial for the cross-linking density and curing kinetics. We evaluated the impact of the UV power and the transparency of the silicone template (Fig. 4). Results for two different power ratings of the UV lamp are shown in (i) and (ii) in Figs. 4a–d. The micrographs reveal that microstructures exposed to 30% UV power exhibited no mushrooms (Fig. 4a). For 50% UV power, mushroom caps were obtained, but their size was still very small. Adhesion tests shown in Fig. 4c confirmed the results; by increasing the UV power to 50%, the pull-off stress was improved by 17% and the work of separation by 29%.

To further explore the effect related to UV exposure, the non-transparent silicone template (nPDMS) used in previous tests was replaced by a translucent silicone template (tPDMS). This modification led to much larger mushroom caps with tip diameters of up to 61 µm, as presented in (iii) and (iv) in Figs. 4a and d. In addition to larger caps, the structures were elongated with pillar heights of 59 µm, compared to 54 µm for structures generated by the non-transparent template. Figure 4e summarizes the adhesion performance of the micropatterned adhesives obtained using the different templates: Micropatterned adhesives replicated by using the translucent templates exhibited significantly higher pull-off stress (83 kPa) and work of separation (1.8 J m⁻²) compared to the adhesives generated by non-transparent templates with a pull-off stress of 70 kPa and a work of separation of 1.3 J m⁻². The larger mushroom caps accompanied with the enhanced adhesion performance are likely related to better penetration of UV light through the translucent template. Thus, the complete mushroom-shaped microstructure including the undercut of the mushroom cap was homogeneously cured. In addition, the transparent mold was less flexible (approx. two times stiffer, see Table 1) compared to the non-transparent template. Thus, the tPDMS template is less compressed during the imprint, which most probably explains the longer pillars. As a disadvantage, we observed that, after several imprint cycles, microstructures ruptured during demolding from the translucent template (tPDMS) and stuck inside the cavities. An explanation for this is that UA16 resin diffused into the tPDMS template upon repeated imprints. This can result in strong covalent bonds between the pre-polymer inside the cavity and the interpenetrated reactive oligomers, which in turn led to cohesive rupture of the microstructures. In contrast, repeated imprinting using the nPDMS was always successful without any residues inside the cavities upon demolding.

Besides process parameters, different material compositions (UA16-based polymer blends) were evaluated to further improve the formation of mushroom caps by enhancing the UV-curing kinetics. The parameters of the replication process for the different blends are summarized in Table 2. The kinetics of the UV-induced curing was modified by adding to UA16 either 1% Miramer SIP900 (polymer blend: U6S9) or 20% Miramer ES4420 (polymer blend: U6E4). Figure 5a shows the results of the photo-differential scanning calorimetry (UV-DSC), which implies that U6E4 cross-linked more quickly compared to U6S9 and UA16. In addition, mercapto-modified acrylate oligomers (U6E4) are well known to reduce the oxygen inhibition during a radical polymerization reaction [48]. In contrast to the tests presented above, the composition of the ground (first) layer was modified by adding 10% Laromer 8887 (polymer blend: U6L7), i.e., a mono-functional acrylic acid ester. Laromer 8887 acted as “thinner” [30, 33] by reducing the cross-linking density accompanied with polymer stiffness reduction and lower elongation at break (see values in Table 1). Micrographs of microstructures obtained from U6S9 (i–iii) and U6E4 (iv–vi) in terms of varying pressing distances are shown in Fig. 5b. Dimensions of the microstructures are reported in Fig. 5c. For U6S9, the diameter of the pillar diameter increased from 58 to 66 µm with decreasing pressing distance from − 245 to − 363 µm. For U6E4, the mushroom caps were larger and less sensitive to the pressing distances. Thus, diameters of the mushroom tips varied between 65 and 67 µm. The adhesion experiments confirmed these trends: The pull-off stresses obtained for U6S9 increased from 56 to 87 kPa with decreasing pressing distance from − 245 to − 363 µm. The work of separation was about 2 ± 0.2 J m⁻² for all U6S9 specimens. The best adhesion performance was obtained for U6E4 with the highest pull-off stress of 116 kPa and the highest work of separation of 3.78 J m⁻² for specimens made with a pressing distance of − 303 µm.

Following process and material optimization, micropatterned adhesive films of U6E4 were continuously fabricated at a pressing distance of −303 µm. The adhesion properties and durability of the adhesive films obtained were further characterized. Figure 6 shows the peel adhesion of the micropatterned U6E4 film in comparison with 3 M Scotch Magic tape. For both films, we found similar peel forces between 1 and 1.4 N cm⁻¹ for a 90° peel angle. Inspection upon tests revealed residue-free detachment for both films. Using lap shear test, a shear strength of the micropatterned U6E4 film of about 71 ± 9 kPa was obtained. In Fig. 7a, the reusability and shear adhesion of a 20-mm-wide U6E4 film were demonstrated by handling a glass object (1.163 kg): Upon attachment of the film, the object was lifted (shear-loaded), placed at the target position and the adhesive film was peeled off and attached again. The procedure was successfully repeated 5 times without any damage of the structure nor residues at the glass surface. Figure 7b demonstrates a similar test for the normal loading direction. Here, upon attachment, the glass object was normally lifted and placed at the target position. The cycle including attachment, lifting and detachment was successfully repeated 5 times without any visible damages nor residues.

The durability of the micropatterned dry adhesive was examined by a long-term pick-and-place test. A circular pad of adhesive film (U6E4, produced with d = − 303 µm; diameter 30 mm) repeatedly picked up and released a glass plate with mass of 145 g for 100,000 attachment–detachment cycles (see Fig. 8a). After these cycles, the micropatterned structure exhibited a certain extent of abrasion and contamination (see Fig. 8b). Similarly, the pull-off stress decreased from nearly 120 kPa to around 90 kPa (see Fig. 8c). However, the pick-and-place action can be still well performed even after 100,000 cycles, demonstrating a good endurance of PUA dry adhesive.