Article pubs.acs.org/Langmuir
Adhesiveless Transfer Printing of Ultrathin Microscale Semiconductor Materials by Controlling the Bending Radius of an Elastomeric Stamp Sungbum Cho,† Namyun Kim,† Kwangsun Song,†,‡ and Jongho Lee*,†,‡ †
School of Mechanical Engineering, and ‡Research Institute for Solar and Sustainable Energies (RISE), Gwangju Institute of Science and Technology (GIST), 123 Cheomdan-gwagiro, Buk-gu, Gwangju 61005, Republic of Korea S Supporting Information *
ABSTRACT: High-performance electronic devices integrated onto unconventional substrates provide opportunities for use in diverse applications, such as wearable or implantable forms of electronic devices. However, the interlayer adhesives between the electronic devices and substrates often limit processing temperature or cause electrical or thermal resistance at the interface. This paper introduces a very simple but effective transfer printing method that does not require an interlayer adhesive. Controlling the bending radius of a simple flat stamp enables picking up or printing of microscale semiconductor materials onto rigid, curvilinear, or flexible surfaces without the aid of a liquid adhesive. Theoretical and experimental studies reveal the underlying mechanism of the suggested approach. Adhesiveless printing of thin Si plates onto diverse substrates demonstrates the capability of this method.
■
controlling the bending radius of a simple, flat, unstructured elastomeric stamp. Mechanical characterization of adhesion and pressure depending upon the bending radius reveals the underlying mechanism. Transfer printing of thin micro Si plate arrays with sizes from tens of micrometers to several millimeters onto various surfaces, including optically smooth, curvilinear, and flexible surfaces, was demonstrated, thus showing the capabilities of the proposed approach.
INTRODUCTION Recent studies have shown the potential uses of highperformance electronic devices in flexible and stretchable forms for various applications, such as displays,1−3 photovoltaics, 4−6 curvilinear electronics, 7−9 biomedical electronics,10−12 transparent electrodes,13,14 and many others.15−18 One of the promising ways to realize these electronics is by transfer printing high-performance electronic devices onto unconventional substrates.19 In many cases, microscale semiconductor devices fabricated on rigid wafers are picked up with a sticky elastomeric stamp, and then the microdevices are transfer printed onto target substrates with the aid of a liquid adhesive.14,20 Although liquid adhesives are convenient to retrieve microdevices from the sticky stamp, transfer printing the microdevices directly to target substrates without an interlayer adhesive relieves limitations associated with the adhesive, such as high processing temperature13,20,21 and high thermal or electrical resistance between the microdevices and substrates.22,23 Retrieval of microdevices occurs at a high speed from a donor substrate and placed/printed at lower speeds24 and higher temperatures25 onto a receiver substrate. In other studies with advanced substrate designs26,27 and manipulations,28 microstructured stamps from nature29 or prepared via microfabrication30−35 enable adhesiveless transfer printing. Some of these approaches make use of a lower peeling speed and the elastic and directional properties of microstructured stamps along with micromanipulation of the stamps. However, where possible, transfer printing microdevice arrays at a relatively higher speed with a simple flat stamp is desirable for a more efficient and simple printing process. Here, we present a method of transfer printing thin micro Si plates by © XXXX American Chemical Society
■
RESULTS AND DISCUSSION Panels a−d of Figure 1 represent schematic illustrations of the procedure for transfer printing micro Si plates by controlling the bending radius of the polydimethylsiloxane (PDMS) stamp (Sylgard 184, Young’s modulus of ∼1.86 MPa36). The thin micro Si plates to be transfer printed are temporarily covered with a photoresist (PR, AZ5214) layer and supported by PR anchors around them through photolithography processes on a silicon on insulator (SOI) wafer. More details can be found in Figure S1 of the Supporting Information. Bringing the PDMS stamp into contact with the micro Si plates (Figure 1a) and then retracting the stamp with a relatively large bending radius picks up the micro Si plates because the PDMS stamp is soft and tacky, generating enough adhesive force to break the temporary anchors (Figure 1b). After any PR residues are cleaned from the broken anchors with drops of acetone, isopropyl alcohol (IPA), and deionized (DI) water, the PDMS Received: May 18, 2016 Revised: July 13, 2016
A
DOI: 10.1021/acs.langmuir.6b01880 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
distribution of adhesion and pressure. Both adhesion and pressure rely on the bending radius during peeling, as illustrated in panels a and b of Figure 2. As the stamp is peeled with a
Figure 1. Schematic illustrations and optical images of the adhesiveless transfer printing procedure by controlling the bending radius of an elastomeric stamp. (a) Schematic illustration of the stamp brought into contact with micro Si plates, which are temporarily held on the original substrate. (b) Peeling the stamp with a relatively large bending radius picks up the micro Si plates from the original substrate. (c) Stamp holding the micro Si plates is brought into contact with the target substrate, which does not have an adhesive layer. (d) Peeling the stamp with a relatively small bending radius prints the micro Si plates on the target substrate. (e) Optical image of the printing process by peeling the elastomeric stamp (thickness of 2 mm, PDMS) with a relatively small bending radius (∼6 mm) to leave the micro Si plates (size, 760 × 760 μm; thickness, 7 μm) on the target substrate (glass slide). (f) Micro Si plates array (7 × 7) transfer printed onto the target substrate.
Figure 2. (a and b) Schematic illustrations of pressure and adhesion induced by the elastomeric stamp depending upon the bending radius. The curved region (peel zone) in contact with a micro Si plate generates adhesive forces, and the flat region provides pressure to the substrate. The peel zone widens as the bending radius increases. (c and d) Schematic illustrations of rolling tests, corresponding to a (c) small and (d) large radius of glass rods to quantify energy release rates. Glass rods with different radii were released on an angled PDMS slab. (e) Measurement results of energy release rates for glass rods with different radii (black square, r = 3 mm; red circle, r = 5 mm; and blue triangle, r = 8 mm). The error bars give the standard deviations. For the same speed, the energy release rate for a larger radius is higher.
stamp holding the micro Si plates is put into conformal contact with a clean target substrate, such as a glass slide, which is not coated with any liquid adhesive layer (Figure 1c). The bottom surface of the micro Si plates that interact with the glass slide can be oxidized during the cleaning process. The top surface of the micro Si plates is covered with a PR layer, which is later cleaned with acetone after the transfer printing process. The PR residues may lower the printing yield (Figure S4 of the Supporting Information). After pressure (∼2.5 kPa) is applied to the stamp for about 3 min, the stamp is peeled with a relatively small bending radius, transfer printing the micro Si plates onto the target substrate, as illustrated in Figure 1d. The optical image in Figure 1e shows the transfer printing process of the micro Si plates (7 × 7; size, 760 × 760 μm; thickness, 7 μm) onto a glass slide without the aid of a liquid adhesive layer by peeling the stamp (thickness of 2 mm) with a small bending radius (∼6 mm). After the transfer printing, the PR layer covering the micro Si plates is removed by cleaning with acetone. Figure 1f shows the micro Si plates transfer printed on a clean glass slide without an interlayer adhesive between the micro Si plates and the glass slide. The ability to selectively pick up or print the micro Si plates between the stamp and substrate is attributed to the
constant radius of curvature, the curved region, called the peel zone,37 is under adhesion, i.e., pulling up the substrate, because the elastomeric stamp stretches before being separated from the substrate, while the flat region is under pressure, pressing down the substrate, because of the bending force of the stamp. For transfer printing, lower adhesion and higher pressure induced from a smaller bending radius (Figure 2a) facilitate printing micro Si plates onto the substrate. In contrast, higher adhesion and lower pressure from a larger bending radius (Figure 2b) facilitate picking up micro Si plates from the substrate. The adhesion characteristics depending upon the bending radius can be evaluated by rolling glass rods with different radii on an angled flat PDMS layer (thickness of 7 mm), as illustrated in panels c and d of Figure 2. More details can be found in the Materials and Methods. For a larger radius, the peel zone becomes wider, resulting in a higher energy release rate (G),37 which is defined as the energy dissipated to separate surfaces per newly created area.38,39 When the rod rolls down freely at a constant speed, the energy dissipated to separate B
DOI: 10.1021/acs.langmuir.6b01880 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Figure 3. Measurements of pressure when peeling the elastomeric stamps whose thickness and bending radius vary. (a) Optical image of the measurement setup. The sensing edge in the slit measures the pressure in the flat region of the elastomeric stamp while peeling the stamp by translating the stage at a constant speed. (b) Pressure distribution depending upon the thickness (t = 1−2 mm) of the stamps with a constant bending radius (4 mm). For a thicker stamp, the pressure is higher. (c) Pressure distribution depending upon the bending radius (r = 3−7 mm) of a stamp (thickness of 1 mm). As the bending radius becomes smaller, the pressure becomes higher. (d and e) Optical images of the elastomeric stamp (thickness of 1 mm) peeling with a (d) higher (8 mm) and (e) lower (3 mm) bending radius at a constant speed (2 mm/s). Dependent upon the bending radii, the elastomeric stamp (d) keeps the micro Si plates on the stamp or (e) transfers them to the glass substrate. These images were taken while pausing the peeling process. (f) Measurements of distances between the transferred micro Si plates depending upon the bending radius and the thickness of the stamp. The error bars indicate the standard deviation. For the stamp with a thickness of 1 mm, the errors are small for various bending radii.
a custom-built motorized translation stage. The sensing edge connected to the force sensor is exposed and leveled with the surface of the custom supporting body to read correct values. As the motorized stage translates at a constant speed (1 mm/s), the force sensor reads pressures delivered by the narrow sensing edge while peeling the stamp with a constant bending radius. The experimental setup is used only for measuring the pressure distribution by applying fine powder on the PDMS stamp because of limitations in the custom setup. More details are provided in the Materials and Methods and Figure S3 of the Supporting Information. The pressure distribution is affected by two parameters, the thickness and bending radius of the elastomeric stamp, as shown in panels b and c of Figure 3. For a fixed bending radius (r = 4 mm), the maximum pressures for the stamps with thicknesses of 1 mm (black line), 1.5 mm (red line), and 2 mm (blue line) are 1.83, 4.14, and 16.6 kPa over 7, 8.5, and 10 mm, respectively. The thicker stamp generates higher pressure over a wider region as the bending stiffness becomes higher for the thicker stamp. As a result of bending the stamp further for the same thickness, the pressure becomes higher, as in Figure 3c. The maximum pressures when peeling the 1 mm thick stamp with the bending radii of 7 mm (green line), 6 mm (magenta line), 5 mm (blue line), 4 mm (red line), and 3 mm (black line) are 0.45, 0.64, 0.91, 1.82, and 4.42 kPa, respectively. These results reveal that the pressure can be controlled by the thickness or bending radius of the stamp. For a stamp whose thickness is fixed, the picking up or printing process can be selected by controlling the bending radius. For picking up, a large bending radius generating higher adhesion and lower pressure is more effective. For printing, a lower bending radius is more effective. Optical images in panels d and e of Figure 3
surfaces at the trailing edge is the same as the potential energy loss of the rod, because the energy required to close the surfaces at the advancing edge can be ignored.40 From the rolling tests, the energy release rates for different rods can be calculated with the following equation:38,39 G = −∂U /∂A = mgh/bl
where U denotes the adhesion energy between the rod and PDMS, A is the newly created area, m and b are the mass and width of the rod, respectively, g denotes the gravitational acceleration (9.81 m/s2), l is the distance the rod rolled, and h is the height change of the rod (h = l sin θ). Figure 2e shows the energy release rates at different radii (black square, r = 3 mm; red circle, r = 5 mm; and blue triangle, r = 8 mm) depending upon the rolling speeds of the glass rods. For the same rolling speed, a rod with a larger radius yields a higher energy release rate. For example, at 2 mm/s, the energy release rates for rods with radii of 5 and 8 mm are 11 and 29% higher than the energy release rate of a rod with a smaller radius (3 mm), indicating that adhesion is adjustable by controlling the bending radius of the viscoelastic stamp. It should be noted that the interface between the glass rods and PDMS layer is not confined27 because the contact widths (Figure S5 of the Supporting Information) are much smaller (
