D4V Silicone: A Replica Material with Several Advantages for

LETTER pubs.acs.org/Langmuir

D4H/D4V Silicone: A Replica Material with Several Advantages for Nanoimprint Lithography and Capillary Force Lithography Peiwen Zheng and Thomas J. McCarthy* Polymer Science and Engineering Department, University of Massachusetts, Amherst, Massachusetts 01003, United States

bS Supporting Information ABSTRACT: Reported are demonstrations that D4H/D4V silicone (the product of the platinum-catalyzed hydrosilylation reaction between tetramethylcyclotetrasiloxane and tetramethyltetravinylcyclotetrasiloxane) is useful and practical as a replica material for both nanoimprint lithography (NIL) and capillary force lithography (CFL). The multiple advantageous properties of this extremely cross-linked material include UV transparency (for photo NIL and photo CFL), thermal stability (for high printing temperatures), high modulus (for high printing pressures), low surface energy (for easy demolding), and low viscosity precursors (for replicating small scale features). The replication performance of this material was tested using Blu-ray discs with sub-25 nm features and anodized aluminum foil with sub-10 nm features. Structures of ∼5 nm length scale on the surface of the anodized Al were replicated using D4H/D4V silicone as a mold material for CFL with a photocurable epoxy resin and for NIL with poly(methyl methacrylate) (PMMA). Features (holes in the anodized aluminum) with aspect ratios of greater than 9 were replicated.

N

anoimprint lithography (NIL) and capillary force lithography (CFL) are now recognized as useful fabrication tools for replicating surface structures with sub-50 nm topographic scale features.1,2 With the absence of the diffraction limitation inherent to photolithography, NIL was demonstrated by Chou’s group to be a low cost lithographic technique that can replicate nanoscale patterns by direct mechanical deformation of imprint resist materials.1,3 Ultrahigh resolution (sub-10 nm) and large throughput have been demonstrated subsequently by many research groups.4,5 UV transparent molds and UV-curable precursors have also been developed that permit NIL at room temperature.6 The requirements of the replicating material from which molds are fabricated, as well as the latitude of materials that can be imprinted, limit this technique.7 For thermal NIL, molds require both hardness and thermal stability and have to withstand pressures of hundreds of psi and temperatures above 100 °C. Photocure NIL has the additional requirement of transparency. Organic polymers are generally not suitable as mold materials, and typically silicon, silicon dioxide, silicon carbide, sapphire, and diamond are used. These materials are patterned with electron beam lithography or focused ion beam patterning and are costly. For release of the mold after imprinting, at least one of either the mold or the resist should exhibit low surface tension. Otherwise a release agent is required to separate the mold from the printed resist.7 CFL is another large-area-printing technique that was developed by Lee’s group.2 In this method, a liquid polymer is deformed by capillary forces to wet a patterned mold surface. Similar to NIL, the CFL process can be effected in two ways: (1) heating a plastic polymer above its glass transition temperature, r 2011 American Chemical Society

applying a mold that is wet by capillary forces, and then cooling and mold separation steps or (2) placing a transparent mold in contact with a low-viscosity and low-modulus-UV-curable polymer precursor that wets the mold due to capillary force, irradiating UV light though the mold to cure the polymer, and subsequently removing the mold. In comparison with NIL, CFL does not require high pressure, which can damage substrates and limits the options of mold materials. In CFL, a commercial silicone elastomer, for example, Sylgard 184 (a Dow Corning product), is effective as a mold material and can pattern highdensity sub-100 nm features.8 Due to the high viscosity of the precursors and the low modulus of PDMS elastomers (1 10 MPa), it is difficult to fabricate molds with sub-20 nm patterns. Incomplete filling of the cavities of the master mold and failure of the silicone during mold removal occur because of these two properties of this system. An obvious challenge for NIL and CFL is to develop new and better mold materials with low cost and the required properties; this has been the goal of several research efforts. A UV-curable prepolymer NOA 63,9 a silsesquioxane-derived material,10 and other hybrid resins11 were developed as replica molds. These materials are expensive, and a release agent is required during the process. Low surface energy (“non-sticky”) materials have also been studied including Teflon AF 2400,12 UV-curable PDMS,13 and polyvinylsilazane-derived silicates.14 We recently reported15 the preparation of extremely cross-linked silicones that were Received: March 28, 2011 Revised: May 24, 2011 Published: May 31, 2011 7976

dx.doi.org/10.1021/la201141k | Langmuir 2011, 27, 7976–7979

Langmuir

LETTER

prepared by the platinum-catalyzed hydrosilylation reaction between tetramethylcyclotetrasiloxane and tetramethyltetravinylcyclotetrasiloxane (eq 1); we abbreviate this material,

“D4H/D4V silicone.” We carefully studied the product prepared from a 2:1 molar ratio of D4H/D4V and suggested that this material may be useful for both NIL and CFL because it exhibits all of the obvious requirements for both techniques. Here we report results of our tests of D4H/D4V silicone as a replicating material and directions for preparing this material from commercially available and inexpensive materials. Two templates (masters) were chosen to study because of their availability and the length scales of their topographic features. Commercial Blu-ray discs16 were split open to reveal a surface that consists of ∼180 nm wide ridges that are separated by ∼180 nm. The valleys in between the ridges are ∼26 nm deep. Electropolished aluminum foil was anodized using conditions17 that form nearly circular 380 nm deep pores with diameters of 40 nm and center-to-center spacing of 110 nm. Both of these surfaces exhibit characteristic smaller topography elements that are shown and discussed below. The D4H/D4V silicone was prepared using a three-stage cure that was optimized for use as a molding material with these templates. The two low viscosity monomers,18 D4H and D4V, were mixed in a 2:1 molar ratio along with 6 ppm Karstedt’s catalyst19 (mass ratio, based on platinum in the product) in a small crystallizing dish. Warning: Using higher concentrations of catalyst can lead to extremely exothermic reactions that have caused explosions and fires.20 This reactive mixture was allowed to cure at room temperature for 24 h. This partially cured21 silicone was poured onto a horizontal sample of either the Blu-ray disk or the anodized aluminum, evacuated for 10 min, and heated at 70 °C for 3 days. After separating the solid cross-linked silicone from the masters (by hand with tweezers), the silicone daughters were further cured at 150 °C for 24 h.22 These daughters of the Blue-ray disk and anodized aluminum foil were used for both UV-CFL with a commercially available acrylate-based UV-curable epoxy resist23 and thermal NIL with spin-coated poly(methyl methacrylate) (PMMA, Tg = 105 °C).24 We note that separation of the “granddaughter” epoxy and PMMA samples from the low surface energy D4H/D4V silicone

Figure 1. AFM height images (3 μm  3 μm) and section analysis data: (a) Blu-ray master, (b) D4H/D4V silicone daughter, (c) epoxy (CFL) granddaughter, and (d) PMMA (NIL) granddaughter.

daughters was very easy; they delaminated spontaneously when a razor blade was inserted into the interface at the edge of the sample. We make several comments concerning the preparation of D4H/D4V silicone: (1) The rate of this reaction can be controlled with platinum concentration, and the time to prepare samples can likely be reduced significantly using higher catalyst concentrations.20 (2) The reaction involves rate-limiting diffusion of platinum atoms (ligand exchange reactions)20 and thus slows with increasing viscosity. (3) The three-stage (room temperature, 70 °C, 150 °C) cure conditions were chosen somewhat arbitrarily, but also by significant amounts of trial and error. The room temperature cure allowed the preparation of a viscous prepolymer that could conveniently be poured onto horizontal substrates. The 70 °C stage formed an elastic sample that separated easily from the master. The catalytic reaction was reproducible using this procedure and not as reproducible using other conditions, particularly if the room temperature stage was omitted. Figures 1 and 2 show atomic force microscopy height images and section analyses of the Blue-ray (polycarbonate) and anodized Al families: masters, daughters, epoxy (CFL) granddaughters, and PMMA (NIL) granddaughters. The dimensions (ridge width, height, and spacing) of the Blue-ray master and granddaughters are indistinguishable. We note that the ridges in the Blue-ray master (Figure 1a) have consistent widths, but their separation distance varies, imparting “waviness” to the structure. The D4H/D4V silicone daughter (Figure 1b) inherits this waviness, showing variation in the ridge width of the replica structure. This defect is passed along to the granddaughters (Figure 1c,d), replicating the defect pattern in the master. The height image of the anodized aluminum master (Figure 2a) shows, in addition to the 110 nm spaced, hexagonally arrayed 40 nm pores, the 7977

dx.doi.org/10.1021/la201141k |Langmuir 2011, 27, 7976–7979

Langmuir

LETTER

Figure 2. AFM height images (1 μm  1 μm) and section analysis data: (a) anodized Al foil master, (b) D4H/D4V silicone daughter, (c) epoxy (CFL) granddaughter, and (d) PMMA (NIL) granddaughter.

hexagonal flower petal-like structure that is characteristic of anodized aluminum. These features are of a length scale of ∼5 9 nm and appear as height differences in the section analysis. The pores are ∼380 nm deep, and the AFM tip does not sample at this depth. The D4H/D4V silicone daughter shows hexagonally packed posts, and the tops of these posts presumably are impressions of the bottoms of the 380 nm pores. This daughter produces epoxy (CFL) and PMMA (NIL) granddaughters (Figure 2c,d) that show the sub-10 nm features of the master. Wider scan AFM height images (Supporting Information) show that the grain boundaries in the original polycrystalline aluminum that are observable in the anodized Al master are transferred to both of the granddaughters. Scanning electron microscopy (SEM) analysis was performed on all of these samples as well. Micrographs of the anodized Al are shown in Figure 3. The SEM images confirm that the center-to-center distance and pore size are replicated master-togranddaughter, but the sub-10 nm topography that is visible using AFM is difficult to discern. This height difference may not be great enough to make contrast in the secondary electron image. The master, daughter, and epoxy (CFL) granddaughter (Figure 2a c) were sputter coated with platinum, and this may have smoothed the original topography. The PMMA (NIL) granddaughter needed further coating with gold to suppress beam damage in order to be imaged, and thus, the pore walls appear thicker in this image (Figure 3d). Higher resolution SEM images (Figure 4) of the D4H/D4V silicone daughter (Figure 4a) show an array of posts of the same height and some individual nanorods (Figure 4b) that were released when a sample was fractured. Figure 4b indicates that rather high aspect ratio (>9) features can be replicated with D4H/D4V silicone. These images show that the depth of the pores in the master is ∼380 nm and

Figure 3. SEM images: (a) anodized Al foil master, (b) D4H/D4V silicone daughter, (c) epoxy (CFL) granddaughter, and (d) PMMA (NIL) granddaughter.

Figure 4. SEM images of (a) a D4H/D4V silicone daughter and (b) several individual nanorods that were separated when a sample was fractured. 7978

dx.doi.org/10.1021/la201141k |Langmuir 2011, 27, 7976–7979

Langmuir that the room-temperature-cured D4H/D4V silicone reaches the bottom of the pores. We summarize with eight statements, some of which contain information that is not mentioned above: (1) D4H/D4V silicone is a material that can be readily made from inexpensive commercially available monomers and parts per million platinum catalyst using simple laboratory procedures. The 2:1 ratio D4H/D4V silicone has been characterized most completely.15 (2) The low viscosity of the monomers permits replication of very small 9) features. The preparation that we report here involves precuring the silicone at room temperature to impart some viscosity, but low molecular weight components are present.21 (3) Fully cured D4H/D4V silicone has a modulus of 1.29 ( 0.04 GPa and hardness of 0.34 ( 0.02 GPa, making it appropriate for high pressure NIL.15 (4) D4H/D4V silicone is thermally stable up to ∼500 °C,15 permitting high processing temperatures for thermal NIL or CFL. (5) UV transparency15 allows both UV-NIL and UV-CFL approaches for UV-curable resists. (6) The low surface energy15 facilitates mold release, and thus, no release agent is needed. (7) The coefficient of thermal expansion of D4H/D4V silicone is ∼167 μm/m/°C from room temperature to 200 °C.25 This rather large (silicone-like) value facilitates mold release. (8) A single D4H/D4V silicone daughter mold of anodized Al was used to prepare two CFL granddaughters and subsequently five NIL granddaughters with no apparent damage. This was the extent to which we investigated the “recyclability” of the daughter molds.

’ ASSOCIATED CONTENT

bS

Supporting Information. AFM height images (3 μm  3 μm) of anodized Al foil master, D4H/D4V silicone daughter, epoxy (CFL) granddaughter and PMMA (NIL) granddaughter. This material is available free of charge via the Internet at http:// pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT We thank Dian Chen for anodized aluminum samples and Prof. Ken Carter’s group for help with printing processes. We thank the Center for Hierarchical Manufacturing (CMMI0531171) and the Materials Research Science and Engineering Center (DMR-0213695) at the University of Massachusetts for support and 3M, Henkel, and Shocking Technologies for unrestricted funding. ’ REFERENCES

LETTER

(8) Suh, K. Y.; Lee, H. H. Adv. Funct. Mater. 2002, 12, 405. (9) Kim, Y. S.; Lee, N. Y.; Lim, J. R.; Lee, M. J.; Park, S. Chem. Mater. 2005, 17, 5867. (10) Lee, B. K.; Cha, N. G.; Hong, L. Y.; Kim, D. P.; Tanaka, H.; Lee, H. Y.; Kawai, T. Langmuir 2010, 26, 14915. (11) Lee, B. K.; Hong, L. Y.; Lee, H. Y.; Kim, D. P.; Kawai, T. Langmuir 2009, 25, 11768. (12) Khang, D. Y.; Lee, H. H. Langmuir 2004, 20, 2445. (13) Choi, S. J.; Yoo, P. J.; Baek, S. J.; Kim, T. W.; Lee, H. H. J. Am. Chem. Soc. 2004, 126, 7744. (14) Park, S.; Park, H. H.; Han, O. H.; Chae, S. A.; Lee, D.; Kim, D. P. J. Mater. Chem. 2010, 20, 9962. (15) Zheng, P.; McCarthy, T. J. Langmuir 2010, 26, 18585. (16) Blu-ray discs used were Memorex BD-R, 4  25 GB, single layer. (17) A two step anodization procedure is a modification of a reported procedure: Masuda, H.; Fukuda, K. Science 1995, 268, 1466. The first anodization used 0.3M oxalic acid, 17 °C, 40 V for 8 h. After removal of the alumina, a second anodization was carried out using 0.3M oxalic acid, 17 °C, 40 V for 3 min. These conditions yield alumina with 380 nm deep pores of 40 nm diameter, spaced (center-to-center) by 110 nm. (18) Tetramethylcyclotetrasiloxane (D4H) and tetramethyltetravinylcyclotetrasiloxane (D4V) were purchased from Gelest and used as received. (19) Karstedt’s catalyst was purchased from Gelest as a divinyltetramethyldisiloxane solution and diluted with D4V to concentrations so that parts per million Pt could be added with a 100 µL syringe. (20) A warning for this reaction has been published (Warning: A violent exotherm was always noted upon addition of platinum, and the contents burst into flames in one case). See: Stein, J.; Lewis, L. N.; Gao, Y.; Scott, R. A. J. Am. Chem. Soc. 1999, 121, 3693. We have performed this reaction in vessels ranging from small vials to 8.5 cm diameter vessels15 using up to 24 ppm Pt and have never observed an exotherm using appropriately diluted catalyst. (21) 1H NMR analysis (Bruker DPX-300) indicates that the reaction extent is ∼33% after 24 h at room temperature. The reaction has thus not proceeded to anywhere near the gel point (reaction extent of 75%). (22) These conditions have been shown to render complete reaction of the vinyl groups.15 (23) EPO-TEK OG 142, an acrylate-based UV-curable epoxy resist was purchased from Epoxy Technology, Inc. and used as received. It was spin-coated on an O2 plasma-cleaned (Harrick Plasma Cleaner) silicon wafer. The silicone daughter (Blu-ray or anodized Al foil) was placed on the top of the resist and after ∼40 s was exposed to 365 nm UV light (500 W OAI UV-lamp) for 15 min at room temperature. The silicone molds were easily separated from the epoxy granddaughters using a razor blade at the edge of the sample at the epoxy-silicone interface. (24) PMMA (Aldrich, Mw = 15 000) was spin-coated on an O2 plasma-cleaned (Harrick Plasma Cleaner) silicon wafer and annealed for 3 h at 150 °C. Silicone daughters were placed on top of the PMMA, and the assemblies were mounted and printed using a Nanonex NX-2000 nanoimprinter at 150 °C and 300 psi. (25) Thermomechanical analysis (TA Instruments model TMA 2940) was carried out from room temperature to 200 °C and gave a coefficient of thermal expansion value of 167 μm/m/°C for D4H/D4V.

(1) Chou, S. Y.; Krauss, P. R.; Renstrom, P. J. Appl. Phys. Lett. 1995, 67, 3114. (2) Suh, K. Y.; Kim, Y. S.; Lee, H. H. Adv. Mater. 2001, 13, 1386. (3) Chou, S. Y.; Krauss, P. R.; Renstrom, P. J. J. Vac. Sci. Technol., B 1996, 14, 4129. (4) Chou, S. Y.; Krauss, P. R. Microelectron. Eng. 1997, 35, 237. (5) Ahn, S. H.; Guo, L. J. Adv. Mater. 2008, 20, 2044. (6) Ruchhoeft, P.; Colburn, M.; Choi, B.; Nounu, H.; Johnson, S.; Bailey, T.; Damle, S.; Stewart, M.; Ekerdt, J.; Sreenivasan, S. V.; Wolfe, J. C.; Willson, C. G. J. Vac. Sci. Technol., B 1999, 17, 2965. (7) Guo, L. J. Adv. Mater. 2007, 19, 495. 7979

dx.doi.org/10.1021/la201141k |Langmuir 2011, 27, 7976–7979