2430
Langmuir 2006, 22, 2430-2432
Three-Dimensional Metallized Features on Polymeric Substrates by Microcontact Printing Vinalia Tjong, Lei Wu, and Peter M. Moran* Institute of Materials Research and Engineering, 3 Research Link, Singapore 117602 ReceiVed September 5, 2005. In Final Form: January 10, 2006 We demonstrate the formation of 3-D metallized features on a polymeric substrate by microcontact printing. A patterned silicon stamp was “inked” with catalytic particles. Thereafter, particles were selectively removed from the raised regions of the stamp. A molten polymer was embossed against the stamp. Upon cooling and separation, the catalytic particles that were within the recessed areas of the stamp were transferred to the polymer. The polymer was then immersed in an electroless plating bath, and metallization occurred selectively on the areas where the catalytic particles were present. We have achieved 3-D metallized columns as small as 500 nm in diameter and about 1 µm tall.
Methods to produce micro- and nanoscale surface structures of various functionalities and topographies have attracted great attention from the microfabrication community because of their potential applications in sensors, optoelectronics, plastic electronics, biomedicine, and biology.1 Metal-patterned polymeric substrates have been widely used in the integration and packaging of electronic devices.2 Although conventional techniques for fabricating planar microstructures have reached a developed stage, techniques to create 3-D microstructures are still in an early stage of development. Potential applications such as separation devices in nanofluidic3 and biomolecular sieving4 have prompted research in creating 3-D submicrostructures and nanostructures. Whereas photolithography with selective etching is commonly used to produce metallized substrates, nonphotolithographic methods have become an alternative in recent years because of the high cost and feature size limitations of photolithography. In addition, photolithography is generally limited to planar substrates. Several nonphotolithographic methods such as direct metal stamping,2 nanoimprinting (NIL),5 microcontact printing (µCP),6 cold welding,7 nanotransfer printing (nTP)8 and hot liftoff9 have been developed for metal patterning at the micro- to nanoscale. These approaches make use of a stamp instead of a mask for pattern formation. Selective electroless deposition (ELD) is a well-known, highyield, cost-effective process used to create thin and thick metal films. In this technique, a catalyst, such as palladium (Pd), is used to catalyze the growth of a metal layer. Preferential metallization can be achieved by selectively depositing the * To whom correspondence should be addressed. E-mail: peter.moran@ singular-id.com. (1) Voldman, J.; Gray, M. L.; Schmidt, M. A. Annu. ReV. Biomed. Eng. 1999, 1, 401-425. (2) (a) Dreyfus, B. A. U.S. Patent 3,230,163, 1966. (b) Wang, Z.; Yuan, J.; Zhang, J.; Xing, R.; Yan, D.; Han, Y. AdV. Mater. 2003, 15, 1009. (c) Bhangale, S. M.; Tjong, V.; Wu, L.; Yakovlev, N.; Moran, P. M. AdV. Mater. 2005, 17, 809-813. (3) Cabodi, M.; Turner, S. W. P.; Craighead, H. G. Anal. Chem. 2002, 74, 5469-5174. (4) Seo, Y.-S.; Luo, H.; Samuilov, V. A.; Rafailovich, M. H.; Sokolov, K.; Gersappe, D.; Chu, B. Nano Lett. 2004, 4, 659. (5) Chou, S. Y.; Krauss, P. R.; Renstrom, P. J. J. Vac. Sci. Technol., B 1996, 14, 4129-4133. (6) Hidber, P. C.; Helbig, W.; Kim, E.; Whitesides, G. M. Langmuir 1996, 12, 1375-1380. (7) Kim, C.; Burrows, P. E.; Forrest, S. R. Science 2000, 288, 831-833. (8) Loo, Y.-L.; Willet, R. L.; Baldwin, K. W.; Rogers, J. A. J. Am. Chem. Soc. 2002, 124, 7654-7655. (9) Wang, Z.; Zhang, J.; Xing, R.; Yuan, J.; Yang, D.; Han, Y. J. Am. Chem. Soc. 2003, 125, 15278-15279.
catalytic particles on certain regions of the substrate or selectively passivating regions of a catalyzed substrate. Hidber et al.6 used µCP and subsequent ELD to form micrometer- and submicrometer-scale copper patterns on surfaces. A patterned elastomeric stamp was used to transfer the catalyst colloidsPd particles stabilized with tetraalkylammonium bromidessselectively to the substrate surface. Moran et al.10 used elastomeric stamps to print patterns of hydrophobic, passivating siloxane molecules on silica substrates. The stamped substrate was exposed to a catalyst and then to an electroless plating solution. This process resulted in directed ELD of the metal on the exposed regions of the substrate. Lee et al.11 selectively adsorbed negatively charged polystyrene beads onto a substrate. Thereafter, the positively charged Pd-complex catalysts were adsorbed onto the beads. Subsequent plating resulting in patterned islands of nickel-coated beads. In our previous work, we demonstrated µCP of catalytic Pd nanoparticles to form metal patterns on nonplanar polymeric substrates.12 With the exception of Lee et al.,11 who were able to form a limited variety of 3-D patterns (i.e., clusters of nickel-coated beads), none of the work described above has been able to pattern 3-D metallic features. Here, we modify our previous method in order to fabricate 3-D metallized polymer features. The process is as follows: First, an OH-terminated silicon stamp was coated with palladium nanoparticles stabilized with poly(vinylpyrolidone), (PVP), by immersing the stamp into a colloidal solution of the particles at room temperature for ∼30 min. The stamp was then removed and dried gently under a stream of nitrogen. The particles adsorbed weakly to the stamp.12 To create a 3-D structure, only catalytic particles in the recessed regions of the stamp were transferred to the substrate. This was achieved by removing catalytic particles from the raised regions of the stamp by gentle wiping with a moist cotton bud.13 Prior to embossing, the stamp was heated to about 100 °C, and the polymer (here we use polystyrene) substrate (10) Moran, C. E.; Radloff, C.; Halas, N. J. AdV. Mater. 2003, 15, 804-807. (11) Lee, I.; Hammond, P. T.; Rubner, M. F. Chem. Mater. 2003, 15, 45834589. (12) Ng, W. K.; Wu, L.; Moran, P. M. Appl. Phys. Lett. 2002, 81, 3097-3099. (13) Removing particles from the stamp in this manner is effective. In our previous paper,12 we showed this experimentally. In that experiment, three Si(100) stamps with adsorbed Pd particles were heated for 2 min to 75, 125, and 200 °C. Thereafter, the surface was wiped gently with a moist cotton bud. The X-ray photoelectron spectra (XPS) show no discernible Pd peaks in the samples, and hence the vast majority of the nanoparticles have been removed from the Si stamp.
10.1021/la052412x CCC: $33.50 © 2006 American Chemical Society Published on Web 02/10/2006
Letters
Langmuir, Vol. 22, No. 6, 2006 2431 Scheme 1. Schematic of the Stamping and Metallization Processesa
Figure 1. TOF-SIMS measurements showing the transfer percentage of nanoparticles on the polystyrene substrate at different stamping temperatures. Five Si(100) stamps were immersed in a PVP-stabilized Pd colloidal solution for 30 min. After drying, these stamps were used to transfer the Pd nanoparticles at various temperatures (75, 100, 125, 140, and 160 °C).
was heated to between 140 and 150 °C, which is a much higher temperature than its glass-transition temperature (∼100 °C). Pressure was then applied slowly to allow the polymer to flow into the cavities of the stamp. After cooling to room temperature, the silicon stamp and polystyrene substrate were separated. Thereafter, the polymer substrate was immersed in a commercially available electroless nickel plating solution at 90 °C and pH 5 for ∼1 min. The success of this method depends on several factors: (i) the relatively weak adsorption of PVP-stabilized Pd nanoparticles on the Si stamp, (ii) the effective and selective transfer of Pd nanoparticles to the polymer substrate, and (iii) the anchoring of Pd nanoparticles on a polymer substrate. According to studies by Boonekamp et al.14 and our previous work using X-ray photoelectron (XPS) spectra,12 Pd nanoparticles stabilized with PVP adsorb weakly to silicon dioxide (SiO2) surfaces. Furthermore, only a thin layer of nanoparticles adsorbs to the substrate surface because steric hindrance between the comparatively long PVP chains prevents a buildup of particles. We found that the adsorption reached a steady value after 30 min of immersion in a PVP-stabilized Pd colloidal solution.15 The effective transfer of Pd catalyst particles can be verified by studying the percentage of Pd transfer using time-of-flight secondary ion mass spectroscopy (TOF-SIMS). This measurement was carried out by normalizing the Pd signal intensity to Ga intensity on the stamped polymer substrate during TOF-SIMS, before and after the stamping process.16 Figure 1 shows the effect of stamping temperature on the transfer percentage of the nanoparticles. As is evidenced from transfer percentage results, heating well above the polystyrene glass-transition temperature is crucial to ensure substantial transfer of the nanoparticles. Maximum transfer (98%) was achieved when the polystyrene was heated to 160 °C. (14) Boonekamp, E. P.; Kelly, J. J.; Fokkink, L. G. J. Langmuir 1994, 10, 4089-4094. (15) Seven Si(100) samples (with native oxide layers) were immersed in a PVP-stabilized Pd colloidal suspension for varying periods from 1 min to 24 h. The adsorption of colloidal Pd particles on the SiO2 surface was studied as a function of time. The amount of adsorption was estimated by measuring the Pd intensity on each sample using TOF-SIMS. (16) Five Si(100) stamps were immersed in a PVP-stabilized Pd colloidal solution for 30 min. After drying, these stamps were heated, and heated polystyrene was embossed against the stamp at various stamping temperatures (75, 100, 125, 140, and 160 °C). After stamping, the polystyrene surfaces were examined by TOF-SIMS.
a
Not drawn to scale.
Scheme 2. Schematic of PVP-Stabilized Pd Nanoparticles Engulfed within the Surface of a Polymer Substratea
a
Not drawn to scale.
If the polystrene substrate was heated beyond about 160 °C during stamping, then the nanoparticles would transfer. However, no plating was observed. We believe that this may be because the particles are being engulfed within the polystrene surface. Scheme 2 depicts nanoparticles being engulfed within the surface of the polymer substrate. As shown in the Scheme, the “surface” of a polymer is difficult to define precisely because it consists of a region of loosely bundled polymer chains that form a “hairy” surface. Even though the stamping process is relatively fast, at elevated temperatures the nanoparticles may have time to become sufficiently engulfed or entwined within this hairy surface to reduce the exposed Pd surface below the threshold needed to catalyze metal deposition. The driving force for the particles to engulf comes from the affinity between the polystrene and the nanoparticles, which is the same affinity that allows the particles to be transferred from the stamp to the polymer. The engulfing of PVP-stabilized Pd particles was studied using a TOF-SIMS depth profiling analysis of polystyrene substrates stamped at various temperatures16 (results not shown here). As a control, two polystrene substrates were examined using TOF-SIMS after being immersed in the colloidal nanoparticle solution: In the first control, the substrate was simply immersed in the colloidal solution and then dried in a stream of nitrogen. In the second control, the substrate was briefly heated to 150 °C after drying. TOF-SIMS of the first control substrate showed the peak Pd
2432 Langmuir, Vol. 22, No. 6, 2006
Letters
instead of pure polystyrene. This was done because the pure polystyrene was too brittle to allow the small, high-aspect-ratio features to be formed consistently; usually the columns would break off during cooling or separation. Petri dish polystyrene required higher stamping temperatures of up to 250 °C. We tried to extend the method to form 200-nm-diameter columns (separated by 200 nm); however, we were not able to get the plating to occur only on the columns, and in general, the entire substrate was coated with nickel. Indeed, it can be seen from Figure 2c that some stray nickel particles are even seen in the gaps between the 500 nm columns (the gaps are 500 nm wide). This may indicate that the method is not able to form features that are closer than about 500 nm to each other because of the chance of stray metallization occurring between the features. In conclusion, microcontact printing was used to fabricate micrometer and submicrometer-scale 3-D metallized features on polymeric substrates. The selective deposition of PVPstabilized Pd nanoparticles was made possible by microcontact printing using rigid silicon stamps. We have accurately produced 3-D metallized columns as small as 500 nm in diameter. Arrays of metallized features such as those demonstrated here have potential applications in areas such as optics, magnetics, sensors and biology, particularly for DNA separation.3,4 Experimental Section
Figure 2. SEM images of 3-D raised polystyrene structures selectively coated with nickel fabricated using our method. The light-gray regions are coated with nickel whereas uncoated areas of the polystyrene substrate appear dark gray/black. (a) Raised 50 µm grid pattern (20 µm tall), (b and c) 2-µm- and 500-nm-diameter columns, respectively (both ∼1 µm tall).
intensity immediately on the surface of the polymer. In the second control substrate (that had been heated to 150 °C), the peak Pd intensity was found only after sputtering had occurred for a few seconds, indicating that the Pd particles were at least partially engulfed within the surface. Results for the stamped substrates also showed similar engulfing (except for the case where stamping was at 75 °C and virtually no particles were transferred16). Although this evidence of engulfing supports our theory, the experiments are not able show the level to which the polystyrene chains entwine with the particle, so we are not able to prove or disprove that it is indeed the level of entwining that inhibits plating. Figure 2 shows scanning electron microscope (SEM) images of a 3-D nickel pattern on a polystyrene polymer fabricated using our method. A 50-µm-wide grid pattern coated with nickelcoated polystyrene is shown in Figure 2a. The dark areas between the grid lines are regions of uncoated polystyrene lying 20 µm below the top of the grid. This demonstrates that Ni plating occurs only where catalytic Pd nanoparticles had been selectively transferred. To show that this method can be extended to include smaller feature sizes, we fabricated 2-µm-diameter and 500-nm-diameter polystyrene columns, shown in Figure 2b and c, respectively. Similar results were obtained, and plating selectivity was maintained. In these experiments, however, we used a polystyrene blend (found in Petri dishes manufactured by Bibby Sterilin Ltd)
Materials and Instrumentation. Silicon stamps were purchased from the Institute of Microelectronics, Singapore. Poly(vinylpyrolidone) (average Mw ≈ 55 000) and polystyrene (product no. 331 651, Mw ) 4000-200 000) were purchased from Sigma-Aldrich. PdCl2 (anhydrous, 59% Pd) and hypophosphorous acid (H3O2P, 50%) were purchased from Merck. Electroless nickel plating was accomplished using a commercially available nickel-plating kit from Plaschem, Singapore, by mixing 1:2.5 v/v PEN-94 A and B solutions. The pH was adjusted using ammonia water and a 10% sulfuric acid solution. TOF-SIMS analysis was performed using a TOF-SIMS IV system built by ION-TOF GmbH. Preparation of the Stamp. The patterned silicon stamps were cleaned thoroughly in boiling acetone before rinsing in boiling isopropyl alcohol and deionized water (∼3 min per step). After drying with compressed air, the stamps were placed in piranha solution (3:1 v/v 98% H2SO4/30% H2O2) at 90 °C for 30 min. They were then rinsed with deionized water and dried under a stream of nitrogen. The process resulted in OH-terminated silicon stamps. Preparation of Colloidal Solution of PVP-Stabilized Pd Particles. A PVP solution was made by dissolving 50 mg of PVP in 500 mL of deionized water. A PdCl2 solution was prepared by mixing 150 mg of PdCl2 with 5.25 mL of hydrochloric acid (fuming 37%). The solutions were stirred overnight. The two solutions were then mixed, and 30 mL of H3O2P was added slowly. The H3O2P reduced the Pd2+ to metallic Pd nanoparticles (roughly 3.8 nm in diameter), which bound to the available PVP chains to form a composite nanoparticle. The total composite nanoparticle (PVPstabilized Pd) is estimated to be about 10-30 nm in diameter. Excess reducing agent was used to prevent the long-term oxidation of the nanoparticles. Finally, deionized water was added to make up 1 L of the nanoparticle solution. These solutions remained stable for months.
Acknowledgment. We thank Ms. Doreen Lai for assistance with TOF-SIMS analysis and The Agency for Science, Technology and Research, Singapore for funding support (grant no. 0221070014). LA052412X
