LETTER pubs.acs.org/NanoLett
Omnidirectional 3D Nanoplasmonic Optical Antenna Array via Soft-Matter Transformation Benjamin M. Ross,† Liz Y. Wu,† and Luke P. Lee* Applied Science and Technology Graduate Group, Biomolecular Nanotechnology Center, and Berkeley Sensor and Actuator Center, and Department of Bioengineering, University of California-Berkeley, Berkeley, California 94720, United States
bS Supporting Information ABSTRACT: Inspired by the natural processes during morphogenesis, we demonstrate the transformation capability of active soft-matter to define nanoscale metal-on-polymer architectures below the resolution limit of conventional lithography. Specifically, using active polymers, we fabricate and characterize ultradense nanoplasmonic antenna arrays with sub-10 nm tipto-tip nanogaps. In addition, the macroscale morphology can be independently manipulated into arbitrary three-dimensional geometries, demonstrated with the fabrication of an omnidirectional nanoplasmonic optical antenna array. KEYWORDS: Nanoplasmonics, soft-matter, active polymers, nanosphere lithography, nanogap, omnidirectional antenna
S
hape-changing polymers, also called active or smart polymers, have been studied in recent years for their ability to deform significantly on the macro-scale when a stimulus is applied. Such stimulus may come in the form of thermal, electric, magnetic, or optical excitation.13 Several research groups have demonstrated microfabrication techniques taking advantage of active polymers.46 However, most work has focused on microscale features, and nanofabrication via active polymers has primarily focused on generation of random structures such as wrinkles7 with the exception of recent work by Odom’s group.8 The control of arbitrarily defined nanoscale structures via active polymers has remained a challenge, and many important questions remain regarding both the limits of this technique and properties of the resulting nanostructures. Here, we extend our earlier work9 and demonstrate that uniform polymer deformations can be engineered to create high-density nanogap structures comparable to the resolution limit of high-resolution lithographic methods. Control of deformation at the nanoscale allows the possibility of fabricating complex nanostructures in a way comparable to that achieved naturally during morphogenesis, as illustrated in Figure 1ad. Specifically, the contraction of an active polymer substrate allows nanostructure repositioning in a way analogous to apical constriction during morphogenesis (process shown in Supporting Information Figure 1). The ability to control substrate dimensions after deposition of structures can confer the significant advantages over standard rigid substrates that biology has achieved,10,11 such as (i) the potential to alter structure-to-structure distance, surpassing both resolution and structural density of the initial placement, and (ii) the potential to deform nanostructures into complex three-dimensional (3D) geometries. The former benefit is of wide interest in reducing feature size of devices, while the latter is of particular interest in microelectromechanical and nanoelectromechanical systems fabrication. r 2011 American Chemical Society
Active polymers can be readily deformed into complex 3D geometries at the macroscale, while the nanoscale morphology is retained on the polymer surface. This provides a platform for integrating nanoarchitectures into stretchable devices ranging from flexible displays, to bioinspired lenses, and to biomedical systems.12 We demonstrate an omnidirectional nanoplasmonic antenna array in Figure 1e. The 3D structure is achieved by heating the contracted polymer until it is moldable (process shown in Supporting Information Figure 2), and its optical properties are discussed below. Both the nano- and macroscale deformations of active polymers are well-suited to the needs of nanoplasmonic molecular sensors. In this application, collective electron oscillations of metallic nanostructures cause extremely high local electric fields, which can then be used to provide a spectroscopic signal in the form of localized surface plasmon resonance (LSPR) shift,13 surface-enhanced Raman scattering (SERS),14 surface-enhanced fluorescence (SEF),15 or plasmon resonance energy transfer (PRET).16 Because the sensitivity of such structures is closely related to the nanostructure geometry, the nanogap distance between structures, and the density of such structures,17,18 active substrates offer the possibility of creating a new class of highly sensitive nanoplasmonic sensors. In addition, the local electromagnetic field controlled by nanoplasmonic antennas can be used to create new light sources, for example, through second harmonic generation effects.19 Because the needs of nanoplasmonics are well-suited to the advantages of active polymer nanofabrication, we use nanoplasmonics as a proof-of-concept and demonstrate active polymer nanofabrication for creating ultradense nanoplasmonic antenna arrays with sub -10 nm tip-to-tip distances. The nanoprisms are defined with nanosphere lithography, which has been demonstrated to be a Received: December 27, 2010 Published: June 10, 2011 2590
dx.doi.org/10.1021/nl201516g | Nano Lett. 2011, 11, 2590–2595
Nano Letters
LETTER
Figure 1. (a,b) Schematic of natural structure definition due to apical constriction; (c,d) biologically inspired transformation of metal-on-polymer structures for defining nanostructure morphology; (e) omnidirectional nanoplasmonic antenna arrays; inset shows the nanoprism arrays, which are present at each “pixel” of the lens. Panels a,b reprinted by permission from Macmillan Publishers Ltd.: Nature Reviews Molecular Cell Biology 8, 633, Copyright 2007.
well-characterized and cost-effective method for patterning nanoplasmonic materials over large areas.20 The size and periodicity of the nanoprisms can be tuned by choice of the nanosphere size. In addition, no high temperatures or corrosive chemicals are required, which makes processing fully compatible with the active polymer substrate used. After the nanoprisms are defined, the polymer is contracted via thermal activation (process shown in Supporting Information Figure 1). In this work, the nanospheres are drop-cast, which provides defect-free regions on the order of 100 μm2. While regions of this size are sufficient for the characterization in this study, much larger areas of nanosphere arrays are achievable, using spin coating, liquid surface selfassembly, dip drying, and colloid confinement methods.21,22 The increase in density of the nanoprisms is defined by the contraction the active polymer, which for the polystyrene polymer substrate studied is approximately 14% of its original area. For the nanoprism arrays studied here, this strain is sufficient to create nanogaps and significant change in optical properties, but we note that larger strains (and hence increase in nanostructure density) are achievable; for example, a 500% strain has been demonstrated using triblock liquid crystal polymers,23 which would
potentially reduce the polymer area to 4% and is a clear area of future focus to improve the technique presented here. The reduction in area from the active polymer contraction allows the formation of nanostructures in densities above that which are initially defined, regardless of the initial fabrication technique. Concomitant with the increase in density is the potential to create nanogaps between lithographically defined structures. In Figure 2df, we illustrate the nanogap formation in more detail for nanoprisms lengths of 120 nm to 1.6 μm. The nanogaps in Figure 2f are clearly sub-10 nm, comparable to the limit of highresolution techniques such as electron beam lithography and focused ion beam (FIB) techniques. While the precision is likely to be higher for these conventional techniques, the low-cost and highly multiplexed capability of active polymer nanofabrication give this technique an advantage for many practical applications. Nanogap formation allows the creation of tip-to-tip nanoprism structures (also called bow-ties), which have shown to be excellent substrates for localizing electromagnetic fields.24 Since the local electromagnetic field increases with decreasing nanogap distance between the nanoprisms, contracting the substrates should allow for significantly increased molecular sensitivity, as well as 2591
dx.doi.org/10.1021/nl201516g |Nano Lett. 2011, 11, 2590–2595
Nano Letters
LETTER
Figure 2. Active polymer fabrication of nanoprism arrays: (ac) finite element simulation of electric field amplitude surrounding initial and contracted nanoprism arrays; (df) SEM images of nanogaps between gold nanoprisms of different sizes. Nanoprism lengths are approximately (d) 1.6 μm, (e) 540 nm, and (f) 120 nm.
enhancing nonlinear effects such as second harmonic generation. Using finite element simulation, we show the order-of-magnitude improvement in local field enhancement upon contraction in Figure 2ac. We next consider how deformation of the nanoprisms varies at different length scales. As shown in Figure 3, the final substrate morphology is sensitive to the initial nanostructure size and thickness, and buckling and delamination regimes can be distinguished. Assuming the mismatch strain between the gold nanoprisms and the polymer to be the strain induced by the contracting polymer, and the modulus E = 78 GPa and Poisson ratio ν = 0.44 for gold, the mismatch stress between the nanostructures and the polymer is σm = 52 GPa. The thin Ti adhesion layer, used to improve adhesion between the substrate and the gold nanoprisms, can be ignored in this approximation, since it is much thinner than the gold layer, and because the relevant mechanical properties cause a very similar buckling response for Ti and Au in the size regime considered. The buckling regimes can be understood using a simple intuitive model. Consider the formula for buckling of a circular patch of a thin film25 σ b ¼ 1:2235
Ef h2f a2
ð1Þ
where σb is the critical stress for the lowest axially symmetric buckling mode, Ef is the plane strain modulus, hf is the film thickness, a is the debonded radius, and the negative sign indicates a compressive stress. This formula is a simplification of the experimental case studied: an infinite film is assumed (compared to the experimental condition of finite nanoprisms) and the debonded region is assumed to be circular. Nonetheless, the theory provides a useful intuitive picture of the physical scenario, and the results agree to an order-of-magnitude with experiment. The results of eq 1 are shown in Figure 3b,c. The stress mismatch allows for buckling to occur until a critical size is reached. Below this critical size, the stress required for buckling is higher than the stress mismatch, and the stress must be relieved by
another mechanism, such as delamination. The effect of increasing the thickness of the gold layer is also predicted by eq 1. Since the critical buckling stress is proportional to the second power of the nanoprism thickness, thicker films result in reduced buckling for a given size. These predictions match well with experiment; as shown in Figure 3a, buckling is clearly visible for larger and thinner nanoprisms, while smaller and thicker prisms tend to delaminate from the surface and “stand up.” We note that the useful increases in density, nanogap creation, and deformation are partially offset by a decrease in the order of the system; variations in the contraction behavior of the active polymer cause some defects to occur in the resulting nanoprism array. However, this is not a fundamental limitation, and future improvements in the uniformity of the polymer substrate may allow a reduction in the defects observed. Future research is needed to clarify the fundamental limits at which active polymers may uniformly contract, and how these limits depend on the form of activation (thermal, electrical, etc.). The changes in substrate morphology are correlated with distinct changes in optical properties. Since the size and periodicity of nanoprisms created by nanosphere lithography are not traditionally independent, active polymer nanofabrication allows a separate handle for tuning nanoprism periodicity, and hence tunability of their spectra. Tunability is an essential characteristic of any plasmonic architecture designed for molecular detection; first, because tuning resonance to the “biological window” is necessary to prevent cellular damage, and second, because tuning to the vicinity of specific molecular resonances can create large increases in both SERS26 and LSPR-shift27 signals. As shown in Figure 4, the extinction peak is blueshifted 200300 nm upon contraction with thicker nanoprisms exhibiting a greater blueshift after contraction. Three effects contribute to the observed blueshift. Radiative dipole interactions between the nanoprisms cause a blueshift as the lattice spacing is decreased.28 In addition, touching nanoprisms are no longer electrically isolated, again resulting in a blueshift. Finally, the deformation of individual nanoprisms may also con2592
dx.doi.org/10.1021/nl201516g |Nano Lett. 2011, 11, 2590–2595
Nano Letters
LETTER
Figure 3. Matrix of the morphology achieved for varying gold nanoprism size and thickness: (a) representative SEM images; the different regimes of buckling and delamination can be understood with a basic analytical model shown in (b) and (c), as predicted by eq 1.
tribute to the observed shift. More study is needed to precisely determine the relative importance of each of these effects. We note that for small and thick nanoprisms (s = 95 nm, h = 45 nm in Figure 4a), the lack of a uniform nanostructure prevents a clear extinction peak from emerging for the contracted polymer, consistent with the lack of buckling deformation in this regime. As discussed above, the contracted polymer substrates can be readily molded into 3D shapes, such as the lens shown in Figure 1e. We show the uniformity of spectra taken from different pixels of this substrate in Figure 4b. The pixel configuration was achieved using a stainless steel shadow mask (Fotofab), which served as a physical barrier during gold deposition, and spectra from individual pixels were achieved using a microscope (Zeiss Axiovert 200) and spectrophotometer (Princeton Instruments Acton SP2300). For each pixel, a spectrum was taken over approximately a 100 μm2 area, and the position of the pixel was recorded. Despite the significant change in structure curvature on the macroscale, the spectra obtained from varying positions on the lens are nearly identical; hence, we have achieved an omnidirectional
nanoplasmonic antenna lens. SEM images verified the local nanoscale morphology is not noticeably affected by macroscopic deformation (Supporting Information Figure 3). By integrating such devices with flexible electronics12 in future device iterations, individual pixels may be addressed simultaneously, thus achieving a “plasmonic compound eye.” To conclude, we have proposed a biologically inspired morphogenetic transformation of hybrid metal and polymer structures for defining nanostructure morphology and have demonstrated its utility in creating omnidirectional nanoplasmonic antenna arrays. Active polymer nanofabrication allows the potential to alter structure-to-structure distance down to sub-10 nm, surpassing both resolution and structural density of the initial placement. In addition, polymer molding offers macroscopic deformation of plasmonic architectures into arbitrary 3D geometries. We have demonstrated the size-dependent mechanical response of nanoplasmonic prism antenna arrays fabricated with active polymer fabrication and have correlated the results with analytic theory. Finally, we have correlated the changes in morphology with optical 2593
dx.doi.org/10.1021/nl201516g |Nano Lett. 2011, 11, 2590–2595
Nano Letters
LETTER
Figure 4. Optical changes with substrate contraction: (a) (left) schematic of nanoprism contraction and (right) representative extinction spectra of nanoprism arrays before (solid line) and after contraction (dashed line) for varying nanoprism size and height; (b) extinction spectra taken with varying ϕ and θ positions demonstrate the uniformity of pixels in the omnidirectional nanoplasmonic antenna lens shown in inset.
response. We believe active polymer-based nanofabrication will have wide utility for developing high-density device architectures for optical MEMS, NEMS, and nanoplasmonic devices and optical systems.
’ ASSOCIATED CONTENT
bS
Supporting Information. Description and figures of experimental and numerical methods. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Author Contributions †
These authors contributed equally to this work.
’ ACKNOWLEDGMENT This work was funded by The National Institutes of Health Nanomedicine Development Center for the Optical Control of Biological Function (PN2 EY018241), and the Center for Nanostructured Materials Technology under “21st Century Frontier R&D Programs” of the Ministry of Education, Science, and Technology, Korea. B.M.R. acknowledges support from The Berkeley Fellowship. L.Y.W. acknowledges support from a Taiwan Merit Scholarship. ’ REFERENCES (1) Bar-Cohen, Y. Electroactive Polymer (EAP) actuators as artificial muscles, 2nd ed.; SPIE: WA, 2004. (2) Mohr, R.; Kratz, K.; Weigel, T.; Lucka-Gabor, M.; Moneke, M.; Lendlein, A. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 3540–3545. (3) Lendlein, A.; Jiang, H.; J€unger, O.; Langer, R. Nature 2005, 434, 879–882. 2594
dx.doi.org/10.1021/nl201516g |Nano Lett. 2011, 11, 2590–2595
Nano Letters
LETTER
(4) Hidber, P. C.; Nealey, P. F.; Helbig, W.; Whitesides, G. M. Langmuir 1996, 12, 5209–5215. (5) Zhao, X.; Xia, Y.; Qin, D.; Whitesides, G. M. Adv. Mater. 1997, 9, 251–254. (6) Chen, C.-S.; Breslauer, D. N.; Luna, J. I.; Grimes, A.; Chin, W. C.; Lee, L. P.; Khine, M. Lab. Chip. 2008, 8, 622. (7) Fu, C.-C.; Grimes, A.; Long, M.; Ferri, C. G. L.; Rich, B. D.; Ghosh, S.; Gosh, S.; Lee, L. P.; Gopinathan, A.; Khine, M. Adv. Mater. 2009, 21, 4472–4476. (8) Lee, M. H.; Huntington, M. D.; Zhou, W.; Yang, J.-C.; Odom, T. W. Nano Lett. 2011, 11, 311. (9) Ross, B. M.; Wu, L. Y.; Lee, L. P. Proc. SPIE 2009, 7401, 74010K. (10) Sanchez, C.; Arribart, H.; Madeleine, M.; Guille, G. Nat. Mater. 2005, 4, 277. (11) Aizenberg, J.; Fratzl, P. Adv. Mater. 2009, 21, 387. (12) Kim, D.-H.; Rogers, J. A. Adv. Mater. 2008, 20, 4887–4892. (13) Willets, K.; van Duyne, R. P. Annu. Rev. Phys. Chem. 2006, 58, 267. (14) Kneipp, K.; Moskovits, M.; Kneipp, H. Surface-Enhanced Raman Scattering: Physics and Applications; Springer: Germany, 2006. (15) Fort, E.; Gresillon, S. J. Phys. D 2008, 41, 013001. (16) Liu, G. L.; Long, Y.-T.; Choi, Y.; Kang, T.; Lee, L. P. Nat. Methods 2007, 4, 1015. (17) Anker, J. N.; Hall, W. P.; Lyandres, O.; Shah, N. C.; Zhao, J.; Duyne, R. P. V. Nat. Mater. 2008, 7, 442–453. (18) Ross, B. M.; Lee, L. P. Opt. Express 2009, 17, 6860. (19) Kim, S.; Jin, J.; Kim, Y.-J.; Park, I.-Y.; Kim, Y.; Kim, S.-W. Nature 2008, 453, 757. (20) Haynes, C. L.; Duyne, R. P. V. J. Phys. Chem. B 2001, 105, 5599. (21) Jiang, P.; McFarland, M. J. J. Am. Chem. Soc. 2004, 126, 13778–13786. (22) Stavroulakis, P. I.; Christou, N.; Bagnall, D. Mater. Sci. Eng., B 2009, 3, 186–189. (23) Ahir, S. V.; Tajbakhsh, E. M. T., A. R. Adv. Func. Mater. 2006, 16, 556–560. (24) Sundaramurthy, A.; Crozier, K. B.; Kino, G. S.; Fromm, D. P.; Schuck, P. J.; Moerner, W. E. Phys. Rev. B 2005, 72, 165409. (25) Freund, L. B.; Suresh, S. Thin Film Materials: Stress, Defect Formation and Surface Evolution; Cambridge University Press: Cambridge, 2003. (26) McFarland, A. D.; Young, M. A.; Dieringer, J. A.; Duyne, R. P. V. J. Phys. Chem. B 2005, 109, 11279–11285. (27) Haes, A. J.; Zou, S.; Zhao, J.; Schatz, G. C.; Duyne, R. P. V. J. Am. Chem. Soc. 2006, 128, 10905–10914. (28) Haynes, C. L.; McFarland, A. D.; Zhao, L.; Duyne, R. P. V.; Schatz, G. C.; Gunnarsson, L.; Prikulis, J.; Kasemo, B.; K€all, M. J. Phys. Chem. B 2003, 107, 7337.
2595
dx.doi.org/10.1021/nl201516g |Nano Lett. 2011, 11, 2590–2595
