Letter pubs.acs.org/NanoLett
Deterministic Symmetry Breaking of Plasmonic Nanostructures Enabled by DNA-Programmable Assembly Matthew R. Jones,†,‡ Kevin L. Kohlstedt,‡,§ Matthew N. O’Brien,‡,§ Jinsong Wu,† George C. Schatz,‡,§ and Chad A. Mirkin*,†,‡,§ †
Department of Materials Science and Engineering, ‡International Institute for Nanotechnology, and §Department of Chemistry, Northwestern University, Evanston, Illinois 60208, United States S Supporting Information *
ABSTRACT: The physical properties of matter rely fundamentally on the symmetry of constituent building blocks. This is particularly true for structures that interact with light via the collective motion of their conduction electrons (i.e., plasmonic materials), where the observation of exotic optical effects, such as negative refraction and electromagnetically induced transparency, require the coupling of modes that are only present in systems with nontrivial broken symmetries. Lithography has been the predominant fabrication technique for constructing plasmonic metamaterials, as it can be used to form patterns of arbitrary complexity, including those with broken symmetry. Here, we show that low-symmetry, one-dimensional plasmonic structures that would be challenging to make using traditional lithographic techniques can be assembled using DNA as a programmable surface ligand. We investigate the optical properties that arise as a result of systematic symmetry breaking and demonstrate the appearance of π-type coupled modes formed from both dipole and quadrupole nanoparticle sources. These results demonstrate the power of DNA assembly for generating unusual structures that exhibit both fundamentally insightful and technologically important optical properties. KEYWORDS: DNA, assembly, plasmonics, nanoparticle, superlattice, symmetry
S
orient strands in a surface-normal direction,6,18,22 constructs known as programmable atom equivalents (PAEs).19 In this work, we exploit three important features of the assembly of PAEs to construct symmetry-broken, plasmonically coupled nanoparticle superlattices: (1) The use of anisotropic building blocks whose shape imposes directional interactions between particles that lead to reduced-dimensionality superlattices;23−28 (2) a DNA design in which two sets of differently sized or shaped particles are modified with oligonucleotides that have complementary sequences, forcing the cocrystallization of structurally dissimilar particles in a manner that breaks symmetry;29 and (3) the use of nonspherical plasmonic nanoparticles that support strongly confined dipolar and quadrupolar resonances,30,31 necessary for the observation of extended π-type coupled modes that are more amenable to metamaterial behavior than the more common σ-type modes.32 This set of design elements allows for the generation of extremely sophisticated assembled superlattices that would be impractical to fabricate with traditional lithographic methods. The particle building blocks that form the basis of this work are two-dimensional platelike nanostructures composed of gold (Figure S1).31 These materials are sufficiently thin (∼7 nm) and anisotropic (lateral dimensions tunable between 50 and
elf-assembly approaches that utilize colloidal nanoparticles as building blocks have a number of advantages over lithographic methods for the construction of plasmonic metamaterials: (1) solution-phase synthesized nanoparticles are often single crystalline and possess superior optical properties due to their low defect density,1 (2) particles can be generated in large quantities and arranged simultaneously, resulting in increased yield and scale,2 (3) the length scales that are routinely accessed via self-assembly methods are considerably smaller than those available to traditional lithographic techniques, as they are defined by molecular interactions,3,4 (4) solution-dispersible materials exist in a three-dimensional environment and can exhibit optical properties that are less sensitive to orientation than those confined to a twodimensional substrate.5 However, the majority of self-assembly methodologies rely on spherical building blocks that interact via nearly isotropic potentials that can result in densely packed, complex structures but with invariably high-symmetry.6−9 Consequently, these approaches offer limited access to nanoparticle arrangements that exhibit metamaterial-like properties.10−14 Directed assembly based on sequence-programmable DNA interactions represents a particularly powerful approach for synthesizing complex particle-based architectures.15−19 Nucleic acids can either be folded into discrete molecular templates with location-specific nanoparticle binding sites,18,20,21 or anchored to rigid nanoparticle cores that act to bundle and © XXXX American Chemical Society
Received: July 18, 2017 Revised: August 11, 2017
A
DOI: 10.1021/acs.nanolett.7b03067 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
used to anchor and orient strands relative to the gold nanoparticle surface and programmable linker oligonucleotides are added to dictate particle interactions via the nucleobase sequence of the terminal “sticky ends” (Figure 1b; see Table S1 for all DNA sequences). Superlattices are characterized using synchrotron-based small-angle X-ray scattering (SAXS) to probe a large number of structures simultaneously and assign symmetries on the basis of the structure factor, S(q).23,29 All superlattices are encapsulated in silica to preserve their lattice symmetry and spacing for imaging via transmission electron microscopy (see Supporting Information for details). In order to rigorously describe the assembled superlattices and quantify how symmetries are broken through rational structural modification, we have classified each system using the nomenclature of three-dimensional group theory (see Supporting Information). Superlattices assembled from self-complementary circular disk particles (sticky end: 5′ GCGC 3′) with different diameters (Figure 1c,d; Figure S1) and DNA lengths (Figure 1e,f) result in monoperiodic arrays with tunable widths and interparticle spacings, respectively (Figures S2 and S3). If particles of two different sizes are modified with complementary sticky-end sequences (5′ AGAGA 3′/5′ TCTCT 3′), biperiodic superlattices are formed (Figure 1g,h; Figures S2 and S3); relative to the monoperiodic array, this results in a decrease in translational symmetry. It is worth noting that the broken symmetries of these nanoparticle cocrystals would be challenging to achieve using traditional self-assembly approaches, wherein particle sizes and shapes must be precisely matched to prevent phase separation.34 In order to understand the plasmonic consequences of programmable symmetry breaking, we utilize the conceptual framework of plasmon hybridization theory.32 When the discrete modes of two nanostructures couple, one expects a hybrid low-energy bonding mode, in which the dipoles on each particle interact favorably, and a hybrid high-energy antibonding mode, in which the dipoles on each particle interact unfavorably. In most cases (e.g., spherical particles),13,14,22,32 these modes arise from dipole alignment in a head-to-tail (bonding) or head-to-head (antibonding) fashion, a geometry known as σ-type hybridization (Figure S4). Because the antibonding configuration possesses no net dipole, it is unable to interact with light (a dipole excitation source) and is described as optically dark. Only the low-energy (redshifted) bonding modes possess a net dipole, are optically bright, and therefore are most frequently observed in cases of nanoparticle assembly.13,14,22 In contrast, the dipolar mode for circular disks is strongly confined to the plane of the particle31 and therefore is oriented orthogonal to the unit lattice vector. This leads to an alternative side-by-side orientation of dipoles that align antiparallel to form a low-energy bonding mode and align parallel to form a highenergy antibonding mode (Figure 2a). As a consequence, the low-energy bonding mode is dark and the high-energy antibonding mode is bright. We observe this effect by collecting solution-phase extinction spectra before and after the assembly of one-dimensional monoperiodic superlattices (Figure 2a,b; Figures S5 and S6). Indeed, the plasmon resonance shows a pronounced blueshift as a result of the antibonding π-type (i.e., side-by-side) dipole coupling in the array. Because the extent to which the antibonding mode is blueshifted should depend on the particle coupling strength and dipole magnitude, the proposed dipole alignment can be tested experimentally by
200 nm) so as to present a dense array of highly oriented oligonucleotides orthogonal to their broad, atomically flat facets, and considerably fewer, poorly oriented oligonucleotides orthogonal to their edges. As a consequence, these structures have been shown to greatly favor face-to-face hybridization interactions both kinetically and thermodynamically,24,25 resulting in the formation of highly ordered one-dimensional superlattices (Figure 1a).23 These particles can be synthesized
Figure 1. DNA-programmable assembly of low-symmetry nanoparticle superlattices. (a) Schematic illustration of the numerous DNA “stickyend” interactions and preferential face-to-face alignment used to drive the formation of one-dimensional superlattices. (b) DNA design in which a nanoparticle-bound strand is hybridized to a linker oligonucleotide presenting sequence-programmable sticky ends. Small-angle X-ray scattering (c,e,g) and transmission electron microscopy (d,f,h) characterization of circular disk superlattices showing control over array width (c,d), interparticle spacing (e,f), and periodicity (g,h) by controlling the particle size, DNA length, and sticky end sequence, respectively. Simulated SAXS data for perfect crystals are shown as dotted lines directly beneath experimental data.
with a circular cross-sectional shape (circular disks), which exhibit a single in-plane dipole resonance, and a triangular cross-sectional shape (triangular prisms), which exhibit both inplane dipole and in-plane quadrupole resonances (Figure S1).31,33 In addition, the small thickness of the particles results in strong in-plane (longitudinal) modes with no measurable out-of-plane (transverse) modes (Figure S1),31 a property that will be important to understand the plasmonic response of the assembled superlattices. The particles are assembled using a previously reported DNA design:23,29 terminal hexyl-thiol-modified oligonucleotides are B
DOI: 10.1021/acs.nanolett.7b03067 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 2. Optical characterization of one-dimensional nanoparticle superlattices. Plasmon mode hybridization diagrams (a,c), and extinction spectra (b,d) for monoperiodic (a, b) and biperiodic (c, d) arrays showing the absence of a low-energy π-type bonding mode (red X, a,b) except when symmetry is rationally broken (c,d). Discrete modes are shown in blue or red while coupled modes are shown in purple. Simulated data (dotted lines) is shown offset adjacent to experimental data (solid lines). Monoperiodic arrays assembled with shorter DNA lengths (e) and larger circular disks (f) show antibonding modes with greater energy shifts with respect to the discrete modes (ΔEnergy = Energyassembled − Energydiscrete). (g) Biperiodic arrays assembled from pairs of circular disks with increasing size ratio show less-efficient bonding and antibonding mode splitting. In (e− g), experimental data are shown in solid symbols while simulated data are shown in open symbols. Electric field distributions and dipole orientations calculated for biperiodic arrays at the antibonding (h) and bonding (i) mode energies confirm the parallel and antiparallel net dipole arrangements, respectively.
orientations at the energies of the blueshifted and redshifted peaks clearly indicate the presence of parallel and alternating net dipole orientations, confirming the assignment of antibonding and bonding mode character, respectively (Figure 2h,i; Figure S6). While π-type hybridized modes are well-known in the literature,32 their observation is almost exclusively limited to structures fabricated or deposited on substrates with optical sources intentionally polarized to excite them.10,35,36 The DNAassembled superlattices described here have been probed optically in a solution-phase, orientation-averaged measurement that negates the role of polarization-dependent effects. In addition, because the circular disks do not possess an observable transverse mode,31 the structure cannot be excited along the long dimension (lattice vector) of the superlattice and thus σ-type coupling is forbidden. As a result, the only bright modes available to the system are those with a π-type antibonding character. This geometry of dipoles is known to possess a magnetic resonance and is considered one of the fundamental building blocks of all metamaterials.32 Indeed, our simulations reveal that the circular disk superlattices presented here possess magnetic and Fano modes at visible frequencies (Figure S7). Manipulation of these coupled resonances is particularly promising for the development of advanced solution-phase metamaterials. Although examples of blueshifted modes have been reported for solution-phase gold nanorods37 and copper sulfide disks,38 the optical consequences of symmetry breaking in these systems have been challenging to
decreasing the particle spacing and increasing the particle size, respectively (Figure 2e,f; Figure S5). As predicted, this results in greater mode splitting. These results are confirmed in silico using the discrete dipole approximation (DDA) method to simulate the extinction spectra of perfectly aligned gold disk arrays (see Supporting Information for computational methods). One of the most powerful features of DNA-mediated assembly is the ability to program particle interactions via sequence-specific hybridization and form cocrystalline biperiodic arrays (Figure 1g,h; Figures S2 and S3).18,29 In this case, the unequal magnitude of the two dipoles suggests that the bonding mode should no longer be dark, as the broken symmetry allows for a net dipole on the structure (Figure 2c). To test this prediction, we compared the extinction spectra for isolated particles and coupled superlattices composed of two differently sized circular disks bearing complementary oligonucleotides (Figure 2c,d, Figure S5). Indeed, biperiodic superlattices support two plasmon modes, one that is redshifted and one that is blueshifted with respect to the isolated particle plasmons. In addition, because the magnitude of the net dipole determines the extent to which light is able to excite a mode, we observe the bonding (redshifted) peak having a lower extinction than the antibonding (blueshifted) peak. Superlattices cocrystallized from combinations of disks with a larger difference in their diameters and thus more disparate discrete plasmon energies show less efficient coupling, a smaller energy shift, and weaker bonding/antibonding modes (Figure 2g; Figure S5). Simulations of the electric field profile and dipole C
DOI: 10.1021/acs.nanolett.7b03067 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 3. Characterization of shape-biperiodic symmetry-broken nanoparticle superlattices. Arrays consist of alternating arrangements of circular disks and triangular prisms (a−d) or circular disks and rod dimers (d−h). (a,e) Images captured from a tomographic series and 3D modeled structures, each tilted by 15° increments out-of-plane illustrate the three-dimensional features of the materials. (b,e) Extinction spectra show discrete particles in red, blue, or green, denoting triangular prisms, circular disks, and rods, respectively, while superlattice modes are shown in purple. Simulated data (dotted lines) are shown below experimental data (solid lines). Relevant modes are denoted by labels in (b,f) with corresponding diagrams (c,g) and simulations for the normalized local electric field distribution and dipole orientations (d,h). Labels denote the mode order (left superscript, dipole l = 1, quadrupole l = 2), bonding or antibonding character (right superscript), and participating particles (right subscript, p = triangular prism, d = circular disk, r = rod dimer). Note: field intensity color scale for the 1π*p/d mode has been decreased by a factor of 2 for clarity.
results in a lack of triangular prism tip registry (Figure 3a, Movie S2).29 The energy alignment of plasmon modes results in dipolar coupling to generate bonding (996 nm, 1πp/d) and antibonding (760 nm, 1π*p/d) modes but also results in a new quadrupole-like mode (618 nm, 2π*p/d, Figure 3b−d). This is surprising because quadrupolar modes on circular disks are only weakly excited with far-field light, as they are symmetryforbidden. Simulations reveal that this resonance arises because the partially bright quadrupole mode on the triangular prisms can excite a quadrupole resonance on the neighboring circular disks as a result of near-field coupling between the two nanostructures (Figure 3d; Figure S10, see Supporting Information for discussion). Triangular prisms are known to have another partially bright multipole, the l = 3 hexapolar mode.33,39 When triangular prisms are cocrystallized with larger circular disks (94 nm, Movie S3), a similar set of coupled dipolar and near-fieldinduced quadrupolar resonances are observed, but with an additional hexapolar-like mode (Figure S10d−f, Figures S2, S3, and S8). In order for a circular disk to support multipolar modes, the circular symmetry of the disk has to be broken by the near-field of the neighboring triangular prism. Although it is
elucidate experimentally because of a lack of programmability in the forces governing particle assembly. Further symmetry breaking is possible via the assembly of nanoparticle building blocks with greater shape anisotropy. Triangular prism nanoparticles have a similar thickness to circular disks (∼7 nm) but have lower in-plane rotational symmetry (3-fold versus continuous) and, as a result, possess a partially bright quadrupolar resonance (Figure S1).33 In addition to the dipolar modes observed for circular disks (see above), monoperiodic arrays of triangular prisms show a coupled π-type antibonding quadrupolar resonance (Figures S8 and S9). A tomographic TEM tilt series illustrates the threedimensional structure of the superlattice and reveals that the tips of the triangular prims are in registry (Movie S1). This is consistent with the hypothesis that the system will assemble in a manner that maximizes the number of interparticle DNA linkages.6,18 When triangular prisms are cocrystallized with circular disks (Figures S2, S3, and S8), the additional shape-derived symmetry breaking results in the appearance of new hybrid π-type modes. As expected, tomographic TEM tilt data demonstrate that the alternating arrangement of particles D
DOI: 10.1021/acs.nanolett.7b03067 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters not surprising that larger disks can more easily accommodate multipolar modes,40 it is important to understand how the symmetry of the coupled modes supported by the shapebiperiodic arrays can be tuned by the size of the components. The lowest-symmetry superlattice we constructed formed via the cocrystallization of circular disks and rods (Figure 3e; Figures S2, S3, and S8). Interestingly, although there is no explicit attractive interaction between rods, they adopt a parallel dimer motif that separates each circular disk in a onedimensional array (see Supporting Information for discussion); tomographic TEM tilt data shows that rod dimers lack orientational correlation along the unit lattice vector (Figure 3e, Movie S4). The modes that result from this structure can be best thought of as arising from coupling between circular disks and rod dimers (Figure 3f−h). A bright antibonding mode on the rod dimer can π-type hybridize with the circular disk dipole in a bonding fashion to generate a redshifted mode (900 nm, 1 πd/r), and in an antibonding fashion to generate a blueshifted mode (700 nm, 1π*d/r, Figure 3f−h). Alternatively, when the rod dimer is dark (bonding), it does not participate in plasmon coupling, resulting in a mode in which the circular disks couple in an antibonding fashion, albeit with an increased effective lattice spacing (800 nm, 1π*d/d, Figure 3f−h). In conclusion, we have demonstrated that fundamental metamaterial properties arise as a direct consequence of deterministic symmetry breaking enabled by DNA-programmable nanoparticle assembly. The ability to engineer colloidal particle interactions in a manner that does not dictate highsymmetry arrangements has been a longstanding goal in chemistry, physics, and materials science.2,15,41 This accomplishment has particular relevance for the scalable synthesis of photonic and plasmonic metamaterials, as low-symmetry arrangements are often compulsory and traditional lithographic methods are low-yielding. Natural extensions of this work to program compositional or dynamically switchable asymmetry are expected to establish these structures as building blocks for future solution-phase metamaterials.
■
Jinsong Wu: 0000-0002-7305-7927 George C. Schatz: 0000-0001-5837-4740 Chad A. Mirkin: 0000-0002-6634-7627 Author Contributions
M.R.J., K.L.K., M.N.O., G.C.S., and C.A.M. designed the experiments. M.R.J. collected and analyzed experimental data for optical and electron microscopy studies. M.R.J. and M.N.O. prepared superlattice samples and collected and analyzed X-ray data. K.L.K. and G.C.S. conducted and analyzed plasmonic simulation results. J.W. and M.N.O. collected and compiled the images for the tomographic tilt series movies. M.R.J. and C.A.M. wrote the manuscript. All authors have given approval to the final version of the manuscript. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This material is based upon work supported by the AFOSR GRANT12204315 and the National Science Foundation’s MRSEC program under award DMR-1121262. Portions of this work were carried out at the DuPont-Northwestern-Dow Collaborative Access Team (DND-CAT) beamline located at Sector 5 of the Advanced Photon Source (APS). DND-CAT is supported by E. I. DuPont de Nemours & Co., Dow Chemical Company, and the state of Illinois. The transmission electron microscopy work was carried out in the EPIC facility of NUANCE Center at Northwestern University, which has received support from the Soft and Hybrid Nanotechnology Experimental (SHyNE) Resource (NSF NNCI-1542205); the MRSEC program (NSF DMR-1121262) at the Materials Research Center; the International Institute for Nanotechnology (IIN); the Keck Foundation; and the State of Illinois, through the IIN. M.R.J. acknowledges the NSF for a graduate research fellowship and Northwestern University for a Ryan Fellowship. M.N.O. also acknowledges the NSF for a graduate research fellowship.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.7b03067. Figures S1−S19 and additional discussion (PDF) Tomographic tilt series of a monoperiodic array of 74 nm edge length triangular prisms embedded in silica (AVI) Tomographic tilt series of a shape-biperiodic array of 74 nm edge length triangular prisms and 67 nm circular disks embedded in silica (AVI) Tomographic tilt series of a shape-biperiodic array of 74 nm edge length triangular prisms and 94 nm circular disks embedded in silica (AVI) Tomographic tilt series of a shape-biperiodic array of 72 nm circular disks and 46 × 12 nm rods embedded in silica (AVI)
■
REFERENCES
(1) Huang, J.-S.; Callegari, V.; Geisler, P.; Brüning, C.; Kern, J.; Prangsma, J. C.; Wu, X.; Feichtner, T.; Ziegler, J.; Weinmann, P.; Kamp, M.; Forchel, A.; Biagioni, P.; Sennhauser, U.; Hecht, B. Nat. Commun. 2010, 1, 150. (2) Gao, B.; Arya, G.; Tao, A. R. Nat. Nanotechnol. 2012, 7, 433. (3) Moreau, A.; Ciraci, C.; Mock, J. J.; Hill, R. T.; Wang, Q.; Wiley, B. J.; Chilkoti, A.; Smith, D. R. Nature 2012, 492, 86. (4) Fan, J. A.; Wu, C.; Bao, K.; Bao, J.; Bardhan, R.; Halas, N. J.; Manoharan, V. N.; Nordlander, P.; Shvets, G.; Capasso, F. Science 2010, 328, 1135. (5) Sheikholeslami, S. N.; Alaeian, H.; Koh, A. L.; Dionne, J. A. Nano Lett. 2013, 13, 4137. (6) Macfarlane, R. J.; Lee, B.; Jones, M. R.; Harris, N.; Schatz, G. C.; Mirkin, C. A. Science 2011, 334, 204. (7) Shevchenko, E. V.; Talapin, D. V.; Kotov, N. A.; O’Brien, S.; Murray, C. B. Nature 2006, 439, 55. (8) Boles, M. A.; Talapin, D. V. J. Am. Chem. Soc. 2015, 137, 4494. (9) Kalsin, A. M.; Fialkowski, M.; Paszewski, M.; Smoukov, S. K.; Bishop, K. J. M.; Grzybowski, B. A. Science 2006, 312, 420. (10) Zhang, S.; Genov, D. A.; Wang, Y.; Liu, M.; Zhang, X. Phys. Rev. Lett. 2008, 101, 047401. (11) Kanté, B.; Park, Y.-S.; O’Brien, K.; Shuldman, D.; LanzillottiKimura, N. D.; Jing Wong, Z.; Yin, X.; Zhang, X. Nat. Commun. 2012, 3, 1180. (12) Soukoulis, C. M.; Wegener, M. Nat. Photonics 2011, 5, 523.
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Matthew R. Jones: 0000-0002-9289-291X Kevin L. Kohlstedt: 0000-0001-8045-0930 E
DOI: 10.1021/acs.nanolett.7b03067 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters (13) Ross, M. B.; Ku, J. C.; Vaccarezza, V. M.; Schatz, G. C.; Mirkin, C. A. Nat. Nanotechnol. 2015, 10, 453. (14) Ye, X.; Chen, J.; Diroll, B. T.; Murray, C. B. Nano Lett. 2013, 13, 1291. (15) Wang, Y.; Wang, Y.; Breed, D. R.; Manoharan, V. N.; Feng, L.; Hollingsworth, A. D.; Weck, M.; Pine, D. J. Nature 2012, 491, 51. (16) Rogers, W. B.; Manoharan, V. N. Science 2015, 347, 639. (17) Yan, W.; Xu, L.; Xu, C.; Ma, W.; Kuang, H.; Wang, L.; Kotov, N. A. J. Am. Chem. Soc. 2012, 134, 15114. (18) Jones, M. R.; Seeman, N. C.; Mirkin, C. A. Science 2015, 347, 1260901. (19) Macfarlane, R. J.; O’Brien, M. N.; Petrosko, S. H.; Mirkin, C. A. Angew. Chem., Int. Ed. 2013, 52, 5688. (20) Kuzyk, A.; Schreiber, R.; Fan, Z.; Pardatscher, G.; Roller, E.-M.; Hogele, A.; Simmel, F. C.; Govorov, A. O.; Liedl, T. Nature 2012, 483, 311. (21) Kuzyk, A.; Schreiber, R.; Zhang, H.; Govorov, A. O.; Liedl, T.; Liu, N. Nat. Mater. 2014, 13, 862. (22) Mirkin, C. A.; Letsinger, R. L.; Mucic, R. C.; Storhoff, J. J. Nature 1996, 382, 607. (23) Jones, M. R.; Macfarlane, R. J.; Lee, B.; Zhang, J.; Young, K. L.; Senesi, A. J.; Mirkin, C. A. Nat. Mater. 2010, 9, 913. (24) Jones, M. R.; Macfarlane, R. J.; Prigodich, A. E.; Patel, P. C.; Mirkin, C. A. J. Am. Chem. Soc. 2011, 133, 18865. (25) Damasceno, P. F.; Engel, M.; Glotzer, S. C. Science 2012, 337, 453. (26) Henzie, J.; Grünwald, M.; Widmer-Cooper, A.; Geissler, P. L.; Yang, P. Nat. Mater. 2011, 11, 131. (27) Boneschanscher, M. P.; Evers, W. H.; Geuchies, J. J.; Altantzis, T.; Goris, B.; Rabouw, F. T.; van Rossum, S. A. P.; van der Zant, H. S. J.; Siebbeles, L. D. A.; Van Tendeloo, G.; Swart, I.; Hilhorst, J.; Petukhov, A. V.; Bals, S.; Vanmaekelbergh, D. Science 2014, 344, 1377. (28) O’Brien, M. N.; Girard, M.; Lin, H.-X.; Millan, J. A.; Olvera de la Cruz, M.; Lee, B.; Mirkin, C. A. Proc. Natl. Acad. Sci. U. S. A. 2016, 113, 10485. (29) O’Brien, M. N.; Jones, M. R.; Lee, B.; Mirkin, C. A. Nat. Mater. 2015, 14, 833. (30) Kelly, K. L.; Coronado, E.; Zhao, L. L.; Schatz, G. C. J. Phys. Chem. B 2003, 107, 668. (31) O’Brien, M. N.; Jones, M. R.; Kohlstedt, K. L.; Schatz, G. C.; Mirkin, C. A. Nano Lett. 2015, 15, 1012. (32) Halas, N. J.; Lal, S.; Chang, W.-S.; Link, S.; Nordlander, P. Chem. Rev. 2011, 111, 3913. (33) Shuford, K. L.; Ratner, M. A.; Schatz, G. C. J. Chem. Phys. 2005, 123, 114713. (34) Paik, T.; Murray, C. B. Nano Lett. 2013, 13, 2952. (35) Funston, A. M.; Novo, C.; Davis, T. J.; Mulvaney, P. Nano Lett. 2009, 9, 1651. (36) Sheikholeslami, S.; Jun, Y.-w.; Jain, P. K.; Alivisatos, A. P. Nano Lett. 2010, 10, 2655. (37) Jain, P. K.; Eustis, S.; El-Sayed, M. A. J. Phys. Chem. B 2006, 110, 18243. (38) Hsu, S.-W.; Ngo, C.; Tao, A. R. Nano Lett. 2014, 14, 2372. (39) Losquin, A.; Zagonel, L. F.; Myroshnychenko, V.; RodríguezGonzález, B.; Tencé, M.; Scarabelli, L.; Förstner, J.; Liz-Marzán, L. M.; García de Abajo, F. J.; Stéphan, O.; Kociak, M. Nano Lett. 2015, 15, 1229. (40) Schmidt, F. P.; Ditlbacher, H.; Hofer, F.; Krenn, J. R.; Hohenester, U. Nano Lett. 2014, 14, 4810. (41) Chen, Q.; Bae, S. C.; Granick, S. Nature 2011, 469, 381.
F
DOI: 10.1021/acs.nanolett.7b03067 Nano Lett. XXXX, XXX, XXX−XXX
