Letter pubs.acs.org/NanoLett
Induced Chirality through Electromagnetic Coupling between Chiral Molecular Layers and Plasmonic Nanostructures Nadia A. Abdulrahman,† Z. Fan,‡ Taishi Tonooka,† Sharon M. Kelly,§ Nikolaj Gadegaard,∥ Euan Hendry,*,⊥ Alexander O. Govorov,*,‡ and Malcolm Kadodwala*,† †
School of Chemistry, Joseph Black Building, University of Glasgow, Glasgow G12 8QQ, United Kingdom Department of Physics and Astronomy, Ohio University, Athens, Ohio 45701, United States § College of Medical, Veterinary and Life Sciences, Institute of Molecular, Cell and Systems Biology, University of Glasgow, Glasgow G12 8QQ, United Kingdom ∥ Division of Biomedical Engineering, School of Engineering, Rankine Building, University of Glasgow, Glasgow G12 8LT, United Kingdom ⊥ School of Physics, University of Exeter, Stocker Road, Exeter, EX4 4QL, United Kingdom ‡
S Supporting Information *
ABSTRACT: We report a new approach for creating chiral plasmonic nanomaterials. A previously unconsidered, far-field mechanism is utilized which enables chirality to be conveyed from a surrounding chiral molecular material to a plasmonic resonance of an achiral metallic nanostructure. Our observations break a currently held preconception that optical properties of plasmonic particles can most effectively be manipulated by molecular materials through near-field effects. We show that far-field electromagnetic coupling between a localized plasmon of a nonchiral nanostructure and a surrounding chiral molecular layer can induce plasmonic chirality much more effectively (by a factor of 103) than previously reported near-field phenomena. We gain insight into the mechanism by comparing our experimental results to a simple electromagnetic model which incorporates a plasmonic object coupled with a chiral molecular medium. Our work offers a new direction for the creation of hybrid molecular plasmonic nanomaterials that display significant chiroptical properties in the visible spectral region. KEYWORDS: Chiral, plasmonic, metamaterials, optics, nanostructures
C
hiral plasmonic nanomaterials1−3 are a remarkable new class of engineered materials, whose unique chiroptical properties have been used to create negative refractive index media,4,5 broadband circular polarisers,1 and superchiral electromagnetic (EM) fields for ultrasensitive structural characterization of biomaterials.6 The optical properties of these lithographically created chiral nanostructures are governed by their forms and symmetries (i.e., they are created with very specific shapes which have no mirror symmetry) and are inherently constrained by design. Such nanomaterials can be complex and, particularly in the case of three-dimensional (3-D) chiral structures, difficult to fabricate.1 Here, we demonstrate a previously unconsidered mechanism by which chiroptical behavior can be induced in the resonances of achiral plasmonic nanostructures, driven by radiative electromagnetic coupling between metallic particle plasmons and a surrounding chiral isotropic medium. This is fundamentally different from previously reported phenomena caused either by orbital hybridization7−9 or near-field, dipole−dipole interactions between chiral molecules and particle plasmons.10−14 Due to © 2012 American Chemical Society
its long range of interaction, the new effect reported here can be three orders of magnitude more effective at inducing circular dichroism (CD) in the absorption resonances of achiral nanostructures than near-field, dipole−dipole interactions. As well as discovering a new and interesting fundamental effect, we suggest that the hybrid nanoplasmonic materials studied, once optimized, could be a route to, for example, negative refractive index media. This approach offers flexibility in both fabrication and application of chiral metamaterials by combining the physical engineering of the metal structure with the molecular properties of the dielectric medium in which they are embedded. Crucially, such an approach is not limited by design constraints of the nanostructures themselves. It has been previously shown10,11,14 that CD can be induced in plasmonic resonances of spherical particles through shortrange coupling between the nanoparticle and a surrounding, Received: November 17, 2011 Revised: January 11, 2012 Published: January 20, 2012 977
dx.doi.org/10.1021/nl204055r | Nano Lett. 2012, 12, 977−983
Nano Letters
Letter
Figure 1. (A) Finite element modeling27 of plasmon absorption in the cross arrays. The inset shows the time-averaged electric fields of the two plasmon modes. Red represents a time-averaged field enhancement of 8 times the incident field, E0. To aid comparison plasmonic resonances are labeled * and **. (B) An extinction spectrum of a plain cross substrate. (C) Extinction spectrum taken from a 70 μg cm−2 FMN film on cross (black) and plain glass (red) substrates. (D) CD spectra collected from a cross substrate (blue) and from 70 μg cm−2 films deposited on both cross (black) and plain glass (red) substrates.
beam lithography (see Supporting Information). The cross dimensions were chosen for the experiment because they have previously been shown to give rise to a strong plasmon resonance in the visible range6 that is convenient for our chiroptical studies. Moreover, such shapes have previously been shown not to give rise to strong dipole−dipole interactions with single layers of chiral molecules.6 The crosses were fabricated using electron beam lithography on a glass substrate in a square lattice with a periodicity of 800 nm. Crosses with nominal thicknesses of 55, 100, and 130 nm and with a 5 nm Ti adhesion layer were used. Flavin mononucleotide (FMN), a biomolecule which displays a strong chiroptical signal in the near UV, was deposited in thin films onto the substrates by evaporation from aqueous solution of FMN (concentration 2.2 mM), using a method similar to that used by Sugawara et al.16 As in this previous study, we have used average molecular surface density (the mass of FMN per cm2) to parametrize the films, since average film thickness scales with surface density. Films of increasing thickness were produced by depositing increasing volumes of FMN solution. The morphology/ heterogeneity of the deposited FMN films was monitored using atomic force microscopy (AFM) (see Supporting Information). The FMN films were continuous and relatively uniform and display a peak to trough roughness of 7 μg cm−2 (∼80 nm film thickness) (see Supporting Information), indicating the localized plasmon near-fields decay on a length scale ∼80 nm. This agrees with the calculated fields shown in the inset of Figure 1A. In Figure 1D we show CD spectrum collected from FMN deposited on the cross substrates. To facilitate comparison between CD spectra of samples with different FMN coverages, we have replotted spectra in terms of the anisotropy factor (g), Figures 2A,D. The g factor characterizes the strength of chiroptical response:
g=
A + − A− I0 , A ± = log A̅ Itransmitted, ±
where the A± are the extinctions for left and right circularly polarized light and A̅ = (A+ + A−)/2 is an average extinction. The g factor of the induced plasmonic resonance in 50 nm thick crosses is plotted as a function of FMN surface density in Figure 2b. There is a clear correlation between increasing g factor and increasing surface density (film thickness). Thus, the level of CD induced in the plasmon resonances of crosses increases with FMN film thickness. It is clear that, for cross substrates with thick FMN overlayers (surface densities ≥70 μg cm−2, ∼800 nm thickness), a broad plasmonic CD feature at 570 nm is observed. The g factors in this region are of the order of 10−3. We have also made measurements for single monolayers of FMN adsorbed from solution on to crosses and found no measurable CD induced. The dependence of the g factor on the thickness of the Au cross was also investigated by measuring CD spectra from 70 μg cm−2 FMN films on 100 and 130 nm thick crosses, Figure 2C,D. The level of chiral anisotropy of the plasmonic resonance increases with the thickness of the cross in each case, reaching values of the order of 10−2−10−1. This demonstrates a correlation between the 980
dx.doi.org/10.1021/nl204055r | Nano Lett. 2012, 12, 977−983
Nano Letters
Letter
considered only the simplest geometry: a spherical core−shell structure consisting of a metallic sphere surrounded by a spherical shell of chiral material. While the experimental geometry is somewhat more complex, we will show that our relatively simple model encompasses most of the important physical behavior. The usual form of Maxwell’s equations given in the Supporting Information should be solved assuming the following important relations between the electromagnetic fields:22,23
Dω = ε(r) · Eω + i ξc(r) · Bω H ω = Bω /μ + i ξc(r) · Eω
(1)
where ξc is the key local parameter describing the chiral property of a molecular layer. This parameter is nonzero only within the molecular shell. In addition, the chiral shell has a dielectric function, εc, producing the scattering and absorption effects of the molecular layer. The nonchiral Au sphere is described with a local dielectric function, εAu, taken from empirical tables.24 The magnetic susceptibility (μ) is assumed to be unity for simplicity. The whole core−shell structure is assumed to exist in vacuum, ε0 = 1. An incident electromagnetic wave approaches the system along the z direction. A general solution for the chiral layered spherical system is found using Mie theory25 adapted to include chiral media.26 We give the details of derivation in the Supporting Information. While the metal function εAu is known, it is important to make a reasonable choice for the important functions ξc and εc for the chiral molecular film. We calibrate these functions in the following way: A single molecule is assumed to have a single optical transition that gives rise to peak CD extinction εC̅ D,Mol = 20/M·cm and peak extinction of εm̅ ol = 104/M·cm. Those are typical values for chiral molecules. We also assume that the molecular transition is similar to the low-wavelength transition of the FMN molecules in Figure 3. We therefore take λmol = 380 nm and FWHMmol = 95 nm, where λmol and FWHMmol are the wavelength and the broadening of the molecular transition, respectively. The resulting medium parameters become
Figure 3. (A) Calculated extinction spectra of the Au sphere and the chiral shell taken separately. (B) Calculated CD spectra of the hybrid core−shell structures with parameters: a = 100 nm and b = 120, 140, 200 nm. The CD shows a plasmon-induced tail at the longer wavelengths and also a plasmon peak at ∼700 nm. The complex structure for the molecular CD transition (at ∼380 nm) comes from the plasmonic and electromagnetic interference effects in the core− shell structure.
μ12 2 ⎛ 1 εc = 1 − 4πnc ⎜ 3 ⎝ ℏω − ℏω0 + i Γ12 ⎞ 1 − ⎟ ℏω + ℏω0 + i Γ12 ⎠ μ · m21 ⎛ 1 ξc = 4πnc i 12 ⎜ 3 ⎝ ℏω + ℏω0 + i Γ12 ⎞ 1 + ⎟ ℏω − ℏω0 + i Γ12 ⎠
CD =
NA × 10−4(Cext, + − Cext, −) 0.23
(3)
where NA is the Avogadro number. The CD spectrum describes the chiroptical response. In Figures 3A,B we plot the results of our model for an Au sphere with a radius a = RAu = 100 nm. The radius of the chiral molecular shell, b = Rshell, is varied. We now observe an interesting structure of the CD spectrum of the system. Along with the features expected from molecular resonances at ∼380 nm, we observe in our model a plasmonic CD resonance at ∼650 nm, as in the experiment. Since, in general, the chiral parameter for molecular media is small (in particular, for our parameters, |ξc| ∼ 10−3at λmol = 380 nm), then we expect, in the absence of the optically active plasmonic component, the molecular CD to be simply ∼Im[ξc(ω)]. In the hybrid structure, the molecular CD contribution is essentially modified near the plasmon resonance, since the electromagnetic fields are strongly perturbed by the plasmon in this region. The wavelength dependence of ξc(ω) is therefore very important since the plasmonic CD signal is ∼ξc(ωplasmon), whereas the molecular contribution depends on ξc(ωmolecule). Importantly,
(2)
where μ12 and m21are the electric and magnetic dipolar matrix elements for the molecular transition, ω0 = 2πc0/λmol, and nc is a density of molecules in the film (see Supporting Information for details). Equation 2 only provide approximations for the optical parameters since they ignore interactions between molecular dipoles. However, the general form of eq 2 is valid even for a dense molecular film with strongly interacting molecules. The electromagnetic response of the combined molecule−nanoparticle system is characterized by the extinction cross sections for circularly polarized incident beams, Cext,± and by the extinction CD: 981
dx.doi.org/10.1021/nl204055r | Nano Lett. 2012, 12, 977−983
Nano Letters
Letter
since Im[ξc(ω)] is relatively small at ω ∼ ωplasmon, the plasmoninduced CD comes mostly from the real part of the chiral parameter, Re[ξc(ω)]. This part of the chiral parameter determines the optical rotatory dispersion (ORD) effect for single molecules: The molecular ORD strength ∼ Re[ξc(ω)]. In our structure, the presence of a chiral medium with a nonzero ORD response (Re[ξc(ωplasmon)] ≠ 0) at the plasmonic wavelength makes the plasmon resonances itself chiral and therefore creates the plasmon peaks in the CD spectra for the achiral Au particle. If we compare the results from our simple model to those from the experiments, we see that the calculated CD spectrum is qualitatively similar to the observed spectra. Our model demonstrates the existence of a mechanism by which CD is induced in a resonant plasmonic nanostructure that is qualitatively and quantitatively different from near-field, dipole−dipole interactions.11,12 In our theoretical model, which incorporates both chiral and plasmonic components, the plasmon-induced electromagnetic mechanism of CD comes naturally from solving the Maxwell’s equations written for a chiral medium interacting with a plasmonic structure. There are important differences between the mechanism proposed in this paper and the near-field, dipole−dipole mechanism from refs 11 and 12. For the radiative mechanism introduced here, the electromagnetic CD is proportional to the thickness of a chiral shell and therefore comes from a large volume of chiral molecules which are located at distances of the order of λ from the plasmonic object. In contrast, the CD induced by near-field, dipole−dipole mechanism comes mainly from the molecules located very close to the plasmonic nanoparticle; the interaction extends only over a few nanometers. Moreover, while the near-field, dipole−dipole mechanism gives rise to plasmonic CD which is proportional to the local electric field at the location of a chiral molecule and therefore is only enhanced in local plasmonic hot spots near the surface of the nanostructure.11,12 The radiative coupling effect described here is an “accumulative” effect that comes from the formation of chiral plasmon−polariton modes in a metal structure covered with a chiral material. The radiative coupling mechanism we propose is also consistent with measurements performed on layers of chiral molecules that have very small ORD (i.e., Re[ξe(ωplasmon)] ∼ 0) at the plasmonic frequency, specifically the protein β-lactoglobulin and the amino acid tryptophan. These layers, as expected, induce no measurable CD in the plasmonic resonances of the crosses (see Supporting Information). To illustrate this effect, we have carried out an additional calculation for the plasmonic CD within our spherical chiral shell model (Figure 5S, Supporting Information) taking a variable real parameter ξc. Figure 5S, Supporting Information, shows the result. As expected, the plasmonic peak CD decreases with decreasing the real parameter ξc. This demonstrates that molecules with weak ORD (i.e., with small Reξc) will not give rise to a strong plasmonic CD, while molecules with large Reξc will. We have established this explicitly in experiments using chlorophyllin films, Figure 4, a molecule which has a larger Re[ξc(ωplasmon)] value than FMN at the wavelength of the plasmon resonance.28 Consistent with our model, a chlorophyllin film induces a significantly larger level of plasmonic CD than an equivalent FMN film. This demonstrates the origin of plasmonic CD in our experiments as arising from the ORD properties of the molecule. To summarize we have demonstrated a new phenomenon by which CD can be induced in the plasmonic resonances of
Figure 4. Spectra collected from 7 μg cm−2 chlorophyllin (inset shows chemical structure of chlorophyllin) films deposited on 100 nm thick crosses (black) and a plain glass substrate (red). The cross substrate shows the induction of chirality into the plasmonic resonance (∼670 nm). Chlorophyllin displays a larger level of optical rotation and hence greater Re [ξ(ω)] than FMN.28
achiral metallic nanostructures via the radiative coupling between the plasmon and the surrounding chiral molecules. Such hybrid molecular-metal plasmonic chiral materials offer the opportunity of greater flexibility in both fabrication and application of chiral metamaterials. For instance, the molecular properties of the surrounding chiral molecular media of the hybrid architecture could be engineered to undergo a physical transformation under a stimulus, thus creating a new route to dynamic metamaterials. Methods. Fabrication. The crosses were designed using L-Edit CAD software with a line width of 80 nm and a periodicity of 800 nm. They were arrayed to cover a total area of 6.4 × 6.4 mm2. The PCMs were fabricated on vitrosil quartz slides 25 × 25 × 0.5 mm3. The slides were cleaned for 5 min in acetone and 5 min in isopropanol both under ultrasonic agitation before being blown dry in a stream of nitrogen. A bilayer of poly(methyl methacrylate was spun to a thickness of about 200 nm and baked at 180 °C for 1 h. A 20 nm Al layer was evaporated as a charge conduction layer during the electron beam lithography. The pattern was exposed in a Vistec VB6 UHR EWF lithography tool. After exposing the samples, the Al layer was removed in a wet etch solution, rinsed in water, and dried in a stream of nitrogen before development in isopropyl alcohol (IPA:methyl isobutyl ketone (3:1). Prior to metal deposition, the samples were exposed for 0.5 min at 40 W to an oxygen plasma. Five nm of titanium was used as an adhesion layer and followed by 100 nm of gold. The final patterns were achieved in a lift-off process by leaving the samples in acetone for 48 h.
■
ASSOCIATED CONTENT
S Supporting Information *
Theory and simulations; fabrication details; extinction; and CD spectra. This material is available free of charge via the Internet at http://pubs.acs.org. 982
dx.doi.org/10.1021/nl204055r | Nano Lett. 2012, 12, 977−983
Nano Letters
■
Letter
(28) Kutyurin, V. M.; Grigorovich, V. I. Chem. Nat. Compd. 1967, 3, 44.
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected];
[email protected];
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the EPSRC (E.H.), the MRC (M.K.), and the National Science Foundation (Z.F. and A.O.G.). N.A.A. acknowledges the Iraqi Government for the award of a Ph.D. studentship.
■
REFERENCES
(1) Gansel, J. K.; Thiel, M.; Rill, M. S.; Decker, M.; Bade, K.; Saile, V.; von Freymann, G.; Linden, S.; Wegener, M. Science 2009, 305, 1513. (2) Schwanecke, A. S.; Krasavin, A.; Bagnall, D. M.; Potts, A.; Zayats, A. V.; Zheludev, N. I. Phys. Rev. Lett. 2003, 91, 247404. (3) Kuwata-Gonokami, M.; Saito, N.; Ino, Y.; Kauranen, M.; Jefimovs, K.; Vallius, T.; Turunen, J.; Svirko, Y. Phys. Rev. Lett. 2005, 95, 227401. (4) Pendry, J. B. Science 2004, 306, 1353. (5) Zhang, S.; Park, Y. S.; Li, J. S.; Zhang, W. L.; Zhang, X. Phys. Rev. Lett. 2009, 102, 023901. (6) Hendry, E.; Carpy, T.; Johnston, J.; Popland, M.; Mikhaylovskiy, R.; Lapthorn, A. J.; Kelly, S. M.; Barron, L. D.; Gadegaard, N.; Kadodwala, M. Nat. Nanotechnol. 2010, 5, 783. (7) Gautier, C.; Burgi, T. J. Am. Chem. Soc. 2006, 128, 11079. (8) Kitaev, V. J. Mater. Chem. 2008, 18, 4745. (9) Xia, Y.; Yunlong Zhou, Y.; Tang, Z. Nanoscale 2011, 3, 1374. (10) Govorov, A. O.; Gun’ko, Y. K.; Slocik, J. M.; Gerard, V. A.; Fan, Z.; Naik, R. R. J. Mater Chem. 2011, 21, 169806. (11) Govorov, A. O.; Fan, Z. Y.; Hernadez, P.; Slocik, J. M.; Naik, R. R. Nano Lett. 2010, 10, 1374. (12) Govorov, A. O. J. Phys. Chem. C 2011, 115, 7915. (13) Slocik, M.; Govorov, A. O.; Naik, R. R. Nano Lett. 2011, 11, 701. (14) Gérard, V. A.; Gun’ko, Y. K.; Defrancq, E.; Govorov, A. O. Chem. Commun. 2011, 47, 7383. (15) Jackson, J. D. Classical Electrodynamics, 3rd ed.; Wiley: New York, 1998. (16) Sugawara, Y.; Kelf, T. A.; Baumberg, J. J.; Abdelsalam, M. E.; Bartlett, P. N. Phys. Rev. Lett. 2006, 97, 266808. (17) Tollin, G.; Edmondon, D. E. Biochemistry 1971, 10, 113. (18) Lieberman, I.; Shemer, G.; Fried, T.; Kosower, E. M.; Markovich, G. Ange. Chem. Int. Ed. 2008, 47, 4833. (19) Goldsmith, M. R.; George, C. D.; Zuber, G.; Naaman, R.; Waldeck, D. H.; Wipf, P.; Beratan, D. N. Phys. Chem. Chem. Phys. 2006, 8, 63. (20) Lidzey, D. G.; Bradley, D. D. C.; Skolnick, M. S.; Virgili Walker, S; Whittaker, D. M. Nature 1998, 395, 53. (21) Ni, W.; Yang, Z.; Chen, H.; Li, L.; Wang, J. J. Am. Chem. Soc. 2009, 130, 6692. (22) Kong, J. A. Theory of Electromagnetic Waves; Wiley: New York, 1975. (23) Post, E. J. Formal Structure of Electromagnetics; North-Holland: Amsterdam, The Netherlands, 1962. (24) Palik, E. D. Handbook of Optical Constants of Solids; Academic Press: New York, 1985. (25) Bohren, C. F. Absorption and Scattering of Light by Small Particles; John Wiley and Sons: Weinheim, Germany, 2004. (26) Bohren, C. F. Chem. Phys. Lett. 1974, 29, 458. (27) Numerical simulations of electromagnetic fields were performed using a commercial finite-element package (Ansoft HFSS, version 11.0) with a mesh size of 4.0 nm. Permittivity values for gold were taken from ref 24. 983
dx.doi.org/10.1021/nl204055r | Nano Lett. 2012, 12, 977−983
