1656 OPTICS LETTERS / Vol. 35, No. 10 / May 15, 2010

Magnetic metamaterials in the blue range usingaluminum nanostructures

Yogesh Jeyaram,1 Shankar K. Jha,1 Mario Agio,2 Jörg F. Löffler,1 and Yasin Ekinci1,3,*1Laboratory of Metal Physics and Technology, Department of Materials, ETH Zurich, 8093 Zurich, Switzerland

2Nano Optics Group, Laboratory of Physical Chemistry, ETH Zurich, 8093 Zurich, Switzerland3Paul Scherrer Institute, 5232 Villigen-PSI, Switzerland

*Corresponding author: yasin.ekinci@mat.ethz.ch

Received November 24, 2009; revised February 17, 2010; accepted March 21, 2010;posted April 15, 2010 (Doc. ID 120413); published May 7, 2010

We report an experimental and theoretical study of the optical properties of two-dimensional arrays of alu-minum nanoparticle in-tandem pairs. Plasmon resonances and effective optical constants of these structuresare investigated, and strong magnetic response as well as negative permeability is observed down to 400 nmwavelength. Theoretical calculations based on the finite-difference time-domain method are performed forvarious particle dimensions and lattice parameters, and are found to be in good agreement with the experi-mental findings. The results show that metamaterials operating across the whole visible wavelength rangeare feasible. © 2010 Optical Society of America

OCIS codes: 160.4670, 160.3918, 160.3900, 240.6680, 310.6628.

In the past decade, a new class of optical materialsknown as metamaterials has emerged and attractedsignificant interest. These are material structuresengineered at the subwavelength scale to exhibitnovel optical properties such as negative refractiveindex or magnetic activity at high frequencies [1]. Af-ter the first experimental demonstration in themicrowave regime, the operational frequency ofmetamaterials has experienced tremendous progresswithin a decade, with novel designs and nanofabrica-tion techniques [2]. In particular, metal/dielectric/metal multilayers, such as cut-wire nanopairs, fish-net nanostructures, and nanoparticle pairs, haveresulted in magnetic and negative-refractivemetamaterials operating at optical frequencies downto 580 nm [3–6]. These structures exhibit strongantisymmetric eigenmodes, which are responsible forthe effective magnetic response [7–10].

Despite the fact that several problems, such as re-ducing the losses, broadening the operational band-width, and extending the nanostructure into thethird dimension, need to be solved, the realization ofmetamaterials operating at shorter visible wave-lengths is of central interest toward applications inbiosensing and imaging. In these aforementioned andmany other works demonstrating metamaterialswith negative refractive index or negative permeabil-ity mostly Au and Ag were used, which are, in fact,the conventional choice in plasmonics. On the otherhand, Al is also a good optical material because of itslow absorption and large real part of the dielectricconstant. While Au and Ag exhibit interband absorp-tions below wavelengths of about 590 nm and350 nm, Al has low absorption down to 200 nm be-cause of its free-electron-like character and high bulkplasmon frequency [11]. These properties have al-ready made Al an ideal candidate for plasmonic ap-plications at short wavelengths [12–15].

In this Letter we study the plasmonic properties ofarrays of Al in-tandem particle pairs and find that or-

dered two-dimensional (2D) arrays of Al/Al2O3/Al

0146-9592/10/101656-3/$15.00 ©

nanoparticles exhibit strong magnetic resonances asa result of the near-field coupling in the in-tandempair. Furthermore, we demonstrate that it is possibleto tune these resonances down to 400 nm and obtaineven negative permeability in the whole visiblerange.

2D arrays of Al nanoparticle pairs on quartz sub-strates were fabricated using extreme-UV interfer-ence lithography (EUV-IL) and sequential depositionof Al, Al2O3, and Al with a subsequent lift-off process.EUV-IL provides high-resolution structures overlarge areas with perfect periodicity [14,16]. Moreover,there is no need for a conduction layer such as in-dium tin oxide, which is used in e-beam lithography.Figure 1(a) shows a top-down scanning electron mi-croscope (SEM) image of a typical sample. Figure 1(b)is another SEM image taken at an oblique angle,demonstrating the sandwich-like geometry of the fab-ricated in-tandem nanoparticle pairs, whose sche-matic cross section is shown in Fig. 1(c). The size

Fig. 1. (Color online) (a) SEM image of one of the fabri-cated samples (diameter 2r=120 nm, period a=200 nm).(b) SEM image at an oblique angle with details on the ver-tical profile of one sample �2r=105 nm�. (c) Model used forthe FDTD calculations: h=21 nm; d=24 nm; a=200 nm; �=20°; nAl2O3

=1.77; nquartz=1.46; nAl is from [11]; 2r=135,120, 105, 87, 82, 72 nm; a 3-nm-thick Al2O3 coating layer is

also taken into account.

2010 Optical Society of America

May 15, 2010 / Vol. 35, No. 10 / OPTICS LETTERS 1657

of the fabricated structures varied from 2r=72 nm to 135 nm (in base diameter) with a fixed pe-riod of a=200 nm. The thicknesses of the Al andAl2O3 layers were 21 nm and 24 nm, respectively.The area of an array was 400 �m�400 �m, simplify-ing optical characterization substantially.

Transmission spectra of the structures were mea-sured with a homemade microspectroscopy setupthat enables spectroscopy in the visible and UVranges. The reflection spectra were measured withthe same homemade setup in the UV range and witha standard optical microscope in the visible range.Since the structures are isotropic on the substrateplane, no special polarization optics was used. Fig-ures 2(a) and 2(b) show the transmission and reflec-tion spectra of the samples with different diameters.The transmission spectra exhibit two dips associatedwith hybridized modes resulting from the dipolarcoupling between the nanoparticles, in addition to ahybridized quadrupolar mode at about 250 nm, dis-cussion of which is beyond the scope of this Letter.The resonance at shorter wavelengths can be identi-fied as a symmetric resonance or electric dipolar reso-nance, while the one at longer wavelengths is attrib-uted to the antisymmetric or magnetic dipolarresonance [8]. These are also evident in the reflectionspectra [see Fig. 2(b)], where the lineshapes appearasymmetric because of a Fano-type interference be-tween the eigenmodes [10]. The antisymmetric modearises as a result of the induced dipole vectors in thetwo metal layers being opposite to each other. Thiscreates a magnetic dipole moment that couples to themagnetic component of the incident field and thusgives rise to an effective permeability diverging fromunity. As seen in Fig. 2(b) the magnetic resonance canbe easily tuned by varying the particle diameter forconstant metal and dielectric thickness.

Finite-difference time-domain (FDTD) simulationswere performed using the schematic model shown in

Fig. 2. (Color online) Transmission [(a) and (c)] and reflec-tion [(b) and (d)] spectra for samples with different disk di-ameters 2r. (a) and (b) experimental, (c) and (d) theoreticalFDTD results. Note that the reflection spectra in (b) weremeasured with two different setups (200–400 nm and400–700 nm). A small range of the spectra (shaded regionbetween 380–420 nm) is not shown, because the resultswere not reproducible owing to a low signal-to-noise ratio

in both setups in this region.

Fig. 1(c), which represents relatively simple but real-istic structural parameters of the fabricated struc-tures. A natural Al2O3 layer of 3 nm is assumed tocover the nanoparticles at the Al–air and Al–substrate interfaces, which is known to be formed im-mediately when Al is exposed to air or oxygen-richsubstrates [17]. As to the dielectric constants of Aland Al2O3 we used values reported in the literature[11]. Periodic boundary conditions were applied to ac-count for the effects of the array on optical response.Figure 2 shows good agreement between the experi-mental and the calculated spectra, especially for thesamples with larger nanoparticles. The small varia-tions between the theoretical and experimental re-sults are due to discrepancies between the exact ge-ometry and optical constants of the modeled andfabricated structures. Moreover, the size and theshape dispersion of the samples resulting from the fi-nite grain size of deposited Al could not be taken intoaccount in the theory.

Since the present structures exhibit large magneticresonances we can expect a strong modulation of theeffective magnetic permeability. We extracted the ef-fective optical constants from the calculated complexreflectance and transmittance coefficients using es-tablished algorithms [18,19], although these shouldbe taken with caution at the present scales. Figure 3displays the complex relative magnetic permeability� for selected samples. Different nanodisk radii rprovide magnetic resonances covering the visible andthe UV spectral ranges with strong variations ofRe��� between 0.5 and 2 and without a significantchange in the figure of merit �Re��� / Im����. These re-sults show how the amplitude and the wavelength ofthe effective permeability of the present metamate-rial structures can be easily tuned.

Having demonstrated the existence of magneticresonances in the blue range, we performed extensiveFDTD simulations to optimize the structural param-eters in order to study the tunability and the ampli-tude of the magnetic response. Particular interestwas paid to achieving negative permeability at shortwavelengths within the structural parameters fea-sible with standard lithographic processes. Transmis-sion spectra and the extracted � [Im��� not shown]for some representative cases are shown in Fig. 4. Astrong magnetic response is obtained at around500 nm, with � as large as 7. A negative permeabilityat about 450 nm is achieved, notwithstanding itsrelatively low figure of merit. As seen in the figure,in changing the lattice parameter from

Fig. 3. (Color online) (a) Real and (b) imaginary part of therelative magnetic permeability � extracted from FDTD

simulations for the fabricated samples.

1658 OPTICS LETTERS / Vol. 35, No. 10 / May 15, 2010

160 nm to 140 nm, the magnetic resonance of thesame nanoparticle pair exhibits a redshift. Similarto electric dipole interaction in nanoparticle arrays[20], this trend might evidence the existence of mag-netic dipole couplings [21], which deserve a detailedinvestigation in future studies.

In summary, we demonstrated that magneticmetamaterials can be obtained down to the bluewavelength range with 2D arrays of Al in-tandemnanoparticle pairs. Furthermore, negative perme-ability is shown to be feasible, paving the way to-wards negative refractive-index metamaterials in thewhole visible range. Even though the figure of meritis relatively low, the promise of the present workshould invoke further studies with designs like nano-hole arrays or fishnet structures [22], which may leadto negative permeabilities with a higher figure ofmerit and even a negative refractive index. Moreover,since the upscaling to longer wavelengths is rathertrivial, the use of Al could also relax the design andfabrication constraints faced by Ag-based metamate-rials [6]. Besides the envisioned exciting applicationsof optical metamaterials [1], the record wavelengthsachievable with Al nanostructures represent an im-portant test bed for exploring the fundamental limitsof high-frequency magnetism [23] as well as the ap-plicability of an effective medium theory in mesos-copic electromagnetic systems.

This work was supported by ETH Research GrantTH-29/07-3. M. A. thanks Vahid Sandoghdar for con-tinuous support and encouragement. Part of thiswork was performed at the Swiss Light Source (SLS),

Fig. 4. (Color online) (a) Transmission spectra and (b) realpart of the relative magnetic permeability � extracted fromFDTD simulations for the structural parameters providedin the legend. The oxide coating has been neglected.

Paul Scherrer Institute, Switzerland.

References

1. V. M. Shalaev, Nat. Photonics 1, 41 (2007).2. S. Linden, M. Wegener, and C. M. Soukoulis, Science

315, 47 (2007).3. H. Schweizer, L. Fu, H. Gräbeldinger, H. Guo, N. Liu,

S. Kaiser, and H. Giessen, Phys. Status Solidi A 204,3886 (2007).

4. U. K. Chettiar, A. V. Kildishev, H.-K. Yuan, W. Cai, S.Xiao, V. P. Drachev, and V. M. Shalaev, Opt. Lett. 32,1671 (2007).

5. H.-K. Yuan, U. K. Chettiar, W. Cai, A. V. Kildishev, A.Boltasseva, V. P. Drachev, and V. M. Shalaev, Opt. Ex-press 15, 1076 (2007).

6. S. Xiao, U. K. Chettiar, A. V. Kildishev, V. P. Drachev,and V. M. Shalaev, Opt. Lett. 34, 3478 (2009).

7. Z. Huang, J. Xue, Y. Hou, J. Chu, and D. H. Zhang,Phys. Rev. B 74, 193105 (2006).

8. T. Pakizeh, M. S. Abrishamian, N. Granpayeh, A. Dmi-triev, and M. Käll, Opt. Express 14, 8240 (2006).

9. W. Cai, U. K. Chettiar, H.-K. Yuan, V. C. de Silva, A. V.Kildishev, V. P. Drachev, and V. M. Shalaev, Opt. Ex-press 15, 3333 (2007).

10. Y. Ekinci, A. Christ, M. Agio, O. J. F. Martin, H. H. So-lak, and J. F. Löffler, Opt. Express 16, 13287 (2008).

11. E. D. Palik and G. Ghosh, eds., Handbook of OpticalConstants of Solids (Academic, 1998).

12. Y. Ekinci, H. H. Solak, and C. David, Opt. Lett. 32, 172(2007).

13. K. Ray, M. H. Chowdhury, and J. R. Lakowicz, Anal.Chem. 79, 6480 (2007).

14. Y. Ekinci, H. H. Solak, and J. F. Löffler, J. Appl. Phys.104, 083107 (2008).

15. A. Mohammadi, V. Sandoghdar, and M. Agio, J. Com-put. Theor. Nanosci. 6, 2024 (2009).

16. H. H. Solak, Y. Ekinci, P. Käser, and S. Park, J. Vac.Sci. Technol. B 25, 91 (2007).

17. V. Y. Gertsman and Q. S. Kwok, Microsc. Microanal. 11,410 (2005).

18. D. R. Smith, S. Schultz, P. Markoš, and C. M. Soukou-lis, Phys. Rev. B 65, 195104 (2002).

19. V. M. Shalaev, W. Cai, U. K. Chettiar, H.-K. Yuan, A. K.Sarychev, V. P. Drachev, and A. V. Kildishev, Opt. Lett.30, 3356 (2005).

20. B. Lamprecht, G. Schider, R. T. Lechner, H. Ditlbacher,J. R. Krenn, A. Leitner, and F. R. Aussenegg, Phys. Rev.Lett. 84, 4721 (2000).

21. I. Sersic, M. Frimmer, E. Verhagen, and A. F. Koen-derink, Phys. Rev. Lett. 103, 213902 (2009).

22. G. Dolling, M. Wegener, C. M. Soukoulis, and S. Lin-den, Opt. Express 15, 11536 (2007).

23. R. Merlin, Proc. Natl. Acad. Sci. USA 106, 1693 (2009).