Design and top-down fabrication of metallic L-shape gap nanoantennas supporting plasmon-polariton modes

Microelectronic Engineering 111 (2013) 91–95

Contents lists available at SciVerse ScienceDirect

Microelectronic Engineering journal homepage: www.elsevier.com/locate/mee

Design and top-down fabrication of metallic L-shape gap nanoantennas supporting plasmon-polariton modes S. Panaro a,b,⇑, A. Toma a, R. Proietti Zaccaria a, M. Chirumamilla a,b, A. Saeed a,b, L. Razzari a, G. Das a, C. Liberale a, F. De Angelis a, E. Di Fabrizio a,c a b c

Nanostructures – Istituto Italiano di Tecnologia, Via Morego 30, Genova I-16163, Italy Università degli Studi di Genova, Genova 16145, Italy Lab. BIONEM, Dipartimento di Medicina Sperimentale e Clinica, Università Degli Studi ‘‘Magna Graecia’’ di Catanzaro, Viale Europa, loc. Germaneto, I-88100 Catanzaro, Italy

a r t i c l e

i n f o

Article history: Available online 16 February 2013 Keywords: L-shape nanoantenna Electron beam lithography Annealing Plasmon-polariton modes Zero-field spot

a b s t r a c t In this work the design, fabrication and optical characterization of a polarization-sensitive L-shape nanoantenna device are reported. Such configuration supports plasmon-polariton modes that are combinations of in-phase and out-of-phase single antenna long-axis surface plasma oscillations. In the former case charges distributions induce in the gap region an intense hot spot while in the latter one a ‘‘zero-field spot’’ occurs in a plasmonic mode which can be referred to a non-zero dipolar momentum. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction Plasmonic nanoantennas have led to huge innovations both for the capability to efficiently interact with free propagating radiation and their property to confine electromagnetic energy inside of subwavelength regions called hot-spots [1–5]. The ability of antenna to enhance the electromagnetic field beyond the diffraction limit leads to several implications and advantages that can find applications in many branches of technology like photoconvertion [6,7], metamaterial engineering [8–10], non-linear optics [11,12], bio-sensing [13–15] and the really challenging single-molecule spectroscopy [16,17]. In particular, antenna dimers [18,19] are optimal systems for the fabrication of plasmon-assisted Raman sensors. Tailoring the antenna morphological parameters it is possible to finely tune the dipolar surface plasmonic resonance (SPR) wavelength around the spectral region of interest employed for enhanced Raman spectroscopy. Moreover the interspacing volume or gap between the apexes of the two antenna arms, under resonance conditions, can experience both strong field amplification and extreme energy localization. If the molecules of interest are placed inside the gap, the plasmonic enhancement effect can make the Raman signal easily detectable, thus showing the spectral fingerprints of the deposited molecules [20,21]. In this work the fabrication, by means of advanced lithographic techniques, and the characterization of L-shaped plasmonic anten⇑ Corresponding author at: Nanostructures – Istituto Italiano di Tecnologia, Via Morego 30, Genova I-16163, Italy. Tel.: +39 010 71781 962; fax: +39 010 720321. E-mail address: [email protected] (S. Panaro). 0167-9317/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.mee.2013.02.014

na arrays are reported and explained in details. L-shape antenna dimer is a system which exploits the combination of single arm long axis plasmon-polariton modes. Single antenna nanostructures show dichroic behavior in the far-field spectrum, according to the polarization of the light impinging on them. In fact, because of the size-dependency of SPR [22–24], the antenna far-field spectrum shows a blue-shift of the resonance peak when the light polarization is rotated from a configuration parallel to the structure long axis to a configuration aligned to the short axis. Differently from what happens to the single antenna case, the L-shape nanostructure presents far-field dichroism which can be ascribed to the in-phase and out-of-phase combination of long axis plasmonic modes induced by different light polarization states. Considering the direct implication of this phenomenon in the electric field distribution, by simply rotating light polarization, it is possible to move from an intense hot-spot configuration to an almost ‘‘zerofield spot’’ in the gap region. In the perspective of improving the Raman signal-to-noise ratio, these results are quite promising and suggest a very flexible SERS functionalized device.

2. Experimental procedure The top-down fabrication technique adopted to produce Lshape gap nanoantennas was Electron Beam Lithography (EBL). Since the fabrication of strongly coupled nano-systems (i.e. with inter-particle distance in the 10 nm range) with large accuracy and homogeneity is an essential requirement for the production of a Raman-active biosensor, such technique appears as one of

92

S. Panaro et al. / Microelectronic Engineering 111 (2013) 91–95

Fig. 1. (a) 3-D sketch of L-shape antennas showing the geometrical parameters representative of the morphology. (b and c), respectively current density and electric field intensity 2-D plots for normal-incidence light polarized at 45° valued at k = 810 nm. (d) Theoretical extinction efficiency spectra of L-shape antennas (L = 190 nm, W = 110 nm, H = 60 nm and G = 20 nm) and of their single antenna arm; in black dashed line the long axis reference extinction spectrum of single antenna, in green dotted/ dashed line the short axis reference extinction spectrum of single antenna, in red double-dotted/dashed line the L-shape antenna extinction spectrum for 45° polarization and in blue continuous line the L-shape antenna extinction spectrum for +45° polarization. (e and f), respectively current density and electric field intensity 2-D plots for normal-incidence light polarized at +45° valued at k = 760 nm. (For interpretation of color in this figure, the reader is referred to the web version of this article).

the optimal choices both for the level of structure shape accuracy and of inter-particle separation that it guarantees. The first step in the fabrication process consists in the nanostructure design. Dimer configuration, whose schematic is shown in Fig. 1(a), is defined by the geometrical parameters length L, width W, height H and inter-particle gap G. Since dimers are fabricated in arrays, another significant parameter to be defined consists in the inter-structure spacing S. Different matrices of L-shape antennas have been produced changing the inter-particle distance G. The substrate on which the nanostructures have been fabricated is CaF2 (1 0 0), particularly addressed in literature for its high optical transparency in the visible and near-infrared (NIR) regions. The EBL nanopatterning procedure involved several steps; substrate-cleaning in an ultrasonic bath of acetone and oxygen plasma exposure have been firstly carried out. Hence PolyMethylMethacrylate (PMMA) electronic resist has been spin-coated on the substrate. After the spin-coat, annealing has been done at 180 °C for 7 min in order to obtain a uniform film. To prevent surface charging effects during the electron exposure, 10 nm Al layer has been thermally deposited on the PMMA surface. Therefore EBL machine (electron energy 20 keV and beam current 45 pA), equipped with a pattern generator (Raith 150-two), has been used for the nanostructure patterning. After the Al removal in a KOH solution, the exposed resist was developed in a conventional solution of MIBK/ isopropanol (IPA) (1:3) for 30 s. Physical Vapour Deposition (evaporation rate 0.3 Å/s), respectively of 5 nm Ti as adhesion layer and 55 nm Au has been performed on the sample. Finally, the unexposed resist was removed with acetone and rinsed out in IPA. O2 plasma ashing at 200 W for 60 s was used to remove residual photoresist and organic contaminants for an improved lift-off. An annealing cycle at 200 °C with a duration of 15 min has been performed on the sample in order to investigate the effect of grain boundaries on the electron relaxation rate [25] and therefore on the optical behavior of L-shape nanoantennas. There is evidence in literature according to which Au deposited on a substrate at room temperature tends to aggregate in grains at a metastable energetic level. By increasing the temperature the internal dynamic of the system can be sped up, thus promoting a polycrystalline grain

growth and grain boundary migration [26]. This process has been demonstrated to clearly affect the electron scattering rate which in turns will induce a blue shift of the plasmon resonance [25]. During the optical characterization, the sample has been illuminated at normal incidence with a linearly polarized visible-NIR (DH-2000-BAL, Ocean Optics) light spot of 30 lm diameter performing far-field transmission spectroscopy (HR4000, Ocean Optics) for different light polarizations. The morphological characterization has been carried out by means of a Dual Beam (SEM-FIB) – FEI Helios NanoLab working at 10 keV beam energy and 0.14 nA electron current. Images have been acquired at normal and 30° to normal tilt angle. 3. Calculations The optical and morphological characterizations have been supported by simulative tools, in order to understand the charge distributions and the modes excited inside of the structures responsible for the observed far-field resonances. Simulations at finite elements have been performed by means of commercial software (CST Studio Suite 2010) which solves Maxwell’s equations in the discrete frequency domain. In order to take under consideration the inter-band transitions of the metals in the visible region, dispersion relations of Au and Ti were loaded in the software in accordance to the model of Lorentz–Drude [27]. The L-shaped nanostructures were assumed to be embedded completely in an effective medium with dielectric constant eeff ¼ ð1 þ n2 Þ=2, where n = 1.43 is the refractive index of the CaF2 substrate [28,29]. Periodic boundary conditions are used to simulate the response of an array of nanoantennas with a spacing of 100 nm in both directions on the plane. From the simulation it has been possible to extract the extinction efficiency Qext defined as the ratio of the extinction cross section rext of the structure to its geometric cross section rgeo and related to the relative transmission measured by:

Q ext ¼

rext Að1  Trel Þ ¼ Na rgeo

ð1Þ

S. Panaro et al. / Microelectronic Engineering 111 (2013) 91–95

93

Fig. 2. ((a) and (b)) Representative SEM images, respectively of 30° out-of-normal tilted L-shape antenna arrays and single L-shape antenna fabricated by EBL technique before annealing process (scale bars corresponding, respectively to 400 and 100 nm). The nanostructures geometrical parameters are L = 190 nm, W = 110 nm, H = 60 nm, G = 20 nm and S = 100 nm. (c) Representative SEM image of 30° out-of-normal tilted L-shape antenna after 200 °C annealing (scale bar corresponding to 100 nm). (d) Transmission optical spectra of the L-shape antenna arrays before (black curve with triangles) and after annealing at 200 °C (red curve with circles) and theoretical extinction efficiency (continuous red curve) for normal-incidence light polarized at 45°. (e) Transmission optical spectra of the L-shape antenna arrays before (black curve with triangles) and after annealing at 200 °C (blue curve with circles) and theoretical extinction efficiency (continuous blue curve) for normal-incidence light polarized at +45°. (For interpretation of color in this figure, the reader is referred to the web version of this article).

where ‘‘A’’ is the area illuminated by the light spot, ‘‘N’’ is the number of array elements inside the spot and ‘‘a’’ is the area of L-shaped antenna projection on the plane where the polarization vector lies. Near-field analysis has been focused on electric field intensity and current density distribution collected on a plane normal to light incidence direction and passing through the center of nanostructures.

4. Results and discussion L-shape gap nanoantennas present a clear polarization-dependency due to the excitement inside of their arms of in-phase and out-of-phase plasmon-polariton resonances. Drawing a parallel to molecular orbital theory, such two plasmonic states could be considered, respectively as bonding and anti-bonding modes resulting from the strong coupling between the two unperturbed single antenna SPRs [1,30]. As it can be well appreciated from the near-field analysis reported in Fig. 1(b) (incoming white light polarized at 45° with respect to the horizontal arm) an intense in-phase charge oscillation mode (i.e. bonding condition) is established inside of the two antenna arms when polarization is parallel to the line that links the two near apexes. Moreover it is possible to appreciate within the inter-particle gap the consequent lighting of an intense hot spot with a corresponding field enhancement factor equal to 20 (Fig. 1(c)). Such light polarization state induces, around a wavelength of 810 nm (Fig. 1(d)), a hot spot intensity comparable to nanosphere or nanoshell dimers values in extreme gap-coupling conditions (2–3 nm) [31]. A result like this is very interesting if it is considered that the G parameter of simulated L-shape dimer is 20 nm. For symmetry reasons, charges inside of the arms can also oscillate out-of-phase (i.e. anti-bonding mode). If light polarization is rotated to +45°, with the same angle convention, indeed a dramatic change in the electric field distribution occurs. Around the resonance wavelength of 760 nm (Fig. 1(d)), the charges start to oscil-

late out-of-phase contemporarily converging towards the gap (Fig. 1(e)) and thus creating a region of almost zero electric field intensity inside of the gap (Fig. 1(f)). This behavior, complementarily to the preceding phenomenon of field enhancement, guarantees an optimal polarization configuration in order to collect precise Raman background signals. For a more complete picture, simulated extinction spectra referring to single antenna arm are reported in Fig. 1(d) showing how the plasmonic coupling considered in this work exclusively involves long axis modes and not short axis resonances which conversely occur in a different spectral portion. On the basis of the present results, L-shape antennas have been fabricated by EBL technique with morphological parameters reported in the caption related to Fig. 2(a and b). In Fig. 2(a), a SEM image shows a cut-out of the fabricated nanostructures representative of the global periodic array, while in Fig. 2(b and c) closeup SEM images of the single configuration (referring to the preand post-annealing case, respectively) are reported in order to appreciate the inter-particle gap. From the optical characterization the transmittance spectra collected for incident light at 45° and +45° polarization angles are reported in Fig. 2(d and e) (black curves). The two curves show the expected excitation of ±45° polarization plasmon-polariton modes resonating, respectively at 850 and 820 nm. In the same plots the correspondent theoretical extinction spectra are reported. Looking at the relative positions of the simulated (red and blue continuous curves) and measured resonance wavelengths (black curves with open triangles), a red-shift of about 60 nm of the measured resonances can be noticed. This fact is well explained if the polycrystalline nature and the surface roughness of the fabricated system are taken under consideration. As exhaustively investigated by K.-P. Chen et al. in Ref. [25], grain boundary effect and consequent electronic multi-scattering contribute both to increase the damping of antenna resonance and to red-shift its peak from the expected homogeneous antenna optical response. After a cycle of annealing on the sample, intended to promote the gold re-organi-

94

S. Panaro et al. / Microelectronic Engineering 111 (2013) 91–95

mon-polariton modes dependent effects.

and not from single antenna size-

5. Conclusions In conclusion, this article has shown the fabrication and the farfield optical analysis of plasmonic L-shape nanoantenna configurations supporting combinations of plasmon-polariton charge oscillation modes strongly polarization-dependent with tunable optical response in visible-NIR region. Their high near-field enhancement within the inter-particle gap, the chance to tune the plasmonic resonance and the electric field intensity just by rotating light polarization and the low-energy wavelengths (around 800 nm) involved make them very interesting as SERS focused devices for biological detection.

Acknowledgments Fig. 3. (a) Measured extinction efficiency intensities of L-shape antenna arrays post-annealing for normal-incidence light polarized at +45°, at different G parameters. ((b)–(d)) Normal angle SEM images showing gap region of post-annealing Lshape antennas for G parameter corresponding, respectively to 20, 10 and 5 nm (scale bar corresponding to 50 nm).

The authors gratefully acknowledge support from European Projects SMD FP7 No. CP-FP 229375-2; Nanoantenna FP7 No. 241818; FOCUS FP7 No. 270483.

References zation, grains tend to melt together reducing the damping effect on the charges. In fact Au grains with (1 1 1) orientation are energetically more favorable with respect to grains differently oriented and, as shown in [25], by heating the system, the former grains tend to grow at the expense of the other ones. This internal reordering leads to a decreasing of electronic damping. This phenomenon is clearly observed in the post-annealing transmission spectra reported in Fig. 2(d and e) (red and blue curves with open circles), which show a shrinking of the resonance peaks and a blue-shift with respect to the spectra collected before the annealing. The post-annealing spectra, corresponding to structures less affected by grain boundary multi-scattering, show a good accordance with theoretical spectra based on the model system. This result confirms the annealing as an effective process in perspective of rendering polycrystalline Au optical response comparable to model bulk. Finally, as stressed in [26], no evident changes in the shape morphology (i.e. edges rounding) of nanostructures are induced by annealing process until 400 °C as it appears by comparing the morphology of L-shape dimers before and after 200 °C annealing (Fig. 2(b and c)). The high enhancement factor expected in near-field simulation analysis for L-shape antennas with G parameter equal to 20 nm and the good accordance of fabricated structures transmittances with calculated far-field extinction spectra suggest the achievement of a good control over the system parameters. Moreover, in perspective of Raman-functionalized improvement on L-shape dimers, such results suggest the reduction of the inter-particle gap in order to increase the hot spot intensity. Capacitive coupling indeed is strongly dependent on inter-particle distances because of the nature of Coulomb interaction [32]. For such reason extinction spectra have been collected at +45° polarization angle on three Lshape antenna matrixes with decreasing gaps and their intensities are reported in Fig. 3(a). If the resonance peak under analysis was simply ascribed to the rising of a single antenna short-axis mode, the intensity of the peak should not depend on G parameter. To the contrary, looking at the plot in Fig. 3(a), it is possible to appreciate a dependency of the resonance peak on the inter-particle distance. Moving from a 20 nm to a 5 nm gap (Fig. 3(b and d)), the extinction intensity shows an increment, confirming that the presented resonances originate from the coupling of long axis plas-

[1] F. Neubrech, D. Weber, J. Katzmann, C. Huck, A. Toma, E. Di Fabrizio, A. Pucci, T. Härtling, ACS Nano 6 (2012) 7326–7332. [2] L. Razzari, A. Toma, M. Shalaby, M. Clerici, R. Proietti Zaccaria, C. Liberale, S. Marras, I.A.I. Al-Naib, G. Das, F. De Angelis, M. Peccianti, A. Falqui, T. Ozaki, R. Morandotti, E. Di Fabrizio, Opt. Express 19 (2011) 26088–26094. [3] R. Proietti Zaccaria, A. Alabastri, F. De Angelis, G. Das, C. Liberale, A. Toma, A. Giugni, L. Razzari, M. Malerba, H.B. Sun, E. Di Fabrizio, Phys. Rev. B 86 (2012) 035410. [4] G. Volpe, S. Cherukulappurath, R.J. Parramon, G. Molina-Terriza, R. Quidant, Nano. Lett. 9 (10) (2009) 3608–3611. [5] J. Ye, F. Wen, H. Sobhani, J.B. Lassiter, P. Van Dorpe, P. Nordlander, N.J. Halas, Nano Lett. 12 (2012) 1660–1667. [6] H.A. Atwater, A. Polman, Nat. Mater. 9 (2010) 205–213. [7] S. Pillai, M.A. Green, Sol. Energ. Mat. Sol. Cells 94 (2010) 1481–1486. [8] W. Dickson, G.A. Wurtz, P. Evans, D. O’Connor, R. Atkinson, R. Pollard, A.V. Zayats, Phys. Rev. B 76 (2007) 115411. [9] S. Zhang, D.A. Genov, Y. Wang, M. Liu, X. Zhang, Phys. Rev. Lett. 101 (2008) 047401. [10] Y. Wang, T. Sun, T. Paudel, Y. Zhang, Z. Ren, K. Kempa, Nano Lett. 12 (2012) 440–445. [11] A. Belardini, M.C. Larciprete, M. Centini, E. Fazio, C. Sibilia, D. Chiappe, C. Martella, A. Toma, M. Giordano, F. Buatier de Mongeot, Phys. Rev. Lett. 107 (2011) 257401. [12] A.M. Malvezzi, G. Vecchi, M. Patrini, E. Di Fabrizio, Phys. Rev. B 68 (2003) 161306. [13] S. Aksu, A.A. Yanik, R. Adato, A. Artar, M. Huang, H. Altug, Nano Lett. 10 (2010) 2511–2518. [14] J.N. Anker, W.P. Hall, O. Lyandres, N.C. Shah, J. Zhao, R.P. Van Duyne, Nat. Mater. 7 (2008) 442–453. [15] H. Im, N.J. Wittenberg, A. Lesuffleur, N.C. Lindquist, S.-H. Oh, Chem. Sci. 1 (2010) 688–696. [16] A. Toma, G. Das, M. Chirumamilla, A. Saeed, R. Proietti Zaccaria, L. Razzari, M. Leoncini, C. Liberale, F. De Angelis, E. Di Fabrizio, Microelectron. Eng. 98 (2012) 424–427. [17] F. De Angelis, R. Proietti Zaccaria, M. Francardi, C. Liberale, E. Di Fabrizio, Opt. Express 19 (2011) 22268–22279. [18] S. Sheikholeslami, Y.-W. Jun, P.K. Jain, A.P. Alivisatos, Nano Lett. 10 (2010) 2655–2660. [19] L. Shao, K.C. Woo, H. Chen, Z. Jin, J. Wang, ACS Nano 4 (6) (2010) 3053–3062. [20] B. Fazio, C. D’Andrea, F. Bonaccorso, A. Irrera, G. Calogero, C. Vasi, P.G. Gucciardi, M. Allegrini, A. Toma, D. Chiappe, C. Martella, F. Buatier De Mongeot, ACS Nano 5 (2011) 5945–5956. [21] C. Farcau, S. Astilean, J. Phys. Chem. C 114 (2010) 11717–11722. [22] P. Bharadwaj, B. Deutsch, L. Novotny, Adv. Opt. Photon. 1 (2009) 438–483. [23] J. Song, R. Proietti Zaccaria, G. Dong, E. Di Fabrizio, M.B. Yu, G.Q. Lo, Opt. Express 19 (2011) 25206–25221. [24] G.W. Bryant, F.J. García de Abajo, J. Aizpurua, Nano Lett. 8 (2008) 631–636. [25] K.-P. Chen, V.P. Drachev, J.D. Borneman, A.V. Kildishev, V.M. Shalaev, Nano Lett. 10 (2010) 916–922. [26] M.J. Rost, D.A. Quist, J.W.M. Frenken, Phys. Rev. Lett. 91 (2003) 026101. [27] A.D. Rakic, A.B. Djurisic, J.M. Elazar, M.L. Majewski, Appl. Opt. 37 (1998) 5271– 5283.

S. Panaro et al. / Microelectronic Engineering 111 (2013) 91–95 [28] F. Neubrech, T. Kolb, R. Lovrincic, G. Fahsold, A. Pucci, J. Aizpurua, T.W. Cornelius, M.E. Toimil-Molares, R. Neumann, S. Karim, Appl. Phys. Lett. 89 (1– 3) (2006) 253104. [29] L.J. Sherry, S.-H. Chang, G.C. Schatz, R.P. Van Duyne, B.J. Wiley, Y. Xia, Nano Lett. 5 (10) (2005) 2034–2038.

95

[30] J.-S. Huang, J. Kern, P. Geisler, P. Weinmann, M. Kamp, A. Forchel, P. Biagioni, B. Hecht, Nano Lett. 10 (2010) 2105–2110. [31] C.E. Talley, J.B. Jackson, C. Oubre, N.K. Grady, C.W. Hollars, S.M. Lane, T.R. Huser, P. Nordlander, N.J. Halas, Nano Lett. 5 (2005) 1569–1574. [32] P. Biagioni, J.-S. Huang, B. Hecht, Rep. Prog. Phys. 75 (2012) 024402.