Plasmonic/metamaterial sensors are being investigated for their high sensitivity, fast response time, and high accuracy. We propose, characterize and experimentally realize subwavelength bilayer metamaterial sensors operating in the near-infrared domain. We measure the figure-of-merit (FOM) and the bulk sensitivity (S) of the two fundamental hybridized modes and demonstrate both numerically and experimentally that the magnetic dipolar mode, degenerate with the electric quadrupolar mode, has higher sensitivity to a variation of the refractive index compared to the electric dipolar mode. In addition, the hybridized system exhibits a four fold increase in the FOM compared to a standard dipolar plasmonic system.
© 2017 Optical Society of America
In recent years, there has been ample interest in realizing practical plasmonic devices due to their unique ability to manipulate light at a sub-wavelength scale . The applications range from plasmon lasers to chemical and biological sensors [2–7]. Plasmonic resonances depend on material properties (i.e. the metal), the geometrical parameters that define the nanostructure, and the index of the surrounding layer. The sensitivity of the latter is often used to determine global or local variations of the refractive index. By quantitatively observing the spectral locations and the linewidths of these plasmonic resonances, enhanced sensing devices can be constructed.
Optical plasmonic sensors, typically based on the refractive index (RI) variation [8–13], are an important tool for the direct analysis of the physiochemical properties of substances. Among the devices developed for the measurement of the refractive index (RI), compared to their electrical or mechanical counterparts, the optical sensors based on plasmonics have the advantage to significantly shrink the wavelength of light, inducing high field confinement, i.e. exalting the electromagnetic field in very small volumes. Moreover, plasmonic sensors have other advantages in terms of their speed and the fact that they can be used many times, considerably decreasing the sensing cost.
In this paper, we propose and investigate a bilayer plasmonic structure that is composed two metallic bars with structural offset (shift-bar) [14–16]. Previous studies have shown that a shift-bar system makes it possible to control the resonances of metamaterials to produce negative index media [17,18]. This is observed when the symmetry of the system is reduced with a shift in one bar until the two hybridized modes are spectrally superimposed. Here, we evaluate the shift-bar system as a sensing platform by numerically and experimentally analyzing its hybridized modes. The effect of refractive index variation on these hybridized plasmonic modes is observed using a polymethyl methacrylate (PMMA) and a methacrylate (MMA) polymer layers. To quantify this variation, we calculate and measure the sensitivity (S) and the figure-of-merit (FOM) of our devices. The sensitivity is characterized as a variation in the resonant wavelength to a finite change in the refractive index. Resonances information is extracted using the complex poles of the scattering parameters so as to accurately calculate the sensitivity and the FOM.
2. Configuration of the hybridized metamaterial system
The proposed multilayered structure of unit cell composed of gold bars is presented in Fig. 1(a). The dimensions of the individual gold bar were chosen such that its fundamental resonance is in the near infrared at 179.5 THz (1.67 μm) . A scanning electron micrograph (SEM) image of a fabricated structure is shown in Fig. 1(b).
Placing two bars in close proximity hybridizes their individual plasmon modes [21–27]. The resulting two fundamental modes of the system correspond to electric dipolar (ω+) and magnetic dipolar (ω-) modes previously investigated [17,24–26]. Spatially displacing one of the bars with respect to the other can be used to modify the spectral position of these two fundamental modes leading to resonance inversion contingent on near-field coupling. A similar hybridization scheme was used to demonstrate negative refraction [17,18]. Here, after the description of the fabrication process involved in realizing the multilayered structure, we experimentally and quantitatively show the inversion of the two fundamental resonances and investigate their sensing capacities using finite-element numerical simulations and experiments.
The fabrication process for this multi-layer structure is detailed in Fig. 2 below. The multilayer metamaterials are fabricated on a glass substrate using high-resolution electron-beam lithography (EBL) (Vistec EBPG5200 writer). First, the glass substrate is cleaned with acetone and isopropyl alcohol (IPA) while sonicating. To minimize sidewall roughness during the lift-off process, high-resolution positive-tone bilayer resists, methyl methacrylate (MMA-EL 8) and polymethyl methacrylate (PMMA-A2) are used for the e-beam resist. MMA resist is spun on first at a thickness of 150 nm and 50 nm of PMMA is spun subsequently [Figs. 2(a) and (b)]. After the writing step and development by MIBK solvent, a 3 nm layer of chromium (adhesion layer) is deposited followed by 37 nm of gold (Au) using an electron beam evaporation system. The e-beam resist is lifted off using a photoresist remover completing the first layer [Figs. 2(c) and (d)].
After the lift-off process, a 100 nm thick SU-8 photoresist is spin-coated onto the sample. Due to the existence of the first layer of metallic structures, the surface of the SU-8 layer is uneven and needs to be planarized for subsequent fabrication steps. This is done by thermally cycling the sample repeatedly followed by SU-8 crosslinking via UV light exposure and a final hard bake step. To confirm the planarization, the roughness of SU-8 layer surface was determined using atomic force microscopy (AFM) and the surface roughness (RMS) was found to be below 5 nm. Thus, the first layer of gold bars on the glass substrate are embedded in SU-8 which also serves as a dielectric spacer [Fig. 2(e)]. EBL, metal deposition, and lift-off steps for the second layer are carried out in a similar manner as the first layer with the requirement of gold alignment marks to ensure the precise stacking of layers [Figs. 2(f)-2(h)]. The completed multilayer structure can be seen in Fig. 2(h). More layers can be added by repeating the process.
As per the experimental characterization, the transmittance and reflectance spectra were measured using a Fourier-transform infrared spectrometer (FTIR, Vertex 70, Bruker Inc.) system coupled with an optical microscope (Hyperion 2000) operated in the near infrared. A tungsten filament source, KBr beam splitter, x15 Cassegrain objective, liquid-N2-cooled mercury cadmium telluride (MCT) detector, and an infrared polarizer were used to match the frequency of interest (1.25-2.8 μm). The measured transmittance and reflectance spectra are normalized with a glass substrate with SU-8 spacer and with a gold mirror, respectively. The transmittance and reflectance spectra of the designed structure are numerically calculated using a full-wave finite-element simulation software with the refractive index of SU-8, nSU-8 = 1.57, and of glass substrate, nglass = 1.50. The gold bars are described by a Drude model with a plasma frequency ωp = 1.367x1016 rad/s and a collision frequency ωc = 6.478x1013 rad/s .
Figure 3 summarizes the simulation (left column) and experimental (right column) results accompanied with SEM images (middle column) for single layer and multilayer structures with varying shifts, dx. Figure 3(a) shows numerical and experimental spectra for a single layer of gold bars on glass substrate with a single observed resonance at 1.64 μm (182.8 THz) which corresponds to the fundamental localized plasmon resonance on an individual bar . This is slightly different from a fundamental resonance of 1.67 μm with ideal geometrical parameters. To better compare the numerical simulations and the experimental results, we simulated thegeometrical parameters extracted from the SEM images of each sample for a direct comparison with the experimental results. There is quantitatively excellent agreement between the numerical simulations and the experimental results for both reflection and transmission in terms of the location of the resonance and the amplitude of the resonance. This validates the quality of the fabricated single layer structures.
Figures 3(b)-3(g) correspond to the numerical and experimental results for the multi-layer structures for varying values of the shift, dx. From Fig. 3(b), it is seen that the spectrum of transmission and reflection comprises two main hybridized resonances as expected: electric dipolar (ω+) and magnetic dipolar (ω-) modes. The electric dipolar mode resides at a higher frequency (smaller wavelength) and the magnetic dipolar mode resides at a lower frequency (higher wavelength). This splitting of the resonances corresponds to a coupling of the plasmon resonances of the different layers, which leads to a lifting of the degeneracy of the fundamental mode. Figures 3(b)-3(g) shows the existence of these two resonances in all cases (represented by vertical arrows in Fig. 3(b)).
With increasing shift, dx, we observe experimentally that the magnetic dipolar mode moves down in wavelength whereas the electric dipolar mode moves up in wavelength as expected (physical shift seen in Figs. 3(b)-3(g)). This is best observed from the quantitative resonances extracted from the scattering parameters for both the simulation and experiment. There is observable inversion between the electric dipolar and magnetic dipolar modes past a shift of ‘dx = 240nm’ indicating strong near-field coupling [24–27]. Other all dielectric platform may also be used for sensing .
Minor discrepancies in the results are mainly due to fabrication imperfections not accounted for in the simulations such as the slightly rounded edges of the bars, misalignment between layers and the surface roughness. It is important to note that in practice, it is difficult to obtain a perfect alignment of the bars and hence the experimental results deviate slightly from the numerical result. The decay rates (i.e. losses) for both the electric dipolar (i.e. larger linewidth) and magnetic dipolar (i.e. smaller linewidth) modes are as anticipated. The same is true for the resonance frequencies. Overall, an excellent agreement between the numerical and experimental results is observed. Our plasmon hybridized sensing platform is now established in the form of the shift-bar system.
3. Sensitivity to the presence of polymer cladding
We first investigate the behavior of a single plasmon resonance based on the single layer structure. After that, we also investigate the electric dipolar and magnetic dipolar modes’ sensitivity [Eq. (1)] and FOM [Eq. (2)] as a function of the coupling strength to compare their sensitivity to the surrounding refractive index.
To better investigate the refractive index sensing capability of our proposed devices, we deposited by way of spin-coating two different and separate cladding layers (h = 70nm) each with different refractive index respectively: PMMA (n = 1.4778) and MMA (n = 1.4118). As seen in Fig. 4, the interaction between the surrounding medium (PMMA and MMA) and the near field significantly affects the resonance wavelength of the fundamental mode (single layer metallic bars).
A sensor is most efficient if it combines both a spectral shift called the bulk refractive index sensitivity (defined as the ratio of resonant wavelength shift, Δλ to the variation of the surrounding refractive index, Δn (RIU)) with a small resonant bandwidth also known as the figure-of-merit (FOM). Hence, we evaluate the quality of our sensors based on their FOM. This FOM is defined as the ratio of the bulk refractive index sensitivity to the full-width-half-maximum (FWHM) of the corresponding resonance. This quantity determines the overall performance of the sensor.
Figure 4 presents the reflection spectra recorded for three different surrounding refractive indices (nair = 1, nPMMA = 1.4778, nMMA = 1.4118) for both numerical simulation (left-column) and experiment (right-column). We experimentally observe that the presence of a refractive index greater than that of air redshifts the resonance by Δλ = 260 nm for PMMA (Δn = 0.4778) top- cladding and by Δλ = 242 nm for MMA (Δn = 0.4118). The linewidth of the single layer dipolar resonance is ~640 nm.
Experimental and numerical data are in good agreement in terms of the resonance position and amplitude. To evaluate the performance of this rudimentary sensor, we experimentally calculate the bulk sensitivity for PMMA (544 nm/RIU) and MMA (587 nm/RIU). The FOM in both cases for the single-layer structure are: FOMPMMA = 0.85 and FOMMMA = 0.91. In comparison, the sensitivity values acquired from simulated scattering spectra are 571 nm/RIU with PMMA and 578 nm/RIU with MMA. The difference in the sensitivities arises from both the fabrication imperfections and the penetration of the field in the material to sense.
Similarly, we evaluate the sensing ability of the hybridized plasmonic system for both the electric dipolar (ω+) and magnetic dipolar (ω-) modes. As previously stated, there is observable inversion between the electric dipolar and magnetic dipolar modes for shifts, dx, larger than ‘dx = 240 nm’ for resonances extracted from scattering spectra for both simulation [Figs. 5(a) and 5(c)] and experiment [Figs. 5(b) and 5(d)]. Here, the resonance positions of the two modes are acquired from the complex poles of the fitted scattering parameters . For an added top-cladding of either PMMA [Figs. 5(a) and 5(b)] or MMA [Figs. 5(c) and (d)], both electric dipolar and magnetic dipolar resonances are redshifted. Moreover, the magnetic dipolar mode is more sensitive compared to the electric dipolar mode judging from the shift in resonances (Δλ) for both PMMA and MMA cladding irrespective of ‘dx’. In both cases for dx = 0, the experimentally calculated FOMs for the electric dipolar mode are FOMPMMA = 0.08 and FOMMMA = 0.27 whereas for the magnetic dipolar the FOMs are FOMPMMA = 2.65 and FOMMMA = 3.82. Only the magnetic dipolar mode is more sensitive than the single-layer fundamental mode which has FOMPMMA = 0.85 and FOMMMA = 0.91. At dx = 0, the linewidths of the electric and magnetic dipolar modes are ~830 nm and ~210 nm respectively. Experimentally, the electric dipolar mode experiences a resonance shift of 39 nm for PMMA and 110 nm for MMA. Similarly, the magnetic dipolar mode experiences a resonance shift of 313 nm for PMMA and 250 nm MMA.
Hence, we plot the FOM for the higher sensitivity magnetic dipolar mode as a function of the shift, dx, in comparison to the single-layer device for both PMMA [Fig. 6(a)] and MMA [Fig. 6(b)] top-claddings. We note here that there is a factor of 4 increase in the FOM for the hybridized magnetic dipolar resonance compared to the fundamental resonance of a nano-bar.
We have numerically and experimentally demonstrated the sensing capability of a hybridized plasmonic system to be superior to a standard single plasmonic resonance system. The FOM is a factor of 4 greater in the hybridized case using the magnetic dipolar mode. Furthermore, the magnetic dipolar mode of the hybridized system has a higher sensitivity and FOM compared to the electric dipolar mode regardless of the shift. A quantitative analysis of the plasmonic resonances allowed for an accurate computation of both sensitivity and FOM. Moreover, specificity can be introduced to the current hybridized label-free sensing scheme by functionalizing the gold nanoparticles with appropriate biological or chemical markers. Intricate hybridized metamaterial resonances will usher the next generation of sensing devices.
This work was partially supported the National Science Foundation Career Award (ECCS-1554021) and the U.S. Department of Energy (DOE) (EE0007341).
The authors thank UCSD’s Nano3 cleanroom staff, Dr. Maribel Montero for assistance with nano-fabrication.
References and links
1. S. A. Maier, Plasmonics: Fundamentals and Applications (Springer, 2007).
2. R.-M. Ma, R. F. Oulton, V. J. Sorger, G. Bartal, and X. Zhang, “Room-temperature sub-diffraction-limited plasmon laser by total internal reflection,” Nat. Mater. 10(2), 110–113 (2011). [CrossRef] [PubMed]
3. P. Berini and I. De Leon, “Surface plasmon–polariton amplifiers and lasers,” Nat. Photonics 6(1), 16–24 (2011). [CrossRef]
4. J. A. Schuller, E. S. Barnard, W. Cai, Y. C. Jun, J. S. White, and M. L. Brongersma, “Plasmonics for extreme light concentration and manipulation,” Nat. Mater. 9(3), 193–204 (2010). [CrossRef] [PubMed]
9. S. Zhan, Y. Peng, Z. He, B. Li, Z. Chen, H. Xu, and H. Li, “Tunable nanoplasmonic sensor based on the asymmetric degree of Fano resonance in MDM waveguide,” Sci. Rep. 6(1), 22428 (2016). [CrossRef] [PubMed]
10. D. Wu, Y. Liu, L. Yu, Z. Yu, L. Chen, R. Li, R. Ma, C. Liu, J. Zhang, and H. Ye, “Plasmonic metamaterial for electromagnetically induced transparency analogue and ultra-high figure of merit sensor,” Sci. Rep. 7, 45210 (2017). [CrossRef] [PubMed]
15. B. Kanté, Y. S. Park, K. O’Brien, D. Shuldman, N. D. Lanzillotti-Kimura, Z. Jing Wong, X. Yin, and X. Zhang, “Symmetry breaking and optical negative index of closed nanorings,” Nat. Commun. 3, 1180 (2012). [CrossRef] [PubMed]
16. R. Ghasemi, X. Le Roux, A. Lupu, A. de Lustrac, and A. Degiron, “On the nonlocal response of multilayer optical metamaterials,” ACS Photonics 2(8), 1129–1134 (2015). [CrossRef]
17. B. Kanté, S. N. Burokur, A. Sellier, A. de Lustrac, and J.-M. Lourtioz, “Controlling plasmon hybridization for negative refraction metamaterials,” Phys. Rev. B 79(7), 075121 (2009). [CrossRef]
18. V. M. Shalaev, W. Cai, U. K. Chettiar, H. K. Yuan, A. K. Sarychev, V. P. Drachev, and A. V. Kildishev, “Negative index of refraction in optical metamaterials,” Opt. Lett. 30(24), 3356–3358 (2005). [CrossRef] [PubMed]
19. P. B. Johnson and R. W. Christy, “Optical constant of the noble metals,” Phys. Rev. B 6(12), 4370–4379 (1972). [CrossRef]
20. L. Novotny and N. Van Hulst, “Antennas for light,” Nat. Photonics 5(2), 83–90 (2011). [CrossRef]
21. P. Nordlander, C. Oubre, E. Prodan, K. Li, and M. I. Stockman, “Plasmon hybridization in nanoparticle dimers,” Nano Lett. 4(5), 899–903 (2004). [CrossRef]
23. J. A. Fan, C. Wu, K. Bao, J. Bao, R. Bardhan, N. J. Halas, V. N. Manoharan, P. Nordlander, G. Shvets, and F. Capasso, “Self-assembled plasmonic nanoparticle clusters,” Science 328(5982), 1135–1138 (2010). [CrossRef] [PubMed]
24. A. Christ, Y. Ekinci, H. H. Solak, N. A. Gippius, S. G. Tikhodeev, and O. J. F. Martin, “Controlling the Fano interference in a plasmonic lattice,” Phys. Rev. B 76(20), 201405 (2007). [CrossRef]
25. A. Kodigala, T. Lepetit, and B. Kanté, “Engineering resonance dynamics of plasmon hybridized systems,” J. Appl. Phys. 117(2), 023110 (2015). [CrossRef]
26. A. Kodigala, T. Lepetit, and B. Kanté, “Exceptional points in three-dimensional plasmonic nanostructures,” Phys. Rev. B 94(20), 201103 (2016). [CrossRef]
27. R. Taubert, R. Ameling, T. Weiss, A. Christ, and H. Giessen, “From Near-Field to Far-Field Coupling in The Third Dimension: Retarded Interaction of Particle Plasmons,” Nano Lett. 11(10), 4421–4424 (2011). [CrossRef] [PubMed]