nanomaterials Article An Inverted Honeycomb Plasmonic Lattice as an Efficient Refractive Index Sensor Javier Rodríguez-Álvarez 1,2, *, Lorenzo Gnoatto 1 , Marc Martínez-Castells 1 , Albert Guerrero 3 , Xavier Borrisé 4 , Arantxa Fraile Rodríguez 1,2 , Xavier Batlle 1,2 and Amílcar Labarta 1,2 1 Departament de Física de la Matèria Condensada, Universitat de Barcelona, 08028 Barcelona, Spain; lgnoatgn26@alumnes.ub.edu (L.G.); martinezcastellsmarc@gmail.com (M.M.-C.); arantxa.fraile@ub.edu (A.F.R.); xavierbatlle@ub.edu (X.B.); amilcar.labarta@ub.edu (A.L.) 2 Institut de Nanociència i Nanotecnologia (IN2UB), 08028 Barcelona, Spain 3 Institut de Microelectrónica de Barcelona (IMB-CNM, CSIC), 08193 Bellaterra, Spain; albert.guerrero@imb-cnm.csic.es 4 Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and The Barcelona Institute of Science and Technology, Campus UAB, Bellaterra, 08193 Barcelona, Spain; xavier.borrise@imb-cnm.csic.es * Correspondence: javier.rodriguez@ub.edu Abstract: We present an efficient refractive index sensor consisting of a heterostructure that contains an Au inverted honeycomb lattice as a main sensing element. Our design aims at maximizing the out-of-plane near-field distributions of the collective modes of the lattice mapping the sensor surroundings. These modes are further enhanced by a patterned SiO2 layer with the same inverted honeycomb lattice, an SiO2 spacer, and an Au mirror underneath the Au sensing layer that contribute  to achieving a high performance. The optical response of the heterostructure was studied by numer-  ical simulation. The results corresponding to one of the collective modes showed high sensitivity Citation: Rodríguez-Álvarez, J.; values ranging from 99 to 395 nm/RIU for relatively thin layers of test materials within 50 and 200 nm. Gnoatto, L.; Martínez-Castells, M.; In addition, the figure of merit of the sensor detecting slight changes of the refractive index of a water Guerrero, A.; Borrisé, X.; Fraile medium at a fixed wavelength was as high as 199 RIU−1 . As an experimental proof of concept, the Rodríguez, A.; Batlle, X.; Labarta, A. heterostructure was manufactured by a simple method based on electron beam lithography and the An Inverted Honeycomb Plasmonic measured optical response reproduces the simulations. This work paves the way for improving both Lattice as an Efficient Refractive Index the sensitivity of plasmonic sensors and the signal of some enhanced surface spectroscopies. Sensor. Nanomaterials 2021, 11, 1217. https://doi.org/10.3390/nano11051217 Keywords: Au plasmonic nanostructures; inverted honeycomb lattice; surface lattice resonances; refractive index sensor Academic Editor: Daniela Iacopino Received: 9 April 2021 Accepted: 1 May 2021 1. Introduction Published: 4 May 2021 The design and manipulation of plasmonic nanostructures have rapidly become one of Publisher’s Note: MDPI stays neutral the most active topics in nanophotonics over the last years. Specifically, several plasmonics- with regard to jurisdictional claims in based approaches have been applied to improve the detection limit and sensitivity of published maps and institutional affil- sensing devices. Good examples of these new capabilities are the potential to prove iations. volumes beyond the diffraction limit [1–3] or to achieve single-molecule sensitivity [4]. Moreover, plasmonic nanostructures have contributed to largely enhancing the signal of surface spectroscopies, such as surface-enhanced Raman scattering or plasmon-enhanced fluorescence [5–9]. Copyright: © 2021 by the authors. One of the most promising niches for plasmonic nanostructures is their integration in Licensee MDPI, Basel, Switzerland. refractive index sensors [10]. This kind of sensor benefits from the fact that the wavelengths This article is an open access article at which the plasmonic excitations arise depend strongly on the features of the medium distributed under the terms and mapping the corresponding near-field distributions. However, the inherent nature of conditions of the Creative Commons localized surface resonances (LSRs) hinders their efficiency as sensing devices as the Attribution (CC BY) license (https:// evanescent electric fields excited in an LSR decrease rapidly with the distance from the creativecommons.org/licenses/by/ metal–dielectric interface. In this way, the regions in which the electric field is enhanced, the 4.0/). so-called “hotspots”, extend only a few tens of nanometers away from the metal–dielectric Nanomaterials 2021, 11, 1217. https://doi.org/10.3390/nano11051217 https://www.mdpi.com/journal/nanomaterials Nanomaterials 2021, 11, 1217 2 of 12 interface and, in many cases, are confined in the space between neighboring plasmonic elements. In contrast, inverted structures, in which dielectric holes are carved through a continuous metallic layer, have proven to be a valid approach to tackle this issue [11] and are known to improve the efficiency of light–matter interaction in some situations [12–16]. Unlike other realizations of inverted plasmonic structures [17–19], such as nanoholes or nanoarrays with a more complex pattern, the heterostructure proposed in this work benefits from the inclusion of a mirror layer. By adding this element, the device can collect all the optical responses solely in the reflection spectrum, given the null transmission. In addition, our system is based on the detection of surface lattice resonances (SLRs) in a honeycomb lattice, which implies sharper peaks associated with higher-energy modes involving the collective excitation of all the elements in the plasmonic array [20–23]. These modes usually present hotspots encompassing larger areas than those associated with local excitations. Besides, the three-fold symmetry of the honeycomb lattice may hamper the excitation of LSR in favor of SLR, making collective modes much more intense [20,21], as will be further discussed in Section 3.2. Overall, the heterostructure presented in this work constitutes a proof of concept for a simple realization of a highly sensitive refractive index sensor that can be manufactured by a relatively easy lithographic process. 2. Materials and Methods As shown in Figure 1, the main part of the studied heterostructure is composed of two layers of Au and SiO2 of 30 and 80 nm in thickness, respectively, where an inverted honeycomb lattice—with a pitch of 866 nm and bar-shaped trenches of 400 nm in length and 30 nm in width—is carved through the whole Au/SiO2 bilayer. On the bottom of the trenches, there is an Au layer of 30 nm in thickness that shapes the direct honeycomb lattice complementary to that at the top. Thus, the two plasmonic honeycomb lattices are placed at an edge-to-edge distance of 50 nm. Between them, there is a sort of two-dimensional photonic crystal that is formed by the contrast in the refractive indices of the SiO2 and the material filling the trenches (air, water, and so on). Consequently, the interactions among the three lattices feed back into each other because they share the same pitch and symmetry. Specifically, one of the collective modes of the two plasmonic structures gives rise to a Nanomaterials 2021, 11, 1217 3 of 12 spatial distribution of the near-field enhancement around the heterostructure that is well suited for sensing applications, as discussed later. This heterostructure rests on top of a second bilayer, composed ofthe periodicity of ansystem, SiO2 periodic spacerboundary and an conditions Au mirror of aof 120 nmunit rectangular andcell100 nm in of the thickness, respectively, which honeycomb enhances lattice are used to the intensity perform of the(see the simulations collective Figure 1b).modes andfunc- The dielectric ensures that the transmission tions through for the materials used insystem the entire these simulations were obtained by fitting analytical func- is negligible. tions to the data from [25] for SiO2 and Au, respectively. Figure 1. (a) Scheme of the heterostructure. The inset shows a schematic cross-sectional view. (b) Top view showing the Figure 1. (a) Scheme of the heterostructure. The inset shows a schematic cross-sectional view. (b) Top honeycomb lattice, the hexagonal Bravais lattice (dotted lines), a rectangular unit cell (dashed grey lines), and some rele- view vant showing parameters the of the honeycomb design. lattice, Regular rhombus arethe hexagonal primitive Bravais cells of the lattice hexagonal lattice.(dotted lines), The cell for a rectangular the simulations is unit the rectangular unit cell. At the bottom left corner of the panel, there is a schematic representation of the electric field and cell (dashed grey lines), and some propagation vectors of the incident radiation. relevant parameters of the design. Regular rhombus are primitive cells of the hexagonal lattice. The cell for the simulations is the rectangular unit cell. At the bottom left corner of the panel,Inthere the manufacturing is a schematic of the samples, a bilayer representation made of the up of Au electric (mirror) field and SiO2 and propagation (spacer) layers of 100 nm in thickness each was deposited on top of a standard Si wafer by vectors of the incident radiation. electron beam evaporation (ATC Orion, AJA International, Inc) and plasma enhanced chemical vapor deposition (PECVD), respectively. A pattern of 500 µm × 500 µm in size following the inverted honeycomb design shown in Figure 1 was exposed by electron beam lithography (EBL) on an additional third layer of poly methyl methacrylate (950 PMMA, MicroChem) with a thickness of 90 nm. After removing the exposed areas, the sample was metallized with 30 nm of Au deposited by electron beam evaporation. The outcome of the evaporation process was a 30 nm Au layer on top of the heterostructure Nanomaterials 2021, 11, 1217 3 of 12 The simulations presented in this work were performed using the finite-difference time-domain (FDTD) method, implemented in the solver provided by LumericalR [24].The method defines a simulation spatial domain, which is discretized. Then, the incident electromagnetic field is set by defining the initial state of the system. The software solves Maxwell’s equations to determine the evolution of the electric and magnetic fields using a discrete time-step. This approximation, as well as the inherent mesh step necessary to discretize the space in the simulation domain, introduce a certain error into the calculations. However, the method allows for direct observation of the physical phenomena taking place without imposing any further assumption on the behavior of the system. The configuration of the simulations was as follows: a short pulse of linearly polarized light was sent perpendicularly to the surface of the heterostructure from a source placed above the metal (see the inset of Figure 1b). Placing different monitors in several spots, we obtained the transmission and reflection cross sections, as well as the electric field distribution as a function of the wavelength of the incident radiation. Owing to the periodicity of the system, periodic boundary conditions of a rectangular unit cell of the honeycomb lattice are used to perform the simulations (see Figure 1b). The dielectric functions for the materials used in these simulations were obtained by fitting analytical functions to the data from [25] for SiO2 and Au, respectively. In the manufacturing of the samples, a bilayer made up of Au (mirror) and SiO2 (spacer) layers of 100 nm in thickness each was deposited on top of a standard Si wafer by electron beam evaporation (ATC Orion, AJA International, Inc) and plasma enhanced chemical vapor deposition (PECVD), respectively. A pattern of 500 µm × 500 µm in size following the inverted honeycomb design shown in Figure 1 was exposed by electron beam lithography (EBL) on an additional third layer of poly methyl methacrylate (950 PMMA, MicroChem) with a thickness of 90 nm. After removing the exposed areas, the sample was metallized with 30 nm of Au deposited by electron beam evaporation. The outcome of the evaporation process was a 30 nm Au layer on top of the heterostructure shaping an inverted honeycomb lattice and the complementary direct version of the same lattice on the bottom of the trenches, as shown in Figure 1a, at an edge-to-edge distance of 60 nm. Between them, a two-dimensional photonic crystal is formed by the contrast in the refractive indices of the 950 PMMA and the material filling the trenches. No further lift-off was necessary owing to the design of the samples. As SiO2 and 950 PMMA have very similar refractive indices, the optical spectra of the simulated and manufactured heterostructures may be compared right away, even though some geometrical parameters are slightly different than in the simulations, namely, the edge-to-edge distances between the inverted Au layer and both the mirror (190 nm) and the direct Au structure (60 nm). The optical characterization was carried out by a Vertex 70 Fourier transform infrared (FTIR) spectrophotometer attached to an optical microscope (Bruker Hyperion). Experi- ments were performed in the reflection configuration with a 4× objective and under the illumination of unpolarized light. The measured signal of an Ag mirror was used as a background in all the experiments. 3. Results 3.1. Optical Response of the Heterostructure Several peaks are exhibited by the simulated reflection spectrum of this heterostructure (see Figure 2). Nevertheless, they can be classified into only two groups depending on either the local or collective nature of the resonances associated with them. The first group includes the peaks located around 729 and 1789 nm, which arise from the excitation of LSR related to local modes of the bar-shaped trenches and the Au bars in the inverted and direct plasmonic structures, respectively. Accordingly, the near-field distributions for these modes are mainly confined inside either the trenches through the Au layer or around the Au bars at the bottom of the SiO2 trenches. For instance, the corresponding near-field distributions for these two peaks show mostly dipolar excitations of the trenches perpendicular to the polarization axis of the incoming radiation with some multipolar or dipolar polarizations ing the Au mirror and varying the edge-to-edge distance between the inv and the Au mirror as there is a certain coupling with the images of the mirror. In such a way, 200 nm was chosen for the edge-to-edge distance Nanomaterials 2021, 11, 1217 at 729 nm was blue-shifted to be on the left-hand side of the sharp 4 of 12peak a is the one we are intending to use for sensing applications. In a similar w edge distance between the inverted Au lattice and the direct one was opt nm oftothe tilted trenches for the peaks at 729 and 1789 nm, respectively (see Figure S1a,c,d,f in reduce the overlap between the peaks at 729 and 765 nm. the Supplementary Materials). Figure 2. Simulated reflection spectrum as a function of the wavelength of the incident radiation. Figure 2. Simulated reflection spectrum as a function of the wavelength of the inc The sharp peaks at about 765 and 1107 nm correspond to the surface lattice resonance (SLR) of the The Au sharp peaks inverted at about honeycomb lattice765 at theand 1107 air–Au and nm correspond Au–SiO to the surface lattice reson 2 interfaces, respectively. Au inverted honeycomb lattice at the air–Au and Au–SiO2 interfaces, respectively It is also worth noting that the positions of these two peaks can be tuned just by adding the Au mirror and varying the edge-to-edge distance between the inverted Au In the lattice andsecond group the Au mirror of excitations as there shown is a certain coupling withintheFigure images of2, there these areatthose w modes the mirror. In such a way, 200 nm was chosen for the edge-to-edge distance so that the peak character that are associated with SLRs. The corresponding peaks are the at 729 nm was blue-shifted to be on the left-hand side of the sharp peak at 765 nm, which is onethe at one 765wenm and thetoones are intending use forlocated about 754 sensing applications. In a and similar1107 nm. way, the These three p edge-to-edge the distance between the constructive inverted Au lattice interference and the direct between one was optimized the incident radiationto be 50 andnm to the coll reduce the overlap between the peaks at 729 and 765 nm. the direct and In the inverted second honeycomb group of excitations shownplasmonic in Figure 2, therelattices. are thoseConsidering with a collective that S at the interface character that arebetween theSLRs. associated with Au Theandcorresponding the surroundingpeaks are thedielectric media, tw aforementioned one at 765 nm and the ones located about 754 and 1107 nm. These three peaks arise from be expected for each honeycomb lattice as both the direct and the inverte the constructive interference between the incident radiation and the collective modes of in contact the directwith two different and inverted dielectric honeycomb plasmonic media, lattices. namely Considering thatair SLRs(above) and SiO are excited at the interface between the Au and the surrounding dielectric media, two SLRs should hexagonal lattice, which is the Bravais lattice of the honeycomb array, th be expected for each honeycomb lattice as both the direct and the inverted Au arrays are the in excitation contact withwith perpendicular two different incidence dielectric media, namely airof the and (above) most SiOintense SLR 2 (below). For a is hexagonal lattice, which is the Bravais lattice of the honeycomb array, the condition for the excitation with perpendicular incidence of the most 𝜆 intense = 𝑛 𝑝SLR cosis 30, λi = ni p cos 30, where 𝜆 is the resonant wavelength of the incident radiation, 𝑝(1)is the p agonal wherelattice (𝑝 = 866 λi is the resonant nm; see wavelength Figure of the incident 1b), andp is𝑛theis radiation, the pitch effective of the hexagonalrefract lattice (p = 866 nm; see Figure 1b), and ni is the effective refractive index of the dielectric dielectric medium forming the interface with the Au layer. Taking avera medium forming the interface with the Au layer. Taking average values of the refractive refractive indices ofindices of2air air and SiO as 1and and 1.45, 2 as 1 and the SiOrespectively, 1.45, respectively, expected wavelengthsthe thatexpected fulfill w Equation (1) are 750 nm and 1087 nm, respectively, in very close agreement with the fulfill Equation (1) are 750 nm and 1087 nm, respectively, in very close agr positions of the peaks found by numerical simulation. positions ofthe Thus, the peaks small found754bynmnumerical peak around simulation. is mainly related to the SLR at the interface with the air of the direct honeycomb lattice on the bottom of the SiO2 trenches. This is supported by both the near-field distribution shown in Figure S1b,e in the Supplementary Materials Thus, the small peak around 754 nm is mainly related to the SLR at the interface with the air of the direct honeycomb lattice on the bottom of the SiO2 trenches. This is supported by both the near-field distribution shown in Figure S1b,e in the Supplementary Materials Nanomaterials 2021, 11, 1217 5 of 12 and the fact that this peak does not appear when the direct Au honeycomb lattice is not present, and the trenches are carved through the whole thickness of the SiO2 layer down to the Au mirror (see Figure S2 in the Supplementary Materials). Interestingly, in the latter and the case, thefact thataround peak this peak729 doesnm notremains appear when almost theunchanged, direct Au honeycomb strengtheninglatticethe is notargument present, and the trenches are carved through the whole thickness that it is only associated with local modes of the trenches in the inverted Au lattice. of the SiO 2 layer down to The the Au mirror (see Figure S2 in the Supplementary Materials). Interestingly, in the latter second SLR mode corresponding to the Au–SiO2 interface of the direct honeycomb lattice case, the peak around 729 nm remains almost unchanged, strengthening the argument that may be likely it is only overlapped associated with localwith modes theofintense peakin the trenches around 1107 nm. the inverted Au lattice. The second On the contrary, the two peaks around 765 SLR mode corresponding to the Au–SiO2 interface of the direct honeycomb and 1107 nm are associated lattice maywithbe the two SLRs at the air–Au and Au–SiO interfaces likely overlapped with the intense peak around 1107 nm. 2 of the inverted honeycomb lattice, respectively BothOn peaks are verythe the contrary, intense and relatively two peaks around 765sharp, and 1107 which makes nm are them suitable associated with the twofor applica- SLRs atItthe tions. air–Au is also and Au–SiO worth noting 2that interfaces of the inverted the overlap betweenhoneycomb the two peaks lattice, atrespectively. 729 and 765 nm is Both peaksreduced somehow are very intense and relatively by the coupling withsharp, which makes them the two-dimensional suitable crystal photonic for applica- underneath tions. It is also worth noting that the overlap between the inverted plasmonic lattice, which is beneficial for sensing purposes. As the two peaks at 729 and 765anm is comparison somehow reduced by the coupling with the two-dimensional photonic crystal underneath Figure S2 in Supplementary Materials shows the spectra for the cases of a continuous SiO2 the inverted plasmonic lattice, which is beneficial for sensing purposes. As a comparison, spacer of 200 nm in thickness and the same spacer, but with trenches carved through its Figure S2 in Supplementary Materials shows the spectra for the cases of a continuous SiO2 entire spacerthickness. of 200 nm in thickness and the same spacer, but with trenches carved through its entireInthickness. addition, simulations of the near-field distributions on the heterostructure at 765 and 1107 Innm reveal the addition, formation simulations of of thepatterns near-field ofdistributions extended hotspots compatible with on the heterostructure the transla- at 765 and 1107 tional nm reveal symmetry of the theformation hexagonaloflattice.patterns For ofthe extended peak athotspots 1107 nm, compatible there arewith also the some extra translational symmetry of the hexagonal lattice. For the peak at 1107 excitations inside the trenches of the inverted Au layer (Figure 3b) arising from the coupling nm, there are also some extra between theexcitations SLR modes inside the direct of the trenches andofthethe inverted inverted lattices Au layerthrough (Figure the3b) polarization arising of from the coupling between the SLR modes of the direct and the inverted lattices through the regions of the trenches in between the two plasmonic lattices (see Figure 3d). Although the polarization of the regions of the trenches in between the two plasmonic lattices (see the two peaks show very similar near-field patterns outside the trenches, it is interesting to Figure 3d). Although the two peaks show very similar near-field patterns outside the note that,itfor trenches, the peak atto765 is interesting note nm,that,theforhotspots the peakfollowing at 765 nm,the thelattice hotspotssymmetry following arethefound on top of symmetry lattice the plasmonic Au layer, are found on top whereas in the case of the plasmonic Au oflayer, the peak at 1107 whereas in thenm, theofelectric case the field enhancement peak at 1107 nm, takes theplace at field electric the bottom interface enhancement with takes the at place SiO 2 layer the bottom(seeinterface Figure 3c,d). with In both the SiOthe cases, 2 layer (see of pattern Figure 3c,d).yields hotspots In both ancases, the pattern out-of-plane of hotspots electric field thatyields an out-of- extends hundreds of plane electricaway nanometers field from that extends hundreds ofinterfaces. the corresponding nanometers away The from the corresponding corresponding distributions of the interfaces. electric The field correspondingspread enhancement distributions of the the deep inside electric fieldthe air and enhancement spread deep SiO2, respectively (see Figure inside the air and the SiO2 , respectively (see Figure 3c,d). It is precisely these out-of-plane 3c,d). It is precisely these out-of-plane near-field distributions associated with the SLR of the near-field distributions associated with the SLR of the inverted honeycomb lattice that inverted honeycomb lattice that provide the heterostructure with strong sensing capabili- provide the heterostructure with strong sensing capabilities, as will be discussed in the ties, as will be discussed in the following subsection. following subsection. Figure 3.Near-field Figure 3. Near-field distributions distributions associated associated with with the theatSLR SLR at 765 765 nm (a,c)nm and(a,c) 1107 and 1107 (a,b) nm (b,d). nm (b,d). cor- (a) and (b) correspond respond to in-plane to in-plane views viewsand 10 nm above 10 nm above below and below the inverted the inverted honeycomb honeycomb lattice, respectively,lattice, at re- spectively, at the heights the heights indicated by theindicated by in dashed lines the dashed the lines in panels corresponding the corresponding panels (c,d). (c,d) depict (c,d). (c,d) de- transversal pict crosstransversal cross sections of the sections of the heterostructure heterostructure along the dashed linesalong the dashed plotted lines plottedin-plane in the corresponding in the corre- panels → (a,b), respectively. ⃗ 𝐸 /𝐸 view panels (a,b), respectively. | E |/E0 stands for the normalized modulus of the electric fieldmodulus sponding in-plane view stands for the normalized with of the electric respect to thefield with of modulus respect to theradiation. the incident modulus of the incident radiation. Nanomaterials 2021, 11, 1217 6 of 12 3.2. Performance of the Refractive Index Sensor The design of the plasmonic heterostructure following a lattice with three-fold sym- metry hampers the excitation of the LSRs that are not fully compatible with the symmetry of a hexagonal lattice, because local modes imply polarizations with an even number of charge poles [20,21]. For instance, the three bar-shaped elements of the honeycomb lattice converging on each corner of the hexagons are not compatible with a simple dipolar mode within the gap between them. This favors the excitation of the SLRs yielding more intense absorption peaks, which in turn may improve the optical response of the system for sensing purposes. While preserving the honeycomb design, the system presented in this work does not follow direct realizations of a honeycomb lattice based on arrays of plasmonic nanoelements [26–28]. Instead, it takes advantage of the inverted version of the direct lattice. The main plasmonic structure presented here is, as opposed to a honeycomb array of bars, a continuous layer of Au where bar-shaped trenches are carved through its full thickness. According to Babinet’s principle, the direct and the inverted structures should be equivalent in terms of their optical properties [14,29], showing the same excitations at the same values of the wavelength, but the nature of the excitations is different in the two structures. The electric dipoles formed in the direct structure are magnetic dipoles in the corresponding inverted realization and vice versa, because the role of the charge accumulation is played by the currents induced in the system in the complementary struc- ture. Previous works show that, despite the similarities in the spectral behavior, inverted structures can perform better in terms of their interaction with the surrounding medium or an externally exciting source [7–9]. One particularly appealing characteristic of the inverted structures is that the near-field distributions tend to spread further away compared with the direct structure. The near-field distributions for the peaks at 765 and 1107 in Figure 3 are good examples of this spreading. In particular, the SLR at about 765 nm exhibits hotspots that extend out-of-plane largely away from the top Au–air interface. Then, any changes in the properties of the surrounding medium above the heterostructure will especially affect this excitation because the electric field is distributed in a relatively large volume that extends up to about 400 nm from the Au–air interface (see Figure 3c,d). This allows both to detect slight changes in the optical properties of a relatively distant region of the medium above and to boost the sensitivity of surface enhanced spectroscopies, such as Raman scattering. In this work, we aim at presenting a sensing application based on the detection of small changes in the refractive index of the medium above the heterostructure by measuring the shift in the wavelength of the SLR at the interface with the top Au layer. More precisely, a 50 nm layer of a test material with a given refractive index is placed on top of the heterostructure. This layer is intended to model the adsorption on top of the heterostructure of some biological species, such as macromolecules or viruses (with refractive index values in the range of our test layer) or any other materials to be detected not fitting inside the trenches because of their relatively small width (30 nm). We decided to use a thin test layer to evidence the high sensitivity of the sensor. In Figure 4, the reflection spectra around the two SLR peaks of the inverted plasmonic structure are depicted for values of the refractive index of the test layer ranging from 1 to 1.45. Although the two peaks show remarkable shifts, the one for the SLR at the top test layer–Au interface is much larger, as expected from the corresponding near-field distributions shown in Figure 3. The fact that the near-field enhancement for this excitation is mostly distributed outside of the heterostructure makes its wavelength highly susceptible to changes in the surrounding medium. Nanomaterials 2021, 11, 1217 7 of 12 Nanomaterials 2021, Nanomaterials 11,11, 2021, 1217 1217 7 of 12 7 of 12 Figure 4. Reflection spectra around the wavelength ranges where the SLRs corresponding to the top (a) and bottom (b) interfaces of the Au inverted honeycomb lattice take place. The spectra were simulated Figure for several 4. Reflection spectravalues aroundwithin 1.1 and 1.45 the wavelength of the ranges refractive where the SLRsindex of a test layer corresponding of 50 to the Figure 4. Reflection spectra around the wavelength ranges where the SLRs corresponding to the nm in(a) top thickness resting and bottom on top ofofthe (b) interfaces theheterostructure. The spectrum Au inverted honeycomb in Figure lattice take 2 without place. The the test spectra were top (a) and bottom (b) interfaces of the Au inverted honeycomb lattice take place. The spectra layer is included simulated as a reference for several (black values within 1.1line). and 1.45 of the refractive index of a test layer of 50 nm in were simulated for several values within 1.1 and 1.45 of the refractive index of a test layer of 50 thickness resting on top of the heterostructure. The spectrum in Figure 2 without the test layer is nm in thickness resting on top of the heterostructure. The spectrum in Figure 2 without the test layerItisisincluded includedworth noting as a reference thatline). (black the shiftline). in the wavelength of this resonance follows an almost as a reference (black perfectly linear dependence on the refractive index of the test layer (see Figure 5a). This It is worth noting that the shift in the wavelength of this resonance follows an almost can beIt easily is worth understood thatconsidering thatwavelength the resonant wavelength satisfying a SLR is perfectly linear noting dependence the shift on the in the refractive index of the of testthis resonance layer (see Figurefollows an 5a). This almost linearly perfectly dependent linear can be easily on the dependence understood effective refractive on the refractive considering index index of that the resonant of the dielectric the test layer wavelength medium (see Figure satisfying that a SLR is forms 5a). This the can interface be easily linearly with the understood dependent Aueffective on the layer, as considering shown refractivethat inthe index Equation ofresonant (1). Hence, wavelength the dielectric medium thethat sensitivity satisfying forms thea of SLRtheis sensor computed interface with thefrom Au the layer, slope as of shown the in linear Equationdependence (1). Hence, shown the in Figure sensitivity linearly dependent on the effective refractive index of the dielectric medium that forms of 5a theis 99 sensornm/RIU (RIU the standsfrom computed interface for refractive the the with slopeAuofindex the unit), linear layer, whichinisEquation as dependence shown a veryin shown remarkable Figure 5a is result (1). Hence, 99the considering nm/RIU (RIU ofthe sensitivity the stands small for refractive thickness of theindex test unit), layer. which is a very remarkable result sensor computed from the slope of the linear dependence shown in Figure 5a is 99 nm/RIUconsidering the small thickness of the test layer. (RIU stands for refractive index unit), which is a very remarkable result considering the small thickness of the test layer. Figure Figure5.5.(a)(a) Wavelength Wavelengthof ofthe thepeak peak corresponding corresponding to tothe theSLR SLRatatthe thetest testlayer–Au layer–Au interface interface as as a function of the a function refractive of the index refractive index for forthree threevalues valuesof of the the test test layer thickness,namely, layer thickness, namely, 50,50, 100, 100, andand 200 nm. 200(b) nm.Wavelength (b) Wavelength of the SLR of the peak SLR peak corresponding corresponding to the test to the testlayer–Au layer–Au interface interface n = 𝑛1.45 for for = 1.45 as as a Figure 5. (a) Wavelength of the peak corresponding to the SLR at the test layer–Au interface as a function of the a function layer of the thickness. layer thickness.TheThe dashed dashed line lineisisonly onlya aguide guide to to thethe eye.eye. function of the refractive index for three values of the test layer thickness, namely, 50, 100, and 200 nm. (b)ToWavelength determine the of the the SLR peak corresponding dependence ofofthe sensitivity to the on test the layer–Auofinterface amount test for 𝑛 = material, the1.45 as a To function determine of thewaslayer dependence thickness. the sensitivity on the amount of test material, the same study repeated forThe dashed sensing line of layers is only a guide 100 and 200 tonmthe ineye. thickness, obtaining same study was repeated for sensing layers of 100 and 200 nm in thickness, obtaining sensitivities of 216 and 395 nm/RIU, respectively (see Figure 5a). These results show an sensitivities almost ofdependence To linear 216 and determine 395 nm/RIU, respectively the dependence of the sensorofsensitivity (see on the sensitivity Figure 5a). the amount on the thickness These results of material, of the test test show material, at an the almost same linear study least in dependence was repeated this range of the sensor forsmall of relatively sensing sensitivity layers of thicknesses. on 100 This the thickness and 200 enables of the test nm in thickness, to quantify material, the amountobtaining of at least in this material sensitivities range of 216ofon deposited relatively the395 and sensor small thicknesses. provided nm/RIU, This its refractive respectively enables index (see to quantify is known. Figure 5a). In Figure These the 5b,amount results show of the an material deposited on the sensor provided its refractive index is known. In almost linear dependence of the sensor sensitivity on the thickness of the test material, at Figure 5b, the wavelength least in this of the SLR range is depicted of relatively as thicknesses. small a function ofThis theenables thickness of the testthe to quantify layer. amountFor ofa thickness larger than 400 nm (560 nm/RIU), the sensitivity significantly departs material deposited on the sensor provided its refractive index is known. In Figure 5b, the from the wavelength of the SLR is depicted as a function of the thickness of the test layer. For a Nanomaterials 2021, 11, 1217 8 of 12 Nanomaterials 2021, 11, 1217 8 of 12 wavelength of the SLR is depicted as a function of the thickness of the test layer. For a thickness larger than linear dependence 400 nmto(560 tending nm/RIU),value a maximum the sensitivity significantly slightly greater departs than that of thefrom 600 the nm linear layer (613 nm/RIU). Therefore, the sensitivity for test layers thicker than about 1000 nm dependence tending to a maximum value slightly greater than that of the 600 nm layer can be(613 nm/RIU). considered Therefore, thickness the sensitivity for test layers thicker than about 1000 nm independent. can be considered thickness independent. The performance of the heterostructure as a detector of an analyte present in a water medium The performance was also tested. of the Forheterostructure this purpose, wateras a detector of an analyte was modelled present by a 600 in a water nm thick layer medium was also tested. For this purpose, water was modelled by a 600 nm thick layer with with a refractive index of 1.33 that also filled the bar-shaped trenches of the inverted hon- a refractive index of 1.33 that also filled the bar-shaped trenches of the inverted honeycomb eycomb lattice. The presence of an analyte was simulated by changing the refractive index lattice. The presence of an analyte was simulated by changing the refractive index of the of the medium to 1.34. To ensure high sensitivity to such small changes of the refractive medium to 1.34. To ensure high sensitivity to such small changes of the refractive index, index, the variation of the intensity of the reflection peak at a given wavelength was meas- the variation of the intensity of the reflection peak at a given wavelength was measured, ured, instead of the peak shift as in the previous cases, as the simplicity in detecting a instead of the peak shift as in the previous cases, as the simplicity in detecting a signal at a signal at a fixed wavelength is of key importance for biosensing. The figure of merit (𝐹𝑂𝑀) fixed wavelength is of key importance for biosensing. The figure of merit (FOM) [30] for [30] for this kind of measurements is defined as follows: this kind of measurements is defined as follows: 𝑑𝐼/𝑑𝑛 𝐹𝑂𝑀 =dI/dn , (2) FOM = 𝐼 , (2) Im where 𝑑𝐼 is the variation in the intensity of the reflection peak corresponding to the SLR where dI is theinterface at the air–Au variationasinathe intensity function of the of the reflection peak wavelength, corresponding 𝑛 is the to theofSLR refractive index the at the air–Au interface as a function of the wavelength, n is the refractive medium, and 𝐼 is the average of the intensities of the two spectra used in the calculation index of the medium, andofImthe as a function is the average ofThus, wavelength. the intensities the shift inofthetheposition two spectra of theused SLRin the as peak calculation 𝑛 varies is translated into a change of its intensity for a fixed wavelength. The computednvalues as a function of the wavelength. Thus, the shift in the position of the SLR peak as varies of is translated into a change of its intensity for a fixed wavelength. The the 𝐹𝑂𝑀 as a function of the wavelength are shown in Figure 6 together with the two computed values of the FOM as a function of the wavelength are shown in Figure 6 together with the two spectra spectra used in the calculations. The maximum value of the 𝐹𝑂𝑀 is 199 RIU−1 at a wave- used in the calculations. The maximum value of the FOM is 199 RIU−1 at a wavelength of length of 974 nm. This value can be compared with the sensitivity computed from Figure 974 nm. This value can be compared with the sensitivity computed from Figure 4 (613 nm 4 (613 nm RIU−1 for a test layer of 600 nm in thickness) divided by the full width at half RIU−1 for a test layer of 600 nm in thickness) divided by the full width at half maximum maximum of the peak, which is about 9 nm [31]. Finally, the minimum detectable change of the peak, which is about 9 nm [31]. Finally, the minimum detectable change in the in the refractive index can be estimated. For a fixed wavelength of 981 nm, for which the refractive index can be estimated. For a fixed wavelength of 981 nm, for which the FOM is 𝐹𝑂𝑀 is maximized, there is a change of 0.48 in the reflection signal induced by a change maximized, there is a change of 0.48 in the reflection signal induced by a change of 0.01 RIU of 0.01 RIU in the refractive index (see Figure 6). Considering a minimum detectable var- in the refractive index (see Figure 6). Considering a minimum detectable variation of 0.1 in iation of 0.1 in the signal intensity, the minimum change in the refractive index that could the signal intensity, the minimum change in the refractive index that could be detectable be detectable by the system would be 0.002 RIU. These results are a solid proof of concept by the system would be 0.002 RIU. These results are a solid proof of concept for further for further applications of these heterostructures for biosensing purposes. applications of these heterostructures for biosensing purposes. Figure 6. Reflection Reflectionpeak peakcorresponding correspondingtotothe theSLR SLRatat the test the layer–Au test layer–Au interface interface forfor a test layer a test of layer 600600 of nmnmin thickness with in thickness values with valuesof the refractive of the index refractive of 1.33 index (black of 1.33 dashed (black line)line) dashed andand 1.341.34 (red(red dashed line). dashed line). The The blue blue solid solid line line corresponds correspondsto tothe computed 𝐹𝑂𝑀 thecomputed FOMfromfromthese thesetwotwospectra spectrausing using Equation (2). Equation (2). In addition to the results of the simulations shown throughout this article, a pre nary experimental realization and optical characterization of the heterostructure were carried out. We aimed at replicating as much as possible the design studied in the s lations while keeping in mind the ease and potential scalability in the manufacturing cess. Hence, the parameters of the modelled system were followed, apart from Nanomaterials 2021, 11, 1217 9 of 12 nanostructured layer of SiO2 that was substituted by a PMMA layer with a thickness nm, giving an edge-to-edge distance between the inverted honeycomb lattice and th mirror of 190 nm instead of 200 nm, as was the case of the simulations. In this way PMMA layer 3.3. Experimental Realization of was etched following a much more direct and easy lithographic method the Heterostructure would have been necessary for the SiO2 layer, while preserving an almost equal valu In addition to the results of the simulations shown throughout this article, a prelim- the refractive index. The final metallization of the heterostructure yielded the Au inve inary experimental realization and optical characterization of the heterostructure were honeycomb lattice on top of the PMMA layer, but also an additional direct lattice on also carried out. bottom We aimed of theat bar-shaped replicatingtrenches as muchcarvedas possible through thethe design PMMA studied in the layer. This was the re simulations whilethe keeping in mind the ease and potential scalability in the manufacturing Au direct lattice was also included in the simulated model, because it is hard to a process. Hence, the parameters its formation of the amodelled following system wereprocess. simple manufacturing followed, apart from Nevertheless, thethe presence o nanostructured layer of SiO extra Au direct 2 lattice does not significantly affect the collective excitationsof that was substituted by a PMMA layer with a thickness of the inve 90 nm, giving an Au edge-to-edge layer, as provendistance between by the the inverted simulation honeycomb of the system without lattice and the the direct lattice show Au mirror of 190 nmFigureinstead S2 in of 200 nm, as was Supplementary the casewhile Materials, of theit simulations. In this way, hinders the excitation theLSR at 729 of the PMMA layer wasby etched following partially shieldinga much more direct the interaction and mode of this easy lithographic method than and its corresponding image at the would have beenror necessary (compare for thethe SiO2 layer, intensity while of this peakpreserving in the spectra aninalmost Figures equal 2 andvalue S2 in ofSupplemen the refractive index. The final Materials). metallization Figure 7a shows of an the heterostructure scanning yielded the electron microscopy Au inverted (SEM) top-view image o honeycomb lattice onheterostructure final top of the PMMA fromlayer, which but thealso an additional overall good quality direct of thelattice on the sample manufactured bottom of the bar-shaped be assessed. trenches carved through the PMMA layer. This was the reason the Au direct lattice was Thealso included optical in the simulated characterization of themodel, because itsample manufactured is hardby to avoid FTIR spectros its formation following showedathe simple manufacturing excitation process. Nevertheless, of the two collective modes associated the with presence the two of interfac the inverted this extra Au direct lattice doeshoneycomb lattice, as shown not significantly affect in theFigure 7b. Theexcitations collective peak corresponding of the to the inverted Au layer, excited in the by as proven air–Au interface was the simulation ofbroader the systemthan without that of the the SLR in the direct Au–PMMA i lattice shown in Figure S2 face. in Experimental Supplementary imperfections on theitmanufactured Materials, while sample together hinders the excitation of the with LSR the exist of other at 729 nm by partially excitations shielding in the vicinity the interaction of the of this modepeak andaround 765 nm may cause its corresponding imagetheir ove giving the at the mirror (compare rise intensity to a singleofbroad peak.in this peak Nevertheless, the spectra in theFigure measured 2 and spectrum Figure S2 and the sim in Supplementary Materials). Figure 7a shows an scanning electron microscopy (SEM) of this tion results were in good agreement, thus proving the experimental feasibility top-view image of erostructure. the final heterostructure from which the overall good quality of the manufactured sample can be assessed. Figure Figure 7. (a) SEM7. (a) SEM image ofimage of the manufactured the manufactured sample. (b)sample. (b) Experimental Experimental (redsimulated (red curve) and curve) and simulated data (black curve) cor- responding datato(black the reflection spectrum as a function curve) corresponding of the wavelength. to the reflection spectrumThe as asimulated functioncurve of theiswavelength. the same as inThe Figure 2 and is included here forcurve simulated the sake of same is the comparison. as in Figure 2 and is included here for the sake of comparison. The optical characterization of the manufactured sample by FTIR spectroscopy showed the excitation of the two collective modes associated with the two interfaces of the inverted honeycomb lattice, as shown in Figure 7b. The peak corresponding to the SLR excited in the air–Au interface was broader than that of the SLR in the Au–PMMA interface. Experimental imperfections on the manufactured sample together with the existence of other excitations in the vicinity of the peak around 765 nm may cause their overlap, giving rise to a single broad peak. Nevertheless, the measured spectrum and the simulation results were in good agreement, thus proving the experimental feasibility of this heterostructure. Nanomaterials 2021, 11, 1217 10 of 12 4. Discussion The results shown in this work evidence the high performance of the presented heterostructure based on an Au inverted honeycomb lattice for sensing purposes. Owing to the chosen lattice pitch and because local excitations are hindered by the three-fold symmetry of the honeycomb array, this heterostructure exhibits very intense and narrow reflection peaks corresponding to its collective excitations at wavelengths in the near infrared that are well suited for sensing applications. The intensity of these peaks is further enhanced by the underneath layers within the heterostructure, namely, the patterned SiO2 layer with the inverted honeycomb lattice, the SiO2 spacer, and the Au mirror that ensures total reflection. Among the collective excitations shown by this heterostructure, the most suitable for sensing the refractive index of the environment is that associated with the peak around 765 nm. This peak is particularly susceptible to changes in the nearby environment because it arises at the top interface of the sensing Au layer, thus yielding an out-of-plane near-field distribution that spreads up and far away from the interface. The sensitivity associated with the shift of this peak was assessed by changing the refractive index of a thin layer of a test material resting on top of the heterostructure. Under these conditions, simulations showed a sensitivity of 99 nm/RIU for a layer of 50 nm in thickness with a full width at half maximum of the peak of only 9 nm. These results become especially relevant when compared with those of previous studies in which the refractive index change was tested in layers of infinite thickness [32,33]. Moreover, a further large shift in the resonant wavelength was also found for an increasing thickness of the test layer. This yielded a rising sensitivity for thicker test layers, opening another possible application of the studied heterostructure as a sensor of the amount of test material, provided its refractive index is known. The sensitivity of the sensor tended to a maximum value greater than about 613 nm/RIU for a thickness larger than about 600 nm, which is a very remarkable performance for this kind of sensor. Finally, the sensitivity of the sensor detecting slight changes of the refractive index at a fixed wavelength of a thick layer of a water medium was also checked. The computed FOM at 974 nm was as high as 199 RIU−1 . This value constitutes a solid improvement over other plasmonic sensors operating in a similar fashion that exhibit values of the FOM about one order of magnitude smaller [30,33–35]. As an experimental proof of concept, the heterostructure was manufactured by a simple method based on EBL. The FTIR spectrophotometry measurements showed intense and relatively narrow excitations around the wavelengths predicted by the simulations. These results show that an easy experimental realization of this heterostructure is possible, in a similar way to in [22], but avoiding the final lift-off process. All in all, this work paves the way for the application of heterostructures based on inverted plasmonic lattices with three-fold symmetry to improve the performance of a wide range of plasmonic sensors and surface enhanced spectroscopies. Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/nano11051217/s1, Figure S1: Near-field distributions for the peaks at 729, 754, and 1789 nm; Figure S2: Reflection spectra for a heterostructure without the Au direct lattice, where the trenches are carved through the whole thickness of the SiO2 spacer, and for a heterostructure with a solid SiO2 spacer of 200 nm in thickness with no trenches carved through it. Author Contributions: Conceptualization, A.L., X.B. (Xavier Batlle), and A.F.R.; methodology, A.L. and J.R.-Á.; numerical simulations, J.R.-Á., L.G., and M.M.-C.; analysis of the results, J.R.-Á. and A.L.; design and manufacturing of the samples, A.G. and X.B. (Xavier Borrisé); writing—original draft preparation, J.R.-Á. and A.L.; writing—review and editing, A.L., X.B. (Xavier Batlle), and A.F.R. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Spanish MINECO, grant number PGC2018-097789-B-I00, and the European Union FEDER funds. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Nanomaterials 2021, 11, 1217 11 of 12 Data Availability Statement: Not applicable. Acknowledgments: We want to thank Pau Molet and A. Mihi, Institut de Ciència de Materials de Barcelona (ICMAB-CSIC), for the FTIR measurements. Conflicts of Interest: The authors declare no conflict of interest. References 1. Frey, H.G.; Bolwien, C.; Brandenburg, A.; Ros, R.; Anselmetti, D. Optimized apertureless optical near-field probes with 15 nm optical resolution. Nanotechnology 2006, 17, 3105–3110. [CrossRef] 2. Kim, J.H.; Song, K.B. Recent progress of nano-technology with NSOM. Micron 2007, 38, 409–426. [CrossRef] 3. Carrara, A.; Maccaferri, N.; Cerea, A.; Bozzola, A.; De Angelis, F.; Proietti Zaccaria, R.; Toma, A. Plasmon Hybridization in Compressible Metal–Insulator–Metal Nanocavities: An Optical Approach for Sensing Deep Sub-Wavelength Deformation. Adv. Opt. Mater. 2020, 8, 1–8. [CrossRef] 4. Kneipp, K.; Kneipp, H.; Bohr, H.G. Single-Molecule SERS Spectroscopy. In Surface-Enhanced Raman Scattering: Physics and Applications; Springer: Berlin/Heidelberg, Germany, 2006; pp. 261–277. ISBN 978-3-540-33566-5. 5. Vo-dinh, T.; Wang, H.; Scaffidi, J. Plasmonic nanoprobes for SERS biosensing and bioimaging. J. Biophotonics 2010, 102, 89–102. [CrossRef] 6. Cho, W.J.; Kim, Y.; Kim, J.K. Ultrahigh-Density Array of Silver Nanoclusters for SERS Substrate with High Sensitivity and Excellent Reproducibility. ACS Nano 2012, 6, 249–255. [CrossRef] [PubMed] 7. Shanmukh, S.; Jones, L.; Driskell, J.; Zhao, Y.; Dluhy, R.; Tripp, R.A. Rapid and Sensitive Detection of Respiratory Virus Molecular Signatures Using a Silver Nanorod Array SERS Substrate. Nano Lett. 2006, 6, 2630–2636. [CrossRef] [PubMed] 8. Article, R.; Dong, J.; Zhang, Z.; Zheng, H.; Sun, M.; Access, O. Recent Progress on Plasmon-Enhanced Fluorescence. Nanophotonics 2015, 4, 472–490. [CrossRef] 9. Bauch, M.; Toma, K.; Toma, M. Plasmon-Enhanced Fluorescence Biosensors: A Review. Plasmonics 2014, 9, 781–799. [CrossRef] [PubMed] 10. Chen, Q.; Liang, L.; Zheng, Q.L.; Zhang, Y.X.; Wen, L. On-chip readout plasmonic mid-IR gas sensor. Opto-Electron. Adv. 2020, 3, 1–13. [CrossRef] 11. Liu, N.; Weiss, T.; Mesch, M.; Langguth, L.; Eigenthaler, U.; Hirscher, M.; Sönnichsen, C.; Giessen, H. Planar metamaterial analogue of electromagnetically induced transparency for plasmonic sensing. Nano Lett. 2010, 10, 1103–1107. [CrossRef] 12. Hrtoň, M.; Konečná, A.; Horák, M.; Šikola, T.; Křápek, V. Plasmonic Antennas with Electric, Magnetic, and Electromagnetic Hot Spots Based on Babinet’s Principle. Phys. Rev. Appl. 2020, 13, 054045. [CrossRef] 13. Horák, M.; Křápek, V.; Hrtoň, M.; Konečná, A.; Ligmajer, F.; Stöger-Pollach, M.; Šamořil, T.; Paták, A.; Édes, Z.; Metelka, O.; et al. Limits of Babinet’s principle for solid and hollow plasmonic antennas. Sci. Rep. 2019, 9, 4004. [CrossRef] [PubMed] 14. Falcone, F.; Lopetegi, T.; Laso, M.A.G.; Baena, J.D.; Bonache, J.; Beruete, M.; Marques, R.; Martín, F.; Sorolla, M. Babinet principle applied to the design of metasurfaces and metamaterials. Phys. Rev. Lett. 2004, 93, 2–5. [CrossRef] [PubMed] 15. Qin, F.; Ding, L.; Zhang, L.; Monticone, F.; Chum, C.C.; Deng, J.; Mei, S.; Li, Y.; Teng, J.; Hong, M.; et al. Hybrid bilayer plasmonic metasurface efficiently manipulates visible light. Sci. Adv. 2016, 2, 1–9. [CrossRef] [PubMed] 16. Huck, C.; Vogt, J.; Sendner, M.; Hengstler, D.; Neubrech, F.; Pucci, A. Plasmonic Enhancement of Infrared Vibrational Signals: Nanoslits versus Nanorods. ACS Photonics 2015, 2, 1489–1497. [CrossRef] 17. Stewart, M.E.; Mack, N.H.; Malyarchuk, V.; Soares, J.A.N.T.; Lee, T.W.; Gray, S.K.; Nuzzo, R.G.; Rogers, J.A. Quantitative multispectral biosensing and 1D imaging using quasi-3D plasmonic crystals. Proc. Natl. Acad. Sci. USA 2006, 103, 17143–17148. [CrossRef] 18. Kannegulla, A.; Liu, Y.; Wu, B.; Cheng, L. Plasmonic Open-Ring Nanoarrays for Broadband Fluorescence Enhancement and Ultrasensitive DNA Detection. J. Phys. Chem. C 2018, 122, 770–776. [CrossRef] 19. Ni, X.; Ishii, S.; Kildishev, A.V.; Shalaev, V.M. Ultra-thin, planar, Babinet-inverted plasmonic metalenses. Light Sci. Appl. 2013, 2, e72. [CrossRef] 20. Conde-Rubio, A.; Fraile Rodríguez, A.; Espinha, A.; Mihi, A.; Pérez-Murano, F.; Batlle, X.; Labarta, A. Geometric frustration in ordered lattices of plasmonic nanoelements. Sci. Rep. 2019, 9, 1–10. [CrossRef] 21. Conde-Rubio, A.; Rodríguez, A.F.; Borrisé, X.; Perez-Murano, F.; Batlle, X.; Labarta, A. Geometric frustration in a hexagonal lattice of plasmonic nanoelements. Opt. Express 2018, 26, 20211. [CrossRef] 22. Chen, Y.; Dai, J.; Yan, M.; Qiu, M. Metal-insulator-metal plasmonic absorbers: Influence of lattice. Opt. Express 2014, 22, 30807. [CrossRef] 23. Humphrey, A.D.; Barnes, W.L. Plasmonic surface lattice resonances on arrays of different lattice symmetry. Phys. Rev. B Condens. Matter Mater. Phys. 2014, 90, 1–8. [CrossRef] 24. Lumerical Inc. Innovative Photonic Design Tools. Available online: https://www.lumerical.com/ (accessed on 11 January 2021). 25. Palik, E.D. Handbook of Optical Constants of Solids; Palik, E.D., Ed.; Academic Press: Washington, DC, USA, 1985; ISBN 978-0-08-054721-3. 26. Johnson, P.B.; Christy, R.W. Optical Constant of the Nobel Metals. Phys. Rev. B 1972, 6, 4370–4379. [CrossRef] 27. Tok, R.U.; Şendur, K. Engineering the broadband spectrum of close-packed plasmonic honeycomb array surfaces. J. Quant. Spectrosc. Radiat. Transf. 2013, 120, 70–80. [CrossRef] Nanomaterials 2021, 11, 1217 12 of 12 28. Chen, Y.; Dai, J.; Yan, M.; Qiu, M. Honeycomb-lattice plasmonic absorbers at NIR: Anomalous high-order resonance. Opt. Express 2013, 21, 20873. [CrossRef] [PubMed] 29. Zhang, T.; He, J.; Xie, W.; Sun, C. Plasmonic properties of a honeycomb structure formed by metallic nanoparticles. Phys. E Low-Dimens. Syst. Nanostructures 2020, 118, 113901. [CrossRef] 30. Hentschel, M.; Weiss, T.; Bagheri, S.; Giessen, H. Babinet to the half: Coupling of solid and inverse plasmonic structures. Nano Lett. 2013, 13, 4428–4433. [CrossRef] 31. Becker, J.; Trügler, A.; Jakab, A.; Hohenester, U.; Sönnichsen, C. The optimal aspect ratio of gold nanorods for plasmonic bio-sensing. Plasmonics 2010, 5, 161–167. [CrossRef] 32. Sherry, L.J.; Chang, S.-H.; Schatz, G.C.; Duyne, R.P.V. Localized surface plasmon resonance of single silver nanocubes. Nano Lett. 2005, 5, 2034–2038. [CrossRef] 33. Agharazy Dormeny, A.; Abedini Sohi, P.; Kahrizi, M. Design and simulation of a refractive index sensor based on SPR and LSPR using gold nanostructures. Results Phys. 2020, 16, 102869. [CrossRef] 34. Wang, X.; Zhu, J.; Wen, X.; Wu, X.; Wu, Y.; Su, Y.; Tong, H.; Qi, Y.; Yang, H. Wide range refractive index sensor based on a coupled structure of Au nanocubes and Au film. Opt. Mater. Express 2019, 9, 3079. [CrossRef] 35. Hicks, E.M.; Zhang, X.; Zou, S.; Lyandres, O.; Spears, K.G.; Schatz, G.C.; Van Duyne, R.P. Plasmonic properties of film over nanowell surfaces fabricated by nanosphere lithography. J. Phys. Chem. B 2005, 109, 22351–22358. [CrossRef] [PubMed]