Strong Coupling between Tamm and Surface Plasmons for Advanced Optical Bio-Sensing

: The total internal reﬂection ellipsometry method was used to analyse the angular spectra of the hybrid Tamm and surface plasmon modes and to compare their results with those obtained using the conventional single SPR method. As such type of measurement is quite common in commercial SPR devices, more detailed attention was paid to the analysis of the p-polarization reﬂection intensity dependence. The conducted study showed that the presence of strong coupling in the hybrid plasmonic modes increases the sensitivity of the plasmonic-based sensors due to the reduced losses in the metal layer. The experimental results and analysis of the optical responses of three di ﬀ erent plasmonic-based samples indicated that the optimized Tamm plasmons ∆ R p ( TP ) and optimized surface plasmons ∆ R p ( SP ) samples produce a response that is about ﬁve and six times greater than the conventional surface plasmon resonance ∆ R p ( SPR ) in angular spectra. The sensitivity of the refractive index unit of the spectroscopic measurements for the optimized Tamm plasmon samples was 1.5 times higher than for conventional SPR, while for wavelength scanning, the SPR overcame the optimized TP by 1.5 times. by ion beam sputtering. Afterwards, thin gold (~40 nm) layers were deposited on both photonic crystal structures by magnetron sputtering. The SPR sample consisted of a gold ﬁlm having a thickness of about 45 nm and about 2 nm underlayer of Ti for better adhesion. The total internal reﬂection ellipsometry (TIRE) experiments were then conducted using a dual rotating compensator ellipsometer RC2 (J.A. Woollam Co., Inc., Lincoln, NE, USA). The spectroscopic ellipsometry experiments were carried out in the 400–1000 nm spectral range. For all the samples with the supported hybrid plasmonic modes and the SPR, the experiments were conducted in a TIRE conﬁguration. These used a half-cylindrical BK7 glass prism in the range of angle of incidence (AOI) 60 ◦ –75 ◦ connected via a refractive index matching ﬂuid with investigated samples To compensate for the light beam e ﬀ ect due to the semi-cylindrical prism, lenses with a 40 mm focal length were used for the pre-focusing of the incoming and reﬂected light to the detector.


Introduction
Sensors based on optical signal interrogation are widely used for the detection of biochemical interactions at solid-liquid interfaces, for the monitoring of various chemicals adsorbed on surfaces [1,2] and for the evaluation of the refractive indexes of liquids [3]. For this purpose, surface plasmon resonance (SPR) optical sensors are most widely used [4,5]. Surface plasmon polariton (SPP) based optical sensors are able to detect the kinetics of various biochemical interactions such as antibody-antigen binding [4] and to determine, for instance, the adsorbed surface mass in monolayers of proteins [5]. The non-destructive, label-free, high sensitivity and real time monitoring ability of this optical signal makes this plasmonic based sensing technique an extremely attractive tool for the analysis of a wide range of surface science areas [6]. The high sensitivity of the optical signal registration is achieved due to the strong localization of the electric field at the metal/dielectric interface when resonance conditions for such light-matter interaction are satisfied [7]. The SPP is a non-radiative electromagnetic waves, which can be excited by light through a glass prism or grating coupler which are necessary to match the in-plane wave vectors of the incident light and the plasmons in the metal layer. This rather strong light-matter interaction leads the properties of the dispersion relation (angular frequency (ω) and in-plane wavevector (k)) of the surface plasmon polaritons which are sensitive to the any perturbation of the refractive index changes on the SPP supported interfaces. However, the rather broad width of such SPP resonances, caused by the large absorption and scattering losses in the metal layer, limits the further improvement of the sensitivity of this type of optical sensor. These losses in the metals are determined by the imaginary part of their relative permittivity [8]. They shorten the The excitation of these hybrid TPP-SPP plasmonic modes require conditions of total internal reflection. These states are also polarization sensitive. Detailed analysis of the polarization properties in such TIR setups can be achieved by using total internal reflection ellipsometry (TIRE) [24]. TIRE is a technique combining spectroscopic ellipsometry and the analysis of surface plasmon electromagnetic waves. The sensitivity of this method is higher compared to that of conventional ellipsometry or SPR [25,26]. In fact, TIRE utilizes the analytical power of ellipsometry and increases its sensitivity by introducing the SPR effect into the operation scheme of the ellipsometer.
Studies employing the strong coupling effect between the TPP and SPP in the spectroscopic mode have been performed before, where changes in the resonant wavelength were registered [22]. However, most commercially available plasmonic sensors use the p-polarized intensity at fixed wavelengths and scanning angles of incidence. This study demonstrates that increased sensitivity to the p-polarization reflection due to the narrowing of the resonance width caused by strong coupling between TPP and SPP can be achieved not only for the wavelength, but also for the angle of incidence scanning.

Materials and Methods
Three samples were investigated. These consisted of two structures supporting optimized hybrid Tamm-surface plasmons modes and a commercially available Xantec surface plasmon resonance (SPR) chip. The first two slightly different photonic structures of hybrid plasmonic modes were optimized for enhanced sensitivity of these particular excitation components (Tamm plasmons or surface plasmons). For the optimized surface plasmon (SP) components in their hybrid modes, distributed Bragg gratings consisting of 6 bilayers of~120 nm TiO 2 /~200 nm SiO 2 and a 30 nm TiO 2 layer on top were formed on BK7 glass substrates. For the optimized Tamm plasmon component, the distributed Bragg grating consisted of five bilayers of~120 nm TiO 2 /~200 nm SiO 2 formed on a BK7 glass substrate. These Bragg gratings were created by ion beam sputtering. Afterwards, thin gold (~40 nm) layers were deposited on both photonic crystal structures by magnetron sputtering. The SPR sample consisted of a gold film having a thickness of about 45 nm and about 2 nm underlayer of Ti for better adhesion.
The total internal reflection ellipsometry (TIRE) experiments were then conducted using a dual rotating compensator ellipsometer RC2 (J.A. Woollam Co., Inc., Lincoln, NE, USA). The spectroscopic ellipsometry experiments were carried out in the 400-1000 nm spectral range. For all the samples with the supported hybrid plasmonic modes and the SPR, the experiments were conducted in a TIRE configuration. These used a half-cylindrical BK7 glass prism in the range of angle of incidence (AOI) 60 • -75 • connected via a refractive index matching fluid with investigated samples (Figure 1). To compensate for the light beam focusing effect due to the semi-cylindrical prism, lenses with a 40 mm focal length were used for the pre-focusing of the incoming and reflected light to the detector.
Coatings 2020, 10, x FOR PEER REVIEW 3 of 11 electromagnetic waves. The sensitivity of this method is higher compared to that of conventional ellipsometry or SPR [25,26]. In fact, TIRE utilizes the analytical power of ellipsometry and increases its sensitivity by introducing the SPR effect into the operation scheme of the ellipsometer. Studies employing the strong coupling effect between the TPP and SPP in the spectroscopic mode have been performed before, where changes in the resonant wavelength were registered [22]. However, most commercially available plasmonic sensors use the p-polarized intensity at fixed wavelengths and scanning angles of incidence. This study demonstrates that increased sensitivity to the p-polarization reflection due to the narrowing of the resonance width caused by strong coupling between TPP and SPP can be achieved not only for the wavelength, but also for the angle of incidence scanning.

Materials and Methods
Three samples were investigated. These consisted of two structures supporting optimized hybrid Tamm-surface plasmons modes and a commercially available Xantec surface plasmon resonance (SPR) chip. The first two slightly different photonic structures of hybrid plasmonic modes were optimized for enhanced sensitivity of these particular excitation components (Tamm plasmons or surface plasmons). For the optimized surface plasmon (SP) components in their hybrid modes, distributed Bragg gratings consisting of 6 bilayers of ~120 nm TiO2/~200 nm SiO2 and a 30 nm TiO2 layer on top were formed on BK7 glass substrates. For the optimized Tamm plasmon component, the distributed Bragg grating consisted of five bilayers of ~120 nm TiO2/~200 nm SiO2 formed on a BK7 glass substrate. These Bragg gratings were created by ion beam sputtering. Afterwards, thin gold (~40 nm) layers were deposited on both photonic crystal structures by magnetron sputtering. The SPR sample consisted of a gold film having a thickness of about 45 nm and about 2 nm underlayer of Ti for better adhesion.
The total internal reflection ellipsometry (TIRE) experiments were then conducted using a dual rotating compensator ellipsometer RC2 (J.A. Woollam Co., Inc., Lincoln, NE, USA). The spectroscopic ellipsometry experiments were carried out in the 400-1000 nm spectral range. For all the samples with the supported hybrid plasmonic modes and the SPR, the experiments were conducted in a TIRE configuration. These used a half-cylindrical BK7 glass prism in the range of angle of incidence (AOI) 60°-75° connected via a refractive index matching fluid with investigated samples (Figure 1). To compensate for the light beam focusing effect due to the semi-cylindrical prism, lenses with a 40 mm focal length were used for the pre-focusing of the incoming and reflected light to the detector. A liquid handling system with a custom-built Teflon chamber was used in which the surfaces of all the samples were placed. This chamber was filled with deionized water, which was then changed to deionized water/ethanol (50%/50%) mixture, whose refractive index is higher than that of pure deionized water. The measured experimental data of the ellipsometric parameters Ψ(λ) and Δ(λ) were then expressed as the p-polarized intensity, using the data acquisition software CompleteEase (J.A. Woollam Co., Inc.), and were then presented as the p-intensity map dependence on the wavelength (λ) vs. angle of incidence (θ°) (Figures 2 and 3). These p-polarized intensity maps, in fact, represent the dispersion relations of the investigated plasmonic excitations. Furthermore, a fixed resonant wavelengths λ = 800 nm (SP), λ = 715 nm (TP), and λ = 820 nm (Hybrid TP-SP) were chosen to demonstrate the optimized sensitivity properties in the angular spectra of the p-intensity polarizations for SPR and Tamm plasmon polaritons and hybrid TPP-SPP mode ( Figure 4). A liquid handling system with a custom-built Teflon chamber was used in which the surfaces of all the samples were placed. This chamber was filled with deionized water, which was then changed to deionized water/ethanol (50%/50%) mixture, whose refractive index is higher than that of pure deionized water. The measured experimental data of the ellipsometric parameters Ψ(λ) and ∆(λ) were then expressed as the p-polarized intensity, using the data acquisition software CompleteEase (J.A. Woollam Co., Inc.), and were then presented as the p-intensity map dependence on the wavelength (λ) vs. angle of incidence (θ • ) (Figures 2 and 3). These p-polarized intensity maps, in fact, represent the dispersion relations of the investigated plasmonic excitations. Furthermore, a fixed resonant wavelengths λ = 800 nm (SP), λ = 715 nm (TP), and λ = 820 nm (Hybrid TP-SP) were chosen to demonstrate the optimized sensitivity properties in the angular spectra of the p-intensity polarizations for SPR and Tamm plasmon polaritons and hybrid TPP-SPP mode ( Figure 4). A liquid handling system with a custom-built Teflon chamber was used in which the surfaces of all the samples were placed. This chamber was filled with deionized water, which was then changed to deionized water/ethanol (50%/50%) mixture, whose refractive index is higher than that of pure deionized water. The measured experimental data of the ellipsometric parameters Ψ(λ) and Δ(λ) were then expressed as the p-polarized intensity, using the data acquisition software CompleteEase (J.A. Woollam Co., Inc.), and were then presented as the p-intensity map dependence on the wavelength (λ) vs. angle of incidence (θ°) (Figures 2 and 3). These p-polarized intensity maps, in fact, represent the dispersion relations of the investigated plasmonic excitations. Furthermore, a fixed resonant wavelengths λ = 800 nm (SP), λ = 715 nm (TP), and λ = 820 nm (Hybrid TP-SP) were chosen to demonstrate the optimized sensitivity properties in the angular spectra of the p-intensity polarizations for SPR and Tamm plasmon polaritons and hybrid TPP-SPP mode ( Figure 4).

Results and Discussions
It has been shown before that by applying spectroscopic ellipsometry in its total internal reflection configuration and the so-called total internal reflection ellipsometry (TIRE), a higher sensitivity to the refractive index and the attached surface mass can be achieved compared with that produced by standard commercially available surface plasmon resonance biosensors [27]. This improved sensitivity was obtained mainly due to the ability to directly measure the phase difference between s-and p-polarized reflected waves in the vicinity of the SPR, where the phases of the reflected waves change drastically [25]. Moreover, TIRE has a better sensitivity to even the ellipsometric parameter Ψ than the standard p-polarized intensity measurement which is obtained by the SPR [24]. The TIRE with a hybrid plasmonic mode of Tamm and surface plasmons were applied for saturated gas [28], graphene influence to the strong coupling [29] and biosensing [22], where the strong coupling effect between these two excitations was employed. However, commercially available SPR biosensors normally use simpler optical schemes, where only the ppolarized intensity of the SPR phenomenon is registered and which use only a single wavelength as the light source. In this study, the TIRE method was used for the analysis of the optical properties and sensitivity features of the Tamm-surface plasmon hybrid modes. Their p-polarized intensity for a single wavelength and the p-polarized signal intensity of the hybrid Tamm-surface plasmon mode

Results and Discussions
It has been shown before that by applying spectroscopic ellipsometry in its total internal reflection configuration and the so-called total internal reflection ellipsometry (TIRE), a higher sensitivity to the refractive index and the attached surface mass can be achieved compared with that produced by standard commercially available surface plasmon resonance biosensors [27]. This improved sensitivity was obtained mainly due to the ability to directly measure the phase difference between s-and p-polarized reflected waves in the vicinity of the SPR, where the phases of the reflected waves change drastically [25]. Moreover, TIRE has a better sensitivity to even the ellipsometric parameter Ψ than the standard p-polarized intensity measurement which is obtained by the SPR [24]. The TIRE with a hybrid plasmonic mode of Tamm and surface plasmons were applied for saturated gas [28], graphene influence to the strong coupling [29] and biosensing [22], where the strong coupling effect between these two excitations was employed. However, commercially available SPR biosensors normally use simpler optical schemes, where only the p-polarized intensity of the SPR phenomenon is registered and which use only a single wavelength as the light source. In this study, the TIRE method was used for the analysis of the optical properties and sensitivity features of the Tamm-surface plasmon hybrid modes. Their p-polarized intensity for a single wavelength and the p-polarized signal intensity of the hybrid Tamm-surface plasmon mode were then compared with the single SPR used in standard optical biosensors. As mentioned before, three samples were used: the single thin gold layer for the SPR and the two photonic crystals (PC)/Au structures with optimized sensitivity for Tamm plasmons and for surface plasmons in their hybrid plasmonic modes, respectively.
In order to determine the wavelength at which the resonant effect of the plasmonic excitation was most efficient, the dispersion relations for all three samples were first measured experimentally and presented as maps of the p-polarized intensity of wavelength dependence on the angle of incidence (AOI). Figure 2A represents the single SPR excitation and Figure 2B,C, the hybrid Tamm-surface plasmon modes for slightly different multi-layered structures. As can be seen, the dispersion relation of the single SPR lies in the 650-700 nm spectral range for angles of incidence 70-75 • , while for the hybrid plasmonic modes, the SPP branch moves down to the longer wavelengths and lies at about 800 nm for the same AOI. Such modification of dispersion relation for SPP branch caused by strong coupling between Tamm optical states and surface plasmon polariton in the hybrid plasmonic mode.
Such behavior is the result of the strong coupling effect between the Tamm plasmons and the surface plasmons, together with a narrowing of the excitation line of the SP branch in this AOI range, compared with that of a single SPR. The narrowing of the dispersion relation line indicates the lower losses of such SP in its hybrid mode. Attention should also be paid to the Tamm plasmon branch in its hybrid plasmonic mode which lies between 600-800 nm in a wide range angle of incidence and has an even narrower width than the SP branch. The cross-section at the resonant wavelength is marked with dashed lines in each spectra of the dispersion relation. In Figure 2, the presented dispersion relations correspond to the optical responses to the ambient medium of deionized water, while in Figure 3, the dispersion relations for all three samples are presented when responding to a mixture of deionized water (50%) and ethanol (50%) which changes the refractive index of the ambient medium. The cross-sections (dash lines) were made at the same resonant wavelength as with the pure deionized water.
The cross-sections at a particular resonant wavelength give the angular spectra of the p-polarized intensity of the corresponding excitations ( Figure 4A-C). The black curves correspond to the optical responses in the vicinity of the plasmonic excitations for deionized water and the red curves show the changes in the optical response of the system when the refractive index changes due to the mixture of deionized water (50%) and ethanol (50%). As can be seen from Figure 4B, the sample optimized for the Tamm plasmons manifested itself at the AOI = 62.5 • and λ TP = 715 nm for the deionized water and AOI = 64.5 • for the mixture of water/ethanol (50%/50%), respectively. This corresponded to a shift of 2 • degrees for the refractive index change δn (λ=715 nm) = 1.3442 − 1.3298 = 0.0144. It should be noted that the full width at half maximum (FWHM) for the Tamm plasmon excitation in the hybrid mode was 0.88 • . In the case of the multilayer sample for optimized surface plasmons in their hybrid plasmonic modes, the angular shift was 1.7 • , i.e., from 66.4 • up to 68.1 • ( Figure 4C) and the FWHM = 1.56 • to λ SP = 820 nm, δn (λ=820 nm) = 1.3425 − 1.32979 = 0.0146. The same measurements were performed with the commercially available SPR chip consisted of a thin (~45 nm) gold layer without any additional multi-layered structures. The SPR shift to the larger AOI due to changes of the refractive index of the ambient (liquid) was 3.5 • , i.e., from 64.9 • up to 68.5 • ( Figure 4A) and the FWHM = 3.5 • for the λ SPR = 800 nm, δn (λ=800 nm) = 1.3428 − 1.3282 = 0.016. However, a much wider FWHM was registered for the SPR. This was 3.6 • , which indicated significantly higher metal losses for these plasmonic excitations compared with the Tamm plasmons or the surface plasmons in their hybrid plasmonic modes. The narrowing of the width of the hybrid plasmonic resonances due to such reduced losses leads to the increased values of the p-intensity reflectance for the TP δR TP ≈ 0.75 and SP δR TP ≈ 0.87 compared with the single SPR, which indicated that this was δR TP ≈ 0.16. As a result, these numbers led to a corresponding sensitivity of the refractive index unit of As variable angle spectroscopic ellipsometry gives the possibility of analysing the optical response by using a full analysis of the polarized light, the p-polarized intensity spectra dependence on the wavelength was also presented ( Figure 5A,B). It should be noted that focusing on only the p-polarized light is related to the studied plasmonic excitations, which are generated only in this state of light polarisation. As was noted above [22], the surface plasmon polariton component in its hybrid plasmonic mode produces a higher sensitivity than was achievable when using the single SPR for the bovine serum albumin protein covalent immobilization.
Coatings 2020, 10, x FOR PEER REVIEW 8 of 11 optimized surface plasmons ∆ ( ) samples produced five and six times better responses than the conventional surface plasmon resonance ∆ ( ) . Meanwhile, the conventional SPR overcame the optimized hybrid plasmonic modes for the registration of changes of the angle of incidence. These were by TP = 1.75 and SP = 2 times, respectively. As variable angle spectroscopic ellipsometry gives the possibility of analysing the optical response by using a full analysis of the polarized light, the p-polarized intensity spectra dependence on the wavelength was also presented ( Figure 5A,B). It should be noted that focusing on only the ppolarized light is related to the studied plasmonic excitations, which are generated only in this state of light polarisation. As was noted above [22], the surface plasmon polariton component in its hybrid plasmonic mode produces a higher sensitivity than was achievable when using the single SPR for the bovine serum albumin protein covalent immobilization.

Figure 5
Experimental spectra of p-intensity dependence on the wavelength for single SPR (A) and optimized Tamm plasmon (TP) (B) samples before (red curves) and after (black curves) deionized water changed to solution of deionized water/ethanol (50%/50%).
Thus, this study was focused on achieving a higher sensitivity for the Tamm plasmon component in its hybrid plasmonic mode. The AOI was optimized for the TP and the conventional SPR samples were chosen so that the highest sensitivity would be achieved for both. The single SPR of the commercially available chip manifested itself as producing a dip in the p-polarized intensity spectra at 680 nm and AOI = 70°. After changing the deionized water with a mixture of water/ethanol (50%/50%), the resonance dip red shifted up to 754 nm ( Figure 5A). For the Tamm plasmon optimized structure, the plasmonic dip was at the 663 nm for AOI = 66° and after increasing the refractive index of liquid, the dip moved to a longer wavelength, i.e., up to 712 nm ( Figure 5B). After changing the deionized water with the water/ethanol mixture, the increased refractive index of the liquid shifted in both plasmonic excitations to longer wavelengths. For the single SPR, the red shift was 72 nm and this corresponded to the p-intensity changes in δ = 0.36 . For the Tamm plasmon optimized sample, the spectral changes were smaller ~50 nm. However, such a red shift induced a change in the relative p-intensity of about  0.54. For both resonances, the differences in the refractive index of the liquid was the same δn = 0.015. Thus, as can be seen from Figure 5B, the red shift of the Tamm plasmon component was significant and the value was evaluated from the simulation of the red shift with the smaller changes of the refractive index of liquid, but taking into account the experimental red shift as a reference. The evaluated sensitivity to the refractive indexes were  Thus, this study was focused on achieving a higher sensitivity for the Tamm plasmon component in its hybrid plasmonic mode. The AOI was optimized for the TP and the conventional SPR samples were chosen so that the highest sensitivity would be achieved for both. The single SPR of the commercially available chip manifested itself as producing a dip in the p-polarized intensity spectra at 680 nm and AOI = 70 • . After changing the deionized water with a mixture of water/ethanol (50%/50%), the resonance dip red shifted up to 754 nm ( Figure 5A). For the Tamm plasmon optimized structure, the plasmonic dip was at the 663 nm for AOI = 66 • and after increasing the refractive index of liquid, the dip moved to a longer wavelength, i.e., up to 712 nm ( Figure 5B). After changing the deionized water with the water/ethanol mixture, the increased refractive index of the liquid shifted in both plasmonic excitations to longer wavelengths. For the single SPR, the red shift was 72 nm and this corresponded to the p-intensity changes in δR p = 0.36. For the Tamm plasmon optimized sample, the spectral changes were smaller~50 nm. However, such a red shift induced a change in the relative p-intensity of about δR p ≈ 0.54. For both resonances, the differences in the refractive index of the liquid was the same δn = 0.015. Thus, as can be seen from Figure 5B, the red shift of the Tamm plasmon component was significant and the δR p value was evaluated from the simulation of the red shift with the smaller changes of the refractive index of liquid, but taking into account the experimental red shift as a reference. The evaluated sensitivity to the refractive indexes were = 74 0.015 = 4933 nm/RIU. These evaluations led to the following estimations of the sensitivity of the hybrid plasmonic modes and the conventional SPR to the RIU. For the optimized Tamm plasmon sample, the p-intensity measurement was 1.5 times higher than for the conventional SPR. For the wavelength scanning, however, the SPR overcame the optimized TP by 1.5 times.

Conclusions
Summarizing, the TIRE method was used for the analysis of the angular spectra of the hybrid Tamm and surface plasmon modes and their comparison with those produced by conventional single SPRs. A more detailed analysis was made of the p-reflection intensity dependence on the AOI because such types of measurements are very common in commercial SPR devices. The conducted study showed that the presence of strong coupling in the hybrid plasmonic modes increases the p-polarized intensity sensitivity of such excitations due to the reduced losses in the metal layers. It should be noted that conventional SPR shows better sensitivity for AOI and wavelength scanning, while the registration of p-intensity changes gives better values for the optimized plasmonic modes. The angular spectra of the hybrid plasmonic modes with the strong coupling effect was analyzed for the first time in our knowledge. The strong coupling effect in the hybrid plasmonic modes allow one to control and tune the dispersion relation of the plasmonic excitations for corresponding purposes. Since modern coating technologies allow the production of nanometric structures with high precision, this produces the possibility of designing nanophotonic devices with advanced properties.
Funding: This research received no external funding.