Bloch Long-Range Surface Plasmon Polaritons on Metal Stripe

Feb 7, 2017 - ... near the end of a waveguide to diffract light upward or by polishing the output facet and allowing the mode to radiate into a free-s...
0 downloads 0 Views 6MB Size
Download PDF
Article pubs.acs.org/journal/apchd5

Bloch Long-Range Surface Plasmon Polaritons on Metal Stripe Waveguides on a Multilayer Substrate Norman R. Fong,†,‡ Matteo Menotti,§ Ewa Lisicka-Skrzek,†,‡ Howard Northfield,†,‡ Anthony Olivieri,†,‡ Niall Tait,∥ Marco Liscidini,§ and Pierre Berini*,†,‡,⊥ †

School of Electrical Engineering and Computer Science, University of Ottawa, 800 King Edward Avenue, Ottawa, Ontario K1N 6N5, Canada ‡ Centre for Research in Photonics, University of Ottawa, 25 Templeton Street, Ottawa, Ontario K1N 6N5, Canada § Department of Physics, University of Pavia, Via Bassi 6, Pavia, Italy ∥ Department of Electronics, Carleton University, 1125 Colonel By Drive, Ottawa, Ontario K1S 5B6, Canada ⊥ Department of Physics, University of Ottawa, 151 Louis Pasteur, Ottawa, Ontario K1N 6N5, Canada ABSTRACT: We propose and demonstrate a thin Au stripe on a truncated 1D dielectric photonic crystal covered with Cytop as a waveguide for Bloch long-range surface plasmon polaritons. High-quality mode outputs were observed and a mode power attenuation of 12−17 dB/mm measured at λ0 = 1310 nm for propagation in the plane of the truncated photonic crystal and within its stopband. The truncated 1D photonic crystal advantageously enables the use of a large range of materials for the substrate, breaking free from the constraint of material symmetry to support long-range plasmons. An input grating coupler implemented as a periodic array of nanoscale Au ridges on a Au stripe was used to excite the mode via perpendicularly incident p-polarized light. The output was provided by adding a second grating coupler near the end of a waveguide to diffract light upward or by polishing the output facet and allowing the mode to radiate into a free-space beam. Advantageously, grating coupling eliminates the need for highquality end facets, and optical alignment is simplified. Given its practicality, the structure proposed is of strong interest for biosensing. KEYWORDS: plasmon, Bloch, long-range, surface, biosensor, multilayer, photonic crystal

L

speeds, namely they must have similar effective indices. This condition is automatically satisfied when one works with symmetric structures, but this can be a serious constraint for some applications. For instance, if one is interested in optical sensing in an aqueous environment, then choices are somewhat limited to working, for example, with a thin metal layer on an optically noninvasive free-standing membrane6 or on a thick free-standing membrane that provides index matching in an effective sense7 or finding a solid substrate having a refractive index similar to that of water over the wavelength range of interest.8−11 While there are materials that satisfy this latter condition (e.g., Teflon or Cytop), their use can pose limitations to device fabrication. A few years ago, Konopsky et al. suggested that a onedimensional photonic crystal (1DPC) can be used on one of the sides of a thin metal layer to mimic, over a limited wavelength range, the optical properties of the medium on the other side.12 This strategy allows one to work with a much wider class of dielectric materials. Indeed, given two materials with a sufficient refractive index contrast, the existence of the LRSPP can be ensured by a careful choice of the geometrical parameters of the multilayer.13−15 The modes supported on

ight confinement near the surface of a photonic structure is appealing for the realization of optical sensors, in which one relies on the interaction between the electromagnetic field and an adlayer forming on the surface as analyte binds thereon. In this respect, surface plasmon polaritons (SPPs), which are propagating modes that exist at the interface between a metal and a dielectric medium, are probably the most studied and utilized.1 This is because they are characterized by a small mode area and large surface fields, which make them very sensitive to changes in the refractive index near the metallic surface.2 In addition, these properties can be exploited to increase the intensity of the Raman signal of a molecule or the efficiency of a fluorescent marker attached to a target molecule located near the surface.3,4 Despite these attractive features, the use of SPPs is limited by propagation losses associated with the absorption of light in the metal. An approach to mitigate losses in SPPs makes use of symmetric structures based on a thin metal layer to support long-range SPPs (LRSPPs). These are plasmonic modes arising from the coupling between the SPPs localized at the two surfaces of the metallic layer. LRSPPs are characterized by propagation distances that can be up to 2 orders of magnitude larger than typical SPPs, albeit with lower field confinement.5 The existence of the LRSPP requires that the SPPs at the upper and lower surfaces of a thin metal layer propagate at similar © XXXX American Chemical Society

Received: November 21, 2016

A

DOI: 10.1021/acsphotonics.6b00930 ACS Photonics XXXX, XXX, XXX−XXX

ACS Photonics

Article

0.448 + i9.228) stripe, on a 0.5 nm thick layer of Cr (nCr = 4.467 + i4.519) for adhesion purposes, on a truncated 1DPC implemented as a SiO2/Ta2O5 multilayer stack composed of N = 12 periods of a unit cell with dSiO2 = 625 nm (nSiO2 = 1.447) and dTa2O5 = 209 nm (nTa2O5 = 2.069). The thickness of each layer was chosen to tailor the field decay in the stack and minimize the losses due to light tunneling into the Si substrate. The stack is capped by a layer of SiO2 of a specific thickness dfirst = 462 nm, selected to obtain a surface wave supported by the stack at λ0 = 1310 nm of the same wavenumber as that of an SPP localized at a Au/Cytop interface:

such a structure are termed Bloch LRSPPs, recognizing the oscillatory decay of mode fields into the multilayer, characteristic of Bloch surface waves propagating in the stopband. Such waves are akin to Tamm SPPs, in analogy with electronic surface states within the bandgap of semiconductors.16 Dielectric-loaded 1DPCs have also been proposed, providing confinement in the plane perpendicular to the propagation direction, which is in-plane to the multilayer.17 Motivated by the higher overall sensitivity of LRSPPs on metal stripes compared to SPPs in prism-coupled biosensors,2 and by recent results that encourage the application of LRSPP biosensors to disease detection,18 we propose metal stripes on a 1DPC as waveguides supporting Bloch LRSPPs. Bloch LRSPPs propagating in the plane of the multilayer with fields confined in the plane transverse to the direction of propagation are expected to have similar sensitivities to LRSPPs in a homogeneous medium.2,18 Furthermore, the waveguides are compact and planar, which facilitates integration with other elements5 and into arrays enabling multiplexed biosensing. Specifically, we show in this paper that a Au stripe on a 1DPC implemented as a SiO2/Ta2O5 multilayer stack and covered with Cytop, can support a guided Bloch LRSPP at infrared wavelengths (λ0 ∼ 1310 nm), having properties comparable to those of the LRSPP in the corresponding symmetric structure. In addition to observing outputs from diced and polished endfacets, we integrate grating couplers at the input and output of stripes to demonstrate broadside (perpendicular) coupling. The next section of the paper summarizes the theory of the proposed structure. The subsequent sections report measurement results, then a summary and concluding remarks.

k SPP(ω0) =

ω0 c

εCytop(ω0)εAu(ω0) εCytop(ω0) + εAu(ω0)

= (6.4846 + i8.3428 × 10−3) μm−1

The structure is on a Si substrate and covered by an optically infinite Cytop layer. This layer forms the upper cladding of the structure and can be etched to form microfluidic channels, exposing the top surface of the Au stripe to analyte in a sensing fluid. For the dimensions selected, the waveguide supports a single long-range mode (Bloch LRSPP). The design procedure used to arrive at the multilayer stack rests on the effective index method, as described in ref 17. In Figure 2 the dispersion of the Bloch LRSPP guided by this structure is plotted (solid green), together with the dispersion

■

THEORETICAL The corresponding structure of interest consists of a metal stripe buried in a dielectric material having a refractive index close to that of water (i.e., ∼1.33), supporting LRSPPs (the low-loss ssb0 mode5) over a wavelength range centered near λ0 = 1310 nm.18 Currently, the realization of biosensors exploiting guided LRSPPs involves a Au stripe buried in Cytop with an etched fluidic channel,10 as Cytop has a refractive index of nCytop = 1.3348 at wavelengths in the telecom range.19 Here we adopt a different strategy and consider a 1DPC implemented as a SiO2/Ta2O5 multilayer stack to replace the Cytop substrate. We choose to work with SiO2 and Ta2O5, as these materials are transparent at telecom wavelengths, widely available, and frequently employed in the fabrication of optical components (e.g., thin film filters). Furthermore, they are robust and facilitate the fabrication of biosensors thereon. The structure of interest and under investigation is shown in Figure 1a and consists of a 5 μm wide, 35 nm thick Au (nAu =

Figure 2. Dispersion relation of the Bloch LRSPP guided by a Au stripe on a truncated 1DPC (solid green) and of the LRSPP guided by a Au stripe buried in Cytop (solid red). The two curves are nearly identical around the target wavelength of λ0 = 1310 nm, as shown in the inset.

of the LRSPP guided by the corresponding Au stripe buried in Cytop (solid red). In an ideal dielectric structure with infinite periodicity, it is possible to observe the opening of a photonic bandgap, whose central position and width depend on the material and geometrical properties of the system. In our case, we deal with a finite number of periods (N = 12), thus, we expect the formation of a high reflectivity spectral region (a stopband), which is reminiscent of the photonic bandgap of the ideal structure. In addition, the presence of a truncated layer leads to the formation of a defect mode within the stopband, which is spatially well-localized near the surface of the structure, namely, at the interface between the SiO2 and Au layers. Around the target wavelength (λ0 = 1310 nm) both LRSPPs (Cytop embedded, red; and Bloch LRSPP, green) are characterized by nearly the same dispersion relation. It is

Figure 1. (a) Front cross-sectional sketch of the metal stripe waveguide of interest, supporting Bloch LRSPPs. The relevant materials and dimensions are identified on the figure. (b) 3D sketch of the metal stripe waveguide with a grating coupler, illustrating an excitation arrangement. B

DOI: 10.1021/acsphotonics.6b00930 ACS Photonics XXXX, XXX, XXX−XXX

ACS Photonics

Article

and Au stripe, highlighting the architecture and dimensions of relevance. The grating design selected is simple, consisting of 16 rectangular ridges of dimensions W and H disposed in a period of Λ. A cross-section of the grating was modeled in the frequency domain via the 2D FEM method using software available commercially, and the coupling efficiency was determined by capturing the power carried by the Bloch LRSPP far away from the grating region (∼150 μm to the left) to eliminate the influence of spatial transients, following the approach described in ref 22. Good dimensions were determined such that the coupling efficiency was maximized, yielding for t = 35 nm dimensions of H = 150 nm, W = 450 nm, and Λ = 1000 nm. The expected coupling efficiency is about 16% (coupling loss of 8 dB) at λ0 = 1310 nm with a 3 dB bandwidth of about 70 nm. Optimization of 3D grating designs using a full model is expected to improve the coupling efficiency.

emphasized that the dispersion relation depends critically on the unit cell composition as well as the design of the cap layer (i.e., the SiO2 layer of thickness dfirst). In practice, the reliability of the design depends on the fabrication accuracy and a precise knowledge of the refractive index of the materials used. The variability of the refractive index that is typical for amorphous materials may result in a shift of the wavelength where the wavenumber matching condition is met. The blue hatched region in Figure 2 represents the expected energy range over which the working point would fluctuate assuming a ±10% variation of the optical thickness of each layer about its nominal value. Within the blue hatched region, the theoretical mode power attenuation (MPA) of the Bloch LRSPP shows little deviation from that of the LRSPP supported by the corresponding structure, with values in the range of 11−12 dB/mm. The gray shaded region in Figure 2 corresponds to the region below the light line of Cytop but outside the photonic bandgap. In Figure 3 we show the computed field amplitude profiles of the modes, obtained by FDE (Finite Difference eigenmode)

■

EXPERIMENTAL SECTION Si wafers (4 in.) each bearing a 12-period SiO 2/Ta2O5 multilayer stack capped with an extra layer of SiO2, following the design and thicknesses given in Figure 1a, were obtained from Iridian Spectral Technologies (Ottawa, Canada). Compliance of the fabricated stacks was verified by measuring wavelength responses in transmission at perpendicular incidence and ensuring the presence of a notch at about 1340 nm with a fwhm bandwidth of about 150 nm. The surface roughness of the stacks was also verified to be 1 nm RMS or less. The waveguides were defined on the stacks by contact photolithography using a standard bilayer lift-off process consisting of LOR1A lift-off resist and S1805 photoresist. A 1 min O2 plasma descum was performed (Technics Plasma Etcher 300 W) prior to thermal evaporation of Cr (0.5 nm) then Au (35 nm) without breaking vacuum. Lift-off was performed with two 10 min baths of Microposit 1165 (MicroChem) at 80 °C with a 10 s ultrasonic agitation between baths, followed by 10 min baths of IPA and DI water. The gratings were defined with a bilayer e-beam lithography process using PMMA 495A6 and PMMA 950A2 as described in ref 24. Immediately before e-beam writing, a charge dissipation layer (ESPACER 300Z) was spin-coated on the sample. After the e-beam write, the e-spacer was rinsed off with DI water and the sample was developed (MIBK:IPA developer at 20 °C) followed by a 1 h post-bake at 90 °C on a hot plate. A short descum step (20 s RIE O2 plasma 100 W) was applied prior to thermal evaporation of 150 nm of Au. Lift off of the gratings was achieved in a static overnight acetone bath at room temperature. Cytop 5% M-grade was spin-coated (1000 rpm) on the wafer, followed by four layers of 9% Cytop S-grade (1500 rpm) and a finishing layer of 5% M-grade. A 30 min bake at 50 °C was used between each layer, and the full stack was ramped to 200 °C at 10 °C/h and baked for 20 h, resulting in ∼10 μm thick upper cladding. Wafers were diced into rectangular die each bearing several test structures. Waveguide end facets were prepared by polishing die using an Ultrapol end and edge polishing machine (Ultratec Manufacturing Inc.) to produce optical quality facets. Optical quality facets were desired on some die in order to observe mode outputs directly on an infrared camera. Briefly, a large sample from a wafer was first covered with a protective layer of photoresist (S1811) then diced on a precision CNC dicing/cutting saw with a diamond blade (D10/20−150414).

Figure 3. Field amplitude profile for (a) the LRSPP on a Au stripe buried in Cytop, with the vertical field profile at x = 0 (b); and (c) the Bloch LRSPP on a Au stripe on a 1DPC with a Cytop upper cladding (Figure 1a), together with the vertical field profile at x = 0 (d).

simulation, for the LRSPP on the corresponding structure (Cytop embedded) and for the Bloch LRSPP on the truncated 1DPC. The calculated mode propagation lengths are LCytop = 638 μm and L1DPC = 389 μm, respectively, corresponding to MPAs of 6.8 and 11 dB/mm. The difference in the propagation losses is mainly attributed to the larger field fraction in the gold stripe for the 1DPC structure. The field distribution in the Cytop upper cladding is very similar in both cases, but the distribution in the 1DPC displays an oscillatory behavior, characteristic of Bloch surface modes propagating in the plane of a truncated 1DPC within the stopband. The lateral extent of the modes is comparable, indicating a similar level of confinement, and enabling bending in plane, leading potentially to Bloch LRSPP integrated optical structures such as S-bends, Y-junctions, couplers, and Mach−Zehnder interferometers, similar to those operating with LRSPPs.5,20,21 Furthermore, the vertical extend of the modes are also comparable, so a sensing performance for Bloch LRSPPs similar to that observed with LRSPPs on waveguides2,18 is expected. The excitation scheme of interest consists of illuminating a Au grating coupler integrated onto the Au stripe with a perpendicularly incident p-polarized Gaussian beam launched from an aligned optical fiber, as shown schematically in Figure 1b. Compared to butt-coupling, such a scheme has the advantages that a high-quality input facet is not required and the optical alignments are simpler, however, the coupling efficiency is typically lower.22,23 Figure 3a shows a longitudinal cross-sectional sketch along a portion of the Au input grating C

DOI: 10.1021/acsphotonics.6b00930 ACS Photonics XXXX, XXX, XXX−XXX

ACS Photonics

Article

close to the nominal ones of t = 35 and 5 μm; however, the Au thickness was not very uniform over the wafer area with some stripes having a thickness of up to 78 nm; stripes of such a thickness support Bloch LRSPPs with a higher attenuation, along with higher-order Bloch LRSPPs. Profilometer scans of the grating couplers revealed a ridge thickness of H ∼ 150 nm, very close to the desired thickness (H = 150 nm). Figure 5e shows a high-magnification microscope image of a polished die end facet. The Au stripe edge is clearly visible as are the Cytop upper cladding and the individual layers of the SiO2/Ta2O5 multilayer stack. The total thickness of the multilayer stack is about 10.5 μm, as expected, and the thickness of the Cytop upper cladding is comparable (also as expected). The experimental arrangements sketched in Figure 6 were used to test several waveguides. In both cases, a polarizationmaintaining single-mode optical fiber (PM-SMF) was used to excite a waveguide under test. A custom 90° curved metallic holder was used to mount the PM-SMF such that the ppolarized beam output therefrom was perpendicularly incident on an input grating coupler. A tunable laser (Agilent 81600B/ 8164A) with a tuning range of about 100 nm near λ0 = 1310 nm was used as the light source coupled to the input of the PM-SMF. In the arrangement of Figure 6a, the output of a waveguide was scattered vertically by a second grating coupler (identical in design to the input grating couplers) and captured by a multimode optical fiber (MMF), also mounted on a curved holder, and connected to a power meter. In the arrangement of Figure 6b, the output light was allowed to exit the polished output facet, then collimated and split, with one beam routed to an infrared camera and the other to a power meter. The fibers were aligned to a die under test using optical positioners and an alignment microscope. Different input PM-SMF’s were used in the measurement. One consisted of a cleaved Panda fiber with a core diameter of 7 μm, launching a Gaussian beam that diverges with distance from the facet. The other fiber was a tapered PM-SMF that focused the beam to a 5.4 μm diameter at a working distance of about 0.5 mm. A better coupling efficiency and less background light were observed using the latter because of improved overlap with the input grating. Figure 7 plots the Bloch LRSPP insertion loss response of several grating-coupled waveguides of identical design but different lengths, measured using the tapered PM-SMF in the arrangement of Figure 6b. The effects of the setup have been removed through calibration, so the insertion loss corresponds to that of the die only. The insertion loss increases with waveguide length (as expected) and in all cases is lowest over the wavelength range from λ0 = 1340−1360 nm. The design wavelength for the grating coupler and waveguide was λ0 = 1310 nm, but the attenuation of LRSPPs is known to decrease with increasing wavelength,5 and the 3 dB bandwidth of the grating coupler spans about 70 nm, so it is not surprising to observe a lower insertion loss at longer wavelengths. The oscillatory features in the wavelength responses are substantially reproducible from one waveguide to the next, suggesting that they are not artifacts, but rather that they originate from the structure. Figure 8 shows a mosaic of Bloch LRSPP outputs emerging from the waveguide of length 1275 μm (and width 5 μm), captured at different wavelengths using an infrared camera; the corresponding wavelength response is plotted as the thick dark blue curve in Figure 7. The output is brighter and more clearly defined at wavelengths longer than 1310 nm, particularly near

Two samples 3.5 mm long and 20 mm wide were diced and then clamped facing inward (face-to-face) to the edge polishing jig. Glass slides were used as spacers, to ensure proper alignment in the jig. The polishing machine was then used to polish the end facets, following a wet polishing process using deionized water and up to 300 rpm rotation speed. Initially, 3 M sandpaper with 9.6 μm grit was used for coarse lapping. Then, a DACE Technologies film bearing 3 μm diamond grit was used for fine lapping. Finally, a no-grit polishing pad was used for the final polishing step. Samples were inspected under microscope several times during the polishing process to assess progress. When the facets achieved a mirror-like finish, the samples were dismounted and cleaned in acetone (in an ultrasonic bath) then rinsed in IPA, followed by N2 drying. Figure 5 shows several microscope images obtained on fabricated structures. Figure 5a shows a low-magnification optical microscope image of several 5 μm wide Au stripes bearing grating couplers. The stripes are spaced sufficiently apart from each other to prevent cross-coupling, and the gratings are offset longitudinally from one waveguide to the next to produce waveguide sections of different lengths, facilitating cutback measurements. The multilayer stack appears greenish under perpendicular white light illumination. Figure 5b shows a high-magnification optical microscope image of a grating coupler comprising 16 periodically spaced rectangular Au ridges. The ridges are 4 μm long, slightly narrower than the underlying Au stripe and roughly centered on the stripe. Figure 5c gives a scanning electron microscope (SEM) image of a portion of a grating coupler, showing the dimensions of the ridges (W = 417 nm, 4.06 μm long) and the period obtained (Λ = 1.01 μm). These dimensions are close to those desired and modeled (Figure 4). Figure 5d gives an atomic force microscope (AFM) image of a Au stripe, revealing a thickness of t = 32 nm, a width of 5.1 μm, and a roughness along the top surface of the stripe of 1.9 nm RMS. These dimensions are

Figure 4. (a) Longitudinal cross-sectional sketch along a portion of the Au input grating and Au stripe. (b) Distribution of the electric field magnitude of the Bloch LRSPP over the longitudinal cross-section computed using the 2D FEM method. A p-polarized 2D Gaussian beam is assumed perpendicularly incident within the Cytop upper cladding, originating from the rectangular region sketched, and the excited Bloch LRSPP propagates in the z direction. The plot spans dimensions of about 25 μm vertically and 100 μm horizontally. D

DOI: 10.1021/acsphotonics.6b00930 ACS Photonics XXXX, XXX, XXX−XXX

ACS Photonics

Article

Figure 5. (a) Optical microscope image of 5 μm wide Au stripe waveguides bearing grating couplers. The couplers are offset longitudinally from one waveguide to the next to produce waveguide sections of different lengths. The multilayer stack produces a greenish hue under perpendicular white light illumination. (b) High-magnification optical microscope image of a grating coupler on a 5 μm wide Au stripe waveguide. (c) Scanning electron microscope (SEM) image of a portion of a grating coupler showing the dimensions of the ridges and the period. (d) Atomic force microscope (AFM) image of a Au stripe on a multilayer stack. (e) Microscope image of a polished end facet, revealing individual layers of the SiO2/Ta2O5 multilayer stack, a 5 μm wide Au stripe waveguide, and the Cytop upper cladding.

Figure 7. Bloch LRSPP insertion loss response of grating-coupled waveguides of the lengths given in the legend, excited with a tapered PM-SMF following the arrangement of Figure 6b.

Figure 6. Experimental arrangements used to test the waveguides; PM-SMF, polarization-maintaining single-mode optical fiber; MMF, multimode optical fiber. The red arrows indicate the direction of light propagation.

1.3348), a portion of the input light is evidently diffracted by the grating into slab modes, possibly due to the limited extent of the grating (∼4 μm, Figure 5c). The width of bands observed at the output is compatible with expectations given the distance between the facet and the input grating, and the waist of the excitation spot on the grating. Furthermore, we observed more background light in slab modes when exciting the input grating with a cleaved PM-SMF, which produces a larger beam at the grating, than with the tapered PM-SMF. Thus, the background light could likely be minimized further by altering and optimizing the grating design to better match the characteristics of the illuminating beam. We also attempted butt-coupling a cleaved PM-SMF to a waveguide input through a polished end facet but with limited success; in this

1340 and 1360 nm, which produce robust guiding. In all cases, background light is observed to be guided by the SiO2/Ta2O5 multilayer stack. Indeed, the finite stack (N = 12) supports a family of leaky slab modes that are radiative into the Si substrate, having effective indices that are close to the refractive indices of the constitutive materials (n ∼ 1.5−2). These modes are not laterally confined by the metallic stripe, and thus light diffracted therein gives rise to dim bands on either side of the bright central spot which corresponds to the guided Bloch LRSPP. Although the grating was designed to preferentially couple perpendicularly incident light into Bloch LRSPPs having an effective index near the refractive index of Cytop (nCytop = E

DOI: 10.1021/acsphotonics.6b00930 ACS Photonics XXXX, XXX, XXX−XXX

ACS Photonics

Article

Figure 9. Cutback plots constructed from the wavelength responses of Figure 7 at (a) λ0 = 1340 nm and (b) λ0 = 1349 nm.

Figure 8. Mosaic of Bloch LRSPP outputs captured using an infrared camera at the wavelengths indicated on each image. The camera is unsaturated and the imaging system unaltered such that all outputs can be directly compared and the relative magnitude of the Bloch LRSPP to the dimmer background radiation can be appreciated. The outputs emerged from the waveguide of length 1275 μm (and width 5 μm) that produced the corresponding response plotted as the thick dark blue curve in Figure 7.

arrangement, far more light was launched in the slab modes compared to input grating coupling, which is neff selective (the former is not). Figure 9 gives two cutback plots constructed from the wavelength responses of Figure 7, along the black and gray dashed cuts thereon, at λ0 = 1340 and 1349 nm, respectively. The linear models fit the measurements with modest fidelity (R2 ∼ 0.9). The measured MPAs (slopes) of the Bloch LRSPP work out to 12.8 and 12 dB/mm at these wavelengths, which compare reasonably well with the theoretical values when the working point is within the blue hatched region of Figure 2. Indeed, propagation losses between 11 and 12 dB/mm are predicted by numerical simulations. The coupling loss (intercept) is attributed entirely to the input grating coupler, neglecting the Fresnel transmission loss through the polished end facet. The measured coupling losses are 8.1 and 9.5 dB, respectively. Figure 10 gives a cutback plot of Bloch LRSPPs on waveguides measured at λ0 = 1310 nm, each bearing input and output grating couplers, and excited using a cleaved PMSMF in the arrangement of Figure 6a. The linear model fits the measurements with high fidelity (R2 ∼ 0.97). The measured MPA (slope) works out to 17.1 dB/mm, which is higher than the theoretical value of 11 dB/mm. This is attributed to the

Figure 10. Cutback plot of Bloch LRSPPs on waveguides at λ0 = 1310 nm, each bearing input and output grating couplers and excited using a cleaved PM-SMF in the arrangement of Figure 6a.

different position of the actual working point (which seems to be near 1340 nm) with respect to that of the modeled structure (1310 nm). Thus, in the sample, the poorer phase match between the SPP at the Au/Cytop interface and the surface wave at the Au/1DPC interface at 1310 nm leads to a larger attenuation of the Bloch LRSPP at this wavelength. The total coupling losses of 22 dB (intercept) are the combined losses of the input and output grating couplers. The cleaved PM-SMF is expected to produce a larger coupling loss through the input grating compared with the tapered PM-SMF (Figure 9) because the mode size of the former is larger and overlaps less with the grating.

■

SUMMARY AND CONCLUDING REMARKS We proposed and demonstrated a thin metal stripe, on a truncated 1DPC implemented as a SiO2/Ta2O5 multilayer F

DOI: 10.1021/acsphotonics.6b00930 ACS Photonics XXXX, XXX, XXX−XXX

ACS Photonics

Article

(11) Chabot, V.; Miron, Y.; Grandbois, M.; Charette, P. G. Long range surface plasmon resonance for increased sensitivity in living cell biosensing through greater probing depth, Sens. Sens. Actuators, B 2012, 174, 94−101. (12) Konopsky, V. N.; Alieva, E. V. Long-Range Propagation of Plasmon Polaritons in a Thin Metal Film on a One-Dimensional Photonic Crystal Surface. Phys. Rev. Lett. 2006, 97, 253904−253907. (13) Alieva, E. V.; Konopsky, V. N. Photonic crystal surface waves for optical biosensors. Anal. Chem. 2007, 79, 4729−4735. (14) Alieva, E. V.; Konopsky, V. N. Long-range plasmons in lossy metal films on photonic crystal surfaces. Opt. Lett. 2009, 34, 479−481. (15) Delfan; Degli-Eredi, I.; Sipe, J. E. Long-range surface plasmons in multilayer structures. J. Opt. Soc. Am. B 2015, 32, 1615−1623. (16) Kaliteevski, M.; Iorsh, I.; Brand, S.; Abram, R. A.; Chamberlain, J. M.; Kavokin, A. V.; Shelykh, I. A. Tamm plasmon-polaritons: Possible electromagnetic states at the interface of a metal and a dielectric Bragg mirror. Phys. Rev. B: Condens. Matter Mater. Phys. 2007, 76, 165415. (17) Liscidini, M. Surface guided modes in Photonic Crystal Ridges: the Good, the Bad, and the Ugly. J. Opt. Soc. Am. B 2012, 29, 2103. (18) Krupin, O.; Wong, W. R.; Béland, P.; Adikan, F. R. M.; Berini, P. Long-Range Surface Plasmon-Polariton Waveguide Biosensors for Disease Detection. J. Lightwave Technol. 2016, 34, 4673−4681. (19) Asahi Glass Company. Cytop Technical Brochure [Online]; http://www.agc.com. (20) Boltasseva, A.; Nikolajsen, T.; Leosson, K.; Kjaer, K.; Larsen, M. S.; Bozhevolnyi, S. I. Integrated optical components utilizing longrange surface plasmon polaritons. J. Lightwave Technol. 2005, 23, 413− 422. (21) Joo, Y. H.; Song, S. H.; Magnusson, R. Demonstration of longrange surface plasmon-polariton waveguide sensors with asymmetric double-electrode structures. Appl. Phys. Lett. 2010, 97, 201105. (22) Chen, C.; Berini, P. Grating couplers for broadside input and output coupling of long-range surface plasmons. Opt. Express 2010, 18, 8006−8018. (23) Mueller, J. P. B.; Leosson, K.; Capasso, F. Polarization-Selective Coupling to Long-Range Surface Plasmon Polariton Waveguides. Nano Lett. 2014, 14, 5524−5527. (24) Fong, N. R.; Berini, P.; Tait, R. N. Fabrication of long-range surface plasmon hydrogen sensors on Cytop membranes integrating grating couplers. J. Vac. Sci. Technol., B: Nanotechnol. Microelectron.: Mater., Process., Meas., Phenom. 2015, 33, 021201.

stack, and covered with Cytop, as a waveguide supporting Bloch LRSPPs. The propagation of Bloch LRSPPs along the metal stripe, confined in the transverse plane, was demonstrated at wavelengths near λ0 = 1310 nm. MPAs of 12−17 dB/ mm were measured, in reasonable agreement with theory; a lower attenuation is possible by optimizing the waveguide design and improving the fabrication processes. Input grating couplers implemented as a periodic array of Au ridges on Au stripes were used to couple perpendicularly incident p-polarized light to the input of waveguides. Coupling losses of 8−10 dB to a tapered PM-SMF were measured; lower coupling losses are possible by optimizing the grating design and its fabrication. Output means were provided by integrating an identical grating coupler near the end of a waveguide to diffract Bloch LRSPPs upward into a receiving MMF or by polishing the output facet and allowing Bloch LRSPPs to radiate into a free-space beam. Implementing metal stripe waveguides on a truncated 1DPC advantageously enables the use of a large range of materials for the substrate, breaking free from the constraint of material symmetry to support LRSPPs. Grating coupling has the advantages that high-quality end facets are not required and optical alignments are facilitated. These advances further the attraction of metal stripe waveguides and LRSPPs for biosensing applications.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Pierre Berini: 0000-0002-6795-7275 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors are grateful to Behnood Ghamsari and Wei Ru Wong for assistance with some of the measurements and to J. E. Sipe for fruitful discussions.

■

REFERENCES

(1) Maier, S. A. Plasmonics: Fundamentals and Applications; Springer Science+Business Media: New York, 2007. (2) Berini, P. Bulk and surface sensitivities of surface plasmon waveguides. New J. Phys. 2008, 10, 105010. (3) Prasad, P. N. Introduction to Biophotonics; Wiley: NJ, 2003. (4) Homola, J. Present and future of surface plasmon resonance biosensors. Anal. Bioanal. Chem. 2003, 377, 528. (5) Berini, P. Long-range surface plasmon polaritons. Adv. Opt. Photonics 2009, 1, 484−588. (6) Berini, P.; Charbonneau, R.; Lahoud, N. Long-range surface plasmons along membrane-supported metal stripes. IEEE J. Sel. Top. Quantum Electron. 2008, 14, 1479−1495. (7) Min, Q.; Chen, C.; Berini, P.; Gordon, R. Long range surface plasmons on asymmetric suspended thin film structures for biosensing applications. Opt. Express 2010, 18, 19009−19019. (8) Wark, W.; Lee, H. J.; Corn, R. M. Long-range surface plasmon resonance imaging for bioaffinity sensors. Anal. Chem. 2005, 77, 3904−3907. (9) Slavík, R.; Homola, J. Ultrahigh resolution long range surface plasmon-based sensor, Sens. Sens. Actuators, B 2007, 123, 10−12. (10) Chiu, C.; Lisicka-Skrzek, E.; Tait, R. N.; Berini, P. Fabrication of surface plasmonwaveguides and devices in Cytop with integrated microfluidic channels. J. Vac. Sci. Technol., B: Nanotechnol. Microelectron.: Mater., Process., Meas., Phenom. 2010, 28, 729−735. G

DOI: 10.1021/acsphotonics.6b00930 ACS Photonics XXXX, XXX, XXX−XXX