Compact on-chip plasmonic light concentration based on a hybrid photonic-plasmonic structure

We present a novel approach for achieving tightly concentrated optical field by a hybrid photonic-plasmonic device in an integrated platform, which is a triangle-shaped metal taper mounted on top of a dielectric waveguide. This device, which we call a plasmomic light concentrator (PLC), can achieve vertical coupling of light energy from the dielectric waveguide to the plasmonic region and light focusing into the apex of the metal taper(at the scale ∼ 10nm) at the same time. For a demonstration of the PLCs presented in this paper, we numerically investigate the performance of a gold taper on a Si 3N4 waveguide at working wavelengths around 800nm. We show that three major effects (mode beat, nanofocusing, and weak resonance) interplay to generate this light concentration phenomenon and govern the performance of the device. Combining these effects, the PLC can be designed to be super compact while maintaining high efficiency over a wide band. In particular, we demonstrate that under optimized size parameters and wavelength a field concentration factor (FCF), which is the ratio of the norm of the electric field at the apex over the average norm of the electric field in the inputting waveguide, of about 13 can be achieved with the length of the device less than 1 μm for a moderate tip radius 20nm. Moreover, we show that a FCF of 5 − 10 is achievable over a wavelength range of 700− 1100nm with the length of the device further reduced (to about 400nm). © 2013 Optical Society of America OCIS codes:(250.5403) Plasmonics; (240.6680) Surface plasmons. References and links 1. J. A. Schuller, E. S. Barnard, W. Cai, Y. Jun, J. S. White, and M. L. Brongersma, “Plasmonics for extreme light concentration and manipulation,” Nat. Materials 9, 193–204 (2010). 2. Z. Liu, J. M. Steele, W. Srituravanich, Y. Pikus, C. Sun, and X. Zhang, “Focusing surface plasmons with a plasmonic lens,” Nano Lett. 5, 1726–1729 (2005). 3. W. Chen, D. C. Abeysinghe, R. L. Nelson, and Q. Zhan, “Plasmonic lens made of multiple concentric metallic rings under radially polarized illumination,” Nano Lett. 9, 4320–4325 (2009). 4. L Yin, V. K. Vlasko-Vlasov, J. Pearson, J. M. Hiller, J. Hua, U. Welp, D. E. Brown, and C. W. Kimball, “Subwavelength focusing and guiding of surface plasmons,” Nano Lett. 5, 1399–1402 (2005). 5. A. Yanai and U. Levy, “Plasmonic focusing with a coaxial structure illuminated by radially polarized light,” Opt. Express17,13150–13157 (2009). 6. L. Novotny, R. X. Bian and X.S. Xie, “Theory of nanometric optical tweezers,” Phys. Rev. Lett. 79, 645–648 (1997). #178837 $15.00 USD Received 5 Nov 2012; revised 30 Dec 2012; accepted 9 Jan 2013; published 17 Jan 2013 (C) 2013 OSA 28 January 2013 / Vol. 21, No. 2 / OPTICS EXPRESS 1898 7. K. V. Nerkararyan, “Superfocusing on a surface polariton in a wedge-like structure,” Phys. Lett. A 237,103–105 (1997). 8. A. J. Babadjanyan, N. L. Margaryan, and K. V. Nerkararyan, “Superfocusing of surface polaritons in the conical structure,” J. Appl. Phys. 87,3785–3788 (2000). 9. M. I. Stockman, “Nanofocusing of optical energy in tapered plasmonic waveguides,” Phys. Rev. Lett. 93,137404 (2004). 10. N. A. Issa and R. Guckenberger, “Optical nanofocusing on tapered metallic waveguides,” Plasmonics 2, 31–37 (2007). 11. K. C. Vernon, D. K. Gramotnev, and D. F. P. Pile, “Adiabatic nanofocusing of plasmons by a sharp metal wedge on a dielectric substrate,” J. Appl. Phys. 101,104312 (2007). 12. E. Moreno, S. G. Rodrigo, S. I. Bozhevolnyi, L. Mart ı́n-Moreno, and F. J. Garc ı́a-Vidal, “Guiding and focusing of electromagnetic fields with wedge plasmon polaritons,” Phys. Rev. Lett. 100,023901 (2008). 13. S. I. Bozhevolnyi and K. V. Nerkararyan, “Adiabatic nanofocusing of channel plasmon polaritons,” Opt. Lett. 35,541–543 (2009). 14. V. S. Volkov, S. I. Bozhevolnyi, S. G. Rodrigo, L. Mart ı́n-Moreno, F. J. Garc ı́a-Vidal, E. Devaux, and T. W. Ebbesen, “Nanofocusing with channel plasmon polaritons,” Nano Lett. 9, 1278–1282 (2009). 15. V. S. Volkov, J. Gosciniak, S. I. Bozhevolnyi, S. G. Rodrigo, L. Mart ı́n-Moreno, F. J. Garc ı́a-Vidal, E. Devaux, and T. W. Ebbesen, “Plasmonic candle: towards efficient nanofocusing with channel plasmon polaritons,” New J. Phys.11,113043 (2009). 16. Z. Fang, H. Qi, C. Wang, and X. Zhu, “Hybrid plasmonic waveguide based on tapered dielectric nanoribbon: excitation and focusing,” Plasmonics 5, 201–212 (2010). 17. K. Tanaka, K. Katayama, and M. Tanaka, “Nanofocusing of surface plasmon polaritons by a pyramidal structure on an aperture,” Opt. Express 18,787–798 (2010). 18. N. C. Lindquist, P. Nagpal, A. Lesuffleur, D. J. Norris, and S. Oh, “Three-dimensional plasmonic nanofocusing,” Nano Lett.10,1369–1373 (2010). 19. C. Ropers, C. C. Neacsu, T. Elsaesser, M. Albrecht, M. B. Raschke, and C. Lienau, “Grating-coupling of surface plasmons onto metallic tips: A nanoconfined light source,” Nano Lett. 7, 2784–2788 (2007). 20. E. Verhagen, A. Polman, and L. Kuipers, “Nanofocusing in laterally tapered plasmonic waveguides,” Opt. Express16,766–788 (2008). 21. A. Normatov, P. Ginzburg1, N. Berkovitch, G. M. Lerman, A. Yanai, U. Levy, and M. Orenstein, “Efficient coupling and field enhancement for the nano-scale: plasmonic needle,” Opt. Express 18,14079–14086 (2009). 22. H. Choi, D. F. P. Pile, S. Nam, G. Bartal, and X. Zhang, “Compressing surface plasmons for nano-scale optical focusing,” Opt. Express 17,7519–7524 (2009). 23. M. Schnell, P. Alonso-Gonz ález, L. Arzubiaga, F. Casanova, L. E. Hueso, A. Chuvilin, and R. Hillenbrand, “Nanofocusing of mid-infrared energy with tapered transmission lines,” Nat. Photonics 5, 283–287 (2011). 24. M. Malerba, A. Alabastri, G. Cojoc, M. Francardi, M. P. Donnorso, R. P. Zaccaria, F. De Angelis, and E. Di Fabrizio, “Optimization of surface plasmon polariton generation in a nanocone through linearly polarized laser beams,” Microelectron. Eng. 97,204 (2012). 25. R. Yang, R. A. Wahsheh, Z. Lu, and M. A. G. Abushagur, “Efficiently squeezing near infrared light into a 21nm-by-24nm nanospot,” Opt. Express 16,20142–20148 (2008). 26. B. Desiatov, I. Goykhman, and U. Levy, “On-chip focusing of light by metallic nanotip,” Frontiers in Optics, OSA Technical Digest (CD) (OSA, 2010), paper FThB7. 27. B. Desiatov, I. Goykhman, and U. Levy, “Nanoscale mode selector in silicon waveguide for on-chip nanofocusing applications,” Nano Lett. 9, 3381–3386 (2009). 28. B. Desiatov, I. Goykhman, and U. Levy, “Plasmonic nanofocusing of light in an integrated silicon photonics platform,” Opt. Express19,13150–13157 (2011). 29. Y. Luo, M. Chamanzar, A. A. Eftekhar, and A. Adibi, “On-chip nanofocusing using a hybrid plasmonic-dieletric tapered waveguide,” Integrated Photonics Research, Silicon and Nanophotonics, OSA Technical Digest (CD) (OSA, 2011), paper ITuD7. 30. X. He, L, Yang, and T. Yang, “Optical nanofocusing by tapering coupled photonic-plasmonic waveguides,” Opt. Express19,12865–12872 (2011). 31. P. Berini, “Plasmon-polariton waves guided by thin lossy metal films of finite width Bound modes of asymmetric structures,” Phys. Rev. B 63,125417 (2001). 32. R. F. Oulton, G. Bartal, D. F. P. Pile, and X. Zhang, “Confinement and propagation characteristics of subwavelength plasmonic modes,” New J. Phys. 10,105018 (2008). 33. D. Dai and S. He, “A silicon-based hybrid plasmonic waveguide with a metal cap for a nano-scale light confinement,” Opt. Express17,16646–16653 (2009). 34. I. Goykhman, B. Desiatov, and U. Levy, “Experimental demonstration of locally oxidized hybrid siliconplasmonic waveguide,” Appl. Phys. Lett. 97,141106 (2010). 35. M. Chamanzar, M. Soltani, B. Momeni, S. Yegnanarayanan, and A. Adibi, “Hybrid photonic surface-plasmonpolariton ring resonators for sensing applications,” Appl. Phys. B 101,263–271 (2010). 36. M. Chamanzar and A. Adibi, “Hybrid nanoplasmonic-photonic resonators for efficient coupling of light to single #178837 $15.00 USD Received 5 Nov 2012; revised 30 Dec 2012; accepted 9 Jan 2013; published 17 Jan 2013 (C) 2013 OSA 28 January 2013 / Vol. 21, No. 2 / OPTICS EXPRESS 1899 plasmonic nanoresonators,” Opt. Express 19, invited for Focus Issue: Collective Phenomena, 22292–22304 (2011). 37. A. D. Rakic, A. B. Djurisic, J. M. Elazar, and M. L. Majewski “Optical properties of metallic films for verticalcavity optoelectronic devices,” Appl. Opt. 37,5271–5283 (1998).


Introduction
Plasmonic metallic structures are well-known for the ability of breaking the diffraction limit and achieving large field enhancements or field confinements in small volumes [1].One mainstream of the approaches is to focus surface plasmon polariton (SPP) modes, which can be further divided into two categories.One is interference-induced plasmonic nanofocusing.For this type of techniques, the plasmonic excitation sites on a metal film are carefully tuned so that the propogation of SPP originated at the sites can get focused into a hot spot on the metal film.Typical planar patterns for plasmonic excitation on the metal film include individual circular rings [2], multiple concentric rings [3], curved arrays of nanoholes [4], and certain coaxial structures [5].The other is geometry-induced plasmonic nanofocusing.In principle, tapering the cross section of the SPP propagation gradually will increase the intensity of the SPP and the slow down its group velocity.At the tip end of the taper, the group velocity becomes zero which means the surface plasmon becomes localized and the energy is highly concentrated [6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24].This type of techniques can accurately control the focus point and usually have a higher concentrating ability.
Conventionally, the nanofocusing techniques have the SPP excited from free space.Very recently, several approaches of realizing plasmonic nanofocusing on integrated-optics platforms have been reported [25][26][27][28][29][30].Using these on-chip light concentration techniques, the light energy is coupled from a photonic waveguide to a plasmonic structure (usually a metallic taper) and then concentrating the light energy to the tip of the taper.The coupling from the photonic waveguide to the plasmonic structure is efficient and robust, while the plasmonic part is usually compact since it only concentrates light energy from a "∼ λ " scale to a "< λ /10" scale.Therefore, the lossy nature of metals has less impacts on the concentration efficiencies.Moreover, many of such structures can be integrated on the same chip in series or in parallel in a very small footprint.This opens up a new avenue for multiplex nanofocusing in applications such as on chip trapping and sensing.
In this paper, we will present the design of a novel compact on-chip plasmonic light concentrator (PLC), and show that there are three major effects (mode beat, nanofocusing, and weak resonance) that govern this light concentration phenomenon and affect the efficiency of the device.By coordinating these effects, this proposed structure can maintain very high concentrating efficiency while the size dimensions (width∼ 400 nm and length< 1 µm) are the smallest among the all the non-resonant on-chip plasmonic-light-concentrators that have been reported.
The basic structure of the device is a small triangle-shaped metal taper on top of a dielectric waveguide, as shown in Fig. 1(a).In our analysis of this structure, the material for the metal triangle is gold (Au), and the material for the dielectric ridge waveguide is silicon nitride (Si 3 N 4 ).The substrate is silicon dioxide (SiO 2 ).In addition, a SiO 2 buffer layer is used to separate the metal triangle and the dielectric ridge waveguide.The ambient material is water.The dimensions of the ridge waveguide are chosen to support only the fundamental TM-like (vertical polarization) mode and the fundamental TE-like (horizontal polarization) mode.Only the fundamental TM-like mode can be used to generate light concentration.We set up the coordinate system with X-axis in the horizontal direction, Y -axis in the vertical direction and Z-axis in the propagation direction of the input light.
Figure 1(b) shows the top view of the metal triangle, which is isosceles.We use W to denote the length of the base of the isosceles triangle, and L to denote the the length of a perpendicular from the center of the curved tip to the middle point of the base side.For simplicity, we will call W the (maximum) width of the triangle and L the length of the triangle throughout this paper.Since the apex of the metal triangle is always rounded up in real fabrications, we introduce the radius of curvature a at the tip as a parameter.

Mode analysis of the hybrid photonic-plasmonic waveguide
To analyze the behavior of the PLC, we first need to find the modes of the corresponding hybrid photonic-plasmonic waveguide.Several techniques of combining plasmonic and dielectric waveguiding with different geometric settings have been reported [31][32][33][34][35][36].The hybrid waveguide we analyze here is formed by placing a metal (Au) strip on top of a dielectric ridge waveguide with a buffer layer as shown in Fig. 2(a).The dimensions of different layers in the the Si 3 N 4 ridge waveguide have been selected for single-mode operation (for each polarization) at wavelength λ of 800 nm with the ambient material being water.The Au strip has thickness of 40 nm and width w as a variable.For λ = 800 nm, the refractive indices of Si 3 N 4 , SiO 2 and water are assumed to be 2.00, 1.46, and 1.33, respectively, and the permittivity of Au is assumed to be −24.02+ j1.18 (computation based on data from [37]).
The analysis technique is graphically shown in Figs.2(b) and 2(c).We first analyze the modes of the corresponding purely photonic waveguide (no Au layer) and of the corresponding purely plasmonic waveguide (by replacing Si 3 N 4 with SiO 2 ).In the next step, we find the modes of the hybrid photonic-plasmonic waveguide by using the supermode analysis through adding and subtracting the modes of the two structures as shown in shown in Figs.2(b) and 2(c).
Based on the designed dimensions, the purely photonic waveguide only support the fundamental TM-like mode (denoted by TM 0 ) and the fundamental TE-like mode (denoted by TE 0 ).The TM 0 mode can only couple with symmetric modes of the purely plasmonic waveguide, and the TE 0 mode can only couple with the asymmetric modes of the purely plasmonic waveguide.
We only consider the fundamental symmetric (S 0 ) and asymmetric (A 0 ) modes of the purely plasmonic waveguide, since higher order modes are cut off with the dimensions of our structure at the wavelength λ of 800 nm. Figure 2(b) shows two supermodes, H TM,0 and H TM,1 , are derived from the superposition of the TM 0 mode and the S 0 mode.Analogously, Fig. 2(c) shows two supermodes, H TE,0 and H TE,1 , are derived from the superposition of the TE 0 mode and the A 0 mode.To give more details about the mode-coupling processes, we use finite element method (FEM, COMSOL Multiphysics) to compute the eigenmodes at the wavelength λ of 800 nm.Note that the effective indices n eff of the propagating modes must have real parts greater than 1.46, which is the refractive index of the SiO 2 substrate.
Figure 3(a) shows the normalized electric field profiles for the TM 0 , S 0 , H TM,0 and H TM,1 modes, with the width w of the Au strip set to 620 nm.The electric field lines are plotted.The effective indices of the modes are also listed in Fig. 3(a).Figure 3(b) illustrates similar data for the modes TE 0 , A 0 , H TE,0 and H TE,1 .The electric field lines in Fig. 3 clearly show the validity of the supermode analysis approach as shown in Fig. 2.
The dispersions of the four propagating supermodes (H TM,0 , H TM,1 , H TE,0 and H TE,1 ) of the hybrid photonic-plasmonic waveguide are plotted in Fig. 4 in the form of the real and imaginary parts of the effective index versus the width w of the Au strip.The coordinate "Real(n eff )" has a minimum at 1.46 as only propagating modes are considered.The imaginary part of the effective index accounts for the propagation loss.
The Real(n eff ) of H TM,0 is greater than the Real(n eff ) of H TM,1 , and as w decreases, the difference between the two becomes larger.The imaginary parts follow the same trend.This  is because as w decreases, the light energy becomes more concentrated around the lossy metal region for H TM,0 , while it becomes more concentrated in the dielectric region for H TM,1 .The mode profiles of both supermodes at w = 60 nm are illustrated in Fig. 5.As w approaches 0, both the real and imaginary parts of the effective index of H TM,0 go to infinity.This means that the phase velocity goes to 0, and the H TM,0 mode becomes more and more like a localized mode.When w = 0, there is no metal strip, and the H TM,1 mode is actually the purely photonic mode TM 0 .Note that as the width of the metal layer continuously and gradually shrinks down, the amplitude of the H TM,0 mode increases since its propagating velocity slows down.Such a process is common to most geometry-induced plasmonic nanofocusing techniques, and is one of the main mechanisms to generate light concentration in our triangle-shaped plasmonic device.Analogously, we observe that the H TE,0 mode evolves into the purely photonic mode TE 0 as w approaches 0.Moreover, when w < 180 nm, the H TE,1 mode is cut off.Therefore we can not take use of the supermodes H TE,0 and H TE,1 for our application of light concentration.The supermodes in the hybrid waveguide can be excited by the modes in the corresponding bare (or purely photonic) waveguide.More specifically, the incident mode TM 0 in the bare waveguide can excite two supermodes (H TM,0 and H TM,1 ) of the hybrid waveguide.The superposition of these two supermodes at the input end has field in the dielectric region to match the incident TM 0 mode.Since they have different effective indices, as can be seen from the dispersion diagram in Fig. 4, the two supermodes undergo a beat effect with light energy bouncing back and forth between the dielectric region and the metal region.For a specific width w of the Au strip, we have where BL(w) is the beat length.In particular, at a distance about half the beat length from the input end, the light energy is vertically coupled efficiently into the metal region.

The performance of a PLC with specific size parameters
The PLCs presented in this paper (see Fig. 1) are designed as tapering the metal strip of a hybrid photonic-plasmonic waveguide.The performance of a PLC depends on the interplay of three effects, mode beat, nanofocusing and weak resonance: the mode beat effect comes from the interference of the two excited supermodes (H TM,0 and H TM,1 ); the nanofocusing effect comes from the triangle taper; and the weak resonance effect comes from the reflections at the input end and the apex of the taper.
To clarify the roles of the those three effects, we consider the field profile in the triangular tapered structure shown in Fig. 6(a).The Si 3 N 4 ridge has a width of 620 nm and a thickness of 200 nm; the SiO 2 buffer has a width of 620 nm and a thickness of 100 nm.The thickness of the Au triangle is 40 nm.The ambient material is also water.These parameters are the same as the hybrid photonic-plasmonic waveguide we analyzed in the previous section.The coordinate system is chosen to be the same as that in Fig. 1.In particular, the coordinate origin is chosen such that the plane X = 0 cuts through the middle of the Si 3 N 4 ridge vertically, the plane Y = 0 coincides with the boundary between the Si 3 N 4 layer and the SiO 2 substrate, and the plane Z = 0 goes through the center of the curved tip.
The electric field patterns in the horizontal (Fig. 6(a) and 6(b)) and vertical (Fig. 6(a)) planes are calculated using a 3D FEM simulation.Here Y = 200 nm +100 nm +20 nm, which is the sum of the thickness of the Si 3 N 4 layer, the thickness of the SiO 2 buffer layer and half of the thickness of the Au layer.The Au triangle has dimensions L = 900 nm, W = 400 nm and a = 20 nm.We let |E 0 | be the average norm of the incident electric field in the Si 3 N 4 region, which is calculated from the the TM 0 mode of corresponding purely photonic waveguide.We estimate the total power flow by a plane-wave approximation with an electric field E 0 homogeneously distributed and restricted within the Si 3 N 4 region.In particular, the norm of E 0 can be computed by the formula Here η 0 = µ 0 /ε 0 ≈ 120π is the wave impedance of free space, n = 2 is the refractive index of Si 3 N 4 , A = 620 nm ×200 nm is the area of the cross section of the Si 3 N 4 slab, and D Sz is the time-averaged power flow through a cross-section plane D. The electric field patterns shown in Fig. 6 are normalized to |E 0 |.The field becomes highly concentrated at the metallic tip.To evaluate the efficiency of field concentration induced by the PLC, we introduce a field concentration factor (FCF) as the electric field norm at the apex point (point Q in Fig. 6 (b) and (c)) normalized to |E 0 |.Our calculations show that FCF = 12.6 for the tapered structure in Fig. 6.Note that the FCF depends on the radius of curvature a.While reducing a can result in higher FCF, we should choose a large enough to accommodate the fabrication requirement.
Here we set a = 20 nm, and write FCF(a = 20 nm) = 12.6.(Many authors use the intensity ratio to evaluate the concentration efficiency, which corresponds to the square of the FCF in this paper.)Compared to other reported results [28,30] of on-chip plasmonic light concentrators, the concentration efficiency (evaluated by intensity ratio) reported here is at least two times better while using a much smaller a.

The three underlying effects of the device
As mentioned above, the field concentration is governed by three effects: mode beat, nanofocusing and weak resonance.
As discussed in Section 2, the incident mode TM 0 in the bare photonic waveguide can excite two supermodes, H TM,0 and H TM,1 , in the corresponding hybrid waveguide with a metal (Au) strip integrated on top.The triangle-shaped PLC can be considered as a taper with the width of the Au strip gradually going down to 0. Therefore, we can still use the two supermodes for the Au taper in our analysis.The mode beat between the two supermodes can make a vertical coupling (or side-coupling), which transfers light energy quickly from the dielectric region to the metal region, and the triangular taper performs plasmonic nanofocusing at the same time.This simultaneous light coupling and focusing in a single triangle-shaped metal taper is a major advantage of our design, resulting in a very compact structure.More accurately speaking, the tapering also induces energy coupling between the two supermodes.In other words, the ratio of energy stored in the two modes will change at different locations along the metal taper, and only the H TM,0 mode becomes localized as the metallic tip.In our analysis based on a simplified model, we consider that the two supermodes have constant mode effective indices which are calculated using the average width of the taper.
The third effect that must be studied in our device is the weak resonance caused by the reflection of light the apex and the input end the the taper.This weak resonance also affects the efficiency of light concentration into the tip.By adjusting the size parameters of the Au taper, we can combine the three effects to optimize the concentration efficiency of the PLC.Fig. 7. FCF versus length L of the Au triangle: the maximum width W is 300 nm for the blue solid curve and 400 nm for the red dashed curve.The wavelength λ is 800 nm.The radius of curvature a at the tip is 20 nm.All other parameters are the same as for the tapered structure shown in Fig. 6.
Figure 7 shows how the FCF varies with respect to the length L of the Au triangle, while the radius of curvature a is fixed at 20 nm.The wavelength λ is 800 nm.The two curves correspond to two cases of the maximum width W of the triangular taper: the blue solid one is for the case W = 300 nm, and the red dashed one is for the case W = 400 nm.Two oscillatory behaviors can be seen in Fig. 7: the slow variations corresponding to the mode-beat effect and the ripples with high oscillation frequency corresponding to the weak resonance effect.
Although Equation ( 1) is derived for calculating the beat length of a hybrid photonicplasmonic waveguide with a fixed metallic width w, we can use it to estimate the beat length of the tapered structure by using the average metallic width over the tapered length for w.For a triangular taper, this corresponds to about 1/2 of the maximum width W of the taper.Using this approach, we can find n eff for the hybrid H TM,0 and H TM,1 modes, and then compute the beat lengths for the two cases (W = 300 nm and W = 400 nm) shown in Fig. 7.For the narrower taper (W = 300 nm and w = W /2 = 150 nm), we obtain Real(n eff (H TM,0 (150 nm))) = 1.9694,Real(n eff (H TM,1 (150 nm))) = 1.5475 and BL(150 nm) = 1896 nm.For the wider taper (W = 400 nm and w = W /2 = 200 nm), we obtain Real(n eff (H TM,0 (200 nm))) = 1.9216,Real(n eff (H TM,1 (200 nm))) = 1.5413 and BL(200 nm) = 2104 nm.We can see that the computed beat length for the wider taper is a little larger than that for the narrower one.
The beat lengths can also be estimated directly from Fig. 7 by locating the the first minimum of the slow variation part of each curve.Using this method, the estimated beat length for the narrower taper is 1.8 µm and for the wider taper is 2.0 µm, agreeing well with the computations based on Equation (1).
When L is about half of the beat length (e.g.L ⋍ 1.0 µm for the wider taper with W = 400 nm), the light energy transfers from the dielectric region to the metal region most efficiently, which results in a higher FCF, as shown in Fig. 7.
The effect of weak resonance can also be explained by our simple theoretical model.The fast ripples in Fig. 7 are due to the limited length of the taper.By approximating the triangular tapered structure with a Fabry-Pérot structure and using the value of n eff (H TM,0 (w)) with w being the average width of the triangular taper, we can calculate the free spectral range (FSR) of these fast oscillation using For example, in the case of wider taper (W = 400 nm and w = 200 nm) in Fig. 7, we have Real(n eff (H TM,0 (200 nm))) = 1.9216, and FSR(200 nm) = 208 nm.This corresponds to about 15 peaks for the red curve, matching well with an actual counting on the figure.The FCF is usually the highest, when L is chosen to be at a resonance peak close to half the beat length.In an actual design, we might choose L smaller than its optimal value to reduce the device dimension while keeping FCF high.For example, by choosing L = 500 nm, an FCF of about 10 can be obtained as seen from Fig. 7.

Analysis of the transmission, reflection and FCF spectra
The field concentration effect of a PLC can affect the transmission and generate some reflection in the underlying photonic waveguide.In this section, we investigate the relations among the amplitudes of the transmission, reflection and FCF by analyzing their spectra.We use the finitedifference-time-domain (FDTD, Lumerical Solutions) method for simulations.
Figure 8 shows the transmission T , reflection R, and T + R spectra for (a) L = 1 µm and (b) L = 2 µm for the triangular taper with W = 400 nm and a = 20 nm.The wavelength range is 600 − 1100 nm.
Let us consider the L = 1 µm case first.As we have computed, when W = 400 nm, the beat length is about 2.1 µm at the wavelength of 800 nm.For wavelengths in a range around 800 nm, the beat length does not change much.Therefore, 1 µm is around half the beat length, when the PLC is expected to be most efficient.The term 1 − T − R indicates the power consumption on the metal structure due to the internal material absorption and the radiative loss.
As shown in Fig. 8(a), a resonance corresponds to the a drop in the transmission spectrum and a peak in the reflection spectrum.At resonance, the field is more concentrated in the metal region, resulting in an increased power consumption.This lowers the transmission T and the overall T + R, while inducing a larger reflection R. It is noticeable that the reflection part contributes less significantly to the overall T + R than the transmission part, and therefore the profile of the T + R spectrum is basically determined by the transmission spectrum.When the structure is off resonance, we have T + R about 0.8, which means a certain amount of power is still coupled to the metal region.This is quite different from the typical spectrum of a plasmonic resonator, for which T + R is very close to 1 when the structure is off resonance.Figure 8(b) shows the spectra in the L = 2 µm case.The spacing between two successive drops in the transmission spectrum is clearly less than that of the L = 1 µm case.Moreover, since 2 µm is close to the beat length, the PLC is expected to be less efficient.As a result, the overall reflection is considerably smaller than that in the L = 1 µm case.
In Fig. 9, we compare transmission spectra and FCF spectra for three groups of metal triangle lengths.Those in the first group (Fig. 9(a)) are 0.4 µm, 0.425 µm and 0.45 µm, those in the second group (Fig. 9(b)) are 1 µm, 1.025 µm and 1.05 µm, and those in the third group (Fig. 9(c)) are 2 µm, 2.025 µm and 2.05 µm.We use W = 400 nm and a = 20 nm for all structures.Recall that by definition, the FCF(a = 20 nm) is the normalized norm of the electric field at the point (0, 320 nm, 20 nm) in the coordinate system.When the wavelength is less than 700 nm, the bare photonic waveguide supports higher order TM modes and the hybrid structure supports more modes in addition to H TM,0 and H TM,1 .Thus our analytic model using two supermodes is no longer valid in this case.The simulations show that the field concentration effect is tremendously suppressed in this case.For wavelengths larger than 700 nm, the vertical dashed lines indicate that a local minima of the transmission corresponds almost exactly to a local maxima of the FCF, and correspondingly the plasmonic structure is at resonance.

Conclusion
In this paper, we have systematically analyzed the behavior and performance of a compact plasmonic light concentrator (PLC) which is a triangle-shaped Au taper integrated on a Si 3 N 4 waveguide.We use the fundamental TM-like mode for the incident light from a photonic waveguide to excite two supermodes in the hybrid structure.We have shown that three major effects (mode beat, nanofocusing, and weak resonance) interplay to generate the light concentration phenomenon and govern the performance of the device.By proper combination of these three effects, the overall structure can be optimized to obtain very large field concentration.In particular, we have demonstrated that after an optimization of the size parameters, a field concentration factor (FCF) of about 13 can be achieved with the length of the device less than 1 µm for a moderate tip radius of 20 nm.Such a field concentration efficiency is about twice better than all reported results of on-chip plasmonic light concentrators in literature with much larger device sizes and an assumption of smaller tip radii.In addition, the dimensions of the device can be reduced further while maintaining a high concentration efficiency (a FCF of 5 − 10 is achievable with the length of triangle about 400 nm).
In addition to realizing a larger field concentration with a smaller device size, our plasmonic device is mounted on top of a photonic waveguide and the waveguide does not terminate after the light concentration as in typical schemes reported earlier [25,26,28,30].This makes our device not only simpler in fabrication, but also more compatible for the integrated optical circuits.Finally, the three major effects commonly appear when different materials are used for the photonic waveguide and plasmonic layer while working at different wavelengths, and we can apply the same philosophy to optimize the design.

Fig. 1 .
Fig. 1.(a) Schematic of an compact PLC that is a hybrid photonic-plasmonic structure with a gold triangle taper integrated on top of a Si 3 N 4 ridge waveguide with a SiO 2 buffer layer.(b) The top view of this hybrid structure.

Fig. 2 .
Fig. 2. (a) The cross section and top view of a hybrid photonic-plasmonic waveguide.(b) Two supermodes (H TM,0 and H TM,1 ) come from the superposition of the fundamental TMlike mode (TM 0 ) of the purely photonic waveguide and the fundamental symmetric mode (S 0 ) of the purely plasmonic waveguide.(c) Two supermodes (H TE,0 and H TE,1 ) come from the superposition of the fundamental TE-like mode (TE 0 ) of the purely photonic waveguide and the fundamental asymmetric mode (A 0 ) of the purely plasmonic waveguide.The electric field lines are sketched for these modes.

Fig. 3 .
Fig. 3. (a) The normalized electric field profiles of TM 0 , S 0 , H TM,0 and H TM,1 .(b) The normalized electric field profiles of TE 0 , A 0 , H TE,0 and H TE,1 .The width w of the Au layer is 620 nm.The wavelength λ is 800 nm.The effective index n eff for each mode is listed.

Fig. 4 .
Fig. 4. Dispersion characteristic of the four supermodes (H TM,0 , H TM,1 , H TE,0 and H TE,1 )of the hybrid photonic-plasmonic waveguide in Fig.2(a) in the form of the real and imaginary parts of the mode effective index versus the width w of the Au strip.All other dimensions are the same as those in Fig.2(a).The wavelength λ is 800 nm.

Fig. 5 .
Fig. 5.The normalized electric field profiles of H TM,0 and H TM,1 .The width w of the Au strip is 60 nm.While the H TM,0 field in highly concentrated in the metallic region, the H TM,1 field has considerable strength outside the metallic region and becomes more similar to the TM 0 mode of the corresponding purely photonic waveguide.

Fig. 6 .
Fig.6.Normalized electric field patterns in the planes horizontally (Y = 320 nm) and vertically (X = 0) cutting through the Au layer.The length L of the Au triangle is 900 nm, and the width W is 400 nm.The calculated field concentration factor (FCF) is 12.6 with the radius of curvature a at the tip being 20 nm.Q is the apex point of the triangular taper.

Fig. 8 .
Fig. 8. Spectra of transmission T , reflection R and the sum of the two for the Au taper length (a) L = 1 µm, and (b) L = 2 µm.The maximum width W of the taper is 400 nm.The radius of curvature a at the tip is 20 nm.

Fig. 9 .
Fig. 9. Spectra of transmission and FCF for three groups of lengths of the Au triangle (a) 0.4 µm, 0.425 µm and 0.45 µm, (b) 1 µm, 1.025 µm and 1.05 µm, and (c) 2 µm, 2.025 µm and 2.05 µm.The width W of the triangle is 400 nm.The radius of curvature a at the tip is 20 nm.All other parameters are the same as for the tapered structure shown in Fig. 6.