Abstract
We investigate electromagnetic scattering from nanoscale wires and reveal the emergence of a family of exotic resonances for source waves close to grazing incidence. These grazing resonances have a much higher Q-bandwidth product and thus, a much higher Q factor and broader bandwidth than the pure plasmonic resonances found in metal nanowires. Furthermore, these grazing resonances are much less susceptible to material losses than surface plasmon resonances. Contrary to the process of exciting surface plasmon resonances, these grazing resonances can arise in both dielectric and metallic nanowires and appear near to the cutoff wavelength of the circular waveguide. This peculiar resonance effect originates from the excitation of long range guided surface waves through the interplay of coherently scattered continuum modes coupled with first-order azimuthal propagating modes of the cylindrical nanowire. These first-order cyclic Sommerfeld waves and associated cyclic Sommerfeld resonances revealed here opens up the possibility of an alternative scheme of enhanced fields with a better merit (higher Q-bandwidth product and lower loss) than conventional surface plasmon resonances in the nano-regime. This nanowire resonance phenomenon can be utilized in broad scientific areas, including: metamaterial designs, nanophotonic integration, nanoantennas, and nanosensors.
©2009 Optical Society of America
1. Introduction
With the ever increasing advances in nanofabrication techniques, many remarkable optical phenomena have been discovered, thus stimulating considerable interest in light scattering at the nanometer length scale. Within the emerging field of nanoplasmonics, metallic nanostructures can be tailored to harness collective optical effects, namely, surface plasmon resonances (SPRs). Localized field enhancement is one of the key underlying physical characteristics in promising new nanotechnological applications. Due to relatively high energy confinement (Q factor) and small mode volume, SPRs have been used in the construction of tightly packed photonic devices [1], and play a major role in such applications [2–9] as, surface-enhanced Raman scattering [2, 3], photoconversion [4, 5], and Metamaterial designs [8, 9]. However, SPRs are extremely sensitive to material loss, which diminishes the Q factor significantly and can limit its benefits for the envisaged applications. Discovery of a new subwavelength field-enhancement mechanism would be greatly beneficial in future nanoscale research.
It is well known that surface plasmons cannot be excited by TM polarization (H is perpendicular to the axis of the rod) [10, 11]. With TE polarization SPRs can be strongly excited at normal incidence, but diminish as the incident angle approaches grazing incidence. Although scattering by a cylindrical object is an old topic [12–16], grazing scattering has a broad applicability [17–19] and is a nontrivial and longstanding problem due to singularities arising at the zero grazing angle, and can be further complicated by the emission of leaky waves from the cylinder [20, 21]. As the grazing angle approaches zero, intrinsic singularities in the scattering solutions reveal possible numerical instabilities. The general solution intended for non-grazing incidence often fails to provide a clear picture of this limiting case due to divergences arising at the zero grazing angle. Even in the macroscopic and long wavelength regime, descriptions of grazing behaviors often lead to contradictions [19].
In this paper, we investigate unexplored and nontrivial grazing phenomena in long nanowires. We undertake the analytic and numerical challenges inherent in grazing scattering at the nanoscale and develop a robust analytic solution to Maxwell’s equations for grazing incidence, from which the essential physics can be extracted. At grazing incidence, scattering and surface wave propagation are interconnected. Physically meaningful concepts like scattering cross section and efficiency Q must therefore be redefined. By taking proper account of the guided surface waves, we generalize the definition of the scattering efficiency. Our calculations reveal a family of non-plasmonic resonances with a much higher Q-bandwidth product and which are more sustainable to material loss than SPRs. It is determined that these resonances, or enhanced fields, originate from the excitation of long-range guided asymmetric surface waves corresponding to the n = 1 azimuthal mode of the nanorods, akin to Sommerfeld waves, traditionally ascribed to the azimuthally symmetric (n = 0) case for TM waves propagating along a conducting cylinder [22]. However, these higher order (n ≠ 0) periodic solutions have been historically overlooked due to the typically high attenuation at long wavelengths [23]. It is remarkable that indeed grazing light can intricately couple to these first order (n = 1) cyclic surface waves, i.e. cyclic Sommerfeld waves (CSWs), and contribute to exotic grazing resonances [24], or cyclic Sommerfeld resonances (CSRs). The cyclic Sommerfeld wave is an azimuthally periodic surface wave and is formed by two azimuthally counter-propagating helical surface waves along the rod. As will be shown, CSRs have a much higher Q-bandwidth product and thus, a much higher Q factor and broader bandwidth than SPRs. CSRs cannot in general be trivially scaled to other spectral regimes due to material dispersion. Contrary to the quasistatic nature [25] of SPRs, these grazing resonances are strictly non-electrostatic. This phenomenon of non-electrostatic resonances in the subwavelength regime was brought to light in a recent work involving the effect of volume plasmons on enhanced fields in ultrasmall structures [26], revealing the richness of resonant scatterings in the nanoscale.
2. General Solution at Oblique Incidence
Consider an infinitely long cylinder of radius a with its rotation axis along the z-direction [use cylindrical coordinates (ρ,ϕ, z)], and with permittivity ε 1 and permeability µ 1. The background permittivity ε 0 and permeability µ 0 are uniform throughout the space. A plane TM wave is incident onto the cylinder with a grazing angle θ with respect to the z-axis as shown in Fig. 1. Without loss of generality, assume the incident wave resides in the x-z plane with a harmonic time dependence, exp(-iωt). Thus, in cylindrical coordinates, the incident plane wave has components,
where x≡k 0 ρ sinθ, ψ(z)≡E 0exp(ik 0 z cosθ), , , and Jn(·) are Bessel functions of the first kind. Due to the non-normal incidence, the TM and TE modes are coupled in the scattered field. The general solution [27] involves an infinite series of Bessel functions. Here we take an alternative approach to derive the general solution. We first solve the Helmholtz equation in cylindrical coordinates for Ez and Hz. The remaining transverse field components are then determined from [28]:
where ⊥ refers to transverse dimensions and β 2 j=ω 2 εjµj-k 2 z,(j=0,1). When the grazing angle θ→0, β0→ 0. Thus, the transverse components have singularities two-orders higher than the longitudinal components. By matching tangential components of the fields at the cylinder surface (ρ=a), i.e., the continuity of E z,ϕ and Hz,ϕ, the general solution for both inside and outside the cylinder and for arbitrary incidence angles can be cast into the form,
where the index ν corresponds to z or ρ. Thus there are six components in Eq. (3). Inside the cylinder, Eq. (3) refers to the total field. Outside the cylinder, Eq. (3) refers to the scattered field, and the total field is the scattered field plus the incident field, E i +E s and H i+H s. For convenience, we define dimensionless quantities: εr≡ε 1/ε 0, µr≡µ 1/µ 0, , , δ≡sinθ, r≡k 0 ρ, r 0≡k 0 a, x≡k 0 ρδ, x 0≡k 0 aδ, x 1≡k 0 aη, y≡k 0 ρη, and z 0≡k 0 aη 0. The radial dependence of the scattered fields outside the cylinder [Eq. (3) with the superscript t=s] are given by
where Hn(x)≡H (1) n(x) is the first kind Hankel function, and
where n=0,1,2, 22EF;, and
Resonant modes mathematically correspond to the poles of the scattering matrix. Physically, resonant modes are related to the excitation of waveguide modes that can be found by determining where the denominators of the coefficients an and bn vanish, i.e. A 2 n+I ε n I µ n=0, which yields,
where k=ω/c and c is the velocity of light in vacuum. The ε 0r, ε 1r, µ 0r, and µ 1r are the relative permittivity and permeability in the corresponding medium.
Inside the cylinder the radial dependence of the fields [Eq. (3) with the superscript t=c] are,
where Jn(·) represents either Bessel or modified Bessel functions depending on the material properties of the cylinder. The coefficients are given by,
and b 0=d 0=0. The general solution involves an infinite series of Bessel functions, and often poses numerical difficulties for small grazing angles.
3. Grazing Incidence Solution
At near-zero grazing incidence, the general solution in Section 2 becomes singular. A wellbehaved grazing solution can be derived by asymptotic expansion of the general solution at the singularity θ = 0. Although the derivation is tedious and lengthy, this approach has the advantage of a simpler analytical solution (written solely in terms of elementary functions). Moreover, the grazing solution provides a simpler picture in which to extract the essential physical quantities of interest. From the grazing solution it is clear that only the n = 0 and n = 1 azimuthal modes have significant contribution to the EM field at the near-zero grazing angles. For convenience, a set of useful properties and asymptotic expansions of Bessel and Hankel functions with small arguments used here are listed in the Appendix. When the grazing angle θ → 0, δ → 0 and x 0→0, from the properties in the Appendix, the small-argument expansion of Eq. (6) yields, The reason for maintaining the third term δ 4 in the above equations is because the first two terms are canceled in the coefficients of an, bn, and the transverse components of the EM fields. Thus,
where n=1,2,3, ⋯, and
where
The quantities B n0, qn, α (1) n (x 0), and α (2) n (x 0) are given in the Appendix.
Substituting Eq. (10) into Eqs. (5) and (9), we obtain, for n≠0,
where
and
The n=0 terms require a separate treatment. From Eq. (5)
which, as θ→0, yields,
where
Substituting Eq. (17) into Eq. (9), we obtain
where
Substituting Eqs. (13)–(17) into Eq. (4), we obtain the asymptotic solution for the radial dependence of the scattered fields outside the cylinder:
where σ denotes ρ or ϕ, δ≡sinθ, x≡k 0 ρ δ, x 0≡k 0 aδ, and
and
In fact, in Eq. (23) only the D 1 remains and all the n>1 terms vanish as the grazing angle approaches zero. The other related coefficients in Eq. (21) are succinctly written as
where g (1) n and g (2) n are given by Eq. (15); χ (±) n is given by Eq. (12); and
Here α (1) n (x) and α (2) n (x) in Eqs. (24) and (25) are given in the Appendix [Eq. (37)]. The solution inside the cylinder is simpler, which can be obtained by substituting cn and dn in Eqs. (13) and (19) into Eq. (8). While the above solution was derived for TM polarization, the TE-mode solution can be obtained by imposing the property of duality of the fields: E→H, H→E, εr→µr, µr→εr, and 𝓨 0→-𝓩 0, where .
The waveguide eigenmode relation, Eq. (7), becomes singular in the limiting case (kz→k 0, β 0→0), and can lead to inaccurate mode solutions when solving Eq. (7) directly. Using the asymptotic expansions in the Appendix, the eigenmodes of Eq. (7) for this limiting case reduces to,
where is the background wave number. By taking the minimum of the absolute value of the left hand side of Eq. (26), the surface modes in the limiting case can be easily identified.
The grazing solution in Eq. (21) indicates that the scattered fields outside the nanorod are long-range guided surface waves. Furthermore, only the n=0 and 1 azimuthal modes have significant contribution at the near-zero grazing incidence. The n=0 term corresponds to the radially symmetric Sommerfeld wave. Historically, only the n=0 term is considered, and the n≠0 terms are neglected due to high losses in the radio frequency (RF) region [23]. However, in the optical regime and at grazing incidences, the n=1 term has a great impact on the resonance effects. Although the grazing solution is apparently complicated, it is written solely in terms of elementary functions, thus providing a clearer physical picture. Also, without the infinite series, the overall field expressions are simpler since only the n=0,1 terms need to be considered. Therefore, it is computationally efficient and more stable. To ensure the accuracy of the above asymptotic expansions, several tests have been conducted. Shown in Figs. 2–4 are the EM field coefficients in Eq. (21) for the case of n=0 and n=1 versus the grazing angle along with their counterparts for the exact result. We also studied the coefficients for other cases (n>1). In each case studied, all six components of the EM fields in the grazing solution accurately approached the general exact solution at low grazing angles. Overall, the grazing solution coincides with the exact result for grazing angles θ≲10°. Although the grazing angle can be arbitrarily small from a strictly mathematical perspective, it is clear that the given geometry sets physical limits for the minimum angle of incidence given by θ min=a/L, where L is the distance between the observation and incidence points; and a is the radius of the rod. Thus in what follows, θ→0 implies θ→θ min.
4. Grazing Resonance
The quality factor of a resonator is usually linked to the photon life time [29, 30]. In the scattering cases, the scattering efficiency Q is related to the scattered field [16, 27]. At grazing incidence, both the scattered waves and excited surface waves are interconnected. To incorporate both effects, it is convenient to define a θ-dependent total Qθ factor,
where
Here is the input intensity (𝓨 0 is the background admittance) and R is the radius of the integration circle around the rod. The Poynting vector components Ssρ and Ssz are along the ρ̂ and ẑ directions respectively, and are calculated from the scattered waves. The quantity Q s ρ describes resonant scattering and the Qsz describes guided surface-wave excitations. The Qθ factor is a useful metric to characterize the resonant scattering and the efficiency of channeling of the EM field through the nanowire. The Qθ appropriately reduces to the traditional Q when θ=90° (normal incidence). Our studies have shown that Qsρ is independent of R for all points outside of the cylinder, as is the case for the traditional Q. Since the radius of the rod is on the order of the skin depth of the metal, the EM field can fully penetrate the nanowire. To describe the effective power of the surface wave inside the nanowire, similar to the definition of the Qsz, we define,
where Scz is the z-component Poynting vector inside the cylinder. Thus, the factor Qcz represents the efficiency of the guided wave propagating inside the nanorod. We also define a dimensionless resonance bandwidth:
where λ 1 and λ 2 are the wavelengths corresponding to the half-maximum value of the Qθ of the resonance; and a is the radius of the rod. Thus, the dimensionless Q-bandwidth product (QBW) is given by QBW≡Qθ×BW. Although the bandwidth of the resonances is in the several tens of THz, due to the subwavelength nature of the problem, i.e. a≪λ, the dimensionless BW defined by Eq. (30) can be very small.
In the simulations, the permittivity of the silver nanorods was obtained from a polynomial fit to experimental data [31], appropriate for the visible spectrum of interest:
where λ is in units of µm. For the results below, we take R=3a with a=15 nm for all the figures. All field quantities are normalized with respect to the input field. For a silver nanorod, the SPR at λ≈309 nm can be excited by a TE wave at normal incidence, as shown in the left plot in Fig. 5. The frequency of the SPR is insensitive to the rod diameter [25], however the Q factor is sensitive to material loss. At low grazing angles, this picture changes dramatically: a series of non-plasmonic resonances emerge as shown in the narrow peak in the Qθ factor. Although Qθ demonstrates the resonant peaks at both normal and grazing incidences, the underlying physical mechanism of these resonances is different. The QBW of these grazing resonances is much higher than that of SPRs (0.7 vs. 0.04). Hence, the grazing resonances can have a much higher Q factor and broader bandwidth than the SPRs, as clearly indicated in Fig. 5. Moreover, by comparing the blue and green curves in Fig. 5, it can be seen that these grazing resonances are much less susceptible to material loss than the usual SPR. As the incident angle transitions from normal to grazing, the SPR continuously transforms into a new resonant mode, i.e. cyclic Sommerfeld resonance (CSR), with a corresponding redshift of the resonant frequency. This result is consistent with the waveguide modes calculated from Eq. (26) for the nanorods, which are shown in the right plot of Fig. 6. As kz/k 0 tends to unity, small changes in the grazing angle result in significant shifts in the resonant wavelength. This explains the features found in the left plot of Fig. 6, where the resonant peak continuously shifts to longer wavelengths when approaching the minimum angle. As indicated in Fig. 6, the grazing resonances appear near to the cutoff wavelength of n=1 azimuthal modes of the circular waveguide. The field enhancement near to the cutoff wavelength was also demonstrated in other geometry [32] and appear to be a universal feature of subwavelength structures. Due to the material dispersion, the grazing resonance cannot in general be trivially scaled to other spectral regimes. The grazing resonances reported here are consistent with a Q factor enhancement reported in previous work [15, 16].
In Fig. 7, we illustrate the robust nature of CSRs for a dielectric and metallic nanorod with |ε 1r|=4 and µ 1r=1. These grazing resonances again correspond to the excitation of first-order (n=1) cyclic Sommerfeld waves. The upper left panel demonstrates how Qθ and the wavelength of grazing resonances increases continuously as the grazing angle approaches θ min. Both the dielectric and metallic cases have resonances with similar magnitudes but with a clear shift in the wavelength. The lower-left panel shows the corresponding n=1 waveguide modes calculated from Eq. (26) and with the same kz values of the CSRs. Again, for both cases, we see that as kz/k 0 approaches unity, the resonant wavelength is highly sensitive to small changes in the grazing angle. The upper-right panel is the calculated Q bandwidth product of the corresponding CSRs as a function of grazing angle. There is a monotonic decline that levels off at larger impact angles. In contrast, the true bandwidth, ΔF, of the corresponding CSRs increases with grazing angle (lower right panel).
The correlation between the grazing resonances and the n=1 cyclic periodic surface waves are clearly shown in the upper-left panel of Fig. 8, demonstrating the consistency of the two curves identifying the resonant wavelengths. Thus, these grazing resonances result from the excitation of two first-order azimuthally counter-propagating guided waves, which travel with a spiraling trajectory along the nanowire and constructive interference significantly enhances the resonant scatterings. The enhancement of the Q factor at near-zero grazing angles is due to guided asymmetric surface waves and a coherent superposition of scattering states, which reinforces the resonance effect. The CSR also complies with the general rule that an increase of the Q factor implies a decrease of the bandwidth. As illustrated in the lower panels of Fig. 8, the increase of the Qθ (lower-left plot) is accompanied by the decrease of the bandwidth ΔF (lower-right plot) in the grazing region.
Another interesting phenomenon is that while the excited surface wave propagates forward outside the nanorod, the power flows backward inside the rod due to the boundary conditions and negative permittivity of silver. This backward propagation is expressed as the negative value of the quantity Qcz since it describes the axial power flow inside the rod, as shown in the upper-left plot of Fig. 9. When the grazing angle approaches zero, more energy in the incident wave is transformed into the guided surface waves, and thus the absolute values of both Qcz and Qsz increase as shown in the upper panels of Fig. 9. The ratio Qcz/Qsz in the lower-left plot of Fig. 9 describes the ratio of the axial power inside and outside the nanorod. As the grazing angle decreases, the scattering in the radial direction becomes weaker and wave guiding along the rod becomes stronger. Since the power inside the rod arises from the radial scattering, the value |Qcz/Qsz | decreases as θ→θ min. Shown in the lower-right plot of Fig. 9 is the ratio Qsz/Qθ versus the grazing angle. This value increases as approaching the minimum angle and the quantity Qsz becomes dominant in the total quality factor Qθ due to the increase of the guiding effect with the decrease of the grazing angle. To better visualize the spatial characteristics of the resonant fields, we show in Fig. 10 the Poynting vector from several perspectives at near-zero grazing incidence. As shown in the left panel, the power (Sρ,Sx,Sy) in the x-y plane is toward the center of the rod and has nearly perfect circular symmetry. The degree of symmetry depends on how small the angle is. However, once the energy penetrates the wire, it travels asymmetrically backward along the silver nanowire, as shown in the upper-right panel. Figure 11 shows the intensity of the total electric field close to the surface ρ=a inside and outside the silver rod. It reveals a two-dimensional standing wave pattern in the azimuthal and propagation directions, as the result of phase-matched spiraling propagation along the nanorod. The geometric influence on the grazing resonances is shown in Fig. 12 for a lossless silver nanowire. In the upper-left panel, both the resonant wavelength and the corresponding QBW increase with the increase of the radius of the nanorod. The lower-left panel shows the total quality factor Qθ and corresponding bandwidth of CSRs versus the radius of the rod. In the lossless nanorod, the total quality factor Qθ decreases with the increase of the radius, indicating this peculiar high Q grazing resonance is a nano-phenonmenon or a phenomenon more pronounced in the nano-region. The smaller the radius, the stronger the guiding of the surface wave along the nanorod. Due to the finite skin depth of silver, when the radius of the nanorod increases, less EM power penetrates the rod, and thus the absolute value of the quantity Qcz reduces, as shown in the upper-right panel of Fig. 12. The lower-right panel shows the ratio Qcz/Qsz versus the radius of the rod, showing again the backward propagation and the decrease of the power inside the silver nanorod as the radius increases.
5. Summary
In summary, we have investigated resonant scatterings of nanowires at near-zero grazing incidences through a newly developed grazing solution. We also interpreted grazing resonances in terms of the excitation of the first-order cyclic Sommerfeld waves and pointed out the non-plasmonic nature of cyclic Sommerfeld resonances. The extraordinary high Q value of these grazing resonances is due to the strong guiding of asymmetric long-range surface waves along the nanorods. This result enriches our fundamental understanding of scattering on the nanoscale and its relevance to many areas within nanotechnology. Since CSRs are associated with the natural modes of the nanorods, they may also be excited by other means. The merit of high Q, broadband, and low loss may render CSRs as an attractive alternative mechanism for enhanced-field applications in the nano-regime.
Appendix
Here we list a set of useful properties and small argument asymptotic expansions of Bessel and Hankel functions. The general properties of the family of Bessel functions used in the main text are:
where 𝓙n(x) represents either Bessel or Hankel functions.
Small argument (x→0) asymptotic expansions of the Bessel functions and their derivatives:
Asymptotic expansions (x→0) of the first kind Hankel functions and their derivatives:
where
where γu=0.5772156649 is Euler’s constant. For n≥3, the asymptotic expansions (x→0) of the first kind Hankel functions and their derivative are given by:
where Γ(·) is Gamma function Γ(n+1)=nΓ(n)=n!, and
It is convenient to define the quantity
Notice that B-n(z)=Bn(z). The derivative is given by
When z≡k 0 aη(δ) and δ≡sinθ, as the grazing angle approaches zero, δ→0, we have
where , B n 0≡B n(z 0), z 0=k 0 a η 0, and
A useful expansion for the ratio of Hankel functions when 0<x 0≤x and x→0:
where
Acknowledgments
The authors gratefully acknowledge G. A. Lindsay, Z. Sechrist, and G. Ostrom for valuable discussions and the support from the Office of Naval Research (ONR), as well as NAVAIR’s ILIR program from ONR.
References and links
1. S. A. Maier, “Plasmonics: The Promise of Highly Integrated Optical Devices,” IEEE J. Sel. Top. Quantum Electron. 12, 1671 (2006). [CrossRef]
2. S. Nie and S. R. Emory, “Probing Single Molecules and Single Nanoparticles by Surface-Enhanced Raman Scattering,” Science 275, 1102 (1997). [CrossRef] [PubMed]
3. H. Xu, E. J. Bjerneld, M. Ḱ’all, and L. B́’orjesson, “Spectroscopy of Single Hemoglobin Molecules by Surface Enhanced Raman Scattering,” Phys. Rev. Lett. 83, 4357 (1999). [CrossRef]
4. K. R. Catchpole and A. Polman, “Plasmonic solar cells,” Opt. Express 16, 21793 (2008). [CrossRef] [PubMed]
5. C. Rockstuhl, S. Fahr, and F. Lederer, “Absorption enhancement in solar cells by localized plasmon polaritons,” J. Appl. Phys. 104, 123102 (2008). [CrossRef]
6. K. Halterman, J. M. Elson, and S. Singh, “Plasmonic resonances and electromagnetic forces between coupled silver nanowires,” Phys. Rev. B 72, 075429 (2005). [CrossRef]
7. S. Feng and J. M. Elson, “Diffraction-suppressed high-resolution imaging through metallodielectric nanofilms,” Opt. Express 14, 216 (2006). [CrossRef] [PubMed]
8. J. Valentine, S. Zhang, T. Zentgraf, E. U. Avila, D. A. Genov, G. Bartal, and X. Zhang, “Three-dimensional optical metamaterial with a negative refractive index.” Nature (London) 455, 376 (2008). [CrossRef]
9. V. M. Shalaev, W. Cai, U. K. Chettiar, H.-K. Yua, A. K. Sarychev, V. P. Drachev, and A. V. Kildishev, “Negative index of refraction in optical metamaterials.” Opt. Lett. 30, 3356 (2005). [CrossRef]
10. J. P. Kottmann and O. J. F. Martin, “Retardation-induced plasmon resonances in coupled nanoparticles,” Opt. Lett. 26, 1096 (2001). [CrossRef]
11. B. S. Luk’yanchuk and V. Ternovsky, “Light scattering by a thin wire with a surface-plasmon resonance: Bifurcations of the Poynting vector field,” Phys. Rev. B 73, 235432 (2006). [CrossRef]
12. L. Rayleigh, “On the Electromagnetic Theory of Light,” Philos. Mag. 12, 81–101 (1881).
13. J. R. Wait, “Scattering Of A Plane Wave From A Circular Dielectric Cylinder At Oblique Incidence,” Can. J. Phys. 33, 189 (1955). [CrossRef]
14. J. R. Wait, “The Long Wavelength Limit In Scattering From A Dielectric Cylinder At Oblique Incidence,” Can. J. Phys. 43, 2212 (1965). [CrossRef]
15. L. G. Guimarães and J. P. R. Furtado de Mendonca, “Analysis of the resonant scattering of light by cylinders at oblique incidence,” Appl. Opt. 36, 8010 (1997). [CrossRef]
16. A. Lind and J. Greenberg, “Electromagnetic Scattering by Obliquely Oriented Cylinders,” J. Appl. Phys. 37, 3195 (1965). [CrossRef]
17. P. Rousseau, H. Khemliche, A. G. Borisov, and P. Roncin, “Quantum Scattering of Fast Atoms and Molecules on Surfaces,” Phys. Rev. Lett. 98, 016104 (2007) [CrossRef] [PubMed]
18. Z.-H. Gu, I. M. Fuks, and M. Ciftan, “Enhanced backscattering at grazing angles,” Opt. Lett. 27, 2067 (2002). [CrossRef]
19. D. E. Barrick, “Grazing Behavior of Scatter and Propagation Above Any Rough Surface,” IEEE Trans. Antennas Propag. 46, 73 (1998). [CrossRef]
20. A. W. Snyder and D. J. Mitchell, “Leaky rays on circular optical fibers,” J. Opt. Soc. Am. 64, 599 (1974). [CrossRef]
21. R. Sammut and A. W. Snyder, “Leaky modes on circular optical waveguides,” Appl. Opt. 15, 477 (1976). [CrossRef] [PubMed]
22. M. J. King and J. C. Wiltse, “Surface-Wave Propagation on Coated or Uncoated Metal Wires at Millimeter Wavelengths,” IEEE Trans. Antennas Propag. 10246 (1962). [CrossRef]
23. J. A. Stratton, Electromagnetic Theory, (McGraw-Hill, New York, 1941).
24. S. Feng and K. Halterman, “Exotic Grazing Resonances in Nanowires,” http://arxiv.org/pdf/0905.1357.
25. D. R. Fredkin and I. D. Mayergoyz, “Resonant Behavior of Dielectric Objects (Electrostatic Resonances),” Phys. Rev. Lett. 91, 253902 (2003). [CrossRef]
26. E. Feigenbaum and M. Orenstein, “Ultrasmall Volume Plasmons, yet with Complete Retardation Effects,” Phys. Rev. Lett. 101, 163902 (2008). [CrossRef] [PubMed]
27. C. F. Bohren and D. R. Huffman, Absorption and Scattering of Light by Small Particles, (Wiley, Weinheim, Germany, 2004).
28. J. D. Jackson, Classical Electrodynamics, 3rd ed. (Wiley, New York, 1999).
29. P. R. Villeneuve, S. Fan, and J. D. Joannopoulos, “Microcavities in photonic crystals: Mode symmetry, tunability, and coupling efficiency,” Phys. Rev. B 54, 7837 (1996). [CrossRef]
30. E. Centeno and D. Felbacq, “Characterization of defect modes in finite bidimensional photonic crystals,” J. Opt. Soc. Am. A 16, 2705 (1999). [CrossRef]
31. M. M. J. Treacy, “Dynamical diffraction explanation of the anomalous transmission of light through metallic gratings,” Phys. Rev. B 66, 195105 (2002). [CrossRef]
32. F. J. García-Vidal, E. Moreno, J. A. Porto, and L. Martín-Moreno, “Transmission of Light through a Single Rectangular Hole,” Phys. Rev. Lett. 95, 103901 (2005). [CrossRef] [PubMed]