Plasmonics (2011) 6:295–299

DOI 10.1007/s11468-011-9203-x

Enhancing the Dual-Band Guiding Capabilities of Coaxial

Spoof Plasmons via use of Transmission Line Concepts
Miguel Navarro-Cía & Miguel Beruete & Mario Sorolla &
Stefan A. Maier

Received: 1 October 2010 / Accepted: 12 January 2011 / Published online: 3 February 2011
# Springer Science+Business Media, LLC 2011

Abstract We derive closed analytical forms for the attainable at optical frequencies by exploiting surface
response of coaxial spoof plasmons, aided by transmis- plasmon polaritons (SPPs) [1]. These localized modes show
sion line concepts under the effective complex surface promise for future nanocircuitry and sensing applications
impedance framework. This constitutes a powerful [2–4]. To extend the range of SPPs and its inherent
platform to improve as well as to elucidate designs with subwavelength localization to THz frequencies, where SPPs
enhanced performances. In particular, we propose a dual- based on metals develop into loosely bound Sommerfeld–
band spoof plasmon waveguiding geometry with the Zenneck surface waves, researchers have started investigat-
higher order slow-wave mode operating well below the ing spoof SPPs [5–10]. At THz frequencies these geomet-
regime governed by dispersion of periodic guides (Bragg rically induced surface waves give rise to tightly bound
reflections at Brillouin zone boundaries), that is, diffrac- modes in analogy to both their optical counterpart SPPs,
tion. The analysis is supported by eigen mode numerical and indeed, also to complex surface waves extensively
calculations. As an example in a waveguide device studied at microwave frequencies in the past [11, 12]. This
context, we demonstrate the dual-band planar routing feature opens up the path for efficient planar guiding in a
ability of elliptical–coaxial cable-based spoof plasmons regime where it was practically limited to Goubau lines [13,
along a straight chain as well as a Y-splitter. 14]. Consequently, sensing techniques at THz frequencies
[15], a spectral region rich of fingerprints, may benefit from
Keywords Spoof plasmons . Surface impedance . this strategy.
Terahertz . Transmission line . Dual-band From the design perspective, it is desirable to have a
complete recipe for analyzing and optimizing spoof SPP
geometries as well as discovering new topologies. For more
Introduction than a century, the microwave engineering community has
successfully employed transmission line concepts and
Strong confinement of electromagnetic energy below the equivalent circuit models to simplify complex problems
diffraction limit along a metal–dielectric interface is [16]. Using these concepts, we have recently been able to
explain the response of several spoof SPP geometries. Also,
M. Navarro-Cía (*) : M. Beruete : M. Sorolla
we proposed a complementary-split-ring-resonator-based
Millimeter and Terahertz Waves Laboratory, spoof SPP, which improves many of the characteristics
Universidad Pública de Navarra, Campus Arrosadía, exhibited by slit arrays or hole arrays spoof SPPs, such as
31006 Pamplona, Spain out-of-plane confinement, substrate thickness, or bandwidth
e-mail: [email protected]
response [17].
M. Navarro-Cía Another strategy for enhancing the bandwidth over
e-mail: [email protected] which spoof SPPs allow high-confinement guiding lies in
the use of holey surfaces sustaining modes in two distinct
M. Navarro-Cía : S. A. Maier
spoof plasmon bands: dual-band spoof SPPs supported by
Experimental Solid State Group, Physics Department,
Imperial College London, arrays of coaxial (annular) holes [18]. While the lowest
London SW7 2AZ, UK order spoof mode, originating from the fundamental
296 Plasmonics (2011) 6:295–299

transversal electromagnetic (TEM) coaxial mode, showed formula Zin ¼ Zs ¼ j  ZTL  tanðbhÞ, where ZTL is the
high confinement (the flat part of the dispersion is reached characteristic impedance, β the wave vector, and h the
well before the Brillouin zone edge) in the experimental height of the perturbation [16] as shown in Fig. 1(a). Notice
demonstration [18], the confinement of the higher order that the time variation is assumed to be ejωt, where ω is the
mode was governed by the edge of the Brillouin zone, angular frequency. For the first two lowest order modes—
hence, attaining its spatial confinement via coherent fundamental TEM and the first higher order transversal
scattering (diffraction). As we will demonstrate below, we electric (TE11)—of the coaxial hole, we can hence write:
can take advantage of transmission line concepts to have a
closed analytical form for the response. And relying on it, Zin TEM ¼ Zs TEM ¼ j  ZTL  tanðk0 hÞ ð1aÞ
we can explain the dual-band spoof SPP behavior in this
coaxial cable-based geometry in terms of changes to the
surface impedance, as well as how to modify the initial 0sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1
 2
geometry to enhance the out-of-plane confinement of the @ 2
Zin TE11 ¼ Zs TE11 ¼ j  ZTL  tan k0 
2 hA
higher order mode. aþb

ð1bÞ
Theory: Transmission Line and Surface Impedance
Framework where k0 is the free space wave vector and 2/(a+b) is the
common approximation used in practice for the cutoff wave
The starting unit cell under study is depicted in Fig. 1(a) vector of the TE11 [16].
along with its parameters: outer and inner radius a and b, It is well-known that p-polarized surface waves like spoof
respectively, and shorted coaxial line height h, which was SPPs are supported when the reactive part of the complex
set to h=59 μm in analogy to the structure studied in Ref. surface impedance is inductive [12, 17, 19]. Moreover, the
[18]. The metal is modeled initially in this letter as a perfect larger the inductive response, the higher the confinement of
electric conductor (which is a good approximation at THz), the p-polarized surface wave to the interface [12, 19].
which prevents the penetration of electromagnetic fields Therefore, we are looking for those values in Eq. (1a) and
into them, making all lengths involved in the system (1b), which make the tangent function go to infinity. For the
scalable. We assume that the shorted coaxial line is TEM mode, that happens when h=λ0/4 [6], thus for our
replicated on a rectangular lattice with periods dx =dz =d= dimensions f=c/4h=1.27 THz, where c is the speed of light
50 μm and that both the coaxial hole dimensions and the in- in vacuum. However, for the TE11 mode, it occurs at higher
plane lattice constant d are much smaller than the frequencies since βTEM =k0 >βTE11. For a relatively narrow
wavelength of radiation, a < d << λ 0. Under this annulus, say a=19 μm and b=23 μm, the cutoff frequency
subwavelength regime approximation, the effective surface of the TE11 mode is 2.27 THz and its spoof surface plasmon
impedance can be readily linked to the input impedance of frequency (the frequency at which the input impedance
the coaxial cable [17, 19]. To analytically compute it, we diverges) is fSSPP_TE11 =2.60 THz according to Eq. (1b). This
calculate the input impedance of the transmission line with is well below λ0/2=3 THz, and therefore, we can safely state
a short circuit at the bottom following the well-known that the surface wave cutoff is not dominated by diffraction.

Fig. 1 a Unit cell along with the

main parameters. b Dispersion
diagram calculated using the
eigen mode solver of CST
Microwave StudioTM. kz is the
wave vector parallel to the surface
in the direction of propagation.
Dotted lines indicate the analytical
fSSPP. Insets electric field distribu-
tion at the first—TEM (bottom
right)—and second—TE11 (upper
left)—fSSPP along the xz cutting
plane just above the surface and
yz cutting plane (color online)
Plasmonics (2011) 6:295–299 297

Numerical Results: Dispersion Diagram and Field

Distribution

To confirm the previous discussion, we compute the

dispersion diagram by using the commercial software
CST Microwave StudioTM, see Fig. 1(b). Periodic bound-
ary conditions are applied along the in-plane periodicities,
and the phase difference between faces of periodic bound-
aries along z (direction of propagation) is varied. An
excellent agreement between the spoof surface plasmon
frequencies calculated analytically previously (shown in
dotted lines in Fig. 1(b)) and those given by the numerical
simulation is obtained, with less than 8% of deviation. Also,
in Fig. 1, the electric field distribution of the eigen mode at
each spoof surface plasmon resonance is plotted, demon-
strating their TEM and TE11 character, respectively. Notice
that the first eigen mode has no y-component in the
coaxial line. Also, its magnetic field (not shown) does not
have y-component. Bear in mind that in a conventional
coaxial cable the direction of propagation would be y, the
first eigen mode shows neither electric nor magnetic field
in the direction of propagation. Thus, it is termed TEM.
Accounting now for the second eigen mode, it displays no
electric field in axial direction of the coaxial, y-direction,
but it does magnetic field (not shown). These features are Fig. 2 Dispersion diagram: a as a function of ax while dz, dx, and az
characteristic of a TE wave. are kept to the fixed values dz =50 μm, dx =100 μm, az =23 μm, and
By inspection of Eq. (1b), one may notice that the larger bx,z =0.7ax,z (the central ellipse is scaled down 0.7 with respect to the
a and b are, the lower the fSSPP is. Thus, it would be desired outer ellipse). Dotted lines indicate the semi-analytical fSSPP; b as a
function of the scale factor bx,z/ax,z between central and outer ellipse
to enlarge them as much as possible. Considering annular with ax =48 μm (color online)
holes, the limit is imposed by the period d. Nevertheless,
we can have one more degree of freedom if we consider
elliptical holes. The lowest order modes remain of TEM two plots, the line of reasoning of the previous paragraph is
and TE11 character [16, 20], described by Eq. (1a) and (1b). demonstrated with a lower TE11 spoof surface plasmon
Needless to say, as an intuitive idea, in this case a and b in frequency for ellipses with larger eccentricity and scale factor.
Eq. (1b) should be substituted by effective parameters aeff Moreover, the fSSPP_TE11 calculated through Eq. (1b) and
and beff, which may be larger than a and b to account for those from the dispersion diagram are in agreement to more
the larger area covered by the ellipse with respect to a circle than 96% of agreement, demonstrating the usefulness of
with radius equal to the minor axis of the ellipse. The transmission line concepts for designing and optimizing spoof
derivation of the exact characteristic equation for the cutoff SPPs. Conversely, the shift of the TEM spoof surface
wave vector of the TE11 in elliptical coaxial waveguide can plasmon frequency as a function of ax and scale factor is
be found in Ref. [20] and references therein. negligible, i.e., the fundamental mode does not suffer an
Figure 2(a) displays the dispersion curves calculated via appreciable frequency shift when ax or bx,z/ax,z changes
CST Microwave StudioTM for several elliptical coaxial holes (shown in Fig. 2(b) but not in Fig. 2(a) for the sake of
with different major axis: ax =23, 28, 33, 38, and 48 μm, clarity) and that is in agreement with the independence of Eq.
while az, h, and dz are kept as before at 23, 59, and 50 μm, (1a) to those parameters.
respectively, dx =100 μm, and the scale factor is bx/ax =bz/az =
0.7. For these parameters, the TE11 cutoff wave vectors are
calculated numerically, and then the cutoff frequencies Numerical Results: Planar Waveguide and Y-Splitter
fSSPP_TE11 according to Eq. (1b), leading to fSSPP_TE11 =2.76, Based on Elliptical Coaxial Spoof Plasmons
2.45, 2.22, 2.05, and 1.82 THz for ax =23, 28, 33, 38, and
48 μm, respectively. On the other hand, the dispersion Finally, we demonstrate dual-band subwavelength (both
diagram for an elliptical coaxial spoof SPP for a different in-plane as well as out-of-plane) planar routing with the
scale factor with ax =48 μm is depicted in Fig. 2(b). In these optimized elliptical coaxial spoof SPPs (h = 59 μm,
298 Plasmonics (2011) 6:295–299

Fig. 3 Cross-sectional view just

in the middle plane (zy cutting
plane) of the electric field Ey at a
f=1.11 THz (0.98fSSPP_TEM) and
c f=1.72 THz (0.98fSSPP_TE11) in
the straight planar waveguide.
Field amplitude Ey as a function
of the out-of-plane distance for:
b 1.088 (red), 1.105 (black), and
1.11 THz (blue) and d 1.68 (red),
1.70 (black), and 1.72 THz (blue;
color online)

dz =50 μm, az =23 μm, ax =48 μm, and scale factor 0.7). frequencies can be distinguished because of the field
To this end, a transient analysis of two different structures enhancement nearby the surface as Fig. 3(b) and (d) show
is performed, employing the finite-integration-time in more detail. There, the amplitude of Ey at the center of
domain-based code CST Microwave StudioTM: a chain the fifth coaxial against the distance with respect to the
of 14 elliptical coaxial holes (Fig. 3) and Y-splitter surface is presented for three different frequencies close to
composed of identical number of annular holes (Fig. 4), both fSSPP_TEM and fSSPP_TE11. In quantitative terms, the 1/e
which means a total structure length of 700 μm, field decay length from the surface is for the TEM mode 32
corresponding to 2.6λ and 4.1λ for the TEM and TE11 (≈0.12λ), 24 (≈0.09λ), and 22 μm (≈0.08λ) for 1.088,
spoof SPP frequency, respectively. Gaussian pulses cov- 1.105, and 1.11 THz, respectively. Also, this expected trend
ering a wide frequency range are launched as stimulation of higher confinement as the frequency approaches fSSPP is
signal from port 1 and perfectly matched layer absorbing confirmed for the TE11 mode whose exponential decay
boundary conditions are used in all the simulation length is 64 (≈0.36λ), 54 (≈0.31λ), and 46 μm (≈0.26λ) for
boundaries together with magnetic field symmetry along 1.68, 1.70, and 1.72 THz, respectively. Notice that the TE11
z. Port 1 is in essence a monopole antenna which radiates mode is less tight confined than TEM in accordance with
omnidirectionally and whose evanescent radiated energy the previous analysis. Conversely, the in-plane confinement
couples to the textured metal surface. This antenna is is 0.36λ and 0.56λ for the TEM and TE11 band, in that
placed on the right of Fig. 3 and on the top of Fig. 4. To order. Accounting now for losses, the computed attenuation
reduce the simulation effort and make the analysis more coefficient for this copper structure is α=0.0047 and
realistic, copper is taken for the metal with conductivity 0.0017 μm−1 for the TEM (1.11 THz) and TE11 spoof
equal to the bulk value of 5.8×107 Ω−1 m−1. SPP (1.72 THz), in that order.
Figure 3(a) and (c) show the y-component of the electric With regard to the Y-splitter, the electric energy density
field for the straight planar waveguide at f = 1.11 is shown in Fig. 4 at f=1.07 THz (0.94fSSPP_TEM)—left—
(0.98fSSPP_TEM) and 1.72 THz (0.98fSSPP_TE11), respective- and f=1.72 THz (0.98fSSPP_TE11)—right. There, the TEM
ly. A clear manifestation of the spoof SPP for both and TE11 characteristic is again revealed since the electric
energy density is distributed uniformly along the ellipse for
the TEM band but it has a null in the middle plane for the
TE11 mode. Finally, this figure nicely shows the ability to
split evenly, and thus guide as desired the TEM and TE11
spoof SPPs.

Conclusion

In conclusion, we have highlighted that engineering

concepts such as complex surface impedance in the
transmission line framework are shown to be simple, yet
arguably effective tools to design and optimize spoof
surface plasmon polaritons for dual-band planar guiding
Fig. 4 Electric energy density at f=0.94fSSPP_TEM (left) and f= with subwavelength confinement. It has been shown that
0.98fSSPP_TE11 (right) in the Y-splitter outline (color online) the analytical result is in very good agreement with
Plasmonics (2011) 6:295–299 299

numerical calculations. This formalism based on equivalent focusing on periodically corrugated metal wires. Phys Rev Lett
97(17):1–4
circuits may be helpful for moving forward in THz
8. Fernández-Domínguez AI, Williams CR, García-Vidal FJ, Martín-
technology as it has been for antenna and microwave Moreno L, Andrews SR, Maier SA (2008) Terahertz surface
engineering in the last century. And this, in turn, may have plasmon polaritons on a helically grooved wire. Appl Phys Lett 93
an impact in sensing techniques at THz frequencies, which (14):41109-1-3
9. Williams CR, Andrews SR, Maier SA, Fernández-Domínguez AI,
is a region of the spectrum that is attracting a lot of
Martín-Moreno L, García-Vidal FJ (2008) Highly confined
attention for sensing application nowadays. Also, this guiding of terahertz surface plasmon polaritons on structured
engineering interpretation could find potential extensions metal surfaces. Nat Photon 2(3):175–179
into the optical regime [21, 22], where it may facilitate the 10. Hibbins AP, Evans BR, Sambles JR (2005) Experimental
verification of designer surface plasmons. Science 308
design of surface plasmon resonance sensors. (5722):670–672
11. Ishimaru A (1990) Electromagnetic wave propagation, radiation,
Acknowledgment The authors are grateful to Dr. F. Falcone for and scattering. Prentice Hall, Englewood Cliffs
useful ideas and discussion. This work has been supported by 12. Collin RE (1991) Field theory of guided waves. Wiley, IEEE
Spanish Government under contract Consolider “Engineering Meta- Press, New York
materials” CSD2008-00066 and by the US Air Force Office of 13. Goubau G (1950) Surface waves and their application to
Scientific Research (AFOSR). Miguel Navarro-Cía also acknowl- transmission line. J Appl Phys 21(11):1119–1128
edges the exchange grant given by the European Science Foundation 14. Akalin T, Treizebré A, Bocquet B (2006) Single-wire transmission
(ESF) within the framework of the ESF Activity entitled “New lines at terahertz frequencies. IEEE Trans Microwave Theor Tech
Approaches to Biochemical Sensing with Plasmonic Nanobiopho- 54(6):2762–2767
tonics (PLASMON-BIONANOSENSE).” 15. Siegel PH (2004) Terahertz technology in biology and medicine.
IEEE Trans Microwave Theory Tech 52(10):2438–2447
16. Pozar D (2004) Microwave engineering. Wiley, New York
References 17. Navarro-Cía M, Beruete M, Agrafiotis S, Falcone F, Sorolla M,
Maier SA (2009) Broadband spoof plasmons and subwavelength
electromagnetic energy confinement on ultrathin metafilms. Opt
1. Raether H (1988) Surface plasmons. Springer, New York Express 17(20):18184–18195
2. Maier SA (2007) Plasmonics: fundamentals and applications. 18. Williams CR, Misra M, Andrews SR, Maier SA, Carretero-
Springer, New York Palacios S, Rodrigo SG, García-Vidal FJ, Martín-Moreno L
3. Homola J, Yee SS, Gauglitz G (1999) Surface plasmon resonance (2010) Dual band terahertz waveguiding on a planar metal surface
sensors: review, Sens. Actuators B 54(1–2):3–15 patterned with annular holes. Appl Phys Lett 96(1):011101-1-3
4. Lal S, Link S, Halas NJ (2007) Nano-optics from sensing to 19. Collin RE, Zucker FJ (1969) Antenna theory. McGraw-Hill, New
waveguiding. Nat Photon 1(11):641–648 York
5. Pendry JB, Martín-Moreno L, García-Vidal FJ (2004) Mimicking 20. Navarro R, Boria VE, Gimeno B, Coves A, Ferrando M (2000)
surface plasmons with structured surfaces. Science 305 Full modal analysis of confocal coaxial elliptical waveguides. IEE
(5685):847–848 Proc Microw Antennas Propag 147(5):374–380
6. García-Vidal FJ, Martín-Moreno L, Pendry JB (2005) Surfaces 21. Engheta N (2007) Circuits with light at nanoscale: optical
with holes in them: new plasmonic metamaterials. J Opt A Pure nanocircuits inspired by metamaterials. Science 317(5845):1698–
Appl Opt 7(2):S97–S101 1702
7. Maier SA, Andrews SR, Martín-Moreno L, García-Vidal FJ 22. Staffaroni M, Conway J, Vedantam S, Tang J, and Yablonovitch E
(2006) Terahertz surface plasmon-polariton propagation and (2010) Circuit analysis in metal-optics, arXiv:1006.3126v5.