IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. 49, NO.

9, SEPTEMBER 2013 777

Coupled Transmission Line Model for Planar
Metal-Dielectric-Metal Plasmonic Structures:
Inclusion of the First Non-Principal Mode
Meisam Bahadori, Ali Eshaghian, Mohsen Rezaei, Hossein Hodaei, and Khashayar Mehrany
AbstractA coupled transmission line model whose parame-
ters are analytically given is provided for fast and accurate
analysis of miscellaneous plasmonic structures with metal-
dielectric-metal (MDM) arrangement. The coupling of surface
plasmons supported by the planar metal-dielectric interfaces is
included in the proposed model and thereby the effects of the rst
non-principal mode of the MDM waveguide are considered. In
particular, MDM bends, junctions, and stubs are studied. Using
rigorous numerical solutions, e.g., the nite-element method and
the nite-difference time-domain, it is shown that the proposed
model is accurate in all those cases.
Index TermsCoupled transmission lines, impedance, optical
waveguides, plasmons, surface waves.
I. INTRODUCTION
I
T IS a well-known fact that the fundamental mode (TM
0
)
of plasmonic metal-dielectric-metal (MDM) waveguides
is strongly similar to the fundamental TEM mode of the
ideal parallel plate waveguide [1], [2]. It is therefore no
wonder that many passive microwave circuits, e.g. trans-
mission lines with multiple stubs, have MDM plasmonic
twins [3][6]. These microwave inspired MDM plasmonic
circuits can usually be analyzed by applying the conventional
techniques of microwave network analysis [7][9]. In par-
ticular, the standard transmission line theory; thanks to its
unparalleled simplicity, has been of much interest. The interest
was spurred by using the concept of impedance matching in
design of MDM splitters (T-junctions) [1]. In this work, each
MDM waveguide was modeled by a transmission line whose
characteristic impedance was given in terms of the propa-
gation constant and geometrical parameters. This idea was
further pursued in analysis and design of miscellaneous MDM
plasmonic devices, e.g. junctions [4], [10], lters [11], [12],
splitters [13], resonators [14], couplers [15], and plasmonic
demultiplexers based on slot cavities [16], [17].
The common feature of all these works is that single-
mode MDM waveguides in the structure are replaced by their
corresponding transmission line segments. In this manner, the
Manuscript received April 18, 2013; revised May 22, 2013 and July 14,
2013; accepted July 24, 2013. Date of publication July 26, 2013; date of
current version August 7, 2013.
The authors are with the Department of Electrical Engineering, Sharif Uni-
versity of Technology, Tehran 11155-4363, Iran (e-mail: meisambahadori@
gmail.com; eshaghian@ee.sharif.ir; mohsenr1986@gmail.com; hhodaei@
ee.sharif.ir; mehrany@sharif.edu).
Color versions of one or more of the gures in this paper are available
online at http://ieeexplore.ieee.org.
Digital Object Identier 10.1109/JQE.2013.2275005
effects of below cut-off higher order modes are neglected.
This negligence can have a crucial effect at waveguide dis-
continuities, e.g. at bends and junctions. Therefore, nding a
simple enough way to include higher order modes is highly
desirable. One such simple approach is to model each disconti-
nuity of the structure by a scattering matrix [18], [19]. Once all
scattering matrices are successfully obtained, the methods of
microwave network analysis remain applicable. Unfortunately,
this is not as simple as it sounds because it requires rigorous
analysis. For this reason, nding a simple transmission line
model that does not incur the unwanted burden of rigorous
analysis is still wanted.
It is the aim of this paper to propose a new approach
based on the transmission line theory to simply and accurately
analyze the effects of higher-order modes in planar MDM
plasmonic structures. To this end, the MDM waveguide is
represented by a coupled transmission line model that includes
the effects of the fundamental mode together with the rst next
non-principal mode. Wave propagation in MDM waveguides
is in this manner modeled by two coupled surface waves
each one traveling along one of the conductors in the coupled
transmission line model.
The organization of this paper is as follows. First, a cou-
pled transmission line model is proposed to consider the
coupling between the surface plasmon waves at the upper
and lower metal-dielectric interfaces of MDM waveguide. The
parameters of the proposed model are all given analytically.
The proposed model is then employed to analyze MDM
bends, asymmetric junctions, and single stub structures in
section III, IV, and V, respectively. Finally, the conclusions
are made in section VI.
II. COUPLED TRANSMISSION LINE MODEL FOR
MDM WAVEGUIDES
In this section, we propose a transmission line model for
surface plasmon propagation at the planar interface between
semi-innite metal and dielectric regions. In this manner, the
MDM waveguide can be represented by a coupled transmis-
sion line model that accounts for the coupling between surface
plasmons at the upper and lower metal-dielectric boundaries.
Fig. 1(a) shows the planar interface between metal
and dielectric semi-innite regions whose permittivities are
denoted by
m
and
d
, respectively. It is a well-known fact that
the complex propagation constant of the transverse magnetic
0018-9197 2013 IEEE
778 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. 49, NO. 9, SEPTEMBER 2013
Fig. 1. (a) The surface wave on a single interface between metal and
dielectric. (b) The per-unit-length equivalent lumped circuit model.
(TM) surface wave is:

s
= k
0
_

d

d
+
m
(1)
where k
0
stands for the free space wave-number [20]. This
complex propagation constant which depends on the complex
permittivity of metal (
m
), takes into account the propagation
loss of the surface wave along the metal-dielectric interface.
The transmission line model of this structure is schemat-
ically shown in Fig. 1(b). The propagation constant of the
transmission line model is obviously
s
. The complex char-
acteristic impedance of the transmission line model; however,
cannot be straightforwardly obtained because the structure is
inhomogeneous and the ratio of the transverse electric to the
transverse magnetic eld is not constant. Since the energy of
surface plasmon waves is mainly conned within the dielectric
region, the characteristic impedance of surface plasmons can
be reasonably approximated by:
Z
s

_
0

E
s
x
(x, z)dx
_
H
s
y
(0, z)dy
. (2)
In this way, the sought-after characteristic impedance can be
easily written in terms of the complex TM wave impedance
in dielectric media and the evanescent x-component of the
wave-vector in the dielectric region:
Z
s

d

d
(3a)

d
= 1
_
_

2
s
k
2
0

d
(3b)
where
d
is the penetration depth of the surface wave [2] in
dielectric normalized to a unit length in the y direction.
Now, the complex per-unit-length lumped circuit elements
of the transmission line model can be written as [2], [7]:
L
s
=

s
Z
s

(4a)
C
s
=

s
Z
s
. (4b)
Given that the transmission line model of planar metal-
dielectric interface is now available, the MDM waveguide
can be modeled by two coupled lossy transmission lines.
In accordance with Fig. 2(a), the MDM waveguide is in
fact nothing more than two adjacent planar metal-dielectric
interfaces. This is schematically shown in Fig. 2(b). The
complex per-unit-length lumped circuit elements of each of the
Fig. 2. (a) An MDM waveguide of width W formed by coupling between
surface waves. (b) The proposed per-unit-length lumped circuit model.
lossy transmission lines corresponding to the metal-dielectric
interfaces then read as:
L
s
=

s
Z

(5a)
C
s
=

s
Z

s
(5b)
where

s
and Z

s
are the perturbed complex propagation
constant and characteristic impedance of the metal-dielectric
interface. In the rst order approximation, the perturbed
propagation constant still follows Eq.(1). The characteris-
tic impedance; on the other hand, can change considerably
because the dielectric region is truncated and thus the per-
turbed complex characteristic impedance is modied as:
Z

s

_
W/2
W/2
E
s
x
(x, z)dx
_
H
s
y
(W/2, z)dy
. (6)
The coupling between the two lossy transmission lines is
governed by the complex per-unit-length mutual inductance;
L
m
, and capacitance; C
m
. The mathematical way to obtain L
m
and C
m
is to appropriately associate the complex propagation
constants of the rst two supermodes of the MDM waveguide
with the eigenvalues of the transmission line differential equa-
tions [15]:
L
m
=
_
_

s
_
2
1
_
L
s
(7a)
C
m
=
1
2
_

2
o
2
2
s

2
e
1
_
C
s
(7b)
where
e
is the complex propagation constant of the funda-
mental mode (TM
0
) calculated from
tanh
_
_

2
e
k
2
0

d
W
2
_
=
_

2
e
k
2
0

m
/
m
_

2
e
k
2
0

d
/
d
(8)
[21], [22], and
o
is the complex propagation constant of the
TM
1
mode calculated from
coth
_
_

2
o
k
2
0

d
W
2
_
=
_

2
o
k
2
0

m
/
m
_

2
o
k
2
0

d
/
d
. (9)
It is worth noting that in case the working frequency is
below the TM
1
cutoff, the real part of
e
is larger than its
BAHADORI et al.: COUPLED TRANSMISSION LINE MODEL FOR PLANAR MDM PLASMONIC STRUCTURES 779
imaginary part while the imaginary part of
o
is larger than
its real part. In case the working frequency is above the TM
1
cutoff, the real parts of
e
and
o
are both larger than their
imaginary parts.
Now that the complex per-unit-length circuit elements of the
coupled transmission line model are all determined, the voltage
and current distribution along the coupled transmission lines
can be written as
_
V
I
_
(z) = A
_
V
e
Y
e
V
e
_
e
j
e
z
+ B
_
V
e
Y
e
V
e
_
e
j
e
z
+C
_
V
o
Y
o
V
o
_
e
j
o
z
+ D
_
V
o
Y
o
V
o
_
e
j
o
z
(10)
where
V
e
=
_
1
1
_
, V
o
=
_
1
1
_
(11a)
V =
_
V
1
V
2
_
, I =
_
I
1
I
2
_
(11b)
and Y
e
and Y
o
stand for the even and odd complex character-
istic admittances, respectively:
Y
e
=
_
C
s
L
s
+ L
m
(12a)
Y
o
=
_
C
s
+2C
m
L
s
L
m
(12b)
The general form of voltage and current distribution as
expressed in Eq.(10) is to be used in the following sections.
III. COUPLED TRANSMISSION LINE MODEL FOR
MDM BENDS
To demonstrate the applicability of the proposed model,
plasmonic MDM bends are studied rst. Fig. 3 shows the
general case, where the bend angle is . This structure is
hereafter referred to as the MDM -bend. The width of the
MDM waveguides is denoted by W, and the permittivities
of the dielectric and metallic regions are represented by
d
,
and
m
, respectively. Two hypothetical reference planes are
indicated by AB- and AC-planes in Fig. 3. The transmission
line model of the structure at the left-side of the AB-plane and
the right-side of the AC-plane is already given in the previous
section. The transmission line model in between the AB- and
AC-planes remains to be found. This part of the model should
take into account the extra phase difference imposed on the
electromagnetic wave propagating along the lower edge of
the bend. Fortunately, this effect can be easily considered
by a single transmission line whose propagation constant and
characteristic impedance are
s
[from Eq. (1)], and Z

s
[from
Eq. (6)], respectively. The length of this transmission line
should be equal to the distance between points B and C in
the gure:
L = 2W tan(/2). (13)
In this manner, the complete model can be constructed in
accordance with the arrangement shown in Fig. 4. The effect
of the phase difference between the upper and lower edges of
Fig. 3. A typical MDM -bend with hypothetical planes AB, and AC.
Fig. 4. The schematic of the proposed coupled transmission line model for
the MDM -bend in Fig. 3.
the bend is reected in the impedance mismatch between the
coupled transmission lines at the left- and right-sides of the
AB- and AC-planes. Using the general propagation solution in
coupled transmission lines; Eq. (10), the scattering coefcients
of the even (S
e
11
, S
e
21
) and the odd (S
o
11
, S
o
21
) supermodes can
be easily obtained. These coefcients are given in Appendix.
Practically, the widths of the MDM waveguides are small
and the odd supermode (TM
1
) is well below the cut-off.
Therefore, the scattering parameters of the odd supermode are
not needed as long as their effects on the scattering coefcients
of the even supermode are considered. It is therefore possible
to replace the 4 4 scattering matrix of the proposed model
by the 2 2 scattering matrix whose elements satisfy the
following relations:
|S
e
11
|
2
+|S
e
21
|
2
1 (14a)
S
e
21
S
e
11
1 exp (j ) (14b)
where

e
W
2
tan(/2). (15)
These relations imply the negligibility of metallic loss
between AB and AC planes [Eq. (14a)] and the possibility
of using a T-circuit model with a capacitive shunt admittance
for the bend [Eq. (14b)]. Therefore, the 4-port coupled trans-
mission line model in Fig. 4 can be replaced by the much
simpler 2-port symmetric T-circuit model shown in Fig. 5.
The input and output ports of this new model can then be
connected to the standard single-conductor transmission line
model of the MDM waveguides [1]. The side arm of the
T-circuit model should impose the phase shift and is thus
made of a transmission line segment whose length, propaga-
tion constant, and characteristic impedance are L/8, Z
e
= Y
1
e
and
e
, respectively. The shunt capacitance of the 2-port model
780 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. 49, NO. 9, SEPTEMBER 2013
Fig. 5. The schematic of the equivalent transmission line model of the MDM
-bend when the odd mode is below the cut off.
Fig. 6. Scattering matrix parameters of a 90-bend whose slit width is
W = 50 nm: (a) The amplitude of S
e
11
and S
e
21
, and (b) the phase of S
e
11
and S
e
21
, each one extracted by the proposed model (solid lines), and the
COMSOL (dashed lines).
is obtained in Appendix, and can be written as:
C
Bend

2
0

e
W
_
||
2
1 (16)
where W is the width of waveguides and
=

1

2
(17a)

1
= (Y
e
Y
o
)
2
+
_
(Y
e
+Y
o
)
2
+4Y
e
Y
o
_
cos(
s
L)
+j 2(Y
e
+Y
o
)(Y
e
Y
o
+Y

s
2
)Z

s
sin(
s
L) (17b)

2
= 2Y
e
_
2Y
o
(1 +cos(
s
L)) + j Z

s
(Y
2
o
+Y

s
2
) sin(
s
L)
_
(17c)
It is worth noting that the 2 2 scattering matrix of the
proposed T-model becomes the identity matrix when the bend-
angle tends to zero.
To demonstrate the accuracy of the proposed model, differ-
ent numerical examples are given and the obtained results are
compared against accurate numerical solutions provided by the
COMSOL package. The simulations are carried out by using
two-dimensional triangular meshes of size 2 nm. It is assumed
that the MDM structure is made of air slits carved within silver.
The complex permittivity of the silver is approximated by the
well-known Drude model [23] and the width of the air slits is
W = 50 nm.
As the rst numerical example, a typical right-angled
90-bend is considered. The amplitude and the phase of
scattering parameters; S
e
11
and S
e
21
, are plotted in Figs. 6(a)
and (b), respectively. The solid and dashed lines in these
gures represent the results obtained by using the proposed
T-circuit model, and the COMSOL, respectively. A very good
agreement is observed between the two.
Fig. 7. (a) A typical MDM S-bend with hypothetical planes AB, and CD.
(b) The schematic of its equivalent transmission line model when the odd
mode is below the cut off.
Fig. 8. Scattering matrix parameters of an S-bend whose slit width is
W = 50 nm: (a) The amplitude of S
e
11
and S
e
21
, and (b) the phase of S
e
11
and S
e
21
, each one extracted by the proposed model (solid lines), and the
COMSOL (dashed lines).
As another example, a typical MDM S-bend made of two
cascade 90-bends is considered. This structure is shown
in Fig. 7(a), and could be encountered in power dividers
and couplers [15], [24]. Using the proposed T-circuit model,
the MDM S-bend can be modeled by the transmission line
circuit shown in Fig. 7(b). The amplitude and the phase
of scattering parameters; S
e
11
and S
e
21
, of this structure are
similarly calculated for d = 100 nm. The obtained results are
plotted in Figs. 8(a) and (b), respectively. The solid and dashed
lines in these gures represent the results obtained by using
the proposed T-circuit model, and the COMSOL, respectively.
Once again, a very good agreement is observed between the
two.
Finally, the reection and transmission at 650 THz are
plotted versus the bend angle. The results from the proposed
T-circuit model (solid lines) and the COMSOL package (cir-
cles) are plotted in Fig. 9. There is an excellent agreement
between the numerical results and the proposed model for
-bends. As expected, the transmission decreases when the
bend angle is increased.
IV. COUPLED TRANSMISSION LINE MODEL FOR
MDM JUNCTIONS
In this section, the proposed coupled transmission line
model is employed for analysis of asymmetric MDM
BAHADORI et al.: COUPLED TRANSMISSION LINE MODEL FOR PLANAR MDM PLASMONIC STRUCTURES 781
Fig. 9. The dependence of the amplitude of the scattering matrix parameters
on the bend angle at 650 THz. The solid lines and the circles indicate the
results of the proposed T-circuit model, and the COMSOL, respectively.
Fig. 10. A typical MDM junction with an asymmetric arrangement that can
support both even and odd modes.
junctions [25]. Due to the presence of asymmetry in the
structure, the rst order non-principal mode of the MDM
waveguide is no longer negligible even if the working
frequency lies below the cut-off. Fig. 10 shows a typical
asymmetric MDM junction. The asymmetric junction is
made of two MDM waveguides whose widths at the left-
and right-sides of the junction are denoted by W
1
, and W
2
,
respectively. The permittivities of the dielectric and metallic
regions are represented by
d
, and
m
, respectively.
The coupled transmission line models of the MDM
waveguides at the left- and right-sides of the junction are
already given in section II. Since the upper edge of the right-
side MDM waveguide (point C in Fig. 10) is connected to
the upper edge of the left-side MDM waveguide (point A
in Fig. 10) and thereby forms a metal-dielectric interface of
length W
2
W
1
at the junction, the coupled transmission line
models of the MDM waveguides at the left- and right-sides
of the junction are expected to be connected to each other via
the transmission line model representing the metal-dielectric
interface. The length of this transmission line is W
2
W
1
and its propagation constant and characteristic impedance
are
s
[from Eq.(1)], and Z
s
[from Eq.(3)], respectively.
Now, the complete model of the junction can be constructed
in accordance with the arrangement shown in Fig. 11.
To demonstrate the accuracy of the proposed model, a
typical asymmetric MDM junction is analyzed. It is assumed
that the MDM structure is made of air slits carved within
silver, and that W
1
= 100 nm, and W
2
= 300 nm. The
complex permittivity of the silver is once again approximated
Fig. 11. The proposed transmission line model for the asymmetric MDM
junction.
Fig. 12. The power transmission spectrum for an asymmetric multi-mode
junction with W
1
= 100 nm, and W
2
= 300 nm. The proposed model (solid
line) agrees with the FDTD result (circles) while the conventional transmission
line model (dashed line) fails.
by the well-known Drude model [23]. The left-to-right power
transmission spectrum of the structure is then calculated by the
proposed model, the conventional transmission line model [1],
and a two-dimensional nite-difference-time-domain (FDTD)
method with grid sizes set to 2 nm and perfectly matched
layers (PML) surrounding the entire geometry. The obtained
results are plotted in Fig. 12. This gure clearly demonstrates
that even though the conventional transmission line model fails
to provide accurate enough results unless the working fre-
quency lies far below the TM
1
cut-off frequency of the right-
side MDM waveguide ( f
c
= 433 THz), the proposed coupled
transmission line model yields accurate enough results even for
the working frequencies lying well above f
c
. It is worth noting
that the frequency at which the transmitted power becomes
zero coincides with the TM
1
cut-off frequency. To further
demonstrate the behavior of this multi-model junction, the
prole of the transverse magnetic eld is plotted in Fig. 13(a)
for f = 200 THz and in Fig. 13(b) for f = 500 THz.
In accordance with the frequency spectrum of Fig. 12, we
see that only the fundamental mode is excited in the right
waveguide when the frequency is below f
c
. However, when
the frequency exceeds f
c
, the odd mode plays the most part
in excitation of the right waveguide.
V. COUPLED TRANSMISSION LINE MODEL FOR
STUBBED MDM WAVEGUIDES
Finally, the proposed coupled transmission line model
is employed for analysis of stubbed MDM waveguides.
Although conventional transmission line models have been
782 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. 49, NO. 9, SEPTEMBER 2013
Fig. 13. The transverse magnetic eld prole for the asymmetric junction
in Fig. 12. (a) at f = 200 THz (below f
c
). (b) at f = 500 THz (above f
c
).
Fig. 14. A typical single stub MDM structure. Coupling of the depicted
surface waves is included in the proposed coupled transmission line model.
quite successful in analysis of MDM waveguides with single
or multiple stubs [3], [4], [11], [26], the proposed cou-
pled transmission line model is still needed because conven-
tional transmission line models could fail to provide accurate
enough results particularly when the working frequency is
increased beyond a certain level. Fig. 14 shows a typical
MDM waveguide with a single stub. The width of the MDM
waveguide and that of the stub section are denoted by W
1
, and
W
2
, respectively. The length of the stub section is represented
by h. The permittivities of the dielectric and metallic regions
are once again represented by
d
, and
m
, respectively. The
coupled transmission line models of the MDM waveguides
at the left- and right-sides of the stub section are already
known. Therefore, the stub section remains to be modeled.
Since the lower edges of the MDM waveguides at the left- and
right-sides of the stub section are directly connected to each
other, the lower terminal points of the coupled transmission
lines representing the MDM waveguides at the left- and
right-sides of the stub section (points B, and F in Fig. 14)
should be connected to each other via the transmission line
model representing the metal-dielectric interface. The length
of this transmission line is W
2
and its propagation constant
and characteristic impedance are
s
[from Eq.(1)], and Z
s
[from Eq.(3)], respectively. The upper terminal points of the
coupled transmission lines representing the MDM waveguides
at the left- and right-sides of the stub section (points A,
and E in Fig. 14) should, on the other hand, be connected
to the transmission line model of the MDM waveguide that
forms the stub section. Since the stub section is in fact a
terminated MDM waveguide of length h, and width W
2
, it
Fig. 15. The coupled transmission line model of the single stub MDM
structure. The complex per-unit-length mutual inductance and capacitance of
the coupled lines are denoted by L
m
, and C
m
, respectively.
Fig. 16. The power transmission spectrum calculated by the proposed model
(solid line), the conventional model [4] (dashed line), and the COMSOL (circle
dotted). Inset shows the transverse magnetic eld prole at 534 THz.
can be represented by the proposed coupled transmission line
model if its upper and lower ports are appropriately terminated.
The end points of the lower port of the coupled transmis-
sion line model which represents the stub section should be
connected to the appropriate terminal points of the coupled
transmission line representing the MDM waveguides at the
left- and right-sides of the stub section (points A, and E in
Fig. 14). The end points of the upper port of the coupled
transmission line model which represents the stub section
should be connected to each other via the transmission line
model representing the metal-dielectric interface because the
stub section is terminated by the metal-dielectric interface (CD
line in Fig. 14). The length of this transmission line is W
2
and
its propagation constant and characteristic impedance are
s
[from Eq.(1)], and Z
s
[from Eq.(3)], respectively. Therefore,
the complete model of the structure is in accordance with the
arrangement shown in Fig. 15.
Once again, the accuracy of the proposed model is
demonstrated by numerical simulation of a typical example.
In accordance with Fig. 14, a single stub structure with
W
1
= W
2
= 100 nm, and h = 300 nm is considered. The
power transmission of the structure is then calculated by using
the conventional transmission line model [4], the proposed
coupled transmission line model, and the COMSOL package
with triangular meshes of size 2 nm. The obtained results
are plotted in Fig. 16. Once again, the proposed model is
in good agreement with the COMSOL results. In particular,
BAHADORI et al.: COUPLED TRANSMISSION LINE MODEL FOR PLANAR MDM PLASMONIC STRUCTURES 783
the zero frequencies extracted by the proposed method are
more accurate than the zero frequencies extracted by the
conventional model.
VI. CONCLUSION
To improve the accuracy of the conventional transmission
line model for MDM waveguides [1], [13], the coupling
between surface plasmon waves supported by each metal-
dielectric interface in MDM-based structures was taken into
account and thereby a new coupled transmission line model
was obtained in this paper. The accuracy of the thus-obtained
model was then examined in miscellaneous MDM-based struc-
tures, viz. MDM bends, junctions, and stubs. The superiority
of the proposed method over the conventional transmission
line model is owed to the fact that the contribution of the rst
non-principal mode of the MDM waveguide is included.
APPENDIX
Using the voltage and current distribution along the coupled
transmission lines [see Eq. (10)], and applying the continuity
condition of voltage and current at point A in the coupled
transmission line model of the MDM -bend [see Fig. 4], it
can be easily shown that:
Q
2
X Q
1
Y = b (A1)
where
X =
_
S
e
21
S
o
21
_
, Y =
_
S
e
11
S
o
11
_
, b =
_
1
Y
e
_
(A2)
Q
1
=
_
1 1
Y
e
Y
o
_
, Q
2
=
_
1 1
Y
e
Y
o
_
(A3)
Two other equations are however needed to calculate the
unknown scattering parameters represented by X and Y vec-
tors. Using the voltage and current distribution of the trans-
mission line connecting B and C to each other, it can be shown
that:
MQ
4
X Q
3
Y = b (A4)
where
M =
_
cos(
s
L) j Z

s
sin(
s
L)
j Y

s
sin(
s
L) cos(
s
L)
_
(A5)
Q
3
=
_
1 1
Y
e
Y
o
_
, Q
4
=
_
1 1
Y
e
Y
o
_
(A6)
The unknown scattering parameters are then obtained as:
X =
_
Q
1
1
Q
2
Q
1
3
MQ
4
_
1
_
Q
1
1
Q
1
3
_
b (A7)
Y =
_
Q
1
2
Q
1
Q
1
4
M
1
Q
3
_
1
_
Q
1
4
M
1
Q
1
2
_
b(A8)
Now that the scattering parameters are at hand, the shunt
capacitance of the symmetric T-circuit model shown in Fig. 5
can be written as [7]:

j C
Bend
Y
e

= 2

S
e
11
S
e
21

(A9)
and with the approximations of Eq. (14), the result of Eq. (16)
immediately follows.
REFERENCES
[1] G. Veronis and S. Fan, Bends and splitters in metal-dielectric-metal
subwavelength plasmonic waveguides, Appl. Phys. Lett., vol. 87, no. 13,
pp. 131102131102, 2005.
[2] M. Staffaroni, J. Conway, S. Vedantam, J. Tang, and E. Yablonovitch,
Circuit analysis in metal-optics, Photon. Nanostruct. Fundam. Appl.,
vol. 10, no. 1, pp. 166176, 2012.
[3] Y. Matsuzaki, T. Okamoto, M. Haraguchi, M. Fukui, and M. Nakagaki,
Characteristics of gap plasmon waveguide with stub structures, Opt.
Exp., vol. 16, no. 21, pp. 1631416325, 2008.
[4] A. Pannipitiya, I. Rukhlenko, and M. Premaratne, Analytical modeling
of resonant cavities for plasmonic-slot-waveguide junctions, IEEE
Photon. J., vol. 3, no. 2, pp. 220233, Mar. 2011.
[5] A. Pannipitiya, I. D. Rukhlenko, M. Premaratne, H. T. Hattori, and
G. P. Agrawal, Improved transmission model for metal-dielectric-metal
plasmonic waveguides with stub structure, Opt. Exp., vol. 18, no. 6,
pp. 61916204, 2010.
[6] Y. Huang, C. Min, L. Yang, and G. Veronis, Nanoscale plasmonic
devices based on metal-dielectric-metal stub resonators, Int. J. Opt.,
vol. 2012, pp. 113, Jul. 2012.
[7] D. Pozar, Microwave Engineering. New York, NY, USA: Wiley, 2009.
[8] M. Swillam and A. Helmy, Feedback effects in plasmonic slot
waveguides examined using a closed form model, IEEE Photon.
Technol. Lett., vol. 24, no. 6, pp. 497499, Mar. 2012.
[9] Y. Wang, X. Zhang, J. Wang, J. Liu, Y. Wang, K. Yang, and Y. Song,
Manipulating surface plasmon polaritons in a 2-D t-shaped metal
insulatormetal plasmonic waveguide with a joint cavity, IEEE Photon.
Technol. Lett., vol. 22, no. 17, pp. 13091311, Sep. 2010.
[10] H. Hodaei, M. Rezaei, M. Miri, M. Bahadori, A. Eshaghian, and
K. Mehrany, Easy-to-design nano-coupler between metalinsulator
metal plasmonic and dielectric slab waveguides, Plasmonics, vol. 8,
pp. 11231128, Jun. 2013.
[11] J. Liu, G. Fang, H. Zhao, Y. Zhang, and S. Liu, Surface plasmon
reector based on serial stub structure, Opt. Exp., vol. 17, no. 22,
pp. 2013420139, Oct. 2009.
[12] P. Lee and Y. Lan, Plasmonic waveguide lters based on tunneling and
cavity effects, Plasmonics, vol. 5, no. 4, pp. 417422, 2010.
[13] H. Nejati and A. Beirami, Theoretical analysis of the characteristic
impedance in metalinsulatormetal plasmonic transmission lines, Opt.
Lett., vol. 37, no. 6, pp. 10501052, 2012.
[14] M. Rezaei, M. Miri, A. Khavasi, K. Mehrany, and B. Rashidian,
An efcient circuit model for the analysis and design of rectangular
plasmonic resonators, Plasmonics, vol. 7, no. 2, pp. 245252, 2012.
[15] M. Rezaei, S. Jalaly, M. Miri, A. Khavasi, A. Fard, K. Mehrany, and
B. Rashidian, A distributed circuit model for side-coupled nanoplas-
monic structures with metalinsulatormetal arrangement, IEEE J.
Sel. Topics Quantum Electron., vol. 18, no. 6, pp. 16921699,
Nov./Dec. 2012.
[16] M. Bahadori, A. Eshaghian, H. Hodaei, M. Rezaei, and K. Mehrany,
Analysis and design of optical demultiplexer based on arrayed plas-
monic slot cavities: Transmission line model, IEEE Photon. Technol.
Lett., vol. 25, no. 8, pp. 784786, Apr. 2013.
[17] F. Hu, H. Yi, and Z. Zhou, Wavelength demultiplexing structure
based on arrayed plasmonic slot cavities, Opt. Lett., vol. 36, no. 8,
pp. 15001502, 2011.
[18] S. Kocabas, G. Veronis, D. Miller, and S. Fan, Transmission line
and equivalent circuit models for plasmonic waveguide components,
IEEE J. Sel. Topics Quantum Electron., vol. 14, no. 6, pp. 14621472,
Nov./Dec. 2008.
[19] G. Veronis, S. E. Kocabas, D. A. Miller, and S. Fan, Modeling of plas-
monic waveguide components and networks, J. Comput. Theoretical
Nanosci., vol. 6, no. 8, pp. 18081826, 2009.
[20] S. Maier, Plasmonics: The promise of highly integrated optical
devices, IEEE J. Sel. Topics Quantum Electron., vol. 12, no. 6,
pp. 16711677, Nov./Dec. 2006.
[21] S. I. Bozhevolnyi, Plasmonic nano-guides and circuits, in Plasmonics
and Metamaterials. Washington, DC, USA: OSA, 2008.
[22] S. Maier, Plasmonics: Fundamentals and Applications. New York, NY,
USA: Springer-Verlag, 2007.
[23] J. Tao, X. Huang, X. Lin, Q. Zhang, and X. Jin, A narrow-band
subwavelength plasmonic waveguide lter with asymmetrical multiple-
teeth-shaped structure, Opt. Exp., vol. 17, no. 16, pp. 1398913994,
2009.
[24] N. Nozhat and N. Granpayeh, Analysis of the plasmonic power splitter
and MUX/DEMUX suitable for photonic integrated circuits, Opt.
Commun., vol. 284, no. 13, pp. 34493455, 2011.
784 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. 49, NO. 9, SEPTEMBER 2013
[25] P. Chen, R. Liang, Q. Huang, Z. Yu, and X. Xu, Plasmonic lters
and optical directional couplers based on wide metal-insulator-metal
structure, Opt. Exp., vol. 19, no. 8, pp. 76337639, 2011.
[26] L. Yang, C. Min, and G. Veronis, Guided subwavelength slow-light
mode supported by a plasmonic waveguide system, Opt. Lett., vol. 35,
no. 24, pp. 41844186, 2010.
Meisam Bahadori was born in Khuzestan, Iran, on
September 17, 1989. He received the B.Sc. degree
(magna cum laude) in electrical engineering in com-
munication systems and the M.Sc. degree (summa
cum laude) in electrical engineering microwaves and
optics from the Sharif University of Technology,
Tehran, Iran, in 2011 and 2013, respectively.
Since 2011, he has been a Research Assistant
with the Integrated Photonics Laboratory, Sharif
University of Technology. He has authored or co-
authored several journal papers. He was a Teaching
Assistant for several courses, such as electromagnetism and microwaves. His
current research interests include nanophotonics, plasmonics, transmission
line modeling, and numerical electromagnetics.
Ali Eshaghian was born in Tehran, Iran, on March
31, 1990. He received the B.Sc. degree in electrical
engineering from the Sharif University of Technol-
ogy, Tehran, Iran, in 2012.
Since 2011, he has been a Researcher with the
Integrated Photonics Laboratory, Sharif University
of Technology. His current research interests include
photonics, plasmonics, and nonlinear optics.
Mohsen Rezaei was born in Razan, Iran, on
April 25, 1986. He received the B.Sc. degree from
the University of Tehran, Tehran, Iran, in 2008,
and the M.Sc. degree from the Sharif University
of Technology, Tehran, in 2011, both in electrical
engineering.
Since 2010, he has been a Research Assistant with
the Integrated Photonics Laboratory, Sharif Univer-
sity of Technology. His current research interests
include photonics, plasmonics, and nonlinear optics.
Hossein Hodaei was born in Esfahan, Iran, on
August 13, 1989. He received the B.Sc. and M.Sc
degrees in electrical engineering from the Isfahan
University of Technology, Isfahan, Iran, and Sharif
University of Technology, in 2011 and 2013, respec-
tively.
Since 2011, he has been a Research Assistant with
the Integrated Photonics Laboratory, Sharif Univer-
sity of Technology. His current research interests
include photonics, plasmonics, and optical devices.
Khashayar Mehrany was born in Tehran, Iran, on
September 16, 1977. He received the B.Sc., M.Sc.,
and Ph.D. (magna cum laude) degrees from the
Sharif University of Technology, Tehran, Iran, in
1999, 2001, and 2005, respectively, all in electrical
engineering.
He is an Associate Professor with the Depart-
ment of Electrical Engineering, Sharif University of
Technology. His current research interests include
photonics, plasmonics, and integral imaging.