You are on page 1of 7

Related content

- Plasmon polaritons in 1D Cantor-like


Plasmon polaritons in photonic superlattices fractal photonic superlattices containing a
left-handed material
containing a left-handed material J. R. Mejía-Salazar, N. Porras-
Montenegro, E. Reyes-Gómez et al.

- Zero-langle n rangle non-Bragg gap


To cite this article: E. Reyes-Gómez et al 2009 EPL 88 24002 plasmon–polariton modes and omni-
reflectance in 1Dmetamaterial photonic
superlattices
C Agudelo-Arango, J R Mejía-Salazar, N
Porras-Montenegro et al.
View the article online for updates and enhancements. - Absorption effects on plasmon polaritons
in quasiperiodic photonic superlattices
containinga metamaterial
E Reyes-Gómez, N Raigoza, S B
Cavalcanti et al.

Recent citations
- Tunable optical response at the plasmon-
polariton frequency in dielectric-graphene-
metamaterial systems
D.M. Calvo-Velasco and N. Porras-
Montenegro

- Gap soliton transparency switch in one-


dimensional Kerr-metamaterial
superlattices
T.P. Lobo et al

- Defect mode in the bulk plasmon-polariton


gap for giant enhancement of second
harmonic generation
F. Reyes Gómez et al

This content was downloaded from IP address 132.255.20.2 on 11/05/2018 at 18:02


October 2009
EPL, 88 (2009) 24002 www.epljournal.org
doi: 10.1209/0295-5075/88/24002

Plasmon polaritons in photonic superlattices containing


a left-handed material
E. Reyes-Gómez1 , D. Mogilevtsev2,3 , S. B. Cavalcanti4(a) , C. A. A. de Carvalho5,6 and L. E. Oliveira2,6
1
Instituto de Fı́sica, Universidad de Antioquia - AA 1226, Medellı́n, Colombia
2
Instituto de Fı́sica, UNICAMP - CP 6165, Campinas-SP, 13083-970, Brazil
3
Institute of Physics, NASB - F. Skarina Ave. 68, Minsk, 220072, Belarus
4
Instituto de Fı́sica, UFAL, Cidade Universitária - 57072-970, Maceió-AL, Brazil
5
Instituto de Fı́sica, UFRJ, Rio de Janeiro-RJ - 21945-972, Brazil
6
Inmetro, Campus de Xerém, Duque de Caxias-RJ - 25250-020, Brazil

received 20 July 2009; accepted in final form 7 October 2009


published online 9 November 2009

PACS 41.20.Jb – Electromagnetic wave propagation; radiowave propagation


PACS 42.70.Gi – Light-sensitive materials
PACS 42.70.Qs – Photonic bandgap materials

Abstract – We analyze one-dimensional photonic superlattices which alternate layers of air and
a left-handed material. We assume Drude-type dispersive responses for the dielectric permittivity
and magnetic permeability of the left-handed material. Maxwell’s equations and the transfer-
matrix technique are used to derive the dispersion relation and transmission spectra for the
propagation of obliquely incident optical fields. The photonic dispersion indicates that the growth
direction component of the electric (or magnetic) field leads to the propagation of electric (or
magnetic) plasmon polaritons, for either TE or TM configurations. Furthermore, we show that if
the plasma frequency is chosen within the photonic n(ω) = 0 zeroth-order bandgap, the coupling
of light with plasmons weakens considerably. As light propagation is forbidden in that particular
frequency region, the plasmon-polariton mode reduces to a pure plasmon mode.

c EPLA, 2009
Copyright 

Over the years, artificial complex materials have been semiconductor. Such resonant surface-plasmon polaritons
increasingly used to shape and manipulate light [1–5]. The may have much shorter wavelengths than that of the radi-
micro-structuring of high quality optical materials yields ation, which enables them to propagate along nanoscale
remarkable flexibility in the fabrication of nano-structures, systems [1–4], opening up a wide range of possibilities for
and allows for the tailoring of electromagnetic dispersions the construction of new optical devices. It is well known,
and mode structures to suit almost any need. Metamate- for example, that one of the most important features
rials [6–13], also known as left-handed materials (LHMs), of surface-plasmon polaritons is to confine light to very
are a remarkable example of such nanostructuring. Light small dimensions, yielding the merging of photonics and
propagation through metamaterials is characterized by a electronics at the nanoscale. Furthermore, considering
phase velocity opposite to the Poyinting vector, which the dispersive character of the LHMs, together with the
corresponds to negative dispersive electric and magnetic enhanced optical magnetism exhibited by them, one might
responses. conjecture the existence of remarkable new phenomena
The advent of metamaterials has opened a new era for such as the excitation of plasmon polaritons of a magnetic
optical devices, and has also given considerable thrust nature, that is, magnetic density waves resulting from
to the recent area of plasmonics, the study of plasmon resonant interactions between the optical field and the
polaritons. In metal-dielectric interfaces, for example, current densities within the metamaterial. Ultimately,
surface-plasmon polaritons are coupled modes that result those developments could lead to increases in the resolu-
from resonant interactions between electromagnetic tion of microscopes, in the efficiency of LEDs, and in the
waves and mobile electrons at the surface of a metal or sensitivity of chemical and biological devices [5–8].
One-dimensional (1D) superlattices which alternate
(a) E-mail: sbessa@gmail.com layers of positive and negative materials have already

24002-p1
E. Reyes-Gómez et al.

where we have assumed that the superlattice was grown


along the z-axis, q is the wave vector component along the
x-direction, and ey is the Cartesian unitary vector along
the y-direction.
Maxwell’s equations lead to the following differential
equations for the electric and magnetic amplitudes:
    
d 1 d ω 2 q2
E(z) = −(z) − 2 E(z), (4)
dz µ(z) dz c n (z)
Fig. 1: (Color online) Pictorial view of the 1D multilayer
photonic superlattice with layers A and B in periodic arrange- and
ment, and the electric and magnetic fields for the TE-like     
d 1 d ω 2 q2
and TM-like electromagnetic waves schematically shown. Note H(z) = −µ(z) − 2 H(z), (5)
that, for normal incidence, the two polarizations are equivalent. dz (z) dz c n (z)
 
where n(z) = (z) µ(z) is the z-position–dependent
exhibited many interesting properties [14–23] that are refraction index.
absent in superlattices composed solely of positive mate- In the sequel, nA and nB are the refraction indices,
rials. In particular, the existence of a non-Bragg photonic while µA and µB are the magnetic permeabilities in
bandgap, also known as a zeroth-order gap, has been layers A and B, respectively. Note that, because of (1),
suggested [12], detected [13,14], and characterized [15,16]. the nB refraction index in layer B is a function of the
In order to investigate the possibility of excitation of frequency, and may be a real positive, real negative or a
electric/magnetic plasmon polaritons, in the present work pure imaginary number. Also, note that q, the incident
we study the oblique incidence of light on a model 1D wave vector component along the x-direction, may be
superlattice composed of layers A of air, and layers B of obtained as a function of the angle of incidence θ ≡ θA
a doubly negative material. Layers A (width a) and B by q = ωc nA sin θ.
(width b) are distributed periodically so that d = a + b is Equations (4) and (5) may be solved by means of the
the period of the superlattice nanostructure (cf. fig. 1). transfer-matrix technique (see, for example, the work by
In the B layers, the electric and magnetic responses are Cavalcanti et al. [16] and references therein). For nB a real
dispersive and may assume negative values. If one neglects number, and n2B − n2A sin2 θ  0, the procedure yields the
losses, they may be described by [9–13] TE polarization dispersion relation from the solution of
the transcendental equation
ωe2 2
ωm
B (ω) = 0 − ; µB (ω) = µ0 − , (1)
ω2 ω2 cos(kd) = cos(QA a) cos(QB b)
 
where B (ω) and µB (ω) are the dielectric permittivity 1 FA FB
and magnetic permeability in slab B, respectively, one − + sin(QA a) sin(QB b). (6)
2 FB FA
may choose [10,11] 0 = 1.21 and µ0 = 1.0, and the elec-
tric/magnetic plasmon modes are at ν = νe = 2πω√e0 and In the formula, k is the Bloch wave vector along the z-
ν = νm = 2πω√mµ0 , which correspond to the solutions of direction which is the axis of the photonic crystal; fields
in consecutive unit cells are related by the Bloch condition,
B (ω) = 0 and µB (ω) = 0, respectively.
i.e., by the phase factor e i k d . QA and QB are defined as
We note that dispersions such as those in (1) hold
in periodically LC loaded transmission lines [10]. These ω ω
QA = nA | cos θ| = nA | cos θA |, (7)
negative-index systems were shown to exhibit good c c
microwave properties, with low loss and broad band-
and
width [11]. 
We shall be interested in studying the properties of ω ω
QB = n2B − n2A sin2 θ = |nB || cos θB |, (8)
both the transverse-electric (TE: electric field parallel to c c
the interface plane, see fig. 1) and transverse-magnetic
where, FA = (QA /µA ), FB = (QB /µB ) and, in the last
(TM: magnetic field parallel to the interface) polarizations
equality, we have made use of Snell’s law nA sin θ =
of a monochromatic electromagnetic field of frequency ω
nB sin θB .
propagating through a 1D periodic system.
For nB a real number, and n2B − n2A sin2 θ < 0, the
In the case of a TE field, one may choose
procedure yields the TE polarization dispersion relation
E(r, t) = E(z) exp[i (qx − ωt)] ey , (2) from the solution of the transcendental equation

whereas, in the case of TM polarization, the magnetic field cos(kd) = cos(QA a) cosh(QB b)
may be considered as  
1 FA FB
H(r, t) = H(z) exp[i (qx − ωt)] ey , (3) − − sin(QA a) sinh(QB b). (9)
2 FB FA

24002-p2
Plasmon polaritons in photonic superlattices containing a left-handed material

5 5 8
(a) TE (b) TM (a)
4 4
νm θ=0
3 3 νe 6 θ = π/12
frequency (GHz)

2 2
4 νm
1 1
a = b = 12 mm a = b = 12 mm
0 0 2
-1.0 -0.5 0.0 0.5 1.0 -1.0 -0.5 0.0 0.5 1.0
5 5
(c) TE (d) TM 0
4 4 -1.0 -0.5 0.0 0.5 1.0

frequency (GHz)
νm
3 3 νe 8
(b)
2 2 θ=0
6 θ = π/6
1 1
a = 12 mm, b = 6 mm a = 12 mm, b = 6 mm
0 0 4 νm
-1.0 -0.5 0.0 0.5 1.0 -1.0 -0.5 0.0 0.5 1.0

kd/π 2
Fig. 2: (Color online) TE and TM dispersion relations ν = ν(k)
ω
in photonic periodic superlattices (ν = 2π ). Calculations were 0
performed for air as slab A (A = 1, µA = 1), a = b = 12 mm -1.0 -0.5 0.0 0.5 1.0
(also for a = 12 mm and b = 6 mm), and ωe /2π = ωm /2π =
8
(c)
3 GHz for the Drude model (cf. eq. (1)) in slab B. Dotted
lines indicate the pure electric/magnetic plasmon modes at 6
ν = νe = 2πω√e0 and ν = νm = 2πω√mµ0 . Solid and dashed lines
correspond to θ = π/12 and θ = 0 incidence angles, respectively. θ=0
4 θ = π/3
νm
where QA is still given by (7), but

2
ω
QB = n2A sin2 θ − n2B . (10)
c
0
Moreover, if n2B < 0, (9) is still valid, with QA still given -1.0 -0.5 0.0 0.5 1.0
by (7), but
ω
 kd/π
QB = n2BI − n2A sin2 θ, (11)
c
where nBI is the imaginary part of nB . Fig. 3: (Color online) TE dispersion relations ν = ν(k) in a =
ω
b = 12 mm photonic periodic superlattices (ν = 2π ) as functions
For the TM polarization, the transcendental equa-
of various incidence angles. Here, dotted lines indicate the
tions (6) and (9), as well as the definitions of QA and QB pure magnetic plasmon modes at ν = νm = 2πω√mµ0 . Dashed lines
for the different cases are also valid, provided one replaces
correspond to θ = 0 and solid lines correspond to (a) θ = π/12,
µA by A and µB by B , where A and B are the dielectric (b) θ = π/6 and (c) θ = π/3 incidence angles, respectively.
permittivities in layers A and B, respectively.
ω
Figure 2 shows the calculated dispersions (ν = 2π ) in
the cases of TE (figs. 2(a) and (c)) and TM (figs. 2(b) for oblique incidence (θ = 0) the electromagnetic field +
and (d)) polarizations, for different layer widths. Note plasmon interaction leads to a pair of coupled modes. We
that νe and νm fall outside the n(ω) = 0 zeroth-order emphasize that, for θ = 0, the electric- and magnetic-field
bandgap. For θ = 0 (dashed curves), the second shown components along the z growth direction are null and
photonic band, just above the n(ω) = 0 zeroth-order there is no longitudinal wave propagation. On the other
gap, is a pure photonic mode and, for θ = π/12, there hand, for θ = 0, in the case of the TM (TE) configuration,
is a coupling between the radiation field and a plasmon the electric (magnetic) field has a component on the
mode which leads to a pair of plasmon-polariton modes. z-direction, which results in the excitation of longitudinal
In other words, while for normal incidence the band just electric (magnetic) waves. For θ = π/12, fig. 2 illustrates
above the zeroth-order gap is a pure photonic mode, that the pair of plasmon-polariton bands are asymptotic

24002-p3
E. Reyes-Gómez et al.

30 2.15
1.0

gap profile (GHz)


0.8
ga p pr of i l e ( G H z )

(a) θ = 0
2.10 0.6
n=0 0.4
20 2.05
n=2 0.2
2.00 0.0
0 π/4 π/2
θ 1 2 3 4 5 6 7 8 9 10
a = b = 12 mm
10 1.0
0.8
νm n=1
0.6
(b) θ = π/12
n=0 0.4
0
0 π/4 π/2 0.2
θ 0.0

|T |
2
1 2 3 4 5 6 7 8 9 10
Fig. 4: (Color online) Gap profiles for TE waves as functions
1.0
of the angle of incidence in a photonic periodic superlattice
(with a = b = 12 mm). The n = 0 zeroth-order gap is labeled 0.8
as n = 0 and is shown in detail in the inset to illustrate its 0.6
behavior with the θ incidence angle. The plasmon-polariton (c) θ = π/6
0.4
gap is labeled as n = 1 and the Bragg gap at the boundary is
denoted as n = 2. The dashed line indicates the pure magnetic 0.2
plasmon mode at ν = νm = 2πω√mµ0 . 0.0
1 2 3 4 5 6 7 8 9 10
to the ν = νe and ν = νm pure electric/magnetic plasmon
values. It is clear from fig. 2(a), for example, that at 1.0
small values of k, the lowest θ = π/12 plasmon-polariton 0.8
mode behaves like an electromagnetic photonic wave. 0.6
(d) θ = π/3
As k increases, the dispersion bends over and reaches
0.4
the limiting value of the pure magnetic plasmon. On
0.2
the other hand, the second θ = π/12 plasmon-polariton
branch behaves like a magnetic plasmon wave, at small k, 0.0
and like an electromagnetic photonic mode as k increases. 1 2 3 4 5 6 7 8 9 10
The same comments are valid for the other panels in fig. 2
frequency (GHz)
(of course, in the cases of figs. 2(b) and (d), the electric
plasmon mode is the one involved). By comparing the
results for normal and oblique incidence, it is clear that, Fig. 5: Transmission plots of TE waves, as functions of the angle
of incidence, through a stack of 50 unit cells composed of air
for θ = 0, resonant plasmon-polariton waves are excited
as slab A (A = 1, µA = 1) with length a = 12 mm, and slab B,
by the coupling of the electric (or magnetic) plasmon with length b = 12 mm, as a LHM with electric/magnetic
modes with the incident eletromagnetic field. One may responses given by the Drude model (i.e., a stack of 50 unit
infer that the plasmon-polariton waves are excited by the cells corresponding to the same superlattice as in fig. 3).
magnetic field, in the TE case, or driven by the electric
field, in the TM case. This is consistent with the fact that
those longitudinal waves, in the TE case, are asymptotic and decreases with increasing incidence angle. Also, the
to the νm pure magnetic plasmon frequency whereas, in plasmon-polariton gap (labeled as n = 1) and the Bragg
the TM case, the plasmon-polariton bands are asymptotic gap at the boundary (denoted as n = 2) increase as
to the νe pure electric plasmon mode. the angle of incidence increases. Figure 5 illustrates the
The calculated dispersions, for TE modes, as a function transmission spectra for a stack composed of fifty unit
of the incidence angle are displayed in fig. 3. We notice cells, with a = b = 12 mm. Null transmission results should
that the essential features persist irrespective of the be compared with the corresponding gaps in figs. 3 and 4.
incidence angle, although both the plasmon-polariton gap It is clear, therefore, that calculated transmission spectra
and Bragg gap at the boundary widen considerably for corroborate the results obtained for the corresponding
higher incidence angles, as expected. Figure 4 depicts the infinite photonic periodic superlattice via the dispersion
gap profiles for TE waves as functions of the incidence relation (cf. figs. 3 and 4).
angle for a = b = 12 mm. One clearly sees that the n = 0 Furthermore, fig. 6 clearly indicates that, by choosing
zeroth-order gap (labeled as n = 0) is not omnidirectional the resonant plasmon frequency within the n(ω) = 0

24002-p4
Plasmon polaritons in photonic superlattices containing a left-handed material

5 5 In conclusion, we have studied light propagation


(a) TE (b) TM
4 4 through a 1D superlattice composed of alternate layers
of a positive constant material and a negative disper-
3 νm 3 sive material, and verified the appearance of coupled
plasmon-polariton modes of electric and magnetic nature.
2 2
frequency (GHz)

νe
The photonic dispersion was calculated from Maxwell’s
1 1 equations using the transfer-matrix approach, with the
LHM modeled by Drude-like dielectric and magnetic
0 0
-1.0 -0.5 0.0 0.5 1.0 -1.0 -0.5 0.0 0.5 1.0 responses. In the case of oblique θ = 0 incidence, the
5 5
photonic dispersion indicates that the magnetic or elec-
(c) TE (d) TM
tric field in the frequency region around the n(ω) = 0
4 4 zeroth-order bandgap leads to coupled magnetic or
electric plasmon-polariton modes for the TE and TM
3 3
νe configurations, respectively. Moreover, present results
2 νm 2 show that the coupling of light with plasmons is weakened
by choosing the plasma frequency inside the zeroth-order
1 1
gap. As light propagation is forbidden in that particular
0 0 gap-frequency region, the coupled plasmon-polariton
-1.0 -0.5 0.0 0.5 1.0 -1.0 -0.5 0.0 0.5 1.0 mode becomes essentially a pure plasmon mode. This
feature permits one to select which type of plasmon
kd/π
polariton (i.e., electric or magnetic) one is willing to
excite, by choosing the magnetic or electric plasmon
Fig. 6: (Color online) TE and TM dispersions in a photonic
periodic superlattice with a = b = 12 mm. Numerical calcula-
frequency within the zeroth-order gap. The possibility of
tions were carried out for air as slab A (A = 1 and µA = 1). For magnetic plasmon-polariton excitations should provide a
slab B, B and µB were chosen by taking plasma frequencies new look towards the implementation of novel techniques
in the Drude model as ωe /2π = 1.5 GHz and ωm /2π = 2.5 GHz, and devices based on the interplay of photonics and
respectively, in (a) and (b), whereas in (c) and (d) we have magnetism at the nanoscale, similarly to the interplay of
taken ωe /2π = 2.5 GHz and ωm /2π = 1.5 GHz. Note that, in photonics and electronics [28,29]. There are still practical
panel (b), the electric plasmon frequency νe = 2πω√e0 is in the challenges to be dealt with in such systems, such as the
n(ω) = 0 zeroth-order bandgap region, leading to a basically challenge of reducing loss which is the common problem
flat θ = π/12 plasmon-polariton band with a frequency essen- related to metamaterials. Here, we have studied an ideal
tially equal to νe for the TM dispersion. A similar situation is system neglecting losses which, one could argue, may taint
observed for the TE dispersion in (c), as νm = 2πω√mµ0 falls in the coupling in such a way as to hinder plasmon-polariton
the zeroth-order gap region. Solid and dashed lines correspond excitations. However, a study on an anisotropic bilayer
to θ = π/12 and θ = 0 incidence angles, respectively. composed of well balanced positive and negative materials
reveals that, compared to a single isotropic meta layer,
the influence of loss on the behavior of the transfer
zeroth-order bandgap, the coupling of light to plasmon function, for various transverse wave vector kx , is greatly
modes, for θ = 0, essentially disappears leading to a reduced [24]. On the other hand, considering that classical
basically pure (electric or magnetic) plasmon mode. electrodynamics predicts a substantially lower radiation
This is a consequence of the fact that the energy of the loss of a magnetic dipole than the radiation loss of an
incident electromagnetic wave lies in a forbidden energy electric dipole of a similar size [30], one might conjecture
gap region and, therefore, the coupling of the incident that the combination of anisotropic stacks with plasmons
light with plasmons is expected to be quite weak. of a magnetic origin should serve as a quite good model
Here we note that the present study has dealt with a towards practical implementations such as magnetic
1D superlattice containing an isotropic LHM layer. Never- plasmonic chips for high data rate processing or sensing
theless, it is important to point out that the realization applications. Future experimental and theoretical studies
of isotropic LHM systems is still a great experimental on layered systems including loss effects and anisotropy
challenge, and that available LHM structures are mostly are certainly forthcoming.
anisotropic [24–27]. Of course, a proper consideration of
the effects of an anisotropic electric/magnetic response for ∗∗∗
the LHM layer would bring additional complexities in the
present theoretical approach. We do not expect, however, We thank A. Bruno-Alfonso for stimulating discus-
that the essential physics of the present study —on the sions, and are grateful to the Brazilian Agencies CNPq,
plasmon polaritons in photonic superlattice containing FAPESP, FAPERJ, and FUJB, the Colombian Agency
left-handed materials with isotropic responses— would COLCIENCIAS, and CODI - University of Antioquia for
change in any substantive way. partial financial support.

24002-p5
E. Reyes-Gómez et al.

REFERENCES [16] Cavalcanti S. B., Dios-Leyva M., Reyes-Gómez E.


and Oliveira L. E., Phys. Rev. B, 74 (2006) 153102;
[1] Shelby R. A., Smith D. R. and Schultz S., Science, Phys. Rev. E, 75 (2007) 026607; Bruno-Alfonso A.,
292 (2001) 77. Reyes-Gómez E., Cavalcanti S. B. and Oliveira
[2] Barnes W. L., Dereux A. and Ebbesen T. W., Nature, L. E., Phys. Rev. A, 78 (2008) 035801.
424 (2003) 824. [17] Shadrivov I. V., Sukhorukov A. A. and Kivshar
[3] Maier S. A. and Atwater H. A., J. Appl. Phys., 98 Y. S., Appl. Phys. Lett., 82 (2003) 3820.
(2005) 011101. [18] Ziolkowski R. W., Phys. Rev. E, 70 (2004) 046608.
[4] Ramakrishna A., Rep. Prog. Phys., 68 (2005) 449. [19] Daninthe H., Foteinopoulou S. and Soukoulis
[5] Ozbay E., Science, 311 (2006) 189. C. M., Photonics Nanostruct. Fundam. Appl., 4 (2006)
[6] Dolling G., Enkrich C., Wegener M., Soukoulis 123.
C. M. and Linden S., Opt. Lett., 31 (2006) 1800. [20] Yuan Y., Ran L., Huangfu J., Chen H., Shen L. and
[7] Li T., Li J.-Q., Wang F.-M., Wang Q.-J., Liu H., Kong J. A., Opt. Express, 14 (2006) 2220.
Zhuy S.-N. and Zhu Y.-Y., Appl. Phys. Lett., 90 (2007) [21] Vasconcelos M. S., Mauris P. W., de Medeiros
251112. F. F. and Albuquerque E. L., Phys. Rev. B, 76 (2007)
[8] Dragoman M. and Dragoman D., Prog. Quantum 165117.
Electron., 32 (2008) 1. [22] Weng Y., Wang Z.-G. and Chen H., Phys. Rev. E, 75
[9] Pacheco J. jr., Grzegorczyk T. M., Wu B.-I., (2007) 046601.
Zhang Y. and Kong J. A., Phys. Rev. Lett., 89 (2002) [23] Zhang L., Zhang Y., He L., Wang Z., Li H. and Chen
257402. H., J. Phys. D: Appl. Phys., 40 (2007) 2579.
[10] Eleftheriades G. V., Iyer A. K. and Kremer P. C., [24] Smith R. and Schurig D., Phys. Rev. Lett., 90 (2003)
IEEE Trans. Microwave Theory Tech., 50 (2002) 2702. 077405.
[11] Grbic A. and Eleftheriades G. V., J. Appl. Phys., 92 [25] Wang S. and Gao L., Eur. Phys. J. B, 48 (2005) 29.
(2002) 5930; Liu L., Caloz C., Chang C. C. and Itoh [26] Sun S., Huang X. and Zhou L., Phys. Rev. E, 75 (2007)
T., J. Appl. Phys., 92 (2002) 5560. 066602.
[12] Li J., Zhou L., Chan C. T. and Sheng P., Phys. Rev. [27] Han P., Chan C. T. and Zhang Z. Q., Phys. Rev. B,
Lett., 90 (2003) 083901. 77 (2008) 115332.
[13] Jiang H., Chen H., Li H., Zhang Y. and Zhu S., Appl. [28] Liu H., Li T., Wang Q. J., Zhu Z. H., Wang S. M.,
Phys. Lett., 83 (2003) 5386. Li J. Q., Zhu S. N., Zhu Y. Y. and Zhang X., Phys.
[14] Smith D. R., Padilla W. J., Vier D. C., Nemat- Rev. B, 79 (2009) 024304.
Nasser S. C. and Schultz S., Phys. Rev. Lett., 84 [29] Liu H., Liu Y. M., Li T., Wang S. M., Zhu S. N. and
(2000) 4184. Zhang X., Phys. Status Solidi B, 246 (2009) 1397.
[15] Liscidini M. and Andreani L. C., Phys. Rev. E, 73 [30] Jackson J. D., Classical Electrodynamics (Wiley,
(2006) 016613. New York) 1999.

24002-p6

You might also like