Hybrid surface waves in semi-infinite metal-dielectric lattices
(N
o
U
Oh'
<
m
(N
CZ5
O
H— > .
Or
q
o ■
• i-H ,
>*
Oh'
>
00
o
in
o
(N
Juan J. Miret, 1 Carlos J. Zapata-Rodrfguez, 2 *
Zoran Jaksic, 3 Slobodan Vukovic, 3,4 and Milivoj R. Belie 4
1 Department of Optics, Pharmacology and Anatomy, University of Alicante, Spain
2 Department of Optics, University of Valencia, Dr. Molmer 50, 46100 Burjassot, Spain
! Center of Microelectronic Technologies and Single Crystals, Institute of Chemistry, Technology and Metallurgy,
University of Belgrade, Njegoseva 12, 11000 Belgrade, Serbia
4 Texas A & M University at Qatar, P.O. Box 23874, Doha, Qatar
* Corresponding author: carlos.zapataQuv.es
Compiled March 3, 2013
We investigate surface waves at the boundary between a semi-infinite layered metal-dielectric nanostructure cut
normally to the layers and a semi-infinite dielectric. Spatial dispersion properties of such a nanostructure can
be dramatically affected by coupling of surface plasmons polaritons at different metal-dielectric interfaces. As a
consequence, the effective medium approach is not applicable in general. It is demonstrated that Dyakonov-likc
surface waves with hybrid polarization can propagate in an angular range substantially enlarged compared
to conventional birefringent materials. Our numerical simulations for an Ag-GaAs stack in contact with glass
show a low to moderate influence of losses © 2013 Optical Society of America
OCIS codes: 160.1245, 240.6690, 260.3910.
Currently, metallodiclcctric superlattices are receiving
increased attention because of their extraordinary opti-
cal properties such as near-field focusing, subwavclcngth
imaging, and negative refraction [1-3]. The inclusion of
metallic elements is responsible for the excitation of plas-
monic resonances on the metal-dielectric (MD) inter-
faces, giving rise to an anisotropic metamaterial. Surface
polaritons may exist also along its optical axis, when the
plasmonic crystal is cut normally to the orientation of
the layers [4].
Hybrid-polarized waves confined to the boundary of
a semi-infinite layered periodic nanostructure, and im-
portantly with oblique propagation with respect to the
lattice optical axis, have first been reported in the con-
text of hyperlensing [5]. The effective-medium approxi-
mation [6] (EMA) was applied to benefit from a simple
approach that leads to Dyakonov surface waves [7]. In
this letter we demonstrate that in general, the use of
the EMA is not justified. Specifically, we focus on sur-
face waves at the boundary between a MD nanostruc-
ture and a semi-infinite dielectric. By solving Maxwell's
equations, we demonstrate the existence of Dyakonov-
like surface waves with hybrid polarization and establish
that they can propagate in a much wider angular range
than the conventional birefringent materials. In our cal-
culations, realistic metal widths and dissipation effects
are utilized.
We consider a bilayered superlattice made of two ma-
terials alternatively stacked along the z axis (Fig. 1). The
unit cell of the ID photonic lattice consists of a trans-
parent material of dielectric constant ed and slab width
Wd, followed by a metallic layer with the corresponding
parameters e rn and w m . In our numerical simulations
we set td — 12.5 and e m = —103.3 + £8.1, which corre-
spond to GaAs and Ag, respectively, at the wavelength
Aq = 1.55 jjjo.. This metamaterial fills the semi-space
plasmonic
superlattice
optical
axis
Fig. 1. Schematic setup under study, consisting of a semi-
infinite Ag-GaAs superlattice (x > 0) and an isotropic
cover (x < 0), either N-BAK1 or P-SF68 [SCHOTT].
x > 0. Adjacent to the periodic medium, filling the space
x < 0, is an isotropic material of dielectric constant e.
At the boundary between the periodic and the isotropic
medium, in the plane x = 0, we expect to find bounded
waves with evanescent amplitudes as \x\ — > oo.
We seek solutions of Maxwell's equations in the form of
surface waves that can propagate at the MD-superlattice
boundary. First, we configure the optical anisotropy of
our periodic structure by employing average estimates.
In particular, the form of birefringence in this type
of plasmonic devices may be modeled by the EMA
[8]. The validity of this approximation is related to
the fact that layer widths are much shorter than the
wavelength, Wd,w m <C A . In this case, the superlat-
tice behaves as a uniaxial crystal whose optical axis
is normal to the layers (here, z-axis). The model esti-
mates the relative permittivities along the optical axis,
£|| = e m ed/ [(1 - /) e m + fed], and transversally, e± =
(1 - /) e d + fe
there f is the metal filling factor,
1
/ = w m j (wd + w m ). If we neglect losses by setting
Im(e m ) = 0, the effective birefringence of the Ag-GaAs
superlattice is An = .yejy — y/e±. Even small filling fac-
tors of the metallic composite lead to enormous birefrin-
gences. If / <C 1 and \e m \ ^> €4, we obtain Ae w —e m f,
which holds up to / max = 0.108, where e± = 0.
Filling factors higher than / max provide negative
values of ej_ and, as a consequence, effective dispersion
becomes hyperbolic [9]. Thus An > 0, since e m < 0. In
spite of / m ax "C 1, the maximum birefringence achiev-
able reaches [An] max = 3.77. For comparison, we quote
the birefringence An = 0.0084 for crystalline quartz and
An = 0.22 for liquid crystal BDH-E7 [10]. Thus, our ar-
tificial uniaxial crystal has birefringence greater by more
than an order of magnitude.
Since we treat the ID lattice as a uniaxial crystal,
we may establish the diffraction equation giving the
wave vector of the surface wave at i = 0. For that
purpose we follow Dyakonov [7], by considering hy-
brid surface modes. In the isotropic medium, we con-
sider homogeneous TE and TM waves in the plane
x = 0, whose wave vector has real components k v
and k z . However, these fields are evanescent in the
isotropic medium, proportional to cxp (— where
K = yjky + k z — fcge and ko = 2tt/\q. On the other
side of the boundary, the ordinary and extraordinary
waves in the effective uniaxial medium also exponen-
tially decay, with rates given by k = \jk 2 + k z — k$e±
and n e = \Jky + fcfej|/ei — fco e lh respectively. The
Dyakonov equation,
(k + « e ) (k + k ) (ek q + e ± n e ) = (e|| - e) (e - ej_) fc 2 K„,
(1)
providing a map of the allowed (k y , k z ) values, has been
derived by applying the standard electromagnetic field
boundary conditions at x = 0. Assuming that en, ej_ and
that all decay rates are positive, the following additional
restriction can be deduced
ex < e < e M , (2)
for the existence of surface waves. As a consequence,
positive birefringence is mandatory to ensure the exis-
tence of surface waves. Therefore, such a layered super-
lattice featuring surface waves cannot be formed by all-
dielectric materials.
To illustrate the difference between using conventional
birefringent materials and plasmonic crystals, we solve
Eq. (1) for a liquid crystal E7 with ei = 2.98 and
e± = 2.31 at a wavelength of Xq = 1.55 /im, and an
N-BAK1 cover of dielectric constant e = 2.42. In this
case, Dyakonov surface waves propagate in a narrow an-
gular region A9 = # max ~ $min; where 9 stands for the
angle between the vector (k y ,k z ) and the optical axis.
Specifically, we estimate A9 — 0.92° around a mean an-
gle 9 = 26.6°. Apparently, the angular range A9 is small;
however it is extremely high if compared with that ob-
tained for other optical crystals, like quartz, exhibiting
2.0 F
Fig. 2. Isofrequency curves evaluated from Eq. (3) for
Ag-GaAs superlattice with filling factor / = 0.10, for dif-
ferent w m ; k t = \Jk~l + ky. Solid lines and dashed lines
correspond to TM and TE polarization. The dashed-
dottcd line represents the isofrequency curve of isotropic
N-BAK1 cover. Inset: Birefringence as a function of w m .
standard birefringence. To increase the angular spread
A9 even more, we consider Ag-GaAs crystal that has
values of £| = 14.08 and e± = 0.92, with a filling fac-
tor / = 0.10; the birefringence now yields An — 2.79.
In this case, Eq. (1) provides solutions in the region of
angles from 9 m - m = 39.0° up to # max = 71.3°. As a con-
sequence, the angular range A9 = 32.3° has increased
by more than an order of magnitude.
The EMA is limited to metallic slabs with w m <C Aq.
However, this condition must be taken into account with
care, since the skin depth of noble metals is extremely
short, A s ~ c/ojp, where uj p is the plasma frequency of the
metal. In the case of silver, we estimate X s = 24 nm. If
the thickness of metallic layer is comparable to the skin
depth, the EMA will substantially deviate from exact
calculations. Note that experimental multilayers rarely
incorporate metallic slabs with a thickness below 10 nm.
To address the problem more thoroughly, we consider
the full-wave solution of Maxwell's equations, leading to
the spectral map of Bloch waves in bulk ID-periodic me-
dia. In this case k z represents the pscudomomcntum of
a Bloch wave. In Fig. 2 we plot the spatial dispersion
relation [8]
cos(fc 2 A) = cos tfm cos ip d - r] ^ e sin ip m sin (p d , (3)
corresponding to TE-polarizcd (ordinary) and TM-
polarized (extraordinary) waves propagating within
the periodic Ag-GaAs structure displayed in Fig. 1.
In Eq. (3) the period is A ~ w rn + Wd,
while r\ = {k\ z + k 2 mz ) /2k dz k mz and n e =
( e m k dz + e d k liz) /^ e m.^dk dz k mz , are the coefficients ap-
plied for ordinary and extraordinary waves, respectively.
Finally, ip q = k qz w q , where k 2 + ky + k^ z = e q k^ repre-
sents the dispersion equation for bulk waves within GaAs
(q = d) and silver (q = m).
In the case of negligible losses, as illustrated in Fig. 2,
we observe that EMA is sufficiently accurate for w m =
3 nm. However, deviations among the contours are ev-
2
[m(/0
Fig. 3. (a) Isofrequency curve corresponding to hybrid
surface waves existing at the boundary between a semi-
infinite P-SF68 cover and a plasmonic Ag-GaAs super-
lattice of / = 0.10 and w m = 12 nm. (b) Three contour
plots of the magnetic field \H X \ computed by COMSOL.
ident for higher widths. Apparently, Eq. (3) is in good
agreement with the EMA in the vicinity of k z = for TM
waves only. In contrast, propagation along the z-axis,
where k x = k y = 0, results in larger discrepancies, as
the Bloch wave number k z increases for higher w m . This
effect is observed simultaneously for TM and TE waves.
Consequently, the size of birefringence displayed by TM
waves is reduced (see inset in Fig. 2). For TE waves,
the isotropy of the isofrequency curve is practically con-
served, e.g. n±_ = 1.70 and nn = 1.67 for w m = 12 nm.
At the same time, ej_ increases with w m .
Moderate changes in birefringence can substantially
impact the existence of Dyakonov-like surface waves. En-
largement of ej_ by increasing w m (with the same /)
leads to a significant modification of the isofrequency
curve derived from Eq. (1). According to Eq. (2), this
may entirely prevent the existence of surface waves.
Up to now we have avoided another important as-
pect of plasmonic devices, that of dissipation in metallic
elements. In order to tackle this problem, we evaluate
numerically the value of k z for a given k y . Since the
imaginary part of e m is not neglected anymore, k z be-
comes complex. This means that the surface wave can-
not propagate indefinitely; there exists an energy atten-
uation length I = (2Im[fc z ]) _1 . Furthermore, we assume
that the real parts of the parameters k, k and n e are all
positive, which correlates with the field decay at |a;| — > oo
and thus with the confinement near x = 0.
Fig. 3(a) shows an isofrequency curve corresponding
to Dyakonov-like solutions, for the Ag-GaAs lattice with
/ = 0.10 and w m = 12 nm. The numerical simula-
tions were performed using COMSOL Multiphysics soft-
ware based on the finite-clement method. In our simu-
lations we did not observe surface waves for a N-BAK1
cover with n = 1.56, since now it is not consistent with
Eq. (2); in our estimations n± = 1.70 and nu = 3.75.
In particular, Fig. 3(a) depicts the isofrequency curves
when n = 1.95, corresponding to P-SF68 [SCHOTT].
The boundaries of such isofrequency curve are estab-
lished according to the propensity of electromagnetic
field to be confined in the neighborhood of 2; = 0.
Fig. 3(a) also shows Im(fc z )/Re(fc z ) in the range of exis-
tence of the surface waves. For paraxial surface waves, in
which k y reaches the minimum (case A) , we observe that
lm(k z ) <C Rc(fc 2 ). This is caused by a large shift of the
field maximum toward the isotropic medium, as shown
in Fig. 3(b), which is consistent with Rc(k) <C Im(re). On
the other hand, for nonpar axial waves, having the largest
values of k y (case C), the fields show slow energy de-
cay inside the plasmonic superlattice. As a consequence,
losses in the metal become manifested by a significant
rise in the values of Im(fc 2 ).
We conclude that oblique surface waves may propa-
gate at the boundary between a plasmonic bilayer su-
perlattice and an isotropic transparent material. These
modes are not TM-polarizcd, since all three spatial com-
ponents of the electric, as well as of the magnetic field
are involved, i.e. the modes are hybrid. Realistic slab
widths lead to solutions that deviate significantly from
the results of EMA and Dyakonov analysis. Our numer-
ical simulations prescribe the use of cover materials of
a higher refractive index, for the existence of surfaces
waves. A wide angular range of surface waves is attain-
able with large to moderate energy attenuation lengths.
Finally, let us point out that the properties of the result-
ing bound states change rapidly with the refractive index
of the surrounding medium (cover), which suggests po-
tential applications for chemical and biological sensors.
This research was funded by the Spanish Ministry
of Economy and Competitiveness under the project
TEC2009-11635, by the Qatar National Research Fund
under the project NPRP 09-462-1-074, and by the
Serbian Ministry of Education and Science under the
projects III 45016 and TR 32008.
References
1. X. Fan, G. P. Wang, J. C. W. Lee, and C. T. Chan,
Phys. Rev. Lett. 97, 073901 (2006).
2. D. de Ceglia, M. A. Vincenti, M. G. Cappeddu, M. Cen-
tini, N. Akozbek, A. D'Orazio, J. W. Haus, M. J. Bloe-
mer, and M. Scalora, Phys. Rev. A 77, 033848 (2008).
3. A. Pastuszczak and R. Kotynski, J. Appl. Phys. 109,
084302 (2011).
4. S. M. Vukovic, I. V. Shadrivov, and Y. S. Kivshar, Appl.
Phys. Lett 95, 041902 (2009).
5. Z. Jacob and E. E. Narimanov, Appl. Phys. Lett. 93,
221109 (2008).
6. S. M. Rytov, Sov. Phys. JETP 2, 466 (1956).
7. M. I. D'yakonov, Sov. Phys. JETP 67, 714 (1988).
8. P. Yeh, Optical Waves in Layered Media (Wiley, New
York, 1988).
9. B. Wood, J. B. Pendry, and D. P. Tsai, Phys. Rev. B
74, 115116 (2006).
10. S.-T. Wu, U. Efron, and L. D. Hess, Appl. Opt. 23, 3911
(1984).
3
Informational Fourth Page
In this section, we provide full versions of citations to
assist reviewers and editors.
References
1. X. Fan, G. P. Wang, J. C. W. Lee, and C. T. Chan, "All-
angle broadband negative refraction of metal waveguide
arrays in the visible range: Theoretical analysis and nu-
merical demonstration," Phys. Rev. Lett. 97, 073901
(2006).
2. D. de Ceglia, M. A. Vincenti, M. G. Cappeddu, M. Cen-
tini, N. Akozbek, A. D'Orazio, J. W. Haus, M. J. Bloe-
mer, and M. Scalora, "Tailoring metallodielectric struc-
tures for superresolution and superguiding applications
in the visible and near-IR ranges," Phys. Rev. A 77,
033848 (2008).
3. A. Pastuszczak and R. Kotyhski, "Optimized low-loss
multilayers for imaging with sub-wavelength resolution
in the visible wavelength range," J. Appl. Phys. 109,
084302 (2011).
4. S. M. Vukovic, I. V. Shadrivov, and Y. S. Kivshar, "Sur-
face Bloch waves in metamaterial and metal-dielectric
superlattices," Appl. Phys. Lett 95, 041902 (2009).
5. Z. Jacob and E. E. Narimanov, "Optical hyperspace
for plasmons: Dyakonov states in metamaterials," Appl.
Phys. Lett. 93, 221109 (2008).
6. S. M. Rytov, "Electromagnetic properties of layered me-
dia," Sov. Phys. JETP 2, 466 (1956).
7. M. I. D'yakonov, "New type of electromagnetic wave
propagating at an interface," Sov. Phys. JETP 67, 714-
716 (1988).
8. P. Yeh, Optical Waves in Layered Media (Wiley, New
York, 1988).
9. B. Wood, J. B. Pendry, and D. P. Tsai, "Directed sub-
wavelength imaging using a layered metal-dielectric sys-
tem," Phys. Rev. B 74, 115116 (2006).
10. S.-T. Wu, U. Efron, and L. D. Hess, "Birefringence
measurements of liquid crystals," Appl. Opt. 23, 3911-
3915 (1984).
4
