Available online at www.sciencedirect.

com

Optical Materials 30 (2008) 1065–1069

www.elsevier.com/locate/optmat

2D photonic crystals on the Archimedean lattices (tribute

to Johannes Kepler (1571–1630))
a,*
R. Gajić , D. Jovanović a, K. Hingerl c, R. Meisels b, F. Kuchar b

a
Institute of Physics, P.O. Box 68, 11080 Belgrade, Serbia
b
Christian Doppler Laboratory, Institute for Semiconductors and Solid State Physics, University of Linz, Altenbergstrasse 69, A-4040 Linz, Austria
c
Institute of Physics, University of Leoben, Franz Josef Strasse 18, A-8700 Leoben, Austria

Available online 28 June 2007

Abstract

Results of our research on 2D Archemedean lattice photonic crystals are presented. This involves the calculations of the band struc-
tures, band-gap maps, equifrequency contours and FDTD simulations of electromagnetic propagation through the structures as well as
an experimental verification of negative refraction at microwaves. The band-gap dependence on dielectric contrast is established both for
dielectric rods in air and air-holes in dielectric materials. A special emphasis is placed on possibilities of negative refraction and left-hand-
edness in these structures. Together with the familiar Archimedean lattices like square, triangular, honeycomb and Kagome’ ones, we
consider also, the less known, (32, 4, 3, 4) (ladybug) and (3, 4, 6, 4) (honeycomb-ring) structures.
Ó 2007 Elsevier B.V. All rights reserved.

Keywords: Photonic crystals; Archimedean lattices; Kepler; Negative refraction; Left-handedness

1. Introduction THz or higher frequencies. Besides many applications,

PhCs are the source of several interesting effects like sup-
So far photonic crystals (PhC) or photonic band-gap pression (enhancement) of spontaneous light emission,
materials have attracted a lot of attention both in funda- localization of light, left-handed (LH) negative refraction,
mental and applied research [1]. The main property of Cs right-handed (RH) negative refraction and light focusing
is a periodically modulated index of refraction (mostly in below the diffraction limit. Band-gaps in PhCs depend on
dielectric materials). The basic concept of PhCs lies in anal- the dielectric contrast, filling factor, topology and structure
ogy with electrons in conventional crystals. Namely, simi- symmetry.
lar as for electrons in a periodic atomic potential, there
are band-gaps for photons in PhCs. Since an electronic 2. Archimedean lattices
band-gap represents the main feature behind all semicon-
ductor devices, PhC structures could provide a tool for a Here we study the PhC properties on the Archimedean
similar control of light. PhCs have already found many lattices. The Archimedean lattices shown in Fig. 1 make
applications, such as antennas, waveguides, resonators, fil- infinite tessellation consisting of regular convex polygons.
ters, lasers, light-emitting diodes or photonic integrated cir- The polygons are placed edge-to-edge to each other and
cuits. It is also worth mentioning there is a potential use of are not necessarily identical. Although the Archimedean
PhCs for the next generation of computers when photons 2D lattices have been known since ancient times, Johannes
can play the role of electrons in the present devices at Kepler was the first who gave a complete description of all
11 possible tilings as in Fig. 1, in his Harmonices Mundi
(Harmony of the Worlds) in 1619 [2] (Fig. 2). The name
*
Corresponding author. is after Kepler’s reference to Archimedes’ regular solid
E-mail address: [email protected] (R. Gajić). polyhedra. Kepler tried to explain the distances between

0925-3467/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.optmat.2007.05.024
1066 R. Gajić et al. / Optical Materials 30 (2008) 1065–1069

Fig. 3. Kepler’s model of the solar system based on platonic solids and his
original drawings.

Fig. 1. The 11 Archimedean lattices designated using the notation of

Grünbaum and Shephard.
shape and number of polygons around each vertex,
ðmn1 n2
1 ; m2 ; . . .Þ.
Starting from the smallest polygon and going clockwise
around a vertex, the numbers mi denote the number of
sides of each polygon, and the superscript ni refers to the
number of equal adjacent polygons. We distinguish vertices
by type and species. For example the (33, 42) and (32, 4, 3, 4)
lattices are of the same species but of different type.
One application of the Archimedean lattices in PhCs
concerns the search for structures with nearly isotropic
optical properties. Especially, the isotropy of band-gaps
is a desirable property for several applications of PhCs.
A number of different solutions have been suggested like
quasicrystals or circular PhCs which have higher order
local symmetries than the Bravais lattices. Some of the
Archimedean lattices, e.g. (32, 4, 3, 4), have also a higher
order local symmetry which together with lattice periodic-
ity could be advantageous over photonic quasicrystals.

3. Archimedean lattice photonic crystals

Fig. 2. Kepler’s Harmonices Mundi published in Linz 1619. We started with the simple Archimedean lattice PhCs, as
are square (Fig. 1i), triangular (Fig. 1j), honeycomb
(Fig. 1k) and Kagome’ (Fig. 1k) lattices. Our study
the planets and sun by spheres of perfect polyhedra placed involved calculations of band structures and band-gap
inside each other as shown in Fig. 3. He associated five maps (plane-wave method), equifrequency contours (EFC
platonic solids with five intervals between the six then – x(kx, ky)), FDTD (finite-difference time-domain) simula-
known planets, Mercury, Venus, Earth, Mars, Jupiter tions and transmission experiments at microwaves (26–
and Saturn. 60 GHz). We were mostly interested in negative refraction
Among the Archimedean lattices, we find the familiar and left-handedness in these PhCs. Veselago [4], in his sem-
square, triangular, honeycomb and Kagome’ structures. inal work, was the first who predicted LH materials inves-
The first three are called regular lattices since they consist tigating hypothetical materials which had both the
of equal polygons. So far the Archimedean lattices have electrical permittivity e and the magnetic permeability l
been mostly studied in mathematics and statistical mechan- simultaneously negative. Veselago found out that the index
ics [3]. Therefore, proper names have not been assigned yet of refraction is negative while the Poynting vector S and
to all Archimedean lattices. Instead, the mathematical the wave vector k (phase velocity) are anti-parallel produc-
Grünbaum–Shephard notation is in use in terms of ing backward propagating waves. Veselago’s work has
 R. Gajić et al. / Optical Materials 30 (2008) 1065–1069 1067

stayed just an interesting idea for more than 30 years until

Pendry et al. [5] proposed and tested the new meta-materi-
als consisting of conducting loops (split-ring resonator
(SRR)) or tubes (‘Swiss rolls’). Their magnetic permeability
l has a resonance and a narrow frequency range with
l < 0. Soon, Shelby et al. [6] experimentally verified nega-
tive refraction at 10. 5 GHz in the first LHM structure
based on SRRs interconnected with a 2D metallic rod lat-
tice (e < 0).
In contrast to Veselago’s meta-materials, photonic crys-
tals (PhC) consist usually of periodically modulated dielec-
tric material where l > 0. Nevertheless, it was shown that
diffraction effects (k  lattice constant) can produce effec-
tive negative refraction or even negative index and LH as
in Veselago’s meta-materials [7–10]. The structures show-
ing negative refraction can be either the LH or RH media.
The losses can be much smaller in PhCs since non-conduct- Fig. 4. FDTD simulation of the 67 GHz TE2 wave incident at 15° across
ing dielectric materials are used. Both negative refraction the CM interface of a square lattice PhC made of Al2O3 rods in air. The
red and yellow arrows denote vgr and vph. (For interpretation of the
and LH in PhCs have been experimentally demonstrated
references in colour in this figure legend, the reader is referred to the web
in the microwave range [11–15]. The refraction experiments version of this article.)
were performed on 2D dielectric and metallic PhCs by mea-
suring the displacement of the transmitted beam at varying
angle of incidence.
The necessary condition to be fulfilled for LH in PhCs is
vph Æ vgr < 0 or equivalently k Æ S < 0, since vgr is parallel to
S in large enough PhCs. The symbols vph, vgr and S indicate
the phase and group velocity and the Poynting vector. We
distinguish between RH and LH behavior of the wave
propagation in PhCs on the sign of the vph Æ vgr product.
The necessary conditions for negative refraction require
that the equifrequency contours (EFCs) in PhCs are both
convex (with an inward gradient) and larger than the cor-
responding EFCs in air. Additionally ko  c/f P 2 Æ as (as
is the surface lattice constant) in order to avoid higher
order Bragg diffractions out of the crystal. Negative refrac-
tion can be observed around the symmetric points in the
Brillouin zone whereas LH behaviour is only possible
around the C point. In the case of the TM2/TE2 bands
around the C point, there is LH negative refraction of
Fig. 5. FDTD simulation of the 36 GHz TM1 wave incident at 45° across
the Veselago type where vgr and vph are opposite and, even
the CM interface of a 2D square lattice PhC made of alumina rods in air.
LH positive refraction takes place in the TM3 band due to The red and yellow arrows denote group and phase velocity, respectively.
anisotropy [10]. It turned out that regular Archimedean lat- The corresponding microwave measurements are given in the inset. (For
tices like square (44) or triangular (36) structures enable interpretation of the references in colour in this figure legend, the reader is
most effects to be observed in PhCs like negative refraction referred to the web version of this article.)
(including all-angle RH negative refraction [8]), left-hand-
edness or superlens focusing.
In a square lattice PhC made of alumina in air, negative of edge-to-edge placement of the adjacent polygons is not
refraction has been studied in the first two TM/TE bands. fulfilled. The PhC in Fig. 6 made of alumina rods has a
As an example, the LH negative refraction of the TE2 large gap-to-midgap ratio of 40%. The FDTD simulation
mode is presented in Fig. 4. and RH negative one of the for propagation at k/a = 4.54 (a is the near-neighbor rod
TM1 wave [10,15] in Fig. 5. distance) revealed negative refraction.
Based on the square lattice we designed the square-par- In the case of hexagonal structure, we studied the Kag-
quet or square-honeycomb structure [12] as shown in ome’ lattice and honeycomb or graphite lattice (dual to tri-
Fig. 6. The lattice is a 2D analog of the well known 3D angular graph). Comparing to the triangular lattice,
woodpile lattice. honeycomb and Kagome’ lattices have two and three atom
At first sight, the structure looks as another Archime- unit cells, respectively. As a result, more complicated band
dean one but close inspection reveals that the condition structures appear. We demonstrated that both lattices
1068 R. Gajić et al. / Optical Materials 30 (2008) 1065–1069

Fig. 8. (32, 4, 3, 4) or ladybug lattice with the four-atom unit cell cut along
the CX axis.

Fig. 6. Square-parquet PhC made of alumina rods in air and its three
atom unit cell. The inset denotes both the measured and calculated
transmission spectra.

(rods in air) had all-angle left-handed refraction for the

TM2 band [16] where both effective indices, npeff (x, hi) =
sgn(vgr Æ kPhC) Æ cjkPhCj/x = ±kPhC/kair and nbeam(x, hi) =
sin(hi)/sin(hr) (x, hi, and hr are the frequency, incident
and, refracted angle, respectively) are close to 1. In
Fig. 7 the LH negative refraction in the TM2 band of the
honeycomb PhC is presented.
Among irregular Archimedean lattices, we studied the
(32, 4, 3, 4) – ladybug and (3, 4, 6, 4) structures. Particularly,
the first one and corresponding derived non-Archimedean
lattices are interesting [17–19] since they possess near and
Fig. 9. LH negative refraction of the TM4 mode in the ladybug lattice
full 12-fold local rotational symmetry enabling isotropic PhC made of Al2O3 rods in air (e = 9.6, r/a = 0.35 and k/a = 4.55) across
band-gaps. The relative band-gap variations as a function the CX interface. The incident angle is 30°. The EFCs of the TM4 band are
of propagation direction are between 1% and 3%. The lady- presented in the inset.
bug lattice is shown in Fig. 8.
In Fig. 9 we present the LH negative refraction in the 4. Conclusion
TE3 band of a ladybug lattice PhC made of alumina
(e = 9.6) rods in air. In review, we described part of our research on the regu-
lar and irregular Archimedean lattice PhCs. Particulary, we
investigated possibilities of negative refraction and left-
handedness in these structures and for the first time LH neg-
ative refraction is demonstrated in the ladybug structure.

Acknowledgements

This work was supported by the Serbian MSEP projects

1469, 141047, the Austrian FWF project 15513 and the
Christian Doppler Laboratory in Linz. K.H. is grateful
for partial support from European Community Project
N2T2. We are also, grateful to Photeon and Heinz Syringer
from Photeon Technologies for financial support and Joh-
ann Messner from the Linz Supercomputer Center for
technical support.

References

Fig. 7. 26 GHz TM2 wave propagating through a honeycomb GaAs rod [1] K. Sakoda, Optical Properties of Photonic Crystals, second ed.,
PhC incident at 30° across the CK interface (FDTD). Springer, 2005.
 R. Gajić et al. / Optical Materials 30 (2008) 1065–1069 1069

[2] J. Kepler, Harmonices Mundi, Linz, 1619. [12] R. Gajic, F. Kuchar, R. Meisels, J. Radovanovic, K. Hingerl, J.
[3] P.N. Sudung, R.M. Ziff, Phys. Rev. E 60 (1999) 275. Zarbakhsh, J. Stampfl, A. Woesz, Zeit. Metall. 95 (2004) 618.
[4] V.G. Veselago, Usp. Fiz. Nauk 92 (1967) 517; [13] K. Guven, K. Aydin, K.B. Alici, C.M. Soukoulis, E. Ozbay, Phys.
V.G. Veselago, Usp. Fiz. Nauk 173 (2003) 790. Rev. B 70 (2004) 205125-1.
[5] J.B. Pendry, A.J. Holden, D.J. Robins, W.J. Stewart, IEEE Trans. [14] P.V. Parimi, W.T. Lu, P. Vodo, J. Sokoloff, J.S. Derov, S. Sridhar,
Microwave Theor. Tech. 47 (1999) 2075. Phys. Rev. Lett. 92 (2004) 127401-1;
[6] R.A. Shelby, D.R. Smith, S. Schultz, Science 292 (2001) 77. P. Vodo, P.V. Parimi, W.T. Lu, S. Sridhar, R. Wing, Appl. Phys.
[7] M. Notomi, Phys. Rev. B 62 (2000) 10696. Lett. 85 (2004) 1858.
[8] C. Luo, S.G. Johnson, J.D. Joannopoulos, J.B. Pendry, Phys. Rev. B [15] R. Meisels, R. Gajic, F. Kuchar, K. Hingerl, Opt. Exp. 14 (2006)
65 (2002) 201104-1. 6766.
[9] S. Foteinopoulou, E.N. Economou, C.M. Soukoulis, Phys. Rev. Lett. [16] R. Gajic, R. Meisels, F. Kuchar, K. Hingerl, Phys. Rev. B 73 (2006)
90 (2003) 107402-1. 165310.
[10] R. Gajic, R. Meisels, F. Kuchar, K. Hingerl, Opt. Exp. 13 (2005) [17] S. David, A. Chelnokov, J.M. Lourtioz, Opt. Lett. 25 (2000) 1001.
8596. [18] M. Rattier, H. Benisty, E. Schwoob, C. Weisbuch, T.F. Krauss,
[11] E. Cubukcu, K. Aydin, E. Ozbay, S. Foteinopoulou, C.M. Soukoulis, C.J.M. Smith, R. Houdre’, U. Oesterle, Appl. Phys. Lett. 83 (2003)
Nature 423 (2003) 604; 1283.
E. Cubukcu, K. Aydin, E. Ozbay, S. Foteinopoulou, C.M. Soukoulis, [19] A. David, T. Fujii, E. Matioli, R. Sharma, S. Nakamura, H. Benisty,
Phys. Rev. Lett. 91 (2003) 207401-1. Appl. Phys. Lett. 88 (2006) 073510-1.