Metamaterials combine the rigidity, practicality and scalability of other classical
approaches with vanishing dimension along the direction of wave propagation.
cells interfere destructively, resulting in
“bright” (strongly radiative) and “dark”
(sub-radiant) states. This leads to very
similar transmission profiles with asymmetric resonance line shapes that are
characteristic of Fano interference controlled by the strength of the coupling
of the metamaterial elements.
As in most classical implementations
of EIT, the operating frequency of these
metamaterials can be tuned using simple
geometrical scaling; however, in contrast
to all other approaches, the metamaterial structures are planar and do not
extend along the propagation direction.
In fact, in the microwave and terahertz
parts of the spectrum, the structure can
have virtually “zero” thickness. On the
other hand, in the optical domain, the
dimension of the metamaterial along the
propagation direction will be limited
by the optical skin depth in metals and
fabrication restrictions.
In both cases, the thickness of the
structures is much smaller than the
wavelength, since the resonators lie on
the plane perpendicular to the propagation direction of the incident wave. This
can lead to very compact realizations of
delay lines and possibly components of
future optical buffers, where the dimensions are basically limited by the number
of resonators required to provide a narrow divergence re-radiated beam.
Addressing limitations
A number of issues must be addressed
before the concept of metamaterial-induced transparency can be transferred
to the optical part of the spectrum and
find its way to practical applications. The
most important limitation—and one that
applies to most optical metamaterials—
is the strong dissipation associated with
metals at these frequencies. This can be
overcome by placing a metamaterial array
with EIT-like properties on an active
substrate, providing moderate gain that
leads to the appearance of a narrow resonance of strong single-pass amplification.
Developing this further, we show that
a steep normal dispersion occurs at the
amplification resonance, thus offering a
solution to the loss problem.
For example, we consider an array of
silver asymmetrically split rings with a
radius of 60 nm and a cross-section of
30 3 50 nm. It is placed on a 200-nm-
thick active substrate with gain levels
that correspond to about 5 percent
single-pass amplification of the substrate without the ring array. At the
frequency of the trapped-mode resonance, almost two-times single-pass
amplification occurs accompanied
by very sharp normal dispersion. The
required levels of gain can be provided
by readily available optically or injec-tion-pumped quantum-well structures
or optically pumped semiconductor
quantum dots. The feature size of the
array is well within the range of the
available fabrication capabilities, such
as electron beam lithography.
Another notable restriction in the
performance of EIT metamaterials is
that their operation is limited to normal
incidence and specific polarization.
However, very recently, we demonstrated
an all-angle, all-polarization metamaterial, which is based on concentric
rings with slightly different radii. This
small difference in ring radius leads to
two closely spaced resonances, each of
which corresponds to one of the rings.
Interference of these resonances leads to
EIT-like behavior similar to the asymmetrically split-ring metamaterial. Due
to the symmetry of the unit cell resonators, such a metamaterial is isotropic.
Weak interactions between different
elements in the array minimize the
metamaterial angle dependence, rendering it practically insensitive to the angle
of incidence for angles up to 45 degrees.
Finally, a common drawback in many
approaches is that the resonant sharp dispersion responsible for slow light behavior
occurs only within a narrow frequency
range. This narrow-band behavior
inherent in most strongly dispersive
resonant phenomena conflicts with the
wide-bandwidth requirements of practical
applications, such as optical buffers.
In planar metamaterials, we showed
that this restriction can be alleviated,
and the bandwidth of operation can be
significantly enhanced by the successive
stacking of metamaterial slabs. Coupling
between adjacent layers leads to multiple
closely spaced trapped-mode resonances
with continuously normal dispersion.
At the same time, the thickness of the
resulting multi-layer structure is still
much smaller than the wavelength.
In conclusion
Metamaterials are very attractive candidates for compact optical delay and slow-light devices. They combine the rigidity,
practicality and scalability of other classical approaches with vanishing dimension
along the direction of wave propagation;
this makes them well-suited to manufacturing using existing planar fabrication technologies. t
Nikolay Zheludev ( niz@orc.soton.ac.uk) and
Nikitas Papasimakis are with the Optoelectron-
ics Research Centre at the University of
Southampton, United Kingdom. Member
[ References and Resources ]
>> U. Fano. Phys. Rev. 124, 1866 (1961).
>> S.E. Harris et al. Phys. Rev. Lett. 64, 1107
(1990).
>> C.L.G. Alzar et al. Am. J. Phys. 70, 37
(2002).
>> V.A. Fedotov et al. Phys. Rev. Lett. 99,
147401 (2007).
>> N. Liu et al. Adv. Mat. 20, 4521 (2008).
>> N. Papasimakis et al. Phys. Rev. Lett. 101,
253903 (2008).
>> S. Zhang et al. Phys. Rev. Lett. 101,
047401 (2008).
>> N.I. Zheludev et al. Nature Photon. 2, 351
(2008).
>> R. Singh et al. Phys. Rev. B 79, in press
(2009).
>> P. Tassin et al. Phys. Rev. Lett. 102,
053901 (2009).
>> N. Papasimakis et.al. Appl. Phys. Lett. 94,
211902 (2009).
