Home Search Collections Journals About Contact us My IOPscience

A dual band terahertz metamaterial absorber

This content has been downloaded from IOPscience. Please scroll down to see the full text.

2010 J. Phys. D: Appl. Phys. 43 225102

(http://iopscience.iop.org/0022-3727/43/22/225102)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 128.248.155.225
This content was downloaded on 15/11/2014 at 13:00

Please note that terms and conditions apply.

IOP PUBLISHING JOURNAL OF PHYSICS D: APPLIED PHYSICS
J. Phys. D: Appl. Phys. 43 (2010) 225102 (5pp) doi:10.1088/0022-3727/43/22/225102

A dual band terahertz metamaterial

absorber
Hu Tao1 , C M Bingham2 , D Pilon3 , Kebin Fan1 , A C Strikwerda3 ,
D Shrekenhamer2 , W J Padilla2 , Xin Zhang1 and R D Averitt3
1
Department of Mechanical Engineering, Boston University, Boston, MA 02215, USA
2
Department of Physics, Boston College, Chestnut Hill, MA 02467, USA
3
Department of Physics, Boston University, Boston, MA 02215, USA
E-mail: xinz@bu.edu and raveritt@physics.bu.edu

Received 29 July 2009

Published 19 May 2010
Online at stacks.iop.org/JPhysD/43/225102

Abstract
We present the design, fabrication and characterization of a dual band metamaterial absorber
which experimentally shows two distinct absorption peaks of 0.85 at 1.4 THz and 0.94 at
3.0 THz. The dual band absorber consists of a dual band electric-field-coupled (ELC) resonator
and a metallic ground plane, separated by an 8 m dielectric spacer. Fine tuning of the two
absorption resonances is achieved by individually adjusting each ELC resonator geometry.
(Some figures in this article are in colour only in the electronic version)

Metamaterials consisting of artificially constructed electro-
magnetic (EM) materials have recently attracted considerable
interest due to their ability to exhibit engineered exotic EM
responses not available in nature, including negative refractive
index, superlensing and cloaking [14]. Many of these ideas
were initially implemented at microwave frequencies due to the
simplicity in fabrication. Nevertheless, there has been substan-
tial progress and successful demonstrations of metamaterials
extending from the terahertz through the visible frequencies
using micro/nano fabrication technologies during the past sev-
eral years [57]. Development of metamaterials at terahertz
frequencies is especially important and attractive since there
is a strong need to create components to realize applications
ranging from spectroscopic identification of hazardous materi-
als to noninvasive imaging. However, it is difficult to find natu-
rally existing materials with strong absorption coefficients that
are also compatible with standard microfabrication techniques.
Thus, metamaterials are a promising approach to manipulate
terahertz waves and to fill this terahertz gap [8, 9].
Sub-wavelength split ring resonators (SRRs) have been
extensively employed to exhibit a magnetic or electric
resonance response at terahertz frequencies [611]. Although
SRR structures are bianisotropic, it is possible to eliminate the
magnetoelectric coupling by making symmetrical structures
such as electric-field-coupled (ELC) resonators with shared
centre gap or shared centre bar. A pure electric resonance
response was experimentally demonstrated at microwave
frequency with the EM wave propagating along the substrate
surface [12]. However, researchers have found that such ELC
resonators could show similar resonance responses with the
EM wave propagating normal to the substrate and the electric
field aligned perpendicular to the resonators gap [10]. To
date, due to the limitations of currently available fabrication
and characterization techniques, most terahertz metamaterials
were made by patterning periodic SRRs in a single planar
form to show effective negative electric permittivity () or
negative magnetic permeability (). Additional layers or more
complex structures need to be introduced for tailoring both
and , which is important to realize negative refractive index.
Tailoring the effective and also provides the means to
engineer the impedance.
Recently, a narrow band microwave resonant metamaterial
absorber with theoretical unity absorptivity realized through
perfect impedance matching was demonstrated [13]. Progress
has been made to extend the absorber design to terahertz
frequencies by using double metamaterial layers [14,
15]. Although those metamaterial absorbers show decent
absorptivity at terahertz frequencies, they are basically single
band absorbers with high absorptivity at a specific frequency.
Multi-band terahertz absorbers with two or more optimized
absorption peaks have yet to be demonstrated, which could
be important, for example, in the development of terahertz
spectroscopic imagers/detectors [16, 17].
We experimentally demonstrate a dual band resonant
metamaterial absorber with two distinct absorption peaks at
1.4 and 3.0 THz. We use an absorber design similar to [15].

0022-3727/10/225102+05$30.00 1 2010 IOP Publishing Ltd Printed in the UK & the USA
J. Phys. D: Appl. Phys. 43 (2010) 225102
h1
h2
(a)
Figure 1. Dual band terahertz metamaterial absorber consisting of a dual band ELC resonator and a metallic ground plane, separated by an
8 m thick dielectric layer. (a) Perspective view of the absorber. (b) Top view of the absorber. (c) Photograph of a portion of the fabricated
absorber. Corresponding dimensions are shown in table 1. (Colour online.)
The main difference is that we combine two single band
ELC resonators together to obtain the dual band resonance
responses, inspired by [18]. Remarkably, the two resonance
responses can be tuned and optimized independently at desired
frequencies with comparably high absorptivity as with single
band metamaterial absorbers. This feature provides more
flexibility in multi-band absorber designs and can be readily
extended to infrared and visible frequency ranges.
As described in [14, 15], maximizing the absorption (A)
is equivalent to minimizing both the transmission (T ) and the
reflectivity (R) simultaneously at the same frequency range.
The near-unity resonant metamaterial absorber can be realized
by manipulating the effective and to get perfect impedance
matching with free space (i.e. Z = Z0 , resulting in R = 0),
with the loss dependent imaginary part of the refractive index
(n2 ) as large as possible.
A dual band metamaterial absorber consists of two
metallic layers separated by a dielectric spacer. The top layer
consists of an array of dual band ELC resonators which is
responsible for determining  (), while the bottom metallic
ground plane layer is added such that the incident magnetic
field drives circulating currents between the two layers, as
shown in figure 1. Two single band metamaterial absorbers
with only one single ELC resonator on the top layer are first
designed and optimized near 1.4 THz (S1) and 2.9 THz (S2),
and are then integrated into a single particle, as shown in
figure 2. Despite coupling between these two ELC resonators
due to the shared bar (see D2), two distinct absorption peaks
are obtained by fine tuning of the overall geometry.
The absorber structures were simulated and optimized
using commercialized full-wave EM simulation software CST
Microwave Studio 2008. Unit cell boundary conditions
with a tetrahedral mesh were used in the frequency solver,
in which case the wave ports were automatically added in the
direction of normal incidence to the substrate. The optimized
geometries are shown in table 1. The Au portions of the
absorbers were modelled as lossy gold with a conductivity
= 4.09107 S cm1 , and the polyimide layer was modelled
using the experimentally determined value of  = 1 + i2 =
2.88 + i0.09. The transmission amplitude S21 and reflection
amplitude S11 were obtained and the absorption was calculated
using A = 1 R T = 1 S11 2
S21 2
, where S21 is zero
H Tao et al
w
g1
g2
b 30 m
30m
a
(b) (c)
across the entire frequency range. Thus, the absorptivity can
be calculated using A = 1 R, which makes optimizing the
absorptivity equivalent to minimizing the reflectivity. The
optimized parameters are those which yielded the lowest
reflectivity at the design frequencies of 1.4 and 2.9 THz.
As observed in [15], a metamaterial absorber based on a
different kind of single ELC resonator, referred to as electric
SRR (e-SRR), also shows two absorption peaks at two distinct
frequencies. However, it is different from the dual band
absorber due to the different origin of the absorption. The
low frequency resonance is due to countercirculating currents
in the e-SRR, while the high frequency resonance derives from
a dipolar response [19], analogous to what has previously
been observed using frequency-selective surfaces (FSSs) at
microwave frequencies [20, 21]. In the dual band absorber,
both resonances come from circulating currents in the ELC
resonators, while the low frequency response is determined
mainly by the upper resonator (S1) and the high frequency
response is contributed mainly by the lower resonator (S2).
This was verified by examining the surface current distribution
on resonance, as shown in figures 2(c) and (d). The two
resonance peaks are labelled as I (at 1.4 THz) and II (at
2.9 THz), respectively. To better understand the absorption
of each resonance, an absorber based on e-SRR with centre
split gap (D1), similar to the design in [15], was designed and
optimized with similar absorption peaks as the dual band ELC
resonator absorber (D2) near 1.4 and 2.9 THz. In comparison
with the e-SRR absorber (D1), the dual band ELC absorber
(D2) has a similar low frequency response. However, it
has significantly higher and narrower resonance absorption
response at the higher frequency.
An additional aspect to consider in the design of
metamaterial absorbers is losses in the constituent materials
making up the structure. As discussed in the introduction, one
of the design criteria is to obtain a large value of the imaginary
part of the effective refractive index. This necessitates having
some losses in the metal. Losses in the dielectric spacer are
expected to contribute as well. For example, in the limit
of a perfect electric conductor and a lossless dielectric, the
absorption in the composite in figure 1 is zero. However,
losses in gold are sufficient to yield a strong narrow band
resonance. The absorption as a function of frequency at various
2
J. Phys. D: Appl. Phys. 43 (2010) 225102 H Tao et al

(a) S1 S2
S2 E

Absorptivity
S1
I

(c)
(b) I II
E
Absorptivity

D2

k
g h II
D1
D1
D2

Frequency (THz) (d)

Figure 2. (a) and (b) show calculated absorptivity of the absorbers composed of S1 (red long dash), S2 (blue short dash), D1 (purple dot
line) and D2 (black solid line). (c) and (d) show simulated surface current density for the two absorptivity peaks on resonance with the
electric field (E) perpendicular to the resonatorsgap. (Colour online.)

Table 1. Dimensions of terahertz metamaterial absorbers (all units in m).

Sample a b w h1 h2 g1 g2 t h g k
S1 33.3 23.3 2 20.7 2 8
S2 33.3 23.3 2 8.7 2 8
D1 42.5 32.5 2 8 25 2 9
D2 33.3 23.3 2 18 7 2 2 8

Note: a, lattice period; b, outer dimension; w, line width; h1, height of S1; h2,
height of S2; g1, gap distance of S1; g2, gap distance of S2; t, thickness of the
polyimide spacer; h, height of D1; g, gap distance of D1; k, length of the gap
of D1.

values of the dielectric loss tangent tan() are calculated for the
metamaterial absorber designs D1 and D2, as shown in figure 3.
The results in the figure start from a lossless dielectric
(tan() = 0, black curve) and gradually increased tan() to 0.2
(while keeping the value of the Au conductivity constant). The
e-SRR absorber (D1) shows different absorption responses at
two resonant frequencies with increasing tan() due to different
resonant origins mentioned above. The peak absorption is 0.94
at 1.4 THz with a lossless dielectric, which is smaller than the
calculated value in figure 2(b). This suggests losses in the
dielectric contribute to increasing the absorption. However,
a point of diminishing returns is reached for larger values
of tan() in that the absorption decreases and the resonance
broadens. For the higher frequency resonance, the absorption
keeps increasing with larger dielectric losses, which conforms
to those FSSs reported previously [22].
The dual band ELC absorber (D2) shows similar
absorption responses with various dielectric loss values at
both resonant frequencies as the e-SRR absorber (D1) does
at low frequency. In the case of tan() = 0, the peak
absorption is 0.86 at 1.4 THz and 0.84 at 2.9 THz, respectively,
which are nearly 15% smaller than the calculation value in
figure 2(b). The absorption can be increased by increasing
the dielectric losses, and it drops back after a certain value,
which is 0.04 in this case. These results suggest that
optimization of tan() of the dielectric spacer can maximize
the metamaterial absorption. Further, it appears the losses
in polyimide (tan() = 0.031) should contribute 0.15 to
the absorption of our dual band metamaterial absorber as is
evident by comparing the black curves in figures 2(b) and
3(b). Besides optimization tan() of the dielectric spacer to
maximize the absorption, we can also optimize the absorber

3
J. Phys. D: Appl. Phys. 43 (2010) 225102 H Tao et al

(PI-5878G, HD MicroSystems) was spin-coated on the

(a) ground plane to form the 8 m thick polyimide spacer. Then
E the ELC resonator arrays were patterned with a conventional
lithography technology on the polyimide spacer. The lift
Absorptivity

off process was accomplished by e-beam evaporating another

D1 50 nm thick Au/Ti film, followed by rinsing in acetone for
several minutes. It is worth mentioning that the whole
fabrication process could be easily repeated at other substrates
including soft substrates such as polyimide to make it a
mechanically flexible dual band metamaterial absorber for
non-planar applications.
A Fourier transform infrared (FTIR) spectrometer was
0.00 used to experimentally verify the behaviour of the absorbers
(b) 0.04 by measuring the reflection response over the frequency range
0.08 of 13.4 THz with a resolution of 15 GHz. A liquid helium
0.12
cooled bolometer detector and a 6 m mylar beam splitter were
Absorptivity

0.20
used to optimize the FTIR performance over the frequencies
E measured. The absorber samples were diced into 1 cm 1 cm
squares and mounted with the electric field perpendicular to
the ELC resonator gaps. The samples were mounted on
D2
a commercially purchased reflection unit and the minimum
achievable incident angle for reflection measurements in
our current setup is constrained to 30 off-normal due to
the experimental limitations [23]. The measurements were
Frequency (THz) performed with an electric field perpendicular to the ELC
Figure 3. Calculated absorptivity as a function of frequency for resonator gap to excite the electric resonance. The absorption
various values of the dielectric loss tangent used in (a) single e-SRR spectrum was easily obtained from the reflection results.
metamaterial absorber D1 and (b) dual band ELC metamaterial The experimental results for all four absorbers are
absorber D2. The electric field (E) is aligned perpendicular to the
displayed in figure 4. For the single ELC resonator absorbers,
resonatorsgap. The curves are labelled with the value of tan() used
in the simulation. (Colour online.) S1 shows an absorptivity of 0.74 at 1.46 THz and S2 shows
an absorptivity of 0.93 at 2.99 THz, which is in reasonable
Table 2. Calculated absorptivity peaks of dual band metamaterial agreement with the simulation though the absorption peaks
absorber as a function of the dielectric loss tangent value and the are lower than expected partly due to fabrication imperfections.
spacer thickness.
For dual band resonator absorbers, D1 shows an absorptivity
Dielectric tan() t A1 (%) A2 (%) B1 (%) B2 (%) of 0.79 at 1.43 THz and another absorption peak of 0.86 at
1 0.00 5.5 85.81 83.97 99.97 99.75 2.95 THz; D2 shows two strong resonance absorption peaks of
2 0.04 8 99.96 99.81 99.96 99.81 0.85 at 1.41 THz and 0.94 at 3.02 THz. There is a noticeable
3 0.08 10 95.34 97.96 99.86 99.93 frequency shift (0.1 THz) at the high resonant frequency
4 0.12 12 87.25 92.85 99.64 99.64 between experimental and simulation results in D2, which
5 0.20 16 72.55 82.98 98.70 98.01
is mainly due to the inaccuracy of the fabricated resonator
Note: tan(), loss tangent of the dielectric spacer; t, modified gap distance. The off-resonance absorptivity (0.3) is quite
thickness of the dielectric spacer in m; A1, absorptivity peak large in disagreement with the simulations (0.1), which may
value at low frequency with 8 m thick dielectric spacer; A2, result partly from the scattering of radiation due to fabrication
absorptivity peak value at high frequency with 8 m thick
dielectric spacer; B1, absorptivity peak value at low frequency imperfections. Although the absorption peaks are lower than
with modified dielectric spacer thickness; B2, absorptivity peak the simulated values, it is clear that the absorber consisting
value at high frequency with modified dielectric spacer of dual band ELC resonators (D2) shows higher absorption
thickness. and better spectral resolution than the single e-SRRs (D1),
and therefore should be a better candidate to be included as
geometry for various dielectric spacers with different losses. an active absorbing element in spectrally selective terahertz
By simply tuning the thickness of the dielectric spacer without detectors.
having to modify the dual band ELC resonators geometry, In summary, a dual band terahertz metamaterial absorber
the absorption peaks can be pulled back to near unity at both consisting of two single ELC resonators combined together and
resonant frequencies, as shown in table 2. a metallic ground plane, separated by a dielectric spacer, has
The above-mentioned absorber structures were fabricated been demonstrated. Two distinct resonant absorption peaks at
on silicon substrates with the parameters shown in table 1. A 1.4 and 3.0 THz were experimentally observed, and the two
50 nm thick Au/Cr film was first e-beam evaporated on a 2 absorption peaks can be individually tuned by optimizing the
inch silicon wafer to form the ground plane. Liquid polyimide corresponding single ELC resonators geometries.

4
J. Phys. D: Appl. Phys. 43 (2010) 225102
(a)
Absorptivity
S1
S1
S2
(b)
D1
Absorptivity
D1
D2
Frequency (THz)
Figure 4. (a) and (b) Experimental absorptivity as a function of frequency. S1 (red long dash), S2 (blue short dash), D1 (purple dot line)
and D2 (black solid line). (Insets: photographs of portions of as-fabricated absorbers.) (c) Photo of the reflection unit used in the
measurement. (d) Schematic of the reflection measurement setup. (Colour online.)
Acknowledgments
This project has been supported in part by the DOD/Army
Research Laboratory under contract no W911NF-06-2-0040,
NSF under contract no ECCS 0802036, AFOSR under
contract no FA9550-09-1-0708, and DARPA under contract
no HR0011-08-1-0044. The authors would also like to thank
the Photonics Centre at Boston University for all the technical
support throughout the course of this research.
Note added in proof. During the course of this work we were
made aware of a paper with similar results [24].
References
[1] Smith D R, Padilla W J, Vier D C, Nemat-Nasser S C and
Schultz S 2000 Phys. Rev. Lett. 84 4184
[2] Shelby R A, Smith D R and Schultz S 2001 Science
292 77
[3] Pendry J B 2000 Phys. Rev. Lett. 86 3966
[4] Schurig D, Mock J J, Justice B J, Cummer S A, Pendry J B,
Starr A F and Smith D R 2006 Science 314 977
[5] Dolling G, Wegener M, Soukoulis C M and Linden S 2007
Opt. Lett. 32 53
[6] Yen T J, Padilla W J, Fang N, Vier D C, Smith D R,
Pendry J B, Basov D N and Zhang X 2004 Science 303 1494
[7] Tao H, Strikwerda A C, Fan K, Bingham C M, Padilla W J,
Zhang X and Averitt R D 2008 J. Phys. D: Appl. phys.
41 232004
[8] Chen H T, Padilla W J, Zide J M O, Gossard A C, Taylor A J
and Averitt R D 2006 Nature 444 597
H Tao et al
S2 (c)
In
Sample
Out
D2 (d)
[9] Chen H T, OHara J F, Azad A K, Taylor A J, Averitt R D,
Shrekenhamer D and Padilla W J 2008 Nature Photon.
2 295
[10] Padilla W J, Aronsson M T, Highstrete C, Lee M, Taylor A J
and Averitt R D 2007 Phys. Rev. B 75 041102(R)
[11] Bingham C M, Tao H, Liu X, Averitt R D, Zhang X and
Padilla W J 2008 Opt. Express 16 18565
[12] Schurig D, Mock J J and Smith D R 2006 Appl. Phys. Lett.
88 041109
[13] Landy N I, Sajuyigbe S, Mock J J, Smith D R and Padilla W J
2008 Phys. Rev. Lett. 100 207402
[14] Tao H, Landy N I, Zhang X, Averitt R D and Padilla W J 2008
Opt. Express 16 7181
[15] Tao H, Bingham C M, Strikwerda A C, Pilon D,
Shrekenhamer D, Landy N I, Fan K, Zhang X, Padilla W J
and Averitt R D 2008 Phys. Rev. B 78 241103(R)
[16] Tonouchi M 2007 Nature Photon. 1 97
[17] Zhang X-C 2002 Phys. Med. Biol. 47 3667
[18] Yuan Y, Bingham C M, Tyler T, Palit S, Hand T H,
Padilla W J, Jokerst N M and Cummer S A 2008 Appl.
Phys. Lett. 93 191110
[19] Padilla W J, Taylor A J, Highstrete C, Lee M and Averitt R D
2006 Phys. Rev. Lett. 96 107401
[20] Sievenpiper D F, Zhang L, Jimenez Broas R F,
Alexopolous N G and Yablonovitch E 1999 IEEE Trans.
Microwave Theory Tech. 47 2059
[21] Hao J M, Zhou L and Chan C T 2007 Appl. Phys. A 87 281
[22] Antonopoulos C, Cahill R, Parker E A, Sturland I M
1997 IEEE Proc. Microw. Antennas and Propagation
144 415
[23] Further information could be found on (http://www.harricksci.
com/variable-angle-reflection-accessory)
[24] Wen Q-Y, Zhang H-W, Xie Y-S, Yang Q-H and Liu Y-L 2009
Appl. Phys. Lett. 95 241111