Huret, J.F. Seaux, D. Cros, and V.

Madrangeas, Ferroelectric thin lms

for applications in high frequency range, Ferroelectrics 316 (2005),
712.
11. K. Kageyama, A. Sakurai, A. Ando, and Y. Sakabe, Thickness effects
on microwave properties of (Ba,Sr)TiO3 lms for frequency agile
technologies, J Eur Ceram Soc 26 (2006), 18731877.
12. V. Laur, A. Rousseau, G. Tanne, P. Laurent, F. Huret, M. Guilloux-Viry,
and B. Della, Tunable microwave components based on KTa1 xNbxO3
ferroelectric material, European Microwave Conference, Paris, France,
October 37, 2005, Vol. 1, pp. 641 644.
13. H.-J. Bae, D.P. Norton, J. Sigman, and L. Boatner, Low dielectric
losses in annealed Ti-doped K(Ta,Nb)O3 thin lms grown by pulsed
laser deposition, J Phys D: Appl Phys 38 (2005), 13311336.

2006 Wiley Periodicals, Inc.

EXTRACTION OF EFFECTIVE
METAMATERIAL PARAMETERS BY
Figure 9 Variation of the gure of merit of TL at 10 GHz as a function PARAMETER FITTING OF DISPERSIVE
of the gap width g. The gure of merit of MS lines has been included for
comparison purposes. Conductors: 3.8 107 (Aluminium), t 2 m;
MODELS
Substrate: sapphire, r 10, tan 104, h 500 m; Ferroelectric lm: G. Lubkowski, R. Schuhmann, and T. Weiland
rf 700, tan f 5 102, hf 0.5 m. 5 m g 30 m. Technische Universitat Darmstadt, Institut fur Theorie
MS: w 436 m. CPW: w 18 m. CS: w 310 Elektromagnetischer Felder (TEMF), Schlossgartenstr. 8, D-64289
m Darmstadt, Germany

Received 6 July 2006

rable devices. We should also mention that a technological pre-
requisite for large integration and development of ferroelectric ABSTRACT: Effective electric permittivity and magnetic permeability
tunable components is to reach a loss tangent of the order of 102. for metamaterial structures are extracted from 3D eld simulation data.
The equivalent representation of the metamaterial is a homogeneous
REFERENCES slab described with parameterized dispersive Drude and Lorentz models.
1. W. Kim, M.F. Iskander, and C. Tanaka, High performance low-cost The parameters of dispersive models are obtained by optimization. Pro-
phase-shifter design on ferroelectric materials technology, Electron posed approach is applied to the extraction of effective material param-
Lett 40 (2004), 13451347. eters for double negative metamaterial cells. 2006 Wiley Periodicals,
2. H. Yoon, K.J. Vinoy, J.K. Abraham, and V.K. Varadan, CPW phase Inc. Microwave Opt Technol Lett 49: 285288, 2007; Published online
shifter using barium strontium titanate thin lm on silicon substrate, in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop.
In: IEEE antennas and propagation society international symposium, 22105
Colombus, Ohio, June 2227, 2003, Vol. 3, pp. 970 972.
3. B. Acikel, T.R. Taylor, P.J. Hansen, J.S. Speck, and R.A. York, A new Key words: metamaterials; effective parameters
high performance phase shifter using BaxSr1 xTiO3 thin lms, IEEE
Microwave Wireless Compon Lett 12 (2002), 237239.
4. P.T. Teo, K.A. Jose, Y.B. Gan, and V.K. Varadan, Beam scanning of 1. INTRODUCTION
array using ferroelectric phase shifters, Electron Lett 36 (2000), 1624 The occurrence of rst works on negative electric permittivity and
1626. magnetic permeability structures [1, 2] gained enormous interest in
5. F.A. Miranda, G. Subramanyam, F.W. Van Keuls, R.R. Romanofsky, the scientic community [35]. One of the concepts for the con-
J.D. Warner, and C.H. Mueller, Design and development of ferroelec- struction of a double negative (DNG) metamaterial (MTM) cell is
tric tunable microwave components for Ku- and K-band satellite to use a combination of a split ring resonator (SRR) and a wire,
communications systems, IEEE Trans Microwave Theory Tech 48 providing negative magnetic permeability and negative electric
(2000), 11811189.
permittivity, respectively.
6. K.-B. Kim, T.-S. Yun, H.-S. Kim, R.-Y. Kim, H.-G. Kim, and J.-C.
Lee, An interdigital capacitor with high tunability and low loss tan-
MTM structures are built of periodically ordered cells, with the
gent, In: 34th European microwave conference, Amsterdam, The assumption that the lattice constant is much less than the wave-
Netherlands, October 1115, 2004, pp. 161164. length in the medium.
7. S. Gevorgian, S. Abadei, H. Berg, and H. Jacobson, MOS varactor There are several methods for the extraction of effective ma-
with ferroelectric thin lms, In: Microwave Symposium Digest, IEEE terial parameters for DNG MTM structures. The most popular
M TT-S international, Phoenix, Arizona, May 20 25, 2001, Vol. 2, pp. approach is the extraction from transmission and reection char-
11951198. acteristics of a MTM, the method known from laboratory as a
8. D. Kuylenstierna, G. Subramanyam, A. Vorobiev, and S. Gevorgian, common way to nd experimentally effective parameters of a
Tunable electromagnetic performance of coplanar waveguides period- material sample under test [6]. However, when applied to MTM
ically loaded by ferroelectric varactors, Microwave Opt Technol Lett
cells, numerical problems occur, e.g. when transmission or reec-
39 (2003), 81 86.
9. D. Kuylenstierna, A. Vorobiev, P. Linner, and S. Gevorgian, Compos-
tion are very small in magnitude [7]. A variation of this approach
ite right/left handed transmission line phase shifter using ferroelectric is the extraction of effective impedance (z) and refractive index (n)
varactors, IEEE Microwave Wireless Compon Lett 16 (2006), 167 of the MTM cell from scattering matrix, and next the computation
169. of the effective permittivity eff and permeability eff from n, z
10. A. Rousseau, M. Guilloux-Viry, V. Bouquet, A. Perrin, G. Tanne, F. values [8, 9]. Since the mentioned method happen to fail in some

DOI 10.1002/mop MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 2, February 2007 285
cases, some improvements based on the determination of effective
boundaries, forced continuity of the dispersive effective refractive
index, and the elimination of the measurement/simulation noise
inuence on effective impedance of the DNG cell were proposed
in Ref. 10.
The approach presented in this Letter is related to the extraction
from scattering parameters. The main difference relies on the fact,
that the shape of parameterized characteristics for effective per-
mittivity and permeability is assumed a priori, and their parameters
are optimized in order to obtain the best tting to the reference
responses.

2. PARAMETER FITTING OF DISPERSION MODELS

Within the presented method, effective material parameters are
found by tting scattering parameters of the equivalent represen-
tation to the scattering parameters of the reference structure. The
reference structure is a detailed DNG geometry simulated in the
electromagnetic solver, while the effective representation is a slab
of a isotropic, homogeneous material described by dispersive
Drude (electric permittivity) and Lorentz (magnetic permeability)
models. The coefcients of the dispersive models are the param-
eters in the optimization process. The optimization goal is to
minimize the difference between the scattering parameters ob-
tained for the reference structure and the homogeneous structure.
The homogeneous cell should provide the same transmission/
reection behaviour as the SRR/wire based DNG cell.
The simulation procedure for the DNG reference cell (structure
from [2], dimensions given in Fig. 1), is similar to the one used in
Ref. 11. An automeshing algorithm is used in CST Microwave
studio [12] to create the computational grid for the SRR/wire
geometry. Hundred mesh points per medium wavelength are cho-
sen, resulting in 38,400 mesh cells. The excitation pulse has a
Gaussian distribution in time domain that is transformed into 1000
intermediate frequencies from 7 to 12 GHz in the frequency Figure 2 Magnitude (top) and angle (in [deg], bottom) of scattering
domain. The ports are at the x limits of the mesh volume where parameters for SRR/wire reference structure (solid line) and for the opti-
open boundary conditions are used. The structure is excited by the mized structure (dashed line)
rst mode of a waveguide port, with the electric eld polarized in
the y direction and propagating along the x direction. Magnetic
y faces of the mesh volume. The ring and the wire are made of
boundary conditions are applied at the faces along the axis of the
copper and placed on a 0.25 mm thick dielectric slab characterized
rings (z limits) and electric boundary conditions are used at the
by R 3.84 and tan 0.018. The numerical problem is solved
by a time domain solver until the residual accuracy is 50 dB.
Obtained scattering parameters S11ref and S21ref are presented in
Figure 2.
The effective representation of the DNG MTM cell is a homo-
geneous slab with its thickness the same as for the MTM unit cell
(5 mm for the structure from Fig. 1) and modelled as an isotropic
medium with dielectric dispersion described by Drude model and
magnetic dispersion characterized with Lorentz model. The Drude/
Lorentz description of a DNG MTM is a common approach [13],
where the Drude model of eff represents an articial medium
composed of a lattice of wires [14], and the Lorentz model of eff
accounts for the effects in SRRs [1].
The scattering parameters for the homogenized DNG cell are
obtained analytically. The effective permittivity and permeability
models are assumed to be of the Drude and Lorentz form, respec-
tively (the assumed time-dependence notation is exp(jwt), the
values of eff and eff are relative to those in free space):

p2
eff (1)
Figure 1 SRR/wire reference structure (SRR at the front side, wire at the ivc
back side of the PCB board), unit cell dimensions: gap width g 0.5 mm,
wire width g 0.5 mm, lattice constant a 5 mm, outer SRR height w where electric permittivity at the high frequency limit, p radial
3 mm, ring spacing d 0.5 mm, strip width c 0.25 mm plasma frequency, c collision frequency;

286 MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 2, February 2007 DOI 10.1002/mop
TABLE 1 Optimized Drude/Lorentz Parameters for SRR/Wire where scattering parameters are tted at i frequencies in the
Effective Structure frequency range of interest.
The optimized parameters of Drude and Lorentz models for the
Parameter Optimized Value
structure in Figure 1 are given in Table 1. A comparison of
1.62 magnitude and angle of scattering parameters obtained by optimi-
p 2 14.63 zation with the reference results in Figure 2 shows a very good
GHz tting. Corresponding eff and eff characteristics given in Figure
c 30.69 MHz
3 show a DNG behaviour in the frequency range 9.69 10.24 GHz.
s 1.26
On the same gure effective permittivity and permeability ex-
1.12
0 2 9.67 GHz tracted according to the method from [9] are presented. The
1.24 GHz obtained eff and eff values agree very well in the given frequency
range. A small discrepancy for electric permittivity characteristic
occurs at the frequency close to the resonant frequency of the
magnetic permeability 0 (note the small negative value of
s 02 which is a common problem of this method [9]).
eff (2)
02 i 2
3. DISCUSSION

where s/ magnetic permeability at the low/high frequency The DNG MTM are generally anisotropic structures and their
limit, 0 radial resonant frequency, damping frequency. electromagnetic properties depend on the polarization of the ap-
From the assumed material parameters effective impedance and plied eld. The simple isotropic model presented in this article
refractive index are obtained (Z0 is the impedance of free space): allows for the extraction of the diagonal components of permittiv-
ity and permeability tensors corresponding to the orientation of the

Z eff Z0
eff
eff
(3)

n eff eff eff (4)

As the MTM structure is a passive medium, the signs in Eqs. (3)

and (4) are determined by the requirement ReZeff 0 and
n eff 0.
The reection coefcient R at the boundary between free space
and MTM slab of effective impedance Zeff:

Z eff Z0
R (5)
Zeff Z0

Transmission T through the homogenized slab of effective imped-

ance Zeff, refractive index neff, and thickness d gives:

T exp j
n d
c eff (6)

where radial frequency and c velocity of light in free space.

The dependence between reection/transmission and scattering
parameters [6]:

1 T 2 R
S11 (7)
1 R 2T 2

1 R 2T
S21 (8)
1 R 2T 2

The Eqs. (1)(8) relating scattering parameters with effective

MTM permittivity and permeability are implemented in Matlab
[15]. An optimization algorithm [16] searches for the values of eff
and eff subparameters (namely: , p, c, s, , 0, ) providing
the best t between the scattering parameters of the homogenized
Figure 3 Effective magnetic permeability eff j (top) and
(S11, S21) and the reference structure (S11ref, S21ref). The opti-
electric permittivity eff j (bottom) obtained by parameter
mizer goal function takes the form: tting of dispersive models (, solid thick line; , solid thin line)

G i
|S11 S11 ref| i |S21 S21ref| i (9)
and by direct extraction from scattering parameters according to the
method from [9] (dotted line); the dashed vertical lines limit the DNG
frequency band

DOI 10.1002/mop MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 2, February 2007 287
excitation eld. For many MTM structures, however, it is suf- Robust method to retrieve the constitutive effective parameters of
cient description. metamaterials, Phys Rev E 70 (2004), 016608.
The presented approach is limited to non and weak bianisotro- 11. T. Weiland, R. Schuhmann, R.B. Greegor, C.G. Parazzoli, A.M. Vet-
pic structures (for Ey and Hz eld excitation, the SRR/wire con- ter, D.R. Smith, D.C. Vier, and S. Schultz, Ab initio numerical sim-
ulation of left-handed metamaterials: Comparison of calculations and
guration from Figure 1 is an example of such a structure [17]).
experiments, J Appl Phys 90 (2001), 5419 5424.
The simple Drude/Lorentz model description does not take into
12. Computer Simulation Technology (CST), Available at http://www.
account magnetoelectric couplings and for bianisotropic MTM cst.com.
more complicated methods of extraction should be used (e.g. [18]). 13. J.B. Pendry and D.R. Smith, Reversing light with negative refraction,
The a priori assumed Drude/Lorentz type shape of the effective Phys Today 57 (2004), 37.
parameters can be regarded as a limitation of the presented 14. J.B. Pendry, A.J. Holden, W.J. Stewart, and I. Youngs, Extremely low
method. However, due to the strong resonant behaviour of DNG frequency plasmons in metallic mesostructures, Phys Rev Lett 76
lattices, the extracted parameters of MTM structures are very often (1996), 4773 4776.
of such a shape [10, 11, 19]. On the other hand, there is no problem 15. The MathWorks, Available at http://www.mathworks.com.
with the continuity of the tted eff and eff characteristics. The 16. R. Storn and K. Price, Differential evolutionA simple and efcient
heuristic for global optimization over continuous spaces, J Global
imaginary parts of permittivity and permeability are positive (
Optim 11 (1997), 341359.
0 and 0), which is not always the case with traditional
17. R. Marques, F. Medina, and R. Rai-El-Idrissi, Role of bianisotropy in
extraction from scattering parameters [9, 19, 20]. Some numerical negative permeability and left-handed metamaterials, Phys Rev B 65
ambiguities known for the extraction from scattering parameters (2002), 144440.
(e.g. multiple branches of the logarithm function for the extraction 18. X. Chen, W. Bae-Ian, J.A. Kong, and T.M. Grzegorczyk, Retrieval of
of the real part of effective refraction index [8, 10]) are also the effective constitutive parameters of bianisotropic materials, Phys
avoided. Rev E 71 (2005), 046610.
The method can be straightforwardly adopted to the tting of 19. B.-I. Wu, W. Wang, J. Pacheco, X. Chen, T. Grzegorczyk, and J.A.
effective description for single negative structures (e.g. MTM with Kong, A study of using metamaterials as antenna substrate to enhance
negative magnetic permeability), as well as to arbitrary types of gain, Prog Electromagn Res PIERS 51 (2005), 295328.
20. S. OBrien and J.B. Pendry, Magnetic activity at infrared frequencies
dispersive models. The delivered dispersive models can be directly
in structured metallic photonic crystals, J Phys: Condens Matter 14
implemented in commercial solvers [12] for higher order simula-
(2002), 6383 6394.
tions of large structures composed of many MTM unit cells.
2006 Wiley Periodicals, Inc.
4. CONCLUSIONS
By optimization of parameterized Drude and Lorentz models the
effective electric permittivity and magnetic permeability for SRR/
wire structure have been found. The scattering parameters obtained DESIGN OF LOWPASS FILTER USING A
for the equivalent representation agree very well with the reference NOVEL SPLIT-RING RESONATOR
results. The proposed approach helps to avoid some numerical DEFECTED GROUND STRUCTURE
problems occuring in the known extraction methods and can be
applied to the extraction of effective material parameters for single Bian Wu, Bin Li, and Changhong Liang
and double negative metamaterials. School of Electronic Engineering, Xidian University, Xian, Shaanxi
710071, Peoples Republic of China

ACKNOWLEDGMENTS
Received 9 June 2006
This work has been supported by DFG (GRK 1037/1 04).
ABSTRACT: A novel square split-ring resonator (SRR) defected
REFERENCES ground structure (DGS) cell model is presented and analyzed in detail.
1. J.B. Pendry, A.J. Holden, D.J. Robbins, and W.J. Stewart, Magnetism Then an improved SRR DGS cell model loaded with open stubs is pro-
from conductors and enhanced nonlinear phenomena, IEEE Trans posed in order to enhance the out-band suppression. Based on this struc-
Microwave Theory Tech 47 (1999), 20752084. ture, an S-band microstrip lowpass lter with low in-band loss and high
2. R.A. Shelby, D.R. Smith, and S. Schultz, Experimental verication of out-band suppression is designed and fabricated effectively. 2006 Wiley
a negative index of refraction, Science 292 (2001), 7779. Periodicals, Inc. Microwave Opt Technol Lett 49: 288 291, 2007;
3. J.B. Pendry (editor), Focus issue on negative refraction and metama- Published online in Wiley InterScience (www.interscience.wiley.com).
terials, Optics Express 11 (2003), 639 760. DOI 10.1002/mop.22111
4. R.W. Ziolkowski and N. Engheta (editors), Special issue on metama-
terials, IEEE Trans Antennas Propag 51 (2003), 2546 2750. Key words: split-ring resonator (SRR); defected ground structure
5. T. Itoh and A.A. Oliner (editors), Special issue on metamaterials, IEEE (DGS); lowpass lter
Trans Microwave Theory Tech 53 (2005), 14131556.
6. A.M. Nicolson and G.F. Ross, Measurement of the intrinsic properties
of materials by time-domain techniques, IEEE Trans Instrum Meas 19 1. INTRODUCTION
(1970), 377382. In the late 1990s, defected ground structure (DGS) was rstly
7. R.W. Ziolkowski, Design, fabrication and testing of double negative proposed by Korean scholar J.I. Park et al. based on the idea of
metamaterials, IEEE Trans Antennas Propag 51 (2003), 1516 1529. photonic band-gap structure, and has found its application in the
8. D.R. Smith, S. Schultz, P. Markos, and C.M. Soukoulis, Determination
design of planar circuits and lowpass lters [13]. DGS is realized
of effective permittivity and permeability of metamaterials from re-
ection and transmission coefcients, Phys Rev B 65 (2002), 195104.
by etching a defected pattern in the ground plane, which disturbs
9. P. Markos and C.M. Soukoulis, Transmission properties and effective the shield current distribution in the ground plane. This disturbance
electromagnetic parameters of double negative metamaterials, Opt can change the characteristics of a transmission line such as
Express 11 (2003), 649 661. equivalent capacitance and inductance to obtain the slow-wave
10. X. Chen, T.M. Grzegorczyk, W. Bae-Ian, J. Pacheco, and J.A. Kong, effect and band-stop property.

288 MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 2, February 2007 DOI 10.1002/mop