3672 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 57, NO.
11, NOVEMBER 2009
Fig. 7. Amplitude and phase of the elevation patterns for the drooped annular
antenna. The measured patterns are shown at the measured resonant frequency
of 1.5703 GHz and the simulated patterns at 1.575 GHz.
particularly by observing the effects of shape and orientation of the
ground plane upon the amplitude and phase patterns. Results showed
that the beam width has been slightly increased for the folded antenna
without altering the bore-sight gain accompanied with improved phase
center stability. A novel drooped antenna operating in the TM30 mode
has been presented using a square annular element. The drooped an-
nular element permits significantly greater control over the radiation
pattern as a result of the interference between the four radiating edges.
Variations of the angle and position of the bend revealed a certain com-
bination, giving complete upper hemispherical coverage with the pat-
(microstrip feed). A mutual coupling of
tern ripple reduced to 2 dB. Notably, in general, there is a tradeoff in
achieving coverage over the entire upper hemisphere and low cross po-
larization. If a broad beam width is of precedence, it may be necessary
to operate the drooped antennas under less than optimal conditions in
regard to cross polarization performance.
REFERENCES
[1] G. Lachapelle, M. Casey, R. M. Eaton, A. Kleusberg, J. Tranquilla, and
D. Wells, “GPS marine kinematic positioning accuracy and reliability,”
Canadian Surveyor, vol. 41, no. 2, pp. 143–172, Oct. 1987.
[2] J. M. Tranquilla, The Experimental Study of Global Positioning Satel-
lite Antenna Backplane Configurations NASA Jet Propulsion Lab., Ra-
diating Systems Research Lab., Univ. New Brunswick, Fredericton,
NB, Canada, Tech. Rep., 1988, Contract 957959.
[3] J. M. Tranquilla and B. G. Colpitts, “GPS antenna design character-
istics for high precision applications,” presented at the ASCE Conf.
GPS-88 Eng. Applicat. of GPS Satellite Surveying Technol., Nashville,
TN, May 11–14, 1988.
[4] J. M. Tanquilla and B. G. Colpitts, “Development of a class of antennas
for space-based NAVSTAR GPS applications,” in Proc. 6th Int. Conf.
on Antennas and Propag. (ICAP 89), Coventry, U.K., Apr. 4–7, 1989,
pp. 65–69.
[5] J. M. Tranquilla and S. R. Best, “A study of the quadrifilar helix an-
tenna for global positioning systems (GPS) applications,” IEEE Trans.
Antennas Propag., vol. 38, pp. 1545–1550, Oct. 1990.
[6] K. G. Clark, “The Finite-difference time-domain technique applied to
the drooped microstrip,” Ph.D. dissertation, Dept. Elect. Eng., Univ.
New Brunswick, Fredericton, NB, Canada, Jul. 1996.
[7] W. Feller, “Three Dimensional Microstrip Patch Antenna,” U.S. Patent
5 200 756, Apr. 1993.
[8] N. Fayyaz, N. Hojjat, and S. Safavi-Naeini, “Rectangular microstrip
antenna with a finite horn-shaped ground plane,” in Proc. IEEE An-
tennas and Propag. Society Int. Symp., Jul. 13–18, 1997, vol. 2, pp.
916–919.
[9] H. Nakano, S. Shimada, J. Yamauchi, and M. Miyata, “A circularly
polarized patch antenna enclosed by a folded conducting wall,” in IEEE
Topical Conf. on Wireless Commun. Technol., Oct. 15–17, 2003, pp.
134–135.
0018-926X/$26.00 © 2009 IEEE
[10] C. L. Tang, J. Y. Chiou, and K. L. Wong, “A broadband probe fed patch
antenna with a bent ground plane,” in Proc. Microw. Conf., 2000, pp.
1356–1359.
[11] C. L. Tang, J. Y. Chiou, and K. L. Wang, “Beamwidth enhancement of
a circularly polarized microstrip antenna mounted on a three-dimen-
sional ground structure,” Microw. Opt. Technol. Lett., vol. 32, no. 1,
pp. 149–153, 2002.
[12] C. W. Su, S. K. Huang, and C. H. Lee, “CP microstrip antenna with
wide beamwidth for GPS band application,” Electron. Lett., vol. 43,
no. 20, Sep. 27, 2007.
[13] T. G. Jurgens, A. Taflove, K. R. Umashankar, and T. G. Moore, “Finite-
difference time-domain modeling of curved surfaces,” IEEE Trans. An-
tennas Propag, vol. AP-40, pp. 357–366, Apr. 1992.
[14] J. M. Tranquilla and S. R. Best, “Phase center considerations for the
monopole antenna,” IEEE Trans. Antennas Propag, vol. AP-34, pp.
741–744, May 1986.
High-Gain Yagi-Uda Antennas for Millimeter-Wave
Switched-Beam Systems
Ramadan A. Alhalabi and Gabriel M. Rebeiz
Abstract—A high-efficiency microstrip-fed Yagi-Uda antenna has been
developed for millimeter-wave applications. The antenna is built on both
sides of a Teflon substrate ( = 2 2) which results in an integrated
Balun for the feed dipole. A 7-element design results in a measured gain
of 9–11 dB at 22–26 GHz with a cross-polarization level of 16 dB.
The antenna is matched to 50
20 dB is measured between two Yagi-Uda antennas with a center-to
center spacing of 8.75 mm (0 7 at 24 GHz), and a two-element array re-
sults in a measured gain of 11.5–13 dB at 22–25 GHz. The planar Yagi-Uda
antenna results in high radiation efficiency ( 90%) and is suitable for
mm-wave radars and high data-rate communication systems.
Index Terms—Automotive radars, endfire antennas, millimeter-wave
antennas, millimeter-wave communication systems, planar antennas,
Yagi-Uda antenna.
I. INTRODUCTION
Planar Yagi-Uda antennas are very attractive for many microwave
and millimeter-wave applications due to their high gain, low cost, high
radiation efficiency and ease of fabrication. The Yagi-Uda antenna is
one of the most popular endfire antennas which can be designed to
achieve a medium gain with relatively low cross-polarization levels.
Previously, Kaneda et al. presented a microstrip-fed Quasi-Yagi
antenna at X-band with a gain of 3–5 dB and a cross-pol. level of
< 0 15 dB [1]. Grajek et al. showed a Yagi-Uda antenna with a
directivity of 9.3 dB at 24 GHz [2]. These antennas utilize planar
microstrip-to-coplanar stripline (CPS) transition which is based on a
Manuscript received July 28, 2008; revised March 12, 2009. First published
July 07, 2009; current version published November 04, 2009. This work was
supported in part by Intel Corporation and in part by the UC-Discovery Program.
The authors are with the Electrical and Computer Engineering Department,
University of California, San Diego, CA 92122 USA (e-mail: ralhalabi@gmail.
com; rebeiz@ece.ucsd.edu).
Color versions of one or more of the figures in this communication are avail-
able online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TAP.2009.2026666
IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 57, NO. 11, NOVEMBER 2009 3673

TABLE I
SIMULATED CROSS-POL. LEVEL VERSUS SUBSTRATE THICKNESS

Fig. 1. Microstrip-fed Yagi-Uda antenna geometry: L = 5:4; Ld= 4:1; L1 =

1:5; L2 = 2:6; Ls = 20; W = 0:4; W1 = 0:4; W2 = 1:0; W3 = 1:2; d =
2:4; dr = 2:7 and ground plane width = 29 (all dimensions are in mm).

half-wave delay line to achieve the 180 phase shift for the balanced
dipole feed, and the frequency dependence of the balun limits the
antenna performance versus frequency. A Yagi-Uda antenna with
one director, a truncated ground plane acting as a reflector and with
a simplified feeding structure was presented by Zheng [3] where the
balun between the microstrip feed and the balanced dipole feed was
built using the top and bottom-sides of the substrate. Lee and Chung
presented a 38 GHz microstrip-fed Yagi-Uda antenna which uses 6 di-
rectors and the microstrip ground plane as a reflector to achieve a gain
of 9.5 dB [4]. DeJean and Tentzeris presented a high gain microstrip
Yagi array with high front to back ratio [5]. Woo et al. presented
a microstrip-fed Yagi-Uda antenna with a new microstrip-to-CPS
transition [6]. This new transition performs the required field and
impedance match between the microstrip line and the CPS feed line
using via holes. Using one director and one reflector, this antenna
showed a gain of 5.2–5.8 dB with a bandwidth of 29.1% from 30 to
40 GHz. H. K. Kan et al. showed a CPW-fed Quasi-Yagi antenna with
a 44% 10 dB impedance matching bandwidth at X-band [7]. Recently,
Hsu et al. showed a 60 GHz CPW-fed on-chip Yagi-Uda antenna with
a gain of 010 dB [8].
This communication presents a seven-element microstrip-fed
Yagi-Uda antenna with high gain (>10 dB), wide bandwidth
and is connected to a microstrip line with
(22–26 GHz) and low cross-polarization levels (018 dB). The an-
). The balun between the microstrip feed
tenna utilizes five directors, and the truncated ground plane acts as
a reflector to maximize the antenna gain. Two-element arrays with
a center-to-center spacing of 8.75 mm (0:70 at 24 GHz) are also
,
presented.
Fig. 2. (a) Fabricated microstrip-fed Yagi-Uda antenna, ground plane width is
29 mm and microstip line length is 20 mm, (b) measured and simulated S .
II. ANTENNA DESIGN AND MEASUREMENTS
A. Antenna Design
The microstrip-fed Yagi-Uda antenna was built on a Rogers
RT/Duroid 5880 substrate ("r = 2:2) with a thickness of 15 mils
(0.381 mm) and utilizes five directors (Fig. 1). The directors are
printed on the top side of the substrate with a director-to-director
spacing of d = 2:4 mm. The initial dimensions of the antenna were
obtained from tables for maximum directivity in air [9] and then
scaled to compensate for the duroid substrate ("e = 1:41) [2]. The
microstrip truncated ground plane is located at dr = 2:7 mm from the
driving dipole and acts as a reflector. The antenna is designed to have
an input impedance of 50
Ws = 1:2 mm (Zo = 50
and the balanced dipole feed is built using the top and bottom-sides
of the Teflon substrate. The driving dipole is fed by a parallel-plate
transmission line of width Wf = 0:4 mm and impedance Zf = 130
and this transmission line becomes a microstrip feed line of length
3674 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 57, NO. 11, NOVEMBER 2009

Fig. 3. Radiation patterns for the microstrip-fed Yagi-Uda antenna:—measured Co-pol, - - - - - simulated Co-pol., -:-:-:- measured Cross-pol.

Fig. 4. Measured and simulated gain of the microstrip-fed Yagi-Uda antenna.

L1 = 1:5 mm and impedance Z1 = 93

followed by another mi-
crostrip section of length L2 = 2:6 mm and impedance Z2 = 56
to
arrive to the 50
microstrip feed.
The driver dipole is built on both sides of the substrate and allows a
wideband balun feed from the single-ended microstrip line to the differ-
ential dipole. However, it also results in an increase in the cross-polar- Fig. 5. Fabricated 2-element arrays of microstrip-fed Yagi-Uda antennas (top
ization level (Table I). The cross-polarization simulations were done side), d = 8:75 mm: (a) with Wilkinson coupler, (b) with matched T-junction.
using HFSS and the antenna dimensions were modified so that the
driving dipole of the Yagi-Uda antenna resonates at the same frequency
for each case. It is clear that a substrate thickness of 15 mils or less The radiation patterns were measured in the receive mode using a
should be chosen for low cross-polarization levels. zero-bias Schottky diode detector (Krytar model 303B) and a lock-in
amplifier (Stanford Research Systems, SR830 DSP Lock-in Amplifier).
B. Impedance and Pattern Measurements The diode detector was connected to the microstrip line using a high
The input impedance of the microstrip-fed Yagi-Uda antenna was performance Southwest microwave 2.92 mm connector [Fig. 2(a)]. The
measured using a microstrip to coaxial line transition [Fig. 2(a)], and RF signal is amplitude modulated with a 1 kHz sine-wave signal and the
shows a good agreement with HFSS simulations with measured S11 < rectified 1 kHz is measured using the lock-in amplifier. The measured
09 dB (simulated S11 < 010 dB) from 22.1 to 25.5 GHz [Fig. 2(b)]. patterns agree well with HFSS simulations and show a front to back
The microstrip to coaxial transition was not included in the HFSS sim- ratio of 20 dB and cross-polarization level of  0 20 dB at 24 GHz
ulations. We believe that the slight difference between the measured (Fig. 3). The patterns are quite symmetric at 24 GHz with an E and
and simulated S11 is due to the effect of this transition. H-plane 3-dB beamwidths of 44 and 50 , respectively.
IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 57, NO. 11, NOVEMBER 2009
Fig. 6. (a) Top layer of the fabricated two microstrip-fed Yagi-Uda antennas
with center to center spacing of d = 8:75 mm (0:7 at 24 GHz), (b) measured
and simulated mutual coupling (S ).
C. Gain Measurements
The absolute gain of the microstrip-fed Yagi-Uda antenna was
measured using a standard gain horn antenna. A 2.92 mm Southwest
microwave connector is used to minimize the reflection at the con-
nector. The received power is measured using a calibrated Agilent
Power Meter (E4417), and the same power meter is used to measure
the transmit power. The Yagi-Uda antenna gain is then obtained
using the Friis transmission formula. The loss of the microstrip line
between the antenna and the Southwest connector is 0.44 dB and is
taken out from the gain measurements. The measurements show a
gain of >10 dB from 22–25 GHz and 10.4 dB 60:5 dB at 24 GHz
(Fig. 4). HFSS reported a gain of 10.9 dB with directivity of 11.2 dB
at 24 GHz and the difference is mostly due to the impedance mismatch
loss. This results in a measured radiation efficiency of 90% within
the 60:5 dB measurement error. The Yagi-Uda antenna gain drops to
1 dB at 26 GHz due to non-optimal phasing of the director elements.
This design is therefore optimal for 21–25 GHz applications with a
gain >8 dB, and results in a bandwidth of 17.5%.
D. Two-Element Array of Microstrip-Fed Yagi-Uda Antennas
Two-element arrays of Yagi-Uda antennas with a center to center
spacing of 8.75 mm (0:70 at 24 GHz) were also built and measured.
The first design utilizes a Wilkinson power combiner to combine the
signals [Fig. 5(a)], while the second design uses a matched T-junction
[Fig. 5(b)]. The mutual coupling between two Yagi-Uda antennas, with
a center to center spacing of 8.75 mm (0:70 at 24 GHz), was mea-
sured using the layout shown in Fig. 6(a). The measurement agrees
3675
Fig. 7. (a) Measured radiation patterns of the 2-element arrays: ------------ with
Wilkinson, - - - - - with T-junction, -:-:-:- simulation, (b) measured S ref-
erenced to plane 1.
Fig. 8. Measured and simulated gain of the two element array of the mi-
crostrip-fed Yagi-Uda antennas (with Wilkinson coupler) at ref. plane 2.
with HFSS simulations and shows a mutual coupling of < 0 16 dB
from 20 to 26 GHz [Fig. 6(b)].
The measured radiation patterns of the two-element arrays show
good agreement with simulations as shown in Fig. 7(a). The measured
24 GHz E-plane patterns has a 3-dB beamwidth of 30 , while the mea-
sured H-plane pattern is similar to the single element pattern and has
a 3-dB beamwidth of 46 . The two-element array has nearly the same
E-plane pattern as a Yagi-Uda antenna with 10 directors. The patterns
were also measured at 22–25 GHz (not shown) and are very similar to
the 24 GHz patterns. The measured S11 of the two-element arrays is
< 0 8 dB from 22.0 to 26.0 GHz as shown in Fig. 7(b).
The two-element array gain at ref. plane 2 in Fig. 5 was measured
at 20, 22, 24 and 25 GHz, where the loss between ref. planes 1 and 2
was estimated to be 0.4 dB. The measured gain is >10 dB from 20 to
25 GHz with a peak value of 12.5 dB and agrees well with simulations
(Fig. 8).
3676 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 57, NO. 11, NOVEMBER 2009
III. CONCLUSION
This communication presented a microstrip-fed millimeter-wave
Yagi-Uda antenna with applications as single element radiator or for
switched-beam systems with medium gain (9–13 dB). The planar
Yagi-Uda antenna can be arrayed for additional gain (+3 dB) and
with low mutual coupling between the elements. The antenna results
in relatively wideband operation (22–26 GHz), low cross-polarization
levels, and high radiation efficiency. This antenna can be scaled to 60,
77, or 94 GHz for automotive radars and high data-rate communication
systems.
REFERENCES
[1] N. Kaneda, W. R. Deal, Y. Qian, R. Waterhouse, and T. Itoh, “A broad-
band planar Quasi-Yagi antenna,” IEEE Trans. Antennas Propag., vol.
50, no. 8, pp. 1158–1160, Aug. 2002.
[2] P. R. Grajek, B. Schoenlinner, and G. M. Rebeiz, “A 24-GHz high-gain
Yagi-Uda antenna array,” IEEE Trans. Antennas Propag., vol. 52, pp.
1257–1261, May 2004.
[3] G. Zheng, A. A. Kishk, A. B. Yakovlev, and A. W. Glisson, “Simplified
feed for a modified printed Yagi antenna,” Electron. Lett., vol. 40, no.
8, pp. 464–465, Apr. 15, 2004.
[4] Y. Lee and S. Chung, “Design of a 38-GHz printed Yagi antenna with
multiple directors,” in Proc. IEEE Antennas Propag. Symp., Jul. 2001,
vol. 3, pp. 606–609.
[5] G. R. DeJean and M. M. Tentzeris, “A new high-gain microstrip Yagi
array antenna with a high front-to-back (F/B) ratio for WLAN and mil-
limeter-wave applications,” IEEE Trans. Antennas Propag., vol. 55, pp.
298–304, Feb. 2007.
[6] D. Woo, Y. Kim, K. Kim, and Y. Cho, “A simplified design of Quasi-
Yagi antennas using the new microstrip-to-CPS transitions,” in Proc.
IEEE Antennas Propag. Symp., June 2007, pp. 781–784.
[7] H. K. Kan, R. B. Waterhouse, A. M. Abbosh, and M. E. Bialkowski,
“Simple broadband planar CPW-fed Quasi-Yagi antenna,” IEEE An-
tennas Wireless Propag. Lett., vol. 6, pp. 18–20, 2007.
[8] S. Hsu, K. Wei, C. Hsu, and R. Chuang, “A 60-GHz millimeter-wave
CPW-Fed Yagi antenna fabricated by using 0.18-m CMOS tech-
nology,” IEEE Electron. Device Lett., vol. 29, pp. 625–627, Jun. 2008.
[9] W. L. Stutzman and G. A. Thiele, Antenna Theory and Design, 2nd
ed. New York: Wiley, 1998.
0018-926X/$26.00 © 2009 IEEE
On the Transmission and Propagation of Low Attenuation
Rate Electromagnetic Pulses in Debye Media
Reza Safian, Costas D. Sarris, and Mohammad Mojahedi
Abstract—In a dispersive medium, the appearance of the steady-state
part of the signal is preceded by oscillations known as precursors. These
early oscillations are the product of the interrelated effects of phase dis-
persion and frequency dependent attenuation. Inside water, the attenua-
tion rate of the Brillouin precursor is sub-exponential, following the inverse
square-root of the distance traveled. Based on that, a near-optimal pulse
that could achieve this attenuation rate, and, hence, would lend itself to
underwater detection and communication applications, was recently pro-
posed. The “optimality” of this pulse is shown to be related to the temporal
support of the pulse and its spectral characteristics, rather than its shape.
A family of alternative pulses is found to have the low attenuation feature of
the “optimal” pulse, as they eventually evolve into the Brillouin precursor
itself shortly after they enter water. In addition, this work considers the
practical case when such a pulse would be generated in air, would impinge
onto an air-water interface and then propagate inside water. It is shown
how the presence of the interface affects the attenuation rate of the pulse
inside water and a simple way to recover its low attenuation rate is sug-
gested. The finite-difference time-domain technique is employed in all the
simulations.
Index Terms—Dispersive media, finite-difference time-domain (FDTD)
methods, wave propagation.
I. INTRODUCTION
The propagation of wideband electromagnetic pulses in a causal,
temporally dispersive dielectric has been studied extensively [1], [2].
Significant contributions in this area have been recently made by Ough-
stun et al., through the investigation of pulse propagation in several
temporally dispersive media [3]. In such media, the phase dispersion
and frequency-dependent attenuation of a wideband pulse excitation
can lead to the evolution of precursor fields, which precede the main
part of the pulse.
For a Debye-type dielectric, the electric field excited by a modulated
pulse with a temporal support T where, T 1=fc (fc is the modu-
lation frequency) evolves into the so-called Brillouin precursor, as the
pulse propagates inside the medium [3]. Moreover, the peak amplitude
of the Brillouin precursor decays as the inverse square root of the prop-
agation distance, as opposed to the exponential decay of the main part
of the pulse.
This property of the Brillouin precursor was harnessed to design a
pulse excitation with low attenuation rate in an infinite Debye medium
[4]. Such an excitation consists of two mutually delayed and opposite in
sign Brillouin precursors. The theoretical investigation of [4], limited
to the case of an infinite Debye medium though, suggested that this
“double Brillouin” pulse could indeed achieve an attenuation rate as the
inverse square of the propagation distance for medium parameters that
corresponded to those of the triply distilled water. This result renders
the double Brillouin pulse a good candidate for applications ranging
from communications to detection in lossy dispersive media that can
be described by the Debye model.
Manuscript received July 09, 2007; revised November 18, 2008. First pub-
lished July 28, 2009; current version published November 04, 2009.
The authors are with the Department of Electrical and Computer Engineering,
University of Toronto, Toronto, ON M5S 3G4 Canada (e-mail: rsafian@waves.
utoronto.ca).
Color versions of one or more of the figures in this communication are avail-
able online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TAP.2009.2028678