A Low Loss Reflectarray Element based on a Dielectric Resonator Antenna (DRA) with a Parasitic Strip M. H. Jamaluddin(1), R. Gillard(1), R. Sauleau*(1), P. Dumon(2) and L. Le Coq(1) (1) Institut d’Electronique et de Télécommunications de Rennes (IETR), UMR CNRS 6164 (INSA and University of Rennes 1), Rennes, France (2) Centre National d’Etudes Spatiales (CNES), Toulouse Cedex, France E-mail:
[email protected] Introduction The reflectarray antennas [1] combine some of the best features of antenna arrays and reflector antennas. The reflectarray concept is based on the reflection properties of the individual cells. The required phase on the radiating aperture can be obtained by varying the parameters of each unit cell. In the literature, several configurations of reflectarrays have been reported in microstrip technology, e.g. [2-4]. However, at millimeter waves, the conductor loss becomes severe, and the efficiency of the antenna is reduced significantly. In contrast to patch antennas, dielectric resonator antennas (DRAs) exhibit low loss, broad bandwidth, small mutual coupling and higher radiation efficiency, e.g. [5]. Therefore, DRA reflectarrays are attractive in that frequency range. A first prototype based on this approach has been introduced previously in Ka-band [6]. In this work, the required phase shift is obtained by varying the height of the DRA, which requires a complex manufacturing process. In this paper, we introduce a new low-loss DRA reflectarray element with a metallic strip printed on the top of the DRA. The main idea of this topology is to vary the phase of the reflected field by changing the size of the strip and not that of the DRA itself. The main characteristics of the unit cell are first described in Ka-band. Then the proposed concept is validated experimentally in C-band. Experimental results are in good agreement with the numerical predictions. Design of the DRA Reflectarray Element in Ka-Band The geometry of the unit cell is represented in Fig. 1. A metallic strip (Wstrip×Lstrip) is printed on the top of a square DRA (εr=10) lying on an infinite ground plane. This unit cell is illuminated by a linearly-polarized plane wave under normal incidence with incident electric field parallel to the strip axis. The strip printed on the DRA acts as a parasitic element. Varying its length enables one to adjust the resonant frequency of the strip-loaded DRA, and thus the phase of the reflected field at a given frequency of operation.
978-1-4244-2042-1/08/$25.00 ©2008 IEEE
Fig. 1. 3-D view of a strip-loaded DRA (a = 5mm, Hdra = 0.9mm, Ldra = 2.7mm, Wstrip = 0.3mm. Lstrip varies from 0 to 2.7mm).
Numerical Results and Analysis in Ka-Band The phase and amplitude of the reflection coefficient of the reflectarray element are represented in Fig. 2 as a function of frequency for several strip lengths. They have been computed using HFSS software. The DRA alone (without a strip) operates at a frequency slightly higher than 35GHz. When the strip length is varied, the resonant frequency decreases, and it is found that a 320° phase variation is obtained at 30GHz. It is also noteworthy to mention that the insertion loss is reasonably small (less than 0.6dB). To make sure that these attractive characteristics do not originate from the metal strip itself and to confirm that the operating principle of the proposed unit cell is based on the perturbation of the fundamental mode of the DRA, it is valuable to compare the performance of the strip-loaded DRA with a ‘full substrate’ configuration of same dielectric constant and height. This configuration is obtained by extending the DRA size until the walls of the TEM waveguide that is used in simulation (Ldra=a). By doing so, the resonance of the DRA is suppressed and the effect of the strip can be studied individually. Fig. 3 compares the frequency response of the DRA and ‘full substrate’ unit cells. Fig. 3a shows that the phase range achieved with the DRA (320°) is slightly smaller than the one of the ‘full substrate’ cell (340°). However, the insertion loss in ‘full substrate’ structure is very high (5dB) compared to that produced by the DRA cells (Fig. 3b). Experimental Validations in C-band To validate the numerical results obtained at 30GHz, a similar unit cell has been designed to operate at lower frequencies (around 5GHz). To facilitate the manufacturing, the strip is etched on the lower face of a Duroid substrate (Fig. 4). This substrate is then glued onto the DRA. This unit cell will be characterized experimentally using a square metallic waveguide. Therefore, in the numerical model, we assume that the four vertical walls surrounding the unit cell are perfect electrical conductors (PEC), in contrast to the previous section where the DRA was excited by a TEM wave. A set of five DRA samples has been fabricated and measured. Comparisons between measurements and simulations are shown in Fig. 5 for the phase and amplitude of the reflection coefficient. A phase variation of nearly 360° is found for each measured sample (Fig. 5b), as predicted numerically (Fig. 5a). The 100-MHz frequency shift observed between experimental and numerical results is attributed to fabrication uncertainties and the adhesive glue used to bond the DRA and the Duroid substrate. Fig. 5d also shows that the measured loss is small (less than 0.8dB) and is in good agreement with the simulations (Fig. 5c). The amplitude response of the unit cell is represented in Fig. 6 versus strip length (Lstrip). The agreement between measured and computed data is considered to be acceptable if we keep in mind that measuring the insertion loss of lowloss reflectarray elements is a challenging task. To further confirm the advantages and low-loss features of DRA reflectarray elements previously highlighted in Ka-band, a set of five ‘full substrate’ unit cells has also been designed and fabricated to operate around 5GHz. Measurements have shown that a 340° phase range can be obtained. Nevertheless, the insertion loss (Fig. 7) is much higher than for DRAs (Fig. 6). Conclusion A strip-loaded DRA has been proposed as a promising reflectarray element in Ka-band. Varying the length of the strip enables one to adjust the phase of each DRA cell. A phase variation of 320° and low loss (less than 0.6dB) are achieved at 30GHz. This concept has
been checked experimentally in C-Band. It was also demonstrated that the DRA structure exhibits much better performance compared to the so called ‘full substrate’ configuration. Acknowledgment The authors would like to thank the CNES (Centre National d’Etudes Spatiales) for its financial support. References D.G. Berry and R. G. Malech, “The reflectarray antenna”, IEEE Trans. Antennas Propagat., vol. 11, no. 6, pp. 645-651, Nov. 1963. J. Huang, “Microstrip reflectarray”, Antennas and Propagat. Soc. Int. Symp., AP-S Digest, pp. 612-615, June 1991. J.A. Encinar and J.A. Zornoza, “Three-layer printed reflectarrays for contoured beam space applications”, IEEE Trans. Antennas Propagat., vol. 52, no. 5, pp. 11381148, May 2004. E. Carrasco, M. Barba, and J.A. Encinar, “Reflectarray element based on aperturecoupled patches with slots and lines of variable length”, IEEE Trans. Antennas Propagat.,, vol. 55, no. 3, Part 2, pp. 820-825, Mar. 2007. K. M. Luk and K. W. Leung, “Dielectric resonator antennas. Overview of the dielectric resonator antenna”, Research Studies Press, pp. 1-2, 2002. A. Petosa, A. Ittipiboon, Y. M. M. Antar, D. Roscoe, and M. Cuhaci, “A Ka-Band dielectric resonator antenna reflectarray”, 30th European Microwave Conference, pp.1-4, Oct. 2000.
[2] [3] [4] [5]
Reflection phase, degree
[6]
180
Lstrip=0 Lstrip=400um Lstrip=800um
150
Lstrip=1,2mm Lstrip=1,6mm
120
Lstrip=2mm Lstrip=2,7mm
90
0
-0,1
Reflection amplitude, dB
[1]
-0,2
60 30
-0,3
0 -30
-0,4
-60 -90
Lstrip=0 Lstrip=0.4mm Lstrip=0.8mm Lstrip=1.2mm
-0,5
-120
Lstrip=1.6mm
-150 -180
Lstrip=2mm Lstrip=2.7mm
-0,6
25
26
27
28
29
30
31
Frequency, GHz
32
33
34
25
35
26
27
28
29
30
31
32
Frequency, GHz
33
34
35
(b)
(a)
Fig. 2. Simulation results (HFSS). Phase (a) and amplitude (b) of the reflection coefficient. 180
0
Full substrate
120
-0,5
DRA
R eflec tion am plitude , dB
Reflection phase,degree
60 0
-1
-1,5
-60
-2
-2,5
-120 -180
-3
-3,5
-240 -300
-4
-360
-4,5
-420
-5
0
0,3
0,6
0,9
1,2
1,5
Lstrip, mm
(a)
1,8
2,1
2,4
2,7
Full substrate DRA 0
0,3
0,6
0,9
1,2
1,5
Lstrip, mm
1,8
2,1
(b)
Fig. 3. Main characteristics of the strip-loaded DRA and ‘full substrate’ unit cell. (a) Phase response, (b) Insertion loss.
2,4
2,7
Fig.4. 3-D view of the strip-loaded DRA in C-band (a = 35mm, Hdra = 14mm, Ldra = 4.9mm, Hduroid = 0.8mm, Wstrip = 1.0mm. Lstrip varies from 0 to 14mm).
Reflection phase, degree
150 120 90 60 30 0 -30 -60 -90
-120
0 -0,1
Reflection amplitude, dB
Lstrip=6mm Lstrip=7,5mm Lstrip=8,25mm Lstrip=9mm Lstrip=11mm
180
-0,2 -0,3 -0,4 -0,5 -0,6 -0,7 -0,8
-150 -180 4,7
4,8
4,9
5
5,1
5,2
5,3
5,4
Freq, GHz
5,5
5,6
5,7
5,8
5,9
-0,9
6
4,7
4,8
4,9
5
5,1
5,2
(a)
5,4
5,5
5,6
5,7
5,8
5,9
6
5,4
5,5
5,6
5,7
5,8
5,9
6
(c) 0
180 150
-0,1
Reflection amplitude, dB
120
Reflection phase, degree
5,3
Freq, GHz
-0,2
90 60
-0,3
30
-0,4
0
-0,5
-30 -60
-0,6
-90
-0,7
-120
-0,8
-150 -180 4,7
4,8
4,9
5
5,1
5,2
5,3
5,4
Freq, GHz
5,5
5,6
5,7
5,8
5,9
-0,9
6
4,7
4,8
4,9
5
5,1
5,2
5,3
Freq, GHz
(b) (d) Fig. 5. Frequency response of the DRA unit cell represented in Fig. 4. (a) Computed phase. (b) Measured phase. (c) Computed amplitude. (d) Measured amplitude. 0
0
Reflection amplitude, dB
Reflection amplitude, dB
-0,1 -0,2 -0,3 -0,4 -0,5 -0,6
Simulation
-0,7
5
6
7
8
9
Lstrip, mm
10
11
-2 -3 -4 Simulation
-5
Measurement
Measurement
-0,8
-1
12
Fig. 6. Insertion loss of the DRA unit cells (5.1GHz).
-6 3
4
5
6
7
8
Lstrip, mm
9
10
11
12
Fig. 7. Insertion loss of the ‘full substrate’ unit cells (5.1GHz).
