The use of uniplanar double-spiral resonant cells embedded in the microstrip feed line can create a bandstop response that effectively suppresses unwanted harmonic radiation in patch antennas.

**Ke Lu, Guang-Ming Wang, Zhong-Wu Yu, and Hai-Yang Xu**

PATCH ANTENNAS are typically light in weight, low in cost, and widely used in communications. On the downside, they suffer from excess harmonic radiation. Fortunately, the authors have developed a new method for suppressing harmonics from patch antennas by means of uniplanar double-spiral resonant cells (UDSRCs). By applying numerical simulation, it can be shown that UDSRCs embedded in a patch antenna's microstrip lines exhibits bandstop filtering properties.

To demonstrate the effectiveness of this harmonic-suppression approach, three cascaded UDSRCs were used to form a bandstop filter on the feedline of a patch antenna. Measured results indicate effective suppression through the sixth harmonic, while the reflection coefficient and radiation pattern at the dominant resonant frequency are not degraded. This improvement in performance comes without increasing the feed line's transversal dimension, a distinct advantage over other patch-antenna harmonic-suppression approaches.

Many techniques have been applied for spurious and harmonic signal reduction in patch antennas. In one method, two-dimensional photonic-bandgap (PBG) patterns in the form of lattice circles were etched into the ground plane beneath the patch and the feedline. The result was suppression of the second and third harmonics but with a penalty of excessive backward radiation.^{1}

Another approach proposed one dumbbell-shaped defected ground structure (DGS) under the feed line and one uniplanar compact microstrip resonant cell (CRMC) embedded in the feed line to suppress second harmonics.^{2} The DGS can introduce backward radiation and result in a squinted antenna beam, although there are no reported problems with the uniplanar CRMC. Another approach proposed placing four pairs of split-ring resonators near both sides of the feed line with limited distance, so as to eliminate the first and second harmonics.^{3}

Additionally, a method involving a lowpass filter composed of a circular head dumbbell shaped DGS structures and a circular head shunt open microstrip stub on the feed line was able to suppress to fourth harmonics effectively.^{4} And second and third harmonics were eliminated for a microstrip patch antenna with proximity coupled feeding line by adjusting the length of feed line and introducing one resonant cell.^{5}

The use of UDSRCs was initially proposed in ref. 6. They were later used to design left-handed transmission lines^{7,8} and to form a phase-shift unit.^{9} In the current work, the UDSRCs are embedded in microstrip lines to form bandstop filters and suppress the harmonic radiation of a patch antenna. The stopband of the proposed structure is adjusted by means of two controllable transmission zeros, providing for great design flexibility.

Based on the harmonics of the reference antenna, a bandstop filter composed of three cascaded UDSRCs and embedded in the feedline was proposed as a means of suppressing the patch antenna's harmonic radiation. The measured results will show that this approach is effective through the sixth harmonic, without degrading performance at the resonant frequency. By not using a DGS approach, the radiation pattern at the dominant frequency was unaffected, and the feed network remained compact.

**Figure 1** shows the structure of a UDSRC formed of two single rectangular spirals connected in series. The proposed bandstop filter is formed by cascading three of these UDSRC structures. A printed-circuit-board (PCB) substrate with relative dielectric constant of 2.2 and thickness of 1.5 mm is used for both simulation and fabrication. The following geometric parameters were fixed throughout: W_{1} = 4.7 mm (the width of 50-Ω microstrip line), W_{3} = 0.3 mm, and W4 = 0.6 mm. Length L_{1} and width W_{2} are the primary geometrical parameters for making adjustments. Through numerical simulation, it is found that an UDSRC embedded in a microstrip line exhibits a broad stopband with two transmission zeros. To get further insight into the influence of varied geometrical parameters, parametric analyses of L_{1} and W_{2}, respectively, were implemented. With the help of rules derived from the analyses, guidelines were developed to effectively and directly control the UDSRC bandstop property in microstrip.

As **Fig. 2** shows, an increase in L_{1} shifts the two transmission zeros downward while the rejection level remains unchanged. As **Fig. 3** shows, the transmission zeros in the high-frequency band (denoted as 1) change insignificantly with W_{2} while the transmission zeros in the low-frequency band (denoted as 2) shift upward with the increase of W_{2}. When the separation of transmission zeros becomes wider, the -10 dB stopband bandwidth becomes wider at the cost of deteriorated rejection level. This analysis provides the needed guidelines for optimizing the performance of the microstrip UDSRCs.

To operate at 1.1 GHz, a square patch with dimensions of 91.2 x 91.2 mm serves as the reference antenna. A 50-Ω microstrip feed line is used for antenna excitation. An inset cut with 32-mm length and 1-mm width was also used for impedance matching. **Figure 4** shows the harmonic distribution from both simulated and measured results.

The reference antenna had a measured dominant resonant frequency of 1.08 GHz, slightly shifted from the simulation frequency, and with a 10-dB bandwidth of about 20 MHz. Measured reflection coefficients of second harmonics at 1.99 GHz, third harmonics at 3.02 GHz, and fourth harmonics at 4.07 GHz were -9.29 dB, -11.59 dB, and -7.07 dB, respectively. The fifth harmonics at 4.28 GHz and sixth harmonics at 4.63 GHz (near the fourth harmonics) were about -10.56 dB and -16.48 dB, respectively. Slight discrepancies between the simulated and measured results were thought to be due to fabrication errors. These discrepancies, which are common, can present some difficulties in suppressing harmonics over a narrow stopband when fabricating a prototype, such as the structures proposed in ref. 3.

The USDRC approach, on the other band, with its comparatively broad stopband, can provide effective suppression of harmonics even with these discrepancies. According to the harmonic distribution of the reference, a bandstop filter can be designed by cascading three different UDSRCs and connecting them with 1-mmlong microstrip lines (**Fig. 1**). In the initial design, unit 1 is responsible for second harmonics and unit 2 is responsible for third harmonics. Unit 3 is for eliminating fourth, fifth, and sixth harmonics. By applying the developed guidelines and the use of a full-wave software simulator, the frequency response of each unit can be optimized by adjusting L1 and W2. The dimensions for the three units, following optimization, are shown in the table, while **Fig. 5** shows the simulated and measured S-parameters for the proposed bandstop filter. **Figure 6** shows the bandstop filter integrated with the feed line for the reference antenna, with simulated and measured reflection coefficients.

Suppression was effective through the sixth harmonic without affecting the reflection coefficients at the dominant resonant frequency. Compared with the reference antenna, the measured reflection coefficients of the proposed antenna at the second, third, and fourth harmonic frequencies were suppressed by 8.84, 11.42, and 6.58 dB, respectively. The fifth and sixth harmonics were suppressed by 10.11 and 14.79 dB, respectively. The dominant resonant frequency was found to be about 1.06 GHz, a slight shift from the 1.08 GHz of the reference antenna. At the same time, the bandwidth and the reflection coefficient roughly remain unchanged. Compared with the reference antenna, there is no increase in terms of the transversal dimension of the feed line because the proposed bandstop filter is completely embedded in the feed line. The total length of the three units is only about 46 mm, less than one-sixth the free-space wavelength at 1.06 GHz. The reflection and transmission coefficients of the prototypes were measured with a model ME7808A microwave vector network analyzer (VNA) from Anritsu.

**Figure 7** presents the simulated and measured E-plane radiation patterns of the reference antenna and the proposed antenna after harmonic suppression at their dominant resonant frequency. The radiation patterns were measured inside an anechoic chamber with a model AV3635 antenna measurement system from Siwi Electronic. The results show that these two antennas have almost identical E-plane radiation patterns at their dominant resonant frequency. The H-plane radiation patterns, while not shown here, are also similar. The discrepancies between the experimental and simulated results can be attributed to fabrication tolerances.

After adding UDSCRs for harmonic suppression, no real changes were found in the radiation patterns compared to the reference design. This compares favorably to the approaches of refs. 1 and 2, in which the back radiation due to the presence of DGSs significantly deteriorated the original radiation patterns. Such deterioration does not occur when using the UDSCRs for harmonic suppression.

In summary, the bandstop properties of UDSRCs can be effectively applied to increase the suppression of harmonics in patch antennas. Full-wave simulations have shown that UDSCRs can be embedded in the antenna's microstrip feed line to achieve an effective stopband response, controllable by modifying the primary geometric parameters of the UDSCRs. The approach was demonstrated by cascading three UDSCRs in the feedline of a reference antenna to significantly increase harmonic suppression through the sixth harmonic, without altering the antenna's basic resonant pattern at the fundamental frequency or increasing the size of the feed network.

REFERENCES

1. Yasushi Horii and Makoto Tsutsumi, "Harmonic Control by Photonic Bandgap on Microstrip Patch Antenna," IEEE Microwave and Guide Wave Letters, Vol. 9, No. 1, January 1999, pp. 13-15.

2. Y. J. Sung and Y.-S. Kim, "An Improved Design of Microstrip Patch Antennas Using Photonic Bandgap Structure," IEEE Transactions on Antennas and Propagation, Vol. 53, No. 5, May 2005, pp. 1799-1804.

3. Jae-GonLee and Jeong-HaeLee, "Suppression of Spurious Radiations of Patch Antennas Using Split-Ring Resonators (SRRs)," Microwave and Optical Technology Letters, Vol. 48, No. 2, February 2006, pp. 284-287.

4. M. K. Mandal, P. Mondal, S. Sanyal, and A. Chakrabarty, "An Improved Design of Harmonic Suppression for Microstrip Patch Antennas," Microwave and Optical Technology Letters, Vol. 49, No. 1, January 2007, pp. 103-105.

5. Luis Incln-Snchez, Jos-Luis Vzquez-Roy, and Eva Rajo-Iglesias, "Proximity Coupled Microstrip Patch Antenna With Reduced Harmonic Radiation," IEEE Transactions on Antennas and Propagation, Vol. 57, No. 1, January 2009, pp. 27-32.

6. Yunchuan Guo, George Goussetis, Alexandros P. Feresidis, and John C. Vardaxoglou, "Efficient Modeling of Novel Uniplanar Left-Handed Metamaterials," IEEE Transactions on Microwave Theory and Techniques, Vol. 53, No. 4, April 2005, pp. 1462-1468.

7. T. Kokkinos, A.P. Feresidis, and J.C. Vardaxoglou, "On the use of spiral resonators for the design of uniplanar microstrip-based left-handed metamaterials," in Proceedings of the European Conference on Antennas and Propagation, 2006, Nice, France.

8. Titos Kokkinos, Alexandros P. Feresidis, and John C. Vardaxoglou, "Equivalent Circuit of Double Spiral Resonators Supporting Backward Waves," Loughborough Antennas and Propagation Conference, Loughborough University of Technology, England, 2007, pp. 289-292.

9. Titos Kokkinos and Alexandros P. Feresidis, "Low-Profile Folded Monopoles with Embedded Planar Metamaterial Phase-Shifting Lines," IEEE Transactions on Antennas and Propagation, Vol. 57, No. 10, October 2009, pp. 2997-3008.