Abstract
It is shown that printed antennas loaded with metamaterial resonators can be designed to exhibit multiband functionality. Two different antenna types and metamaterial loading are considered: (i) printed dipoles or monopoles loaded with open complementary split ring resonators (OCSRRs) and (ii) meander line or folded dipole antennas loaded with split ring resonators (SRRs) or spiral resonators (SRs). In the first case, multiband operation is achieved by series connecting one or more OCSRRs within the dipole/monopole. Such resonators force opens at their positions, and by locating them at a quarter wavelength (at the required operating frequencies) from the feeding point, it is possible to achieve multiple radiation bands. In the second case, dual-band functionality is achieved through the perturbation of the antenna characteristics caused by the presence of the metamaterial resonators. This latter strategy is specially suited to achieve conjugate matching between the antenna and the chip in radiofrequency identification (RFID) tags at two of the regulated UHF-RFID bands.
1. Introduction
Metamaterials are effective media made of periodic (or quasiperiodic) inclusions of conventional materials (typically metals and dielectrics) with controllable acoustic, electromagnetic, or optical properties. Indeed, by properly structuring these artificial materials, it is possible to achieve unique and exotic properties, such as negative refraction or subwavelength focusing, among others, and it is potentially possible to implement acoustic and optical cloaks. There has been an intensive research activity in this field since 2000 [1–10], when the first metamaterial structure by Smith and co-workers was reported [11]. Key to the success of this research field was the synthesis of negative effective permeability media by means of split ring resonators (SRRs) [12]. These particles are electrically small resonators that can be excited by means of an axial magnetic field. Hence, if an array of SRRs is illuminated by means of an electromagnetic radiation with the magnetic field axial to the SRRs, the structure behaves as an effective medium with negative permeability in a narrow band above SRR resonance. By combining SRRs with metallic posts (exhibiting a negative effective permittivity up to the so-called plasma frequency), the first structure simultaneously exhibiting negative permittivity and permeability was synthesized [11]. These structures are called left-handed, or double negative, materials, and the main relevant feature of such media, derived from the negative sign of the constitutive parameters, is the anti-parallelism between the phase and group velocities [13].
The metamaterial concept and some metamaterial properties were soon transferred to the microwave engineering field, since it was demonstrated that by loading a transmission line with series capacitances and shunt inductances, left-handed wave propagation (i.e., backward waves) is achieved [14–16]. These artificial lines loaded with series capacitances and shunt inductances and exhibiting some metamaterial properties are called metamaterial transmission lines. In 2003, it was demonstrated that metamaterial transmission lines can also be implemented by means SRRs [17]. Specifically, by properly combining SRRs and shunt strips in a coplanar waveguide (CPW) transmission line, backward wave propagation in a narrow band arises, and by merely loading the CPW with SRRs, the structure exhibits a stop-band behavior that has been interpreted as due to the negative value of the effective permeability in the stop band [18]. It is important to mention that even though effective permeability and permittivity in metamaterial transmission lines can be defined [1, 2, 4], the relevant parameters in transmission lines (fundamental for design purposes) are the phase constant and the characteristic impedance. Thanks to the presence of reactive elements (inductances, capacitances, or electrically small resonators) in metamaterial transmission lines, it is possible to tailor these parameters to some extend in these artificial lines. Therefore, the key advantage of metamaterial transmission lines over conventional lines is the controllability of the phase constant and characteristic impedance [19]. In many applications few (or even a single) unit cells suffice to achieve the required line specifications. Hence, periodicity and homogeneity (necessary for the implementation of effective media metamaterials) are not a due in metamaterial transmission lines. Indeed, some times, rather than tailoring the characteristic impedance or the phase of a transmission line, it might be simply necessary to use coupled resonators in order to achieve a transmission zero or to perturb somehow the line characteristics in the vicinity of particle resonance. As long as the considered resonators are electrically small and related to the SRR, these resonator-loaded lines are called transmission lines with metamaterial loading. This concept can be made extensive to other structures such as printed antennas, which is the case considered in this paper.
The paper is organized as follows. In Section 2, the metamaterial resonators considered in this paper are reviewed. Section 3 is focused on multiband printed dipole and monopole antennas loaded with open complementary split ring resonators (OCSRRs), and Section 4 is devoted to the implementation of dual-band UHF-RFID tags based on meander line and folded dipole antennas with metamaterial loading. Finally the main conclusions are highlighted in Section 5.
2. SRRs and Other Related Metamaterial Resonators
The typical topology of the SRR is depicted in Figure 1(a). It consists on a pair of coupled split rings with the apertures on opposite sides. The coupling between both rings drives the first resonance frequency to low values, as compared to that of the individual rings, and for this reason the SRR is electrically small. By virtue of the small electrical size of the particle, a quasistatic analysis can be applied to the SRR [4, 20], and the capacitance and inductance are given by the series connection of the distributed (edge) capacitances of both SRR halves and by the inductance of a single loop of identical width and average radius . Pairs of SRRs etched in the back substrate side of CPW transmission lines can be inductively driven by the magnetic field generated by the line, resulting in negative effective permeability metamaterial transmission lines [17, 18].
(a)
(b)
(c)
(d)
By applying duality, the complementary split ring resonator (CSRR) results [21, 22] (Figure 1(b)). This resonator exhibits roughly the same resonance frequency to that of the SRR (provided the same dimensions and substrate are considered) and is typically driven by an axial time-varying electric field. CSRRs have been used for the implementation of negative effective permittivity metamaterial transmission lines in microstrip technology, where the CSRRs are etched in the ground plane beneath the conductor strip (more details on the calculation of particle inductance and capacitance are given in [23]).
Let us now consider the open versions of the previous particles, that is, the open SRR (OSRR) [24] and the open CSRR (OCSRR) [25]. The typical topologies are depicted in Figures 1(c) and 1(d), respectively. Notice that as long as these particles are open resonators (the connecting pins are indicated), they can be excited by means of a voltage or a current source. As compared to SRRs and CSRRs, OSRRs and OCSRRs are electrically smaller by a factor of two. The reason is simple: for the OSRR, the inductance is identical to that of the SRR, but the capacitance is the edge capacitance of the whole circumference, and hence it is 4 times higher than that of the SRR. This reduces the resonance frequency of the OSRR to half the resonance frequency of the SRR. For the OCSRR, the capacitance is identical to that of the CSRR, but the inductance is four times larger since in the CSRR there are two inductive paths (in parallel) between the inner region of the particle and the outside metallic region, whereas for the OCSRR there is a single inductive path corresponding to the whole circumference of the particle. As can be seen in the circuit models depicted in Figure 1, the OSRR is an open series resonator, whereas the OCSRR is an open parallel resonator. It has been demonstrated that combined OSRRs and OCSRR in CPW and microstrip lines are useful for the implementation of metamaterial transmission lines (and many circuits based on them) and broadband bandpass filters [26].
By connecting the external pins of the OSRR, a 2-turn spiral resonator (2-SR) results [27, 28] (other multiple-turn spirals have been studied in the literature [29]). Therefore the 2-SR is a closed resonator that exhibits the same resonance frequency as that of the OSRR. It can also be concluded from duality arguments that the complementary counterpart of the 2-SR exhibits also this resonance frequency (which is also that of the OCSRR).
3. Multiband Printed Dipole and Monopole Antennas with Metamaterial Loading
Let us now focus on the design of antennas loaded with metamaterial resonators. Indeed, the improvement of antenna performance by means of metamaterials has been a subject of intensive research in recent years. Metamaterial substrates and superstrates to improve the antenna directivity in planar antennas, or leaky wave antennas with scanning capability based on metamaterial transmission lines, are two of the multiple applications of metamaterial technology in the field of antennas [1–5, 30–33] (an in-depth review of this topic is out of the scope of this paper).
In this section, the focus is on metamaterial-loaded printed antennas [34–36]. The idea of this technique is based on loading a conventional printed antenna with a set of resonant particles. For example, in [34], it is shown that a dual-band antenna is achieved by coupling a set of SRRs to a printed dipole. Using this approach, the benefits of printed antennas are kept while dual-band antennas are achieved by using a simple design technique. The SRRs produce open circuits in their positions at the resonance frequencies. Hence the antenna resonance is achieved not only when the effective length of the dipole arms is λ/4 (λ being the guided wavelength), but also when the different locations of the SRR are λ/4 from the antenna feeding point (the SRRs must be tuned at these frequencies). By using SRRs, narrow bandwidths are reported in [34] for the bands associated with the SRR loading. Even the use of SRRs with different resonance frequencies leads to bandwidths smaller than 5% [35], which might not be useful for most applications. However, an open circuit in the dipole arms can also be obtained by means of OCSRRs. The advantage is that by series connecting OCSRRs within the dipole, broadband responses can be achieved (as compared to SRRs), due to the relative values of capacitance and inductance.
To illustrate the potential of OCSRR-loaded printed antennas, we report here a dual-band dipole [37] and a tri-band monopole [38]. The layout of the printed dipole (an antipodal structure) and the details of the OCSRR are depicted in Figure 2, whereas the photograph of the antenna is shown in Figure 3, and compared to the conventional monoband dipole. The parameters of each dipole strip are the length and the width (which must be engineered to optimize matching). This configuration has been chosen because it avoids the use of a balun to feed the antenna. This is possible because the antipodal printed dipole is fed through a paired strips transmission line with a SMA connector soldered to the end of the line. The dimensions of the feeding line are the length and the width . An OCSRR is connected in series to each dipole strip at a distance from the center of the antenna. The different dimensions of the proposed antenna have been optimized to simultaneously operate at the L1-GPS frequency (1.575 GHz) and the WiFi band of 2.4–2.48 GHz. The final dimensions of the design are mm, mm, mm, mm, mm, mm, and mm. The used substrate is the Rogers RO3010 with and mm.
(a)
(b)
In the case of the conventional unloaded antenna, there is only one series resonance within the band of interest. This resonance frequency is 2.2 GHz. This resonance corresponds to the fundamental mode of the dipole antenna. Moreover, in the vicinity of the resonance, the value of the real part of the impedance is close to 50 Ω (not shown), which produces proper matching, as it can be observed in Figure 4. On the other hand, the OCSRR loading introduces a parallel resonance in the proposed dual-band antenna input impedance. The addition of this parallel resonance has a double effect on the input impedance of the proposed dual-band antenna. The first one is a slight shift on the series resonance of the dipole antenna towards higher frequencies. In this case, this resonance is found at 2.45 GHz. The second effect is the appearance of an additional series resonance below the parallel resonance of the OCSRRs. This additional series resonance is found at 1.6 GHz. The real part of the input impedance is around 50 Ω in the vicinity of this additional series resonance. This allows achieving an additional band with proper matching in the proposed antenna (Figure 4). As it can be seen, this additional frequency is below the fundamental frequency of the unloaded dipole antenna, achieving some degree of miniaturization.
Considering below −10 dB, the first working band is centered at 1.56 GHz with 5% bandwidth and the second one is centered at 2.46 GHz with 9% bandwidth. It is important to note that this is a considerable improvement with respect to other works in which the bandwidth of one of the bands was always below 5% [34, 35]. The obtained results cover the bandwidth of the proposed applications (GPS and WiFi). The gain of the antenna has been estimated in a TEM cell from the power received by the antenna and the incident field measured by a probe [39]. The measured gain of the proposed antenna is 0.85 dB at the GPS band (1.575 GHz) and 2 dB at the WiFi band (2.45 GHz). These experimental results agree with the CST-simulated ones: 0.9 dB at the first band and 2.2 dB at the second one.
The CST-predicted radiation patterns for the proposed dual-band antenna are shown in Figure 5 for completeness. A dipolar-like radiation pattern is obtained at both working bands. The typical figure of eight is obtained in the XZ plane and an omnidirectional pattern is obtained in the YZ plane at both frequencies (1.575 and 2.45 GHz). The only difference between both working bands is the cross-polarization level, which is around −10 dB in the first band while it is −20 dB in the second one.
(a)
(b)
Following the ideas of the antipodal dipole antennas with metamaterial loading, we have also designed multiband printed monopole antennas. Figure 6 shows a photograph of two fabricated prototypes: a dual band and a tri-band printed monopole. The considered substrate in this case is the low-cost FR4 ( and mm). The dual-band antenna covers the bands of 2.40–2.48 GHz (Bluetooth and WiFi) and 5.15–5.80 GHz (WiFi). The final dimensions of the monopole are mm, mm. The parameters of the feeding line are set to obtain a 50 Ω CPW. Hence, mm and mm. The dimensions of each ground plane are mm and mm. The gap between the ground planes and the monopole is mm. The OCSRR is placed at a distance mm and its parameters are mm, mm. The gap is set to 0.50 mm.
(a)
(b)
(c)
The tri-band monopole antenna is an extension of the previous dual-band antenna and it covers the previous bands and the IEEE 802.11y band of 3.65–3.70 GHz. According to the layout of the tri-band monopole (Figure 6), the additional OCSRR is placed at a distance mm. Its design parameters are mm, mm. This corresponds to a resonance frequency of 3.65 GHz. The gap is set to 0.40 mm. These values have been optimized to only cover the desired bandwidth and not interfere with other systems. The other parameters of the antenna remain unchanged with respect to the dual-band design, except the length of the monopole which is reduced to mm to compensate the inductive behavior of the OCSRRs below their resonance frequencies.
The simulated and measured reflection coefficients of the dual-band and tri-band monopoles are depicted in Figure 7. The dual-band printed monopole antenna exhibits good matching ( dB) from 2.29 GHz to 2.52 GHz at the lower frequency band. This corresponds to a 9.6% bandwidth. In the upper band, the antenna is well matched from 4.66 GHz to at least 7 GHz. Thus, the fabricated antenna satisfies the specifications of Bluetooth and WiFi (bands of 2.40–2.48 GHz and 5.15–5.80 GHz). The fabricated tri-band monopole antenna is well matched from 2.30 GHz to 2.52 GHz for the first band. Its reflection coefficient is below −10 dB between 3.56 GHz and 3.78 GHz for the second band, and between 5.06 GHz and 6.71 GHz for the third band. Hence, the fabricated prototype is well matched within the regulated bandwidths of Bluetooth and WiFi including IEEE 802.11y (3.65–3.70 GHz band).
The proposed antennas present monopolar radiation characteristics at all the bands. As an example, the normalized measured radiation patterns of the tri-band monopole antenna are shown in Figure 8. A monopolar radiation pattern is obtained at the three frequencies. The cross-polar component (XPOL) has low values (below −20 dB in all of the cases, except the XY plane of the third frequency which is below −15 dB). The gains of this design are 1.4 dB, 1.2 dB, and 1.7 dB at the first, second, and third bands, respectively. These results are in good agreement with simulations, in which the radiation efficiency is 92%, 83%, and 94% and the overall efficiency is 91%, 82%, and 93% at the central frequency of each band.
(a)
(b)
(c)
4. Dual-Band UHF-RFID Tags
Metamaterial loading is also interesting for the implementation of dual-band antennas with closed frequency bands, as is the case of UHF-RFID, where the different regulated bands worldwide are contained in the spectral region between 860 MHz and 960 MHz. In this case, however, the approach is based on a perturbation method which was reported in [40]. The key aspect in the implementation of long read-range UHF-RFID tags is to achieve conjugate matching between the antenna and the integrated circuit (or chip). The input impedance of the chip is provided by the manufacturer and varies with frequency. Therefore, the implementation of dual-band UHF-RFID tags means to design the antenna (and the matching network, if it is present) so that the chip “sees” its conjugate impedance at the required frequencies. This can be done by cascading a dual-band impedance matching network between the antenna and the chip, consisting on a transmission line loaded with a metamaterial resonator [40]. The resonator produces a perturbation in both the characteristic impedance and the phase constant of the transmission line, and conjugate matching at two frequencies can be obtained (the details are given in [40]). However, it is possible to directly actuate on the characteristic impedance of the antenna, by loading it with metamaterial resonant particles [41] (avoiding thus the matching circuitry). The principle is very similar to that reported in [40] for matching networks.
Following the above-cited perturbation approach, several dual-band UHF-RFID tag antennas have been implemented. One prototype device consists on a meander line antenna (MLA) loaded with a spiral resonator (SR). The layout of the antenna, compared to the one without the SR is depicted in Figure 9. The dual-band antenna has been designed to be operative at the European (867 MHz) and USA (915 MHz) UHF-RFID regulated bands. The antenna has been designed on the Rogers RO3010 with dielectric constant and thickness mm. The considered RFID chip is the SL31001 from NXP semiconductors. The impedance of the chip at the intermediate frequency is Ω. This impedance is considered as reference impedance so that the antenna is designed to roughly exhibit the conjugate impedance of the chip at this intermediate frequency, and then the perturbation (by means of the SR) is introduced in order to achieve conjugate matching at the required frequencies. The dimensions of the MLA are 48 mm × 48 mm, and the strip width is 1.4 mm. The other relevant dimensions are mm, mm, mm, mm, and mm. The input impedance of the MLA is depicted in Figure 10, whereas Figure 11 depicts the input impedance and matching of the MLA loaded with the SR. As can be appreciated, the dual-band functionality in the SR-loaded MLA is achieved.
(a)
(b)
(a)
(b)
(a)
(b)
Both antennas have been fabricated, and the read range, given by has been measured through the experimental setup available in our laboratory (Figure 12). In (1), is the wavelength, EIRP, determined by local country regulations, is the product of which are the transmission power and the transmission gain, respectively, is the minimum threshold power necessary to activate the RFID chip, is the gain of the receiving tag antenna, and is the power transmission coefficient. The value of EIRP in European frequencies is 3.3 W; whereas in American frequencies it is 4 W. The power transmission coefficient is inferred from the simulation of the return loss of the antenna, using as port impedance that of the chip. The tag gain is also obtained from simulation; hence the theoretical read range can be calculated.
The experimental setup consists of a N5182A vector signal generator which creates RFID frames. Such generator is connected to a TEM cell by means of a circulator. The tag under test is located inside the TEM cell and it is excited by the frame created by the generator. Then the tag sends a backscatter signal to a N9020A signal analyzer through the circulator. The RFID frame frequency is swept with a specific power. Once the operation frequencies are identified, the output power is decreased until the tag stops working. Finally, a probe is placed into the TEM cell in order to determine the incident electric field intensity at each frequency, which is related to the received power by the chip according to where is the incident power density, is the effective area of the tag antenna, and η is the free-space wave impedance (which is equivalent to 120π Ω). The measured read range can be inferred by introducing (2) in (1), resulting in
As indicated before, the EIRP European value is lower than the USA counterpart, so the read range in Europe would be roughly reduced by a 0.9 factor for the same incident electric field intensity.
The fabricated tags, as well as the simulated and measured read ranges, are depicted in Figure 13, where it can be seen that the read range obtained in the dual-band tags (at the frequencies of interest) is superior to that of the monoband tag. The read range of the monoband MLA tag is roughly 4 m at the frequencies of interest, whereas almost 6 m and 8 m at the European and USA frequency bands, respectively, are achieved by means of the designed dual-band MLA. This enhancement in the read range is due to an improved matching between the antenna and the integrated circuit, since it has been verified by simulation that the radiation efficiency of the monoband MLA is almost constant from 867 MHz to 950 MHz. In fact, if it was possible to achieve a perfect matching between the monoband antenna and the integrated circuit in all the band of interest, the obtained read range would be around 8 m at all frequencies. Nevertheless, this assumption is not possible due to the particularities of the chip impedance.
(a)
(b)
(c)
The perturbation method reported above can also be applied to folded dipole RFID tag antennas [42]. Figure 14 depicts a prototype, fabricated on the Rogers RO3010 with dielectric constant and thickness h mm. The dimensions are 62.7 mm × 17.9 mm. The simulated and measured read ranges are depicted in Figure 15 and compared to those of the tag without the presence of the SR. Again, the dual-band functionality at the regulated RFID European and USA frequency bands is achieved.
We would like to mention that in the designed RFID tags, the resonator frequency is set at the intermediate frequency (891 MHz), and as a consequence, the resonator impact on the antenna radiation efficiency can be neglected at the frequency bands of interest (867 MHz and 915 MHz).
5. Conclusions
In conclusion, two different approaches for the implementation of multiband printed antennas have been reviewed: one of them consisting on the introduction of series connected OCSRRs in the arms of printed dipole or printed monopole antennas; the other one based on a perturbation method achieved by loading printed antennas, such as meander line antennas (MLAs) or folded dipoles, with spiral resonators (SRs). The former approach has been revealed to be useful for personal area networks (PANs) and wireless local area networks (WLANs); the second one is of interest in the field of UHF-RFID. In both cases, the multiband functionality has been demonstrated through the fabrication and characterization of several prototype devices.
Acknowledgments
This work has been supported by Spain-MICIIN (Project contracts TEC2010-17512 METATRANSFER and CSD2008-00066) and Spain MITyC through the Projects TSI-020100-2009-778 and TSI-020100-2010-493. Thanks are also given to the Catalan Government for giving support through Project 2009SGR-421.