W-Band Filtering Antenna Based on a Slot Array and Stacked Coupled Resonators Using Gap Waveguide Technology

This letter proposes a new design approach for filtering antennas. The novel matching reflection coefficient-based method allows the integration of filters and antennas without compromising the frequency behavior of either of these components. Moreover, this integration is done by avoiding the need for lengthy optimization processes and provides a high degree of flexibility in the types of antennas that can be used. In order to validate it, two examples are provided. In both cases, a 4th-order Chebyshev bandpass filter at 101.5 GHz implemented in a stacked groove gap waveguide configuration is used, firstly along with a single aperture antenna and, subsequently, with a slotted ridge gap waveguide array. This second example has been manufactured to demonstrate the usefulness of the new design methodology. Excellent measured performance has been obtained for a filtering antenna at W-band for the first time.


I. INTRODUCTION
F ILTERING antennas have gained significant attention in recent years due to their ability to integrate the functionalities of both an antenna and a filter in a single device, eliminating the need for any extra interface or transition.The main objective of a filtering antenna is to exhibit the aimed radiation characteristics within a specific passband while effectively suppressing undesired frequencies at designated stopbands.These components have found applications in diverse fields such as wireless communication, radar systems, satellite communication, and Internet-of-Things (IoT) devices, where simultaneous filtering and transmission/reception capabilities are essential.
Extensive research has been conducted on the integration of antennas and filters [1].Different planar technologies have been employed for filtering antennas such as microstrip [2], [3] the substrate-integrated waveguide (SIW) technology [4], [5], which are mainly suitable for low-frequency applications since their transmission properties greatly depend on the substrate parameters, suffering from high insertion loss at the millimeterwave frequency range.Rectangular waveguide is well known as the technology that provides the lowest losses and has also been employed for this kind of device [6], [7], [8].Nevertheless, ensuring electrical contact becomes very crucial when assembly and manufacturing needs to be done in a clamshell configuration, especially at high mmWave frequencies.Because of that, gap waveguide technology [9] emerged as an alternative at millimeter waves and has also been used in the design of filtering antenna devices, although not at frequencies higher than the Ka-band [10], [11], [12] and V-band [13].
Different design approaches for filtering antennas have been documented in the literature.The simplest strategy is to consider the antenna as a load element, as in [14], where a filter and an antenna, initially designed independently, are later interconnected using a coaxial line.An interesting approach is to embed the filter in the feeding network of the antenna, as done for instance in [11].Another alternative is to integrate both components into one structure, sharing resonating cavities or other circuit elements.In order to do it, long optimization processes of the whole structure may be required, as in [8], where the S 11 -parameter of the whole filtering antenna must be optimized to obtain the required dualband response.Aiming to reduce the optimization time, other approaches based on the coupling matrix have been proposed in [6], [7].In these works, a design parameter Q r (radiation quality factor) is defined to model the filtering antenna and then, it is made to coincide with the external quality factor, Q ext1,N , of the filter (where N is the number of resonators).Although this allows the integration of both the antenna and the filter, the frequency responses that can be achieved are constrained by the need to have a value of Q r equal to Q ext1,N and its validity is limited to a reduced frequency range.Moreover, in order to have this equivalence, it is often necessary to modify the dimensions of the antenna, compromising its radiating behavior.
This letter presents a novel matching reflection coefficientbased method for the design of filtering antennas using gap waveguide technology, which allows the integration of the antenna and the filter in one device without compromising the radiating characteristics of the antenna and facilitates accomplishing wide band frequency devices if required.First, we will describe the technique employing a single radiating element and then we will validate its usefulness employing a more complex radiation structure such as a slotted ridge gap waveguide (RGW) array antenna.Excellent characteristics have been demonstrated for the first time for a filtering antenna at W-band in gap waveguide technology.

II. DESIGN METHOD
The design procedure begins with the design of the filter as if it were an isolated structure.Therefore, a variety of filters implemented with different configurations can be considered.In this case, we will employ stacked resonator cavities coupled through an iris, as in [15], making the circuit footprint more compact.However, unlike [15], where RGW resonators are employed, higher-order-mode groove gap waveguide (GGW) resonators are used in our case.This allows the use of larger cavities that provide more robustness during fabrication using classical low-cost methods such as computer numerical control (CNC) milling, as explained in [16], where the TE 103 mode was employed for this purpose using inline couplings.Additionally, higher-order modes in a stacked configuration allow us to modify the position of the iris, keeping the same coupling and, hence, having more flexibility and an extra variable to use during the design and further optimization, if needed, due to their field distribution along the cavity.In this step, we need to fix the filtering specifications determined by the order of the filter, N, the in-band return losses, RL, and the passband defined by its lower and upper frequencies, f 1 and f 2 , respectively.For a bandpass between f 1 and f 2 , the coupling between the resonating cavities, K i, i+1 , and the external quality factor, Q ext1,N , are calculated as in [17] where BW is the fractional bandwidth and g i are the equivalent low-pass filter coefficients.Once K i, i+1 and Q ext1,N values are known, the dimensions and position of the irises (w i,i+1 , t i,i+1 , Offset) located in the bottom walls of the cavities can be determined as explained also in [17].A cross-section of the stacked filter topology is shown in Fig. 1(a) along with a single cavity and an iris in Fig. 1(b).
Once the filter has been designed as an isolated structure, the second step involves focusing on its last stage, which we will refer to as the Filter Output Stage (FOS).Since the reflection coefficient of the FOS, denoted as Γ, is a key parameter in the design procedure, it is necessary to define the FOS precisely.The FOS in coupled cavity filters may consist of the filter output iris (responsible for the Q extN ), the last resonating cavity, and the last coupling iris (responsible for the coupling coefficient k N-1, N ) located on the bottom wall of the last cavity, as depicted in Fig. 1(c).The reflection coefficient of the FOS, Γ, is calculated over the bandwidth of interest using a full-wave electromagnetic simulator, such as CST MWS, and it will be the frequency response of an isolated resonator if it has been defined as previously explained.Actually, it would also be possible to evaluate Γ at any other position of the device, provided that the rest of the procedure is also performed consistently.
Once the FOS is evaluated, the third step of the method involves replacing the filter output of the FOS with the radiation element that will be used in the filtering antenna.The new structure, referred to as the Radiation Output Stage (ROS), is composed of the radiation element (for example, a single aperture), the last resonating cavity, and the last coupling iris, as illustrated in Fig. 1(d).The reflection coefficient of the ROS denoted as Γ r , is calculated, and the aim in this step is to keep the reflection coefficient unaltered, ensuring that Γ r = Γ over the frequency of interest.This method is general for any kind of radiation element and valid as long as Γ r replicates Γ, which would preserve the overall filtering characteristics.In order to achieve it, it is necessary to identify which variables of the ROS can be modified without altering the radiation characteristics, which must also remain unchanged.This will also depend on the radiating element used, as we will see in the examples below, and will ensure that the radiating characteristics of the antenna will not be modified when integrated with the filter.
After successfully replicating the reflection coefficient, the last step of the method consists of substituting the FOS of the filter previously designed with the final ROS since the rest of the filter remains unchanged in the filtering antenna.No further optimization is necessary, as the filter characteristics have already been restored.The final configuration is depicted in Fig. 1(e).

A. Design and Simulation of the Filter
As said above, the first step of the procedure consists of the design of an isolated bandpass filter that, in this example, will have N = 4, f 1 = 98.5 GHz, f 2 = 104.5 GHz, and RL better than Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.As discussed in Section II, the higher-order TE 103 mode has been chosen to design a device more robust to manufacturing tolerances and the couplings have been implemented using an iris in the bottom wall of the cavities.A cross section of the filter can be observed in Fig. 1(a), and the corresponding dimensions are provided in Table I.The simulated filter response is shown in Fig. 2

B. Design and Simulation of the Filtering Antenna With a Single Aperture as the Radiation Element
Following the matching reflection coefficient-based method explained in Section II, a filtering antenna will be implemented, integrating the previously designed BPF and a radiation element based on an aperture [19].Once the filter has been designed, the second step involves isolating the FOS, see Fig. 1(c), to obtain its reflection coefficient, Γ, shown in Fig. 2(a).In the third step, we consider the ROS, which includes the desired radiation element, i.e., a radiation aperture in this case.The ROS, see Fig. 1(d), is evaluated to obtain its reflection coefficient, Γ r , and the aim now is to adjust the physical parameters of the ROS to replicate the reflection coefficient of the FOS without altering the radiation characteristics.To this end, we identify in this case the length of the last resonating cavity, L 4 , and the last coupling iris width, w 3,4 , as the parameters that can be adjusted without modifying the radiating characteristics of the antenna used.Fig. 3 shows how Γ r varies as the cavity length and iris width are modified.The influence of the iris width in the magnitude of the Γ r is shown in Fig. 3(a), while Fig. 3(b) shows how to shift the frequency resonance by adjusting the cavity length.The comparison between the reflection coefficient of the FOS (red line) and the ROS (black line) after adjusting these two parameters is presented in Fig. 2(a), where a good agreement is obtained.Finally, we physically replace the FOS with the ROS to have the final component.As explained previously, when Γ r closely matches Γ, no further optimization is necessary.Fig. 2(b) illustrates the comparison between the return loss of the BPF and the filtering antenna shown in Fig. 1(e).The dimensions of the final design are included in Table I.

C. Design and Simulation of the Filtering Antenna With a Slotted RGW Array as the Radiation Element
As shown in Fig. 4, an 8-element center-fed slot array antenna based on RGW [18] will be used in this example to explore the feasibility of the matching reflection coefficient-based method with a more complex radiation element.The radiating characteristics are primarily determined by the dimensions of the slotted aperture, such as the slot length l, slot offset d s2 and the longitudinal spacing d s1 , which can be well chosen to achieve in-phase radiation at each slot within a desired bandwidth.
Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.Therefore, the length of the last resonating cavity, L 4 , the last coupling iris width, w 3,4 , and the parameters of the ridge and center-located coupling hole, including L r1 , L r2 , H r1 , H r2 , w t , and L t have been identified to adjust Γ r without affecting the radiation characteristics.Although in this case the number of variables and, therefore, the complexity is also larger, a good similarity between the reflection coefficients of the FOS and the ROS can be observed in Fig. 5(a).In order to consider the effect of the rounded corners included in the geometry to be fabricated by CNC milling (radius equal to 0.2 mm) and compensate that the similarity between the reflection coefficients Γ and Γ r is not as identical as in the previous example, a minor optimization is required in this case.The final dimensions are in Table I.Fig. 5(b) illustrates the comparison between the return loss of the BPF and the filtering antenna, while Fig. 6 shows the gain of the device across the passband.

IV. MANUFACTURING AND MEASUREMENTS
The second filtering antenna design has been manufactured in brass employing traditional CNC milling and, afterwards, gold plated.Photographs of the fabricated filtering antenna and the measurement setup are shown in Fig. 7.In Fig. 6, a good measured filtering performance is shown, with an average return loss level of 14 dB and a centered passband.Furthermore, the filtering antenna shows a measured gain of more than 14.5 dBi, close to the simulations along the whole passband.Fig. 8 shows the simulated and measured radiation patterns of the filtering antenna at 98.5 GHz, 101.5 GHz, and 104.5 GHz in both  H-and E-planes.As observed, the sidelobe level is below 13 dB in the H-plane.Generally, there is a good match between the measured and simulated results, and the discrepancies can be caused by fabrication inaccuracies and the employment of a metal frame to measure the radiation pattern.A comparison with previous works on filtering antennas implemented in GGW is presented in Table II, where it can be seen that the proposed device with a slotted RGW array merits high capabilities and is very suitable for millimeter-wave frequency range applications, being the first filtering antenna operating at W-band.In contrast to the previous works referenced in the table, the use of a single higher-order mode cavity configuration ensures the required robustness against the manufacturing tolerances.

V. CONCLUSION
This letter introduces a matching reflection coefficient-based method for filtering antennas at W-band.To demonstrate the usefulness of the proposed technique, two different prototypes were designed in GGW technology, each one incorporating different radiation elements.The first design featured a radiation aperture, while the second prototype used a slotted RGW antenna.The second prototype was manufactured using CNC milling.The agreement between measurements and simulations highlights the usefulness of this approach and paves the way for the design of high-performance and high-frequency filtering antennas.Moreover, the method can be used in other microwave technologies, such as microstrip or SIW, also allowing the design of filtering antennas in other frequency ranges.

Fig. 1 .
Fig. 1.Steps of the matching reflection coefficient-based method: (a) Schematic of a stacked BPF implemented in GGW; (b) schematic of a single cavity; (c) schematic of the filter output stage (FOS); (d) schematic of the radiation output stage (ROS); and (e) schematic of the stacked coupled-resonator filtering antenna using an aperture as a radiation element.

Fig. 2 .
Fig. 2. Filtering antenna using a radiation aperture as a radiation element: (a) Comparison of the reflection coefficient of the FOS and ROS; (b) comparison between the return losses of the filter and of the filtering antenna.
(b), where the |S 21 | is shown in the green line and the |S 11 | in the red line.

Fig. 5 .
Fig. 5. Simulated S-parameters of the filtering antenna using a slotted RGW array as a radiation element: (a) Comparison of reflection coefficient of the FOS and of the ROS; (b) comparison between the return losses of the filter and of the filtering antenna.

Fig. 6 .
Fig. 6.Comparison between the simulated S-parameters of the filtering antenna (black line) and its measurements (blue line): |S 11 | in solid line and realized gain in dashed line.
Mu Fang and Ashraf Uz Zaman are with the Department of Electrical Engineering, Chalmers University of Technology, 412 96 Gothenburg, Sweden.Digital Object Identifier 10.1109/LAWP.2024.3399269

TABLE II COMPARISON
OF THE FILTERING ANTENNA WITH OTHER PREVIOUS WORKS IN GGW TECHNOLOGY