Monolithically Integrated Coaxial Resonator-Based Filtennas Using SLA Three-Dimensional Printing

This letter reports on a new class of highly-miniaturized coaxial-resonator-based filtennas alongside a monolithic integration concept using digital additive manufacturing. The filtenna approach is based on the co-design of a coaxial cavity resonator and an annular slot antenna that collectively perform as a radiating resonating node within a coupled-resonator-based bandpass filter. Beam direction can be controlled by altering the position of the radiating slot. A stereolithography apparatus (SLA) monolithic integration concept is proposed for size compactness and reduced weight. For proof-of-concept validation purposes, first-, second-, and third-order filtennas were analyzed through electromagnetic simulations. Experimental validation was performed through the manufacturing and testing of a second-order top-slot filtenna, a second-order and a third-order side-slot filtenna with the following characteristics: center frequency: 4.75 GHz, 4.8 GHz and 4.6 GHz, fractional bandwidth: 7.4%, 8.3%, and 7.6% and realized gain: 5.5 dBi, 2.6 dBi, and 2.6 dBi.


I. INTRODUCTION
T HE rapid growth of wireless technologies is increasingly creating the need for small form factor and low weight RF transceivers. RF co-design has been identified as one of the most effective size reduction techniques due to reducing the number of components in RF front end. Typical examples of this trend include co-designed bandpass filters (BPFs) and power dividers [1], BPFs and RF circulators [2], and BPFs and antennas [3], [4], [5], the so-called filtenna concept.
The majority of the filtennas to date are based on microstrip or substrate integrated waveguide (SIW) integration approaches [3], [4], [5], which result in low antenna efficiency. For example, the SIW-based filtenna in [5] exhibits an efficiency of 59% at 3.7 GHz. Metallic waveguide-based filtennas exhibit lower loss as shown in [6], for a cavity-backed slot antenna with an efficiency of 88%. However, these types of approaches exhibit large volume (>0.25 λ 3 0 ). Furthermore, they consist of multiple CNC-machined parts that need to be assembled with screws which further increase the complexity, size, and overall weight. Manuscript  In the recent years, digital additive manufacturing (DAM) has become a low-cost alternative for prototyping antenna elements. Metal-based AM process such as direct metal laser sintering (DMLS) [7] or selective laser melting (SLM) [8], [9] offers high mechanical robustness, but with relatively higher cost and low surface roughness. Ceramic-based or plastic/resin-based AM process [10], e.g., fused deposition modeling (FDM) [11], [12], [13], [14], [15], [16], [17], PolyJet [18], and stereolithography apparatus (SLA) [19], [20], [21] exhibit significantly lower cost and weight. However, due the need for internal metallization they can only be implemented as split-blocks. On the other hand, monolithic SLA integration concepts have only been demonstrated for open-ended waveguide structures [20], [21]. Considering the aforementioned limitations, this letter reports on the first time a monolithic SLA based-integration concept for a new class of coaxial-resonator-based filtenna. Size compactness is achieved through the: 1) RF co-design of two RF components, 2) use of miniaturized capacitively loaded coaxial resonators, and 3) monolithic integration enabled by SLA manufacturing.

II. THEORETICAL FOUNDATIONS
The conceptional coupling routing diagram (CRD) and power transmission coefficient |S 21 | of an Nth-order BPF, and the CRD and gain of an Nth-order filtenna are depicted in Fig. 1. The filtenna is created by replacing the last resonator and the load in the CRD of a conventional coupled-resonator-based BPF by a radiating resonating node. To achieve the same power reflection response, the radiating resonator must present the same impedance loading as the last resonator (N) and its preceding impedance inverter (Q ext ) in the conventional BPF. In what follows, the designs of the radiating resonator and the coaxial-resonator-based filtenna are provided.

A. Radiating Resonator Concept
The coaxial cavity radiating resonator concept is based on two annular slot antenna configurations, namely the top-slot and the side-slot in Fig. 2(a) and (b)-and a capacitively loaded coaxial cavity resonator. The frequency of the resonator is determined by the capacitive gap g between the upper wall of the cavity and the apex of the post, the radius ratio of the cavity b and the post a and the resonator height h [22]. The RF signal is coupled into the cavity by connecting the SMA connector to the post. The radiation pattern direction can be altered by placing the slot in different locations. When the slot is placed on the upper wall [see Fig. 2(a)], the main lobe is directed in the z-direction with a maximum gain of 5.2 dBi. In yet another configuration, the slot can be placed on the front side wall [see Fig. 2(b)] and in this case the antenna pattern is directed towards the x-axis with a gain of 2.6 dBi. As shown, the top-slot design exhibits higher gain due to the presence of a larger ground plane around the radiating slot and the bottom wall of the resonator that acts as an additional ground plane. Depending on the desired direction of radiation/gain, either of these configurations can be used.
As an initial design, the length of the slot is chosen around λ 0 /2. However, when added on the resonator its resonance frequency will change due to the slot loading. For the top-slot configuration, the resonant frequency is altered by changing the radius r slot and the angle θ slot of the, as shown in Fig. 3(a) and (b). In particular, the operating frequency decreases when the r slot or θ slot increases. Alternatively, the frequency can be tuned by changing g of the coaxial-cavity resonator [see Fig. 4(a)]. For the side-slot configuration, the frequency is dependent on the θ slot , and the slot height does not affect the frequency, as shown in Fig. 3(e) and (f). Fig. 3(c) and (g) illustrates the electric  (E)-field distribution on the surface of the upper and side wall of the radiating resonator at 4.8 GHz. As noticed, the quasi-TEM mode of the coaxial cavity is excited [22], where the E-field is strong around the post area due to the capacitive loading of the resonator, and the strongest E-field is observed around the slot. Furthermore, the E-field is transversal to the slot width, indicating that the antennas are linearly polarized.

B. Filtenna Design: First Order
The CRD of a first-order filtenna is shown in Fig. 4(b). It is comprised of a radiating resonator with a loaded quality factor Q L that is connected to the source via an external coupling element Q ext whose values can be calculated using where m 01 is the normalized coupling coefficient and Q U is the unloaded quality factor of the resonator. For impedance matching, the Q L of the radiating resonator must be equal to the Q L of the resonator in the conventional BPF. Thus, the first-order filtennas [see Fig. 2  using [5]  (2) where k is the coupling coefficient between the SMA connector and the under-coupled resonator. |S min 11 | is the minimum |S 11 | at the f 0 . f 1 and f 2 are frequencies where |S 11 | = |S Φ 11 |, as shown in Fig. 5(b). Using full-wave EM simulations, Q L and the operating frequency f 0 is extracted for different values of r slot and g as depicted in Fig. 6(a) for the case of the top-slot antenna. Similarly, Q L and f 0 for different values of θ slot and g for the side-slot configuration is shown in Fig. 6(b). As it can be seen, Q L and f 0 are altered by both the size of the slot and g. The Q ext for different tapping locations of the SMA connector h e is shown in Fig. 6(c).
For demonstration purposes, a first-order top-slot filtenna with fractional bandwidth (FBW) of 8%, center frequency f 0 of 4.8 GHz and maximum gain of 5.2 dBi is designed with Q ext = 18.8 and Q L = 17.1 and using (1), (2). For the desired response, an annular slot with r slot = 14 mm, θ slot = 180°, and tapping location h e = 3 mm needs to be selected. The simulated realized gain and |S 11 | are plotted in Fig. 4(c).

C. Higher Order Filtenna Design
To demonstrate scalability of the concept to higher order gain functions, a second-order top-slot BPF response is designed with: FBW of 8%, f 0 = 4.8 GHz and maximum realized gain of  5.2 dBi. To realize this gain function, Q ext and Q L in Fig. 7(a) are calculated using (1) (Q ext = 18.8, Q L = 17.1) and the inter-resonator coupling is calculated as k 12 = m 12 FBW = 0.06. The normalized coupling coefficients are m 01 = 0.85 and m 12 = 0.75. Having specified the desired coupling element values, the geometric parameters r slot = 14 mm, θ slot = 180°, and h e = 3 mm are specified using the design curves in Fig. 6(a)-(c). Next, the distance d between of the two resonators is specified so that the desired k 12 in the CRD in Fig. 5(a) is obtained using the design process in [24]. The relationship of k 12 as a function of d is provided in Fig. 6(d), which indicates that a d equal to 24 mm needs to be selected.
Following the aforementioned design method, a second-order and a third-order side-slot filtennas can be designed, as shown in Fig. 8(a) and (b). In the case of the third-order filtenna, the coupling coefficients are k 12 = k 23 = 0.064, Q ext = 16.1 for FBW = 8% and f 0 = 4.6 GHz, The normalized coupling coefficients are m 01 = 0.88 and m 12 = m 23 = 0.80.
To facilitate monolithic integration using SLA-based manufacturing, nonradiating holes need to be opened on the outer walls of the filtenna so that the Cu-platting solution can flow into the cavity and allow for the internal walls to be metalized. Their size needs to be carefully selected so that they do not affect the Q U of the coaxial resonators [24]. Fig. 9 depicts the surface H-field distribution of the second-order top-slot filtenna at 4.8 GHz. As it can be seen, it remains the same after adding the nonradiating holes, which indicates that those holes do not affect its EM performance.

III. EXPERIMENTAL VALIDATION
For proof-of-concept validation purposes, a second-order topslot filtenna, a second-order and a third-order side-slot filtenna with f 0 of 4.8, 4.65, and 4.65 GHz, and an FBW of 8% were prototyped using SLA 3-D printing. For monolithic integration, the filtenna model (Fig. 10) needs to be appropriately oriented with an angle of 30°within the DAM system and support structures need to be carefully added so that the model is fully supported and the supports can be readily removed after manufacturing. A commercially available Cu-plating process was employed that creates a Cu thickness of 50 μm, which is 20× larger than the skin depth at the designed frequency. Fig. 10 depicts the CAD model for the SLA manufactured second-order top-slot filtenna before and after Cu-plating. The RF measured and EM simulated |S 11 | and maximum gain are depicted in Fig. 11 Table I provides a comparison of the proposed coaxial resonatorbased filtenna with other state-of-the art 3-D filtennas using SIW and CNC manufacturing. As it can be seen, the proposed concept is the only one using 3-D coaxial cavity resonators and the very first one that is fabricated monolithically through SLA. Overall, a fair agreement has been achieved between the simulation and measurement, successfully validating the filtenna concept.

IV. CONCLUSION
The design, manufacturing, and measurement of a new class of highly miniaturized coaxial cavity-resonator-based filtennas is reported. Size compactness is achieved through the: 1) RF co-design of two RF components, 2) use of capacitively loaded coaxial resonators, and 3) monolithic integration enabled by SLA process. A second-order top-slot, a second-order and a third-order side-slot filtenna prototypes were manufactured and tested, exhibiting a fair agreement with EM simulations.