SIW-Based Devices

Abstract

With the development of microwave wireless communication technology, microwave passive frequency selectivity circuits are developing toward multi-function and miniaturization. Substrate-integrated waveguide (SIW) devices have good development prospects in miniaturization and integration. The research on devices based on SIW is lack unified theoretical combing. This chapter describes the novel circuits based on substrate-integrated waveguide, including filter, power divider, switch, phase shifter, diplexer, and crossover. This chapter describes the design methods and special properties of these circuits. The display of these circuits is expected to give some inspiration to your research.

1. Introduction

Microwave devices are generally designed using metallic waveguides and planar technology. Traditional metallic waveguide is low insertion loss (IL), and high-quality factor (Q value), but the components are bulky and nonplanar, which cannot be integrated into RF integrated circuits. Planar technology (microstrip transmission lines or coplanar waveguides) could be easily integrated, but the insertion loss and Q value are worse. At higher frequencies (>30 GHz), planar structures are prevented to apply due to high transmission losses [1].

To tackle the above-mentioned problems, the concepts of SIW are proposed [2, 3]. The SIW is considered a quasi-waveguide, developed by two parallel copper sheets with rows of conducting cylinders embedded in a dielectric substrate on each side to connect the two plates, as shown in Figure 1. When the parameters of the design conform following rules, SIW can be equivalent to a conventional rectangular waveguide [4].

Figure 1.

Three-dimensional view of SIW.

$$s>d$$E1

$$s/{\mathrm{\lambda }}_c<0.25$$E2

$$l/k_o<1\times {10}^{-4}$$E3

$$s/{\mathrm{\lambda }}_c>0.05$$E4

Where s is the gap between the adjacent vias, d is the diameter of a via, ${\lambda }_c$ is the cutoff wavelength, $l$ is the total loss, and $k_o$ is the wave number in a vacuum.

In this way, a rectangular metallic waveguide filled with a dielectric material is constructed in planar form, the most key advantage, thus making a complete integration with other planar transmission-line circuits on the same substrate. Meanwhile, SIW structures maintain most of the benefits of classical metallic waveguides. Thus, SIW is a suitable choice for less lossy, compact, and simple microwave and mm-wave systems.

After 20 years of development, SIW has been used to design all kinds of passive circuits, such as filters, power dividers, switches, phase shifters, diplexers, and crossovers, and presents high performance. At the same time, many techniques have been developed in recent times. This chapter provides a brief, yet necessary, understanding of current research on the SIW components.

2. SIW filters

Traditional planar transmission structures, such as microstrip and coplanar waveguide (CPW), have been widely used in filter design. Compared with these two structures, the substrate-integrated waveguide technology has the advantages of high Q coefficient, low cost, and high integration, and is widely used in the design of narrowband and broadband band-pass filters. The substrate-integrated waveguide technology can realize a variety of filters with different characteristics, including multi-band filters, wide stop-band filters, and reconfigurable filters. In the early days, SIW filters focused on the realization of circuit functions, and reports on the miniaturization of SIW filters appeared around 2005 [5]. There are various miniaturization technologies for SIW filters, and miniaturization has become the main trend in the development of SIW filters. This section introduces the application of SIW technology in filters from five aspects.

2.1 Conventional substrate-integrated waveguide filters

The design of the SIW bandpass filter can follow a similar approach to the design of gas-filled waveguide filters, or it can be based on the coupled matrix approach. In [6], a three-pole Chebyshev filter is designed through the inductive column synthesis technology, and its insertion loss and return loss are more than 1 dB and 17 dB, respectively. This SIW-based filter reduces size, weight, and cost significantly.

A novel wide-stopband SIW filter using an angular cavity is performed in Ref. [7]. For each corner cavity, two transmission zero points can be introduced, and the position of each transmission zero point can be easily controlled by adjusting the appropriate geometric dimensions. Figure 2 shows the configuration of the proposed SIW filter.

Figure 2.

Configuration of proposed SIW filter [7].

2.2 Multilayer substrate-integrated waveguide filters

The emergence of multi-layer technology can extend the traditional two-dimensional planar structure to a three-dimensional direction, greatly increasing the flexibility and freedom of filter design, and can increase the coupling mode. At present, the realization process of multi-layer structure mainly includes low temperature cofired ceramic (LTCC) process and multi-layer PCB process. The second part mainly introduces the multi-layer structure design of the filter applied to the substrate-integrated waveguide.

The first multilayer technology is the substrate-integrated-folded waveguide (SIFW), which was proposed by Grigoropoulos et al. [8]. The use of substrate-integrated folded waveguides in filter design can greatly reduce the overall circuit size of the filter.

Figure 3 shows a filter based on a SIFW resonator design [9]. The filter was developed using LTCC technology to achieve miniaturization. Resonators of the same layer, such as resonator 1 and resonator 4, resonator 2 and resonator 3, are horizontally coupled through the corresponding sensing window of the through hole. The different layers of resonators, such as Resonator 1 and Resonator 2, Resonator 3 and Resonator 4, are vertically coupled through a slot in the common wall between the two layers. Compared with the traditional planar direct coupled waveguide filter, the area of the filter can be reduced to 26.3%.

Figure 3.

Configuration of proposed SIFW filter [9].

The second multilayer technology is the ridged substrate-integrated waveguide (RSIW). As shown in Figure 4 [10], RSIW is the introduction of a longitudinal metal ridge into a classical waveguide without any degradation in RF performance. At the same time, the use of ridge waveguides can increase the bandwidth by 37%.

Figure 4.

The dimensional structure of the RSIW [10].

2.3 Fractional mode substrate-integrated waveguide filters

The electric field distributions in the conventional SIW, half-mode SIW (HMSIW), quarter-mode SIW (QMSIW), eighth-mode SIW (EMSIW), sixteenth-mode SIW (SMSIW), and 32-mode SIW (TMSIW) cavities at the dominant (TE101) mode are shown in Figure 5. The symmetry planes A-A1, B-B1, C-C1, and D-D1 are considered to be magnetic walls, and the remaining planes are electric walls. The full-mode SIW cavity can be cut into half-modes by cutting along the symmetry line B-B1, as shown in Figure 5b. It is obvious that the fundamental mode remains unchanged and the volume is reduced by half. The HMSIW cavity can be cut into quarter modes along the O-A1 line, as shown in Figure 5c. Compared with HMSIW, QMSIW achieves a 50% reduction. The QMSIW can be reduced to the EMSIW by cutting the QMSIW cavity along the O-D1 line, as shown in Figure 5d. Similarly, the electric field distribution is unchanged. SMSIW and TMSIW can be obtained by cutting the EMSIW cavity along the O-E line and O-F line, respectively, as shown in Figure 5e and f. The resonant frequency and electric field distribution of all sub-modes remain the same.

Figure 5.

Electric field distributions of (a) full-mode SIW, (b) HMSIW, (c) QMSIW, (d) EMSIW, (e) SMSIW, (f) TMSIW [11].

It is worth noting that since the topology of the fractional mode technique is inherently open-structured, there will be undesired radiation leakage, which can sometimes significantly reduce the quality factor. To solve this problem, some methods are applied. Taking quarter mode as an example, the shielded QMSIW is used to design the filter in Ref. [12]. As shown in Figure 6, the shielded quarter-mode resonator is made by placing two rows of metal through holes in the opening, which are called shield walls. Adding shielding walls can partially block the propagation of waves along the surface, reducing the efficiency of the radiating side of the structure. Compared with the conventional QMSIW, the shielded QMSIW exhibits better performance.

Figure 6.

Electric field distribution for the fundamental mode of the SIW resonators. (a) Conventional QMSIW cavity, (b) shielded QMSIW cavity [12].

2.3.1 Half-mode SIW (HMSIW) filters

The half-mode substrate-integrated waveguide was proposed by Wei Hong et al. in 2006 [13]. This waveguide structure is obtained by cutting the SIW so that the symmetry plane along the transmission direction is equivalent to the magnetic wall. As shown in Figure 7, the HMSIW is 50% smaller in size while keeping the electric field distribution unchanged. Since the volume of HMSIW is reduced by half and the propagation characteristics remain unchanged, it has attracted the attention of many scholars.

Figure 7.

Dominant field distribution in HMSIW and SIW [13].

A compact dual-band filter based on HMSIW is shown in Figure 8 [14]. The dual-band filter is designed by the quasi-TEM mode and the TE102 mode of the novel HMSIW resonator. The resonant frequency of TE102 can be adjusted by the slot line, which has little effect on the performance of the quasi-TEM mode. To improve the out-of-band rejection, the source negative coupling structure is applied. The measurement results show that the first and second passbands are mainly concentrated at 2.41 and 3.51 GHz, the relative bandwidths are 10.8% and 6.4%, respectively, and the minimum insertion losses of the two passbands are 1.45 and 1.74 dB, respectively.

Figure 8.

Configuration of the proposed dual-band filter [14].

2.3.2 Quarter-mode SIW (QMSIW) filters

The authors of Ref. [15] proposed the bandpass filter based on QMSIW for the first time. The geometries of the two filters are shown in Figure 9. The bandpass filter consists of cascaded QMSIW cavities, and its overall size is greatly reduced. For verification, two filters are designed. The prototype I is designed at the center frequency of 5.85 GHz with a fractional bandwidth of about 14%. Prototype II is designed at the center frequency of 5.5 GHz with a fractional bandwidth of about 26%.

Figure 9.

The geometries of the two filters. (a) The prototype I, (b) the prototype II [15].

The authors of Ref. [16] reported a QMSIW for designing a series of bandpass filters. Quarter-mode SIW cavities can be connected in various ways, such as sharing a side or sharing a corner, with some metal vias removed to allow cavity coupling. The authors introduce novel filter topologies based on the proposed two different techniques, side-coupling, and corner-coupling. Taking the filter shown in Figure 10 as an example, the resonator is laterally coupled by removing some through holes in the common wall. By using a plus-single topology, it can be used to implement filters with any number of poles. Figure 11 shows the filter composed of the new coupling. By using a coplanar resonator and quarter-mode cavity together, a three-pole filter with a center frequency of 4 GHz is designed, and its relative bandwidth is 16%.

Figure 10.

Configuration of the side-coupled four-pole filter [16].

Figure 11.

Configuration of the corner-coupled four-pole filter [16].

2.3.3 Eighth-mode SIW (EMSIW) filters

The QMSIW can be further bisected into two parts along the symmetrical plane with one electric wall and two magnetic walls forming the EMSIW.

The authors of Ref. [17] reported a systematic research campaign on the EMSIW filter. The authors conducted a series of studies on eighth-mode filters. Compared with the traditional SIW cavity, the one-eighth mold cavity has a compact size. This paper first explores the four feeding modes of the eighth mode, then designs a variety of bandpass filters for the eighth mode utilizing electrical coupling and magnetic coupling, and finally designs two bandpass filters. The first filter takes the form of a combination of an eighth mode and a coplanar waveguide. Through the mutual coupling between the TE101 mode of the two EMSIW cavities and the quasi-TEM mode of the folded coplanar waveguide, a band-pass filter with a center frequency of 9.1 GHz is finally designed. The second filter adopts a multi-layer process, and the upper and lower substrates are coupled through two rectangular slots. Compared with filter 1, this filter has more transmission zeros. The proposed EMSIW BPFs have the merits of being compact in size and high selectivity. The geometry of the two filters is shown in Figures 12 and 13.

Figure 12.

Geometry of the triple-order EMSIW BPF with CPW [17].

Figure 13.

Geometry of the triple-order EMSIW BPF with QMSIW cavity [17].

2.3.4 High-mode SIW (SMSIW and TMSIW) filters

A miniaturized bipolar bandpass filter based on a SMSIW was investigated [18]. The geometry of the filter is shown in Figure 14. The filter operates in fundamental mode TM01. Under the same resonant frequency, the size of the SMSIW cavity is only 1/16 of that of the traditional SIW cavity. To suppress the higher-order mode TM02, a rectangular slot is etched in the ground plane below the transmission line. The overall size of the filter is compact and has good out-of-band performance.

Figure 14.

Geometry of the proposed filter [18].

A TMSIW bandpass filter is presented in Ref. [19]. The geometry of the filter is shown in Figure 15. By adjusting the position of the feed and the coupling distance, the frequency position of the passband can be selectively achieved. The TM0n0 mode has a good unloaded quality factor in addition to providing different frequency options. In addition, the perturbation of the arc slot can effectively tune the coupling and achieve high selectivity.

Figure 15.

Geometry of the triple-order EMSIW BPF with QMSIW cavity [19].

2.4 SIW filters loaded with complementary Split-ring resonators (CSRR)

As an electromagnetic metamaterial, CSRR is often used in the miniaturization of filters. CSRR is an electrical resonator that can be excited by an axial electric field, which is proposed by Pendry [20]. In Ref. [21], two HMSIW filters loaded with the stepped-impedance CSRRs are presented. The geometries of the two filters are shown in Figure 16. By loading both CSRR and stepped-impedance complementary split-ring resonator (SICSRR) into the two sides of the HMSIW cavity, a dual-band resonator is formed. By adjusting the size of CSRR and SICSRR, the position of the two passbands can be flexibly adjusted. Compared with other dual-band filters, the HMSIW-SICSRR filter size is reduced by 48.5% and 50%, respectively.

Figure 16.

Geometry of the proposed filters. (a) Loaded with CSRR and SICSRR, (b) loaded with SICSRRs [21].

Novel bandpass filters are reported by using HMSIW loaded with CSRR and a capacitive metal patch [22]. The geometry of the proposed filter is shown in Figure 17. Based on this combined structure, independent double passbands can be generated. In addition, the external quality factor and coupling factor of the two frequency bands can be adjusted independently.

Figure 17.

Geometry of the triple-order EMSIW BPF with QMSIW cavity [22].

2.5 Tunable SIW filters

A new method to design tunable filters is shown in Figure 18 [23]. The proposed filter consists of a conventional SIW resonator and an additional via. The vial contains an open-loop slot in the top metal wall, where the size, location, and orientation of the open-loop slot determine the current and field distribution paths, which in turn controls the tuning range, resulting in an 8% tuning range.

Figure 18.

Geometry of the proposed filter [23].

A tunable filter based on microstrip patch resonators is proposed [24]. The overall layout of the filter is shown in Figure 19. The designed filter is similar to the SIW structure, that is, an electrical wall composed of through-holes is placed around the microstrip patch to reduce radiation, thereby reducing the insertion loss of the filter. In addition, the varactor diode is selected as the tuning element because of its small size, fast tuning speed, and large tuning range. The final designed filter covers six different states from 2.0 GHz to 2.53 GHz, and its insertion loss is 1.68 dB–3.90 dB.

Figure 19.

Geometry of the proposed filter [24].

Table 1 shows the performance parameter of the filters following different technological approaches.

Device	Technology	Size; (λg * λg)	Center frequency (GHz)	FBW (%)	Insertion loss (dB)	Return loss (dB)
[7]	SIW	—	20	27.5	2.78	16
[9]	SIFW	0.3*0.88	30.2	11.3	3.7	13.5
[14]	HMSIW	0.45*0.67	2.41/3.51	10.8/6.4	1.45/1.74	20/20
[15]	QMSIW	0.5*0.5	5.85	14	2	13
[15]	QMSIW	0.5*0.5	5.5	26	1.2	13
[17]	EMSIW-CPW	0.33*0.33	9.1	22	0.9	23
[17]	EMSIW-Multilayer	0.32*0.32	9.1	19.8	1.3	20
[18]	SMSIW	—	2.05	7	1.2	16
[19]	TMSIW	0.468*0.198	7.26	6.8	1.5	—
[21]	HMSIW+ SICSRR	0.046*0.046	5.51/8.83	9.4/5	1.3/1.8	18/18
[21]	HMSIW+ SICSRR	0.045*0.045	5.52/8.81	9/5.4	1.3/2	23/14

Table 1.

Comparison of different filters.

3. SIW power divider

Power divider (PD) can be used in phased array antennas, multiplexers, power amplifiers, and so on. Substrate-integrated waveguides have attracted attention in the design of power dividers due to their strong anti-interference, smaller radiation loss, and higher quality factor (Q value). There are three main research directions on SIW power dividers in recent years. Firstly, SIW PDs are developed with higher isolation and lower insertion loss. The main isolation networks include one or more resistors and the microstrip-based network [25, 26]. Secondly, the filtering power dividers (FPD) are proposed to combine the power divider and filter to realize the miniaturization of the circuit [27, 28, 29]. Thirdly, ultra-wideband and wideband isolation PDs are fabricated [30, 31]. The main technical parameters of the power divider include insertion loss, return loss, isolation between ports, amplitude balance and phase balance, etc.

3.1 PD with high isolation

Wilkinson PD is used to achieve a high degree of isolation between output ports. An architecture of Wilkinson PD is based on fixed-width SIW lines with isolation resistors as shown in Figure 20 [25]. The resistors are formed of various surface-mounted parallel resistors to reduce the effect of SIW-resistance discontinuity. The present topology shows enhanced isolation and adaptation in a wide bandwidth.

Figure 20.

The novel Wilkinson PD.

Another SIW PD with high output isolation is shown in Figure 21 [26]. The power divider is constructed by using two substrate layers and three metal layers. Five microstrip lines are placed on the top metal layer and coupled with the SIW through three-slot lines located on the middle metal layer (common ground). Three resistors are used to improve the output isolation. This device has input return loss and insertion loss greater than 17 dB and 1.6 dB, respectively. However, this structure increases the complexity of the process.

Figure 21.

Structure of proposed PD (a) 3D view. (b) Top view [26].

3.2 Filtering power divider

Loading a resonator or multiple resonators on the power divider can realize the dual function of filtering and power division. Figure 22 shows a triple-band FPD and a quad-band FPD [27]. Two unequal CSRRs are engraved face to face on the top of the SIW rectangular cavity to achieve two operating frequencies. An outer pair of U-shaped slots (USS) or two pairs of USS is engraved on the bottom of the SIW to achieve the third or the fourth operating frequency. The fabricated FPD presents good isolation and return loss.

Figure 22.

Prototypes of Triple-band and quad-band PDs [27].

Compared to the above FPD, the balanced SIW FPD is relatively less explored, especially for dual-band applications. Figure 23 shows a single-layer dual-band SIW FPD which has three bisected substrate-integrated cavities [28]. By adjusting the position of metal vias, each cavity is designed to not only form three-pole dual passbands but also attain high in-band common-mode (CM) rejection. An isolation resistor is loaded across each output pair to attain high isolation easily.

Figure 23.

Fabricated dual-band balanced SIW FPD.

3.3 Ultra-wideband and wideband isolation PD

Another filtering method is to load the slot line. As shown in Figure 24, two-slot lines are loaded on the SIW to get a third-order filtering response [30]. The isolation network consisting of the coplanar waveguide feeding line of Port 1, the interdigital capacitor, and the resistor is embedded in SIW, which expands the bandwidth of isolation and achieves lower insertion loss and good isolation. The proposed SIW FPD has isolation higher than 21 dB from 5.59 to 6.4 GHz.

Figure 24.

Proposed SIW FPD [30].

Table 2 shows the performance parameter of the power dividers following different technological approaches.

Device	Technology	Bandwidth; (GHz)	Isolation (dB)	Insertion loss (dB)	Return loss (dB)
[25]	Isolation network	8.85–10.6	20	3.3	15
[26]	Isolation network	9.2–10.8	20	1.6	17
[27]	FPD	3.47/4.73/6.31; 3.34/4.82/6.17/7.66	14; 12	3.8/4.85/4.06; 3.5/3.95/4.26/4.85	20; 19
[28]	FPD	28/39	20/14.9	1.7/2.2	—
[29]	FPD	10.64–11.84/; 13.37–13.97	16.1	1.4/2.2	17.8/; 18.4
[30]	Wideband isolation	5.59–6.4	21	1.05	20

Table 2.

Comparisons with different power dividers.

4. SIW switch

The working principle of the SIW switch is to control the transmission of waves in the needed SIW path by various control methods, mainly including electronic control switches, magnetic control switches, and mechanical switches. Because of the ease of integration and low cost, PIN diodes are used as tunable devices loading on the surface of SIW for most electronic control switches. There have been many examples of designing switches using the difference of characters in diode forward and reverse bias [32, 33, 34, 35]. Magnetic control switches are mostly loaded with ferrite [36, 37]. The permeability of the ferrite slabs loaded in the SIW cavity is tuned by an external transverse magnetic field, so that the cutoff frequency of the SIW mode is changed and the switching function is realized. Magnetic control switches have higher power handling capability and longer service life. The mechanical switch has the deficiencies of a short operation lifetime and slow switching speed [38]. However, it can avoid the problem of parasitism. A switch mainly focuses on three factors: insertion loss, return loss, and isolation.

4.1 Electronic control switch

An electronic control single-pole single-throw (SPST) switch is shown in Figure 25. Firstly, make a narrow slot in the SIW top wall, then, the diodes are capacitively coupled to the slot by metal pads on a very thin insulating layer. With the pin diodes being switched between on-state and off-state, the propagation could be transformed between two different types of modes. When the diodes are turned on, the top slot is closed and the waveguide becomes an ordinary SIW. When the diodes turn off, the electric field and propagation characteristics of the structure can be seemed as HMSIW, as long as the slot has an appropriate size [32]. The difference in cut-off frequency between SIW and HMSIW gives a switching bandwidth from 3 to 4.7 GHz. The fabricated switch is presented with isolation of 50 dB and 3 dB IL in 3–3.5 GHz.

Figure 25.

Structure of switched waveguide [32].

In addition, PIN diodes are also used in a single pole double throw (SPDT) switch. Inductive posts with rectangular slots are embedded in the SIW to increase the impedance of the SIW, which can disturb the incident signal flow. The PIN diodes and bias networks are placed to control the operating mode of the inductive posts, as shown in Figure 26. The proposed SPDT switch has measured isolation(S31) of greater than 10 dB, isolation(S32) of greater than 15 dB, and an IL of less than 2.55 dB from 8.24 to 10.36 GHz [33]. However, SIWs are narrow at higher frequency ranges and cannot use switched vias. Thus, this method is no longer suitable at higher millimeter-wave frequencies.

Figure 26.

Structure of the SIW SPDT switch [33].

Different from the above internal through-hole design, a SIW switch with a resonant slot is made innovatively [34]. The proposed switch in Figure 27 has an IL of less than 1.3 dB and isolation of more than 10 dB in the switchable frequency band. The resonant slot vanishes when the PIN diode is in the forward-biased state, while the resonant slot presents a high impedance load on the SIW when the PIN diode is in the reverse-biased. This novel device can be particularly suitable for higher millimeter-wave frequency due to its simple structure.

Figure 27.

A switch with a resonant slot [34].

4.2 Magnetic control switch

There are many attempts to magnetically control using ferrite loading. A way to load is shown in Figure 28, which consists of rectangular ferrite slabs loaded on the sidewall slots of SIW [36]. By applying external magnetic bias on the ferrite slabs, the cutoff frequency of the waveguide can be changed. The proposed switch has an IL of less than 1 dB and an isolation of 20 dB.

Figure 28.

Ferrite-loaded SIW switch.

Figure 29a shows another magnetic control SIW switch with loading ferrite. A dc magnetic bias is applied to control the propagation capability. A Ferrite slab is placed on the surface of the SIW. The surfaces of the ferrite slab are covered with silver except for the one touching the substrate. As shown in Figure 29b, when the switch is off-state, e-fields present a strong transverse field displacement, causing the cutoff in TE10 mode. This way is easy to be handled by PCB technology and achieves better performance.

Figure 29.

(a) The proposed SIW switch (b) simulated e-fields within the switch(off-state) [37].

4.3 Mechanical switch

The following introduces a high-isolation SPDT SIW mechanical switch in V and W bands. Four via holes with the ring and cross are used in each SIW path to block the waves, as shown in Figure 30. A copper pad is put on top of the via holes. When a copper pad touches the SIW, via holes connect the copper pad and top layer copper. Via holes begin to block waves. When lifting the copper pad, via holes are invisible, and the switch is off-state. The proposed switch has a more than 50 dB isolation [38].

Figure 30.

(a) Mechanical SPDT switch (b) schematic of the proposed SPDT.

Table 3 shows the performance parameter of the switches following different technological approaches.

Device	Technology	Bandwidth (GHz)	Isolation (dB)	Insertion loss (dB)	Return loss (dB)
[32]	PIN diode / electric control	3–4.7	50	3	10
[33]	PIN diode/ electric control	8.24–10.36	15	2.55	10
[34]	PIN diode/ electric control	20–25	10	1.3	10
[35]	PIN diode/ electric control	4.6–5.3	23	1.35	15
[36]	Ferrite/magnetic control	9.5–10	20	—	10
[37]	Ferrite/magnetic control	9.5–11	27	1.6	15
[38]	Mechanical switch	50–75; 75–105; 105–110	50; 50; 50	3.5; 7.5; 10	10; 10; 10

Table 3.

Comparisons with different switches.

5. SIW phase shifter

Phase shifters are very important devices for signal phase adjustment and antenna beam steering. A SIW phase shifter has the advantages of low insertion loss, high power handling capability, and excellent immunity to electromagnetic interference (EMI). The classical design methods of phase shifters based on SIW are mainly to control the equivalent dielectric constant of the transmission line [39, 40, 41], make self-compensating wideband shifters [42, 43, 44], and take advantage of the slow-wave effect [45, 46]. Controlling the equivalent dielectric constant of transmission is the most widely used method. There are many designs such as using air-filled slabs and loading extra metallic posts. The designed phase shifters have good performance in a narrow band, but all of them are sensitive to frequency, which has poor performance in a wide bandwidth. Self-compensating technology is used to realize wide-band phase shifters. A phase shifter that takes advantage of the slow-wave effect can get a more compact size. In addition, dynamic tunable phase shifters are also popular. The main criteria for SIW phase shifter include insertion loss, operating bandwidth, phase deviation, and power handling capabilities. Next, we will introduce several phase shifters with excellent performance.

5.1 Self-compensating wideband shifter

A phase shifter just using a delay line cannot be used to achieve the phase shift in a wide frequency range because it will cause a large phase deviation. To solve this problem, we can use the phenomenon that the delay line and the unequal-width structure to make an opposite dispersive phase shift and realize a wideband compensation phase shifter [42]. As shown in Figure 31, a broadband SIW self-compensating phase shifter is proposed. The differential topology reduces the frequency sensitivity. The fabricated phase shifter achieves a 90° phase shift from 25.11 to 39.75 GHz, and the phase imbalance between the two paths is within 0.2 dB and 2.5 dB, respectively [43].

Figure 31.

Structure of the proposed SIW self-compensating phase shifter.

Another compensation method is slotting compensation [44]. By taking advantage of the effects of the slots on the value and slope of the phase shift, slotted SIW is proposed to produce the phase shift slopes which are opposite to the microstrip delay lines to achieve a wide bandwidth phase shifter. The concept of such a technique is verified by designing −90°, −45°, 45°, and 90° phase shifter prototypes, as shown in Figure 32.

Figure 32.

Photograph of five prototypes [44].

5.2 Dynamic tunable phase shifter

By incorporating switches into the circuits, a phase shifter can achieve multiple phase changes easily. According to the different types of switches, controllable phase shifters are divided into electrically controlled tunable phase shifter, magnetically controlled tunable phase shifter, and mechanical tunable phase shifter. Compared with the fixed phase shifter, the variable phase shifter is more flexible to meet the requirements of the system, which facilitates the evolution of phased array techniques strongly. Early SIW multiple phase shifters were designed with a multi-phase channel side by side, which increased the size [47]. At present, most of the adjustable phase shifters are single-channel structures, and the phase shift is adjusted by switching the equivalent dielectric constant of the transmission line in the channel.

A cylindrical metal post inserted in SIW is equivalent to a T-network high-pass filter [48]. When a PIN diode is buried within a SIW, the diode shows the same behavior as the metal post [49]. The difference is that diode on or off can change the value of the phase shift on the SIW. Accordingly, a new approach to designing a dynamic phase shifter is proposed. A bias circuit is placed on top of the SIW to control the diode states, as shown in Figure 33.

Figure 33.

Proposed SIW controllable phase shifter (a) structure. (b) Biasing circuit of the NIP diodes [49].

Different from diode tuning, a SIW phase shifter which can be reconfigured using liquid metal is presented [50]. A series of holes that can be filled or emptied of liquid metal is placed on the SIW. The holes with filled liquid metal can be seen as a wall to block the passage of energy. Figure 34 illustrates the operating paths in different states. By using the different electrical lengths among the three paths, the device achieves coarse steps of phase change, from 0° up to 180°, in steps of 60°. Utilizing reactive loading (placing a single hole in each path), the phase shifter can achieve a phase shift in steps of 10°, from 0° up to 180°.

Figure 34.

The three operating states of the proposed phase shifter [50].

5.3 Slow wave effect

Based on the slow-wave effect, an excellent phase shifter is fabricated using a CSRR-loaded SIW, as shown in Figure 35. CSRR provides the slow-wave effect in the TE10 mode of SIW. The phase velocity reduction resulting from the slow-wave effect provides a phase shift [45]. By controlling the phase velocity and the physical length of the transmission line, different degrees of phase shift can be achieved. The CSRR-loaded SIW realizes a direct conversion between the fast wave and the slow wave, which makes the phase shifter integrate with other devices easily.

Figure 35.

Proposed a slow-wave structure.

Table 4 shows the performance parameter of the phase shifters following different technological approaches.

Device	Technology	Bandwidth (GHz)	Phase shift (。)	Amplitude imbalance (dB)	Return loss (dB)
[39]	Equivalent dielectric constant	26.5–40	43 ± 6	0.23 ± 0.2	12.5
[40]	Equivalent dielectric constant	26.5–40	40.8 ± 2.6	—	10
[41]	Equivalent dielectric constant	20–32	45 ± 5; 90 ± 5	0.1	12
[42]	Self-compensating	25.11–39.75	90.5 ± 2.5	0.2	12
[44]	Self-compensating	21.2–32.7	−90 ± 5; −45 ± 2.5; 45 ± 2.5; 90 ± 5	—	15
[45]	Slow-wave effect	7–10	90 ± 5	—	15
[46]	Slow-wave effect	10.5–11.5	180 ± 10	—	12
[49]	Tunable/PIN diode	10	0–45	—	10
[50]	Tunable/liquid metal	10	0–180	—	15

Table 4.

Comparisons with different phase shifters.

6. SIW diplexer

The diplexer is a three-terminal device used for transmitting and receiving signals. One port is connected to the antenna and the other two are used for transmitting and receiving signals, respectively. The two processes are completed independently without any effect on each other. Thus, isolation, insertion loss, cost, and size should be paid more attention to. Diplexers can be divided into the following classes according to the shunting method of signals. Traditional diplexers mainly consist of two filters in different frequency bands and a T-junction [51, 52, 53, 54], which increase the design complexities and the insertion loss. In recent years, there are two main ways to replace the T-junction. Firstly, multiple dual-mode resonators (MDMR) are used to replace the T-junction [55]. At the same time, good isolation is achieved through the orthogonality of the dual modes. However, the design flexibility of these diplexers is limited by the dual-mode coupling mechanisms. The second way is employing the scheme of a common dual-mode resonator (CDMR) and multiple single-mode resonators [56, 57, 58, 59]. This method solves the problem of flexibility and achieves circuit miniaturization. Three structures are shown in Figure 36. In addition, there are also some other hot spots and trends, such as the miniaturization design of diplexers based on sub-mode SIW, balanced SIW diplexers, and the tunable design of SIW diplexers.

Figure 36.

(a) T-junction scheme (b) multiple dual-mode resonators scheme (c) a common dual-mode and multiple single-mode resonators scheme.

6.1 T-junction structure

A SIW diplexer with CSRRs loaded is shown in Figure 37 [52]. A direct microstrip line insert-feeding is adopted on the top working as a T-junction. The left and right structures are two bandpass filters operating in different frequency bands.

Figure 37.

Topology of the proposed SIW CSRR diplexer.

6.2 Multiple dual-mode resonators

Figure 38 shows a six-port balanced SIW diplexer by using dual-mode resonators [55]. Because the electric field responds differently to each mode and the characteristics of differential signals, there is better isolation performance between two channels. The electric field distribution of the lower channel and the higher channel is shown in Figure 39. Additionally, differential input can enhance common-mode rejection and increase the stability of circuits.

Figure 38.

Structure of the balanced diplexer.

Figure 39.

Electric field distributions in (a) lower channel. (b) Higher channel [55].

6.3 A common dual-mode and multiple single-mode resonators

A wide passband SIW diplexer with a common half-mode dual-mode resonator and multiple single-mode resonators is shown in Figure 40. The first half-mode dual-mode resonator is used to flexibly allocate fractional bandwidths. The rest of the resonators are single-mode couplings, which almost be mapped independently. Figure 41 shows the electric field distributions in this diplexer at 3.5 and 5.0 GHz, respectively. Different channels are dominated by different modes.

Figure 40.

Structure of the fabricated prototype [58].

Figure 41.

Electric field distributions in the proposed SIW diplexer [58].

The schematic of a tunable SIW diplexer with various single-ended and balanced ports is shown in Figure 42 [59]. Resonator A is a common dual-mode resonator, Resonator B and A form the lower channel, while Resonator C and resonator A form the upper channel. The bottom of each resonator is covered with a silver disk, which attaches to a piezo disk. Piezoelectric actuator is used to tune the operating passband. The single-end structure and balanced structure are both fabricated and have state-of-the-art performance.

Figure 42.

The schematic of the tunable diplexer [59].

Table 5 shows the performance parameter of the diplexers following different technological approaches.

Device	Technology	Lower channel/; Upper channel (GHz)	Isolation; (dB)	Size; (mm × mm)	Insertion Loss; (dB)
[51]	T-junction	24.31–25.66/; 25.96–27.34	50	—	1.95/2.09
[52]	T-junction	4.66/5.8	30	17.5 × 14	1.6/2.3
[54]	T-junction	24.3–25.65/; 25.95–27.35	50/40	—	2/2.5
[55]	MDMR	8.62–8.98/9.17–9.63	35	47.45 × 62.5	2.2/2.3
[56]	CDMR	8/9	40	31.13 × 31.13	2.86/3.04
[57]	CDMR	11.8–12.2/; 13.24–14.26	27	33.6 × 23.4	1.34/1.41
[58]	CDMR	3.44–3.56/5.41–5.59	10.1/21.6	87.2 × 46.4	2.77/2.55

Table 5.

Comparisons with different diplexers.

7. SIW crossover

As a microwave element, the crossover is not only able to make two signals cross-transmit and maintain high isolation, but also usually appears in the array antenna beamforming network, and is an important part of the Butler array. The use of SIW technology to design the crossover has become a research hotspot for many scholars. According to the circuit structure, the existing SIW crossovers can be divided into two categories, one is the traditional cascaded 3-dB directional coupler structure, and the other is based on the coupling theory, utilizing the orthogonality of the modes in the SIW resonators to achieve isolation of two signals.

7.1 Cascaded 3-dB directional coupler structure

The planar structure of cascading two 3-dB SIW couplers is the earliest SIW crossover. As shown in Figure 43, the proposed crossover has a measured return loss of lower than 13-dB, isolation of better than 20 dB, and an insertion loss of less than 0.5 dB within 5% bandwidth [60]. However, cascading inevitably brings the problem of excessive circuit size. To reduce the size, a two-layer cascaded 3-dB coupler is proposed based on the two-layer folded SIW directional coupler [61]. The electric field and magnetic field are coupled through the middle slot.

Figure 43.

The crossover at 60 GHz [60].

7.2 SIW resonators

For the SIW resonator, on the one hand, using signal resonance to cascade multiple cavities on one channel can realize signal filtering; on the other hand, using the orthogonality of the resonant modes in the cavity, the signals in two channels can be isolated. The combination of the two can realize the dual-function integration of filtering and crossover in one device, which will reduce the complexity of the circuit and help reduce signal transmission loss.

In a general filter crossover circuit structure, according to the duality principle, each channel needs three cavities, and the design and implementation of the crossover require at least five SIW cavities, as shown in Figure 44.

Figure 44.

Traditional filter crossover.

The commonly used dual-mode SIW resonators are mostly TE201 and TE102 modes. As shown in Figure 45, the electric field distributions of the two modes are orthogonal to each other. The strongest electric field of one mode is the weakest of the other mode. The four ports are arranged according to the strength of the electric field, and the two signals are transmitted in the TE201 and TE102 modes respectively, which can realize the cross isolation of the signals. On this basis, a filter crossover circuit with cascading five dual-mode resonators is designed [62].

Figure 45.

Distributed electric fields of TE102 and TE201 in rectangular SIW cavity.

From the perspective of reducing circuit area and heat dissipation loss, the industrial field has higher and higher requirements for device performance. SIW crossovers are developing in the direction of better filtering performance, smaller size, and higher integration. Three ways to improve development are summarized below.

7.3 Filtering crossover

In the traditional filter crossover circuit with five cavities, to further reduce the circuit size, the four peripheral dual-mode cavities are changed to single-mode resonators whose fundamental mode is TE101 [63, 64], and the cavity arrangement is different, but the signal filtering selectivity of the crossover is enhanced in [64]. By increasing the number of cavities, and flexibly arranging single-mode and dual-mode resonator cavity arrangements, a variety of filter crossovers with different center frequencies and bandwidths are realized [65]. The circuit only retains one dual-mode resonator, and four CPW half-wave resonators are cascaded at the four ports, which further reduces the circuit size [66]. In addition, square SIW cavities with TE102 and TE201 modes are used as building blocks, and multi-channel filtering crossover is designed and fabricated by rationally arranging coupling ports and feed ports [67].

Figure 46 shows the electric field distribution of a filter crossover circuit with a single resonator. The circuit is based on a four-mode resonant SIW rectangular resonant cavity, and the middle through holes and slots are used to adjust the electric field distribution. Compared with the above-mentioned circuits with multiple resonant cavities, on the one hand, the structure only uses a single resonant cavity, which further reduces the size, and on the other hand, the operating frequencies of the two channels in the above structure are the same, but the center frequencies of the two channels in this structure are different, which realizes the dual-frequency crossover of the two-channel signals and expands the application range.

Figure 46.

Electric field distribution in the crossover: (a) TE103 mode, (b) TE104 mode, (c) TE201 mode, (d) TE202 mode [68].

In [69], the circuit structure is still in the traditional cross form, but the SIW evanescent-mode (EVA) resonator is applied. The coupling coefficient between the resonators is controlled by diodes, which makes the coupling coefficient of the adjacent cavities 0, and the diagonal cavities are not 0, thus completing the isolation of two signals. Piezoelectric actuators are utilized to form a tunable filtering channel, and the application range of the crossover is expanded.

7.4 Crossover with balanced structure

The above-mentioned crossover circuits are all used for single-ended circuits, but considering the reduction of signal noise interference, the crossover circuits applied to balanced circuits are also studied. The multi-layer circuit topology is the main structure.

Based on the traditional filter crossover circuit, five SIW cavities are cascaded as shown in Figure 47. The CM is suppressed in this structure. Under differential mode (DM), the crossover part is realized by orthogonal modes TE102 and TE201.

Figure 47.

The crossover circuit layout [70]. (a) 3D figure; (b) Planar graph.

Another balanced structure is shown in Figure 48. The 1–3 channels and the 2–4 channels are arranged in different layers, and the signals are transmitted through the middle slots. TE202 and TE204 in a single rectangular cavity are orthogonal to each other to complete the cross isolation of the two signals.

Figure 48.

3D diagram of crossover [71].

Table 6 shows the performance parameter of the crossovers following different technological approaches.

Device	Technology	Bandwidth (GHz)	Isolation (dB)	Insertion loss (dB)	Return loss (dB)
[60]	3-dB Directional coupler	5	20	0.5	13
[65]	Filtering crossover	11.57–12.43	21.8	1.61	18.5
[67]	Filtering crossover	8.83	23	2.29	20
[68]	Filtering crossover	7.11/8.77	14.65	2.7/2.57	13.1
[70]	Balanced structure	10.2	30	3.2	—

Table 6.

Comparisons with different crossovers.

8. Conclusions

As an easy-to-integrate planar circuit, the SIW technology lays the foundation for the design and implementation of high-performance, planar, low-cost, and easy-to-integrate microwave passive circuits. This chapter describes the novel circuits based on substrate-integrated waveguide, including filter, power divider, switch, phase shifter, duplexer, crossover, and so on.