Circularly polarized antenna

Abstract

A circularly polarized antenna includes a plurality of radiation units configured in a rotational symmetry manner, and a feed network connected to the plurality of radiation units. Each radiation unit comprises a patch element and a short circuit wall which is connected with the patch element and is suitable for enabling the patch element to be short-circuited. Different from the traditional sequential phase feed, the sequential phase feed only depends on the physical length of the delay line to realize the phase change, and the reactance loading feed line ingeniously utilizes the equivalent capacitance and inductance to change the phase. The method does not need to depend on a long delay line, and stable phase difference can be realized in a wide frequency band.

Description

Technical Field

[0001] The present invention relates to an on-chip antenna, in particular to an on-chip circularly polarized antenna. Background Art

[0002] As one of the core technologies in integrated circuit manufacturing, complementary metal-oxide-semiconductor (CMOS) technology has been widely adopted and rapidly gaining popularity due to its advantages, including low power consumption, high speed, process maturity, and good scalability. With the continuous improvement of transistor cutoff frequency and maximum oscillation frequency, CMOS technology is gradually becoming the preferred solution for terahertz (THz) integrated circuits, sparking significant interest in system-on-chip (SoC) and antenna-on-chip (AoC) solutions. Circularly polarized (CP) antennas play a vital role in many applications, offering multiple advantages, such as effective mitigation of multipath interference, eliminating the need for polarization alignment, and reducing system susceptibility to ghost targets and receiver interference. Leveraging CMOS technology and the extremely short wavelength of THz frequencies, CP antennas can be seamlessly integrated with RF and digital circuits on the same chip. The combination of THz technology, CP antennas, and CMOS processes offers new research and development opportunities in areas such as high-resolution radar imaging, short-range communications, automotive systems, biomedicine, nondestructive testing, and inter-satellite communications.

[0003] However, the design of on-chip CP antennas faces many challenges. Due to the limited thickness of the silicon dioxide layer in the back-end-of-line (BEOL) process, the axial ratio (AR) bandwidth of traditional on-chip CP antennas is typically limited to 5-6%. To address this issue, some researchers have attempted to utilize silicon-based structures in CMOS processes to increase antenna bandwidth, achieving AR bandwidths greater than 16%. However, the use of silicon also brings a variety of negative effects, including reduced radiation efficiency, surface waves, distortion of the radiation pattern, and interference with active devices in the front-end-of-line (FEOL) process.

[0004] Continuous phase feeding technology is a promising alternative approach for extending antenna bandwidth without using silicon. This approach eliminates the need for CP characteristics within the antenna element itself. Its AR bandwidth is primarily determined by the sequential phase feeding network, significantly simplifying the design of wide AR bandwidth. This technique has been widely used in monostatic and quasi-monostatic radars, typically in conjunction with a duplexer or coupler to distinguish transmit and receive signals. However, due to the requirement for large antenna arrays and complex feeding networks, this approach often results in excessively large antenna size. Phase differences between adjacent output ports of the feeding network are typically achieved using delayed transmission lines. For example, a 90° phase difference requires a quarter-wavelength transmission line. Therefore, this approach is generally not cost-effective, which presents disadvantages in terms of both cost and miniaturization. Furthermore, due to the thickness of the oxide layer, the impedance bandwidth is typically narrow. Even with an optimized feeding network, the overlap between the impedance bandwidth and the AR bandwidth is typically less than 10%. In summary, the development of on-chip CP antennas faces numerous limitations, and research progress is significantly less than that of on-chip linearly polarized (LP) antennas. Therefore, designing compact and broadband on-chip CP antennas has become an important issue that needs to be addressed urgently. Summary of the Invention

[0005] Therefore, in one aspect, the present invention provides a circularly polarized antenna comprising: a plurality of radiating elements arranged in a rotationally symmetrical manner, and a feed network connected to the plurality of radiating elements. Each radiating element comprises a patch element, and a short-circuit wall connected to the patch element and adapted to short-circuit the patch element.

[0006] Preferably, the circularly polarized antenna further comprises a ground layer. For each radiating element, the short-circuit wall is located between the corresponding patch element and the ground layer, and is connected to the patch element and the ground layer.

[0007] More preferably, the short-circuit wall comprises a plurality of layers.

[0008] Most preferably, the short-circuit wall comprises ten layers.

[0009] In a variant of the preferred embodiment, the patch element of each radiating element has a generally rectangular shape.

[0010] Preferably, for each radiating element, the short-circuit wall is substantially connected to a first side of the patch element, and the feeding network is substantially connected to a second side of the patch element, the second side being opposite to the first side.

[0011] Additionally or alternatively, for each radiating element, a projection of the short-circuit wall on the patch element has a generally T-shaped shape. The short-circuit wall has a first section connected to a first side of the patch element, and a second section perpendicular to the first section and extending from the first section toward the center of the patch element.

[0012] Preferably, for each radiating element, the feeding network comprises a corresponding output arm connected to the radiating element, wherein the output arm extends in a direction substantially parallel to the first section.

[0013] In a variant of the preferred embodiment, the first section of the short-circuit wall has a length that is the same as the dimension of the first side of the patch element.

[0014] In another variation of the preferred embodiment, the patch element of each radiating unit forms a rectangular notch at a corner of the rectangular shape, and the feeding network is connected to the patch element near the rectangular notch.

[0015] In another variation of the preferred embodiment, for each radiation element, a corresponding hole is formed on the ground layer, the shape of the hole corresponds to the shape of the short-circuit wall, and the hole partially receives the short-circuit wall.

[0016] In another variation of the preferred embodiment, the circularly polarized antenna includes four of the aforementioned radiating elements, and the input phase difference between adjacent radiating elements is 90°.

[0017] In another aspect of the present invention, a circularly polarized antenna is provided, comprising a plurality of radiating elements arranged in a rotationally symmetrical manner, and a feed network connected to the plurality of radiating elements. The feed network includes a core portion and a plurality of output arms. Each output arm corresponds to one of the plurality of radiating elements, and the plurality of output arms are coupled to the core portion. The plurality of output arms are arranged in a rotationally symmetrical manner.

[0018] Preferably, the core portion has a generally square annular shape.

[0019] More preferably, the core portion has unequal widths along the extension direction of the core portion.

[0020] In a variation of the preferred embodiment, the core portion comprises a parallel plate capacitor.

[0021] In another variant of the preferred embodiment, a parallel plate capacitor is located in front of one of the output arms along the signal transmission path of the circularly polarized antenna.

[0022] In another variation of the preferred embodiment, the parallel plate capacitor includes a first portion located in the same layer as the rest of the core portion, and a second portion located parallel to and below the first portion.

[0023] In another variation of the preferred embodiment, the output arm is located on the same layer as the first portion of the parallel plate capacitor.

[0024] In another variation of the preferred embodiment, the circularly polarized antenna includes four radiating elements, the input phase difference between adjacent radiating elements is 90°, and the circularly polarized antenna includes three parallel plate capacitors corresponding to the three radiating elements along the signal transmission path of the circularly polarized antenna.

[0025] In another variant of the preferred embodiment, the core section is fed by a feeder line located in the same layer as the core section.

[0026] In another variation of the preferred embodiment, the feed line extends along a first direction, and the core portion is short-circuited by a short-circuit line extending along a second direction perpendicular to the first direction.

[0027] Preferably, the short-circuit line passes through a short-circuit hole formed at the center of the core portion and is connected to a ground layer below the feed network.

[0028] Embodiments of the present invention therefore provide a broadband on-chip circularly polarized antenna with high compactness, wide bandwidth and small size. The antenna involves significant improvements in the feed network and the radiating element. A reactance-loaded sequential-phase feed network is proposed to generate a phase difference. Unlike traditional sequential phase feeding that relies solely on the physical length of the delay line to achieve phase change, the reactance-loaded feed line cleverly utilizes equivalent capacitance and inductance to change the phase. This approach eliminates the reliance on long delay lines and achieves a stable phase difference over a wide bandwidth. The reactance loading concept is universal and is applicable not only to sequential phase feeding, but also to any design that requires phase delay, such as hybrid devices and couplers. Antennas according to embodiments of the present invention can be used in future 6G wireless communications, providing enhanced spectrum and energy efficiency. In addition, they can also be used in the Internet of Things (IoT), sensing, imaging and short-distance high data rate communications.

[0029] Other features and aspects of the present invention will become clear from the following description of specific embodiments and the accompanying drawings. Any feature described in this disclosure in relation to one aspect or embodiment may be combined with any other feature described in this disclosure in relation to any other aspect or embodiment as needed. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Exemplary embodiments of the present invention will now be described with reference to the accompanying drawings, in which:

[0031] Figure 1 FIG. 1 is a top view of a broadband on-chip circularly polarized antenna according to an embodiment of the present invention.

[0032] Figure 2 yes Figure 1 Side view of the antenna in Figure 2 Also shown are the passivation layer and the silicon substrate.

[0033] Figure 3a yes Figure 1-2 Exploded view of one of the radiating elements in the antenna.

[0034] Figure 3b yes Figure 3a A top view of the radiating element in FIG, showing the dimensions of the radiating element.

[0035] Figure 4a yes Figure 1-2 Exploded view of the antenna’s feed network.

[0036] Figure 4b yes Figure 4a A top view of the feed network in Figure 1, showing the dimensions of the feed network.

[0037] Figure 5a Shown in full TM 01 Heat map of a patch antenna resonating in one mode.

[0038] Figure 5b Shown in TM 01 Heat map of the patch antenna shown above resonating in half mode.

[0039] Figure 5c Shows Figure 1-4b Heat map of the radiating element in the antenna, which is in TM 01 Resonance occurs in the half mode.

[0040] Figure 6a Shows that in full TE 201 Thermal image of the cavity where the mode resonates.

[0041] Figure 6b Shows a quarter cavity at a quarter TE 201 Heat map of resonances in the modes.

[0042] Figure 6c Shows Figure 1-4b The radiating element of the antenna is in a quarter of the TE 201 mode resonance.

[0043] Figure 7a It shows that when f = 390 GHz, Figure 1-4b Time-averaged E-field size distribution of the radiating element of the antenna on the xoy cross section.

[0044] Figure 7b The time-averaged E-field size distribution of the above-mentioned radiation unit on the xoy cross section is shown when f=440 GHz.

[0045] Figure 8 The simulated reflection coefficient of the radiating element is shown.

[0046] Figure 9a Shows l s From l s = Simulated reflection coefficient of different sizes from 0μm to 60μm (reference value: l s =60μm).

[0047] Figure 9b Shows different sizes of l p From l p =Simulated reflection coefficient from 185μm to 215μm (reference value: l p =200μm).

[0048] Figure 9c Shows different sizes w p (From w p =114μm to 126μm) (reference value: w p =120μm).

[0049] Figure 10a Shows Figure 4a-4b Simplified equivalent circuit of outputs 1 and 2 of the feeding network.

[0050] Figure 10b Shows Figure 4a-4b Simplified equivalent circuit of outputs 3 and 4 of the feed network.

[0051] Figure 11a Simulation output showing the magnitude of the feed network.

[0052] Figure 11b Simulation output showing the phases of the feeding network.

[0053] Figure 12 Shows Figure 1-4b Simulated and measured S-parameters of the antenna, before and after TRL (through-reflection line) calibration.

[0054] Figure 13a Shows Figure 1-4b Measured and simulated axial ratio curves of the antenna in .

[0055] Figure 13b Shows Figure 1-4b Measured and simulated gain curves of the antenna in .

[0056] Figure 14a Shows Figure 1-4b Measured and simulated YOZ plane modes of the antenna at 360 GHz.

[0057] Figure 14b Shows Figure 1-4b Measurement and simulation of the antenna in the YOZ plane mode at a frequency of 380 GHz.

[0058] Figure 14c Shows Figure 1-4b Measurement and simulation of the antenna in the YOZ plane mode at a frequency of 400 GHz.

[0059] Figure 14d Shows Figure 1-4b Measured and simulated YOZ plane modes of the antenna at 420 GHz.

[0060] Figure 14e Shows Figure 1-4b Measured and simulated YOZ plane modes of the antenna at 440 GHz.

[0061] Before explaining any embodiments of the present invention in detail, it should be understood that the invention is not limited in its application to the details and arrangement of components of the embodiments described below or illustrated in the accompanying drawings. The invention is capable of other embodiments and of being carried out in various ways. Furthermore, it should be understood that the phraseology and terminology used herein are for descriptive purposes only and are not to be construed as limiting. DETAILED DESCRIPTION

[0062] Figure 1-4b Figure 1 shows a broadband on-chip circularly polarized antenna according to a first embodiment of the present invention. This antenna is a 425GHz broadband on-chip circularly polarized antenna with a highly compact sequential phase feed network. The antenna utilizes equivalent capacitance and inductance to change the phase, thereby achieving a very compact feed network and a wide AR bandwidth. In addition, the radiating element in the antenna is designed as a unique semi-TM 01 and a quarter TE 201 mode, resulting in significant size reduction and wider impedance bandwidth.

[0063] like Figure 1 As shown, the antenna comprises two main components: a reactance-loaded sequential phase feed network comprising a core 22 and four output arms 24 (connected to the core 22), and a plurality of radiating elements 20a, 20b, 20c, and 20d. The four radiating elements 20a, 20b, 20c, and 20d are arranged rotationally symmetrically around the core 22, as are the four output arms 24. Radiating element 20a is the first radiating element in the antenna's signal transmission path (i.e., the path from the feed line 28 to the terminal stub 44), followed by radiating element 20b, and so on. The four radiating elements 20a, 20b, 20c, and 20d have similar constructions, and each of the radiating elements 20a, 20b, 20c, and 20d is connected to the feed network via its respective output arm 24.

[0064] The core portion 22 has a generally square-ring shape, but it is not a completely closed square shape, as will be described in detail later. The four output arms 24 are connected to different sides of the square shape via connecting lines 36. The feed network ensures that each pair of consecutive radiating elements 20a, 20b, 20c, 20d has a predetermined input phase difference, thereby achieving sequential phase. Specifically, as Figure 1 The antenna shown includes four radiating elements 20a, 20b, 20c, and 20d, with the input phase difference between adjacent radiating elements being 90°, resulting in sequential rotation of circular polarization. All components of the antenna are placed above a ground plane 26, with the feed network short-circuited to the ground plane 26, as described below. The antenna is fed by a feed line 28, which is connected to on-chip circuitry (not shown) for integration or to a GSG (ground-signal-ground) probe (not shown) for testing.

[0065] like Figure 2 As shown, for example, during the BEOL process, the antenna is arranged in the first SiO2 layer 40, and the antenna structure mainly uses four metal layers: TM1 for the radiating elements 20a, 20b, 20c, 20d, M8 for the connecting line 36, M9 for the feeding network and other transmission lines, and M1 as the ground layer 26. However, it should be noted that there are eight layers between M1 and M9, including M8, but Figure 2 For simplicity, the other layers (M2-M7) are not shown. TM1 is the topmost layer, and M1 is the bottommost layer (i.e., ground layer 26). In other words, the antenna has ten layers, including TM1 and M1. Above the first SiO2 layer 40 is a passivation layer 38, a pre-protective layer added to the antenna surface to prevent impurities from entering and damaging the antenna. In actual processing, depending on the specific technology, the first SiO2 layer 40 is typically composed of multiple sublayers. Below the ground layer 26 is a second SiO2 layer 54. All metal layers from M1 to TM1 are embedded in a silicon dioxide (SiO2) dielectric layer. The passivation layer 38 and the second SiO2 layer 54 define the passive region 56 of the on-chip device, which, as described above, can be manufactured using a back-end-of-line (BEOL) process. Below the second SiO2 layer 54 is a silicon substrate 42, which can be a general substrate for any components on the on-chip device other than the circularly polarized antenna. The substrate 42 defines the active region 58 of the on-chip device and can be manufactured using, for example, a feed-end-of-line (FEOL) process.

[0066] Now see Figure 3a-3b The structure of each radiation unit 20a, 20b, 20c, 20d is described. As mentioned above, these radiation units 20a, 20b, 20c, 20d have similar internal structures, so Figure 3a-3bThe structure shown is applicable to all radiating elements 20a, 20b, 20c, and 20d. The radiating elements 20a, 20b, 20c, and 20d in this embodiment are designed using a standard 65-nanometer CMOS process, which consists of the above-mentioned ten metal layers, from the bottom metal layer M1 to the top metal layer TM1. In each radiating element, there is a patch element 30 and a T-shaped short-circuit wall 32. The projection of the short-circuit wall 32 on the patch element 30 is generally "T"-shaped. The patch element 30 serves as the main body of the radiating element and is designed to have the largest size in TM1. The patch element 30 is generally rectangular and short-circuited on the side. A rectangular notch 34 is formed at one corner of the rectangular shape (close to the corner of the feed network), so that the shape of the patch element 30 is an incomplete rectangle.

[0067] The corresponding output arm 24 of the feed network is coupled to the patch element 30 near the rectangular notch 34. The rectangular notch 34 is mainly used in conjunction with the output arm 24 and provides greater design freedom to achieve appropriate coupling and better impedance matching between the output arm 24 and the patch element 30. The patch element 30 is located on the TM1 layer, while the output arm 24 is located on the M9 layer directly below the TM1 layer. Therefore, the enlarged portion 24a of the output arm 24 partially overlaps with the portion 30a of the patch element 30 in the top view (see FIG. Figure 3b ), and the metal through hole 46 is configured in the portion 30a to electrically connect the output arm 24 with the patch element 30, as shown Figure 3a This feed design is to better excite the two different modes, and the output arms 24 with segments of different widths are to better match the impedance.

[0068] The rectangular notch 34 (and the output arm 24) are located on the first side of the rectangular shape of the patch element 30. In contrast, the T-shaped short-circuit wall 32 is generally arranged on the second side opposite to the first side. In particular, the first section 32a is aligned with the edge of the second side of the patch element 30, which is opposite to the rectangular notch 34. The first section 32a has the same length as the second side of the patch element 30, as shown in FIG. Figure 3b As shown in FIG. 3 . The second section 32b of the short-circuit wall 32 extends from approximately the middle point of the first section 32a, and the second section 32b is perpendicular to the first section 32a. The second section 32b extends from the second side of the patch element 30 to the center of the patch element 30. Figure 3a-3b As shown, the output arm 24 and the first section 32a are parallel to each other.

[0069] The short-circuit wall 32 is composed of ten metal layers from M1 to TM1, with vias (not shown) stacked between them. Therefore, the short-circuit wall 32 grounds one side of the patch element 30. The structure of the short-circuit wall 32 is similar to the sidewall of the on-chip substrate integrated waveguide (SIW) structure. Given the extremely small side length and narrow spacing in the CMOS process, the gap energy leakage between adjacent vias is extremely small. Since the short-circuit wall 32 has a portion in the M1 layer, Figure 3a The corresponding portion of the ground layer 26 shown in FIG. 1 forms a hole 26 a that matches the “T” shape of the shorting wall 32 , so as to receive a portion of the shorting wall 32 in the M1 layer in the hole 26 a .

[0070] Figure 3b The dimensions of the radiating element are shown. For the 425GHz broadband on-chip circularly polarized antenna, the optimized parameters of the radiating element are as follows: p =125μm,l p =200μm, w s =20μm,l s =

[0071] 60μm, w g =10μm,l g =90μm, x f =118μm,y f =11μm, w1=8μm, w2=15μm, l2=58μm.

[0072] Now refer to Figure 4a-4b Will describe Figure 1-2 The structure of the antenna feed network in the traditional sequential phase feed network mainly relies on the physical length of the delay line to achieve phase change. For example, a quarter-wavelength delay line is required to achieve a 90° phase difference. This reliance on physical length greatly increases the size of the traditional sequential phase feed network. In contrast, Figure 1-4b The reactance-loaded sequential phase feed network in the antenna uses equivalent capacitance and inductance to achieve phase shift, significantly reducing the size. This approach paves the way for the design of highly compact and efficient circularly polarized antenna systems.

[0073] like Figure 4a-4bAs shown, three metal layers are used in the feed network, including a ground layer 26 (represented as the M1 layer in these figures), an M9 layer for most of the components in the feed network, and an M8 layer for connecting lines 36. Two types of vias are also used, including metal vias (M8-M9, not shown) connecting the M8 and M9 layers, and short-circuit vias 60 that short-circuit the terminal stub 44 to the ground layer 26 (M1-M9) at the center of the feed network (i.e., the center of the antenna). Within the short-circuit vias 60 are short-circuited lines 48 for short-circuiting the core portion 22, extending perpendicularly to the feed line 28. The M8-M9 vias electrically connect the output arms 24 to their respective connecting lines 36.

[0074] As mentioned above, the feed network mainly includes a square ring structure, namely the core part 22 and four output arms 24, all of which adopt a microstrip line structure. The core part 22 integrates multi-level microstrip lines with varying widths to facilitate power distribution and impedance matching. In other words, the width of the core part 22 is unequal when it extends in the square ring structure. Figure 4b As shown, the core portion 22 has five different widths in different parts. The feed line 28 is arranged along a first direction ( Figure 4b The vertical direction in the middle) extends to the core part 22 and is at right angles to the core part 22. The feed line 28 is located on the same layer as the core part 22. The geometry of the core part 22 from the feed line 28 to the terminal stub 44 can be described as: turn left 90° - turn right 90° - turn right 90° - turn right 90° - turn right 90°. The feed line 28 is parallel to the fourth section 22a of the core part 22, which forms a side of the square adjacent to the fourth output arm 24. The length of the fourth section 22a is different from the other sides of the square, and it forms a right angle with the terminal stub 44 at half the length of one side of the square. Therefore, the square ring shape of the core part 22 is not closed, and the terminal stub 44 is perpendicular to the feed line 28. In addition, each connecting line 36 is perpendicular to a side of the core part 22 adjacent to it, and is perpendicular to the output arm 24 corresponding to it. As shown Figure 4b As shown, not all connecting lines 36 have the same width.

[0075] In traditional designs, the output is directly connected to the ring structure, and the phase difference is achieved through the delay line inside the ring structure. Figure 4a-4bThe feed network shown uses another approach, where the first three output connection lines 36 along the signal transmission path are located on the M8 layer, separated from the core portion 22 on the M9 layer. The connection lines 36 are physically separated from the core portion 22 by a small gap between the M8 and M9 layers. Utilizing the thin interlayer separation inherent in the CMOS process, these gaps form parallel plate capacitors 50a, 50b, and 50c. Each of the parallel plate capacitors 50a, 50b, and 50c is located in front of a corresponding arm of the output arm 24 on the signal transmission path of the circularly polarized antenna. Figure 4a As shown, the first portion of each parallel plate capacitor 50a, 50b, 50c is located on the same layer as the rest of the core portion 22 and the output arm 24. The so-called first portion here is actually a portion of the core portion 22. In addition, there is a second portion that is parallel to and below the first portion. The capacitance value of the capacitor 50a, 50b, 50c can be adjusted by changing the size of the overlapping area between the first portion and the second portion. In this embodiment, the capacitance value of the capacitor 50a, 50b, 50c is adjusted by adjusting the size of the overlapping area between the first portion and the second portion. Figure 4b The l shown c1 、l c2 and l c3 The length of the capacitors is adjusted. These capacitors have a crucial influence on the energy coupled from the ring structure to the output end, thereby determining the output phase. Uniquely, the last connecting line 36 adjacent to the fourth section 22a is connected in parallel with the terminal short stub 44, and the terminal short stub 44 is grounded through the above-mentioned short-circuit through-hole 60. This last connecting line 36 utilizes the principle that a transmission line with a terminal short circuit will exhibit inductive characteristics when its length is less than a quarter wavelength. Therefore, the last connecting line 36 is electrically connected in parallel with an equivalent inductor 52 composed of the terminal short stub 44 and a portion of the fourth section 22a, and its inductance is mainly composed of the short-circuit line (l 18 ) is determined by the length of the parallel plate capacitor. Through careful design, these reactive elements (parallel plate capacitors and equivalent inductors) can be used as phase shifters to achieve the desired phase difference without using long delay lines. These passive loads, especially the parallel plate capacitors 50a, 50b, and 50c, cleverly utilize the multi-layer metal in CMOS technology, where the extremely thin distance between different metal layers is usually considered an obstacle to antenna and transmission line design. However, this becomes an advantage in designing parallel plate capacitors 50a, 50b, and 50c, because the closer the two metal layers are, the easier it is to design the capacitance value of the parallel plate capacitor to be larger. In 65nm CMOS technology, the gap between the M8 and M9 layers is only 0.8 microns. This tiny gap makes the design of capacitors 50a, 50b, and 50c easy.

[0076] Figure 4bThe dimensions of the feed network are shown. For a 425 GHz broadband on-chip circularly polarized antenna, the optimized parameters of the feed network are as follows: w1 = 4 μm, l2 = 20 μm, w2 = 8 μm, l3 = 30 μm, w3 = 4 μm, w4 = 8 μm, l5 = 22 μm, w5 = 2 μm, l6 = 26 μm, w6 = 4 μm, l7 = 16 μm, w7 = 8 μm, l8 = 18 μm, w8 = 2 μm, w9 = 8 μm, l 10 =48μm,w 10 =2μm,l 11 =44μm,w 11 =6μm,l 12 =22μm,w 12 =4μm,w 13 =8μm,l 14 =25μm,w 14 =2μm,l 15 =24μm,w 15 =3μm,l 16 =31μm, w 16 =3μm, w 17 =8μm,l 18 =52.5μm, w 18 =3μm, l c1 =8.5μm,l c2 =16.5μm,l c3 =21.5μm.

[0077] In describing Figure 1-4b After describing the physical structure of the antenna, the model analysis, design process and working principle of the antenna will be introduced. Among them, the radiating elements 20a, 20b, 20c, and 20d resonate in two different modes: half TM at low frequency and half TM at low frequency. 01 mode and quarter TE at high frequencies 201 mode. Both modes are not complete, but rather half or a quarter of a complete mode. This unique characteristic mainly stems from the realization of equivalent electric and magnetic walls. Figures 5a-5c The design process of the first mode of the radiating element is demonstrated. Figure 5a A conventional patch antenna is shown with a feed structure similar to Figure 1-4b The antenna in is the same, but the patch length in the x-direction is approximately twice as long. Figure 5a The antenna in TM 01 mode, which is the most common way to excite a patch antenna to emit side radiation. Due to the inherent characteristics of this mode, the potential in the middle of the patch is close to zero, equivalent to a virtual ground. Therefore, if a physical short-circuit wall is created in the middle of the patch and the patch size is halved, such as Figure 5b As shown, TM 01The resonant frequency of the mode should remain unchanged. At this stage, the antenna resonates at half TM 01 mode. This mode is often used to miniaturize antennas because the resonant frequency is the same but the size is halved. Figure 5b The short-circuit wall established in Figure 5c Another short-circuit wall (called Figure 3a-3b The second paragraph 32b) makes the whole structure look like the letter "T" rotated 90° counterclockwise, that is, Figure 1 、 3a -3b The shape of the short-circuit wall 32. The main function of this additional short-circuit wall is to 01 The resonant frequency of the mode is adjusted higher and closer to the second resonant mode, thereby expanding the impedance bandwidth. The reason for choosing an additional short-circuit wall instead of changing the patch size to control the semi-TM 01 mode, because changing the patch size will affect both resonance points at the same time. However, this additional short-circuit wall can independently control the first resonance mode, namely the semi-TM 01 mode. The subsequent simulation section will further discuss more details of the resonance control. It is worth noting that in the semi-TM 01 In the TM mode, most of the energy is fed into the patch through the via 46 at the end of the output arm 24 because 46 is strategically placed in the half TM 01 The peak energy point of the mode.

[0078] The design process of the second model is Figures 6a-6c Detailed presentation is given in . Figure 6a shows a completely enclosed metal cavity that is 201 According to the inherent electromagnetic distribution of this mode, the positions of the virtual electric (E) wall and the virtual magnetic (H) wall are easy to identify, such as Figure 6a As shown by the dotted line in the figure. A quarter of the metal cavity is extracted to form Figure 6b The quarter cavity shown. This quarter cavity retains the original physical E wall on its left side, but the other three sides are open. The metal-air interfaces of these three open sides can be equivalently regarded as magnetic walls. Given the consistent boundary conditions, the electromagnetic field distribution in this extracted quarter cavity is almost the same as that of the original full metal cavity. At this stage, the quarter cavity resonates at a quarter TE 201 In this design process, the metal-air interface is used to form an equivalent magnetic wall, which is the same as the principle of half-substrate integrated waveguide (HSIW). After adding the T-shaped short-circuit wall and feeding structure, the quarter cavity evolves into Figure 6c The radiation unit in Figure 3a It can be observed that the central short-circuit wall (corresponding to the second section 32b) is located exactly at one quarter of the TE 201mode, and thus has the least effect on the resonant frequency of this mode. Therefore, by manipulating the extension length of the central short-circuit wall, the semi-TM 01 mode resonant frequency, and one quarter TE 201 Mode is not affected.

[0079] Equivalent E and H walls are used in the radiating element design to create a compact structure that can resonate in two imperfect modes. Using these two modes extends the impedance bandwidth while maintaining a compact size, which is very beneficial for cost control and further integration with sequential phase feed networks.

[0080] The design of the radiation elements 20a, 20b, 20c, and 20d is simulated using the commercial software HFSS. The overall size of the element is no greater than 150 μm × 250 μm, which is equivalent to 0.21λ0 × 0.35λ0 at 425 GHz. Figure 8 As shown in Figure 2, the simulation predicts a -10dB reflection coefficient between 377 and 460GHz. Since the distance from M1 to TM1 is 8.8μm, the size of this component is extremely small, only 0.012λ0. 11 |The curve shows that there are two different resonances at 390GHz and 440GHz. At 390GHz and 440GHz, the time-averaged E field size distribution on the xoy cross section is as follows: Figure 7a-7b As expected, the antenna operates at half TM at lower frequencies. 10 Mode works like Figure 7a As shown, at higher frequencies, the TE is one quarter 201 Mode works like Figure 7b This is consistent with the above theoretical analysis.

[0081] To demonstrate the resonance control, several key parameters of the radiating element are swept, and the results and discussion are presented. First, the length of the center short-circuit wall (l s ) is scanned from 0 microns to 60 microns. The results are as follows Figure 9a As shown. It can be seen that l s The first resonance is significantly affected, while the second resonance is only slightly deviated. s The longer the first resonant frequency is, the higher the first resonant frequency is. This is because the semi-TM 01 The mode is directly related to the length of the patch in the x-direction, and its change will lead to a corresponding change in the resonant frequency. s will not actually change the patch length, but it will move half the TM 01 The virtual ground position in the mode is used to shorten the surface current path in disguise. At the same time, the central short-circuit wall is located at one-quarter of the TE 201mode, so the effect on the resonant frequency of this mode is small. Therefore, by changing l s The first resonance can be manipulated individually, while the quarter TE 201 The pattern is barely affected.

[0082] Then, the patch length in the y direction (l p ) was scanned from 190 μm to 210 μm in 5 μm increments. Figure 9b As shown. It was observed that the patch length l p The change of l will affect the first and second resonances, which is manifested as the frequency of the two resonance points increases with l p Given that TE 201 The mode is related to the length and width of the cavity. The larger the cavity size, the lower the resonant frequency. For the first resonance, in theory, the mode is only related to the patch width w p However, due to the unique feeding mechanism used in this design, the patch length l p The change of will also affect the first resonance point.

[0083] Finally, the patch width (w p ) on the reflection coefficient. Figure 9c The parameter sweep results are shown. It can be seen that the width of the patch also affects the two resonances. The increase in patch width is associated with a decrease in the two resonant frequencies. Given that TM 01 and TE 201 The dependence of the mode on the patch width is well understood. From the parameter sweep analysis above, it can be seen that both the length and width of the patch affect both resonant frequencies. Therefore, bandwidth adjustment can be achieved by varying the length of the center short-circuit wall. By bringing the first resonance closer to the second, the impedance bandwidth can be expanded.

[0084] Since the radiating element has multiple polarization directions, it cannot be used alone to achieve LP or CP radiation. However, applying sequential phase feeding can convert the final radiation into CP, regardless of the inherent polarization of the radiating element.

[0085] Next, we will discuss Figure 1 、 4a -4b The working principle of the sequential phase feed network. In the design of circularly polarized antennas, the sequential phase feed network plays two main roles: ensuring uniform power distribution and maintaining a fixed phase difference between adjacent output terminals. From the perspective of power distribution, Figure 1 、 4a The feed network in -4b can be viewed as three power dividers and a through line. For ease of reference, Figure 1 、 4a - The four output arms 24 in 4b are Figures 10a-10b The first output arm 24 along the signal transmission path is labeled as output 1, and so on. The output 1 section operates as a 1:3 power divider: one-quarter of the feed energy passes through capacitor 1 ( Figure 4a The parallel plate capacitor 50a in the output is coupled to output 1, and the remaining energy is transferred to the subsequent stage. Similarly, the output 2 part also acts as a 1:2 power divider. The power distribution ratio of the second part is 1:2, which is higher than the 1:3 of the previous stage. Therefore, a larger coupling coefficient is required, and thus a larger capacitance is required. Therefore, capacitor 2 (l c2， Right now Figure 4a The length of the parallel plate capacitor 50b) is greater than that of the capacitor 1 (l c1 ) is longer. Based on similar principles, the output 3 part acts as a 1:1 power divider, requiring capacitor 3 (i.e. Figure 4a The parallel plate capacitor 50c in the circuit has a higher coupling coefficient. Therefore, capacitor 3 is designed to have the highest capacitance and is therefore longer than the first two capacitors. Output 4 is directly connected to the square ring and connected to the equivalent inductor 1 ( Figure 4a 52) in parallel. This output differs from the previous ones in that it requires no power splitting. All energy should be directed to output 4, thus requiring the highest possible coupling coefficient. Achieving this with capacitors is challenging. Therefore, the final output uses an equivalent inductor rather than a capacitor to shift the phase.

[0086] In order to quantitatively understand how these reactive components change the amplitude and phase, a simplified circuit model of output 1 and output 2 is first given, as shown in Figure 10a As shown in Figure 1. In this model, all resistive losses are neglected. Output 1 and output 2 are connected to the square ring structure through series capacitors c1 and c2, respectively. Due to the open-circuit nature of the microstrip arms, they also have parallel parasitic capacitance c. g1 and c g2 The transmission lines and radiating elements after outputs 1 and 2 are simplified to loads with impedances of Z1 and Z2, respectively, and phases of θ1 and θ2, respectively. The multi-level microstrip line between outputs 1 and 2 is simplified to a transmission line with a characteristic impedance of Z B , the phase constant is β B , length l B The capacitor network S matrix of output 1 and 2 (in the gray dashed box) is [S 1 ] and [S 2 ], and their phase delays are θ s1 and θ s2 As mentioned above, point A can be regarded as a power divider with a ratio of 1:3. Assuming that the subsequent circuit is well matched, the impedance of the subsequent transmission line from point A is approximately Z B, and the impedance of output 1 can be expressed by the following formula.

[0087]

[0088] Given that c g1 The value of is very small, the power P1 of output 1 can be approximated as P s1 .

[0089]

[0090] As shown in (2), the energy of output 1 is proportional to the energy of capacitors c1 and c g1 is proportional to. Although c g1 represents the inherent parasitic capacitance of the open microstrip line, but it is difficult to adjust. However, by changing l c1 The value of c1 can be controlled, thus providing a method to control the power distribution ratio of point A. Based on the same principle, the power distribution ratio of subsequent outputs 2 and 3 can also be adjusted by adjusting the length l c2 and l c3 to control.

[0091] Regarding phase shift, the capacitance network S of output 1 21 and the resulting phase change θ s1 It can be expressed as follows:

[0092]

[0093] Similarly, the capacitor network S of output 2 21 and the resulting phase change θ s1 It can be expressed as follows:

[0094]

[0095] Therefore, the phase difference (θ 12 ) can be expressed as:

[0096]

[0097] Considering c g1 and c g2 The values of are relatively small and their influence can be ignored, so (5) can be rewritten as:

[0098]

[0099] The above formula shows that the phase difference between outputs 1 and 2 is related to the capacitances c1 and c2 and the length of the transmission line between the two output ports (l B). Due to the similar design, the equation also applies to the phase difference between output 2 and output 3. In the sequential phase feed design of four outputs, the phase between adjacent outputs must be 90° to generate CP waves. It is necessary to ensure that the difference between the first two terms in (6) is positive to shorten the transmission line (l B ) while maintaining a phase difference of 90°. Since “arctan” is a monotonically increasing function, to ensure that the difference between the first two terms in (6) is positive, c1 should be smaller than c2. The larger the difference between the two capacitors, the shorter the transmission line. Interestingly, this conclusion is completely consistent with the power division principle mentioned earlier. As mentioned earlier, output 1 and output 2 can be regarded as two power dividers with power division ratios of 1:3 and 1:2 respectively. According to formula (2), in order to achieve this increasing power division ratio, the values of capacitors c1 and c2 must also increase. These increasing capacitance values also ensure a positive difference between the first two terms in formula (6), which helps to achieve a positive phase shift and shorten the required transmission line length.

[0100] Then, Figure 10b The simplified circuit model of outputs 3 and 4 is given in Figure 2. Again, all resistive losses are discarded in this model. The transmission lines and radiating elements after outputs 3 and 4 are simplified to loads with impedances of Z3 and Z4, respectively, and output phases of θ3 and θ4, respectively. The multi-stage microstrip line between outputs 3 and 4 is simplified to a transmission line with a characteristic impedance of Z D , the phase constant is β D , length l D Output 3 is loaded with series capacitor c3 and parallel capacitor c g3 Output 4 and terminal short circuit (impedance is Z ind , the phase constant is β ind , length l 18 ) in parallel, equivalent to the inductor l ind The S matrices of the reactive networks of outputs 3 and 4 (in the gray dashed boxes) are [S 3 ] and [S 4 ], and their phase delays are θ s3 and θ s4 .

[0101] The power distribution principle of output 3 is similar to that of output 1 and will not be described here. The power expression of output 4 is as follows:

[0102]

[0103] Since the final stage requires that all energy be fed into output 4, rather than shorting it to ground, the first term in the denominator of (7) must be as close to zero as possible. This can be achieved by adjusting several parameters, including the short-circuit length l 18, output impedance Z4 and short-circuit characteristic impedance Z ind Short-circuit θ s4 The phase change caused can be expressed as follows:

[0104]

[0105] Therefore, the phase difference (θ 34 ) can be expressed as follows:

[0106]

[0107] Ignore c g3 , then (9) can be rewritten as follows:

[0108]

[0109] Due to the short-circuited microstrip line (l 18 ) is less than a quarter of the wavelength, so all terms in Equation (9) are positive. This shows that the introduction of capacitor C3 and equivalent inductance l ind The transmission line length between outputs 3 and 4 can be shortened D , while maintaining a 90° phase difference. In addition, c3 and l ind By adjusting the parallel plate capacitor (l c3 ) and short-circuit (l 18 ) length to control.

[0110] Figure 4a-4b The feed network in the simulation was simulated using the commercial HFSS software. The overall size of the feed network is approximately 140 μm × 140 μm, which is equivalent to 0.2λ0 × 0.2λ0 at 425 GHz. Figure 11a The simulated values of S parameters are shown. Impedance matching is particularly difficult in this design due to the use of reactive components. However, the final simulated |S 11 The curve remains below -15dB, indicating good impedance matching. Furthermore, the power level at the center frequency ranges from 7.1dB to 7.3dB, indicating even power distribution across the four ports. Figure 11b The simulated output phase is shown. At a center frequency of 425 GHz, the phases of the four output ports are 76 degrees, -16 degrees, -103 degrees, and -195 degrees, respectively. The phase difference between each pair of adjacent ports is approximately 90 degrees. Furthermore, these phase curves are nearly parallel to each other, indicating excellent sequential phase output over a wide frequency range. This characteristic is crucial for circularly polarized antennas, as maintaining a constant phase difference over a wide frequency range significantly improves the AR bandwidth.

[0111] Next, we will describe Figure 1-4bThe antenna shown in the figure was experimentally verified. To demonstrate its purpose, a prototype conforming to the proposed antenna was fabricated using a 65nm CMOS process. The entire structure consists of four main components: radiating elements, a sequential phase feed network, a GSG (Ground Signal Ground) pad, and a ground plane. The overall dimensions of the antenna are approximately 390 μm × 390 μm, equivalent to 0.55λ0 × 0.55λ0 at 425 GHz. Notably, the feed network is small enough to fit entirely within the gaps between the radiating elements, demonstrating its exceptional miniaturization. This is a significant advantage over conventional sequential phase feed networks, which typically require considerable space due to their long delay lines. To comply with design rules, numerous holes were introduced in the M1 and TM1 layers. Due to their extremely small size, these holes have limited impact on antenna performance. A GSG pad structure was designed in TM1 for connecting a 50Ω GSG probe with a 50μm pitch. The central pad (S) is directly connected to the microstrip line, while the two side pads (G) are grounded. However, the GSG pads are not designed for 50Ω impedance and may introduce unwanted parasitic effects, which may distort the S parameters and reduce gain. A TRL calibration is performed using a specially designed TRL kit to address this issue.

[0112] The S-parameters of the sample were measured using a wafer probe station. The measurement setup consists of the following components: (1) a GSG probe feeding the on-chip antenna; (2) a semi-automatic wafer probe station for fixing and positioning the probe; and (3) a probe for measuring the antenna reflection coefficient (S 11 ) Agilent vector network analyzer; (4) signal generator; (5) OML extender with a frequency range of 325 to 500 GHz. The calibration process consists of two steps: SOLT and TRL calibration. First, SOLT (Short-Open-Load-Thru) calibration is performed to eliminate errors from the VNA (vector network analyzer), cables and probes. It effectively moves the reference plane to the end of the probe tip. Then TRL calibration is performed to eliminate parasitic effects caused by the GSG pads. The second calibration step is to move the reference plane further to the AUT (antenna under test), thereby eliminating interference from the GSG pads and ensuring more accurate measurements. SOLT calibration uses the commercial SOLT calibration kit CS-15 calibration substrate from GGB Industries. For TRL calibration, a TRL kit was custom designed with the same GSG pad structure and manufactured using the same 65 nm CMOS process as the antenna. This consistency in design and manufacturing ensures maximum effectiveness of TRL calibration.

[0113] The radiation pattern is measured using a spherical wafer antenna measurement setup. The system consists of four main components: a transmitter, a feeder, a receiver, and a mechanical assembly. The transmitter utilizes a signal generator and a multiplier chain to generate signals in the frequency range up to 500 GHz. The generated THz signal is then fed to the AUT via a dedicated RF probe. The receiver is a standard linearly polarized THz waveguide horn antenna connected to a VDI mixer (WR 2.2, 325-500 GHz) at a far-field distance to capture the signal radiated by the AUT. The power received from the VDI mixer is then transferred to a spectrum analyzer for direct reading. A robotic arm can rotate within the AUT's y-axis (y-axis) plane to measure the radiation pattern. Due to probe obstruction and measurement setup limitations, only a portion of the y-axis plane pattern can be measured. Furthermore, the VDI mixer is mounted on a separate rotator to facilitate rotation around the mixer itself. In the final stage, the feeder is replaced with a standard horn antenna with known gain and an attached 90° waveguide elbow to calculate the AUT's gain according to the comparative method.

[0114] Given that the AUT is a CP and the THz horn antenna used for reception is a LP, direct measurement to obtain information such as the left-handed CP (LHCP) and right-handed CP (RHCP) gains is not feasible. In the case of testing a CP antenna with an LP antenna, a typical approach is to use phase-corrected measurements of φ and θ polarizations to calculate the RHCP and LHCP components. However, this approach requires precise phase information of the received signal, which is particularly difficult to achieve at THz frequencies. A simpler approach is to use a multi-amplitude component technique, which involves rotating the receiving antenna by 0° (horizontal), 45°, 90° (vertical), and 135°. By collecting amplitude data at these four different angles, the AR, RHCP, and LHCP can be calculated. The measurement setup presented here integrates a rotator at the receiving end, which is capable of precisely controlling the rotation angle of the receiving LP horn antenna. This greatly simplifies the application of the multi-amplitude component method. Once the received powers at these four angles are obtained (0° is P1, 90° is P2, 45° is P3, and 135° is P4), the tilt angle τ of the ellipse and the axial ratio of the antenna can be calculated using the following formula:

[0115]

[0116]

[0117] The left and right gains can be calculated using the following formulas:

[0118]

[0119] Among them, the total gain G total It can be calculated by the sum of two orthogonal gains, namely G0+G 90 or G45 +G 135 .

[0120] Figure 12 The simulated reflection coefficient and the reflection coefficient directly measured using the fabricated sample are shown. The original measured reflection coefficient deviates significantly from the simulation result, showing the average |S 11 The level is approximately -5 dB. After TRL calibration using a custom-designed TRL kit, the calibrated reflection coefficient (represented by the solid line with a "·") closely matches the simulation results, demonstrating an absolute impedance bandwidth of 125 GHz and a relative value of 29.4% over the frequency range of 357 to 482 GHz. However, the calibrated reflection coefficient curve exhibits some irregularities, including ripples and fluctuations, most notably between 350 and 420 GHz. These discrepancies may be partially attributed to deformation of the feed probe after multiple probing attempts. Other potential sources of these observed inconsistencies may include inherent coupling and interference caused by the complexity of the feed network, systematic errors in the measurement setup, and environmental interference.

[0121] Figure 13a The measured frequency-varying AR in the broadside direction is shown. The results are very consistent with the simulation results, except for an overall low-frequency shift of approximately 20 GHz. This indicates some discrepancies in the feed network between the simulation and measurement. This discrepancy is likely due to slight variations in the impedance of the parallel plate capacitors and equivalent inductors used in the feed network. Possible causes for this variation include differences between the complex CMOS process and the simulation, changes in the dielectric constant, and manufacturing errors. Despite the frequency shift, the antenna achieves an impressive 3 dB AR bandwidth (centered at 425 GHz). Because the antenna's impedance bandwidth fully encompasses the AR bandwidth, the antenna's impedance-AR bandwidth overlap reaches 19%. To the best of the inventors' knowledge, this represents the widest AR bandwidth on an on-chip circularly polarized antenna ever documented. Notably, the antenna is only 8.8 μm thick (only 0.012 λ0). The experimental results demonstrate that the proposed sequential phase feed network delivers superior performance across a wide frequency band, even in an extremely thin and compact form factor. This demonstrates that the proposed feed network outperforms conventional delay-line-based feed networks in terms of size, performance, and form factor. Most importantly, the concept of reactive loading is universal and applicable not only to sequential phase feeding but also to any design requiring phase delay, such as hybrids and couplers. This is significant because it provides a flexible and efficient solution for achieving broadband phase control while maintaining a highly compact size and low form factor.

[0122] Figure 13bThe curve of the measured antenna gain and frequency is shown. It can be seen that there is a certain offset at the two local maximum points of the measured LHCP curve. The first local maximum point is caused by the semi-TM 01 The mode is generated, shifts to lower frequencies, and reaches a local maximum of 1.2dBic at 380GHz. The second local maximum point is formed by a quarter of the TE 201 The mode is generated and shifts to higher frequencies, reaching a local maximum of 0.8 dBic at 475 GHz. The gain near the local maximum is several dB higher than the simulated gain, but due to frequency offset, the gain between these two local maxima decreases significantly, reaching a local minimum of -5.9 dBic at 405 GHz. It can be seen that the measured gain curve fluctuates significantly compared to the simulation results. In addition to the reasons mentioned above, some of the errors may come from the multi-amplitude component technology used. Because this method requires measurements at different angles, certain measurement errors and systematic errors may appear multiple times in the calculation, resulting in a deterioration in the final result. The calculated right-hand circular polarization curve is highly consistent with the AR curve, indicating a low level of right-hand circular polarization, especially in the frequency range of 360 to 440 GHz, which demonstrates the excellent broadband performance of the design.

[0123] Due to the 20GHz frequency offset, when comparing the measured and simulated radiation patterns, the simulation results with a frequency 20GHz higher than the measured results were used for comparison. Figures 14a-14e As shown in the figure, the measured antenna exhibits excellent LHCP radiation across a wide frequency band while maintaining low RHCP levels. However, at 420 GHz, there is a slight difference in RHCP levels, primarily due to an overall shift in the AR bandwidth, which causes the axial ratio to shift from a local minimum to a local maximum. Overall, the measured antenna radiation pattern closely matches the simulation results, validating the excellent LHCP radiation performance of the proposed antenna.

[0124] In summary, Figure 1-4b The antenna in the paper is a 425GHz broadband on-chip circularly polarized antenna combined with a highly compact load reactance sequential phase feed network. The antenna involves significant improvements to the radiating element and feed network. By cleverly utilizing the equivalent electric and magnetic field walls, the radiating element is tuned to a unique semi-TM 01 and a quarter TE 201mode, thereby extending the impedance bandwidth while maintaining a compact size. This design adopts a breakthrough approach that cleverly utilizes multi-layer metal in CMOS technology and integrates reactive elements into the feed network to change the phase. This innovative technology significantly reduces the size of the feed network and provides impressive broadband performance. The final antenna measures only 0.55λ0×0.55λ0 and is only 0.012λ0 thick. Despite its compact structure, the antenna still achieves 19% impedance-AR bandwidth overlap and a maximum gain of 1.2dBic. Its unique advantages include substantial miniaturization, extremely small size, and wide impedance and AR bandwidth, making it an attractive candidate for fully integrated THz sensing / imaging and communication applications.

[0125] In the above, the embodiments of the present invention are fully described. Although specific embodiments are mentioned in the description, it will be clear to those skilled in the art that the present invention can be implemented in variations of these specific details. Therefore, the present invention should not be construed as being limited to the embodiments described herein.

[0126] Although these embodiments have been described and illustrated in detail in the drawings and the foregoing description, they should be considered illustrative rather than restrictive, as only exemplary embodiments are shown and described and the scope of the invention is not limited in any way. It is understood that any feature described herein can be used with any embodiment. The illustrative embodiments are not mutually exclusive, nor are they exclusive of other embodiments not described in this specification. Therefore, the present invention also provides embodiments that include combinations of one or more of the above-mentioned illustrative embodiments. The invention described in this specification may be modified and altered without departing from the spirit and scope of the invention, and therefore, only the limitations described in the accompanying claims should be applied.

[0127] For example, in Figure 1-4b In the exemplary embodiment shown, there are four radiating elements, and the input phase difference between adjacent radiating elements is 90°. However, those skilled in the art will appreciate that, according to the present invention, there is no limitation on the number of radiating elements or the input phase difference of a circularly polarized antenna. For example, the present invention can be applied to a circularly polarized antenna consisting of four to eight radiating elements, arranged in a rotationally symmetrical manner.

[0128] In the exemplary embodiment described above, the feed network includes three equivalent capacitors and one equivalent inductor. However, the present invention should not be limited to any specific number of inductors and capacitors. For example, the number of capacitors and / or inductors may be increased or decreased depending on the number of radiating elements.

[0129] Figure 1-4bThe antenna in the exemplary embodiment shown is designed for operation at 425 GHz and is fabricated using 65-nanometer CMOS technology. However, the present invention is not limited to any particular frequency spectrum or manufacturing technology. Rather, the present invention can be applied to other frequency bands or semiconductor manufacturing processes. For example, the International Telecommunication Union (ITU) WRC-19 agenda item 1.15 is considering the use of the 275-450 GHz frequency range for next-generation terrestrial wireless and fixed service applications.

Claims

1. A circularly polarized antenna, comprising: a plurality of radiating elements arranged in a rotationally symmetrical manner; Each of the radiation units comprises: SMD components; a short-circuit wall connected to the patch element and adapted to short-circuit the patch element; and A feeding network connected to the plurality of radiating elements.

2. The circularly polarized antenna according to claim 1 further comprises a ground layer; for each of the radiating elements, its short-circuit wall is located between the corresponding patch element and the ground layer, and is connected to the patch element and the ground layer. The circularly polarized antenna according to claim 2 , wherein the short-circuit wall comprises a plurality of layers. The circularly polarized antenna according to claim 3 , wherein the short-circuit wall comprises ten layers. 5 . The circularly polarized antenna of claim 1 , wherein the patch element of each of the radiating elements has a generally rectangular shape.

6. The circularly polarized antenna according to claim 5, wherein for each of the radiating elements, the short-circuit wall is substantially connected to a first side of the patch element; the feeding network is substantially connected to a second side of the patch element; and the second side is opposite to the first side.

7. The circularly polarized antenna according to claim 5, wherein: For each of the radiating units, the projection of the short-circuit wall on the patch element has a generally "T" shape; the short-circuit wall has a first section connected to a first side of the patch element, and a second section perpendicular to the first section and extending from the first section toward the center of the patch element.

8. The circularly polarized antenna according to claim 7, wherein for each of the radiating elements, the feeding network comprises a corresponding output arm connected to the radiating element; the output arm extends in a direction substantially parallel to the first section. 9 . The circularly polarized antenna according to claim 7 , wherein the first section of the short-circuit wall has a length that is the same as a dimension of the first side of the patch element.

10. The circularly polarized antenna according to claim 5, wherein the patch element of each of the radiating units forms a rectangular notch at a corner of the rectangular shape; and the feeding network is connected to the patch element near the rectangular notch.

11. The circularly polarized antenna according to claim 2, wherein: For each of the radiation units, a corresponding hole is formed on the ground layer, the shape of the hole corresponds to the shape of the short-circuit wall; the hole partially receives the short-circuit wall. 12 . The circularly polarized antenna according to claim 1 , comprising four radiating elements, wherein an input phase difference between adjacent radiating elements is 90°.

13. A circularly polarized antenna, comprising: a plurality of radiating elements arranged in a rotationally symmetrical manner; A feed network connected to the plurality of radiating elements, the feed network comprising: core part; and a plurality of output arms, each of the output arms corresponding to one of the plurality of radiating elements; The plurality of output arms are coupled to the core portion; Wherein, the plurality of output arms are arranged in a rotationally symmetrical manner.

14. The circularly polarized antenna according to claim 13, wherein: The core portion has a generally square annular shape.

15. The circularly polarized antenna according to claim 14, wherein: The core portion has different widths along the extension direction of the core portion.

16. The circularly polarized antenna according to claim 13, wherein the core portion comprises a parallel plate capacitor. 17 . The circularly polarized antenna according to claim 16 , wherein the parallel plate capacitor is located in front of one of the output arms along a signal transmission path of the circularly polarized antenna. 18 . The circularly polarized antenna according to claim 16 , wherein the parallel plate capacitor comprises a first portion located at the same layer as the rest of the core portion, and a second portion located parallel to and below the first portion. 19 . The circularly polarized antenna according to claim 18 , wherein the output arm and the first portion of the parallel plate capacitor are located on the same layer.

20. The circularly polarized antenna according to claim 16 comprises four said radiating elements; the input phase difference between adjacent said radiating elements is 90°; and the circularly polarized antenna comprises three said parallel plate capacitors, which correspond to the three said radiating elements along the signal transmission path of the circularly polarized antenna.

21. The circularly polarized antenna according to claim 13, wherein: The core part is fed by a feeder line located in the same layer as the core part.

22. The circularly polarized antenna according to claim 21, wherein: The feed line extends along a first direction; and the core portion is short-circuited by a short-circuit line extending along a second direction perpendicular to the first direction.

23. The circularly polarized antenna according to claim 22, wherein: The short-circuit line passes through a short-circuit hole formed at the center of the core portion and is connected to a ground layer below the feed network.

Images