High Performance Electrically Small Huygens Rectennas Enable Wirelessly Powered Internet of Things Sensing Applications: A Review

1. An introduction to wireless power transfer (WPT) technologies for Internet of Things (IoT) applications

WPT technologies enable the transfer of electromagnetic (EM) energy from a power source to the load of an electronic device without any conducting wires or physical contact. WPT is much safer and more reliable than conventional wire-powered systems, in particular, for application in industrial activities such as mining and gas exploration. Moreover, WPT has become extremely important for the current and future IoT systems associated with the rapid developments of fifth generation (5G), sixth generation (6G) and beyond wireless technologies. Future ubiquitous wirelessly powered IoT devices are anticipated to negate the current need for short-life, bulky, and non-degradable chemical batteries. WPT is a highly attractive green-technology that has attracted considerable attention from both industry and the academic worlds.

The concept of WPT can be traced back to Nikola Tesla’s experiments at the end of the 19th century [1], [2], [3]. He invented the first resonant inductive near-field WPT technique and successfully lit a lamp bulb wirelessly. Far-field WPT has a long history dating back to the 1960s. Several very nice reviews of the past and recent developments are available [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18]. Solar power satellites (SPSs) and wirelessly powered aircraft and high altitude platforms are examples of very long distance applications of WPT technology. With the much deeper and faster development of both science and engineering in applied electromagnetics in the 21st century, WPT technology has already been widely adopted in many commercial electronics products, such as wireless-chargeable AppleTM products, electric toothbrushes, radio frequency identification (RFID) systems, and electric cars.

Based on their distinctive operating mechanisms, WPT systems can be categorized into two major types: near-field [19] and resonant mid-range [20], [21], [22] magnetically-coupled systems, and far-field EM-wave power transfer systems. The magnetically-coupled systems have generally adopted coils to transfer energy via magnetic fields. The physics of these systems enables only short-distance WPT. The sole approach that enables long-distance WPT, as far as hundreds of meters up to thousands of kilometers, is the transmission and reception of the power carried by propagating EM waves. Fig. 1 illustrates the power flow diagram associated with a typical far-field WPT system facilitated by EM waves. The transmitting antenna emits the EM waves that are energized from an alternating current (AC) power source. A receiver located in the far field of the transmitting antenna captures the emitted EM waves and converts that AC captured energy into direct current (DC) power via a rectifier circuit. The longest realized WPT has been from a space satellite; the satellite’s solar panels collected sunlight (EM waves) and its microwave transmitters beamed that energy to the Earth’s surface to be captured by a rectenna array, i.e., an SPS system [5], [23].

Far-field WPT is the major enabling technology for realizing battery-free IoT ecosystems in the incoming 5G and future wireless generations [14]. WPT opportunities now widely exist owing to the recent significant advances in ultra-low power electronics. For example, a battery-free cellphone powered solely by wireless fildelity (WiFi) signals has been successfully demonstrated [24]. Moreover, it has been concluded [25] that the power consumption of many IoT sensors can reach less than 0.1 milliwatts, which means that WPT base stations would be able to broadcast at very low power levels. Fig. 2 illustrates an application example of a wirelessly powered IoT sensing system for smart agriculture and farming. Many IoT sensors can be deployed around a single transmitter (power beacon) which itself is powered by a renewable energy source. It is an environmental-friendly arrangement that requires significantly less human intervention and labor. This type of battery-free ecosystem will help increase agricultural productivity and quality. The automatically collected data from the IoT sensors is transmitted by them back to the base station which also acts as the data gateway. The data would include important information vital to crops such as the soil pH level, moisture level, temperature, and humidity. It can then be analyzed for optimal decision making, for example, when to plant and when to harvest. Given the large number of anticipated IoT sensors and the current battery costs and sizes associated with this application, far-field WPT technology is highly desired for future smart agriculture and other wireless sensor networks.

2. Miniaturized rectenna designs for far-field WPT

The most critical aspect of a WPT-enabled IoT system is the performance of its rectennas, i.e., the integration of the receiving antennas and their rectifier circuits. While numerous studies on rectenna designs have been published, such as those on multi-band rectennas [26], multi-port rectennas [27], low input-power designs [28], [29], rectenna arrays [30], [31], and metasurface-based designs [17], [32], [33], these systems are not suitable for battery-free IoT devices and large numbers of elements in applications such as sensor networks because of their large sizes. Specifically, their ka values are larger than 1, where k is the free-space wavenumber and a is the smallest sphere that encloses the entire rectenna. Note that a design is considered to be electrically small if ka ≤ 1.

Several miniaturization techniques have been developed to achieve much more compact rectenna designs. These are summarized in Table 1 [24], [34], [35], [36], [37], [38]. The meandered planar inverted-F antenna (PIFA)-based rectennas are one of the most popular designs [24], [39], [40], [41]. The concept is to bend the arm of the receiving PIFA to realize size miniaturization. For example, a meandered PIFA rectenna was adapted in Ref. [24] to achieve a wirelessly-powered cell phone. Although the size was reduced in comparison to conventional PIFA or dipole-based rectennas, its wireless power capture capability was limited because the radiation pattern of the receiving antenna is omnidirectional with a low realized gain. Moreover, its polarization purity was low. Another miniaturization technique includes the design of meandered monopole- or dipole-based rectennas [34], [42], [43]. Although these designs achieve a higher polarization purity compared to the PIFA-based designs, their realized gains are still low and their sizes are not electrically small. Moreover, it is clear from Table 1 that it is difficult to seamlessly integrate these designs with the rectifier circuit. To address this issue, a fractal loop rectenna was developed in Ref. [35]. The entire system was very compact because the rectifier circuit was integrated inside the receiving loop antenna. Nonetheless, its wireless power capture capability was still low as a result of the low-gain omnidirectional pattern of its receiving loop antenna. To enhance the wireless power capture capability, i.e., to attain a higher realized gain of the receiving antenna, miniaturized fractal patch-based rectennas were developed in Refs. [36], [44], [45]. Although the area of each patch was reduced, the realized gain of the receiving antenna remained dependent on the size of the ground plane. Thus, the overall size of these designs remained large. To further reduce the rectenna size, dielectric-loaded versions were developed in Refs. [37], [46], [47]. However, their wireless power capture capability decreased owing to the dielectric loading, and their sizes were still not electrically small. One rectenna design in Ref. [38] achieved an electrically small size. However, the radiation pattern of its receiving antenna was omnidirectional with only a 1.0 decibel (dB) peak realized gain.

Table 1. Typical examples of miniaturized rectenna designs.

Meandered PIFA rectenna [24]	Meandered monopole/dipole rectenna [34]	Fractal loop rectenna [35]

Fractal patch rectenna [36]	Dielectric-loaded rectenna [37]	Electrically small near-field resonant parasitic rectenna [38]

The figure of meandered PIFA rectenna is reproduced from Ref. [24] with permission of ACM, ©2017. The figure of fractal loop rectenna is reproduced from Ref. [35] with permission of IEEE, ©2017. The figure of dielectric-loaded rectenna is reproduced from Ref. [37] with permission of IEEE, ©2014. The figure of electrically small near-field resonant parasitic rectenna is reproduced from Ref. [38] with permission of AIP Publishing, ©2011.

All of these “miniaturized” rectenna designs remain unsuitable for the targeted IoT applications. They do not simultaneously meet the associated challenges of being very compact in size, having large and broad-angle wireless EM wave capture capabilities, and having high AC-to-DC conversion efficiencies, while also being easy to fabricate and assemble with low costs.

As will be demonstrated in this paper, these challenges have been successfully met with the development of a series of high-performance electrically small Huygens rectenna systems that enable wirelessly powered IoT applications. The electrically small Huygens linearly-polarized (HLP) rectenna systems are presented first. Their design methodology smartly integrates two metamaterial-inspired electrically small radiating near-field resonant parasitic (NFRP) elements: the Egyptian axe dipole (EAD) and the capacitively-loaded loop (CLL), which are excited by an elemental, driven radiating element into a highly compact entity [48], [49]. The Huygens cardioid radiation patterns are obtained by achieving balanced, in-phase radiated electric, and magnetic fields. A low-profile HLP version is realized using three printed circuit board (PCB) substrate layers. An ultrathin HLP system is realized on a single PCB substrate [50]. The seamless integration of both electrically small antennas with IoT temperature and light sensors has been accomplished and confirmed with measured prototypes [51]. Two electrically small Huygens circularly-polarized (HCP) rectenna systems that address polarization mismatch issues are then described. One version is a low-profile HCP rectenna again realized with three PCB substrate layers [52]. The other is a modified HCP rectenna realized with four PCB substrate layers that achieves the maximum AC-to-DC conversion efficiency [53]. A dual-functional HLP rectenna and antenna system that is capable for both WPT and communications functions is discussed [54]. It meets the needs of emerging applications, such as wirelessly powered sensors, that require simultaneous wireless information and power transfer (SWIPT) performance. Finally, two HLP rectenna array systems are developed to further enhance the capture of available wireless power. Both the DC and radio frequency (RF) power-combining schemes are described.

The remainder of this paper is organized as follows. Section 3 illustrates the design methodology of the electrically small Huygens rectennas. Section 4introduces the two HLP rectenna systems and how they are tailored and integrated effectively with IoT temperature and light sensors. Section 5 presents the two HCP rectenna systems that can be similarly tailored for sensor and other IoT applications. Section 6 discusses the dual-functional HLP rectenna and antenna system. Section 7 demonstrates rectenna arrays that facilitate both RF and DC combining schemes. Finally, Section 8 is the conclusion. All numerical simulations and their optimizations reported herein were performed using the commercial software: Advanced Design System (ADS, Keysight Technologies, Inc., USA), ANSYS Electromagnetics Suite (HFSS, ANSYS, USA), and CST Studio Suite (CST, Germany). The reported simulation model results employed the known, real properties of all of the dielectrics and conductors.

3. Design methodology of electrically small Huygens rectennas

The rectennas presented herein are composed of Huygens dipole antennas (HDAs) and ultracompact rectifier circuits. The basic EM concepts associated with electrically small HDAs are presented first to understand the underlying design methodologies. The associated rectifier circuits and layouts are then described.

3.1. Electromagnetics of electrically small Huygens antennas

The basic HDA consists of a pair of complementary radiating elements, for example, an electric dipole (ED) and a magnetic dipole (MD). If the two dipoles are arranged orthogonal to each other and radiate in-phase balanced fields, a cardioid-shaped Huygens pattern will be realized as shown in Fig. 3(a). While the basic electromagnetics of a Huygens antenna is the same as the magnetoelectric (ME) dipole antennas that have been invented and extensively developed by Luk and Wong [55], Ge and Luk [56], Luk and Wu [57], Wang et al. [58], and Li et al. [59], they achieve it with significantly more compact structures. For example, the entire volume of our developed HLP receiving antenna (π(0.115λ0)2 × 0.04λ0 = 0.00166λ03, where λ0 is the wavelength of the resonance frequency) [49] is 150 times smaller than the volume (λ0 × λ0 × 0.25λ0 = 0.25λ03) of the original half-wavelength ME dipole design in Ref. [55]. Although the bandwidth of that ME dipole (43.8%) is much larger, our narrowband electrically small design (∼0.6%) is more suitable for WPT applications that do not require a wide bandwidth. The gain of the ME dipole (8 dBi) is higher simply because it has a large ground plane and, hence, a notably larger overall size. On the other hand, the half-power beamwidth of our HLP antenna (> 150°) is significantly larger.

Fig. 3(a) presents the ideal HDA. An infinitesimal ED is oriented along the +x-axis. An infinitesimal MD is oriented orthogonal to it along the +y-axis. The far-field expressions for their electric fields, ${\vec{E}}_{E D}$ and ${\vec{E}}_{M D}$ , are [60](1) ${\vec{E}}_{ED}^{ff} = + j k η [I_{e} L] \cdot \frac{e^{- j k r}}{4 π r} \cdot [\hat{r} \times \hat{r} \times \hat{x}] = + j k η [I_{e} L] \cdot \frac{e^{- j k r}}{4 π r} \cdot [- cos θ cos φ \hat{θ} + sin φ \hat{φ}]$ (2) ${\vec{E}}_{MD}^{ff} = + j k [I_{m} L] \cdot \frac{e^{- j k r}}{4 π r} \cdot [\hat{r} \times \hat{y}] = + j k [I_{m} L] \cdot \frac{e^{- j k r}}{4 π r} \cdot [- cos φ \hat{θ} + cos θ sin φ \hat{φ}]$ where θ is the elevation angle relative to the z-axis; φ is the angle in the azimuth x–y plane; r is the distance from the origin to the observation point; Imis the magnetic current; Ie is the electric current; L is the length of the electric and magnetic current elements; $I_{e} L$ is the ED current moment; $I_{m} L$ is the MD current moment; $k = 2 π / λ_{0}$ is the free-space wavenumber, $λ_{0}$ being the free-space wavelength corresponding to its resonance frequency; and $η = \sqrt{μ_{0} / ε_{0}}$ is the free-space wave impedance where μ0 and ε0 are the permeability and permittivity of free space, respectively. The total field radiated by both dipoles in concert is the sum of the fields they radiate individually. With Ie and Im being orthogonal and in-phase, the balanced condition: $I_{m} = η I_{e}$ , yields the Huygens far-field radiation pattern:(3) ${\vec{E}}_{ED}^{ff} + {\vec{E}}_{MD}^{ff} = j k η [I_{e} L] \frac{e^{- j k r}}{4 π r} (1 + cos θ) [- cos φ \hat{θ} + sin φ \hat{φ}]$

As illustrated, a cardioid behavior is attained with this HDA concept, i.e., the factor $(1 + \cos θ)$ yields a factor of two in the broadside direction, θ = 0°, and zero in the opposite one, θ = 180°. Notably, the patterns in both principal vertical planes, φ = 0° and φ = 90°, are identical. The peak directivity of the HDA is twice that of either dipole element individually and its front-to-back ratio (FTBR) is infinite instead of being unity as it is for each dipole individually as illustrated in Fig. 3(b).

The metamaterial-inspired electrically small NFRP elements, the EAD and the CLL, are ideal candidates for the ED and MD, respectively [61], to realize a practical electrically small HDA [48], [49], [50], [51], [52], [53], [54]. As illustrated in Fig. 3(b), the EAD functions as an ED and the CLL acts as an MD, both of which individually have a donut-shaped omnidirectional three-dimensional (3D) radiation pattern. If they are seamlessly integrated and properly balanced to yield the behavior described by Eq. (3), the Huygens cardioid radiation pattern in Fig. 3(a) is realized. Note that the dipole directions are purposely rotated by 90° in Fig. 3(b) from those in Fig. 3(a), i.e., the ED is now oriented along the −y-axis and the MD is oriented along the +x-axis. This choice emphasizes that in both Figs. 3(a) and (b), the cross-product of the ED and MD directions is along the +z-axis, the direction of the maximum of the cardioid pattern. The practical realization of this Huygens behavior will be demonstrated in the configuration shown in the figure below in Section 4.1. The actual realizations of an HDA with these metamaterial-inspired NFRP elements that was achieved in practice with low-cost PCB technology is described in the next section.

3.2. Compact rectifier circuit design

The other critical part of an electrically small rectenna is the rectifier circuit. It must be compact and highly efficient. Fig. 4 presents the circuit model of the two-diode full-wave rectifier developed in Ref. [49]. The main components of the rectifier are the two Schottky diodes arranged in a parallel configuration. An inductor L and the two capacitors C1 and C3 form an input impedance matching network for a standard 50 Ω source. The inductor also acts as a high-stop filter to reflect back any higher-order harmonics generated by the nonlinear diodes. Capacitor C2 also acts as an energy storage device during one half-cycle. The vertical diode facilitates charging it during each negative portion of the sinusoidal signal. This stored energy is released during the positive portion, which doubles the output voltage. The capacitor C3 smooths the output DC voltage delivered to the load, represented by the resistor RL. The rectifier circuit is easily realized on a compact PCB substrate as shown in Fig. 5.

Because the circuit is realized as a parallel microstrip line structure, the thickness of the PCB is not important and can be ultrathin. Our targeted operating frequency is in the free 915 MHz industrial, scientific, and medical (ISM) band. For optimal operation in this band, HSMS286C Schottky diodes (Broadcom Inc., USA) were selected. The inductors, including two RF chokes(LRF_choke), were added to prevent any AC signals arising from the twin-line transmission line from coupling with the DC output. The capacitors were from Murata Manufacturing Co., Ltd., Japan and the resistor RL was obtained from Yageo, China. All of these lumped components were in 0403 (1 mm length) surface-mount device (SMD) packages. The optimized values for each component were: L = 39 nH, C1 = 0.4 pF, C2 = 100 pF, C3 = 100 pF, RL = 10 kΩ, and LRF_choke = 560 nH. The PCB implementation of the rectifier is compact with its length being only 7.3 mm.

Fig. 6 [49] presents the measured performance of the rectifier [49]. The input parallel microstrip line of the rectifier was connected to a differential source that is realized by a balun-choked coaxial cable. The cable was connected to a signal generator from Keysight Technologies, Inc., USA. Fig. 6(a) [49] shows the measured magnitude of the S-parameter S11, |S11|, values as a function of the source frequency under different input power levels. The 10 dB impedance bandwidth covers 120 MHz, from 825 to 945 MHz for a variety of input power levels.

Fig. 6(b) [49] presents the measured and simulated AC-to-DC efficiencies as functions of the input power at our targeted 915 MHz frequency. The AC-to-DC conversion efficiency is defined as the output DC power divided by the input RF/AC power. The measured results agree well with their simulated values. The measured peak AC-to-DC efficiency was 80.3% for a 10 dBm (10 mW) input power, which is very close to the simulated peak value of 81.8%. Moreover, the efficiency is greater than 50% when the input power ranges widely from −5 dBm (0.3 mW) to 13.5 dBm (22.4 mW). Thus, the developed compact rectifier circuit demonstrated a high AC-to-DC conversion efficiency. The small discrepancy in the received power level for the peak efficiency value is due to the difference between the actual parameters of the real HSMS286C diodes and those values used in the ADS simulation model. Note that the matching inductor L in this circuit introduces a nontrivial loss to the system. As shown in our other designs below, the AC-to-DC conversion efficiency of this rectifier can be further improved if the inductor L is removed and its absence is properly counterbalanced by the input impedance of the antenna.

4. Highly efficient, electrically small HLP rectennas

Two measured, highly efficient, electrically small HLP rectennas are discussed. Their simulated and measured results are presented. Their practical versions tailored to realize wirelessly powered temperature and light IoT sensors are described in detail.

4.1. Low-profile HLP rectenna realized with three PCB substrate layers

The presented design methodology was used to realize a low-profile electrically small HLP rectenna implemented with low-cost PCB technology [49]. Its simulation models are shown in Fig. 7 [49]. The entire rectenna consists of three PCB disks as depicted in Fig. 7(a) [49]. Each layer was realized as a Rogers DuroidTM 5880 copper-clad substrate. Its relative permittivity, permeability, and loss tangent are 2.2, 1.0, and 0.0009, respectively. The thickness of Substrate #1 and Substrate #3 is 0.787 mm and that of Substrate #2 is 0.508 mm. The metallic traces of the EAD element are etched on the center Substrate #2, and the CLL is formed by two metallic strips connected by two vertical copper posts. One strip is printed on the upper surface of Substrate #1 and the other is printed on the upper surface of Substrate #3. As shown in Fig. 7(b) [49], the short-driven dipole and the metallic traces of the rectifier are seamlessly integrated and etched on the bottom surface of Substrate #3. The gap between Substrate #1 and Substrate #2 determines the coupling between the EAD and the CLL NFRP elements and, thus, the realization of the receiving antenna’s Huygens cardioid pattern. It was optimized to be 4.0 mm in this design. The diameter of the HLP rectenna is only 0.23λ0 and its profile is only 0.04λ0. It is electrically small with its ka value being 0.723.

Another significant feature of this design is the aforementioned absence of the lossy inductor in the rectifier. It was eliminated by properly tuning the input impedance of the broadside-radiating HLP antenna. This was achieved by adjusting the length of its elemental dipole antenna to have an inductive value that is directly matched to the rectifier, which has a capacitive impedance. Fig. 8(a) [49] presents the simulated impedance of the HLP antenna (without the rectifier) as a function of the source frequency. The impedance has an inductive value of (77 + j129) Ω at the targeted frequency of 915 MHz. Fig. 8(b) [49] shows the simulated radiation patterns at 915 MHz. It is clear that cardioid-shaped Huygens patterns are realized. The peak broadside gain is 3.8 dBi and a significantly high FTBR, 23.5 dB, is obtained. Broad and similar realized gain patterns in both principal vertical planes are obtained. The radiation efficiency of the antenna is 80%. Notably, while the ideal HDA has identical patterns in both principal vertical planes, the realistic system has slightly different ones simply because of the finite, different sizes of its elements in those planes.

Fig. 9 [49] presents the current distributions and directions on the EAD and CLL radiators over a period of time, T, at 915 MHz. They clearly demonstrate the resonant behavior of this NFRP element pair. The currents on the CLL (MD) element are the strongest at the times: t = 0 and t = (2/4)T. The currents on the EAD (ED) element are dominant at the times: t = (1/4)T and t = (3/4)T. Because the phase of the equivalent MD is 90° ahead of its corresponding loop currents, the ED and MD are in-phase. As the NFRP elements are already orthogonal to one another by construction, the conditions for producing balanced fields; and, hence, the desired Huygens radiation performance was satisfied.

The fabricated HLP rectenna prototype is shown in Fig. 10 [49]. It is noted that two short copper rods were attached to the short-driven dipole for tuning purposes after the prototype’s dimensions were discovered to not be the specified values. The rectenna was measured in an anechoic chamber as shown in Fig. 11 [49]. The far-field WPT measurement setup consisted of a signal generator (Keysight Technologies, Inc., USA), a power amplifier (Mini-Circuits, USA), a standard wideband horn antenna as the transmitting antenna, a DC power supply, a multimeter, and several 50 Ω cables. The rectenna was located 1.2 m from the horn antenna’s aperture in its far field at 915 MHz.