A Comprehensive Review on Wireless Capacitive Power Transfer Technology: Fundamentals and Applications

Capacitive power transfer (CPT) technology is becoming increasingly popular in various application areas. Due to its limitations, such as low frequency, low coupling capacitance, and the high voltage stress on metal plates, the studies on high power CPT applications fell behind previously. Therefore, the wideband gap (WBG) semiconductor devices and the compensation topologies are further adopted to tackle these limitations. The main purpose of the paper is to review CPT applications in terms of performance parameters, advantages, disadvantages, and also challenges. Initially, the basic principles of CPT technology are examined, which cover compensation topologies, coupler structures, transfer distance, power electronic components, and system control methods. Then, CPT applications are evaluated for performance parameters (i.e., power level, operation frequency, system efficiency, transfer distance) along with compensation types, inverter types, and coupler types. The applications are categorized into six main groups according to industrial topics as safety, consumer electronics, transport, electric machines, biomedical, and miscellaneous. Herein, power level changes from $\mu \text{W}$ to kW ranges, the operation frequency varies from 100s of kHz to 10s of MHz ranges as well. The maximum system efficiency is recorded as 97.1 %. The transfer distance varies from $\mu \text{m}$ range to 100s of mm ranges. The full-bridge inverter topology and four-plate coupler structure are noticeable in CPT applications. Finally, advantages, disadvantages, and challenges of CPT applications are evaluated in detail. This review is expected to serve as a reference for researchers who study on CPT systems and their applications.


I. INTRODUCTION
Wireless power transfer (WPT) technology was put forward to transmit electrical energy using coils by Tesla [1]. In 1961, Schuder contributed to WPT technology in biomedical applications [2]. Brown and Glaser are the pioneers of contactless microwave power transfer [3], [4]. WPT can be categorized into two scenarios: far-field and near-field wireless power transfer technologies depending on the transfer distance. Far-field WPT is based on electromagnetic radiation. It can be classified into two types as optical The associate editor coordinating the review of this manuscript and approving it for publication was Nagesh Prabhu .
(laser) power transfer applications [5] and microwave (radio) power transfer applications [6]. On the other hand, near-field WPT is based on non-radiative energy. The non-radiative energy can be transferred depending on the coupling techniques that are magnetic resonant coupling [7], inductive coupling [8], capacitive coupling [9], and magneto-dynamic coupling [10]. Furthermore, near-field and far-field categories are comprehensively reviewed in WPT [11]. Among the coupling techniques, inductive and capacitive coupling techniques become prominent. The foremost technique uses an alternating magnetic field to transfer power wirelessly and is hereafter called inductive power transfer (IPT). The basic inductive power transfer scheme is formed with a high frequency inverter on the primary side, a high frequency rectifier with a load on the secondary side, and an inductive coupling interface as well. The latter technique utilizes an alternating electric field to transfer power wirelessly and is hereafter called capacitive power transfer (CPT). A basic scheme of the CPT system includes a high frequency inverter on the primary side, a high frequency rectifier with a load on the secondary side, and a capacitive coupling interface. Compared to IPT technology, CPT has many advantages: lower cost, lightweight, lower eddycurrent loss, higher reliability, and lower EMI [12]. Considering review articles regarding IPT and CPT techniques, IPT has many alternatives. For instance, electric vehicles (EVs) charging in static and dynamic charging concepts [13], in terms of state-of-the-art research [14], [15], and compensation topologies [16], [17]. Jiang et al. reviewed resonant and non-resonant inverter structures in terms of their key features and characteristics [18]. The authors concluded that class EF 2 resonant inverter is a good option to obtain high efficiency power transfer, and the direct single-stage AC-AC non-resonant converter is a good candidate to reach the desired high frequency operation. The typical applications and future trends are also presented in a review concept for IPT [19]. Herein, typical applications are represented by Electric Vehicles (EVs), biomedical implants, and portable electronics.
Considering CPT applications, Dai and Ludois [20] presented a comparative study of inductive and capacitive couplings based on small air-gap applications. Herein, performance parameters (i.e., power level, transfer efficiency, operating frequency, transfer distance, transmitter, and receiver areas, etc.) are reviewed in detail. Furthermore, challenges and limitations are mentioned in both techniques with recommendations. Low coupling capacitance and high voltage stress on capacitive coupler are referred as the most critical ones. In a review by Rozario et al. [21], compensation networks in small air-gap and large air-gap applications are investigated to find the best compensator for the designer. Moreover, comparative studies on compensation networks are carried out according to the performance metrics (i.e., sensitivity to distance, tolerance on frequency, power level, and control capability). The authors concluded that while the LLC resonant network is the best compensator for small air-gap CPT applications as they provide a flexible design, doublesided LCLC, and double-sided LC compensation networks should be used in large air-gap CPT applications as they provide reduced voltage stress. Lu et al. [22] reviewed compensation topologies, capacitive coupler structures, and recent developments in CPT technology. In addition, a comparative study is conducted for CPT and IPT technologies in terms of their advantages and disadvantages. The authors identified that CPT has better large misalignment performance for EVs charging applications. The conventional coupler structures are also reviewed as two-plate, four-plate, six-plate, and electric field repeater. Mustapa et al. [23] presented a literature review on CPT technology in a biomedical implantable device. The authors point out that CPT is more robust against exposure to EMI. This paper presents a different perspective on CPT applications. Hence, CPT applications are categorized into six main groups as safety applications, consumer electronics applications, transport applications, electric machines applications, biomedical applications, and miscellaneous applications. Then, sub-applications for each main application are determined in this paper. Firstly, we have safety applications that involve foreign object and voltage stress. The second is consumer electronics applications that embody LED lighting, portable device charging, and integrated circuits (ICs). The third one is that transport applications include electric vehicle charging, drone (UAV) charging, and underwater charging. Additionally, electric vehicle charging applications are classified as static, dynamic, bidirectional, and robotics. The fourth one is electric machines applications which cover synchronous machines, three-phase, and rotary. The fifth one is biomedical applications, and the last one is miscellaneous applications that contain hybrid and multiple transmitters/receivers. Furthermore, applications are reviewed with performance parameters (i.e., power level, operation frequency, system efficiency, and transfer distance) along with compensation types, inverter types, and coupler types. Finally, advantages, disadvantages, and also challenges of CPT applications are reviewed in detail.
Following the introductory section, this paper is organized as follows. Section II introduces the basic principles and operation of CPT. Section III elaborates the applications in detail. Finally, conclusions and suggestions are drawn in section IV.

II. BASIC PRINCIPLES AND OPERATION OF CAPACITIVE POWER TRANSFER
A typical scheme of a CPT system is represented in Fig. 1. Herein, a CPT system consists of the primary side and the secondary side. The primary side is represented by a high frequency inverter and the primary compensation network. The secondary side is represented by the secondary compensation network, rectifier, and load. The capacitive coupler is a medium, which provides a power flow loop. A high frequency inverter is used to provide AC excitation for the primary and secondary resonant components with the help of displacement current in the coupling interface. A rectifier is ultimately used to supply DC for the load. The compensation networks enable a minimum impedance to flow the current via resonance condition with improved system efficiency. Among the compensation networks, filter-based topologies are drawn attention in CPT systems, such as LC, LCL, LCLC, etc. The coupling interface plays an essential role for both primary and secondary sides. Although two-plate, four-plate, six-plate, and electric field repeater are available for the conventional capacitive coupler structures, four-plate structures are generally preferred in CPT applications. In a typical fourplate coupler structure, two plates form a forward path, and the other two plates generate a return path for the resonant VOLUME 10, 2022 current. Different materials are used for the capacitive plates, such as aluminum, zinc, copper, etc. Among the materials, aluminum metal plates are usually utilized due to the low cost. In contrast to IPT technology, alternating electric fields are used to transfer power wirelessly. Moreover, the system weight and cost are reduced due to the metal plates compared to Litz wire coils in IPT. There is no necessity for the plate thickness and shape. However, the shape of the metal plates can provide a flexible design for different applications. The electric field emission can be reduced via circular-shaped metal plates as well. Coupling capacitance depends on plate area, transfer distance, and the dielectric material between the metal plates [24].
As the transfer distance increases, the value of coupling capacitance decreases, and hence the switching frequency needs to be increased to obtain a sufficient power level. However, the system efficiency will drop due to increase in the switching losses. Furthermore, higher resonant inductors are required to satisfy low coupling capacitances. Higher resonant inductors lead to higher conduction losses and also lower system efficiency. Therefore, compensation networks affect system efficiency and hence they should be designed properly for high power large air-gap applications.
A. COMPENSATION TOPOLOGIES 1) THE BASIC STRUCTURE Compensation topologies are an essential part of the CPT system as given in Table 1. Due to the limited size of metal plates and low permittivity of air, the coupling capacitances tend to have very low value, and it leads to a large coupling impedance. Thus, coupling impedance is generally much larger than the load impedance [25]. Compensation topologies are designed to increase the voltage across the capacitive coupler for the primary side and to reduce the voltage on the secondary side. Therefore, compensation topologies provide higher power transfer capability, higher efficiency, and minimized power rating of the converter.

2) PROPOSED DEVELOPMENTS IN LITERATURE
Double-sided L compensation is a filter-based topology, which provides a simplified circuit structure. Herein, the capacitive coupler acts as a resonant capacitor to compensate the series inductor [26]- [30] present in multi-cell applications [31], multi-phase applications [32], and also mutual capacitance effect [33]- [35]. Furthermore, one pulse switching active capacitor (OPSAC) is used instead of a resonant inductor to resist parameter variations [36]- [38]. Doublesided LCL compensation is proposed by Theodoridis to increase power transfer for portable device charging applications [39]. Zhang et al. presented double-sided LCL compensation for EV charging applications [40]. In addition, an optimization study is carried out to improve the parameters in double-sided LCL compensation [41]. Lu et al. proposed double-sided LCLC compensation [42]. Herein, the number of passive components is a disadvantage of the compensator, where the output power and system efficiency are well proportioned. Afterwards, double-sided CLLC compensation is put forward to reduce the required resonant inductance value [43]. A comparative study is carried out between LCLC and CLLC topologies considering output power and efficiency under the variations [44]. Although efficiency and output power linearly increase with the load resistance for CLLC compensation, the efficiency of LCLC compensation does not linearly increase with the load resistance. In addition, LCLC and CLLC compensation topologies have higher coupling capacitances and output power capability compared to L and LCL compensation topologies. A comparative study is carried out for LCL, LCLC, and CLLC compensation topologies in terms of voltage gain considering load, frequency, and coupling variations [45]. Thus, the CLLC compensator is the best one for system performance variation. A double-sided LC compensation is suggested for large air-gap applications [46], [47]. Herein, output power and coupling coefficient terms are inversely proportional to each other [47]. Hence, optimization studies are carried out to improve system efficiency [48]- [53]. The steady output voltages and higher system efficiencies are obtained due to the optimization studies. Then, a comparative study is performed to realize the characteristics of the LC, LCL and, LCLC topologies [54]. Herein, the output current, current gain, and output power terms are compared for different operating frequencies. Thus, the LCLC compensation is more robust when the operation frequency changes. Hybrid compensation networks are also reviewed, such as LC-CLC compensation to enhance the coupler misalignment performance [55] and LCL-L compensation to improve the system performance reducing the number of components [56], [57]. Multi-stage networks are proposed to provide higher voltage and current gain for high power applications [58].
Impedance-based compensation networks are proposed in CPT systems. For instance, Z impedance compensation has a symmetrical Z shape structure, which includes two identical inductors and two identical capacitors. Moreover, it has advantages encountered in short circuit and open circuit conditions [63]. F type compensation topology acquires an F shape structure, which comprises a resonant capacitor along with two resonant inductors. It has benefits in case of a sudden change in pick up [64]. Lastly, analysis and comparative works on compensation topologies are comprehensively discussed [21], [25], [61], [65]- [70]. The results show that conventional compensation topologies are typically used in CPT applications. Among them, L compensation has a simplified structure, typically used in small air-gap applications. However, higher resonant inductance and more sensitive to misalignments are its drawbacks. LCLC and CLLC topologies are typically used in large air-gap and high power applications since system power is proportional with the coupling coefficient value. CLLC compensation is the best one during system performance variations.
B. CAPACITIVE COUPLER STRUCTURES 1) THE AIM Fig. 2 represents the classification of conventional capacitive coupler structures. The main purpose of the coupler that consists of metal plates is to provide power transfer generating electric fields.

2) PROPOSED DEVELOPMENTS IN LITERATURE
Initially, two-plate structures are reviewed in detail. Herein, one metal plate serves as a power transmitter and the other metal plate serves as a power receiver. Furthermore, two-plate structures are categorized into three types according to application areas in this review. Firstly, the quasi-wireless application consists of a power transmitter and a power receiver with a returning path [71]- [74]. In this concept, only two metal plates are required, and the other two metal plates are replaced by the earth ground, which provides the current-returning  [75], [78], [86] (b) four-plate [40], [47] (c) six-plate [104] (d) electric field repeater [110].
path. Zhang and Lu propose an air-gap distance of 2m quasiwireless application [75]. According to that, it can be applied VOLUME 10, 2022 in both indoor and outdoor CPT applications. Neste et al. suggest a large surface quasi-wireless application [76], [77]. Thus, the system cost is reduced using inexpensive surfaces to transfer power. Secondly, electric vehicle charg-ing application can be mentioned in the two-plate concept. Using parasitic capacitances between the vehicle chassis and earth ground to provide the current-returning path is the special feature for this concept. In addition, the stray 3120 VOLUME 10, 2022 capacitance provides current-returning path [78], [79]. Metal wheels also play an essential role to provide current-returning path [80], [81]. The third one is a single-wire application that includes a power transmitter, power receiver, and load [82]- [86]. In this concept, a physical current-returning path is not used. Moreover, the working principle of single-wire applications is depending on the voltage difference between the terminals, and hence displacement current flows through. For this condition, large misalignment tolerance is provided between the coupled plates [86]. Electric field and magnetic field distributions between two coupled circular plates are also investigated in a single-wire concept. [87]. Hence, the electric field is uniformly distributed in the middle of two coupled plates whereas the magnetic field is unevenly distributed because of the lead wires. Four-plate coupler structure can be classified into two types as two pairs of metal plates in a parallel position [88] focusing on its different grounding conditions [89] and a vertical position reducing system size and improving coupling capacitance [40], [90] along with its input and output ports configurations [91]. Moreover, misalignment conditions are discussed in fourplate structures [92]- [94]. Thus, the horizontal structure is more sensitive to misalignments. Besides, the port capacitance term is proposed to derive the cross-coupling capacitors in modeling of the coupler [56], [95]. The modeling of the coupling interface is presented based on the coupling coefficient term [96]- [98]. Herein, the reduced cross-coupling effect provides higher power transfer. Both [47], [48] mention the π type and two-port equivalent models to simplify the complicated coupler structure. A current source-based two-port model is typically used in CPT systems. However, Wang et al. propose an induced voltage source model to cope with coupling variations [99], [100]. The T-type model is also proposed to easily realize optimal load conditions in contrast to the π type model [101], [102]. Abramov et al. present a model to realize cross-coupling relationships between input and output ports of the CPT system [103]. The six-plate coupling structure includes active four plates inside and larger passive two plates behind to reduce electric field emissions [104]- [109]. The electric field repeater is proposed to increase the transfer distance [110]. In addition, the electric field repeater enables the utilization of multiple transmitters/receivers in the WPT system [22].
Besides the conventional coupler structures, different coupler designs have been proposed. Dai and Ludois propose a conformal bumper structure for EV charging applications [111]. Herein, the coupling capacitance increases with reduced air-gap during charging process. A sandwich shaped coupler is proposed to improve the system misalignment ability [112]. However, the angular misalignment is the drawback side of the coupler. A separated circular coupler is suggested for drone (UAV) applications [113]. The coupler provides enhanced mutual capacitance. The sleeve-type coupler is suggested for rotary applications to enhance the system performance [114]. The sleeve-type coupler also provides smaller coupler size, lower resonant inductance, and lower voltage on sleeves. A hybrid coupler, which consists of inductive and capacitive couplers is proposed to provide space-saving in WPT applications [115]. One of the advantages of the coupler is to provide compact and flexible structure. Nevertheless, the system model becomes complex with the integration of twoparallel frame. The role of dielectric materials is discussed for capacitive coupler structures [24], [116]- [118]. Therefore, the effective coupling capacitance can be considerably enhanced with the integration of dielectric materials, which have higher relative permittivity than air.
The features of conventional capacitive coupler structures are given in Table 2. Misalignment is also a crucial issue to consider in capacitive coupler structures. There are four possible misalignment types called vertical, horizontal, angular, and rotational discussed in Table 3.

3) COUPLER METHODOLOGY
Calculation or estimation methods of coupling capacitance play a significant role to keep the resonance better in primary and secondary resonant circuits, then the compensation components are tuned accordingly [74]. Basic capacitance formula does not consider the fringing field effect, specifically large air-gap applications. Hence, finite element analysis (FEA) and response surface analysis (RSA) are used to analyze the coupler and estimate the coupling capacitance. An analytical approach is also proposed to evaluate the coupling capacitance matrix as an alternative to FEA [121].
The compatibility becomes unavoidable for the capacitive coupler structures in interoperability conditions, specifically for multiple transmitters and multiple receivers applications, which have different coupler areas or coupler configurations. For this reason, a compatibility test is carried out between different coupler configurations in terms of coupling capacitance and coupling coefficient terms [79]. Thus, the ring shaped coupler shows a good compatibility feature with the ring shaped coupler, on the other hand, the square shaped coupler shows a good compatibility feature with the square and disc shaped couplers.  In summary, the two-plate structure has advantages in terms of the reduced number of metal plates. Among the four-plate coupler structures, the four-plate vertical structure provides a compact solution, rotational misalignment tolerance, and higher self-capacitances. However, the four-plate horizontal structure is typically preferred in CPT applications due to the easily-realized equivalent circuit. Among the conventional coupler structures, the six-plate coupler is prominent in the safety concept. Nevertheless, the increased number of coupling capacitors is the drawback in realizing such couplers. The electric field repeater provides long-distance wireless power transfer. However, increased number of plates is the negative side of the coupler.

C. TRANSFER DISTANCE
Transfer distance or air-gaps classified as a small air-gap and a large air-gap play a significant role in determining output voltage and also output power in IPT and CPT applications. According to application areas, transfer distance changes crucially. For instance, the small air-gap applications are regarded as less than 10 mm, and the large air-gap applications are regarded as between 100 and 200 mm in CPT sys- tems [118]. Transfer distance is typically selected as 150 mm, indicating the ground clearance of the vehicle chassis in large airgap-based electric vehicle charging applications [51]. Lu et al. [42] achieved a maximum system efficiency with 90.8% for 150 mm transfer distance using double-sided LCLC compensation. Thus, the limitations and challenges are discussed based upon the increased transfer distance [122]. These are determined as human safety, system efficiency, the required power level, air breakdown, and allowable fringing field levels. In summary, transfer distance drastically affects the system performance in WPT applications.

D. POWER ELECTRONIC COMPONENTS AND SYSTEM CONTROL STRUCTURES
In this section, power electronic components and system control structures are thoroughly reviewed. According to that, power electronic components can be categorized into two groups as inverters and rectifiers in a general CPT system.

1) INVERTERS
Inverter structures occupy an essential place to excite resonant components for primary and secondary circuits in CPT systems. Considering inverter structures, half-bridge, full-bridge, class E, class D, class ϕ 2 , push-pull, and quasi-resonant topologies are available in CPT systems. Among them, full-bridge and class E inverters become prominent in CPT applications, as represented in Fig. 3. The half-bridge inverters, typically use WBG devices in CPT applications [123]- [126]. A grid integrated structure is presented along with a boost PFC [127]. Both [128], [129] introduce the ZVS technique in low power CPT applications with 3122 VOLUME 10, 2022 half-bridge structure and asymmetric half-bridge structure, respectively. On the other hand, full-bridge inverter structures are typically used in medium power and high power CPT applications using WBG devices [78], [130], [131].
Kumar et al. [132] propose a full-bridge-based inverter structure to provide variable compensation in WPT systems. It is possible to provide low cost, low resonant inductance, and higher switching frequency via power amplifier topologies. However, class E, class D, and class ϕ 2 inverter structures have figured in CPT applications among other power amplifier-based inverters. Firstly, design methods are presented in a basic class E inverter structure [133], [134]. Herein, the system consists of an ideal active switch with a shunt capacitance, a resonant inductor, and a capacitive coupler structure with a load side, as shown in Fig. 3 (a). Saat et al. carry out analysis and design studies on the side of CPT technology, specifically low power applications [135]- [140]. Resulting from one active switch and low coupling capacitance, class E inverter is typically preferred in low power CPT applications [141]- [143]. To improve the power level and the transfer distance, studies are carried out using class E inverter structure [144]- [148]. Moreover, studies are conducted to enhance the power transfer capability and minimize the system size [149]- [152]. Considering class D inverter, both half-bridge and full-bridge structures are put forward [153]- [156]. Class D structures have lower voltage stress on the switching devices than class E structures; however, a complicated gate driver circuit and increased THD of the output voltage are the disadvantages [153]. Additionally, soft-switching methods are discussed in class D inverter structures to provide peak efficiency [157]- [159]. Among the CPT applications, the class D inverter is utilized in consumer electronics [60], [160]. Kim and Choi propose a class ϕ 2 inverter structure for CPT systems [161]. Although the class ϕ 2 inverter is resemble to class E inverter, it has an additional circuit formed by an inductor and a capacitor in parallel with the switch. It provides low voltage stress compared with the class E inverter structure. The push-pull converter is involved in autonomous structure for comparison with class E converter in terms of performance characteristics [162] and the purpose of underground data transmission [163]. A single switch quasi-resonant converter is then proposed to reduce the losses and cost [164]. In summary, the full-bridge inverter structure is typically used in high power CPT applications. Power amplifier-based inverter structures are mainly utilized in low power CPT applications.

2) RECTIFIERS
Rectifiers are of great importance in wireless power transfer applications to supply regulated output voltage for a load. Half-bridge, full-bridge, and class E structures are available in CPT rectifier topologies. Among them, full-bridge passive rectifier and class E structures become prominent, shown in Fig. 4. Initially, a half-bridge rectifier is preferred in dynamic charging railway applications as they provide low cost and low loss [80]. Full-bridge passive rectifier struc- tures are typically used in both small air-gap and large airgap CPT applications. A passive rectification is presented to evaluate the capacitive effect of diode rectifiers [165]. On the other hand, full-bridge active rectifier structures are utilized in bidirectional CPT applications. Class E half-bridge rectifier structures are generally used in class E 2 DC-DC converter applications that consist of class E inverter and class E rectifier. Class E full-bridge rectifier has lower diode conduction loss, higher power density, and higher frequency rectification than class E half-bridge rectifier [166]. Thus, Domingos et al. propose a class E full-bridge rectifier for CPT applications [167]. Active variable reactance (AVR) rectifier structure is suggested in wireless capacitive power transfer applications to compensate for large misalignment conditions and distance variations among couplers [168], [169]. In summary, class E rectifier topology is suitable for multi-MHz CPT applications due to its soft-switching capability.
On the other hand, the full-bridge passive rectifier topology becomes challenging for multi-MHz applications. Due to the capacitive effect of the diodes, there are changes in resonant frequency in high frequency operation. Although it has disadvantages also, it is typically used in medium power and high power CPT applications as it does not include complex structure. Class E rectifier structure is mainly used in low power CPT applications.

3) SYSTEM CONTROL STRUCTURES
System control structures are of great importance in properly operate CPT systems and are typically put forward for load and coupling capacitance variations.
Considering load variations, power flow control method [170], auto frequency tuning control method [127], [171], perturb & observe (P&O)-based tracking control method [129], VOLUME 10, 2022 linear quadratic gaussian (LQG) control method [172] are proposed in CPT systems. The power flow control provides automatic regulation of the output voltage against circuit parameter variations. The auto frequency tuning control provides automatic control of switching frequency to make it equal to resonant frequency during load variations. The P&Obased control method creates optimal equivalent load against the load variations via maximum power tracking. The LQG control provides constant output voltage and low switching loss by automatically adjusting the switching frequency to the resonant frequency. Additionally, PI control is used to regulate the output voltage in the case of physically separated transmitter and receiver sides [173].
Considering coupling capacitance variations, frequency bifurcation approach in control methods [174], quasi-sliding mode control [175], hybrid control [176] are suggested in CPT systems. The frequency bifurcation approach provides coupling-independent operation, and hence relatively constant efficiency and power transfer. The quasi-sliding mode control provides system robustness during parameter variations. The hybrid control method is used to provide constant output voltage and soft-switching operation. Herein, the output current of the inverter is first detected, and the mode of the inverter is determined by the controller.
Considering both coupling capacitance and load variations, adaptive multi-loop control [177], the on-off keying modulation control [28], [178], and output feedback control called robust H ∞ method [179] are proposed in CPT systems. The adaptive multi-loop control includes frequency tracking and matching network tuning for the primary and secondary sides. It provides power transfer regulation via variable matching networks. The on-off keying modulation control senses the average value of the supply current, and it is compared with a reference value to drive the circuit. It enables optimum frequency and maximum efficiency operation during the variations. The robust H ∞ control is an output feedback control method. It provides a regulated output voltage during the load variations. According to that, the robust H ∞ control shows a better performance as compared with the PI control.

III. APPLICATIONS ON CAPACITIVE POWER TRANSFER
In this section, applications of CPT are considered. The safety, consumer electronics, transport, electric machines, biomedical, and miscellaneous applications indicate the main CPT applications as shown in Fig. 5. Advantages and disadvantages of CPT applications are thoroughly examined in Table 12. At last, challenges of CPT applications that are variable load, transfer medium, and safety issues mentioned in detail.

A. SAFETY APPLICATIONS
Safety applications are classified as the foreign object and voltage stress. The most important features of safety applications are given in Table 4. Based on this, it is aimed to show which application type is more suitable.

1) FOREIGN OBJECT
The foreign objects affect the system performance and absorb the energy through heat formation in WPT applications [180]. Therefore, foreign objects can lead to the proper shutting down of the operating system. Foreign objects can be classified as living objects [181], metal objects [182], and dielectric materials [183] in wireless capacitive charging applications. The foreign objects are typically modeled as an equivalent circuit to observe their effects on system performance. The equivalent circuit models of a living object and a metal object are presented for EVs charging applications [184]. Additionally, their detection methods are also presented. An algorithm is proposed in low power CPT applications based on target detection [185]. Herein, the value of transmitter side capacitance is observed depending on whether foreign objects enter the system or not. The effects of foreign objects are evaluated in capacitive wireless electric vehicle charging applications [186]. It is concluded that the foreign objects significantly impact the system's performance when the object gets close to the capacitive coupler. All in all, foreign object detection is a critical issue in CPT systems that should be meet the safety requirements. Otherwise, performance degradation emerges due to the detuning of system operating frequency and resonant frequency.

2) VOLTAGE STRESS
Voltage stress applications take part in both low power and high power CPT systems for safety concept. The main principles of the voltage stress applications are to decrease electric field emissions on capacitive metal plates. For instance, 3124 VOLUME 10, 2022 the effects of electric field emissions are considered in living tissues for biomedical applications [187]- [189], and the human safety concept for electric vehicle charging applications [190]. Mostafa et al. aimed at reducing voltage stress on metal plates in secondary side of the CPT system using CL network [191], using step-down transformer [192], and utilizing buck converter [193] to provide less system sensitivity in low power applications. Luo et al. suggest an optimization method to improve the component values [194], [195] and investigate network design to obtain maximum power transfer in high power CPT applications [196]. Both [197], [198] present a predesigned voltage stress approach on capacitive couplers to decrease the electric field emissions. Choi et al. propose a method, which comprises transformer usage in CPT systems. The authors proposed a double-sided transformer usage to reduce the system sensitivity and voltage stress [35], and one transformer usage to decrease the volume of the system [199]. Lastly, a multi-objective optimization method is also evaluated to improve the parameter values [41]. In the method, the power and signal channels are modeled in a mathematical form, and objective functions are given based on the models. In summary, voltage stress applications are mainly considered for human safety reducing electric field emissions.

B. CONSUMER ELECTRONICS APPLICATIONS
Consumer electronics applications are categorized as LED lighting, portable device charging, and integrated circuits (ICs). The most important features of consumer electronics applications are given in Table 5. Based on this, it is aimed to show which application type is more suitable.

1) LED LIGHTING
Due to the development of LED lighting technology in terms of lower energy consumption, lower cost, and enhanced life span operation, LED lighting applications have gained more importance for WPT applications in recent years [200]. Shmilovitz et al. propose a topology, which provides both power transfer and galvanic isolation. The authors concluded that lower cost, higher efficiency, and a straightforward structure are obtained [201]- [203], and further lower current distortion for the CPT system [204]. Considering CPT applications, LED lighting applications are used along with unipolar CPT applications [205], rotary [206], and multiple transmitters [207]. Lastly, a design structure is proposed for decorative purposes in LED lighting CPT applications [208]. In summary, LED lighting applications are mainly used in low power CPT systems. Among the converter topologies, the resonant converter is prominent in LED lighting applications.

2) PORTABLE DEVICE CHARGING
Portable device charging applications are classified as mobile phone charging and laptop charging. Initially, wireless mobile phone charging applications are evaluated in this section. Herein, a power amplifier structure with a step-up and step-down transformer is proposed to enhance the coupling capacitance and decrease the quality factor [60]. A cellshaped capacitive coupler design is suggested for positionindependent [209].
The frequency and duty cycle control method is applied to provide stable operation during coupling and load variations [28]. Both [210], [211] are proposed to achieve CPT through metal barriers. Hence, the eddy current loss is reduced with a safety operation. The constant output voltage is intended for output independence between receivers in mobile phone charging [212]. On the other hand, CPT laptop charging application is considered by Theodoridis with a design algorithm [39]. It provides minimum size and decreases voltage stress on the capacitive interface. The authors of [213], [214] discuss design methods for resonant inductors in high frequency laptop charging applications. The air-core inductor structure becomes prominent with the reduced cost. In conclusion, multi-MHz wireless charging is required to provide space-saving for portable device charging applications. The WBG semiconductor devices are drawn attention in converter topologies to enable higher efficiency. Considering compensation topologies, L compensation is widespread in portable device charging applications.

3) INTEGRATED CIRCUITS
The capacitive coupling enables electrical insulation between the ICs inside a chip. This condition makes CPT special for the projected sensitive instruments, which need to be decoupled from the noise [218]. Wireless capacitive power and data transfer are considered using bidirectional communication [217], using a subharmonic resonant system [216], and using multiple receivers [215]. In summary, capacitive coupling provides reliable and efficient energy transmission in the ICs. The reduction of power consumption is another critical issue. Additionally, transfer distance typically varies µm ranges for integrated circuits applications.

C. TRANSPORT APPLICATIONS
Transport applications are categorized as electric vehicle charging, drone charging, and underwater charging, as represented in Fig. 6. CPT for EV charging is represented in Fig. 6 (a). CPT for underwater charging is depicted in Fig. 6 (b). CPT for drone charging is shown in Fig. 6. (c). EVs charging applications are also grouped as static, dynamic, bidirectional, and robotics. The most important features of transport applications are given in Table 6. Based on this, it is aimed to show which application type is more suitable.

1) ELECTRIC VEHICLE CHARGING a: STATIC CHARGING
Electric vehicle charging applications can be classified into two types as static wireless charging and dynamic wireless charging. Static wireless charging is a stationary charging type that overcomes hazardous cable charging problems. Major design principles of stationary wireless VOLUME 10, 2022  charging can be stated as the selection of switching frequency, coupler design, compensation circuit design, and converter design considering safety, efficiency, volume, and cost. When it is necessary to increase the transfer distance between the plates, low coupling capacitance becomes the most challenging problem, especially in high power CPT systems. To tackle this problem, higher resonant inductances, higher switching frequency, and kV level voltages on metal plates are required [90]. Thus, different compensation topologies are proposed to increase the voltages on metal plates, such as double-sided LCLC compensation [42], [219], double-sided CLLC compensation [43], double-sided LC compensation [78], [105], [220]- [223], double-sided LCL compensation [40] and LCL-L compensation [57]. To adapt EV charging applications in CPT systems, switching frequency is also increased to MHz ranges using WBG semiconductors [220], [224]. Furthermore, multi-modular inverter structures [131], [225], [226] and mutual capacitance effect [33], [71] are evaluated in CPT systems to increase the power level. Resonant inductances are also investigated in EV charging applications for CPT systems [221], [227], [228]. When the voltages on metal plates are increased to kV level, electric field emission rises that should otherwise be compatible with the standards. Hence, the design of capacitive coupler structures is considered to reduce the electric field strength in EV charging applications [78], [79], [104], [229]- [231]. Finally, EV charging applications need to satisfy power transfer requirements. These are related to the switching frequency, coupling coefficient value, and voltages on the coupler. In addition, full-bridge inverter topology and high-order compensation topologies become prominent in this application concept.

b: DYNAMIC CHARGING
The structure of a dynamic CPT system includes the roadway side and vehicle side. The roadway side comprises the power transmitter plate inside the roadway connected with the inverter and primary compensation network. The vehicle side includes the power receiver plate connected with the compensation network and rectifier with a battery. Dynamic wireless charging applications can be categorized as capacitive coupler-based studies [77], [232]- [234] and roadway investigation-based studies [80], [81], [235]- [237] in CPT systems. The reduction of electric field emissions is emphasized on coupler-based studies. The design of roadway-powered CPT system is highlighted in roadway investigation-based studies. The benefits and the drawbacks of the dynamic wireless charging system are highlighted in [238], [239]. The increased driving range, reduced battery size, decreased traffic congestion and pollution are the benefits of the dynamic wireless charging, the high cost of the system makes up the drawback side. Although, dynamic capacitive wireless charging has its unique benefits such as low cost, high-reliability metal plates in a long-track system, good misalignment performance, and no eddy-current loss, but the self-inductance of the transmitter metal plates should be given attention in a high frequency operation.

c: BIDIRECTIONAL CHARGING
Despite CPT technology mainly used in unidirectional EV charging applications, it has also been used for bidirectional EV charging applications in recent years. A phase-shifted control method is proposed in a bidirectional CPT system with a compact coupler structure [240]. The control method provides the maximum power transfer efficiency by controlling the output voltage of the inverter in the secondary side. A bidirectional converter is proposed to provide both inductive and capacitive power transfer [241]. Herein, the bidirectional converter is a DC-DC converter, which enables the modularity of the system. A parameter design method is suggested for maintaining constant output voltage in coupling capacitance and load variations [242]. Both [243], [244] utilize a phase-shifted control method for bidirectional electric vehicle charging and underwater charging applications, respectively. The method enables to control the power flow direction and magnitude. As a result, the power flow is controlled in a bidirectional CPT system, and hence effective power transfer is achieved in these applications. In addition, full-bridge inverter topology with WBG semiconductor devices is mainly used to enhance the power transfer capability.

d: ROBOTICS CHARGING
As robot technology advances, the demand for it has increased accordingly, and also our life has become easier. However, when it comes to wireless robots charging technology, there are a couple of studies in CPT applications. This is because of its limitations and challenges that are thoroughly reviewed [245]. The challenges are expressed as the required power transfer, system efficiency, human safety, transfer distance, and transmission medium. A shielded design is proposed to reduce electric field emissions along with analysis and design studies in wireless EV robotics charging applications [246]. A bidirectional structure is proposed to provide energy stability in different robots [247]. Furthermore, energy sharing and balancing are considered in robotics charging applications [248]. In conclusion, robotics applications are regarded as mini EVs charging applications. For this application concept, the high frequency operation and L compensation topology become prominent.

2) DRONE (UAV)
Considering drone applications, many techniques are put forward to increase the flight range, such as battery damping or the use of high voltage power headlines [249]. Limitations and challenges are evaluated in drone charging applications [250]. According to that, the main limitations of electric UAVs are their limited battery capacity and their less robustness. Both inductive and capacitive power transfer technologies are investigated in drone charging applications [251]. Considering CPT applications, the receiver side of the system is minimized to ensure compatibility [252], [253], and to obtain the maximum efficiency [254]. A capacitive coupler design is proposed to decrease parasitic capacitances with a rotational misalignment capability [113], [255]. A master/slave approach between drones is suggested for increasing flight range by using different coupling configurations [256]. Herein, the matrix arrangement structure has a higher mutual capacitance than the horizontal and vertical structures. As a result, unmanned electric vehicles (UAV) are used in a wide range of areas from military to farming. Thanks to the CPT technology, the limited operating time is not a critical point.

3) UNDERWATER
Capacitive WPT technology is a good candidate for underwater charging applications. The reason for this is the low cost of metal plates and their endurance to the high pressure of the underwater environment. Moreover, the high permittivity of seawater could improve the coupling capacitance, and hence increased power transfer capability [257]. Taking of a long time at charge/discharge process of underwater devices, limited range, and higher cost of system maintenance bring out wireless underwater charging approach [258]. Tamura et al. present a capacitive coupler structure for underwater CPT applications considering coupling coefficient and quality factor terms depending on the frequency [259], [260].
An improved conductivity of capacitive coupler structure is suggested for underwater CPT applications [261]. Zhang and Lu contributed to underwater CPT applications investigating high power long-distance electric ship charging [257], [262]. Thus, MW level power transfer and hundreds of millimeter ranges for transfer distance can be achieved in wireless capacitive electric ship charging. Electric vessel charging is also evaluated using equivalent circuit models to realize the CPT system [263]. Herein, CPT provides a cheaper, lighter, and more reliable charging solution compared to IPT technology for small vessels. In summary, capacitive coupler designs are mainly considered in underwater charging applications to use the high conductivity of seawater. Moreover, double-sided LC compensation topology and full-bridge inverter topology have become prominent in this application concept.

D. ELECTRIC MACHINES APPLICATIONS
Electric machines applications are classified as synchronous machines, three-phase, and rotary. Among them, rotary   application has become most popular, depicted in Fig. 7. The most important features of electric machines applications are given in Table 7. Based on this, it is aimed to show which application type is more suitable.

1) SYNCHRONOUS MACHINES
With the use of CPT technology in synchronous machine applications, many advantages are offered such as maintenance is not required, no arcs, and non-sensitivity to speed [264]. Ludois et al. [265] suggest a method that couples power to the rotor in replacement of slip rings. A rotating rectifier board as a coupler is suggested for wound field synchronous machines [266]. On the other hand, the synchronous generator field excitation is elaborated using CPT technology considering minimized cost [267] and voltage regulation [59]. In summary, CPT technology is preferred in VOLUME 10, 2022 synchronous machines to feed the rotor field winding without using slip rings and rotary transformers.

2) THREE-PHASE
A single-phase structure is typically available in CPT applications. Conversely, the three-phase structure is not still as sufficient as it should be. Contactless CPT is achieved between stationary and moving parts by utilizing the capacitance of the linear bearings [32], [268]. A three-phase structure is applied to rotary applications using e-GaN FETs [269]. The three-leg inverter with a four-wire model is proposed to enhance the system misalignment ability [270]. As a result, higher coupling capacitance and improved power transfer capability are provided via three-phase coupler structure.

3) ROTARY
Rotary applications are formed by using stationary primary plates and rotating secondary plates [179]. Capacitive coupler structures are evaluated in rotary applications [114], [206]. The capacitive coupler structure provides freedom of movement between the primary and secondary metal plates.
Aerodynamic fluid bearings are proposed to increase the coupling capacitance by decreasing the distance between the stationary and moving parts [271]. To improve the system efficiency and robustness, an auto-tuning circuit using PI control method [272], the output feedback control method [179], and a cascaded boost-class E converter structure [273] are suggested for rotary applications. A three-phase CPT system is suggested for rotary applications to provide balanced and good coupling between the electrodes during rotations [269]. In summary, CPT technology provides 360 degrees free rotation in this application concept eliminating power cables.

E. BIOMEDICAL APPLICATIONS
CPT technology has distinguishing features such as low eddycurrent loss and lightweight in biomedical applications, and, therefore, there is a growing interest. CPT technology in biomedical applications is evaluated in Table 8. Based on  this, it is aimed to show which application is more suitable for biomedical applications. Capacitive wireless charging method is used in deeply implanted biomedical applications [274], [275]. Deeply implanted biomedical devices that become a challenging application for WPT technologies because it required minimum implant size and low tissue heating. Herein, CPT technology provides safe charging with reduced EMI and low tissue heating. Erfani et al. present rectifier structures [276], [277], and capacitive network effect in CPT biomedical applications [189]. Herein, the capacitive link is formed with two pairs of parallel plates using tissue as a dielectric material for the external side (transmitter) and implant side (receiver). A circuit model is proposed to investigate the effects of different parameters on self-capacitance (SC)-based power transfer approach using a mouse cadaver [278]. An analog front-end structure is designed in wireless capacitive pressure sensors [279]. Electrocardiogram (ECG) is considered in terms of signal isolation [280] and power consumption [281]. Capacitive wireless power and data telemetry are suggested for biomedical implants [178], [282]. Capacitive wireless power and data transmissions decrease the risk of infection, complexity, and frequent intervention [283]. A comparative study is carried out for neurostimulation regarding performance analysis of inductive, capacitive and ultra-sound coupling techniques [284]. According to that, the most suitable technique varies depending on the application type due to its unique benefits. A coil shape design is presented to improve the coupling capacitance between the transmitter and receiver antennas for biomedical implant devices [285]. An optimization study for class E inverter topology is presented in biomedical implants [286]. Herein, the optimization study provides good performance during load and coupling variations. In summary, CPT technology provides reduced EMI and more robust performance nearby metallic objects for biomedical applications compared to IPT technology. Capacitive coupling mainly uses two pairs of parallel plates with tissue as a dielectric material. The multi-MHz frequency operation is typically selected to minimize the size for wireless capacitive biomedical applications.

F. MISCELLANEOUS APPLICATIONS
Miscellaneous applications are classified as hybrid and multiple transmitters/receivers in Fig. 8. Fig. 8 (a) represents multiple transmitters, Fig. 8 (b) depicts multiple receivers, Fig. 8 (c) shows multiple transmitters and multiple receivers, and Fig. 8 (d) shows a hybrid application. The most important features of miscellaneous applications are given in Table 11. Based on this, it is aimed to show which application type is more suitable.

1) HYBRID
Hybrid WPT technology is put forward to combine higher efficiency characteristics of IPT technology and specifically lower cost of CPT technology [22]. Power is transferred via electric and magnetic fields, simultaneously. The output power is equivalent to the sum of the IPT and CPT systems. Firstly, Lu et al. suggested a hybrid WPT system using double-sided LC compensation for high power applications [287] and double-sided LCL compensation for low power applications [62]. A hybrid coupler is proposed to utilize the full advantages of the resonant components [288]. Herein, the hybrid coupler provides unity current gain against the variation of impedance. A WPT system design is proposed to transfer power across the metal barrier by using capacitive and inductive couplings [289]. For the design, a two-plate coupler structure is used in the primary side to provide a current flow in the metal barrier, and a coil structure is utilized to generate a magnetic field for the power transfer. A method is suggested for capacitive power and data transmissions [290]. While the data is transferred via electric fields, the power is transferred via magnetic fields. A conjugate image theory is evaluated to obtain maximum transfer efficiency considering series and parallel compensation topologies [291], [292]. Luo et al. present a hybrid WPT system for railway applications [293], [294] and high power applications [295]. The authors improve the misalignment ability of the system. A hybrid WPT system is considered in wireless EV charging applications [296], [297]. Thus, it provides increased power level, improved efficiency, and reduced system cost. A converter called MagCap is proposed to provide bidirectional power transfer in hybrid WPT system [298]. Table 9 presents the performance comparison  and Table 10 lists out the features of IPT, CPT and HWPT systems.
In summary, the hybrid coupler has the merits of IPT and CPT technologies. It combines both the good features by means of inductive coils and capacitive metal plates resonate together along with compensation components. Meanwhile, power transfer is provided. The hybrid system also utilizes specific advantages of electric and magnetic fields in resonant circuits to improve the power transfer.

2) MULTIPLE TRANSMITTERS/RECEIVERS
Multiple transmitters/receivers applications can be categorized as multiple transmitters with a single receiver, multiple receivers with a single transmitter, and multiple transmitters and multiple receivers in CPT systems. The CPT with multiple transmitters is proposed to improve the system efficiency [300]. According to that, the system efficiency can VOLUME 10, 2022  be enhanced by adding more transmitters. On the other side, CPT with multiple receivers are carried out with an optimum load determination [301]- [303], electric vehicle charging applications [304], conveyors [305], using mixed resonant topology [306], integrated circuits applications [215], and also portable device charging applications [170]. The main purpose is to charge multiple receivers at once with maximum efficiency using a single transmitter. The multiple transmitters and multiple receivers are used to increase the power transfer in CPT systems [307], [308]. In this concept, the optimum loads are determined to maximize the power transfer. In summary, it is expected that the focus will lie on charging multiple receivers simultaneously at once in CPT applications. The reasons for that can be stated as follows: reduced cost as it requires only an inverter, suitable to limited space applications, and shortened charging time for several outputs.

G. CHALLENGES OF CPT APPLICATIONS 1) VARIABLE LOADS
The fundamental challenge of CPT technology is the low coupling capacitance, specifically transport applications. This makes it hard enough to work efficiently at high power. Hence, compensation topologies are adopted to address the low coupling capacitance, mainly at a fixed load. However, the variable loads create crucial problems such as low efficiency and variable coupling capacitance. For this reason, Su et al. [306] proposed a hybrid compensation topology formed of a π-CLC on the primary side to increase the voltage of the coupling and T-CLC on the secondary side to provide constant output current against load variations for multiple receiver applications. Kline et al. [28] suggested an automatic frequency and duty cycle control method to keep maximum efficiency during load and coupling variations for portable device charging applications. In summary, variable loads adversely affect the performance of the system and, therefore, they should be controlled properly to operate CPT system in efficient manner.

2) TRANSFER MEDIUM
Transfer medium plays a significant role in forming power transfer mechanism in WPT systems. Air is typically used as a medium in CPT applications. However, underwater charging applications utilize water as a medium [257]. The biggest challenge of underwater charging applications is the dynamic nature of the seawater or ocean, which varies the coupling [258]. Moreover, misalignment become a critical issue because of this reason. Thus, power transfer efficiency is significantly affected. In summary, transfer medium affects the behaviour of power transfer mechanism, especially in the feature of high power, high frequency and high electric field strength.

3) SAFETY ISSUES
Safety is a significant concern in capacitive power transfer technology. The safety issue is evaluated in CPT applications in terms of high voltage stress on metal plates, electric field emissions, and foreign object impact. Due to the integration of compensation topologies, the plate voltage can rise to kV level to success high power transfer. This kV level is much more dangerous for all living creatures. Thus, the capacitive coupler structure should be insulated properly based on the application type.
Electric field emissions should be compatible with the international standards, for instance electric field strength may be lower than 614 V/m at 1 MHz for human safety according to IEEE C95.1 standard [310]. The effects of electric field emissions are evaluated in terms of human safety for biomedical [189] and electric vehicle charging applications [190]. Furthermore, a six-plate capacitive coupler is proposed to reduce electric field emissions for EV charging applications [104]. Herein, the external two plates serve as shielding plates to keep the electric field in the safety range.
The effects of foreign objects, that is, living and non-living objects, can detrimentally affect the power transfer when they are near the transfer medium [186]. In summary, safety issues are an inevitable factor to design an appropriately operating CPT system.

IV. CONCLUSION AND SUGGESTIONS
The CPT technology has been a significant candidate among WPT technologies in recent years. In this paper, basic principles of CPT technology are reviewed in terms of compensation topologies, coupler structures, transfer distance, power electronic components and control methods. Then, CPT applications are classified according to different industrial topics such as safety, consumer electronics, transport, electric machines, biomedical and miscellaneous applications. These topics are categorized within themselves and evaluated considering their performance parameters (i.e., power level, operation frequency, system efficiency, transfer distance) along with compensation types, inverter types, and coupler types. It is concluded that CPT technology is preferred from low power biomedical applications to high power transport applications, low transfer distance integrated circuits applications to large transfer distance transport applications. In addition, compensation topologies vary from L compensation topology to high-order compensation topologies. Besides conventional coupler structures, different coupler designs are figured in CPT applications. It has been established that the coupler designs are mainly proposed to either reduce electric field emissions or increase coupling capacitances. Finally, the benefits, drawbacks and challenges of CPT applications are reviewed in detail. The benefits typically observed are cost effective, reduced size, high tolerance to misalignments, high tolerance to parameter variations, reduced electric field emission. The drawbacks mainly listed out are high occupied space, increased cost, requires extra circuitry, control complexity, and design complexity. The challenges are variable loads, adverse effect of transfer medium, and safety issues with high voltage stress, high electric field emissions, and foreign object impact. With the further development of CPT technology, it will become preferred in many application areas. For this reason, it is predicted that CPT applications will play an important role not only in academic research but also in industrial areas.