A resonant current-mode wireless power transfer for implantable medical devices: an overview

Abstract

Implantable Medical Devices (IMDs) have been developing in ways to be lighter and lower-power systems. In the view of such developments, the battery recharging capacity to ensure the stable operation of the system is essential. Wireless power transfer (WPT) was proposed as a solution to recharge the battery without complex metallic contacts. However, due to limitations such as threshold voltage of power switches and minimal input power of the multi-stage structure (Rectifier + Regulator/DC-DC converter) of conventional voltage-mode (VM) WPT, there are drawbacks of an input power range above a certain threshold level and limitations due to strict regulations on the human body. These issues make the design of the IMD battery charger much harder and prevent IMDs from being a more viable option for people-in-need. This paper introduces distinguishing characteristics of resonant current-mode (RCM) WPT technology to overcome the aforementioned issues. It also describes the basic theory, conventional circuits of VM/RCM, comparisons, and major challenges of RCM. Finally, advanced and efficiency-enhancing techniques of the-state-of-art works among the RCM topologies will be discussed to follow up the trend of RCM WPT.

Introduction

Implantable Medical Devices (IMDs) are widely adopted for wireless sensor networks [1–4], artificial heart/pacemaker [5, 6], retinal implant, capsule endoscopes, neural recording for brain diagnosis [3, 7–9], deep brain stimulation for neurological diseases [8, 10–14, 18–23], cochlear implant [4, 15], and electro-encephalography (EEG) monitoring for epilepsy control [16, 17], which are shown in Fig. 1(a). Battery recharging capability is indispensable in battery-operated IMD systems. However, if the IMDs are charged with a cord or need surgery to replace the battery, it would not be preferred for people-in-need.

Fig. 1.

a IMDs, b the structure of wireless IMD charging system

As a solution, a wireless battery recharging system is suggested, called wireless power transfer (WPT) technology.

As shown in Fig. 1(b), WPT enables to transfer energy from a source to single or multiple loads via electro-magnetic field/wave [18], where there are two critical challenges to implant electronic devices in a human body.

First, received/transmitted power is strictly limited with safety regulations due to heating issues. Federal Communication Commission (FCC) and other standard-setting organizations limited power exposure on human tissues below 4 W/kg. Likewise, non-thermal effects like cell membrane permeability or central nervous system effects must be less than 10 mW/cm² [19].

Second, the physical size/volume of the IMD systems should be large enough to receive sufficient power and convert the energy forms that the loads require. It is obvious that people are reluctant to implant huge foreign substances in their bodies. In general, tens of millimeter diameter systems are suggested in [2, 3, 20], it is still large for implantation into the retina, human organs, etc. For the reasons, there is a demand for WPT technology with low harvestable input power while having the size of the system that can be implanted without repulsion.

Previous review papers that focus on fundamentals of WPT [21–24], high-power applications [18, 25–27], electric-vehicle (EV) [22, 24, 28, 29], coil-design [22, 28, 29], compensation network [18, 21, 29], standards [29], and specific IoT applications [30] of the voltage-mode (VM) have been published for last decade. However, it is difficult to find a review paper focusing on resonant current-mode (RCM) that has recently been proposed. This paper analyzes not only basic operations of the conventional VM, RCM WPT systems but the latest RCM trends and advanced technologies [19].

This paper is organized as follows, Sect. 2 demonstrates fundamentals of inductive power transfer (IPT), simple characteristic equations, compensation networks, and the VM, RCM topologies. Section 3 introduces various challenges facing WPT systems for IMDs. Advanced efficiency-enhancing techniques will be discussed in Sect. 4. Finally, Section 5 concludes the paper with a comparison table about the-state-of-art RCM works.

Fundamentals

1. Inductive Power Transfer

   WPT technique transmits energy to the loads wirelessly unlike the traditional metallic contact way. WPT can be classified into two topologies, (1) a radiative way for far-field applications by using microwave-power carriers and antennas [31], (2) an inductive coupling way for near-field applications with mutually-coupled coils. IMD applications mainly use the inductive method that can achieve higher energy transfer efficiency. Figure 2(a) shows a basic inductive power transfer (IPT) circuit model, and the circuit can be demonstrated as an equivalent circuit of a transformer [32]. As shown in Fig. 2(b), $M_{12}$ denotes the mutual inductance, $R_1$ and $R_2$ are the equivalent series resistance (ESR) of the transmitter (Tx) and receiver (Rx) coils, respectively.

   By neglecting the ESR of the coils, the power delivered to the load ($P_L$) at the IPT topology can be expressed with the angular frequency of power carrier (ω) and the RMS value of magnetic flux density at the center of the Rx coil (B), which is given in (1).

   $$P_L=V_2{I}_2=\omega A_{c}B{\mathrm{\zeta }}_1{J}_1{\mathrm{\zeta }}_2{J}_2{A}_w^2$$ 1
   As shown in Fig. 2(c), $A_c$, $A_w$ are the parameters of the Rx coil, which are the area viewed from the top and the cross-sectional area of the conductor, respectively. ${\mathrm{\zeta }}_1$, ${\mathrm{\zeta }}_2$, and Φ are the space factor of the Tx/Rx coil and the orientation of the coils, respectively. Under the assumption that the number of turns and physical/electrical parameters of the Rx/Tx coils are equal, (1) can be simplified as

   $$P_L=\omega {\text{M}}_{12}{I}_1{I}_2=\omega {\text{M}}_{12}{\mathrm{\zeta }}_1{J}_1{\mathrm{\zeta }}_2{J}_2{A}_w^2$$ 2

   As known from (2), the $P_L$ is proportional to ω, $A_w$, and the mutual inductance ($M_{12}$), which is the function of intra-coupling (k, Φ) of the Rx/Tx coils and distance between the Rx/Tx coils (d) [33]. In conclusion, the frequency of the power carrier and the size of the receiver coil should be increased, and the Rx/Tx coils should be well-aligned for higher power transfer efficiency (PTE).

2. Compensation Topology

   In order to increase the efficiency of WPT systems, proper selections of L, C values of the circuit and a compensation network that suits each application are critical design steps. However, when designing the WPT system with the RCM or VM, there are one similarity and difference between the two topologies.

   At the beginning, no matter which VM or RCM topology is used, designing the Rx/Tx coils with high-Q is important. A frequency-selectivity of the high-Q transmitter ($Q_{\mathit{Tx}}$) and receiver ($Q_{\mathit{Rx}}$), which is expressed with the values of R, L, and C, guarantees that it delivers only a desired power carrier frequency component to the loads without the internal or external interference. Generally, $Q_{\mathit{Tx}}$ should be greater than $Q_{\mathit{Rx}}$ in terms of stability [23]. In addition, high-Q can minimize the reactive power loss generated by the reactance components of the circuits from the viewpoint of reflected impedance.

   In contrast, the compensation networks and the loading effect are not critical considerations when the RCM WPT topology is designed. The RCM is independent of compensation networks, unlike the VM WPT topology for which a compensation network needs to be carefully selected. It is mainly because the RCM WPT topology has two separated operating modes, which will be described in detail on Section 3. Since the only LC tank is reflected to the transmitter when the Rx coil receives energy from the Tx side, it makes the design and calculation easier than the VM topology. Fig. 3 and Table 1 summarize circuits and characteristics of four compensation networks, Series-Series (SS), Series-Parallel (SP), Parallel-Series (PS) and Parallel-Parallel (PP) [21, 34, 35].

3. Voltage-Mode WPT

   A conventional VM WPT circuit is illustrated in Fig. 4. Most of the VM WPT systems are composed of a multi-stage structure such as the rectifier, regulator/DC-DC converter for the safe charging of the battery [19, 36–50]. However, since the VM WPT systems require that the received power must exceed the diode threshold voltage ($V_{\mathit{th}}$) of the rectifier and minimum input power ($P_{in,min}$) of the regulator and DC/DC converter, it makes inappropriate to apply for IMDs where transmission power is strictly limited due to regulations on human tissues. Thus, the VM WPT is more appropriate for high-power systems that can easily neglect the limitations of the multi-stage structure. For the reasons, the VM WPT technology is widely utilized to a battery charger for portable devices, mobile phones, mining/underwater systems, electric vehicles (EVs), and so forth [22, 24, 28, 29].

4. Resonant Current-Mode WPT

   As a solution to resolve the issues of the VM topology, resonant current-mode (RCM) WPT topology is proposed that can harvest -32 dBm sensitivity for implantable systems [19]. Figure 5(a) shows the RCM WPT circuit and sub-circuits of each operating mode. The RCM WPT stores the energy via a LC tank and directly transfers the energy to a load as a current source. The RCM WPT has two operating modes. The key waveforms of each mode in the RCM are illustrated in Fig. 5(b).

   At the start of the RCM operation, called the resonant mode (RESO), the switch $S_1$ turns on and the switch $S_2$ turns off to isolate the load, and a LC tank is formed. The power received via the Rx coil freewheels for multiple cycles ($N_{\mathit{cycle}}$) through the LC tank to accumulate energy unlike the VM WPT that receives power and rectifies/regulates power simultaneously. There is no power delivered to the load in the resonant mode.

   In the charging mode (CHG), the switch $S_1$ turns off and switch $S_2$ turns on to connect the inductor and the battery in series. The energy stored in the inductor of the LC tank is delivered to the battery when voltage of capacitor ($V_C$) becomes zero, the time of zero-crossing ($t_1$), until the time of non-residual energy in the LC tank ($t_2$). In this mode, the inductor of the LC tank operates as a current source.

   Figure 5(c) shows a common control mechanism of the reported RCM WPT works. A zero-crossing detector, which is a comparator with $V_c$ and GND as inputs, gives a signal to the N-cycle counter when $V_c$ crosses zero. Since $V_c$ has a 90 degrees phase delay from the coil current ($I_L$), the point that $V_c$ becomes zero is the equal point that $I_L$ reaches its maximum ($t_1$). Therefore, zero-crossing detector serves to ensure that there is no residual energy in the capacitor of the LC tank when the mode is changed from RESO to CHG. Next, the N-cycle counter changes the mode from RESO to CHG when the number of resonance cycle ($N_{\mathit{cycle}}$) meets the pre-determined number of resonance cycles ($N_{\mathit{RESO}}$).

   The end-of-charging detector (ECD) detects that $I_L$ becomes zero to start RESO again ($t_2$). The duration of CHG ($t_2-t_1$) is relatively short compared to RESO’s, and two modes are continuously repeated within this mechanism. The ECD consists of a comparator that keeps comparing $V_{\mathit{BAT}}$ and a reference voltage to monitor that the $V_{\mathit{BAT}}$ reaches the desired voltage as the I_L flows.

   Based on this operation, the RCM WPT is considered as a very suitable topology for the battery charger of IMDs because the RCM has the following two advantages compared to the VM [19, 51]. First, minimum harvestable input power decreases. Even a tiny amount of energy can be accumulated and amplified in RESO. It can be a solution of coil misalignment, distance variations, and strict power exposure safety regulations on human tissues. Moreover, it is no longer limited by $V_{\mathit{th}}$ and $P_{in,min}$, which are the biggest problems of the VM WPT. Second, the RCM WPT topology has no voltage conversion stage. Therefore, the unwanted power losses are eliminated thanks to the absence of the rectification circuit, DC/DC converter, and linear regulator.

Fig. 2.

a IPT circuit model, b IPT equivalent circuit model, c parameters of Rx/Tx coil

Fig. 3.

Compensation networks a SS, b SP, c PS, d PP

Table 1.

Compensation topologies – SS, SP, PS, PP

	Compensation topologies
	SS	SP	PS	PP
Transmitter capacitance	$\frac{C_R{L}_R}{L_T}$	$\frac{C_R{L}_R^2}{L_T{L}_R-M^2}$	$\frac{C_R{L}_R}{\frac{M^4}{L_T{C}_R{L}_R{R}_L}+L_T}$	$\frac{(L_T{L}_R-M^2)C_R{L}_R^2}{\frac{M^4{C}_R{R}_L}{L_R}+{\left(L_T,L_R,-,M^2\right)}^2}$
Transmitter quality factor ($Q_{\mathit{Tx}}$)	$\frac{L_R{R}_T}{\omega M^2}$	$\frac{\omega L_T{L}_R^2}{M^2{R}_L}$	$\frac{L_T{R}_L}{\omega M^2}$	$\frac{\omega L_T{L}_R^2}{M^2{R}_L}$
Receiver quality factor ($Q_{\mathit{Rx}}$)	$\frac{\omega L_R}{R}$	$\frac{R}{\omega L_R}$	$\frac{\omega L_R}{R}$	$\frac{R}{\omega L_R}$
Reflected resistance	$\frac{{\omega }^2{M}^2}{R}$	$\frac{M^{2}R}{L_R^2}$	$\frac{{\omega }^2{M}^2}{R}$	$\frac{M^{2}R}{L_R^2}$

Fig. 4.

Conventional Voltage Mode WPT system

Fig. 5.

a The RCM WPT structure and sub-circuits of each mode, b key waveforms, c common control mechanism of the RCM WPT

Major Challenges

1. Power Carrier Frequency

   As mentioned in Section 2(a), the higher power carrier frequency enhances PTE of the WPT systems. However, most of the RCM WPT systems have 50–1 MHz frequency ranges due to the delay of the ECD comparators [19, 51–53]. To resolve the delay issue, [54] suggests 13.56 MHz RCM WPT system with a diode that has a passive turn-off characteristic when the current becomes zero as the ECD. As a result, size of receiver coil can be reduced up to 7× 7 mm, which were tens of millimeter diameter in previous works. Figure 6 shows the correlation tendency of normalized coil size and carrier frequency among reported sub-watt RCM works.

2. Breakdown voltage

   The RCM WPT has been introduced that it has advantages for applications that need to harvest low input power. However, when the Rx coil received excessive power, then the mode transition switch $S_2$, as shown in Fig. 7(a), suffers from a negative voltage of $V_C$ [19, 56]. The voltage across the switch $S_2$ ($V_{S2}$) is $V_C-V_{\mathit{BAT}}$ in RESO while $V_C$ keeps increasing its amplitude as AC voltage. For example, the maximum amplitude of $V_C$ is limited to 1.7 V for 3.3 V Li-Ion battery charging system, which is designed with 5 V I/O transistors. Such a structure has not only the risk of switch break-down but the limitation of the efficiency for the power-hungry systems. To resolve the issue, a modified LC tank structure is suggested in [56], which is shown in Fig. 7(b). Thanks to this structure, it becomes possible to design the WPT system with a wider input range. Even if the circuit has changed, the reflected impedance of the LC tank in RESO is the same. Also, no matter how much $V_c$ increases, it has the advantage that the $V_{S2}$ is fixed to $V_{\mathit{BAT}}$ in RESO. In addition, the capacitor divider is connected in parallel with the resonance capacitor ($C_2$) so that exceeding voltage of $V_c$, which is one of the most crucial input parameters of the controller, cannot damage any control circuits. With this structure, extended input power range and improved efficiency are achieved.

3. Efficiency

   Two main efficiencies are justified in the WPT, power transfer efficiency (PTE), defined as transferred load power ($P_L$) divided by the power of the transmitter ($P_{\mathit{Tx}}$), power conversion efficiency (PCE), defined as $P_L$ divided by the power of receiver coil ($P_{\mathit{Rx}}$). The most common RCM WPT circuit, illustrated in Fig. 5(a), can be transformed into several ways with schemes of reported works to enhance PTE and PCE. In Table 2, the techniques and causes of improving and degrading efficiencies are summarized according to efficiency types, schemes of each circuit topology, and controller.

Fig. 6.

Normalized coil size and carrier frequency of the RCM works (3-dimensional coils [19, 55] are calculated with the largest part of the coil)

Fig. 7.

a A common receiver structure of the RCM WPT, b a modified receiver structure of [4]

Table 2.

Summary of the causes of efficiency increment and decrement among the reported RCM works

	Type	Scheme	Content	References
Increment	PTE	Circuit	Resonance topology: High-Q design of the Rx/Tx coils	Common
			Inductive coupling method: efficient for near-field transmission
			Extending input power range: low harvestable input power with -32 dBm sensitivity	[19]
			Extending input power range: increasing the voltage of input power via the modified receiver	[56]
		Controller	Extending input power range: reconfigurable VM/CM power management topology	[51, 53, 58]
		Circuit	Reflected impedance: only the LC tank is reflected to the transmitter thanks to Quasi-Resonant Boost-Converter (QRBC) topology	[54, 59]
		Controller	Reflected impedance: Q-modulation technique that the equivalent loaded-Q is changed itself at $R_L$ variation by modifying duty ratio	[43, 60]
		Circuit	Time-interleaved scheme that can proceed RESO and CHG simultaneously with two parallel capacitors structure	[54, 59]
	PCE	Circuit	Threshold voltage cancellation of the rectifier switches	[43, 54, 59, 60]
		Controller	Optimal mode transition: the optimal number of resonance cycles and precise duration of CHG to prevent over-discharge or residual energy in the LC tank	[19, 51, 53, 54, 58, 59]
			Zero-Current Detection as a single diode: not based on comparators	[54, 56, 59]
Decrement	PTE	Circuit	Rx/Tx coils: mis-alignment	[61, 62]
			Rx/Tx coils: huge distance	[51, 61, 62]
			Rx/Tx coils: losses by low-Q and self-resonant frequency	[61]
			Human tissue losses: EM wave absorption	[63]
	PCE	Controller	High-speed comparators: offset-controlled and limited bandwidth	[43, 52, 54, 59, 60]

Discussion

In this section, the-state-of-art RCM technologies for IMD applications and simplified operation analysis of the RCM WPT will be provided.

1. Reconfigurable VM/CM

   The conventional VM WPT systems pursue the higher PCE regardless of the voltage conversion efficiency (VCE) with sufficient $P_{\mathit{Rx}}$. However, some VM WPT systems tried to achieve the higher VCE with rectifiers or voltage multipliers under the circumstance of insufficient $P_{\mathit{Rx}}$ [36, 37, 57]. The effort to increase VCE had led to decrement of PCE by the voltage conversion stage. Likewise, in the RCM, even though PCE is relatively low, VCE greater than unity must be obtained by utilizing small $P_{\mathit{Rx}}$. Considering both characteristics of the VM and RCM, the reconfigurable VM/CM integrated power management (VCIPM) was suggested [53, 58]. The VCIPM system senses the received voltage of the Rx coil ($V_{\mathit{Rx}}$), thus it achieved PCE and VCE by mode transition to VM and RCM, respectively. When $V_{\mathit{Rx}}$ is greater than desired load voltage ($V_L$), it operates in the VM for higher PCE without considerations of the VCE. On the contrary, when $V_L$ exceeds $V_{\mathit{Rx}}$, it operates in the RCM to convert to desired $V_L$ with higher VCE. Developed from [57], it also achieved rectification, self-regulation and over-voltage protection (OVP) simultaneously with a single off-chip capacitor by letting it flow intentional reverse current from the load to the source.

2. Shared-Inductor Boost-Converter/CM

   The reconfigurable shared-inductor boost converter/current-mode topology is introduced [52]. The proposed system has a dual output capability, where the priority is determined depending on the magnitude of the received power. There are three power transfer priorities to the loads, (1) $R_L$, (2) $R_{\mathit{Hv}}$, (3) energy storing capacitor $C_s$, and six modes to optimally provide limited input power. First, as shown in Fig. 8(a), priority (1) and (2) can be satisfied with a shared-inductor boost-converter (SBC) topology that forms a series LC-$C_s$ tank to extract energy in $C_s$, which was pre-stored in previous operations, even though insufficient power is delivered to the receiver. Second, if sufficient power is received enough to satisfy priorities (1) and (2) simultaneously, it forms the LC tank except for $C_s$ to charge $C_s$ again with surplus energy, which means satisfying priority (3). At this point, the circuit operates as the conventional RCM topology with controller, called adaptive switching control (ASC), as shown in Fig. 8(b). Finally, the Rx coil will be opened not to receive additional energy when it already meets all priorities. Hysteresis comparators are used in control circuits for unnecessarily sensitive mode transitions causing PCE degradation and it also achieved extended output-power range via large external $C_s$ and voltage-power regulation with the six modes.

3. Time-interleaved Resonant Voltage-Mode

   In terms of the conventional RCM topology, an inductor, which is the Rx coil, served as a current source that directly charges a battery in CHG as well as a power receiver from the transmitter in RESO. However, the Rx coil of the RCM topology cannot receive the transmitted power continuously. The Rx coil receives the power only in RESO, and transfers energy to the load in CHG. Due to this discontinuous power transmission of the conventional RCM, a time-interleaved resonant voltage-mode (RVM) scheme with two parallel resonant capacitors was proposed in [54, 59]. The proposed RVM architecture has two phases of alternative capacitor connections so that the Rx coil only has the role of receiving energy, not a power source of the receiver circuit. As shown in Fig. 9, a circuit that receives power and a circuit that transfers received power to the load are separated. At this point, power sources are the capacitors of the LC tank, and these capacitors operate as voltage sources of the following a quasi-resonant boost-converter (QRBC) to charge the battery. With the proposed structure, in the conventional RCM, a comparator-based zero-voltage detection (ZVD) was required to detect the point when all energy in the LC tank was stored in the inductor, whereas a diode-based zero-current detection (ZCD) was used to detect the time of all energy stored in the capacitor, dramatically increasing the power carrier frequency. In terms of Rx/Tx coils, the load is always isolated, thus it can be designed with high-Q and have higher PTE compared to the other RCM systems. From the controller’s point of view, using digitally controlled feedback increases mode transition accuracy and PCE.

Fig. 8.

a A shared-inductor boost-converter, b an adaptive switching control supporting the RCM operation

Fig. 9.

A time-interleaved RVM scheme

Conclusion

This paper reviewed the WPT techniques for IMDs, especially focuses on resonant current-mode topologies. According to the best of our knowledge, it is the first review paper, which introduces the RCM topology, fundamentals, and operating characteristics. The RCM WPT systems show specific advantages that can overcome the limitations of the VM topology. Moreover, this paper suggested why the RCM WPT is advantageous to be implemented to the IMDs that have been developed toward low-power design. It also summarized the issues of the RCM for IMD applications, which have the goal of minimizing damage to human tissues and maximizing the power efficiency transmitted to the loads at the same time. This paper is expected to contribute to the research of a safer and more feasible battery recharging system for implantable devices, which are already embedded in the bodies of many global-wide patients. Finally, Table 3 summarizes the technical details of the-state-of-art RCM WPT works.

Table 3.

Summary of the-state-of-art RCM works

	[19]	[43]	[51]	[53, 58]	[52]	[54, 59]	[56]
Topology	CM	CM	CM	VM/CM	SBC/CM	RVM	CM
Process (μm)	0.18	0.35	Discrete	0.35	0.35	0.18	0.18
Active area (mm2)	0.554	4.8*	Discrete	0.52/1.56*	1.35	1.25	0.54
Power carrier frequency (MHz)	0.05	2	1	1	1	13.56	5
Rx coil size (mm)	2.6 × 3.5 × 11.7	65 × 65	30 × 30	30 × 30	22 × 22	7 × 7	21 × 21
Distance (mm) @Tx power	85 @20mW	80	70	60	-	10	10,23
Load voltage (V)	> 1.2	4.7	3.1	3.2	2.6/3.9***	0.5–3.0	1.8
$P_{in,min}$	0.6 $\mathrm{\mu }$ W	–	–	–	–	–	3.47mW
Max. PCE (%) @$P_{\mathit{in}}$	67.7 @4.2 $\mathrm{\mu }$ W	76 @1.45 W	–	77 @10mW	75.3 @4.7mW	95 @11.1mW	84.9 @16.8mW
Off-chip components C/L**	1/0	2/0	–	2/0	4/0	2/1	1/0
$L_{\mathit{Rx}}$(nH)	7,200	20,000	4,400	4,400	4,500	137	374
$C_{\mathit{Rx}}$(nF)	1.4	–	6****	6****	–	1.2	2.9
Quality factor $Q_{\mathit{Rx}}$	51	41.2	29	29	–	24	–
$L_{\mathit{Tx}}$(nH)	6,500	40,000	250,000	250,000	42,000	70	1,290
$C_{\mathit{Tx}}$(nF)	1,560	–	0.123	0.123	–	–	0.640
Quality factor $Q_{\mathit{Tx}}$	–	42.2	62.5	62	–	200	–
Power source to the load at CHG	L	L	L	C(VM)
/L(CM)	L	C	LC tank

*Including pads area

**Excluding the Rx coil

***Dual-output

****Values used for the simulation

Funding

This study was funded by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (2020R1G1A1006181).

Declarations

Conflict of interest

Kim J.-H. declares that he has no conflict of interest. Hassan NU declares that he has no conflict of interest. Lee S.-J. declares that he has no conflict of interest. Jung Y.-W. declares that she has no conflict of interest. Shin S.-U. declares that he has no conflict of interest.

Ethical Approval

This article does not contain any studies with human participants or animals performed by any of the authors.