Compact size antenna for skin implantable medical devices

Abstract

This research work presents an implantable antenna that operates at 5.8 GHz. By using a radiator with a loop-based design, the antenna can be made smaller. Radiator is made up of three connected rectangular loops. On the substrate’s back side, an I-shaped ground plane is used. As substrate and superstrate, polydimethylsiloxane (PDMS) with dimensions of 7 mm × 5 mm × 0.3 mm is used. The conducting sections are made using copper foil that is 30 µm thick. The suggested antenna is examined by the implantable medical device using realistic human scalp phantom models and a homogenous skin box. Simulated study revealed that it operates around 5.8 GHz with a bandwidth from 5.69 to 5.92 GHz. The specific absorption rate was 0.28 and 0.26 W/kg for skin box and human scalp phantoms, respectively, at 1 mW input power across 1 g volume tissue.

Introduction

Implantable medical devices (IMDs) proved their applications in medical field for detecting and monitoring biomedical health information of the patient and transmitting the data to external communication device for storage and processing. Implantable devices find their applications in neural recording, pressure sensing, glucose monitoring, pacemakers, etc.

The main aspect of these IMDs is to have an effective communication link that can be provided by a suitable compact sized antenna. Over the last years, several implant antennas have been developed based on the IMDs’ modeling and the required operational bands. While developing these implantable antennas, one of the main considerations is its size. Using high dielectric materials as substrates and superstrates, addition of the shorting pins between patch and ground can help in achieving lower resonant frequency while maintaining compact size. Improving the current path in the patch can also help in achieving the compact size. The modeling of the antenna needs to be developed based on the intended position of implantation because the presiding human tissues’ dielectric properties affect the operating frequency of the antenna. The performance of these IMDs depends on how well different design and surrounding aspect effect on the antenna performance are mitigated. The simulation setup needs to be based on the tissue type to be implanted. Moreover, the IMDs’ overall modeling can change basing on whether the implant can be just underneath the skin or for deep implant applications.

Over the years, several kinds of the implantable antennas are developed to serve single-band to multiband applications. In paper [1], a spiral-shaped planar inverted-F antenna (PIFA) antenna is developed to achieve the electrically small-scale antenna to be used for deep implant applications. Double split rings are used to design an extremely small antenna for brain implant applications in paper [2]. Rectangular spiral–shaped model is used in designing a compact sized antenna for leadless cardiac pacemaker systems in paper [3]. In paper [4], a capacitive loaded loop and a complementary split-ring loop along with interdigital capacitor loop are used to design a compact sized antenna, which serves as both sensor and antenna for cerebrospinal fluid monitoring. The changes in the dielectric properties of these fluids are detected due to the slits and gaps in antenna design which act as capacitor for sensing. A split-ring resonator model is used for developing the implant antenna for cortical and head-worn parts for the effective far-field implant communication applications in paper [5]. A meander lines-based slots are used in antenna development to achieve compact size to operate at ISM band around 2.45 GHz in paper [6]. Another meander line-based antenna with differential feed technique to enhance its performance is designed in paper [7].

For any implantable antennas, the affects of the surrounding encase circuitry and other elements inside the device on resonant frequency and operational band need to be mitigated to achieve better performance. In paper [8], the step impedance matching of designed model, using circular ring models, is used to have wider bandwidth and detuning the affects for multiple biotelemetry applications. IMDs having wide bandwidth can help in high data rate transmissions. Metamaterial-inspired structure is loaded as the radiating element to achieve the miniaturized wideband antenna for implantable applications in paper [9]. Loop model along with printed capacitor models are used to achieve the compact sized efficient wideband implantable antenna in paper [10]. In paper [11], loop radiator model incorporated with split-ring resonator models are used to develop the wideband miniaturized implantable antenna. A sigma-shaped radiator positioned inside the C-shaped ground model is printed on a thin flexible substrate to achieve the wideband antenna with low specific absorption rate (SAR) in paper [12]. Some of these antennas can have dual operational bands. In paper [13] a fractal-based antenna with differentially fed mechanism is used to achieve the dual-band implantable antenna. A serpentine-shaped radiating element along with open-end slot on ground is used to design miniaturized dual-band antenna for scalp implant applications in paper [14]. Circular-shaped antennas with slots, vias, and open-end slots on ground plane along with superstrate are used to achieve compact sized dual-band antenna in paper [15]. In paper [16], a simple radiating model with unique slots etched on radiator and a circular slot with one side open is made on ground helped in achieving the miniaturized dual-band antenna. A dual-band antenna is designed by using a novel radiating element along with continuous T-shaped slots on ground plan for skin and scalp implantable antenna in paper [17]. Meander line–shaped structures are incorporated to design a skin implantable dual-band operational antenna for intracranial pressure monitoring in paper [18]. A metamaterial-inspired structure is loaded as antenna, and a different metamaterial structure is used on superstrate to enhance the antenna performance and the circular polarization behavior in paper [19]; this antenna can serve dual-band for skin implant applications. A feed network based on rotational symmetrical stubs along with ground having L-shaped slots is used together to develop a dual-band antenna in paper [20]; this unique design achieved dual polarization capabilities.

Similarly, different antenna designs are used to achieve tri-band or multiband operational implant antennas. In paper [21], a meander line model and an open slot along with multiple slots on ground are used to achieve the miniaturized triple band antenna for implants. A novel PIFA with a slot on ground plan is used to develop multiband skin implantable antenna in paper [22]. Spiral-shaped structure is used to implement a multiband antenna that can serve in scalp implantable and leadless pacemaker applications in paper [23]. This implantable antenna can communicate with external antennas to convey the monitored data. The wearable antennas which can be used for telemedicine applications [24] can be used as external antennas. These antennas used high dielectric materials as substrates to achieve the low frequency of operation. Some literature metallic vias are also incorporated. The high dielectric human tissues also help in achieving the low resonant frequency. These human tissues equivalents can be prepared to evaluate the antennas [25]. The current work used a polydimethylsiloxane (PDMS) material which is hydrophobic in nature won’t react with any fluids, and no metallic pin vias are used in this design. It is more designed to be more compact than the existing models, and the detuning of the shift in operating frequency is addressed to improve its robustness toward surroundings.

In our current work, the loop-based structure is used to develop a compact sized implantable antenna. While three loops form the radiating element, an I-shaped model is used as ground plane. The loop-based patch design helped in miniaturing the antenna. The detuning is done using one more layer of Rogers 6010 material with one-side copper cladding placed below the antenna. The resonance frequency at 5.8 GHz helps in making the antenna compact sized. The loop designs especially center loop and the ground plane dimensions helped in tuning the operating frequency. The antenna is achieved operating frequency at 5.8 GHz with 240 MHz bandwidth. The dimension of the antenna is restricted to only 5 mm × 7 mm.

Design scenarios for simulation

The proposed antenna design, IMD modeling, and the considered simulation environment setup are discussed in this section.

Antenna design

The detailed antenna design is presented in Fig. 1, and the designed antenna along with proposed simulation model are shown in Fig. 2. The dimensions of the antenna are listed in Table 1. The antenna radiator design is proposed based on loop structures. Three rectangular loops are connected to form the radiator as shown in Fig. 1(a). The length and widths of the loop structures and the gaps between them help in tuning the antenna and also improving the impedance matching for the required operating frequency. In this work, a low dielectric PDMS material is used; however, the compact size is achieved by utilizing the loop-based structure. An “I”-shaped partial ground is used in this design. The shape of ground can easily help in tuning the antenna to required operating frequency. The ground plan is shown in the Fig. 1(b), and the center width of the I-shaped ground plays a major role in tuning the operational band easily from lower to upper frequencies compared to other geometrical parameters of the antenna design. Conducting parts of the antenna are implemented using copper foil of thickness 30 µm using electrical discharge machining tool. PDMS having dielectric constant of 2.7 and loss tangent of 0.314 is used as substrate and superstrate. Ansys HFSS v19.2 is used for design and simulation analysis.

Figure 1. Detailed design of the implantable antenna design. (a) Top view, (b) Back view and (c) Side view.

Figure 2. Overview of the IMD along with antenna.

Table 1. Dimensions of loop implant antenna (Unit: mm)

Considered IMD model

The IMD arrangement along with antenna is illustrated in Fig. 2. The device is made of antenna, printed circuit board (PCB) circuitry having electronics, and sensory packs and a battery; all of these components are enclosed into a biocompatible encase [21]. It contains battery of size 3.3 mm (diameter) and 2.4 mm (height), Rogers 6010 with dielectric constant 10.2 is used as PCB board, and other components are given perfect electric boundary to take into consideration of their effects on the antenna performance.

Phantom models and simulation setup

The simulation setup for skin tissue and scalp skin tissue is presented in Fig. 3. For skin tissue homogeneous skin phantom (HSP), the wet skin dielectric properties at 5.8 GHz are reported from paper [26]. Heterogeneous human model is used to analyze the scalp skin implant. The IMD along with antenna is placed 2 mm under the skin or scalp skin part of the human phantom to carry out the simulation.

Figure 3. Simulation setup for scalp implant and skin implant analysis.

Parametric studies

By using the HSP model shown in the Fig. 3, the parametric study for various dimensions of the antenna described in Fig. 1 is carried out. The affect of the variation of these dimensions on reflection coefficient is presented in Fig. 4.

Figure 4. Parametric study of dimensions L3, W3, L7, and W6.

From Fig. 4, it can be seen that, the internal loop dimensions L3 and W3 tune the operating frequency by 200 MHz range and the impedance matching is varying based on change in dimension. The dimensions for which the resonant frequency is exactly maintained at 5.8 GHz and the impedance matching is maintained good are selected. The ground dimensions L7 shows less affect, whereas W6 affects more. Here the W6 indicates the width of I-shaped ground. If the width of the ground is reduced then it affects the impedance matching and degrades the reflection coefficient. The variation of parameter W6 of the ground part showed more shift in operating frequency. I-shaped ground can be adjusted to tune operating frequency more easily. A similar parametric study is carried out for all dimensions that produce required operating frequency (Table 1). For further understanding of the antenna design, step-by-step implementation under HSP and its effect on reflection coefficient are presented in Fig. 5.

Figure 5. Step-by-step implementation of skin implantable antenna.

Antenna-simulated performances under HSP condition

The proposed antenna simulated under HSP showed the following performance illustrated in Fig. 6. From Fig. 6(a), we can observe that, the antenna operates at 5.8 GHz with return loss of −19.49 dB, it maintained operational band from 5.69 to 5.93 GHz with a bandwidth of 240 MHz. Figure 6(b) shows that the antenna maintained a gain around −32dBi when it is placed inside HSP. The electric field distribution in the IMD when the antenna is active is shown in Fig. 6(c).

Figure 6. Simulated analysis. (a) S₁₁ vs frequency, (b) 2 D gain, and (c) E field.

Circuit affect and detuning

When the antenna is simulated under phantom model, the efficacy of the device mainly depends on how the presence of the other components affects on performance is mitigated. The change in antenna performance with respect to reflection coefficient can be detuned for the without and with circuit cases. From Fig. 7(a), it can be noted that the operational band shifted when the circuit part is removed from the simulation setup. Additional designs made of copper foil can be used to detune the shifted operating frequency. The designs can use slots, stubs, or metamaterial-inspired structures. These detuning plates can be placed under antenna. The dimensions of these plates can be verified using parametric study to verify their effectiveness in detuning the shifted operating frequency when used. In this design, an additional copper plate is placed below the antenna with gap and the analysis showed reflection coefficient maintained with or without circuitry, as shown in Fig. 7(b). Similarly, different detuning approaches can be applied to mitigate the affects of the antenna positioning inside the IMDs, as they affect the performance. By detuning these affects, the efficacy of the antenna can be improved.

Figure 7. S₁₁ vs frequency comparison (a) Without circuit – without detuning and with detuning (b) After detuning without circuit and with circuit.

Circuit and tissue affects

From Fig. 7(b), it can be observed that after the detuning method is applied the reflection coefficient response for no circuit and circuit case are maintained similar. Additionally, we discuss two more cases here in this section: The selected two cases are no circuit and circuit cases for without tissue condition. The comparison between these four cases is shown in Fig. 8.

Figure 8. S₁₁ vs frequency comparison for cases of circuit and tissue presence or absence.

From Fig. 8, one can see that, apart from the analysis that is retaken from Fig. 7, the additional two cases for without tissue condition showed shift in operating frequency toward 6 GHz. Now the antenna radiation patterns and SAR analysis under HSP and realistic scalp phantom are analyzed and presented in the following sections to prove its suitability for IMDs.

Results and discussions

The antenna analyses under HSP and scalp model are presented in the previous section. The analysis showed similar reflection coefficient curve response for both HSP and scalp skin phantoms. The antenna radiation patterns and SAR analysis under HSP and realistic scalp phantom are analyzed and presented to prove its suitability for IMDs.

Radiation patterns

The radiation patterns in XY, YZ, and ZX planes of the proposed IMD model under skin layered phantom are analyzed using Ansys SAVANT, as shown in Fig. 9.

Figure 9. Simulated radiation patterns in XY, YZ, and ZX planes under HSP simulation setup.

The radiation patterns in Fig. 9 show that, most of the radiation energy is present in forward direction. Here the implant device is positioned in a way that the patch antennas are facing toward skin.

SAR analysis

SAR is the amount of energy that is absorbed by human tissue when the relevant antenna is in actively radiating. It can be evaluated by (1):

(1)\begin{equation}SAR = {{\sigma {{\left| E \right|}^2}} \over {\rho}}\end{equation}

where E is the electric field strength within the tissue (V/m) [27], σ is conductivity (S/m) of the relevant tissue, and ρ is the tissue density (kg/m³). The SAR is analyzed following Federal Communications Commission (FCC) standards over 1 g volume tissue. These IMDs are very low powered around 25 µW to avoid the damage to human tissues by radiation. In general, the allowable SAR value by FCC is 1.6 W/kg. The SAR analysis comparison is presented in Figs. 10 and 11 for the HSP and the realistic scalp models. For the realistic scalp model, a portion of the scalp part of the human phantom is used. As the implant is intended to be placed under skin of this scalp part, the skin tissue properties are assigned for this realistic phantom part. The SAR values over different input power level for HSP and realistic scalp models are listed in Table 2. For both simulation setups, at input power 1 mW, the antenna maintained low SAR values of 0.28 and 0.26 W/kg, respectively.

Figure 10. HSP model: simulated SAR vs input power.

Figure 11. Realistic scalp model: simulated SAR vs input power.

Table 2. SAR vs input power for HSP model and realistic scalp model

Fabrication and measurements

The fabricated implant antenna prototype is shown in Fig. 12(a). Here the PDMS is developed by mixing the base and curing agents (sylgard 184) in 10:1 ratio and annealed at 60°C. Electrical Discharge Machining (EDM) wire cutting tool is used for preparing the conductive parts with precision. The measuring setup is shown in Fig. 12(b). The chicken breast purchased from grocery shop is used for measuring setup as its dielectric properties are equivalent to human skin [28, 29]. The fieldfox microwave analyzer N9915A is used for measuring S₁₁ vs frequency curve. The simulated reflection coefficient curve in HSP and realistic scalp models are compared with measured data under chicken breast, which is presented in Fig. 13.

Figure 12. (a) Fabricated loop implantable antenna and (b) Measuring setup.

Figure 13. S₁₁ vs frequency comparison for simulated under HSP and scalp phantom with measured under chicken breast (without circuit case).

Figure 13 indicates the reflection coefficient comparisons for no circuit case. Here while the simulated and experimental validation shows that antennas operate at 5.8 GHz. Some discrepancy is observed as in simulation the HSP box is used, but for experimental validation the chicken breast geometry differs.

Comparison with previous works

The implantable loop antenna performance is compared with previous existing models in Table 3.

Table 3. Implantable loop antenna comparison with previous works

In the current work, XY direction has same size as that in paper [30], but the addition of the superstrate and the change in thickness made the current antenna volume slightly more than paper [30]. But in paper [30], the antenna alone developed without any consideration of the IMD. In the current work, the surroundings affects and the detuning method are presented to show its capability to maintain the intended operating frequency. Compared to paper [30], the current work may have less gain but size of the current model is reduced 32 times.

Conclusion

In this paper, a compact sized antenna for 5.8 GHz ISM band skin implantable applications is implemented. The design is mainly based on loop-based structures. The dimensions of the loops in radiator element and the “I”-shaped ground are tuned to achieve the required frequency of operation at 5.8 GHz with good impedance matching. The overall size of the skin implantable antenna is only 19.6 mm³. The simulated analysis showed that, the proposed antenna operated at 5.8 GHz with 240 MHz bandwidth in both HSP and scalp phantom models. The measured reflection coefficient data showed in Fig. 13 had small shift in operating frequency, with more bandwidth’s size compared to simulated analysis. The specific absorption analysis at 1 mW showed that the proposed IMD model has only 0.28 and 0.26 W/kg in HSP and scalp phantom models which is under the prescribed SAR standard value by FCC of 1.6 W/kg. The antenna maintained gain around −32dBi with Omni coverage opposite to the phantom presence. The developed antenna is suitable candidate for implanting at skin or scalp skin positions to operate at 5.8 GHz ISM band for monitoring applications.