A Novel 1-9 Ghz On-Body Antenna for Low-Loss Biomedical Telemetry with Implants

Home , Popeye

This is the author's version of an article that has been published in this journal. Changes were made to this version by the publisher prior to publication. The final version of record is available at http://dx.doi.org/10.1109/TAP.2018.2889159 > AP1804-0782< 1

Bio-Matched Horn: A Novel 1-9 GHz On-Body Antenna for Low-Loss Biomedical Telemetry with Implants

John Blauert, Student Member, IEEE, and Asimina Kiourti, Member, IEEE

monitoring/control equipment. Such wireless communications Abstract— We present the Bio-Matched Horn antenna: a are typically performed in approved frequency bands, novel 1.773 cm3 on-body antenna for very-low-loss including the Medical Device Radiocommunications Service communication with wireless implants across the 1 to 9 (MedRadio) band [17] of 401-406 MHz or the Industrial, GHz bandwidth. To achieve this unique performance, the Scientific and Medical Radio (ISM) bands of 2.4-2.5 GHz and Bio-Matched Horn is composed of water-filled holes used 5.725-5.875 GHz [18]. to mimic the frequency-dependent permittivity of the Accordingly, design of on-body antennas that are underlying tissue over its entire bandwidth. The horn specifically optimized for biomedical telemetry with wireless shape further allows the proposed antenna to implants has attracted considerable scientific interest [19]- communicate efficiently through biological tissue for both [21]. Expectedly, design of such on-body antennas is subcutaneous and deep-tissue wireless implants. associated with a number of challenges, including: Measurement results at 2.4 GHz are in good agreement (a) Mismatch at the biological tissue and antenna interface. with simulations and show a transmission loss of 8.01 dB to The relative permittivity of human skin is noticeably higher (ε a 4-mm-deep implant and 19.21 dB to a 2-cm-deep r = 46.79 at 400 MHz and ε = 38.06 at 2.4 GHz) [22] than that implant. Compared to previous designs, this is a r remarkable improvement by 14.5 dB and 10.8 dB for of the surrounding free-space (εr = 1). Exterior antenna designs subcutaneous and deep-tissue communication, reported to date for telemetry with wireless implants do not respectively. In this paper, we discuss the operation take this high permittivity into account [7]-[9]. Instead, the principle and design of the Bio-Matched Horn, its most common approach entails the use of an interrogating performance for subcutaneous and deep-tissue implant antenna separated from the human body through free-space telemetry, sensitivity to rotational and positional [4]-[6], [11]-[12]. This approach has shown to introduce a misalignment, and Specific Absorption Rate (SAR) noticeable mismatch loss of 17.42 dB at the skin-air interface performance. at 2.4 GHz [9]. On-body antennas have been reported, but the relative permittivity of their substrate is typically εr ~10 [9], Index Terms— Biomedical telemetry, deep tissue implying, again, considerable mismatch losses. coupling, wearable antenna, wireless implants. (b) Environmental and inter-subject variability. Electrical properties of biological tissues can vary based upon factors I. INTRODUCTION such as temperature, age, gender, daily activities, etc. [23]. In HE rapid development of implantable medical devices for fact, research has demonstrated possible changes in Tsensing (brain implants, glucose monitors, etc.) and permittivity and conductivity as high as ±20% from their stimulation (pacemakers, defibrillators, etc.) [1]-[5] is poised nominal values [24]. Such changes may cause detrimental to reinvent medical healthcare as we know it. Developing detuning of the on-body antennas. To account for this, wireless implanted medical devices requires the development broadband antenna designs are needed [12]. Adaptive of efficient implantable antennas [4]-[16]. The latter are impedance matching techniques could potentially be an integrated within the implant and, depending on the alternative to using broadband designs [25]. Nevertheless, to application, they may operate in subcutaneous [4]-[8] and/or date, research on adaptive matching has mostly be limited to deep-tissue environments [9]-[12]. In turn, the utmost aim of inductive coupling rather than antenna-based communications these implantable antennas is to enable efficient wireless [26]-[27]. communication, namely biomedical telemetry, with exterior (c) Frequency-dependent tissue properties. Electrical properties of biological tissues are frequency-dependent, with Manuscript received April 27, 2018, revised August 28, 2018. This work permittivity and conductivity of skin changing by as much as was supported in part by the Juvenile Diabetes Research Foundation under 21% and 87% in the 1-9 GHz range, respectively [22]. Even if grant 2-SRA-2016-237-Q-R and by the Consortium on Electromagnetics and Radio Frequencies (CERF). matching of the antenna substrate is pursued to overcome J. Blauert and A. Kiourti are with the ElectroScience Laboratory, Dept. of challenge (a) raised above, such matching will only be valid at Electrical and Computer Engineering, The Ohio State University, Columbus, a given frequency. That is, broadband matching of an OH, USA (e-mail: [email protected], [email protected]).

TABLE I. SUBCUTANEOUS COUPLING: COMPARISON OF PROPOSED DESIGN VS. THE STATE OF THE ART Implantation Transmission Ref. Frequency Air Gap Size / Antenna Type Bandwidth Depth Loss 2.4 GHz 3.3 mm 4 mm π × 72.52 × 13.635 mm3 / Spiral 0.6-6 GHz 26 dB [5] 4.8 GHz 3.3 mm 4 mm π × 72.52 × 13.635 mm3 / Spiral 0.6-6 GHz 19 dB 400 MHz 4 mm 5 cm 375 mm / Dipole NA 34 dB [6] 2.4 GHz 4 mm 5 cm 62.4 mm / Dipole NA 32 dB [7] 2.4 GHz 4 mm 0 cm 26.3 × 30 × 1.6 mm3 / Spiral 2-11 GHz 22.5 dB This work 2.4 GHz 4 mm 0 cm 22 × 22 × 10 mm3 / Horn 1.4-8.5 GHz 8.01 dB

TABLE II. DEEP TISSUE COUPLING: COMPARISON OF PROPOSED DESIGN VS. THE STATE OF THE ART Implantation Transmission Ref. Frequency Air Gap Size / Antenna Type Bandwidth Depth Loss [8] 400 MHz 9 mm 0 cm 28 × 26.8 × 0.635 mm3 / Patch 380-470 MHz 24 dB π ×17.52 × 0.76 mm3 / Exponentially [9] 2.4 GHz 2 cm 0 cm 2.35-2.7 GHz 30 dB Tapered Slot [10] 2.4 GHz 3.5 cm 10 cm NA / Free-Space Horn 0.9-2.45 GHz 56.5 dB [11] 400 MHz 1 cm 1.5 cm 70 × 60 × 1.6 mm3 / Patch 350-450 MHz 50 dB [12] 2.4 GHz 5 cm 2.55 m NA / Free-Space Horn NA 81 dB This work 2.4 GHz 2 cm 0 cm 22 × 22 × 10 mm3 / Horn 1.4-8.5 GHz 19.2 dB antenna’s permittivity to the underlying tissue environment As highlighted in Tables I and II, the Bio-Matched Horn proves challenging via traditional approaches that rely on offers a promising alternative to existing on-body antenna ceramic laminates (i.e., substrates with frequency-independent designs for wireless telemetry with both subcutaneous (<5- properties). mm-deep) and deep-tissue (>5-mm-deep) implants. As seen, (d) Inherent material loss of biological tissues. Biological the proposed antenna is remarkably efficient, offering 14.5 dB tissues are inherently lossy, implying detrimental effects in the lower transmission loss vs. previous designs [7] at 2.4 GHz for associated on-body-to-implant wireless communication [28]. a 4 mm implantation depth. Concurrently, the Bio-Matched For example, the loss tangent of skin and muscle tissues at 2.4 Horn offers 10.8 dB lower transmission loss vs. previous GHz is equal to 0.284 and 0.242, respectively [22]. As a designs [9] at 2.4 GHz for a 2 cm implantation depth. comparison, the loss tangent of free-space is typically Although the proposed antenna has a 1 cm height, which is not approximated as zero, while conventional ceramic and optimal for wearables, the antenna was constructed as proof- polymer substrates exhibit loss tangents of <0.02. Efficient, of-concept design for the heterogeneous water-based dielectric high-gain on-body antenna designs are, thus, needed to matching technique. overcome the associated wireless path losses. The rest of the paper is organized as follows. Section II To mitigate these issues, we propose the Bio-Matched discusses the operation principle and design of the Bio- Horn antenna: a miniaturized (1.773 cm3) on-body horn Matched Horn. Section III presents the employed antenna that employs an anisotropic material configuration to measurement and simulation set-ups. Section IV discusses the mimic the frequency-dependent electrical properties of antenna performance. The paper concludes in Section V. biological tissues across its 1 to 9 GHz bandwidth. Notably, developing a 1-9 GHz bandwidth allows for several benefits, II. OPERATION PRINCIPLE AND ANTENNA DESIGN as it: (a) covers the 1.4 GHz WMTS band commonly used in The proposed Bio-Matched Horn is shown in Fig. 1(a) with ocular implants [4], (b) covers the 2.4 GHz ISM band which is design parameters given in Table III. The antenna is compact, emerging as extremely popular for wireless implants (brain maintaining a volume of 1.773 cm3, and is intended for on- implants, glucose monitors, and so on [4]-[7],[9]-[10],[12]), body operation and telemetry with subcutaneous and/or deep- (c) covers the 5.8 GHz ISM band, which has already been tissue wireless implants. The key novelty of the Bio-Matched utilized in numerous implantable applications [29]-[30], (d) Horn design is its effective permittivity that has been covers several other unlicensed frequencies that have been engineered to match the frequency-dependent permittivity of used to date for wireless implants (e.g., 4.8 GHz [5], UWB the underlying skin tissue across the entire 1 to 9 GHz neural implants [7]), and (e) enables extreme tolerance in bandwidth. This is accomplished by including 2-mm-diameter matching against inter-subject and environmental variability cylindrical holes inside of the 3D-printed polylactic acid (εr = [12]. 3.549, tanδ = 0.001 [31]) horn structure, that are,

(a) (b) Fig. 1. Proposed Bio-Matched Horn antenna: (a) side 1 view with dimensions shown in Table III and fundamental unit cell shown in yellow, and (b) fabricated prototype.

(a) (b) TABLE III. DIMENSIONS OF THE BIO-MATCHED HORN ANTENNA Fig. 3. Experimental set-ups employed in this study: (a) POPEYE leg Parameter Value Parameter Value phantom, and (b) 80% lean ground beef phantom.

La 2 mm Lc 22 mm fundamental unit cell and shown to match the permittivity of Lb 17.3 mm D 2 mm skin [21] over a broad bandwidth, as shown in Fig 2. As seen, the difference between the two permittivities is < 2.86%

across the entire 1-9 GHz bandwidth. While this approach is 41

Skin limited by assuming infinite periodicity, it can be used to 40 Engineered Dielectric accurately develop an estimate of the effective permittivity of

39 the engineered lattice structure for electrically small unit cells [33]. As the frequency increases and the unit cell is no longer 38 electrically small (> ~ ), the antenna begins to no longer 37 휆 r appear effectively homogeneous. As a result, the antenna will 10 36 become mismatched close to 9 GHz, when the unit cell

35 approaches the aforementioned length. This will be discussed in Sec. IV.A. 34 The Bio-Matched Horn was 3-D printed using a Raise3D 33 N2 printer [34] and polylactic acid filament (εr = 3.549, tanδ =

32 0.001 [31]). Once the plastic container was printed, the holes 1 2 3 4 5 6 7 8 9 of the Bio-Matched Horn were filled with distilled water. Frequency (GHz) Fig. 2. Relative permittivity of skin tissue vs. effective permittivity of Super glue was then applied along the edges of the horn, and the Bio-Matched Horn antenna’s engineered dielectric, as a function of copper tape was placed on top of the adhesive to seal in the frequency. water. In this particular case, copper tape serves as the conducting flares of the Bio-Matched Horn. The fabricated subsequently, filled with distilled water. Referring to Fig. 1, Bio-Matched Horn antenna prototype is shown in Fig. 1(b). To the cylinders penetrate from side ‘1’ to side ‘3’ of the horn, excite the antenna, an SMA connector was used. While this is running in parallel to the yz-plane. The fundamental not optimal (size of the SMA is comparable to the antenna hypothesis of the design is that as the skin’s permittivity size), it is adequate in assessing antenna performance. The decreases with frequency, the permittivity of the distilled latter was confirmed via simulations with and without the water in the Bio-Matched Horn will also decrease accordingly, SMA connector, indicating insignificant discrepancies thus matching the horn to human tissue over a broad between the two. bandwidth. Notably, biological tissues are rich in water concentration; hence their frequency-dependence follows very III. MEASUREMENT AND SIMULATION SET-UPS much the same pattern/trend as that of water. Performance of the Bio-Matched Horn was experimentally In order to use the minimum amount of distilled water, the evaluated upon tissue-emulating phantoms, and specifically engineered dielectric was designed to be anisotropic, upon: a) SPEAG’s POPEYE leg phantom, (Zurich developing a higher permittivity parallel to the cylinders and a Switzerland [35]), Fig. 3(a), and b) an 80% lean ground beef lower permittivity orthogonal to the cylinders [32]. For this phantom, Fig. 3(b). The POPEYE phantom enables accurate proof-of-concept design, the size of the water holes was representation of the tissue anatomy and average electrical chosen to be 2 mm in order to be electrically small enough to properties. However, it is sealed and does not allow be effectively homogeneous within the ISM bands, but large implantation of wireless implants for biomedical telemetry enough to be 3D printed. The overall antenna size was chosen evaluation. To achieve the latter, a ground beef phantom was to fit in multiples of the fundamental unit cell. The effective employed, that has long been used in the literature to emulate relative permittivity between the flares was calculated using a the frequency-dependent electrical properties of biological single harmonic Fourier convolution matrix of the tissues [5]. Permittivity and conductivity of both phantoms

40 8 -5 Simulated with Calculated Effective Permittivity Beef Permittivity Simulated with Actual PLA/Water POPEYE Permittivity 7 Beef Conductivity -10 35 POPEYE Conductivity 6

-15 5 30

r 4 -20

25 3 Conductivity (S/m) -25 2 20 Reflection Coefficient |S11| (dB) 1 -30

15 0 1 2 3 4 5 6 7 8 9 -35 Frequency (GHz) 1 2 3 4 5 6 7 8 9 Fig. 4. Electrical properties of the POPEYE and ground beef phantoms. Frequency (GHz) Fig. 6. Simulated reflection coefficient (|S11|) of the Bio-Matched Horn using the calculated effective permittivty from Fig. 2 vs. using the actual heterogeneous structure.

-5

Beef Simulated Beef Measured -10 POPEYE Simulated POPEYE Measured

-15

-20 Fig. 5. Simulation set-up showing the Bio-Matched Horn placed against a rectangular tissue-emulating phantom. -25

measured via Keysight’s 85070E high-temperature probe at 1- Reflection Coefficient |S11| (dB) -30 9 GHz are shown in Fig. 4. Expectedly, the ground beef phantom exhibits higher permittivity and conductivity values, -35 since it does not include bone tissue which is inherently low- 1 2 3 4 5 6 7 8 9 permittivity and low-loss. For comparison, the permittivity of Frequency (GHz) 2/3 muscle tissue that is often used to emulate the average Fig. 7. Simulated vs. measured reflection coefficient (|S11|) of the Bio- Matched Horn. tissue properties [36] varies from 36 to 28 in this frequency range [22]. Accordingly, the conductivity of 2/3 muscle tissue varies from 0.65 to 7 S/m [22]. A. Reflection Coefficient Finite Element (FE) simulations for the Bio-Matched Horn To assess the suitability of our engineered dielectric were performed in Ansys High Frequency Structure Simulator calculations in Fig. 2, we first contrasted the antenna reflection (HFSS). As shown in Fig. 5, the antenna was placed against a coefficient assuming: a) the intended heterogeneous structure 7 cm × 7 cm × 6.4 cm rectangular tissue-emulating box. of Fig. 1, and b) an effective medium block with properties Permittivity and conductivity of the aforementioned box equal to those calculated in Fig. 2. Both antennas were represented the measured electrical properties of the POPEYE simulated using the set-up of Fig. 5, with the phantom and/or the ground beef phantom, per the study requirements. exhibiting properties equal to those reported for beef in Fig. 4. As seen in Fig. 6, good agreement exists between the two, IV. PERFORMANCE RESULTS particularly at frequencies below 5 GHz. Concurrently, the Simulation and measurement results are hereafter presented reflection coefficient is lower than the target -10 dB across the for the Bio-Matched Horn antenna: a) reflection coefficient, b) largest portion of the 1-9 GHz bandwidth. Any discrepancies electric fields, c) transmission loss with subcutaneous and are attributed to the fact that our effective permittivity deep-tissue implants, d) sensitivity to misalignment, e) calculation assumed infinite periodicity. As also discussed in sensitivity to water leakage, and f) SAR performance. Focus is Sec. II, the unit cell becomes electrically large at higher on the 2.4 GHz ISM band [18], which has been repeatedly frequencies and is no longer effectively homogeneous. reported in the literature for biomedical telemetry with The simulated vs. measured reflection coefficient wireless implants [5]-[7], [9]-[12]. Of course, similar results performance of the Bio-Matched Horn antenna is shown in can readily be derived at any of the operating frequencies of Fig. 7. Two simulation set-ups were considered with the the Bio-Matched Horn across its 1-9 GHz bandwidth. tissue-emulating box of Fig. 5 emulating the POPEYE and

-5

-10 Simulated Measured -15

-20

-25

-30 Reflection Coefficient |S11| (dB) -35 1 1.5 2 2.5 3 3.5 4 Frequency (GHz) (a) (b) Fig. 9. Implantable patch antenna used in the transmission loss studies: (a) fabricated prototype, and (b) measured vs. simulated reflection coefficient performance.

To account for the wireless implant, a proof-of-concept implantable antenna was employed, as shown in Fig. 9(a). This antenna exhibited a patch geometry with a footprint of

Fig. 8. Antenna electric fields at 2.4 GHz, showing directive radiation 18.5 mm × 11 mm, and was not miniaturized or optimized in normal to the antenna/tissue surface. any way. The prototype was milled upon a 1.27 mm thick

Rogers TMM 13i substrate (εr = 12.85 and tanδ = 0.0019), and ground-beef properties of Fig. 4, respectively. Accordingly, was further coated with a 0.5 mm layer of two measurement set-ups were considered that employed the polydimethylsiloxane (PDMS) (ε = 2.8 and tanδ = 0.001) to POPEYE and ground-beef phantoms of Fig. 3. r ensure biocompatibility [37]. The simulated vs. measured As seen in Fig. 7, the Bio-Matched Horn matches reflection coefficient of this implantable patch is shown in Fig. remarkably well across the 1-9 GHz bandwidth, despite the 9(b). As expected, the antenna operates in the 2.4 GHz ISM POPEYE phantom having only ~60% of the permittivity of band, with good agreement between simulations and the ground beef. Notably, measurements are in very good measurements. The slight detuning in the measured reflection agreement with simulations, with the Bio-Matched Horn only coefficient most likely stems from inaccuracies in the being slightly detuned, most likely due to slight water leakage. thickness of the manually-employed PDMS coating. Indeed, measurements were repeated multiple times with Transmission loss results between the Bio-Matched Horn varying amounts of water filling the holes to confirm the and the patch antenna of Fig. 9(a) implanted subcutaneously latter. When placed upon the POPEYE phantom (Fig. 3(a)), and deep into tissue are shown in Fig. 10(a) and Fig. 10(b), the reflection coefficient remains below -10 dB across the respectively. In measurements, layers of beef were added bandwidth of 1.0-8.01 GHz, with the exception of a small incrementally upon the patch to achieve the desired section that is -9 dB from 4.72-5.25 GHz. When placed upon implantation depth. Expectedly, as implantation depth the ground beef phantom (Fig. 3(b)), the reflection coefficient increases, transmission loss to the underlying implant remains below -10 dB across the bandwidth of 1.35-8.81 GHz, increases linearly on a dB scale [38]. Notably, great agreement with the exception of a small section that is -9 dB from 4.60- is observed between simulations and measurements. Measured 5.65 GHz. loss was, on average, ~1 dB higher than simulations for B. Electric Field Distribution subcutaneous testing. This falls within the range of anticipated The electric fields radiated by the Bio-Matched Horn inside measurement error, most likely attributed to slight water the tissue-emulating phantom of Fig. 5 at 2.4 GHz are shown leakage. For deep-tissue testing, measured loss was, on in Fig. 8. For this particular example, the POPEYE electrical average, ~3.3 dB higher than simulations. Compared to the properties were used in simulations. As seen, the Bio-Matched subcutaneous scenario, discrepancies increase, as expected Horn keeps the orientation of the fields normal to the given the added difficulty in properly placing and aligning the placement of the antenna, allowing them to propagate deep antenna at deep depths. into the tissue in a directive manner. Importantly, and as also highlighted in Tables I and II, the proposed Bio-Matched Horn considerably outperforms state- C. Transmission Loss with Wireless Implants of-the-art on-body antenna designs for communication with Simulations and measurements were carried out to assess wireless implants. Considering the subcutaneous implant the feasibility of the Bio-Matched Horn antenna in scenario of Fig. 10(a), the Bio-Matched Horn offers at least communicating with both subcutaneous (<5-mm-deep) and 14.2 dB lower loss than what has been reported in [5]-[6]. At a deep-tissue (>5-mm-deep) wireless implants. The ISM depth of 2 cm (see Fig. 10(b), the Bio-Matched Horn has a frequency of 2.4 GHz was again considered in our studies. To mere 19.21 dB of loss as compared to [9] which had 30 dB of enable implantation, the ground beef phantom of Fig. 3(b) was loss. Although [10]-[12] have an additional mismatch loss employed. Accordingly, the simulation set-up of Fig. 5 was set between free space and biological tissues, the Bio-Matched to emulate the ground beef properties of Fig. 4.

(a)

(b) Fig. 10. Transmission loss between the Bio-Matched Horn antenna and the implanted patch for: (a) subcutaneous, and (b) deep-tissue

Depth: 4 mm 70 Depth: 20 mm

(b) 60 Fig. 12. Transmission loss at 2.4 GHz between the Bio-Matched Horn

50 and an implantable patch antenna immersed at: (a) 4 mm, and (b) 20 mm, as a function of positional misalignment. 40 TABLE IV. BIO-MATCHED HORN SENSITIVITY TO WATER LEAKAGE 30 Reflection Transmission Transmission Loss (dB) 20 Water Content Coefficient at Coefficient at 2.4 GHz 2.4 GHz 10 Air-Filled Horn -2.55 dB -19.6 dB

0 0 10 20 30 40 50 60 70 80 90 0.5 mm Water Holes -2.97 dB -19.2 dB Rotation Angle (Degrees) 1 mm Water Holes -5.24 dB -17.8 dB Fig. 11. Tranmission loss at 2.4 GHz between the Bio-Matched Horn and an implantable patch antenna immersed at 4mm and 20mm depth, as a 1.5 mm Water Holes -13.0 dB -16.6 dB function of rotational misalignment. 2 mm Water Holes -22.3 dB -16.5 dB (Proposed Design) Horn has at least 12.78 dB lower transmission loss even when subtracting out any losses associated with mismatch [10]. (subcutaneously), and b) 20 mm (deep-tissue) inside the D. Sensitivity to Rotational and Positional Alignment phantom. The Bio-Matched Horn was subsequently rotated The Bio-Matched Horn maintains a linear polarization, about its central axis and the transmission loss between the which causes it to be dependent on rotational alignment for horn and implanted patch was recorded, Fig. 11. communication with wireless implants. To assess this effect As seen, perfect alignment yields 7.4 dB of transmission on the associated transmission loss, simulation studies were loss in the subcutaneous scenario and 16.9 dB of transmission performed at 2.4 GHz. To do so, the set-up of Fig. 5 was loss in the deep-tissue implantation scenario. For both depths, employed, with the ground beef electrical properties (see Fig. the Bio-Matched Horn can be misaligned by up to 45° and 4) assigned to the tissue-emulating box. The Bio-Matched remain within 3 dB of the perfectly matched polarization. At Horn was placed right upon the air-tissue interface, while the an angular offset of 90°, the transmission loss increases implantable patch antenna of Fig. 9(a) was immersed: a) 4 mm significantly. This is attributed to the fact that the Bio- Matched Horn’s polarization and the patch antenna’s

Copyright (c) 2019 IEEE. Personal use is permitted. For any other purposes, permission must be obtained from the IEEE by emailing [email protected]. This is the author's version of an article that has been published in this journal. Changes were made to this version by the publisher prior to publication. The final version of record is available at http://dx.doi.org/10.1109/TAP.2018.2889159 > AP1804-0782< 8 this antenna in communication systems to various implanted antennas using pigs as model animals,” IEEE Antennas and Wireless medical sensors, stimulators, and/or energy harvesters, b) Propagation Letters, vol. 11, pp. 1686-1689, 2012. [16] A. Alemaryeen and S. Noghanian, “Crumpling effects and specific developing more reliable manufacturing with smaller absorption rates of flexible AMC integrated antennas,” IET Microwaves, excitation connectors and improved water sealing capabilities, Antennas, & Propagation, vol. 2, iss. 4, pp. 627-635, Mar. 2018. c) expanding to conformal antenna designs, and d) conducting [17] FCC Service Rules, 47 C.F.R. Part 95, “Medical Device parametric and statistical analyses of antenna design Radiocommunications Service (MedRadio),” Current as of April 16, 2018. parameters and related measurement errors. [18] OET Bulletin.65, 1997. “ITU-R Radio Regulations section 5.138 and 5.150,” Int. Telecommun. Union-Radiocommun., Geneva, Switzerland. ACKNOWLEDGMENTS [19] S. Rao et al., “Miniature implantable and wearable on-body antennas: towards the new era of wireless body-centric systems” IEEE Antennas The authors would like to acknowledge Matthew and Propagation Magazine, vol. 56, no.1, pp. 271 - 291 Feb. 2014. Patterson of Rose-Hulman Institute of Technology for 3D [20] S. Ma, L. Sydänheimo, L. Ukkonen, and T. Björninen, “Split-Ring printing the Bio-Matched Horn antenna base. Resonator Antenna System With Cortical Implant and Head-Worn Parts for Effective Far-Field Implant Communications,” IEEE Antennas and Wireless Propagation Letters, vol 17, iss. 4, pp. 710 – 713, April 2018 REFERENCES [21] L. Ukkonen, L. Sydänheimo, and Y. Rahmat-Samii, “Embroidered [1] V. Ranganathan, et al, “A high-voltage compliant neural simulator with textile antennas for wireless body-centric communication and sensing,” HF wireless power and UHF backscatter communication,” in Proc. Antennas & Propagation Conference (LAPC), 2-3 Nov. 2015 Wireless Power Transfer Conference, May 2016. Loughborough, UK. [2] P. Anacleto, P. M. Mendes, E. Gultepe, and D. H. Gracias, “3D small [22] C. Gabriel, “Compilation of the dielectric properties of body tissues at antenna for energy harvesting applications on implantable micro- RF and microwave frequencies,” Report N.AL/OE-TR- 1996-0037, devices,” in Proc. 2012 Loughborough Antennas & Propagaion Occupational and environmental health directorate, Radiofrequency Conference, Nov. 2012. Radiation Division, Brooks Air Force Base, Texas (USA), June 1996. [3] C. Poon, B. Lo, M. Rasit Yuce, A. Alomainy and Y. Hao, “Body Sensor [23] K. Bocan, M. Mickle and E. Sejdic, “Tissue Variability and Antennas Networks: In the Era of Big Data and Beyond,” IEEE Reviews in for Power Transfer to Wireless Implantable Medical Devices,” IEEE Biomedical Engineering, vol. 8, pp. 4-16, April 2015. Journal of Translational Engineering in Health and Medicine, vol. 5, pp. [4] K. Gosalia, M. Humayun and G. Lazzi, “Impedance Matching and Aug. 2017. Implementation of Planar Space-Filling Dipoles as Intraocular Implanted [24] A. Kiourti and Κ.S. Nikita, “Numerical Assessment of the Performance Antennas in a Retinal Prosthesis,” IEEE Trans. on Antennas and of a Scalp–Implantable Antenna: Effects of Head Anatomy and Propagation, vol. 53, no. 8, pp. 2365-2373, Aug. 2005. Dielectric Parameters,” Wiley Bioelectromagnetics, vol. 34, issue 3, pp. [5] C.W. L. Lee et al., “A high-sensitivity fully-passive neurosensing 167–179, Apr. 2013. system for wireless brain signal monitoring,” IEEE Trans. Microw. [25] S. O’Driscoll, A. S. Y. Poon and T. H. Meng, “A mm-sized implantable Theory Tech.,vol. 63, no. 6, pp. 2060–2068, Jun. 2015. power receiver with adaptive link compensation,” ISSCC, pp. 293-295, [6] Y. Liu, Y. Chen, H. Lin, and F. Juwono, “A novel differentially fed Feb. 2009. compact dual-band implantable antenna for biotelemetry applications,” [26] C. Lu et al., “A sub-mW antenna-impedance detection using electrical IEEE Antennas and Wireless Propagation Letters, vol. 15, pp. 1791- balance for single-step on-chip tunable matching in 1794, Mar. 2016. wearable/implantable applications,” ISSCC, pp. 298-299, Feb. 2017. [7] H. Bahrami et. al., “Flexible, Polarization-Diverse UWB Antennas for [27] F. Mirzavand, V. Nayyeri, M. Soleimani and R. Mirzavand, “Efficiency Implantable Neural Recording Systems,” IEEE Transactions on improvement of WPT system using inexpensive auto-adaptive Biomedical Circuits and Systems, vol. 10, no. 1, pp. 38-48, Feb. 2016. impedance matching,” Electron. Lett., vol. 52, no. 25, pp. 2055-2057, [8] L.-J. Xu, Z. Duan, Y.-M. Tang, and M. Zhang, “A dual-band on-body Dec. 2016. repeater antenna for body sensor network,” IEEE Antennas Wireless [28] A. J. Skrivervik, “Implantable antennas: The challenge of efficiency,” in Propag. Lett., vol. 15, pp. 1649–1652, 2016. Proc. Eur. Conf. Antennas Propag., Gothenburg, Sweden, Apr. 2013, pp. [9] J. Felicio, J. Costa and C. Fernandes, “Dual-Band Skin-Adhesive 3627–3631. Repeater Antenna for Continuous Body Signals Monitoring,” IEEE [29] D. Karnaushenko, D. Karnaushenko, D. Makarov and O. G. Schmidt, Journal of Electromagnetics, RF and Microwaves in Medicine and “Compact helical antenna for smart implant applications,” NPG Asia Biology, vol. 2, no. 1, pp. 25-32, March 2018. Materials 7, e188, 2015, doi:10.1038/am.2015.53;. [10] E. Chow, Y. Ouyang, B. Beier, W. Chappell and P. Irazoqui. [30] S. Radiom, M. Baghaei-Nejad, G. Vandenbosch, L. Zheng and G. “Evaluation of Cardiovascular Stents as Antennas for Implantable Gielen, “Far-field RF Powering System for RFID and Implantable Wireless Applications,” IEEE Transactions on Microwave Theory and Devices with Monolithically Integrated On-Chip Antenna,” 2010 IEEE Techniques vol. 57, no. 10, pp. 2523-2532, Oct. 2009. Radio Frequency Integrated Circuits Symposium, 23-25 May 2010. [11] A. Kiourti, J. R. Costa, C. A. Fernandes, and K. S. Nikita, “A broadband [31] E. Huber, M. Mirzaee, J. Bjorgaard, M. Hoyackm S. Noghanian and I. implantable and a dual-band on-body repeater antenna: design and Chang “Dielectric property measurement of PLA,” 2016 IEEE transmission performance,” IEEE Trans. Antennas Propag., vol. 62, no. International Conference on Electro Information Technology, Grand 6, pp. 2899–2908, Jun. 2014. Forks, ND, USA, 19-21 May 2016. [12] M. Magill, G. Conway and W. Scanlon, “Tissue-Independent [32] R. C. Rumpf, C. R. Garcia, H. H. Tsang, J. E. Padilla, and M. D. Irwin, Implantable Antenna for In-Body Communications at 2.36–2.5 GHz,” "Electromagnetic isolation of a microstrip by embedding in a spatially IEEE Transactions on Antenans and Propagation, vol. 65, no. 9, Sep. variant anisotropic metamaterial," Progress In Electromagnetics 2017. Research, Vol. 142, 243-260, 2013. [13] A. RamRakhyani and G. Lazzi, “On the design of efficient multi-coil [33] R. C. Rumpf, “Maxwell’s Equations in Fourier Space,” Lecture 17, telemetry system for biomedical implants,” IEEE Transacations on Computational Electromagnetics Webseries, CEM Lectures, YouTube, Biomedical Circuits and Systems, vol. 7, iss. 1, pp. 11-23, Feb. 2013. Sep. 2013. [14] M. Manoufali, K. Bialkowski, B. Mohammed, P. Mills and A. Abbosh, [34] Raise3D, Inc., Costa Mesa, CA, www.raise3d.com, 2018. “Near-Field Inductive-Coupling Link to Power a Three-Dimensional [35] Schmid & Partner Engineering AG, www.speag.com, 2018. Millimeter-Size Antenna for Brain Implantable Medical Devices,” IEEE Transactions on Biomedical Engineering, vol. 65, iss.1, pp. 4-14, Jan. [36] T. Onishi and S. Uebayashi, “Biological Tissue-Equivalent Phantoms 2018. Usable in Broadband Frequency Range,” NTT DoCoMo Technical Journal, vol. 7, no. 4, pp. 61-65, Mar. 2006. [15] T. Karacolak, R. Cooper, E. Unlu and E. Topsakal, “Dielectric properties of porcine skin tissue and in vivo testing of implantable

[37] S. Koulouridis, G. Kiziltas, Y. Zhou, D. J. Hansford, and J. L. Volakis, IEEE Engineering in Medicine and Biology Society (EMB-S) “Polymer-ceramic composites for microwave applications: Fabrication and performance assessment,” IEEE Microw. Theory Techn., vol. 54, Young Investigator Award for 2014, the IEEE Microwave no. 12, pp. 4202–4208, Dec. 2006. Theory and Techniques Society (MTT-S) Graduate [38] A. Demir et al., “Anatomical Region-Specific In Vivo Wireless Fellowship for Medical Applications for 2012, and the IEEE Communication Channel Characterization,” IEEE Journal of Biomedical Antennas and Propagation Society (AP-S) Doctoral Research and Health Informatics, vol. 21, no. 5, pp. 2154-2162, Sept. 2017. Award for 2011. She is currently serving as an Associate [39] W. Lei, H. Chu and Y. Guo, “Design of a circularly polarized ground Editor for the IEEE Transactions on Antennas and radiation antenna for biomedical applications,” IEEE Transactions on Antennas and Propagation, vol. 64, iss. 6 pp. 2535-2540. Apr. 2016. Propagation. [40] J. Kim and Y. Rahmat-Samii, “Implanted antennas inside a human body: simulations, designs, and characterizations,” IEEE Transactions on Microwave Theory and Techniques, vol. 52, iss. 8, pp. 1934 – 1943, Aug. 2004. [41] “Evaluating compliance with FCC guidelines for human exposure to radio frequency electromagnetic fields,” FCC, Washington, DC, USA.

John “Jack” Blauert (M’14) graduated Summa Cum Laude with B.S. degree in electrical engineering from Rose-Hulman Institute of the Technology in Terre Haute, IN, in 2017. He is currently pursuing his PhD in electrical engineering at The Ohio State University ElectroScience Laboratory in Columbus, OH under the supervision of Dr. Asimina Kiourti. From 2015 to 2018, he completed internships with Texas Instruments Kilby Labs, Marathon Petroleum and ArcelorMittal. To date, he has authored one journal paper and three conference papers. His research interests include biomedical applications of electromagnetics research, implanted medical devices, and electromagnetic metamaterials. Mr. Blauert has been a member of IEEE HKN since 2014. He is also serving as a Reviewer for both the IEEE Journal of Electromagnetics, RF and Microwaves in Medicine and Biology and IEEE Transactions in Antennas and Propagation. In 2018, he was awarded the IEEE Antennas and Propagation Society Doctoral Research Grant.

Asimina Kiourti (S’10–M’14) received the Diploma degree in electrical and computer engineering from the University of Patras, Patras, Greece, in 2008, the M.Sc. degree in technologies for broadband communications from University College London, London, U.K., in 2009, and the Ph.D. degree in electrical and computer engineering from the National Technical University of Athens, Athens, Greece, in 2013. Dr. Kiourti is currently an Assistant Professor of Electrical and Computer Engineering at The Ohio State University and the ElectroScience Laboratory, Columbus, OH, USA. From 2013 to 2016 she served as a Post-Doctoral Researcher and then a Senior Research Associate at the ElectroScience Laboratory. During her career, she has (co-)authored over 40 journal papers, 80 conference papers, and 7 book chapters. Her research interests include wearable and implantable sensors, antennas and electromagnetics for body area applications, and flexible textile-based electronics. Dr. Kiourti has received several awards and scholarships, including the URSI Young Scientist Award for 2018, the