1
Design of an Implantable Slot Dipole Conformal
Flexible Antenna for Biomedical Applications
Maria Lucia Scarpello
1
, Divya Kurup
2
, Hendrik Rogier
1
, Senior Member, IEEE,
Dries Vande Ginste
1
, Member, IEEE, Fabrice Axisa
3
, Jan Vanfleteren
3
, Member, IEEE, Wout
Joseph
2
, Member, IEEE, Luc Martens
2
, Member, IEEE, Gunter Vermeeren
2
1 Ghent University, Department of Information Technology (INTEC), Electromagnetics Group, Sint-Pietersnieuwstraat 41,
B-9000 Gent, Belgium.
e-mail: marialucia.scarpello@ugent.be.
2 Ghent University, Department of Information Technology (INTEC), UGent-WiCa, Complex Zuiderpoort Blok C0 bus 201 Gaston
Crommenlaan 8, B-9050 Gent, Belgium.
3 Ghent University, ELINTEC-TFCG, Technologiepark 914 B-9052 Gent, Belgium.
Abstract—We present a flexible folded slot dipole
implantable antenna operating in the Industrial, Sci-
entific, and Medical (ISM) band (2.4-2.4835 GHz) for
biomedical applications. To make the designed antenna
suitable for implantation, it is embedded in biocompat-
ible Polydimethylsiloxane (PDMS). The antenna was
tested by immersing it in a phantom liquid, imitating
the electrical properties of the human muscle tissue.
A study of the sensitivity of the antenna performance
as a function of the dielectric parameters of the en-
vironment in which it is immersed was performed.
Simulations and measurements in planar and bent state
demonstrate that the antenna covers the complete ISM
band. In addition, Specific Absorption Rate (SAR)
measurements indicate that the antenna meets the
required safety regulations.
Index Terms—Implantable antennas, Industrial, Sci-
entific and Medical (ISM) band, Specific Absorption
Rate (SAR), bent antenna, muscle tissue sensitivity.
I. Introduction
Implantable devices are becoming widely researched
for different fields of applications, both for humans and
animals. Some examples of applications are: monitoring
blood pressure and temperature, tracking dependent people
or lost pets, wirelessly transferring diagnostic information
from an electronic device implanted in the human body
for human care and safety, such as a pacemaker, to an
external RF receiver [1]. Small implantable biomedical
devices placed inside the human body may improve the
lives of numerous patients. Patients with the antenna
implanted in the body regularly return to the hospital for
checkups, where their status and the status of the implant
are verified. With the use of RF technology, data recorded
by the implanted antenna can be transmitted wirelessly to
the receiving station, while the patient is waiting in the
lounge. Some patients may require checks every day. In such
case a home care unit can be placed in the patient’s home.
The unit can communicate with the medical implant and
can be connected to the telephone system, or the internet,
and send data regularly to the responsible person at the
hospital [2].
The state-of-the-art of research in implantable antennas
shows that microstrip or planar-inverted F antennas
(PIFA), operating in the 402-405 MHz Medical Implant
Communications Service (MICS) band, were simulated [3]
and also fabricated and measured [4]–[6]. The main issue
with this kind of antennas, operating at 403 MHz and
thus corresponding to a wavelength of 744.4 mm in free
space, is that it is not practical to put them into a living
human body without performing thorough miniaturization.
Indeed, taking the effect of the body into account during
the design, reducing the antenna size by around seven
times, without miniaturization it still remains too big to
be implanted. In [5], the designed antenna has a cylinder
height of 22.72 mm and an external radius of 10.5 mm,
and so it is a small size antenna. However, the reflection
coefficient value shown in the paper is only simulated
and not measured and the simulated MICS bandwidth
is only partially covered. In [6], the designed PIFA antenna
is also small, i.e. 22.5 mm
×
18.5 mm
×
1.9 mm, but
its fractional bandwidth is 20% lower than the one of
the antenna we propose and it is not embedded in any
insulating biocompatible material during measurements.
Choosing a higher resonance frequency, corresponding to
the Industrial, Scientific, and Medical (ISM) band (2.4-
2.4835 GHz), is one way to reduce the antenna size and
making it available to be implanted. Another advantage
has to do with the radio communication link: the larger
bandwidth allows for higher bitrates. Implantable H-shaped
slot cavity antennas are studied for 2.45 GHz applications in
[7], [8]. In [7], the antenna was simulated, whereas in [8] the
same antenna was reduced in size. But since the size was too
small to be fabricated (2.8 mm
×
4.0 mm
×
1.6 mm), the an-
tenna was rescaled to larger dimensions to be manufactured
and measured in order to be able to compare measurements
with the finite-difference time-domain (FDTD) simulations.
Radiation patterns, gain pattern, and radiation efficiency
value are related to the rescaled antenna and to a rescaled
resonance frequency equal to 980 MHz. Moreover in [8], no
biocompatible material was used to embed the antenna and
the cable during the measurements. In [9], [10], two dual
band implantable antennas are presented, working properly
in the MICS band and in the ISM band. Measurements
2
were performed in a human skin mimicking gel tissue and
in a rat skin mimicking gel tissue, respectively. In [10],
the gain pattern is simulated and its maximum value is
−
10 dBi in the ISM band. However, the two antennas
are not embedded in any biocompatible material and they
are not flexible. In [11] a cardiovascular stent, working at
2.4 GHz has been designed, fabricated, and implanted
in a live porcine subject. The stents are left without
any insulation material and they are in direct contact
with the tissue. In [12] three different inhomogeneous
digital phantoms are considered to check the different
radiation performances of wireless implants. Here again
only simulations are performed but no measurements.
In this paper, we present the design, characterization,
and measurements of an implantable antenna operating in
the 2.45 GHz ISM band, recommended by the European
Radiocommunications Committee (ERC) for ultra-low-
power active medical implants [13]. The antenna is a flexible
slot dipole. To make the antenna suitable for implantation,
it is embedded in biocompatible Polydimethylsiloxane
(PDMS). The reflection coefficient was simulated and
measured in the MSL2450 liquid, provided by Speag
(Zurich, Switzerland) [14], mimicking a 100% human muscle
tissue, having well-defined dielectric values at 2.45 GHz. To
investigate how different human tissues, surrounding the
implanted slot dipole, affect its radiation characteristics, a
study on the sensitivity of the liquid mimicking the human
muscle tissue was performed. Thereto its dielectric nominal
values were varied from 50% larger to 50% smaller. These
simulations were performed to ensure that the antenna is
functioning properly in any type of body environment. The
radiation characteristics of the antenna in terms of E-field
and gain were simulated by means of FDTD calculations.
For the evaluation of performances and safety issues related
to implanted antennas, the 10-g and 1-g averaged specific
absorption rate (SAR) are measured and compared with
the ICNIRP [15] and with the FCC guidelines [16].
First, in Section II, the antenna design and its fabrication
are presented. Next, in Section III, the performance in
terms of reflection coefficient is reported. Good agreement
is demonstrated between the simulated and measured
reflection coefficient. In Section IV, the sensitivity of the
antenna as a function of the dielectric properties of the
muscle tissue is analyzed, verifying that the antenna can
be placed close to different kinds of tissue. In Section V,
the radiation characteristics of the antenna, including the
measured SAR distribution, are shown. In Section VI
conclusions are summarized.
II. Biocompatible Folded Slot Dipole Antenna
Design and Manufacturing Process
The antenna, presented in this paper, is a flexible
folded slot dipole embedded in biocompatible PDMS, as
folded slot dipole geometries can provide significantly
larger bandwidths than patch antennas [17]. The top and
frontal view of the antenna are shown in Figs. 1 and 2,
respectively. The dimensions of the folded slot dipole
antenna are shown in Table I. The antenna is designed
h
l
H
y
x
z
L
Gap
G
G
S
Wg’
Wg’’
Wg’’’
Ws
Figure 1: Top view of the coplanar waveguide-fed antenna.
by means of the 2.5-D EM field simulator Momentum of
Agilent’s Advanced Design System (ADS). The antenna
design procedure consists of three steps. First, the folded
slot dipole antenna was designed, using ADS’s optimization
routines, to operate in the 2.45 GHz ISM band in free
space. Second, one superstrate and one substrate of PDMS
were added to the design and, after the characterization
of the PDMS, we redesigned and reoptimized the antenna
embedded in silicone so that it covers the ISM band. Third,
in a last optimization step, on top of the superstrate
and below the substrate one layer of liquid, mimicking
the dielectric characteristics of human muscle tissue at
2.45 GHz, was added. Finally, the antenna so designed
have good simulations results, working properly in the
ISM band. To check more accurately, the antenna was
also simulated with the 3-D simulator CST Microwave
Studio and the simulation results were still satisfactory. To
manufacture the antenna we rely on a flexible electronic
technology. A photoresist film was spin-coated on a copper
foil and patterned by UV radiation through a photomask;
the patterned shape is shown in Fig. 1; two PDMS layers
are used as substrate and superstrate, each with a thickness
of 2.5 mm, to mould the antenna [18], [19]. The dielectric
properties of the PDMS were characterized at 2.45 GHz,
to be
r
= 2.2 and tan δ = 0.013.
The feeding structure of the slot dipole antenna consists
3
Table I: Size of the folded slot dipole antenna
[mm]
H 8.5
L 25.9
h 3.2
l 8.3
Wg’ 1.2
Wg” 1.5
Wg”’ 1.0
Ws 0.3
G 1.8
S 1.7
Gap G-S 0.1
Copper 9 µm
Nickel 50 nm + Gold 200 nm
Polyimide 25 µm
PDMS 2.5 mm
Figure 2: Frontal view of the coplanar waveguide-fed
antenna.
of a coplanar waveguide (CPW) with a 50 Ω mode
impedance. Matching the mode impedance of the CPW to
50 Ω is obtained by tuning the distance between the tracks
G and S, as well as the width of the tracks (Fig. 1). The
CPW is fed by a U.FL connector and an ultra-fine Teflon
coaxial cable supplied by Hirose [20]. The U.FL connector
is chosen (specifically for measurements purposes) because
it is compact and it suits the small CPW size. Both the
connector and the cable are also embedded in PDMS. Fig. 3
shows the antenna prototype before being embedded in the
PDMS. Fig. 4 shows a side view of the antenna prototype
with its connector and cable after being embedded in the
PDMS.
Figure 3: Top view of flex antenna without PDMS.
Figure 4: Side view of antenna and cable embedded in
PDMS.
III. Simulation and Measurements
In real-life applications, the antenna is intended to be im-
planted into the human body, subcutaneously, particularly
inside the muscle. Hence, the measurement setup, using a
phantom, is as follows. The antenna is placed at the center
of a plastic container of dimensions 80 cm
×
50 cm
×
20 cm
filled with 30 liters of the Human Muscle Tissue liquid
MSL2450 [14]. This liquid mimics the dielectric character-
istics of human muscle tissue at 2.45 GHz, standardized
in [21] to be
r
= 52
.
7,
σ
= 1
.
73 S/m,
ρ
= 1000 kg/m
3
.
Dielectric values of the liquid at 2.45 GHz measured by
the manufacturer result to be:
r
= 50
.
8,
σ
= 2
.
01 S/m,
ρ
= 1
.
030 kg/m
3
[14]. In Tables II and III permittivity
and conductivity values of the liquid MSL2450 at different
frequencies, measured by the manufacturer, and of the
human muscle tissue [21], [22], [23] are listed, respectively.
Table II: Dielectric values of MSL2450, at different frequen-
cies.
Frequency[GHz]
r
σ S/m
2.0 52.15 1.47
2.4 52.06 1.94
2.45 50.8 2.01
2.5 50.51 2.02
3.0 48.10 2.77
4
Table III: Dielectric values of human muscle tissue, at
different frequencies, as reported in [21], [22], [23].
Frequency[GHz]
r
σ S/m
2.0 53.29 1.45
2.4 52.79 1.70
2.45 52.72 1.73
2.5 52.66 1.77
3.0 52.05 2.14
The values in Tables II and III cover the band (2.0 GHz
and 3.0 GHz) in which the measurements were performed,
and as such also encompass the ISM band (2.4-2.485 GHz).
Fig. 5 reports the
S
11
values for each couple of dielectric
values reported in Tables II and III, valid at the specified
frequency. It can be observed that the small differences
in
r
,
σ
, reported in Tables II and III, do not lead
to very different antenna behavior. A third simulation,
where a fixed relative permittivity
r
= 50
.
8 and a fixed
conductivity
σ
= 2
.
01 S/m were used within the complete
band, also indicates that the antenna is rather insensitive
to changes of the surrounding medium. This will be further
illustrated by considering other human tissues (Figs. 9 and
10). Moreover, a detailed study of the sensitivity of the
antenna as a function of the dielectric properties of the
liquid is reported in Section IV.
2 2.2 2.4 2.6 2.8 3
−25
−20
−15
−10
−5
0
Frequency [GHz]
|S11| [dB]
ISM Band
|S11|, [21]
|S11|, MSL2450
|S11|, MSL2450
with ε
r
=50.8, σ=2.01 S/m
Figure 5: S11 values using the MSL2450 dielectric proper-
ties [14] (dotted line) and to human muscle tissue dielectric
properties [21] (dashed line), compared to the return
loss value of the antenna immersed in a medium with
fixed values of permittivity and conductivity, i.e. those of
MSL2450 at 2.45 GHz (full line).
The antenna is connected to a Rhode and Schwarz
ZVR Network Analyzer, as shown in Figs. 6 and 7. Inside
the phantom,
S
11
measurements are performed when the
antenna is both in planar and bent state.
Figure 7: SAR and reflection measurement setup with the
implantable antenna inside the liquid during a measure-
ment.
Network Analyzer
Muscle Tissue Liquid
30 liters
Antenna
80 cm
50 cm
20 cm
Figure 6: Schematic representation of the reflection mea-
surement setup for the designed implantable antennas using
muscle tissue simulating liquid.
First, Fig. 8 displays a comparison between the measured
and simulated reflection coefficient of the antenna in planar
state. The simulations are performed using the EM field
simulator Momentum of Agilent’s Advanced Design System
5
(ADS) and CST Microwave Studio simulator. Since ADS
Momentum is a 2.5-D simulator, it does not account for
the finite size of the PDMS layers. This finite size typically
results in a shift of the resonance frequency to lower
frequencies, so for the initial design in Momentum, to
cover the ISM band, the antenna is designed to resonate
at 2.5 GHz. Once fabricated and measured, this design
ensures that our antenna will cover the complete requested
bandwidth, as also verified by CST simulations in Fig. 8.
The required
-10
dB impedance bandwidth of the antenna
is 83.5 MHz in the 2.45 GHz ISM band. Simulations and
measurements satisfy the requirements: (i) the antenna
was simulated in planar state with ADS. Its bandwidth is
very wide, 1.23 GHz (2.22-3.45 GHz), it includes different
resonances of the antenna and the fractional bandwidth at
the target frequency (2.45 GHz) is approximately 50.2%.
(ii) The antenna was also simulated in planar state with
CST: its bandwidth is 270 MHz (2.30-2.57 GHz) and the
fractional bandwidth at the target frequency (2.45 GHz)
is approximately 11.0%. (iii) Antenna measurements are
performed to validate the simulations: the measured band-
width in planar state is 350 MHz (2.20-2.55 GHz) and the
fractional bandwidth at the target frequency (2.45 GHz)
is approximately 14.2%.
2 2.2 2.4 2.6 2.8 3
−30
−25
−20
−15
−10
−5
0
Frequency [GHz]
|S11| [dB]
ISM Band
ADS Simulation
Measurement
CST Simulation
Figure 8: Antenna in planar state: reflection coefficient,
simulations vs measurement.
Second, to demonstrate that the antenna can work within
different human tissues, the antenna is placed into two
human body structures, as reported in [24], [25] and
shown in Figs. 9 and 10, and simulated by means of ADS
Momentum. The dielectric properties (
r
,
σ
), for each layer
of the two structures, at
f
= 2
.
45 GHz, were obtained
from [21], [22] and [23], and shown in Table IV.
Table IV: Dielectric values of four different human tissues,
at 2.45 GHz, as reported in [21], [22], [23].
r
σ S/m
Muscle 52.72 1.73
Skin 38.00 1.46
Fat 5.28 0.10
Bone 18.54 0.80
Figure 9: Three layers geometry for the design of the slot
dipole antenna, placed in between fat and muscle.
Figure 10: Five layers geometry for the design of the slot
dipole antenna, placed in between two layers of muscle.
In Fig. 11, the return loss value of the antenna placed
in the structures of Figs. 9 and 10 is shown. A slight shift
of the resonance frequency towards higher frequencies can
be observed, but the return loss value still remains below
- 10 dB in the whole ISM band.
Third, the performance of the antenna when it is bent,
as such making it conformal to curved parts of the body,
is verified. Thereto, the antenna is bent around its x-axis
2 2.2 2.4 2.6 2.8 3
−25
−20
−15
−10
−5
0
Frequency [GHz]
|S11| [dB]
ISM Band
|S11|, MSL2450
|S11|, [23]
|S11|, [24]
Figure 11: Simulation of the return loss of the antenna
placed in the structures of Figs. 9 (dashed line) and 10
(dotted line), compared to the antenna placed in the
MSL2450 (full line).
![Table V: Resonance frequency [GHz] and reflection coefficient [dB] for some couples of r and σ values.](/figures/table-v-resonance-frequency-ghz-and-reflection-coefficient-8lyf7t5t.webp)
