Journal of Medical Engineering & Technology, Vol. 35, No.

1, January 2011, 4753

Innovation
A multi-bundle concentric coil wirelessly transferring power
to in vivo implantable devices
H. M. AMASHA*{, J. I. AL-NABULSI{, O. M. ALOQUILIx and B. O. AL-NAAMI{
{Biomedical Engineering, Damascus University, FMEE, Airport Road, Damascus, Syria, Jordan,
Damascus, Syrian Arab Republic
{Biomedical Engineering, Hashemite University, PO Box 330127, Zarqa, 13511 Jordan
xElectrical Engineering, Hashemite University, PO Box 150459, Zarqa, 13115 Jordan
(Received 18 July 2010; revised 10 August 2010; accepted 17 September 2010)

Biomedical devices implanted inside the human body have a heavy demand on battery
power. The internal batteries are charged wirelessly through two coils. The primary is
placed outside the chest and is fed with an electromagnetic eld, while the implanted
secondary delivers current to the batteries. Increasing the number of turns in the internal
secondary induces an increased amount of localized heat. A new approach proposed by
the authors involves implanting a specically designed multi-bundle concentric coil inside
the body. It is shown that this newly proposed coil produces less localized heat. The total
number of turns in the proposed coil is the same as that in the single-bundle coil except
that it is divided into four equal bundles. Each bundle has a dierent diameter and is
spatially concentric. Since the turns are divided into thinner bundles, they are easier to
isolate with a biocompatible material and oer much better heat dissipation and fewer
hotspots. Electromagnetic simulation using nite element analysis proved that the
performance of the proposed coil is no lower than the single-bundle ordinary coil.
Thermal simulation showed the improvement of temperature distribution using the multibundle coil, compared to the single-bundle coil.
Keywords: Implanted devices; Wireless power; Battery charging; Multi-bundle coil

1. Introduction
Currently, there are several schemes used to wirelessly
transfer energy to implanted devices. In one method, the
implanted target comprised of a piezoelectric element
bounded to a metallic material. Induced eddy currents
vibrate the metal and the resulting stress generates
charges in the piezoelectric elements [1, 2]. An improved
method utilizes a piezoelectric material which also
incorporates magneto-resistive layers [3].
A dierent class of methods that rely on magnetic
induction modes uses induction coils in vivo [4, 5]. Some are
restricted to very close range or very low power [69]. Some

of these methods employ low frequencies, e.g. 50 Hz, while

others use radio frequency (RF) and microwave signals
[1014]. The output voltage harvested on the secondary coil
depends on the frequency, distance, coil structure, and the
media separating these coils. For instance, in biomedical
application, the choice of a suitable frequency is crucial and
depends on safety and many physiological and anatomical
factors. Figure 1 shows the positioning of the two coils in
space and the distance between them. Recent works and
research has shown that this power transfer is possible even
at long distances [15].
The proposed method is based on the well-known
principle of coupling. The transfer function of current

*Corresponding author. Email: haniamasha@gmail.com

Journal of Medical Engineering & Technology
ISSN 0309-1902 print/ISSN 1464-522X online 2011 Informa UK, Ltd.
http://www.informahealthcare.com/journals
DOI: 10.3109/03091902.2010.525685

48

H. M. Amasha et al.

Figure 1. A standard arrangement of the coils in space, with

the eld lines and the output current from one turn of the
coil only, for illustration purposes.
from the rst coil to the second is governed by the mutual
induction between the two coils. An appropriate analytical
framework for modelling this resonant energy exchange is
coupled-mode theory (CMT) [15]. In Figure 1, the eld of
the system of two resonant objects 1 and 2 is approximated
by equations (1) and (2) as follows:
Fr; t  a1 tF1 r a2 tF2 r;

where F1(r) and F2(r) are the eigenmodes of 1 and 2 alone,

and the eld amplitudes a1(t) and a2(t) can be shown to
satisfy, to the lowest order as shown in equation (2):
da1 io  i
1
dt
da2 io  i
2
dt

1 a1

i ka2 ;

2 a2

i ka1 ;

where both o1 and o2 are the individual eigenfrequencies,

1 and
2 are the resonance widths due to the objects
intrinsic (absorption, radiation, etc.) losses, and k is the
coupling coecient [15]. Equations (2) show that at exact
resonance (o1 o2 and 1 2), the normal modes of the
combined system are split by 2k; the energy exchange
between the two objects takes place in time *p/2k with
some losses, which are minimal when the coupling rate is
much faster than
all loss rates (k 44 1,2) [15]. It is exactly
p
this ratio k= 1 2 that has been set as the gure-of-merit
for any system under consideration for wireless energytransfer, along with the distance over which this ratio can
be achieved [15].
In our new proposed approach, the two coils are at a
distance of 1 cm, which indicates a near electromagnetic
eld case. The non-uniformity of barrier between the two
coils (various thicknesses of fat, skin and muscle) makes it
more dicult to predict the transfer function methodically.
Furthermore, the decreasing radius of each of the bundles
contributed to our decision to adopt an approximate power
transfer coecient, which was empirically determined by

averaging many practical measurements to achieve the

desirable one.
In all the methods mentioned above, heat dissipation in
the vicinity of the receiver is of high concern, since this
can lead to tissue damage, which progressively limits heat
clearance, causing the problem to escalate [16, 17]. Even
the integrated battery charging circuit was reported to
dissipate power in the charging phase [18]. Furthermore, if
the proposed coil is implanted in vivo, it must not be
bulky in size and must be easy to insulate with a biocompatible material. It must also be exible and easy to
x to the chest wall. The power transfer performance of
the new proposed coil must not be less than the
performance of existing coils. The implanted devices
require extended charging time for the internal battery.
The internal coil produces localized heat in its vicinity
which will harm the surrounding tissue if not dissipated
eectively.
An external coil (primary) is placed on the surface of the
skin, while the secondary coil is placed under the skin
opposing the primary coil. It is suggested that the internal
(secondary) coil be implanted at a site as far as possible
from the heart to eliminate interference, and as close to the
skin surface as possible to decrease distance between the
two coils. This will enhance electromagnetic coupling and
lead to a better power transfer ratio. The primary is fed
with an electromagnetic sinusoidal waveform with a
frequency higher than the frequency of bio-signals, and
with an amplitude not exceeding the permitted safety levels
of currents induced in the human body [1921]. The
internal coil could also be used as an antenna to report on
the status of the implanted device.
2. Method and materials
The multi-bundle coil which is intended to be implanted
in vivo as a secondary coil is constructed using a thin single
wire. This is wound 50 times to form the rst bundle and
then wound another 50 times to form the second bundle,
but with a radius 1 cm larger, and the third and fourth
bundles are formed in a concentric manner. Each bundle
has a diameter 2 cm wider than the previous bundle. Figure
2(a) shows a schematic of the multi-bundle coil while gure
2(b) illustrates a practical coil implemented for experimental verication.
3. Simulation models
Here the proposed multi-bundle coil is compared with a
single bundle coil by two simulations: electromagnetic and
thermal. The proposed multi-bundle coil is expected to be
superior at least in the thermal criterion, and to be equal
in electromagnetic performance. A nite element model
using ANSYS (ANSYS Inc., Canonsborough, PA, USA)
is constructed to simulate the external coil and a skin/fat

Wireless transfer of power to in vivo implantable devices

49

Figure 3. A lateral central cross section through the two

opposing coils and magnetic eld lines.

Figure 2. (a) A 2D model of the multi-loop coil conguration. (b) An experimental coil with four bundles.

layer of 0.5 cm thickness that separates it from the

internally implanted single-bundle or multi-bundle coils.
3.1. Electromagnetic simulation
A sinusoidal waveform is applied to the primary coil and
the potential is measured at the output of the secondary.
Figure 3 shows the simulation of the eld lines distribution
with a lateral cut through the two coils. The single-bundle
primary coil on the left and the secondary multi-bundle on
the right are spaced 0.5 cm apart. Two sets of measurements at the secondary coil are compared when the same
input was applied to the primary coil. The rst set is
measured when there is a single-bundle coil implanted as

the secondary, and the second set is measured when a

multi-bundle coil is implanted as the secondary. Results
show that there is no signicant dierence in the output of
the single-bundle and multi-bundle secondary coils. Primary signals are applied with various wave-shapes (sinusoidal, triangular, and rectangular) and a range of
frequencies to decide on the frequency and wave-shape
that transfer power eectively from the primary to the
secondary. The frequencies range from 10 kHz up to 100
kHz with steps of 10 kHz. All waveform shapes are tested
for the whole of the frequency range applied in the
experiments.
Three types of separators with varying thicknesses are
placed between the two coils; skin, fat and both. First a
2 mm layer of skin and a 4 mm layer of fat are used
separately; later the two layers are combined. Normally,
these layers are not uniform in thickness and as time
progresses after implantation a layer of fat tends to
accumulate between the bundles. The location of implantation might introduce further anomalies in the thickness of
these layers.
3.2. Thermal simulation
In order to evaluate the thermal performance of the two
coil congurations, nite element analysis (FEA) is used.
The ANSYS1 package is used for the analysis; twodimensional (2D) models are made for the two coil
congurations, one of which is shown in gure 2(a), and
another similar model is made for the single loop
conguration.
The rings illustrated in gure 2 are the copper loops while
the remaining areas adjacent to the coils are skin. The
elements used are the 2D six-node triangle (PLANE35) for
the meshing of the two models. Values of 400 W m72 K
and 0.21 W m72 K are used respectively for copper and
skin thermal conductivities [22, 23]. The skin and copper

50

H. M. Amasha et al.

regions of the model are then glued in ANSYS1 to ensure

node connectivity at the interfaces. A temperature of 310 K
is assigned to the outer and inner skin areas of the model,
and a heat ow of 5 W is assigned to the nodes within the
copper loops.
The issue of implant thermo-compatibility arises in at
least two contexts: rst, the active production of waste heat
by individual systems implanted within the human body;
and, second, the physiological eects of medical implants
that may result from the passive thermo-physical characteristics of those implants, or from the materials with
which they are constructed [24, 25, 26].
Previous discussions of thermally active systems involve
waste heat conduction, the local and global in vivo
thermogenic limits of implanted systems, thermographic
navigation, and the thermal safety of in vivo electrical and
mechanical systems. For instance, excessive waste heat
generation (e.g. creating localized temperatures 4 428C)
can stimulate thermo-sensitive channels in keratinocytes
and in a specialized group of heat-sensing sensory neurons
terminating in the skin. Aside from blackbody radiation,
sweating, capillary sphincter control, and behavioural
thermoregulation (including respiratory cooling), the body
regulates its temperature and ooads excess heat principally via two mechanisms, as follows [27, 28].
First, in passive conduction heat travels by pure
conduction through fat and muscle from the body core
out to the periphery. The average thermal conductivity of
human tissue is Kt * 0.5 W m71 K, so for a typical
L 10 cm path length (*half-torso thickness), heat ow
Hf * (Kt/L) 5 W m72 K], or *10 W K71 for a 2 m2
human body. In a cold room, the mean temperature
dierential between core and periphery DT is almost 11 K,
so Hf is approximately 100 Watts, which is approximately
the basal metabolic rate. Experiments conrm that 59 W
m72 K is the minimum heat ow in very cold conditions
(the actual value depending largely upon the thickness of
subcutaneous fat layers). In this case, the peripheral
capillary blood ow has slowed to a trickle, producing
the minimum thermal conductivity of the human body in
cold conditions. On the other hand, in a warm room or
during heavy exercise, DT is only 1 K, so Hf is only about
10 Watts. Thus, paradoxically, at warmer temperatures
when the human body is generating considerable surplus
heat, the bodys passive heat ow is actually very low
because of the smaller temperature dierential between
core and periphery [28].
Second, heat is transported via active blood ow. In
warm rooms, the peripheral capillary sphincters are fully
dilated, allowing more blood to ow through the peripheral
capillaries relative to the core capillaries, and also the total
volume of blood ow may increase. During heavy exercise,
total blood ow volume may rise by a factor of 4 or 5.
Diathermy experiments suggest that the active blood ow
mechanism alone may carry o 100200 Watts of heat

before core temperature starts to rise. In cold rooms and in

the absence of heavy exercise, peripheral capillary sphincters are maximally contracted; thus minimizing blood ow
(and hence heat transport) to the periphery [28].
The passive conduction mechanism can throw o about
100 Watts of waste heat when the human body is in a cold
room but only 10 Watts when the body is in a warm room.
The active conduction mechanism can throw o negligible
heat in a cold room but up to 100200 Watts in a warm
room. Thus, as the external environment warms up, the
human body shifts from passive conduction to active
conduction via increased blood ow and capillary sphincter
widening [28].
Here, we are investigating the dissipation of heat
generated not very deep in the torso. Heat will circulate
rapidly in the single bundle, emanate from the bundle and
propagate through the surrounding tissue, remaining in the
vicinity of the bundle until removed by blood ow and
conduction. However, on the other hand, the multi-bundle
coil generates heat in a more distributed manner.
Since we wish to investigate the thermal distribution in
the coil plane before dissipation to dierent planes, the
situation can be simplied into a basic conduction problem
in the coil plane, which is governed by the following
dierential equation [29]:
qk K

dT
;
dr

where qk is the conductive heat transfer per unit area

[W m72], K is the thermal conductivity, and dT/dr is the
radial gradient of heat transfer [K m71]. The temperature
distributions for the two models are shown in gures 4 and 5.
The results showed a maximum increase in skin temperature of 0.358C for the multi-loop conguration and a
0.58C for the single-loop conguration. This highlights the

Figure 4. Temperature distribution for the multi-bundle

coil conguration.

Wireless transfer of power to in vivo implantable devices

advantage of using a multi-loop coil compared to a single

loop, as in biological systems a dierence of 0.158C rise in
temperature is signicant since this might go on for a long
time (the whole charging time period), with the heat energy
accumulating in that area, causing a hot spot that can cause
damage to cells.
Figures 6 and 7 show the thermal gradient for the two
coil congurations. It can be concluded that the multi-loop
system yielded a smaller thermal gradient, which has the
advantage of reducing heat ow from the coil into adjacent
skin. In addition the multi-loop conguration showed a
more uniform heat ow and reduced thermal spot regions.

51

The results obtained from the thermal analysis of the

coils can be partly explained by the fact that the magnetic
ux lines penetrating the multi-loop conguration are more
distributed compared to the single loop, as shown in gure
2. This would result in a lower temperature rise in the
multi-loop system, in addition to a lower thermal gradient
due to heat distribution over a larger area, which is likely to
be dissipated more quickly, whether by conduction or
blood ow.
4. Experimental results
The results obtained on the multi-bundle coil were better
than results from a single-bundle coil whose number of
turns is equal to the total number of turns of the multibundle, and whose diameter equals the diameter of the
outermost bundle.
4.1. Practical electromagnetic measurement

Figure 5. Temperature distribution for the single-bundle

coil conguration.

Figure 6. Thermal gradient for the multi-bundle coil

conguration.

The results obtained using electromagnetic rectangular or

triangular waveforms applied to the primary did not always
produce reliable values on the secondary, and the output
was not of a sucient magnitude. A sinusoidal waveform
produced enough current on the output to charge the
battery without the need to increase the applied input to
harmful levels. A sinusoidal constant voltage source was
applied to the input and maintained by varying the
frequency from 10 kHz to 100 kHz. A layer of 2 mm skin
is used in this procedure to separate the two coils and
measurements were collected 50 times under similar
conditions and environments; values were averaged and
errors calculated and analysed.

Figure 7. Thermal gradient for the single-bundle coil

conguration.

52

H. M. Amasha et al.

The results of these experiments are summarized in table

1. Errors are calculated and found to be similar for groups
of frequencies, as shown in the table.
This procedure was repeated with a 0.4 cm fat layer
placed between the two coils. The results of these
experiments are summarized in table 2. Errors are
calculated and given in the table for every range of
frequencies.
This procedure is then repeated with a 1.0 cm skin/fat/
muscle layer between the two coils. The results of these
experiments are summarized in table 3.
Careful scanning of results collected at dierent frequencies indicates that we can choose a frequency around 70
kHz without losing much output voltage. In fact, in a few
measurements, the results did not vary more than 5%
across the whole range of the study (10100 kHz). This
would keep us well within the preferred zone of frequencies
applied to the human body. The output diered slightly
when measured with skin alone, fat alone and when both
were combined together, keeping the same thickness.

bundle and the fth thermistor in the centre of the coils.

Temperature was measured with a multi-bundle coil as
shown in gure 8 and with the single bundle coil as gure 9
illustrates.
The measured temperature increment values for this
newly proposed multi-bundle coil were 65% of the
measured increment temperature values near the single
Table 3. Voltages measured in the inner coil for 10100 kHz
when 10 mm of skin/fat/muscle separates two coils.
Error

f (Hz)

Vout (mV)

2%

10
20

800
880

1%

30
40
50
60

870
1020
1010
1070

1%

70
80
90
100

1070
1090
1090
1070

4.2. Practical temperature measurement

The temperature distribution was measured using ve
thermistors; three in the middle between each two bundles;
the fourth thermistor was placed outside the external
Table 1. Voltages measured in the inner coil for 10100 kHz,
when 2 mm of fat separates two coils.
Error

f (Hz)

Vout (mV)

3%

10
20

800
1000

12%

30
40
50
60

1000
1120
1210
1215

51%

70
80
90
100

1210
1205
1200
1210

Figure 8. Measurement of temperature with a multi-bundle

coil.

Table 2. Voltages measured in the inner coil for 10100 kHz

when 5 mm of fat separates two coils.
Error

f (Hz)

Vout (mV)

2.5%

10
20

900
980

2%

30
40
50
60

970
1100
1170
1180

70
80
90
100

1170
1190
1180
1150

12%

Figure 9. Measurement of temperature with a single bundle

coil.

Wireless transfer of power to in vivo implantable devices

bundle. The longer the charging period the more

the likelihood of this dierence increasing, i.e. heat
clearance near the single bundle would accumulate
more. In vivo results may vary due to clearance by blood
ow and are expected to work in favour of the multi-bundle
coil.
5. Discussion and conclusions
The proposed implanted multi-bundle coil has advantages
over a single-bundle coil and has shown comparable
electromagnetic results. Having multiple bundles and
spacing between them is preferable for several reasons:
rst, they allow for better isolation of each bundle with the
biocompatible material, and second, these spaces allow for
tissue to regenerate between the bundles and, hence, give
better accommodation and xation of the coil to the chest
wall. Most importantly, there is better heat dissipation,
reducing the chances of producing hot spots close to the
internal coil. Computer simulation has conrmed the
practical electromagnetic temperature measurements and
assumptions.
Declaration of interest: The authors report no conicts of
interest. The authors alone are responsible for the content
and writing of the paper.
References
[1] OHandley, R.C., Huang, J.K., Bono, D.C. and Simon, J., 2008,
Improved wireless, transcutaneous power transmission for in vivo
application. IEEE Sensors Journal, 8, 5762.
[2] Shroeppel, E.A. and Spher, P.R., 1998, Transcutaneous energy
coupling using piezoelectric device. US Patent 5 749 909.
[3] Huang, J.K., Bono, D. and OHandley, R.C., 2007, Novel sensors
based on magneto-resistive/piezoelectric laminations. Sensors Letters,
5, 15.
[4] Mann, C., 1999, Transcutaneous transmission patch. US Patent 5 948
006.
[5] Hahn, T.W., 2001, Method and apparatus for power link detection
with implantable medical devices. US Patent 6 321 118.
[6] Hirai, J., Kim, T.W. and Kawamura, A., 2000, IEEE Transactions on
Power Electronics, 15, 21.
[7] Fernandez, J.M. and Borras, J.A., 2001, Contactless battery charger
with wireless control link. US patent number 6 184 651.
[8] Scheible, G., Smailus, B., Klaus, M., Garrels, K. and Heinemann, L.,
2003, System for wirelessly supplying a large number of actuators of a
machine with electrical power. US patent number 6 597 076.

53

[9] Ka-Lai, L., Hay, J.W. and Beart, P.G.W., 2006, Contact-less power
transfer. US patent number 7 042 196.
[10] Sample, A.P., Yeager, D.J., Powledge, P.S., Mamishev, A.V. and
Smith, J.R., 2008, Design of an RFID-based battery-free programmable sensing platform. IEEE Transactions on Instrumentation and
Measurement, 57, 26082615.
[11] deMuro, D.M. and Meadows, V., 1997, Wireless battery charging
system. US patent number 5 596 567.
[12] Krishnan, R., 2006, Method and apparatus for wireless powering and
recharging. US patent number 6 127 799.
[13] Calhoon, J.C., Keely, L.B., Michell, W., 2008, Inductively charged
battery pack. US patent number 7 375 492.
[14] Haller, M.I., 2003, Voltage converter for implantable micro-simulator
using RF-powering coil. US Patent 6 631 296.
[15] Karalis, A., Joannopoulos, J.D. and Soljacic, M., 2008, Ecient
wireless non-radiative mid-range energy transfer. Annals of Physics,
323, 3448.
[16] Amasha, H.M., 1997, The feasibility of sucient wireless power
transfer using an implanted single-bundle coil. Biomedical Engineering
Library, FMEE, Damascus University Press, Damascus, Syria.
[17] Amasha, H.M., Ghazawi, Z. and Al-Nabulsi, J., A multi-bundle concentric coil for charging the battery of total articial heart or a
pacemaker. 34th IEEE Computers in Cardiology International Conference,
CINC2007, North Carolina, USA, 30 September3 October 2007.
[18] Pengfei, L., Bashirullah, R., 2007, A wireless power interface for
rechargeable battery operated medical implants. IEEE Transactions on
Circuits and Systems, 54, 912916.
[19] Webster, J.G., 2009, Medical Instrumentation and Design, 4th edn
(John Wiley, NJ, USA).
[20] Pething, R., 1979, Dielectric and Electronic Properties of Biological
Material (John Wiley, New York, USA).
[21] Brown, B.H., Smallwood, R. and Barber, D., 1999, Medical Physics
and Biomedical Engineering (CRC Press, UK).
[22] Risegari, L., Barucci, M., Olivieri, E., Pasca, E. and Ventura, G., 2004,
Measurement of the thermal conductivity of copper samples between
30 and 150 mK. Cryogenics, 44, 875878.
[23] Cohen, M.L., 1977, Measurement of the thermal properties of human
skin: A review. Journal of Investigative Dermatology, 69, 333338.
[24] Challoner, A.V.J., 1975, Accurate measurement of skin blood ow by
a thermal conductance method. Medical & Biological Engineering, 13,
196201.
[25] Khaled, A.R.A. and Vafai, K., 2003, The role of porous media in
modeling ow and heat transfer in biological tissues. Journal of Heat
and Mass Transfer, 46, 49895003.
[26] Van den Berg, J., 1975, Thermal conductivity and heat transfer of the
human skin. Bibl Radiology, 6, 166177. PMID 1180847.
[27] Giering, K., Minet, O., Lamprecht, I. and Muller, G., 1995, Review of
thermal properties of biological tissues. In: G. Muller and A. Roggan
(Eds) Laser Induced Interstitial Thermotherapy, (SPIE Press, Bellingham, WA, USA).
[28] Freitas Jr, R.A., 2003, Nanomedicine: Biocompatibility (Landes
Bioscience Press, Austin, TX, USA).
[29] Kreith, F. and Black, W., 1980, Basic Heat Transfer, 1st edn (Harper
and Row).

Copyright of Journal of Medical Engineering & Technology is the property of Taylor & Francis Ltd and its
content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for individual use.