1846 IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 58, NO. 6, JUNE 2011
Wireless Implantable Electronic Platform for Chronic
Fluorescent-Based Biosensors
Pietro Valdastri∗, Member, IEEE, Ekawahyu Susilo, Thilo Fo¨rster, Christof Strohho¨fer,
Arianna Menciassi, Member, IEEE, and Paolo Dario, Fellow, IEEE
Abstract—The development of a long-term wireless implantable
biosensor based on fluorescence intensity measurement poses a
number of technical challenges, ranging from biocompatibility to
sensor stability over time. One of these challenges is the design of a
power efficient and miniaturized electronics, enabling the biosen-
sor to move from bench testing to long term validation, up to its final
application in human beings. In this spirit, we present a wireless
programmable electronic platform for implantable chronic moni-
toring of fluorescent-based autonomous biosensors. This system is
able to achieve extremely low power operation with bidirectional
telemetry, based on the IEEE802.15.4-2003 protocol, thus enabling
over three-year battery lifetime and wireless networking of mul-
tiple sensors. During the performance of single fluorescent-based
sensor measurements, the circuit drives a laser diode, for sensor
excitation, and acquires the amplified signals from four differ-
ent photodetectors. In vitro functionality was preliminarily tested
for both glucose and calcium monitoring, simply by changing the
analyte-binding protein of the biosensor. Electronics performance
was assessed in terms of timing, power consumption, tissue ex-
posure to electromagnetic fields, and in vivo wireless connectivity.
The final goal of the presented platform is to be integrated in a
complete system for blood glucose level monitoring that may be
implanted for at least one year under the skin of diabetic patients.
Results reported in this paper may be applied to a wide variety of
biosensors based on fluorescence intensity measurement.
Index Terms—Biotelemetry, calcium sensing, chronic implants,
fluorescence-based biosensor, fluorescence resonant energy trans-
fer (FRET), glucose monitoring, implantable devices.
I. INTRODUCTION
R ECENT achievements in miniaturized electronics andsmart system integration have led to a proliferation of
wearable and implantable devices for the monitoring of vital
signs. These systems hold the promise to lower the threshold
for early diagnosis and to dramatically enhance the quality of
Manuscript received September 22, 2010; revised November 27, 2010 and
January 24, 2011; accepted February 20, 2011. Date of publication March 7,
2011; date of current version May 18, 2011. This work was supported by FP6
European Project P. Cezanne under Grant IST-2006-IP031867. Asterisk indi-
cates corresponding author.
∗P. Valdastri is with the BioRobotics Institute, Scuola Superiore Sant’Anna,
56127 Pisa, Italy (e-mail: pietro@sssup.it).
E. Susilo, A. Menciassi, and P. Dario are with the BioRobotics Institute,
Scuola Superiore Sant’Anna, 56127 Pisa, Italy (e-mail: e.susilo@sssup.it;
arianna@sssup.it; paolo.dario@sssup.it).
T. Fo¨rster and C. Strohho¨fer are with Fraunhofer EMFT, 80686 Munich,
Germany (e-mail: thilo.foerster@emft.fraunhofer.de; christof.strohhoefer@
emft.fraunhofer.de).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TBME.2011.2123098
life, toward the ultimate implementation of personalized health-
care [1].
Pressure sensors embedded in disposable contact lenses [2] or
in cardiovascular stents [3] are just two examples of how smart
system integration may enable continuous monitoring of a phys-
ical parameter. Similarly, implantable chemical biosensors may
have a disruptive impact in this field [4]. In particular, fluores-
cence techniques provide high sensitivity and ease of operation
for chemical sensing, thus making fluorescence-based sensor
commonly adopted worldwide [5]. A number of fluorescence
microscopy and spectroscopy techniques based on the lifetime,
anisotropy, or intensity of the emission of fluorescent probes
have been developed over the years [6]. These enormously sen-
sitive techniques are usually implemented in a lab-bench setting.
The development of a portable, or even implantable, platform
based on fluorescence intensity measurement could pave the
way to a new generation of monitoring systems.
Among other chemical species that can be detected in this
manner [7], [8], glucose is by far the most relevant one in diag-
nostics. The availability of a reliable long-term blood glucose
level (BGL) sensor would have a disruptive impact on health-
care and diabetes management, considering that over 5% of the
world’s population suffers from diabetes [9] and this number is
increasing rapidly [10]. Wireless electronic systems have been
suggested for amperometric [11], [12] and near-infrared spec-
troscopy based [13] glucose sensors. However, problems with
existing continuous BGL monitors have encouraged alterna-
tive approaches to glucose sensing [14]. In particular, solutions
based on fluorescence intensity and lifetime have special ad-
vantages, including sensitivity and potential long-term stability.
Special attention has been paid to the use of biosensors gen-
erating fluorescence intensity signals proportional to glucose
level with the help of fluorescence resonant energy transfer
(FRET) [15]–[17]. A transdermal FRET-based system for con-
tinuous BGL monitoring is reported in [18]. To further reduce in-
vasiveness, functionalized fluorescent microparticles [19], [20]
can be injected in the patient and used to transdermally moni-
tor glucose concentration. Despite promising results, this latter
approach is not self-contained, since it requires external light
excitation and a video camera for image analysis.
In this spirit, a miniaturized implantable system for chronic
fluorescence-based glucose sensing would allow continuous
monitoring, without affecting the patient’s daily lifestyle. Sev-
eral challenges must be faced to achieve this relevant goal, most
of which are related to the biosensor itself (e.g., biocompatibil-
ity and sensor stability over time). Concerning electronics, high
energy excitation pulses must be provided to drive the sensor
0018-9294/$26.00 © 2011 IEEE
VALDASTRI et al.: WIRELESS IMPLANTABLE ELECTRONIC PLATFORM FOR CHRONIC FLUORESCENT-BASED BIOSENSORS 1847
Fig. 1. Implant concept design.
and low noise photodetectors are required to collect the amount
of energy transferred during the FRET event. Therefore, design-
ing a low power, miniaturized and self-contained electronics is
not straightforward. On the other hand, such electronics could
easily be extended to the fluorescence-based detection of other
chemical species, thus broadening its potential impact.
The aim of this paper, therefore, is to describe a miniatur-
ized electronic platform for implanted long-term fluorescence
biosensors. The platform allows fluorescence excitation and de-
tection by driving a laser diode light source and phototransistors
as detectors. Signals are amplified and transmitted across the
skin to a personal mobile device, e.g., a cell phone or a laptop PC,
where data can be displayed instantly. The electronic platform
was designed bearing in mind the main features of implantable
devices, namely small size, low power consumption, and safety.
The functionality of the electronics was demonstrated in vitro
with a FRET-based biosensor responding to changes in glu-
cose concentration. In order to show how the proposed platform
can be adapted to the monitoring of different chemical species,
we repeated the in vitro experiment using a calcium-binding
molecule. Finally, telemetry performance was assessed in terms
of both safety and in vivo operation.
II. SYSTEM OVERVIEW
A. Sensing Principle
Despite the focus of this paper is on the electronics, an
overview of the complete system is provided in this section,
while a more detailed description is reported in [21]. The com-
plete system is conceived to work as an under-skin implant,
detecting analytes in interstitial fluids. It comprises a biosensor
for signal generation, measurement, control and communication
electronics, and a lithium intercalation cell as power supply. A
schematic representation of the implant is reported in Fig. 1.
The biosensor is based on the principle of FRET [14], [22] be-
tween two fluorescent proteins, cyan fluorescent protein (CFP)
and yellow fluorescent protein (YFP). These two proteins are
covalently linked to an analyte binding protein (ABP). Analyte
detection and transduction with this protein is schematically
shown in Fig 2. Binding of the analyte to the bioreceptors in-
duces a conformational change in the proteins and an incidental
Fig. 2. (a) Schematic representation of analyte detection with the sensor pro-
tein. The fluorescent proteins CFP and YFP are covalently attached to the analyte
recognition protein ABP. (b) Upon analyte binding, ABP changes its confor-
mation and alters the distance between the fluorescent proteins. The resulting
change in fluorescence resonant energy transfer is proportional to the analyte
concentration.
variation in the distance between the two fluorescent proteins.
This translates into a change in intensity of the fluorescence
emissions from the individual fluorescent proteins. The com-
bined protein construct is immobilized inside a hydrogel by
physical entrapment. The hydrogel has the shape of a cylindri-
cal rod with diameter below 1 mm and length of around 8 mm.
It is suspended at its ends in a liquid-filled chamber, and fluo-
rescence photodetectors are located at its sides. The hydrogel
serves as optical waveguide for the fluorescence excitation light
(405 nm) provided by a laser diode and coupled into the hydro-
gel through one of its end facets. Fluorescence is detected in two
wavelength regions centered around 480 nm (CFP) and 527 nm
(YFP). Wavelength discrimination is achieved by placing in-
terference filters in front of the photodiodes. When the analyte
concentration changes, the energy transfer between the fluores-
cent proteins changes accordingly, as also the relative intensity
of fluorescence on the photodetectors. A schematic representa-
tion of the sensor head is included in Fig. 1. The sensor measures
changes in analyte concentration inside the liquid-filled cham-
ber, i.e., the measurement chamber. When the sensor is used
in vivo or in liquids contaminated with larger molecules, bacte-
ria or cells, access to the measurement chamber is closed by a
membrane, which allows ions and glucose to freely diffuse into
the sensor chamber while retaining cells and larger molecules.
As common for implanted biosensors, stability over time is
the main hurdle to be faced with. As regards the described sens-
ing strategy, the deactivation of ABP, CFP, and YFP, the decay of
the florescence signal over time, pollution of the measurement
chamber by serum proteins or blood clots, and the need for cal-
ibration are issues that need to be carefully addressed in future
works. In the following, we focus on the design of an electronic
platform enabling the implementation of this strategy. As the
ones listed earlier, this is a relevant issue to be addressed, due
to the severe specifications described in the next section.
B. Requirements for Implant Electronics
The in vivo repetitive sampling of a fluorescent-based biosen-
sor poses several special requirements on the electronics, which
we will address in the course of this article.
1848 IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 58, NO. 6, JUNE 2011
Fig. 3. Typical timing sequence for fluorescence-based sensor reading.
I/O features: The electronics must properly drive a laser light
source and acquire light intensity from four different photode-
tectors. One pair of the photodetectors provides readings for the
fluorescence intensities emitted from CFP and YFP, from which
the FRET ratio is calculated. The second is a reference pair. It is
worth mentioning that driving a laser light source, required for
biosensor excitation, within a long-term wireless implant is ex-
tremely challenging in terms of power consumption and current
requirements.
Timing: As highlighted in Fig. 3, each sensor reading is sub-
ject to specific timing requirements. First the laser diode must
be switched on. Then, after a user adjustable delay time Tdel
allowing stabilization of the signal, the sampling of the four
photodetectors may start. Two other important parameters that
the user may need to adjust are sampling time Tsampling and
number of samples to be acquired N. Once the measurement is
over, the averages of the four signals over N samples are made
available to the end user. The last parameter is repetition time
(Trep ), which represents the user adjustable period for a single
measurement.
Dimensions: The overall dimensions of the electronics must
be as small as possible, thus minimizing discomfort once the
sensor has been implanted under the skin. The target dimen-
sions of the overall sensing system, including the measurement
chamber, are 50 mm in length, and 20 mm in diameter, thus
comparable to an implantable pacemaker.
Energy consumption: At least one year of operative battery
lifetime is the minimum requirement to justify implantation.
Once this time has elapsed, the implant may be replaced with a
new one, in a day-hospital procedure.
Wireless connectivity: The wireless technology should guar-
antee data transfer reliability between the implant and a data col-
lector located within a radius of a few meters from the patient.
Bidirectional communication is preferable, in order to allow the
user to interact with the implant and to adjust measurement pa-
rameters. Furthermore, by adopting a wireless technology that
enables sensor networking, future scenarios with complex ar-
chitectures may be envisaged (e.g., a single patient moving in a
hospital environment and connecting to the closest data collector
available for downloading sensor data, or a cluster of sensors, ei-
ther implanted or worn by the patient, communicating together
within a personal area network). Finally, the wireless energy
density of telemetry passing through human tissues must com-
ply with safety regulations [23] in order to prevent any possible
risk for the patient.
C. Architectural Overview of the Electronics
In order to fulfill the aforementioned requirements, the system
architecture needs to comprise a programmable device with an
adequate amount of I/O, low power consumption, a compact
package, and, possibly, bidirectional wireless communication
features. This core component must drive the laser diode to
provide excitation to the fluorescence biosensor and acquire
the four properly amplified analog signals coming from the
photodetectors. A switch driven by the programmable unit can
be placed between the battery and the sensing circuitry. This
would allow the analog circuitry to be switched off whenever
possible, thus saving battery power.
A receiving unit—located outside the patient—must be able
to communicate with the implant, monitoring its status and en-
abling the end user to display and store the sensor readings and
to share them with medical staff. The most straightforward way
to achieve this goal is to develop a universal serial bus (USB)
dongle that connects the implant to a laptop PC. This is the
approach pursued in this article. Further development may in-
tegrate the receiving unit in a mobile device, e.g., a personal
digital assistant or smart phone.
D. Hardware Description
An ideal candidate as core component of the implant elec-
tronics is the CC2430 wireless microcontroller from Texas
Instruments. This device embeds an 8051 programmable micro-
controller together with an IEEE 802.15.4-2003 [24] compatible
transceiver in a 7 mm × 7 mm package. An eight-channel ΣΔ
analog to digital converter (ADC) provides 12-bit resolution
sampling of the amplified photodetector signals. The CC2430
includes a five-channel direct memory access (DMA) controller,
which can be used to relieve the microcontroller core of handling
data movement operations and to achieve high overall perfor-
mance with good power efficiency. In particular, the DMA con-
troller can move data from a peripheral unit, such as the ADC,
to memory with minimum CPU intervention. The on-board
16-bit timer can be used to obtain the desired Tsampling during
the acquisition of the N photodetector samples. By exploiting
the combined use of these three peripherals, a single sensor
reading can be acquired minimizing microcontroller operations
and consequently reducing power consumption. A 4-kB block
of nonvolatile memory can be used to store the acquired data
until receipt of an acknowledgment of correct transmission.
As regards telemetry, the CC2430 features a 2.4-GHz
transceiver complying with the IEEE 802.15.4-2003 standard
[24] with a 250-kbps data rate. Despite tissue absorption at
2.4 GHz, previous work [25], [26] has demonstrated that this
can still be a viable solution for implantable devices. A full
ZigBee stack [27] can be implemented, when required, without
any further hardware modification. Thanks to the availability
VALDASTRI et al.: WIRELESS IMPLANTABLE ELECTRONIC PLATFORM FOR CHRONIC FLUORESCENT-BASED BIOSENSORS 1849
of program flash up to 128 kB, the platform may also support
application with a complex sensor network scenario. By using
adjustable transmission power, the microcontroller can be pro-
grammed to always use the minimum level of energy required.
This allows it to achieve reliable communication [26] and save
battery. Few external components, e.g., a 32-MHz crystal and a
surface mount balun (BD2425N50200A00, Anaren) for antenna
impedance matching, are required for the correct operation of
the radio section. In the present application, a 31-mm long wire
serves as λ/4 whip antenna.
Four different power modes can be used for the CC2430 to
manage power consumption. In particular, a deep sleep mode,
featuring nearly zero power consumption, can be used until the
system is implanted under skin. An external magnetic switch
can be used to wake up the microcontroller from the deep sleep
mode. Once the implant is in place, the microcontroller can al-
ternate operation between full functional mode and sleep mode,
as explained in Section II-E. A watchdog timer is also provided,
thus preventing the implant from code failure.
Other features of the CC2430 used in this application include
temperature and battery level monitoring. Whenever the device
is implanted and activated, the built-in temperature sensor can
monitor the patient’s body temperature, while the battery level
monitor can provide real time information on the state of the
power source and report to the central station when approaching
a critical level.
The analog front end is responsible for the conversion of the
fluorescence signals from CFP and YFP to a voltage that is
passed on to the wireless microcontroller. As first step, the pho-
tocurrent generated by the fluorescence of the FRET proteins
is converted to a voltage with a conversion factor of 106 V/A.
This is achieved with the help of a first operational amplifier
(AD8034, Analog Devices) stage applied to the photodetector
(TEMT6000, Vishay). As second step, the voltage undergoes
further amplification in a differential amplifier. A reference volt-
age supply (ADR361, Analog Devices) included in the analog
front end helps stabilize the output voltage and minimize the
offset during the second stage amplification.
Two identical photodetectors are combined into a single unit,
sharing common voltage supply, ground and reference. How-
ever at the signal conditioning stage, each of them has its own
independent circuit. Output voltages from these circuits are pro-
portional to the photocurrents and can be fed directly to the mi-
crocontroller ADC inputs, necessitating a mere four connections
between the main control board and the analog front end.
A blue-violet (405 nm) laser diode (DL-4146-101S, Sanyo) is
used to excite the fluorescent proteins. A dedicated laser diode
driver (iC-WJ, iC-Haus) allows for power output regulation,
providing a simple on/off interface for the microcontroller.
Both the analog front end and the laser diode driving circuitry
require 5.5 V as power supply, while the selected battery (devel-
oped for this specific application by Tadiran Batteries in a 1520
package, providing 360 mAh and supplying current spikes up
to 750 mA) operates at 4 V. To match the two voltage levels, a
stepup converter (MAX643, Maxim Integrated Circuits) is used
in the adjustable-output configuration, thus requiring two addi-
tional external resistors. The oscillation frequency of the stepup
Fig. 4. (a) Miniaturized electronics. (b) Sensor device before encapsulation.
(c) Final packaging of the sensor.
is 37 kHz. In order to provide enough current to the laser diode
after voltage conversion, the stepup is operated in a high output
current mode. This requires a 10.4 mm × 9.4 mm × 5.8 mm
surface mount power inductor (B82475A1, Epcos), a high-
speed switching N-channel enhancement mode MOS transistor
(BSH103, Philips Semiconductors) and a fast switching diode
(BAR64-04, Siemens). Finally, an analog switch (MAX4736,
Maxim Integrated Circuits) is integrated that allows permanent
or temporal disconnection of all the aforementioned analog parts
by the microcontroller, in order to save power.
Three different boards were designed to accommodate all
the components. The main control board hosts the wireless mi-
crocontroller, the analog switch, the stepup converter and the
connectors to the battery and to the other boards. The board
is circular with 17-mm diameter and is built upon four layers
in rigid flexible circuit material reaching a total thickness of
12.3 mm after the components have been mounted on the top
and bottom layers. Regarding layer functionality, all the layers
are used as signal planes. Only a small area on the third layer is
left out as a ground plane in compliance with the Anaren applica-
tion note [28]. This ground plane is recommended for correctly
matching the impedance of the balun to the whip antenna.
A second board comprises the laser diode and its driver, and
plugs into the main board by a two-pole connector. The board
measures 13 mm in diameter and 4 mm in thickness (with
mounted components) and is shaped to fit into the bottom of
the measurement chamber.
The analog front end is fabricated on a double-sided rigid-flex
printed circuit board measuring 10.5 mm × 5.5 mm. Two front
end boards, implementing two photodetectors each, are con-
nected to the main control board by means of custom designed
four-pole flexible circuit connectors (CviLux Corporation).
The completely assembled electronics is represented in
Fig. 4(a). The two analog front end boards and the laser diode
board are designed to be first integrated and packaged with the
measurement chamber, and then connected to the main board,
as shown in Fig. 4(b). As final step of system integration, a
polyarylether ether ketone can be screwed on the bottom of the
measurement chamber, thus sealing the implant [Fig. 4(c)]. The
final dimensions are 63 mm in length and 21 mm in diameter.
As regards the external unit, a CC2430 is interfaced to a
serial-USB signal converter (FT232R, FTDI Chip) through the
1850 IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 58, NO. 6, JUNE 2011
asynchronous serial port (UART). This dongle is connected to
the USB port of a standard PC, where a bespoke user interface,
developed in Labview 8.2 (National Instruments, Austin, TX),
is installed.
E. Code Description
Regarding firmware, the architecture comprises three differ-
ent programmable units: the first unit is the CC2430 integrated in
the implant. It communicates with the CC2430 of the USB don-
gle, which is the second programmable unit. The firmware codes
for both these units were developed and debugged by using IAR
Embedded Workbench (IAR Systems). The third programmable
unit is the PC, where the user interface is installed. As a result,
three independent codes running in parallel, but interacting with
each other, were developed.
Since the platform needs to be able to execute different com-
mands issued by the user (e.g., fluorescent-based sensor read-
ing, temperature measurement, battery level enquiry, adjustment
of measurement parameters, and setting of transmission power
level), each command is codified with an ASCII symbol. Sim-
ilarly, all the possible system failures that may occur within
the operation (e.g., implant out of communication range, USB
dongle failure) are properly codified and presented on the user
interface.
During the code development phase, top priority was given
to the implant firmware, especially to minimizing power con-
sumption and to maximizing battery lifetime. For this reason,
the implant is programmed to cyclically wake up every Tsleep
and to request a command from the external unit. If the USB
dongle is within communication range and the user has already
issued a command, the implant receives it and acts accordingly.
Once the command has been executed, the implant goes back
to sleep mode. If the USB dongle is not nearby or the user has
not issued any command within a Ttimeout (set to 25 ms in the
present application), the implant goes back to sleep mode as
well. Tsleep can be adjusted by the user by making a tradeoff
between responsiveness of the implant and its battery lifetime.
The USB dongle code and the user interface are built around
the implant code, allowing for a reliable communication and an
effective failure management. The user interface allows the user
to select the identification number of a specific implant, to set
all the different parameters of a single measurement (i.e., Tdel ,
Tsampling , N, Trep ), to visualize and store the measurements, and
to gather temperature and battery voltage. The received signal
strength may also be displayed thanks to the receive signal
strength indicator CC2430 feature.
III. EXPERIMENTAL RESULTS
A. In Vitro Functionality Test
Preliminary functionality tests were carried out with the elec-
tronic platform and a hydrogel waveguide biosensor [21] con-
taining glucose binding proteins [29], [30] in the FRET com-
pound molecule.
Tests were performed by dipping the hydrogel waveguide sen-
sor into a solution mimicking human interstitial fluid. It consists
Fig. 5. FRET response to glucose binding and release, obtained from the
fluorescence signals measured by the sensor. Measurement parameters: N =
256, Tdel = 255 ms, Tsampling = 100 μs, Trep = 10 s.
of 10-mM HEPES buffer solution adjusted to pH 7.4 with HCl,
and containing 122-mM NaCl, 25-mM NaHCO3 , 3-mM KCl,
1.2-mM CaCl2 , 1.2-mM MgCl2 , 0.5-mM K2HPO4 . A 30-mg/ml
bovine serum albumin was added to simulate the protein content
in interstitial fluid. The sensor was placed repeatedly in these
solutions containing 2-mM and 10-mM glucose, to evaluate the
sensor response. All experiments were performed at 37 ◦C. Mea-
surements were performed every 10 s, adopting the following
parameters: N = 256, Tdel = 255 ms, Tsampling = 100 μs.
A typical measurement is shown in Fig. 5. The FRET signal
was calculated as the ratio between the fluorescence emission
from YFP and CFP. Fig. 5 shows an initially stable signal at
2-mM glucose concentration. At time t = 10 min, the sensor
was immersed in 10-mM glucose solution. After a transitory
period of roughly 10 min, the signal stabilizes at a higher FRET
level. The signal returns to the lower level when the sensor is
placed back into the solution with low glucose concentration.
The transition between low and high FRET levels and vice
versa is caused by glucose diffusion into and out of the sensor.
It shows exponential behavior (drawn line in Fig. 5) with a
time constant of around 4–5 min, which is determined by the
analyte’s diffusion into the hydrogel waveguide.
Key to the detection of glucose is the particular FRET protein
employed in the hydrogel waveguide. By simply replacing that
with a similar one sensitive to a different analyte, other vital
parameters may be monitored with the same electronic platform.
To prove this, we replaced the hydrogel used for the previous
test with a similar one, containing a calcium sensitive FRET
compound.
Similar tests were performed by dipping the hydrogel wave-
guide sensor into 50-mM TRIS buffer solutions adjusted to pH 8
with HCl, and containing either 0- or 5-mM CaCl2 . Initially,
the sensor was placed into a buffer solution without CaCl2 for
15 min, then for 15 min into a buffer solution containing 5-mM
CaCl, and finally once again into the solution without CaCl2 .
Measurements were performed every 10 s, adopting the same
VALDASTRI et al.: WIRELESS IMPLANTABLE ELECTRONIC PLATFORM FOR CHRONIC FLUORESCENT-BASED BIOSENSORS 1851
Fig. 6. FRET response to calcium binding and release, obtained from the
fluorescence signals measured by the sensor. Measurement parameters: N =
256, Tdel = 255 ms, Tsampling = 100 μs, Trep = 10 s.
Fig. 7. Typical voltage output from a single channel of the analog front end
for the following parameters setting: N = 256, Tdel = 255 ms, Tsampling =
100 μs.
measurement parameters as aforementioned. A typical measure-
ment is shown in Fig. 6, demonstrating similar performance of
calcium sensing as demonstrated in Fig. 5 for glucose sensing.
These results demonstrate the electronic platform’s potential
beyond in vivo sensing, in applications such as distributed au-
tonomous environmental monitoring, where it can serve as the
sensor control and communications unit of a sensor network.
B. Timing
A plot showing the typical voltage output from a channel of
the analog front end during a single measurement is reported
in Fig. 7. As previously explained, the microcontroller starts
sampling the voltage Tdel after the laser diode has been switched
on. N samples are acquired, each every Tsampling , and then
averaged to provide a single value per channel. The delay time
Tdel should therefore be chosen so that the sampling starts only
once the photodetector voltage has reached its steady state value.
On the other end, Tdel should be minimized in order to save
Fig. 8. Current consumption for the following parameters setting: N = 256,
Tdel = 255 ms, Tsampling = 100 μs.
energy. N and Tsampling must be adjusted considering the noise
of the voltage signal during the measurement time. With the help
of a plot as the one shown in Fig. 7, the measurement settings
can be optimized. In particular, Fig. 7 shows the signal pulse
from an in vitro glucose test with the fluorescence sensor. The
mean value of the voltage output reaches 98% of its regime after
81 ms; therefore, Tdel can be reduced by about 70% to increase
battery lifetime.
C. Power Consumption
In order to quantify the current consumption of the complete
electronic platform, the voltage drop across a 1-Ω resistor placed
in series to the positive voltage supply terminal was acquired
during a single measurement of the biosensor. This measure-
ment was performed with a 2-GS/s digital storage oscilloscope
(TPS2024, Tektronix, Inc.). The plot shown in Fig. 8 shows
the single phases of the measurement and quantifies each sin-
gle contribution in terms of power consumption. In particular,
the implant exits from sleep mode and requests a command.
In this case, the request for a single biosensor measurement
and the related parameter setting (N = 256, Tdel = 255 ms,
Tsampling = 100 μs) are received. Then, the analog circuitry,
including the laser diode, is switched on for Texc , as defined in
Fig. 3 and the measurement is performed. Directly following the
measurement, the averaged 12-bit values of the four amplified
photodetector signals are transmitted to the external unit and the
implant goes back to sleep mode. It is worth mentioning that
the periodic peaks and valleys of current reported in the inset
of Fig. 8 are due to the stepup voltage converter, working at
37 kHz. These peaks show a periodic current drain, required to
hold the voltage level to 5.5 V, starting from 4 V.
As detailed in [26], battery lifetime can be estimated by using
the following average current consumption:
I =
∑
i IiTi∑
i Ti
(1)
where Ii and Ti are the current consumption and the time in-
terval related to a single action performed by the system within
1852 IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 58, NO. 6, JUNE 2011
a periodic operation. Given the parameter setting used, a sin-
gle sensor measurement requires an Isensing = 94 mA for a
total time of Tsensing = 392 ms. In the case, of chronic BGL
monitoring, clinical considerations require that a single mea-
surement should be acquired every 30 min. Therefore, Tsleep =
1799.6 s may be assumed. As regards Isleep , since this is below
the resolution of the test bench equipment used, the value was
estimated from the datasheet of the components used as Isleep =
1.5 μA. By applying (1), an average current consumption of I =
22 μA is obtained, which translates into over 22 months of con-
tinuous operation using the previously mentioned battery with
360-mAh capacity. A duration of 40 months can be achieved by
optimizing Tdel as reported earlier.
D. Tissue Exposure to Electromagnetic Fields
Once the described electronics demonstrated that it could
effectively drive a fluorescence-based sensor for sufficient time
to justify implantation, further tests were performed to assess
the telemetric link.
The maximum value of plane wave power density emitted
by the implantable unit was measured in order to compare it
with international safety regulations. The power density was
measured by a portable field strength meter (8053, PMM, Italy)
located in the closest proximity to the module antenna. Its probe
was oriented along the maximum radiation of the antenna, re-
sulting in 0.002 W/m2 . This value is considerably lower than
the ICNIRP reference level for general public exposure to time-
varying electric and magnetic fields [23], fixed at 10 W/m2 for
a signal frequency of 2.4 GHz. Thus, the implant can be consid-
ered safe in electromagnetic terms.
E. Wireless Connectivity
Finally, the system was implanted under the skin of a pig in
order to assess telemetry performances in an in vivo scenario.
The experiment was performed in an authorized laboratory, with
the assistance and collaboration of a specially trained medical
team, in accordance to all ethical considerations and regulatory
issues related to animal experiments. A 35-kg domestic swine
was sedated with a Ketamine injection in addition to Propofol
and Stresnyl dosed intravenously. After intubation, anesthesia
was maintained with Enflurane. The sensor, in its final packag-
ing configuration, was implanted under the skin (3-cm depth)
in the pelvic region of the swine [see Fig. 9(a)] and the incision
was sutured. The USB dongle—shown in close proximity to
the implant in Fig. 9(b)—was placed 2 m from the animal, in
the operating room environment. A temperature sensor reading
request was issued by the user interface every 10 s for a total of
4 h. At the end of the experiment over 98% of telemetric trans-
missions of the temperature reading were successfully achieved.
The 2% loss may be due to electromagnetic interferences from
other instrumentations used for measuring physiological param-
eters such as electrocardiogram and pressure. These failures do
not represent a critical issue in an operative scenario, because
a second sensor reading can be requested by the external unit
should the first one fail.
Fig. 9. In vivo test of telemetry. (a) Sensor before under-skin implantation.
(b) USB dongle placed in proximity to the implanted sensor.
IV. CONCLUSION
The main goal of this paper was to describe a miniaturized
wireless electronic platform, suitable for in vivo monitoring
of chemical species, such as glucose or calcium. Our solu-
tion enables long-term wireless monitoring of an implanted
fluorescent-based sensor in a miniaturized package, measur-
ing 63 mm in length and 21 mm in diameter. Over three years,
operational battery lifetime can be obtained by properly adjust-
ing measurement parameters. The proposed electronic platform
can be easily applied to the monitoring of a different analyte by
changing the binding protein embedded in the FRET compound
molecule, as demonstrated by in vitro tests performed first with
a glucose sensitive FRET compound, and then with a calcium
sensitive one. Wireless networking of multiple sensors can be
implemented without any further hardware modification, thanks
to a IEEE802.15.4-2003 protocol compliant technology. Further
tests demonstrated the safety of the platform as regards exposure
to time-varying electric and magnetic fields and the functional-
ity of the telemetric link once the system was implanted under
skin.
Despite these encouraging results, some current limitations to
the study described here require comments. As a first important
step, the complete system needs to be validated by long-term
in vitro and in vivo trials in order to assess sensor stability
over time, with particular emphasis on both sensor packaging
and the interface of the sensor with interstitial fluids. While
a purposely developed membrane may shield the measurement
chamber from pollutant, calibration will likely involve referenc-
ing from time to time with fingerprick measurements, as well
as the use of an internal calibrant, which mimics the drift of
the glucose sensitive protein. Thanks to bidirectional wireless
communication, possible adjustments of sensor calibration can
be implemented straightforwardly, thus preventing the need of
explanting the system.
Another direction of future work will focus on further minia-
turization, in order to meet the specified dimensional require-
ments. Reduction in volume can be achieved by optimizing the
arrangement of the internal electronic components or by adopt-
ing a smaller battery that would still provide one year lifetime.
Finally, despite performing preliminary safety validation, all
the steps for clinical approval must be taken before applying the
platform to humans.
To conclude, it is worth mentioning that all these further steps
will originate from the electronic platform described in this
VALDASTRI et al.: WIRELESS IMPLANTABLE ELECTRONIC PLATFORM FOR CHRONIC FLUORESCENT-BASED BIOSENSORS 1853
paper, which represents a valid reference for any long-term im-
plantable wireless system exploiting a fluorescence-based sens-
ing principle.
ACKNOWLEDGMENT
The authors would like to thank T. Economou (IMBB-
FORTH, Iraklion, Greece) for providing the sensing proteins,
I. Lacik (Polymer Institute of the Slovak Academy of Sciences,
Bratislava, Slovakia) for the hydrogel waveguides, E. Elster
from Tadiran Batteries, Kiryat Ekron, Israel, for providing the
power source, and R. Di Leonardo and R. Lazzarini for their
invaluable help.
REFERENCES
[1] K. Malasri and L. Wang, “Securing wireless implantable devices for
healthcare: Ideas and challenges,” IEEE Commun. Mag., vol. 47, no. 7,
pp. 74–80, Jul. 2009.
[2] M. Leonardi, E. M. Pitchon, A. Bertsch, P. Renaud, and A. Mermoud,
“Wireless contact lens sensor for intraocular pressure monitoring: Assess-
ment on enucleated pig eyes,” Acta Ophthalmologica, vol. 87, no. 4,
pp. 433–437, 2009.
[3] E. Y. Chow, A. L. Chlebowski, S. Chakraborty, W. J. Chappell, and
P. P. Irazoqui, “Fully wireless implantable cardiovascular pressure monitor
integrated with a medical stent,” IEEE Trans. Bio.-Med. Eng., vol. 57,
no. 6, pp. 1487–1496, Jun. 2010.
[4] P. Tathireddy, L. Rieth, A. Sharma, and F. Solzbachert, “Implantable
microsystems and neuro electronic interfaces,” in Bionic Health: Next
Generation Implants, Prosthetics and Devices, London, U.K.: IET, 2009,
pp. 1–22.
[5] L. Basabe-Desmonts, D. N. Reinhoudt, and M. Crego-Calama, “Design
of fluorescent materials for chemical sensing,” Chem. Soc. Rev., vol. 36,
pp. 993–1017, 2007.
[6] J. R. Lakowicz, Topics in Fluorescence Spectroscopy. New York:
Springer-Verlag, 1994.
[7] J. R. Siqueira, Jr., L. Caseli, F. N. Crespilho, V. Zucolotto, and O. N.
Oliveira, Jr., “Immobilization of biomolecules on nanostructured films for
biosensing,” Biosens. Bioelectron., vol. 25, no. 6, pp. 1254–1263, 2010.
[8] C. R. Taitt, G. P. Anderson, and F. S. Ligler, “Evanescent wave fluorescence
biosensors,” Biosens. Bioelectron., vol. 20, no. 12, pp. 2470–2487, 2005.
[9] G. A. Urban, “Micro-and nanobiosensors—state of the art and trends,”
Meas. Sci. Technol., vol. 20, no. 1, p. 012001, 2009.
[10] S. Wild, G. Roglic, A. Green, R. Sicree, and H. King, “Global prevalence
of diabetes,” Diabetes Care, vol. 27, no. 5, pp. 1047–1053, May 2004.
[11] R. D. Beach, R. W. Conlan, M. C. Godwin, and F. Moussy, “Towards a
miniature implantable in vivo telemetry monitoring system dynamically
configurable as a potentiostat or galvanostat for two- and three-electrode
biosensors,” IEEE Trans. Instrum. Meas., vol. 54, no. 1, pp. 61–72, Feb.
2005.
[12] M. M. Ahmadi and G. A. Jullien, “A wireless-implantable microsystem
for continuous blood glucose monitoring,” IEEE Trans. Biomed. Circ.
Sys., vol. 3, no. 3, pp. 169–180, Jun. 2009.
[13] K. S. Kanukurthy, M. B. Cover, and D. R. Andersen, “Data acquisition
unit for an implantable multi-channel optical glucose sensor,” in Proc.
IEEE Electro. Inform. Technol. (EIT), 2007, pp. 32–37.
[14] J. C. Pickup, F. Hussain, N. D. Evans, O. J. Rolinski, and D. J. Birch,
“Fluorescence-based glucose sensors,” Biosens. Bioelectron., vol. 20,
no. 12, pp. 2555–2565, 2005.
[15] B. Tang, L. Cao, K. Xu, L. Zhuo, J. Ge, Q. Li, and L. Yu, “A new
nanobiosensor for glucose with high sensitivity and selectivity in serum
based on fluorescence resonance energy transfer (FRET) between CdTe
quantum dots and Au nanoparticles,” Chem.—Eur. J., vol. 14, no. 12,
pp. 3637–3644, 2008.
[16] F. Khan, L. Gnudi, and J. C. Pickup, “Fluorescence-based sensing of glu-
cose using engineered glucose/galactose-binding protein: A comparison
of fluorescence resonance energy transfer and environmentally sensitive
dye labelling strategies,” Biochem. Biophys. Res. Commun., vol. 365,
no. 1, pp. 102–106, 2008.
[17] A. Bartolome, L. Smalls-Mantey, D. Lin, G. Rao, and L. Tolosa, “Fret-
based glucose monitoring for bioprocessing,” in Genetically Engineered
Probes for Biomedical Applications, vol. 6098, no. 1. San Jose, CA: SPIE,
2006, p. 60980L.
[18] R. Ballerstadt, C. Evans, A. Gowda, and R. McNichols, “In vivo perfor-
mance evaluation of a transdermal near-infrared fluorescence resonance
energy transfer affinity sensor for continuous glucose monitoring,” Dia-
betes Technol. Therapeut., vol. 8, no. 3, pp. 296–311, 2006.
[19] H. Shibata, Y. J. Heo, T. Okitsu, Y. Matsunaga, T. Kawanishi, and S.
Takeuchi, “Injectable hydrogel microbeads for fluorescence-based in vivo
continuous glucose monitoring,” Proc. Nat. Acad. Sci. USA, vol. 107, no.
42, pp. 17 894–17 898, 2010.
[20] A. Chaudhary, M. McShane, and R. Srivastava, “Glucose response of
dissolved-core alginate microspheres: Towards a continuous glucose
biosensor,” Analyst, vol. 135, no. 10, pp. 2620–2628, 2010.
[21] T. Fo¨rster, C. Strohho¨fer, K. Bock, P. Kasak, M. Danko, Z. Kronekova,
T. Nedelcev, I. Krupa, and I. Lacik, “Biosensor for calcium based on
a hydrogel optical waveguide with integrated sensor proteins,” in Proc.
Int. Solid-State Sens., Actuat. Microsyst. Conf. (TRANSDUCERS 2009),
Denver, CO, Jun., pp. 1218–1221.
[22] T. Fo¨rster, “Zwischenmolekulare Energiewanderung und Fluoreszenz,”
Annalen der Physik, vol. 437, no. 1–2, pp. 55–75, 1948.
[23] ICNIRP Guidelines, “Guidelines for limiting exposure to time-varying
electric, magnetic, and electromagnetic fields (up to 300 GHz),” Health
Phys., vol. 74, pp. 494-522, Apr. 1998.
[24] Wireless medium access control (MAC) and physical layer (PHY) specifi-
cation for low-rate wireless personal area networks (LR-WPANs),” IEEE
Standard 802.15.4, 2003.
[25] P. Valdastri, S. Rossi, A. Menciassi, V. Lionetti, F. Bernini, F. Recchia,
and P. Dario, “An implantable zigbee ready telemetric platform for in vivo
monitoring of physiological parameters,” Sens. Actuat., A: Phys., vol. 142,
no. 1, pp. 369–378, 2008.
[26] P. Valdastri, A. Menciassi, and P. Dario, “Transmission power require-
ments for novel zigbee implants in the gastrointestinal tract,” IEEE Trans.
Biomed. Eng., vol. 55, no. 6, pp. 1705–1710, Jun. 2008.
[27] ZigBee Alliance. (2009). The ZigBee alliance website [Online]. Available:
http://www.zigbee.org
[28] N. R. Subramanian and N. Kirkeby, Anaren 0404 balun optimized for
Texas Instruments CC2430 Transceiver, Aug. 31, 2007.
[29] K. Ye and J. S. Schultz, “Genetic engineering of an allosterically based
glucose indicator protein for continuous glucose monitoring by fluores-
cence resonance energy transfer,” Anal. Chem., vol. 75, pp. 3451–3459,
2003.
[30] W. Frommer, M. Davidson, and R. Campbell, “Genetically encoded
biosensors based on engineered fluorescent proteins,” Chem. Soc. Rev.,
vol. 38, pp. 2833–2841, 2009.
Pietro Valdastri (M’05) received the Master’s
(Hons.) degree in electronic engineering from the
University of Pisa, Italy, in 2002, and the Ph.D. de-
gree in bioengineering from the Scuola Superiore
Sant’Anna, Pisa, Italy, in 2006.
He is working on several European projects re-
garding the development of minimally invasive and
wireless biomedical devices. He is currently an Assis-
tant Professor of Biomedical Robotics at the Scuola
Superiore Sant’Anna, where he is involved in re-
search on implantable robotic systems and active cap-
sule endoscopy.
Ekawahyu Susilo received the Bachelor’s degree in
Electrical Engineering from Universitas Surabaya,
Indonesia, in July 2000, and the Ph.D. degree in bio-
engineering from the Scuola Superiore Sant’Anna,
Pisa, Italy, in 2009. His major is in wireless real time
embedded system design.
He is currently with Decawave, Ltd., Dublin Com-
pany, Fingal, Ireland, developing PHY/MAC and net-
work layer of the new extension of IEEE 802.15.4a
device using ultra wide band for better precision in
3-D real time location system.
1854 IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 58, NO. 6, JUNE 2011
Thilo Fo¨rster received the Master’s degree in physics
from the Technical University of Munich, Munich,
Germany, in 2005, and the Ph.D. degree in electrical
engineering from the Technical University of Berlin,
Berlin, Germany, in 2010.
From 2006 to 2010, he was with the Fraun-
hofer Research Institution for Modular Solid State
Technologies Fraunhofer-Einrichtung fu¨r Modulare
Festko¨rper-Technologien (EMFT), Munich, on sev-
eral national and international research projects in
the biomedical field regarding the development of
biosensor systems. Since 2011, he has been a Technology Scout at Ascenion
GmbH an intellectual property asset management company.
Christof Strohho¨fer received the Master’s degree in physics (Diplom-Physiker)
from the University of Karlsruhe, Germany, in 1997, and the Ph.D. degree in
physics from Utrecht University, The Netherlands, for work performed at the
FOM-Institute for Atomic and Molecular Physics, in 2001.
In 2002, he joined Fraunhofer IZM, Munich, which became the independent
Fraunhofer research institution EMFT in July 2010. In his capacity as a Senior
Scientist and Project Manager, he was involved in national and European projects
in the field of biosensors, biosensor fabrication technology as well as polymer
technology. In May 2011, he joined BBraun Avitum AG as a Manager of New
Technologies.
Arianna Menciassi (M’00) received the Master’s de-
gree in physics from the University of Pisa, Pisa, Italy,
in 1995, and the Ph.D. degree in Biomedical Engi-
neering from the Scuola Superiore Sant’Anna, Pisa,
in 1999.
She is currently an Associate Professor of Biomed-
ical Robotics at the Scuola Superiore Sant’Anna. Her
current research interests include biomedical micro-
robotics for the development of innovative devices
for surgery, therapy, and diagnostics.
Dr. Menciassi serves as a Co-Chair in the IEEE-
RAS Technical Committee of Surgical Robotics and in the Editorial Board on
the IEEE/ASME TRANSACTIONS ON MECHATRONICS.
Paolo Dario (F’02) received the Master’s degree in
Mechanical Engineering from the University of Pisa,
Pisa, Italy, in 1977.
He is a Professor of Biomedical Robotics at the
Scuola Superiore Sant’Anna, Pisa, where he super-
vises a team of about 150 young researchers. He
is the author of more than 160 ISI journal papers,
many international patents, and several book chapters
on medical robotics. His main research interests in-
clude biorobotics, including mechatronic and robotic
systems for rehabilitation, prosthetics, surgery, and
microendoscopy.
Prof. Dario is a recipient of the Joseph Engelberger Award as a Pioneer of
Biomedical Robotics.