Abstract—There is often a strong correlation between

elevated levels of intraocular pressure (IOP) and glaucoma;

however, the underlying mechanisms that lead to blindness are

not well understood. The key may lie in the study of genetic

factors which determine IOP and lead to glaucoma-related

blindness. Mice are typically used for genetic research due to

their short generation time and accelerated lifespan,

manageability, the availability of established and pure lines,

and the ability to manipulate the genome. Post genetic

manipulation, IOP monitoring at regular intervals is needed

and for large scale testing, on the order of thousands of mice, it

is crucial to have at least a partially automated data collection

scheme. This work presents a fully wireless system on a chip

that measures 300 µm in its widest dimension, has a wireless

microwave-based data and power link, and is capable of

relaying digitized pressure recordings to a nearby base-station.

I. INTRODUCTION

LAUCOMA is a condition that stems from the buildup of

intraocular pressure (IOP) that causes the death of

retinal ganglion cells and degeneration of the optic nerve

which eventually leads to blindness. This disease is the

second leading cause of blindness, affects over 60 million

worldwide, and has caused bilateral blindness in over 8.4

million people [1].

Work done at Jackson Laboratory is exploring the

mechanisms behind cell death due to elevated levels of IOP

in glaucomatous subjects [2]. The mouse is the typical

animal model for genetic studies due to their quick

generation times, established lines, fully known genome, and

feasibility of genomic manipulations [3]. Furthermore, their

small size, relatively low cost, and ease of manageability

facilitates comprehensive studies requiring large matrices of

parametric variations. The genetic studies begin with the

breeding of a large matrix of genetically varied mice whose

genomes are altered through the addition of transgenes or the

variation of endogenous genes. IOP is then monitored

Manuscript received April 23, 2010. Funds are provided by a Howard

Hughes Medical Institute Collaborative Innovation Award.

E. Y. Chow was with Purdue University, West Lafayette, IN 47907

USA. He is now with Cyberonics, Inc., Houston, TX 77058 USA (phone:

719-640-5171; fax: 281-283-5522; e-mail: eric.chow@cyberonics.com).

S. W.M. John is an investigator of the Howard Hughes Medical Institute

and professor at The Jackson Laboratory, Bar Harbor, ME 04609 USA (e-

mail: simon.john@jax.org).

W. N. deVries is with Jackson Laboratory, Bar Harbor, ME 04609 USA.

D. Ha, T. Lin, and W. J. Chappell are with the Electrical and Computer

Engineering Department, Purdue University, West Lafayette, IN 47907

USA (e-mail:dha@purdue.edu, lin128@purdue.edu, chappell@purdue.edu).

P. P. Irazoqui is with the Biomedical Engineering Department, Purdue

University, West Lafayette, IN 47907 USA (e-mail: irazoqui@purdue.edu).

periodically for each mouse. Currently, the method for IOP

monitoring involves a manual measurement using a micro-

needle system [4]. The time and resources required for these

periodic tests are significant when performed on the

thousands of mice used in these genetic studies.

An automated approach would greatly facilitate mouse

IOP monitoring in glaucoma genetic studies and could

produce more accurate, thorough, and less costly results.

This work proposes an ultra-miniature ocular device, as

shown in Fig. 1, which is small enough for implantation

inside a mouse eye, fully wireless in terms of powering and

data transfer, and can perform periodic measurements of

IOP. The device consists of a 300 µm x 300 µm x 50 µm

application specific integrated circuit (ASIC), a similarly

sized microelectromechanical systems (MEMS) pressure

sensor, and a 1 – 2 mm diameter loop antenna. The ASIC,

implemented using the Texas Instruments 130 nm CMOS

process, derives its power from a microwave-based far-field

source, captures 0.5 mmHg resolution measurements from

the MEMS sensor, amplifies and digitizes the data into an 8

bit word, and wirelessly transmits the data to the external

base-station.

II. DESIGN AND METHODS

To facilitate these glaucoma genetic studies, it is desirable

to design a pressure monitoring method that is capable of

capturing periodic IOP measurements from 7 to 8 mice per

cage and robust enough to function while the animals are

Sub-Cubic Millimeter Intraocular Pressure Monitoring Implant to

Enable Genetic Studies on Pressure-Induced Neurodegeneration Eric Y. Chow, Dohyuk Ha, Tse-Yu Lin, Wilhelmine N. deVries, Simon W.M. John, William J.

Chappell, and Pedro P. Irazoqui, Members, IEEE

G

Fig. 1. Conceptual illustration of wireless intraocular pressure

monitor implanted in the eye of a mouse

32nd Annual International Conference of the IEEE EMBSBuenos Aires, Argentina, August 31 - September 4, 2010

978-1-4244-4124-2/10/$25.00 ©2010 IEEE 6429

mobile and moving about their enclosure. One IOP sensing

method is to implant a device that exhibits a change based

on pressure and whose change is detected and quantified by

an external device. Several works have achieved this passive

sensing approach through monitoring the resonance or

impedance of an implanted inductor-capacitor (LC) network

whose capacitance or inductance varies in response to

pressure [5]. Unfortunately, these passive measurement tools

all require a nearby (within a few mm) external device and

are not robust, work only in a very stable and controlled

environment, and cannot realistically deal with multiple

implanted animals [6]. This work presents an active sensing

method that captures microwave energy from a far-field

source, processes/amplifies/digitizes pressure measurements,

and wirelessly transmits the data to an external receiver.

A. Sensor Interface

A custom designed liquid-crystal-polymer (LCP) based

MEMS capacitive sensor is used to convert pressure

readings into capacitance value [7]. This device is connected

to the first stage of the sensor interface consisting of two

matched 1 µA current sources, shown in Fig. 2, which feed

the MEMS pressure sensor and a reference capacitor. Upon

the start of a measurement cycle, the two current sources

charge up the capacitors until the reference capacitor reaches

a voltage set by a Schmitt trigger based threshold detector.

Once the threshold detector is triggered, a control signal is

set to turn the current sources off, after which the voltages

on the two capacitors will remain steady. Since the charging

rate of the capacitors is proportional to its size, the voltage

on the MEMS capacitor, at the time the reference capacitor

voltage crosses the threshold, will depend on its capacitance

and therefore the pressure it is sensing. Now the reference

capacitor is set to the capacitance of the MEMS sensor at

atmospheric pressure. Thus, the difference between two

voltages at the top plate of the capacitors will correspond to

the pressure above atmospheric pressure that is sensed.

These voltages, or signals, are then fed through a

transmission gate, which allows the signal to pass once the

current sources have stopped, and then through a custom

sized source follower to remove the dc offset. This signal is

then amplified through a difference amplifier whose gain is

set to 51.25 which amplifies the voltage range,

corresponding to a pressure range of 50 mmHg, to the full

output swing of the amplifier, 0 V to 900 mV. Following

amplification, the signal is fed into a single-slope integrator-

based analog-to-digital convertor where it is digitized into an

8-bit word.

B. Wireless Powering, Charge Storage, and Regulation

An active implant has significant advantages in terms of

on-board signal processing, robustness, and overall

functionality; however, it does require a power supply. For

the target application, a power storage unit, such as a battery,

is not an option due to the size constraints, so wireless

powering schemes are explored. Inductive powering is a

commonly used technique; however, since it requires perfect

alignment and close proximity, it is only useful in stationary

situations [8]. To accommodate the situation of several

mobile mice in an enclosure, microwave-based far-field

powering is used to achieve a method that delivers sufficient

power over a distance of several 10s of centimeters and is

relatively orientation independent [9].

The radio-frequency (RF) energy is radiated from an

external source and picked up by the implant’s nitinol-based

loop antenna. The high frequency ac wave is then rectified

into a dc supply using a Cockcroft-Walton-based ac-to-dc

converter topology [10]. Schottky diodes are used in the

circuit due to their high frequency performance, which is a

result of their metal-on-semiconductor nature. Furthermore,

their typically low turn-on voltages enable to rectifier to

achieve much higher efficiencies than with other diode

types. This particular Cockcroft-Walton-based multiplier-

type topology is chosen because even for low input power

levels, the rectifier is still able to achieve a high enough

voltage to sufficiently bias the CMOS node. This ability

greatly enhances the robustness of the wireless powering

link; however, this operation is only achieved if the system

that this rectifier feeds is designed to consume relatively low

levels of power.

To minimize the average power consumption of the

system, a sleep-wake cycle is utilized. The system is

designed to take a measurement and transmit the data once

every 0.5 seconds; however, it only requires a couple 100 µs

for a measurement cycle, which includes capturing the

MEMS pressure sensor reading, amplifying, digitizing the

data, and wirelessly transmitting out the serial bit stream.

Therefore, the system is put into “sleep” mode between the

measurement cycles and a nanowatt timer is used to track

the slee-time and “wake” the system back up. The timer is is

achieved by first using the drain-to-soruce leakage current of

a transistor to source an on-chip capacitor. A threshold

detector is placed on the capacitor node such that when the

capacitor is charged to a specified level, the timer triggers

the wake-up of the system. The transistor and capacitor are

both sized such that the time it takes to charge the capacitor

Fig. 2. Simplified block diagram of sensor interface

6430

up to the threshold is designed to be about 0.5 seconds.

A regulator is needed to provide a stable voltage supply to

the rest of the circuitry to ensure reliable operation. The

regulator’s primary goal is to maintain a constant voltage

independent of the varying current consumption of the

system. In this application, the regulator also has to tolerate

large variations in its input source since it is fed by the

rectifier whose dc output varies as a function of the RF

energy it sees at its input. In this work, a negative feedback

low-dropout regulator topology is used with a self-

compensating voltage reference to enable stable supply

regulation over a wide dc input swing.

C. Wireless Transmitter and Antenna

An np complementary cross-coupled LC voltage-

controlled-oscillator (VCO) is used to modulate the digital

data onto a 2.45 GHz RF carrier for wireless transmission.

The symmetry and bandpass characteristic of this LC

negative resistance oscillator gives it the lowest phase-noise

for a given power consumption [11]. A fully symmetric

inductor is designed to take advantage of the architecture’s

symmetry and optimize the overall phase noise. The signal is

modulated via on-off-keying (OOK), which is chosen for its

minimal complexity and low power consumption. The

digital data is fed into a switch which controls the bias

current supply for the oscillator switching it between on and

off states. The start time for the oscillator is significantly

shorter than a bit-duration and thus will have a negligible

effect on the data transfer and bit-error-rate. The transmitter

output is then fed to the nitinol-based loop antenna.

A loop antenna is chosen for this implantable device

because its closed form feature reduces the risk of damaging

the surrounding tissue and it also holds a relatively fixed

position inside the eye. To conform to the available space of

the mouse anterior chamber, the maximum diameter of the

loop antenna should not exceed 2 mm. To minimize damage,

the surgical incision length must be less than 1 mm,

therefore, to take full advantage of the available area within

the anterior chamber, a material with a self-expandable

characteristic is desired for the antenna. Our antenna utilizes

nitinol, a material comprised of half nickel and half titanium,

that has the property of shape memory and reverts back to its

original shape after deformation. Therefore after a surgeon

collapses the antenna for the insertion process, the loop will

expand back to its original shape after it is released in the

anterior chamber.

III. RESULTS AND DISCUSSION

An assembled conceptual prototype is shown in Fig. 3,

which consists of a 300 x 300 x 300 µm3 silicon substrate,

representing the size of the final ASIC bonded with the

MEMS sensor, integrated with a 2 mm diameter nitinol loop

antenna.

The fabricated MEMS sensor has a sensitivity of 0.75

fF/mmHg and is currently 500 × 500 × 160 µm3; however,

work is being done to maintain this sensitivity and reduce

this length and width down to under 300 µm to meet the

surgical requirements [7]. At atmospheric pressure (760

mmHg), the sensor has a capacitance of 1.1765 pF, and at 60

mmHg above atmosphere, the capacitance is about 1.214 pF.

Therefore, the sensitivity of the MEMS device is about 0.75

fF/mmHg.

The MEMS capacitive sensor is integrated with the ASIC,

layout shown in Fig. 4, using a flip-chip-like bonding

method with anisotropic conductive adhesive [12]. The

sensor interface circuit consists of 1 µA current sources

which charge up the MEMS and reference capacitors to the

450 mV threshold in approximately 530 ns, depending on

the actual MEMS capacitance value. The charging time

difference is then converted to analog voltages and for a 0.75

fF change, the voltage difference is about 287.05 µV. These

voltages are then fed through a difference amplifier to

improve the signal integrity for the following digitization.

The pressure range of interest is 0 – 60 mmHg above

atmosphere and analytical calculations are used to determine

gain of 51.25 to utilize the full rail-to-rail output swing of 0

V to 900 mV. Post-amplification, the voltage value

corresponding to a 0.75 fF change is 14.711 mV. This

amplified signal is then sent to an 8-bit analog-to-digital

converter (ADC). The final result is an 8 bit digital word

representing pressures in the range of 0 – 60 mmHg above

atmosphere where each step size corresponds to a resolution

of 0.25 mmHg.

The digital data is then sent to the wireless transmitter for

modulation on the 2.45 GHz carrier frequency. The parallel

bits are first serialized through a switched multiplexer that

steps through each bit at a rate of 100 kbps. For error

correction and synchronization, 4 start and 4 end bits are

concatenated onto the 8-bit word. This 16-bit stream is then

fed into the 2.45 GHz transmitter. The phase noise of the

transmitter is simulated as -81.73 dBc/Hz at 100 KHz offset

and -102.52 dBc/Hz at 1 MHz offset. The simulated output

power is about 5.919 µW (-22 dBm). The transmitter

consumes a peak power of 585.1 µW; however, since OOK

Fig. 3. Conceptual prototype consisting of a 300 x 300 x 300 µm3

silicon substrate integrated with a 2 mm diameter nitinol loop antenna

6431

modulation is used, the average power consumption is half

the peak value. The transmitter startup times between the on-

off switching are less than 100 ns which are low enough to

cause negligible interference with the 100 kbps data transfer.

This 292.55 µW average transmitter power is only

consumed for the 160 µS of time it takes to transmit out the

16-bits of data at 100 kbps.

A summary of the specifications including timings and

power consumptions are tabulated in Table I.

IV. CONCLUSION

There is a need for an automated testing method to

facilitate large-scale genetic testing of mice to determine the

root causes of glaucoma. In this work, we proposed a

miniature device which surgeons can implant into the mouse

eye to continuously monitor intraocular pressure. This active

device is remotely powered, amplifies and digitizes pressure

readings, wirelessly transmits the data to an external base

station. The ASIC captures pressure measurements with a

resolution down to 0.25 mmHg in a range of 0 to 60 mmHg

above atmosphere, digitizes the data into an 8-bit word,

periodically samples twice a second, and consumes less than

136 nW of average power. The far-field microwave-based

wireless techniques used to power and retrieve data provide

this implant with the robustness necessary to operate in a

cage with multiple mobile mice.

ACKNOWLEDGMENT

The authors would like to thank Texas Instruments for

their design help and fabrication services.

REFERENCES

[1] H. A. Quigley and A. T. Broman, "The number of people with

glaucoma worldwide in 2010 and 2020." vol. 90, 2006, pp. 262-

267.

[2] S. W. John, R. S. Smith, O. V. Savinova, N. L. Hawes, B. Chang,

D. Turnbull, M. Davisson, T. H. Roderick, and J. R. Heckenlively,

"Essential iris atrophy, pigment dispersion, and glaucoma in

DBA/2J mice," Invest Ophthalmol Vis Sci, vol. 39, pp. 951-962,

1998.

[3] K. Paigen, "A miracle enough: the power of mice," Nat Med, vol. 1,

pp. 215-220, 1995.

[4] S. W. John, J. R. Hagaman, T. E. MacTaggart, L. Peng, and O.

Smithes, "Intraocular pressure in inbred mouse strains." vol. 38,

1997, pp. 249-253.

[5] C. Po-Jui, D. C. Rodger, S. Saati, M. S. Humayun, and T. Yu-

Chong, "Microfabricated Implantable Parylene-Based Wireless

Passive Intraocular Pressure Sensors," Microelectromechanical

Systems, Journal of, vol. 17, pp. 1342-1351, 2008.

[6] U. Schnakenberg, P. Walter, G. vom Bögel, C. Krüger, H. C.

Lüdtke-Handjery, H. A. Richter, W. Specht, P. Ruokonen, and W.

Mokwa, "Initial investigations on systems for measuring

intraocular pressure," Sensors and Actuators A: Physical, vol. 85,

pp. 287-291, 2000.

[7] D. Ha, K. Musick, P. P. Irazoqui, and W. J. Chappell, "Parylene on

LCP Flexible Capacitive Pressure Sensor for Intraocular Pressure

Sensor," (Unpublished Work).

[8] N. Donaldson and T. Perkins, "Analysis of resonant coupled coils

in the design of radio frequency transcutaneous links," Medical and

Biological Engineering and Computing, vol. 21, pp. 612-627, 1983.

[9] W. C. Brown, "The History of Power Transmission by Radio

Waves," Microwave Theory and Techniques, IEEE Transactions

on, vol. 32, pp. 1230-1242, 1984.

[10] E. Edgar and P. Lorrain, "The Cockcroft-Walton Voltage

Multiplying Circuit," Review of Scientific Instruments, vol. 24, p. 6,

1953.

[11] A. Hajimiri and T. H. Lee, "Design issues in CMOS differential LC

oscillators," IEEE Journal of Solid-State Circuits, vol. 34, 1999.

[12] E. Y. Chow, Y. Chin-Lung, A. Chlebowski, M. Sungwook, W. J.

Chappell, and P. P. Irazoqui, "Implantable Wireless Telemetry

Boards for In Vivo Transocular Transmission," Microwave Theory

and Techniques, IEEE Transactions on, vol. 56, pp. 3200-3208,

2008.

Fig. 4. Layout of 300 µm x 300 µm ASIC with sensor interface, voltage

regulator (Vreg), timer (Tim), wireless transmitter (TX), and RF

rectifier

TABLE I

ASIC SPECIFICATIONS

Specification Simulated Value Units

Sensor interface power

consumption

32.4 µW

Single-slope 8-bit ADC power

consumption

27 µW

Sensor interface

measurement time

500 ns

Digitization time 0 – 256 µs

Sleep time 0.5 s

Sampling rate 2 Hz

Average sleep power

consumption

26.78 nW

TX frequency 2.45 GHz

TX data-rate 100 kbps

TX output power -22 dBm

TX phase noise -81.728 @ 100 KHz

-102.524 @ 1 MHz

dBc/Hz

TX startup time < 100 ns

TX Power Consumption 585.1 (peak)

292.55 (average)

µW

Average total power

consumption

120.401 to 135.710 nW

6432