WIRELESS INTRAOCULAR PRESSURE SENSOR USING STRETCHABLE

VARIABLE INDUCTOR
Mohammad Hossein Mazaheri Kouhani1, Arthur Weber2, and Wen Li1
1
Electrical and Computer Engineering Department, Michigan State University, East Lansing, USA
2
Physiology Department and Neuroscience Program, Michigan State University, East Lansing, USA

ABSTRACT precision surgery to anchor the device on the iris, which

This paper reports the design, fabrication, and requires highly skilled operators. Also, the fabrication
characterization of a wireless, flexible, passive pressure process of its spiral tube is complex and expensive.
sensor that enables real-time intraocular pressure (IOP) Furthermore, its optical external reading mechanism
monitoring. It employs a novel approach that measures the requires a large module that interferes with the vision of the
strain on the sclera of eye-ball using integrated planar patient, therefore, making it unsuitable for wearable, long-
stretchable variable inductor. Major reduction in fabrication term measurements.
complexity is achieved in contrast to the state-of-the-art To date, most devices that are excited using
solutions which are based on variable capacitors. Our device electromagnetic (EM) coupling employ variable capacitors
potentially improves the stability of long-term measurement [4-6] as a pressure sensitive element. Variable capacitors
by eliminating baseline drift due to leaks within sealed need a pressurized reference chamber that hardly sustain
capacitive chambers. Implantation in the sclera reduces their baseline pressure over-time due to packaging
energy absorption by tissue, thereby improving inductive imperfections. This can cause significant signal drift mainly
coupling efficiency between the implanted sensor and the to their leaking. For instance, the Implantdata Eyemate
external coil. This will significantly facilitate the suffers from 3.47 mmHg of drift on average [2].
commercialization of minimally invasive, in-vivo IOP
sensing devices for Glaucoma patients. Devices are Our solution to the above challenge is a passive strain
packaged using Parylene-C, which ensures the mechanical gauge that serves as both a pressure sensitive element and a
flexibility and biocompatibility of the sensor. wireless interface. In our design, the strain gauge is
configured as a variable inductor instead of a variable
INTRODUCTION capacitor. The variable inductor will be implanted in the
sclera of the Glaucoma eye. The EM response of the sensory
Emergence of advanced bioinformatics calls for new implant change enabling us to correlate those changes with
technologies monitoring circadian patterns on a 24-h fashion pressure fluctuations.
for each individual patient reliably, accurately, and mobile.
Glaucoma is the second leading cause of blindness in the The method reported here differs from most previous
world [1]. It is an asymptomatic, progressive, and works. In contrast to variable capacitive sensors and the
irreversible disease that is usually associated with elevated variable inductor sensor [7] that is based on a movable
IOP. Accurate, real-time, and portable monitoring of IOP ferrite magnet, our variable inductor does not employ any
can be of great use while coping with Glaucoma. sealed chamber, which prevents the inaccuracy of
measurement due to baseline drift over time. It is simple,
Most previous implantable devices, reported over the stable, and cheap to fabricate, which makes it a great
past two decades, can be categorized into three groups in candidate for chornic implantation in Glaucoma eyes. If it is
terms of their energy transferring mechanisms: active, embedded in a contact lens, the inductive strain gauge
passive, and radio-frequency-powered devices. sensor has the potential to become a non-invasive device.
Active devices usually employ application-specific-
integrated-circuits (ASIC) that can store, process, and
transmit data. However, they make the overall implant
device large, heavy and inflexible. Also, active systems
require energy either from an integrated battery or a power-
receiving coil, which both add to the weight and size of the
final implant. Implantdata Eyemate is a relatively successful
example of its kind offering a wireless intraocular
transducer (WIT) [2] and yet suffering from disadvantages
discussed above.
Among fully-passive implants, optically read Bourdon
tube is a well-designed representing example of its category
[3]. It requires no external power delivery since the sensing Figure 1: The conceptual diagram of the device and an
implant is completely passive. However, it requires high equivalent circuit model.

978-1-5090-5078-9/17/$31.00 ©2017 IEEE 557 MEMS 2017, Las Vegas, NV, USA, January 22-26, 2017
DESIGN AND FABRICATION The fabrication process is schematically shown in Fig.
3. The fabrication of the sensor prototype is carried out
A concept diagram of the sensor is shown in Fig. 1 using two masks. First, a 3-inch glass wafer is coated with 5
along with its equivalent circuit. The wireless passive µm Parylene-C, a 20 nm layer of titanium and 700 nm of
sensing used in this work is based on the mutual inductance copper (Fig. 3A and 3B) respectively. Parylene-C is
between two coils: the reader and the sensor. The reader deposited in a chemical vapor deposition (CVD) system
(transmitter) coil is connected to an impedance analyzer. (PDS2010 Parylene Coater, Specialty Coating System).
The diameter of the reader coil is comparable to the Metal is evaporated using a thermal evaporator. After that,
diameter of the sensor coil which is coupled with. The mask one is used to pattern the geometric shape of the
sensor can be modeled as a planar and circular LC passive device metal core using ultra-violet (UV) photolithography
resonator that has a constant capacitor and a stretchable and then wet etching (Fig. 3C and 3D). At this point, surface
variable inductor. The self-inductance of the inductor, the mounted (SMD) capacitor is soldered to the contacts of the
parasitic capacitance between the segments, and the Q- inductor using silver epoxy. After the full wafer is coated
factor are all changeable by the expansion of the sensor with another 5 µm Parylene-C, aluminum is deposited on
diameter. This phenomenon is employed to measure the top of Parylene-C and patterned using the second mask to
strain in the eye tissues induced by IOP elevation. The form an aluminum mask to protect the sealed metal core
change of self-inductance and parasitic capacitance result in during plasma dry etching of Parylene-C (Fig. 3 E-G).
the change of resonance frequency of the LC loop. Eventually, devices are released from the glass substrate
Therefore, we are able to read the pressure variance by before implantation (Fig. 3H). Fig. 4 show the photo image
means of frequency drift. of a fabricated device and the microscope image of the “S-
Finite element simulation is performed in the Magnetic shape” stretchable wire.
and Electric Fields (mef) module of COMSOL Multiphysics
(Ver. 5.2 COMSOL, Inc.). The shift of the mutual
inductance between the reader and sensory coils is studied
in response to the diameter expansion of the sensory coil
(Fig. 2). The geometries of the reader and sensory coils are
provided in Table I. Simulation results show a decreasing
mutual inductance as the diameter of the sensory coil
expands, suggesting an upshift of the resonance frequency
based on the resonance frequency equation shown in Fig. 1.
Table I details the inductance decrease by the increase of
diameter of the sensory coil.

Figure 3: Fabrication process flow of the sensor. A.

Parylene-C deposition on full wafer. B. Metal
deposition. C. First UV Photolithography. D. Wet
etching metal. E. Full wafer Parylene-C Deposition.
F. Forming aluminum mask using the 2nd mask. G.
Plasma dry etching Parylene-C. H. Releasing device.

Figure 2: Simulation in COMSOL Multiphysics. The

Inductance variation of reading-coil is measured
with respect to diameter variation of the sensory coil.

Table I: Simulation Results for Inductance variation

Reading Coil (mm) 15.5 15.5 15.5 Figure 4: A photo of a fabricated stretchable sensor
Sensory Coil Diameter (mm) 15.5 19.5 23.5 and a close-up microscope image showing the “S-
Inductance (pH) 43.34 43.20 42.89 shape” wire geometry.

558
CHARACTERIZATION AND RESULTS higher than the eye tissue. As a result, the strain gauge
sensor expands way larger, faster, and easier in response to
Device is characterized under two circumstances. One lower pressure elevations comparing with the eye.
uses a latex balloon as a model for eye. The pressure of the Therefore, the pressure sensitivity measured using the
balloon is controlled by adjusting the air pressure inside of it balloon is higher than the one characterized later using a pig
using a syringe pump. Fig. 5 shows the experimental setup eye. A pressure sensitivity of 2667 KHz/mmHg is observed
for device characterization using the balloon. The detected in case of balloon. The Q-factor of the sensory coil is
resonant frequency versus the separation between the coils measured as the sharpness of the resonant peak. As shown
and the chamber pressure are shown in Fig. 6 and Fig. 7, in Fig. 6, the Q-factor drops as the separation of the two
respectively. The elasticity the balloon membrane is much coils increases, mainly due to the weakening of the
electromagnetic coupling between the two coils.

Figure 5: Setup for testing the stretchable inductor

on a balloon membrane

Figure 8: A. Setup for testing the stretchable inductor

on a pig eye. B. Top-view of the pig eye with a device
implanted around the cornea
Figure 6: Results showing the dependence of signal
strength on the distance between the reading coil and
sensor.

Figure 7: Resonance frequency shift due to pressure Figure 9: Resonance frequency drift due to pressure
elevation. elevation.

559
 Table II: Summary of device design and pressure CONCLUSION
sensing performance. In this work, we design and implement a variable
Properties Device on Device inductive sensor that can non-linearly respond to pressure
balloon on pig variation by changing its geometry due to the strain caused
eye by the pressure elevation in the eye. We demonstrate that an
Coil diameter (cm) 1.4 1.4 upshift of resonant frequency occurs in the sensing LC tank
Overall thickness (µm) 10.8 10.1 when the pressure is elevated. We analytically speculate and
Capacitance (pF) 2.2 1.8 verify with simulation results that this frequency shift is
Inductance (nH) 50 50 caused by the diameter increase of the sensory coil.
Parasitic resistance (Ω) 50 25 However, we do not yet exactly know the detailed equation
Theoretical resonant freq. 411 551 ruling this relationship. Indeed, the parasitic capacitance
(MHz) between the “S-shape” winding segments of the sensory coil
also play an important role, along with the variations of the
Pressure sensitivity 2667 57
overall Q-factor and mutual inductance. Further work is in
(kHz/mmHg)
progress to study various other designs and their EM
response to precisely explain the fundamentals behind this
In order to become one step closer to in-vivo studies, relationship and optimize the sensing mechanism.
we report a characterization experiment on a pig eye. Fig. 8
illustrates the experimental setup for device testing and Fig.
ACKNOWLEDGEMENTS
9 shows the frequency shift versus the pressure elevation. In
this case, a small conjunctival flap is cut around the cornea. The authors would like to acknowledge the financial
The inductor wire is embedded under the conjunctival flap support from the Michigan State University Foundation
of the eye. The eye is pressurized with an initial IOP of under the Strategic Partnership Grant.
around 15 mmHg. A small needle is inserted to the bottom
of the eye in order to inject saline to the interior chamber of REFERENCES
the eye through a syring pump. This enables us to mimic the
pressure elevation that occurs in the human eye. Since the [1] S. Kingman, “Glaucoma is second leading cause of
outflow of the liquid is limited, we are able to control he blindness globally” Bulletin World Health Organization,
pressure by only increasing the inflow of saline. This vol. 82 pp. 887–8.
experiment shows a pressure sensitivity of 57 KHz/mmHg [2] S. Melki, et al., "An implantable intraocular pressure
based on the frequency shift correlated with the eye pressure transducer: initial safety outcomes.” JAMA ophthalmology,
in Fig. 9. 2014, vol. 132(10), pp. 1221-1225.
Table II summarizes the performance of the sensor [3] P. J. Chen et al., “Implantable micromechanical
characterized under different conditions. Results from both parylene-based pressure sensors for unpowered intraocular
balloon and pig eye are in agreement with each other, in a pressure sensing.” Journal of Micromechanics and
way that, they both present an upshift of resonant frequency Microengineering, 2007, vol. 17(10), pp. 1931.
when the pressure elevates (Fig. 7 and Fig. 9). The elasticity [4] K. C. Katuri et al. “Intraocular pressure monitoring
of the balloon membrane is significantly higher than the eye sensors,” Sensors, IEEE Journal of, 2008, vol. 8, no.1, pp.
tissue, therefore, exhibiting a much higher sensitivity. The 12-19.
Q-factor of the sensor drops when moving from the balloon [5] P. J. Chen et al., “Microfabricated implantable parylene-
to the pig eye due to the electromagnetic absorption of the based wireless passive intraocular pressure sensors,”
tissue. The change in Q-factor may act as an additional Journal of Microelectromechanical Systems, 2008, vol. 17,
indicator signal correlated with pressure variation in our no. 6, pp. 1342-1351.
further studies.
[6] B. Crum et al., “A Flexible Parylene-Based Intraocular
When the sensor expands, the upshift of the resonant Pressure Sensor,” Proceeding of the 9th IEEE International
frequency is speculated to be caused by two main factors. Conference on Nano/Micro Engineered and Molecular
The mutual inductance between the reader and sensory coils Systems, 2014.
decreases as indicated in Table I. This might be due to the [7] B. Kang, et al., “A wireless intraocular pressure sensor
variation of the overall magnetic flux density through the with variable inductance using a ferrite material.” Journal of
coil loops, as a result of the change in the air core area. Semiconductor Technology and Science, vol. 13(4), pp. 355-
Another important factor is the parasitic capacitor between 360.
adjacent winding segments, which decreases as the
separation of the segments increases. The ruling analytical
equation has not yet completely understood and will be CONTACT
investigated in the future work. *M. H. M. Kouhani, tel: +1-517-9748327;
[email protected]

560