Metglas-Elgiloy Bi-Layer, Stent Cell Resonators for Wireless Monitoring of Viscosity and Mass Loading

Metglas-Elgiloy Bi-Layer, Stent Cell Resonators for Wireless Monitoring of Viscosity and Mass Loading
Home
Search
Collections
Journals
About
Contact us
My IOPscience
Metglas–Elgiloy bi-layer, stent cell resonators for wireless monitoring of viscosity and mass
loading
This article has been downloaded from IOPscience. Please scroll down to see the full text article.
2013 J. Micromech. Microeng. 23 025010
(http://iopscience.iop.org/0960-1317/23/2/025010)
View the table of contents for this issue, or go to the journal homepage for more
Download details:
IP Address: 68.60.46.21
The article was downloaded on 22/12/2012 at 02:23
Please note that terms and conditions apply.
IOP PUBLISHING
JOURNAL OF MICROMECHANICS AND MICROENGINEERING
doi:10.1088/0960-1317/23/2/025010
J. Micromech. Microeng. 23 (2013) 025010 (9pp)
Metglas–Elgiloy bi-layer, stent cell
resonators for wireless monitoring of
viscosity and mass loading
Anupam Viswanath 1 , Scott R Green 1 , Jürgen Kosel 2
and Yogesh B Gianchandani 1
1
Department of Electrical Engineering and Computer Science and Center for Wireless Integrated
MicroSensing and Systems (WIMS2), University of Michigan, Ann Arbor, MI, USA
2
Computer, Electrical and Mathematical Sciences and Engineering Division, King Abdullah University
of Science and Technology, 4700 KAUST, Thuwal 23955, Saudi Arabia
E-mail: [email protected]
Received 14 October 2012, in final form 18 November 2012
Published 21 December 2012
Online at stacks.iop.org/JMM/23/025010
Abstract
This paper presents the design and evaluation of magnetoelastic sensors intended for wireless
monitoring of tissue accumulation in peripheral artery stents. The sensors are fabricated from
28 μm thick foils of magnetoelastic 2826MB MetglasTM , an amorphous Ni–Fe alloy. The
sensor layer consists of a frame and an active resonator portion. The frame consists of 150 μm
wide struts that are patterned in the same wishbone array pattern as a 12 mm × 1.46 mm
Elgiloy stent cell. The active portion is a 10 mm long symmetric leaf shape and is anchored to
the frame at mid length. The active portion nests within the stent cell, with a uniform gap
separating the two. A gold-indium eutectic bonding process is used to bond MetglasTM and
Elgiloy foils, which are subsequently patterned to form bi-layer resonators. The response of
the sensor to viscosity changes and mass loading that precede and accompany artery occlusion
is tested in vitro. The typical sensitivity to viscosity of the fundamental, longitudinal resonant
frequency at 361 kHz is 427 ppm cP−1 over a 1.1–8.6 cP range. The sensitivity to mass loading
is typically between 63000 and 65000 ppm mg−1 with the resonant frequency showing a
reduction of 8.1% for an applied mass that is 15% of the unloaded mass of the sensor. This is
in good agreement with the theoretical response.
(Some figures may appear in colour only in the online journal)
1. Introduction
within the extracranial carotid circulation, this work is directed
at chronic arterial occlusive disease occurring in the legs.
The most common cause of PAD is hardening of the arteries
(atherosclerosis), the gradual buildup of fatty deposits (plaque)
on the walls of the arteries that slow or block blood flow
(figure 1). Plaque buildup also causes the artery walls to stiffen
and become unresponsive to varying levels of blood supply.
Intermittent claudication, defined as pain in the muscles of
the leg with ambulation, is the earliest and the most frequent
presenting symptom in patients with lower extremity PAD.
Although claudication symptoms are typically localized in the
calf or the thigh, ‘rest pain’ is characteristically in the foot [1].
In the late stages of PAD, tissue hypoperfusion progresses to
ischaemic ulceration and gangrene, and major amputation is
Stents are metal mesh tubular structures used to treat narrowed
vessels and ducts in the body that have been constricted as
a result of stenosis. Balloon angioplasty is an established
procedure for the implantation of these stents. Although
angioplasty relieves the symptoms of stenosis, it carries with it
a risk of a reappearance of the narrowing, typically due to the
reaction of the body to the presence of the stent—a condition
known as in-stent restenosis.
Peripheral arterial disease (PAD) comprises those
pathologies which result in obstruction to blood flow in the
arteries, exclusive of the coronary and intracranial vessels.
Although the definition of PAD technically includes problems
0960-1317/13/025010+09$33.00
1
© 2013 IOP Publishing Ltd
Printed in the UK & the USA
J. Micromech. Microeng. 23 (2013) 025010
A Viswanath et al
Figure 2. Conceptual diagram of magnetoelastic monitoring of
peripheral stents located in the posterior tibial artery. External
circuitry drives the external interrogation coils to wirelessly measure
the response of the implanted sensor.
Figure 1. Conceptual diagram showing restenosis occurring in the
posterior tibial artery and the integration of sensor on stent. A
schematic of the MetglasTM /Elgiloy single stent cell resonator used
for sensing is also shown.
between magnetization response to a magnetic field and the
strain of the sensor material. There are at least two important
advantages of this method over purely electromagnetic
methods. Firstly, magnetoelastic materials have high coupling
coefficients and high magnetic permeability, negating the need
for an internal antenna within the stent geometry. This is
valuable in this space-limited application. Secondly, because
the typical resonant frequency of these sensors can be designed
to lie within a range of <400 KHz, the attenuation posed by
bodily tissue when interrogating the sensors is low. Antennas
of sizes similar to the sensors used in this work have working
frequencies in the 0.1–1 GHz range. Bodily tissue has much
higher attenuating properties at those frequencies.
Our previous work demonstrated magnetoelastic sensors
incorporated with stents for sludge accumulation in a biliary
stent [6–8]. Initial designs were tethered to the stents with
anchors, separating sensor and the stent [6]. Although a
reasonable architecture for biliary stents, this separation could
lead to problems such as undesirable blood turbulence and
blood clotting in peripheral stents [9]. A closer integration of
the sensor with the stent to form a bi-layer resonator would
help reduce clotting potential. A sensor design is needed that
permits better conformity to the expanding stent structure,
while still providing adequate sensitivity. An effective bonding
method for sensor-stent integration is also needed.
Magnetoelastic resonant sensing has attracted
considerable interest in the sensor community as the
method can be used to measure a wide variety of
environmental parameters including pressure [10],
humidity [11], temperature [12] and liquid viscosity
[13]. These sensors have the relative advantages of immunity
to noise and changes in the amplitude or orientation of the
interrogation field. Another important advantage of these
sensors in the context of implantable devices is the ability
to be wirelessly interrogated. This paper presents a sensor
integrated with a stent for wirelessly monitoring the restenosis
in a peripheral artery stents. The design, modeling and material
selection for the sensors are described in section 2 followed by
the fabrication approach in section 3 and experimental results
in section 4. The discussions and conclusions are presented in
section 5.
eventually required in more than a third of the patients. In most
severe cases of PAD, surgical interventions such as angioplasty
are necessary for treatment.
Once a stent is implanted, the patient is at risk for
restenosis, a process in which the open lumen of the
stent is occluded by hyperproliferation of endothelial cells.
Various methods for the prevention of restenosis have
been investigated. These methods include administering the
patient with medications such as statin to reduce low-density
lipoprotein cholesterol or clopidogrel to reduce risk for blood
clots. Stents coated with a pharmacologic agent (drug eluting
stents) are known to interfere with the process of restenosis.
Since these methods do not assure the prevention of restenosis,
the patient must be monitored to ensure patency of peripheral
stents on a regular basis.
Current techniques for the diagnosis of PAD and
restenosis in stented peripheral arteries include quantitative
Doppler ultrasonographic measurements of blood flow
velocity [2–4]. Duplex ultrasound has been used to define the
anatomic extent of PAD. Although these methods have been
useful in documenting the patency of a single arterial segment,
such as a stented superficial femoral artery or a bypass graft,
assessment of the entire lower extremity arterial tree remains
imprecise and its adequacy as the sole diagnostic modality for
planning a percutaneous or open surgical intervention remains
controversial [5].
Magnetoelastic resonance angiography, as used in this
work, serves as a direct method for the detection of blockages
in stents and can thus enable timely medical intervention.
The method is outlined in figure 2. A magnetoelastic sensor
integrated with the stent resonates at a frequency that responds
to changes in mass loading, and to a lesser extent to changes in
the blood viscosity. The mechanical resonance is interrogated
by a set of external coils. The transmit, pickup and dc
magnetic coils are placed in a co-axial coil configuration in
a sleeve around the limb. This interrogation technique utilizes
magnetic inductive coupling for a short interrogating distance
(near field). Unlike other purely electromagnetic wireless
techniques, magnetoelastic coupling exploits the relationship
2
J. Micromech. Microeng. 23 (2013) 025010
A Viswanath et al
Table 1. As-cast material properties of MetglasTM 2826 MB alloy.
Density (kg m–3)
Thickness (μm)
Saturation magnetostriction (ppm)
Saturation bias field (Oe)
7900
28
12
8
2. Design and modeling
2.1. Material selection for the resonant sensors
Figure 3. Sensor and stent geometry showing important dimensions.
A sensor bonded to a single stent cell is also shown.
Magnetoelastic sensors are typically made of amorphous
ferromagnetic ribbons or wires, mostly iron-rich alloys
such as Fe40Ni38Mo4B18 (brand name MetglasTM 2826MB),
Fe90Si5B5 (brand name MetglasTM 2605SA1) and
Fe67Co18B14Si1 (brand name MetglasTM 2826CO) [14, 15].
The MetglasTM ribbons listed above have a high
magnetoelastic coupling coefficient, as high as 0.98,
and magnetostriction on the order of 10−5 [16–18]. A high
magnetoelastic coupling allows efficient conversion between
magnetic and elastic energies and vice versa. In addition, the
magnetomechanical coupling coefficient of the sensor can
be controlled by annealing the magnetoelastic material in a
transverse magnetic field [19–25], or changing the bias field
and/or sensor aspect ratio.
In our previous work we studied the properties of each of
the MetglasTM alloys (2826MB, 2605SA1 and 2605CO) with
simple sensor designs [7]. Along with maintaining a usable
wireless signal under loading, the sensor must present a low
profile (to maintain the open flow channel of the stent) and must
not hinder the mechanical operation of the stent (especially
expandability and bending flexibility). The amorphous alloys
mentioned above can meet these requirements, as they are
thin (≈25 μm), can be located along the sidewall of the stent
and can also be patterned such that the sensor expandability
matches that of the stent.
In an effort to improve the elasticity of the stent, chromenickel Elgiloy is used. Elgiloy is advantageous in the sense that
it has a much higher yield strain than stainless steel: ≈1% for
Elgiloy versus ≈0.15% for 316 L stainless steel [26]. Elgiloy is
also less prone to corrosion. In order for the sensor to conform
better to the stent, a sensor material would have to possess
appropriate elastic properties. Fortunately, MetglasTM alloys
meet the requirements for both magnetostrictive and elastic
properties. For instance, the 2826MB alloy, as used in this
effort, is reported to have a yield strain of 1.6%, which is even
higher than that of cold-reduced Elgiloy [27]. Some important
material properties of MetglasTM 2826MB alloy are listed in
table 1.
patterning into the shape of conventional stent cells with
overhanging bi-layer resonators formed at particular locations
along the stent. The bonding strategy used for this integration
is described in section 3. The dimensions of the stent cell
and the sensor active area are shown in figure 3. The sensor
layer comprises a frame and an active resonator portion. The
frame consists of 150 μm wide struts that are patterned in
the same wishbone-array pattern as a 12 mm × 1.46 mm
stent cell. The frame is bonded to the stent struts. The
active portion is a 10 mm long symmetric leaf shape and
is connected to the frame with a small anchor at mid-length.
The leaf shape nests within the frame and stent cell, with a
uniform gap separating the active portion from the frame. This
gap is 125 μm wide and allows for mechanical decoupling
between the sensor and the frame. The typical active area of
a sensor is approximately 4.5 mm2. The resonator is excited
in its fundamental, longitudinal extensional mode of vibration
which produces movement of the ends of the active area of the
sensor.
The stent application calls for a generally tubular shape
for use in angioplasty. This sensor design allows for the easy
coiling of stents into this shape without excessive mechanical
strain on the magnetoelastic material, which may lead to
unwanted shifts in resonance response.
2.3. Modeling
The magnetomechanically coupled finite element analysis
(FEA) tool presented in [6] is used to estimate frequency
responses and expected signal amplitudes of the sensors. The
desired sensor geometry is modeled in the FEA program,
along with the geometry of the transmit and receive coils.
In the simulations, the transmit and receive coils have an
internal diameter of 14 and 13 cm respectively. The sensor is
positioned along the central axis of the transmit/receive coils
with a radial distance of 6.5 cm separating the sensor from
the receive coil. The current in the transmit coil was measured
and applied in the model to generate the field at the sensor.
The magnetic flux density emanating from the sensor was
integrated numerically (with appropriate scaling factors) over
the volume of the receive coil to calculate the induced EMF.
The apparent modulus, permeability and magnetostrictivity
used in the model were based on available literature values.
These values were modified slightly to improve the fit with
experimental frequency responses obtained from stent cell
resonators fabricated from the actual material used. A baseline
2.2. Structural design
As shown in figure 3, the sensor design conforms to the cell
of a conventional stent structure. The stent design follows a
wishbone-array pattern that is favored for its flexibility during
expansion. A similar pattern is used in commercial stents such
R
R
R
as the PRECISE
PRO RX
and the CYPHER
(Cordis
Corporation, a Johnson and Johnson company). The bonding
of magnetoelastic material to stent material is followed by
3
J. Micromech. Microeng. 23 (2013) 025010
A Viswanath et al
Figure 4. Frequency response of unloaded sensor in air. The
measured resonant frequency is 361 kHz while the custom
magnetomechanical FEA model resonates at 346 kHz.
Figure 6. Process flow for the fabrication of bi-layer stent cell
resonators integrated with the stent. (1) MetglasTM 2826MB and
Elgiloy foils are aligned and bonded using the Au–In eutectic
bonding process to form the bi-layer. (2) Batch patterning of the
bonded foils is performed using μ-EDM. (3) Bi-layer stent cell
resonators at specific locations along the stent frame are fabricated.
Parylene deposition is then performed on the resonators to passivate
them and make them bio-compatible.
Figure 5. Stent cell resonator response to mass loading in water
flow (velocity of 15 cm s−1) and at a temperature of 37◦ C. Mass
loading is provided by paraffin wax. Mo denotes the unloaded sensor
mass and m the mass load on the sensor.
signal was taken in the absence of the sensor, which was
simulated by setting the magnetoelastic coupling coefficient to
zero. This baseline was subtracted from the overall signal, with
the sensor present. Figure 4 shows the simulated frequency
response of the sensor. The simulated resonant frequency for
the stent cell resonators in its fundamental longitudinal mode
of vibration is 346 kHz. The simulated response has a signal
amplitude of 2.3 mV at resonance. In addition, the theoretical
response of magnetoelastic sensors to uniform mass loading
is given by the following characteristic equation [14]:
1
.
floaded = fo,unloaded
(1)
1 + m
Mo
3. Fabrication
The process flow for the fabrication of bi-layer stent cell
resonators integrated with the stent is shown in figure 6
(this process is intended to provide rapid prototyping for
research investigations and will need to be modified for
final production). MetglasTM 2826 MB and Elgiloy foils are
aligned and bonded using the Au-In eutectic bonding process.
This results in bi-layer foils comprising the sensor and stent
material. Batch patterning of the bonded foils is then performed
using μ-EDM (micro-electrode discharge machining) to result
in bi-layer stent cell resonators at specific locations along the
stent frame.
A number of bonding techniques were evaluated to
integrate MetglasTM 2826MB onto the Elgiloy stent material.
These include methods such as parylene–parylene bonding,
dental cements, hot-plate and NanofoilTM soldering. The most
successful technique was the gold–indium (Au–In) eutectic
bond [28]. This process involves the deposition of multiple
layers of Au and In on the component metal surfaces. A
low temperature bonding process is carried out between the
In this equation, f loaded and fo,unloaded denotes the mass
loaded and unloaded resonant frequencies of the sensor
respectively. Mo denotes the un-loaded mass of the sensor
and m denotes the mass load on the sensor. The simulated
frequency response of the sensor to mass loads is shown in
figure 5. The FEA simulated, unloaded resonant frequency of
346 KHz is used in the analytical model.
4
J. Micromech. Microeng. 23 (2013) 025010
A Viswanath et al
Table 2. Bonding process parameters.
Bonding temperature
Bonding time
Heating ramp rate
Cooling ramp rate
Bonding pressure
Vacuum pressure
200◦ C
60 mins
10◦ C min–1
1.65◦ C min–1
≈0.95 MPa
29.4 mTorr
27% In and 73% Au in the final eutectic bond. This particular
stoichiometry allows for a low temperature formation of the
eutectic at 200 ◦ C while having a high re-melting temperature
of 450 ◦ C [28]. The minimum allowable thicknesses of
bonding layers were set in order to overcome the maximum
surface roughness of metals. The surface roughness of both
metals is measured using an Olympus LEXT Interferometer.
The maximum peak-to-peak variation of surface roughness is
measured to be ≈6 μm for MetglasTM 2826MB and 3 μm for
Elgiloy. The bonding layers thicknesses chosen (figure 7(a))
add up to 12.9 μm and hence overcome the inherent surface
roughness of the component metals.
The Au–In eutectic bonding process is performed in a
vacuum oven at a temperature of 200 ◦ C, with 1 MPa pressure
applied. Bonding pressure is applied using custom designed,
spring-loaded C-clamps which are calibrated for pressure
loading. Glass slides are used to uniformly distribute pressure
along the area of the bonding materials. A heating ramp rate of
10 ◦ C min−1 and a cooling rate of 1.65 ◦ C min–1 ensure
adequate inter-diffusion and solidification of the mixture to
form the Au–In bond. Table 2 lists the parameters used in
the bonding process. The Au–In eutectic bonding allows for a
bonding layer that is ≈12.9 μm thick. Combining the 28 μm
thick MetglasTM and 150 μm thick Elgiloy foils, this results
in a total thickness of ≈190 μm for the bi-layer. Figure 7(b)
shows an SEM image of a test structure that has been sectioned
by μEDM after bonding. EDX analysis was performed to show
weight percentages of gold and indium across the bond.
Isolated, single stent cell resonators are patterned from
a pre-bonded MetglasTM 2826MB to Elgiloy piece. Batch
patterning of these resonators is carried out by serial μEDM
of pre-bonded MetglasTM to Elgiloy pieces. Tungsten tool
electrodes of 125 μm diameter provide a good compromise
between machining speed and minimum feature sizes
achievable and were thus used in the μEDM process. The
machined sensors are released and cleaned thoroughly to
remove any debris as a result of the machining. The resulting
single stent cell resonators are bi-layers of MetglasTM 2826MB
and Elgiloy. A fabricated, isolated, single stent-cell, bi-layer
resonator is shown in figure 8.
Magnetoelastic alloys are known to corrode in aqueous
environments due to its high iron content. To passivate the
material, the sensors are coated in a conformal layer of
200 nm thick Parylene-C using a standard vacuum deposition
technique. This process results in sensors that are more robust
in corrosive environments while causing negligible shifts in
resonator frequency and amplitude response.
(a)
(a)
Figure 7. (a) Initial deposited layer thicknesses prior to Au–In
eutectic bonding. (b) SEM image of bond cross-section showing
Au–In eutectic alloy formation. EDX analysis was performed to
show weight percentages of gold and indium across the bond.
two components up to 200 ◦ C. Above 157 ◦ C the indium layer
melts and dissolves the gold layers to form a mixture of liquid
and solid. The solid-liquid interdiffusion process continues
until the mixture solidifies to form the Au–In bond. This
technique has established high bond strengths ranging from
4.4–10 MPa in literature, and is a relatively low temperature
process having a high bond re-melting temperature [29]. The
low temperature process is beneficial in avoiding thermal
stresses induced in the magnetoelastic material that may
otherwise lead to recrystallization or shifts in performance.
The final fabrication sequence is as follows: Elgiloy
foils (125 μm thick) are first subjected to a back-sputtering
treatment to improve Cr/Au adhesion to its surface. This
involves plasma treatment of the Elgiloy at a power of 150 W in
oxygen gas flow for a time period of 60 s. Subsequent chrome
deposition then results in chrome oxide formations leading to
better adhesion, which were confirmed by simple scotch tape
peel tests. The Elgiloy foils are then pre-coated with a 0.6 μm
thick Cr/Au adhesion layer (using an evaporation technique)
and a 6 μm thick indium layer (using electroplating). As an
added precaution, a 0.2 μm thick Au layer helps prevent
oxidation of indium prior to bonding. The 28 μm thick
MetglasTM foil surfaces are coated with a 100 nm Cr adhesion
layer (using evaporation technique) and a 6 μm gold layer
(using an electroplating technique). These make up the eutectic
ingredient layers on the metal surfaces required for the bonding
process. A schematic of the bonding layer thicknesses is shown
in figure 7(a). Thicknesses of electroplated/sputtered gold and
indium are chosen to result in specific weight percentages of
5
J. Micromech. Microeng. 23 (2013) 025010
A Viswanath et al
This measurement setup is consistent with the
configuration needed for an implanted sensor in one of
the posterior arteries (i.e. arteries in the legs or hands) of the
patient. Additionally, the co-axial coil configuration achieves
a higher signal-to-noise ratio than other configurations and
does not have stringent requirements for magnetoelastic sensor
positioning within the coil [6].
Flow tests were performed using either a positivedisplacement, dc-operated pump (B&D Mfg., Inc.) or a
peristaltic pump, depending on the liquid flowing from the
reservoir. Tubes having an inner diameter of 4.0 mm were used
in the flow setup to mimic the average diameter of the posterior
tibial artery [30]. The sensor was placed in the tube at the center
of the interrogation coils. The temperature of the test liquid
was maintained at 37 ◦ C (mimicking normal human body
temperature). Polyimide thermofoil heaters (Minco Products,
Inc., Minnesota, USA) were placed on the outside of the tube
to maintain a local temperature of 37 ◦ C around the sensor.
Water has an average density that is close to that of blood
and was consequently used in flow velocity sensitivity and
mass loading tests. Evaluation of viscosity sensitivity involved
the use of varying concentrations of sugar water solutions to
mimic conditions of blood flow before and after restenosis. The
sensitivity to flow velocity changes was evaluated by varying
the power supply to the dc pump and hence the velocity of
water. The velocities were varied between 20 and 11 cm s−1 to
replicate conditions of blood flow in peripheral arteries during
diastolic and systolic cycles respectively [31]. As viscosity
changes of blood flow accompany the constriction leading
to restenosis in an artery, the sensitivity to viscosity change
was evaluated. Viscosity levels of the test liquid were varied
between 1.1 and 15 cP using varying concentrations of sugar
(sucrose) water solutions [32]. The velocity of flow for the
sugar solutions was 15 cm s−1. The volume of water used for
each sample was 200 mL and the molar mass of sugar used in
the calculations was 342.3 g mol−1.
The sensitivity to mass loading was evaluated by
application of paraffin wax onto the overhanging resonator
portion of the sensor. The mass loaded sensors were subjected
to water flow (velocity of 15 cm s−1) at room temperature. A
precise weighing scale (with 0.1 mg resolution) was used to
measure the unloaded and loaded sensor mass. A repeatability
assessment of the sensors was performed in water flow. The
frequency response of the sensors was measured 15 times with
a time interval of 10 min between successive measurements.
The amplitude of the resonant peak (Vres) and the anti-resonant
peak (Vanti_res) were measured, and the average of the two
magnitudes was calculated. In all experiments the fundamental
longitudinal mode of resonance was studied, unless otherwise
noted.
The effect of the bonding process on the magnetic
properties of magnetoelastic Metglas 2826MBTM material was
assessed. The magnetic properties of the Metglas 2826MBTM
material were studied using a vibrating sample magnetometer
(VSM), (Princeton Corp., New Jersey, USA). Two sets of
samples were tested: ‘as-cast’ ribbons and ribbons subjected
to temperatures and pressures encountered in the bonding
process (table 2). The magnetization curves for these two sets
Figure 8. Fabricated resonators (a) Isolated sensor comprising of
bi-layer MetglasTM –Elgiloy resonators. (b) Perspective view of the
anchor of the bi-layer resonators.
Figure 9. Co-axial coil configuration used for resonance
measurements during in vitro tests.
4. Experimental methods and results
A co-axial coil configuration was used for benchtop testing
and evaluation of the sensors (figure 9). Dual-Helmholtz coils
were used to provide a uniform and well-controlled dc bias
field to the sensors. The dc magnetic bias was used to set the
operating point of the sensor response [6–8]. The transmit coils
were located on both sides of the receive coils. The transmit
and receive coils were aligned in a co-axial and concentric
configuration within the DC bias field coils. The inner diameter
of all coils was chosen to be 13 cm. The number of turns in
the dual-layered transmit and receive coils were chosen to be
48; the dc coils were wound with 102 turns. Wires of 22 AWG
type were used in the coils.
The sensor to be evaluated was positioned at the center
of this co-axial coil configuration, with the long axis of the
sensor oriented along the coil axis. The transmit and receive
coils were driven and measured by an Agilent 4395A network
analyzer. The output signal from the network analyzer was
amplified using a power amplifier (Model 7500, Krohn-Hite
Corporation, Massachusetts, USA). The dc coils were powered
by a 6.2 V, 4.2 A signal from a dc power supply. A Hall probe
(AD22151, Analog Devices, Inc.) was used to measure the
resulting magnetic bias field strength across the sensor.
6
J. Micromech. Microeng. 23 (2013) 025010
A Viswanath et al
Figure 11. Stent cell resonator response to changes in viscosity
levels. Viscosity is varied from 1.1 to 15.4 cP using varying
concentrations of sugar (sucrose) in water. The resonant frequencies
measured are normalized to the unloaded, sensor resonant frequency
in air.
Figure 10. Measured resonance plots of bi-layer resonators in flow
at 37 ◦ C. Diastolic (flow velocity of 20 cm s−1) observed f res =
356.5 kHz while systolic (flow velocity of 11cm s−1) observed
f res = 356.6 kHz.
Table 3. Magnetic properties of Metglas2826MBTM ribbons in
“as-cast” condition and ribbons subjected to bonding process
temperature and pressure.
Metglas2826MBTM
ribbon type
“As-cast” ribbon
Exposed to bonding
process
Coercivity
(mOe)
37.33
−142.6
Remanence
(memu)
0.330
−3.671
of samples were measured with the VSM. The magnitude of
coercivity increased from 37.33 to 142.6 mOe. The measured
magnitude of remanence also increased from 0.330 to
3.671 memu (table 3). Although there is slight increase in
the magnitude of the coercivity and remanence, it is evident
that the material remains magnetically soft and the practical
effects on the sensor are insignificant.
Isolated sensors were tested in vitro for resonance
response to various parameter changes. The unloaded response
of a typical sensor in air is presented in figure 4. For this device,
the typical unloaded resonant frequency is 361 kHz for the
fundamental, longitudinal mode of vibration. The sensitivity
was evaluated for changes in flow velocity of water. The flow
velocity was varied between 20 and 11 cm s−1 to mimic
systolic and diastolic conditions of blood flow. The measured
frequency response for each condition, at 37 ◦ C, is shown
in figure 10. The maximum increase in resonant frequency,
due to 9 cm s−1 decrease in flow velocity, fell within the
measurement error of the network analyzer. The measured
sensitivity of the fabricated sensors to flow velocity was less
than 155 ppm cm s−1. This is a favorable attribute because the
sensors are not intended to respond to flow velocity.
The typical viscosity sensitivity of the sensors to varying
viscosity levels of sugar water flow is presented in figure 11.
The resonant frequencies measured are normalized to the
unloaded, sensor resonant frequency in air. For viscosity levels
of 1.084 and 8.596 cP, the measured resonant frequency was
357.65 and 356.505 kHz respectively. The maximum change
in frequency observed is 0.32% over a 1.1–8.6 cP range. This
corresponds to a viscosity sensitivity of 427 ppm cP−1 for the
sensor.
Figure 12. Repeatability assessment for stent cell resonators.
The sensors were characterized for sensitivity to mass
loading using paraffin wax to simulate the plaque/tissue
depositions. The unloaded sensors were found to have an
average weight of 8.5 mg. Mass loads upto 15% of the
unloaded mass of the sensor were evaluated. A typical
measured resonance response after mass loading is shown
in figure 5. Also shown in this figure is the theoretically
expected decrease in resonant frequency, assuming uniform
mass loading on the sensor.
The measured sensitivity to mass loading was found to
range from 630 00 to 650 00 ppm mg−1 with a maximum
resonant frequency change of 8.1% for 15% mass loading
on the sensors. Additionally, the trend observed in measured
response agrees with that seen in the theoretical response
within 3.5% error.
An assessment of repeatability involved resonance
measurements for 15 trials with a time interval of 10 min
in between trials with the sensor position and interrogation
parameters maintained constant for all trials. The results of
the repeatability experiments are detailed in figure 12. The
maximum change in resonant frequency measured between
trials was around 0.01% or 100 ppm over a time period of
140 min. This corresponds to a mass load of 0.02% of the
unloaded sensors mass.
7
J. Micromech. Microeng. 23 (2013) 025010
A Viswanath et al
5. Discussion and conclusions
The parylene coating used in this work is intended to
both reduce the corrosion of the sensor, as well as to provide
a surface with improved biocompatibility to the endothelial
cells and blood cells. The long term efficacy of parylene in
maintaining biocompatibility of the sensor is of interest and
should be evaluated in future testing. Other biocompatible
coatings such as titanium may be an alternative in this
application.
The sensors show a negligible cross sensitivity to
flow velocity changes, allowing for reliable measurement
of viscosity and mass loading. Although the sensitivity
to viscosity is relatively low when compared to that of
mass loading, sensors may be operated reliably as viscosity
detectors to measure changes between 1.1 and 8.8 cP. As
expected, the sensitivity to mass loading ranges from 630 00
to 650 00 ppm mg−1. Although this paper presents the sensors
in the context of peripheral artery stents, such devices may be
useful for coronary artery diseases as well.
This paper presents the sensor design and evaluation for
wireless monitoring of tissue accumulation in stents used
to treat PAD. Magenotelastic sensors resonating in their
fundamental, longitudinal extensional mode of vibration are
integrated to a single stent cell. The Au–In eutectic bonding of
sensor material to stent allows for low temperature integration
of resonators to stent without affecting its magnetoelastic
properties. Benchtop in vitro testing of sensors was performed
to characterize the sensors for flow velocity, viscosity and mass
loading sensitivities.
The FEA simulated resonant frequency of the sensors is
within 4.2% of the measured response. The mismatch may
be attributed to an error in the assumed material properties
or surface irregularities and geometrical imperfections of the
tested sensors. A more accurate FEA model taking into account
these factors for complex geometries is expected to reduce the
mismatch even further.
As a result of the constriction brought about by restenosis,
a sensor placed within a stent is subjected to changes in
blood flow velocities, viscosity levels and mass accumulation
of endothelial cells simultaneously. The sensors should thus
be capable of differentiating between these three parameters
of interest. We have previously reported how a measurement
of quality factor can be used to separate viscosity and mass
response for magnetoelastic sensors [7]. However, the sensors
should be insensitive to flow velocity for reliable use in
this application, which is consistent with the measured results.
The sensitivity to viscosity was typically 427 ppm cP−1. The
sensors, however, show a large sensitivity to mass loading,
with loads as small as 6% of the unloaded sensor mass,
producing significant changes in the resonant response. Also,
as seen in figure 5, the frequency sensitivity of the sensors
has not saturated at the maximum mass load, suggesting that
the full scale range of these sensors is even greater than that
reached in these tests. A relative assessment of mass loading
and viscosity sensitivities favors the use of these sensors as
mass detectors in stenosed arteries. However, in the absence
of stenosis, these sensors may be used for estimating changes
in viscosity as well. This suggests a dual application for these
sensors depending on the requirement. Repeatability studies
show that the stent cell resonators exhibit a typical, maximum
change in unloaded resonant frequency of 0.01% over a period
of 140 min. This, in turn, corresponds to a mass load of 0.02%
of the unloaded sensor mass, or 1.7 μg for the bi-layer stent cell
resonators (which have an average unloaded weight of 8.5 mg).
The total drift of the sensors over time is thus negligible in
comparison to the mass loading of interest due to restenotic
pathologies.
It is to be noted that for this particular application,
the measurand of interest includes bio-fouling agents. Stent
restenosis occurs due to buildup of ‘bio-fouling’ endothelial
hyperproliferation. The buildup of these ‘bio-fouling’ agents
collectively provides a mass load to the sensors which is
interpreted as a resonant shift. These sensors may therefore
be used in other applications that are affected by bio-bouling,
such as implanted glucose sensors.
Acknowledgments
The authors acknowledge Dr Christine Eun and Jun Tang
for assisting with the thin film layer depositions required
for the eutectic bonding process. Dr Tao Li assisted with the
parylene coating steps. Metglas Inc. provided samples for this
project. This work was supported in part by the King Abdullah
University of Science and Technology (KAUST, Saudi Arabia)
and the University of Michigan.
References
[1] Ouriel K 2001 Peripheral arterial disease Lancet 358 1257–64
[2] Ouriel K and Zarins C K 1982 Doppler ankle pressure: an
evaluation of three methods of expression Arch. Surg.
117 297–300
[3] Ouriel K, Veith F J and Sasahara A A 1988 A comparison of
recombinant urokinase with vascular surgery as initial
treatment for acute arterial occlusion of the legs New Engl.
J. Med. 338 1105–11
[4] Ubbink D T, Tulevski II, Hartog D, Koelemay M J,
Legemate D A and Jacobs M J 1997 The value of
non-invasive techniques for the assessment of critical limb
ischaemia Eur. J. Vasc. Endovasc. Surg. 13 296–300
[5] Visser K and Hunink M G 2000 Peripheral arterial disease:
gadolinium-enhanced MR angiography versus color-guided
duplex US—a metaanalysis Radiology 216 67–77
[6] Green S R and Gianchandani Y B 2009 Wireless
magnetoelastic monitoring of biliary stents IEEE/ASME
J. Microelectromech. Syst. 18 64–78
[7] Green S and Gianchandani Y B 2010 Tailored magnetoelastic
sensor geometry for advanced functionality in wireless
biliary stent monitoring systems J. Micromech. Microeng.
20 075040
[8] Green S, Kwon R, Elta G and Gianchandani Y B 2010 In situ
and ex vivo evaluation of a wireless magnetoelastic bilary
stent monitoring system Biomed. Microdevices 12 477–84
[9] Park J Y and Davies J E 2000 Red blood cell and platelet
interactions with titanium implant surfaces Clin. Oral
Implants Res. 11 530–9
[10] Kouzoudis D and Grimes C A 2000 The frequency response of
magnetoelastic sensors to stress and atmospheric pressure
Smart Mater. Struct. 8 885–9
8
J. Micromech. Microeng. 23 (2013) 025010
A Viswanath et al
[11] Jain M K, Schmidt S, Ong K G, Mungle C and Grimes C A
2000 Magnetoacoustic remote query temperature and
humidity sensors Smart Mater. Struct. 9 502–10
[12] Jain M K, Cai Q Y and Grimes C A 2001 A wireless
micro-sensor for simultaneous measurement of pH,
temperature, and pressure Smart Mater. Struct.
10 347–53
[13] Grimes C A, Kouzoudis D and Mungle C 2000 Simultaneous
measurement of liquid density and viscosity using remote
query magnetoelastic sensors Rev. Sci. Instrum.
71 3822–4
[14] Grimes C A, Mungle C, Zeng K, Jain M K, Dreschel W R,
Paulose M and Ong K G 2002 Wireless magnetoelastic
resonance sensors: a critical review Sensors
2 294–313
[15] Cai Q Y and Grimes C A 2000 A remote query magnetoelastic
pH sensor Sensors Actuators B 71 112–7
[16] Hernando A, Vazquez M and Barandiaran M 1988 Metallic
glasses and sensing applications J. Phys. E 21 1129–39
[17] Modzelewski C, Savage H T, Kabacoff L T and Clark A E
1981 Magnetomechanical coupling and permeability in
transversely annealed Metglas 2605 alloys IEEE Trans.
Magn. 17 2837–9
[18] O’Handley R C 2000 Modern Magnetic Materials: Principles
and Applications (New York: Wiley)
[19] Brouha M and van der Borst J 1979 The effect of annealing
conditions on the magnetomechanical properties of
Fe–B–Si amorphous ribbons J. Appl. Phys 50 7594–6
[20] Soyka V, Kraus L, Zaveta K and Jurek K 1999
Magnetoelastic properties of stress/field annealed
Fe80Cr2B14Si14 amorphous alloy J. Magn. Magn. Mater.
196–197 262–3
[21] Clark A and Wun-Fogle M 1989 A new method of
magnetostrictivity and magnetostriction measurement IEEE
Trans. Magn. 25 3611–3
[22] Kim M H, Lee K S and Lim S H 1999 Magnetostriction
measurements of metallic glass ribbon by fiber-optic
Mach–Zehnder interferometry J. Magn. Magn. Mater.
191 107–12
[23] Zhai J, Dong S, Xing Z, Li J and Viehland D 2006 Giant
magnetoelectric effect in Metglas/polyvinylidene-fluoride
laminates Appl. Phys. Lett. 89 083507
[24] Anderson P III 1982 Magnetomechanical coupling, E effect,
and permeability in FeSiB and FeNiMoB alloys J. Appl.
Phys. 53 8101–3
[25] Modzelewski C, Savage H, Kabacoff L and Clark A 1981
Magnetomechanical coupling and permeability in
transversely annealed Metglas 2605 alloys IEEE Trans.
Magn. Mag 17 2837–9
[26] ASM International Handbook Committee 1990 Metals
Handbook (Bilthoven, The Netherlands: ASM Int.)
[27] Lin J J and Perng T P 1995 Embrittlement of amorphous
Fe40Ni38Mo4B18 alloy by electrolytic hydrogen Metall.
Mater. Trans. A 26 197–201
[28] Lee C C, Wang C Y and Matijasevic G 1993 Au–In bonding
below the eutectic temperature IEEE Trans. Compon.
Hybrids Manuf. Technol. 16 311–6
[29] Aktakka E E, Kim H and Najafi K 2009 Wafer level
fabrication of high performance MEMS using bonded and
thinned bulk piezoelectric substrates IEEE Int. Conf. on
Solid State Sensors, Actuators, and Microsystems
(Transducers) 849–52
[30] Sabatier M J, Stoner L, Reifenberger M and McCully K 2006
Doppler ultrasound assessment of posterior tibial artery size
in humans J. Clin. Ultrasound 34 223–30
[31] Fronek A, Coel M and Bernstein E F 1976 Quantitative
ultrasonographic studies of lower extremity flow velocities
in health and disease Circulation 53 957–60
[32] Haynes W M (ed) 2011 CRC Handbook of Chemistry and
Physics 92nd edn (Boca Raton, FL: CRC Press)
9
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Related manuals

Download PDF

advertisement