Sensors 2013, 13, 4581-4597; doi:10.

3390/s130404581
OPEN ACCESS

sensors
ISSN 1424-8220
www.mdpi.com/journal/sensors
Review

Molecular Electric Transducers as Motion Sensors: A Review

Hai Huang 1, Vadim Agafonov 2 and Hongyu Yu 1,3,*
1
School of Electrical, Computer and Energy Engineering, Arizona State University, Tempe,
AZ 85287, USA; E-Mail: [email protected]
2
Center of Molecular Electronics, Moscow Institute of Physics and Technology, Moscow 141700,
Russia; E-Mail: [email protected]
3
School of Earth and Space Exploration, Arizona State University, Tempe, AZ 85287, USA

* Author to whom correspondence should be addressed; E-Mail: [email protected];

Tel.: +1-480-727-7454; Fax: +1-480-965-8102.

Received: 5 March 2013; in revised form: 2 April 2013 / Accepted: 2 April 2013 /
Published: 8 April 2013

Abstract: This article reviews the development of a new category of motion sensors including
linear and angular accelerometers and seismometers based on molecular electronic
transducer (MET) technology. This technology utilizes a liquid not only as an inertial
mass, but also as one of the main elements in the conversion of mechanical motion into
electric current. The amplification process is similar to that in a vacuum triode. Therefore,
it is possible to achieve signal amplification close to 108. Motion sensors demonstrating
wide frequency and dynamic range and sensitivity that are one to two orders of magnitude
better than MEMS devices of the same size have been developed.

Keywords: molecular electronic transducer; accelerometers; seismometers

1. Introduction

Motion sensors, including accelerometers and gyroscopes, provide measurement of movement in at

least six degrees of freedom. The simplest way to do motion sensing is with a solid-state mass-spring
system, creating a damped simple harmonic oscillator. The movement of the solid-state proof mass can
be measured with respect to displacement, velocity, or acceleration by suitable mechanical-electrical
transducers. Efforts to miniaturize linear accelerometers and gyroscopes for inertial systems are mostly
concentrated around Micro-Electro-Mechanical Systems (MEMS) technology. Similarly, in terms of
Sensors 2013, 13 4582

design, fabrication, and readout, accelerometers and gyroscopes are the current leaders in
commercially successful MEMS technology. Among a variety of transduction mechanisms underlying
solid-state MEMS motion sensors, the most successful types are based on capacitive transduction due
to the simplicity of the sensor element itself, no need for exotic materials, low power consumption, and
good stability with respect to temperature. Although capacitive transducers have a characteristic
nonlinear capacitance vs. displacement response, feedback is commonly used to convert the signal to a
linear output. MEMS motion sensors in combination with other sensors, such as compass, pressure
sensor, and GPS, have created a consumer electronics sensing package that works as the intelligent
interface for users to interact with their electronics and, further on, with external environments, and
have reasonable performance with low cost CMOS-compatible silicon microfabrication technology.
However, in addition to high-volume consumer electronics markets requiring low-to-medium
performance motion sensors, there are huge markets for high-performance motion sensing devices,
with applications designed for military inertial navigation/guidance, high-resolution seismic sensing
and high-g sensors. The key requirements for these high-performance applications include small size,
wide bandwidth, low noise floor, low cross-axis sensitivity, low drift, wide dynamic range, high shock
survivability, and low power consumption. There has been plenty of work done along the direction of
scaling down the device size while maintaining low noise, high sensitivity and high resolution using
MEMS techniques [1–3]. However, the design and fabrication of these solid-state MEMS devices are
complicated, which can result in low reliability, low reproducibility, and high cost. More importantly,
they have yet to prove satisfactory for specific applications, especially in low frequency seismic
sensing, because of the inherent limitation of the working mechanism of the solid-state mass-spring
system. For some applications they also have limited successes due to their fragility to high shocks.
As opposed to a solid inertial mass, a molecular electronic transducer (MET) is sensitive to the
movement of a liquid electrolyte relative to fixed electrodes. METs are part of a third class of
fundamental electronic devices, characterized by charge transfer via ions in solution—hence the name
“Solion”. This is in contrast to solid-state electronics (charge transfer by electron/hole pairs in a solid
conductor or semiconductor) and vacuum tubes (charge transfer by free electrons in an ionized gas or
vaccum). Solion technology was first developed in the 1950s by US-Navy sponsored research. Early
applications of Solion devices were for detection of low-frequency acoustic waves, either in the form
of an infrasonic microphone or limited-band seismometer [4–7]. Significant work on Solion motion
detectors was continued in Russia, where the term “Molecular Electronic Transducer” was introduced
to describe such a device [8]. Inspired by the exceptionally high rate of mechanical signal conversion
to electric current in MET involving mass and charge transport, pioneering MET studies [9–15]
provide an alternative paradigm in the development of motion sensors in wide variety of applications
including nuclear explosion monitoring and seismic sensing in planetary exploration [16–18]. The
advantages of MET motion sensors include their small size, lack of fragile moving parts (thus high
shock tolerance), high sensitivity and low noise especially at low frequencies, and independence of the
response on installation angle.

2. Molecular Electronic Transducer-Principle of Operation

S
Sensors 2013, 13 45883

Figure 1 illustrates thet basic cooncept of a MET senssing elemennt, consistinng of an electrochemical
c
cell with twwo anode-ccathode paiirs. An eleectrolyte-fillled channeel allows thhe electroly yte to movve
innertially allong the chhannel’s lenngth. Four electrodes configuredd as anodee-cathode-caathode-anodde
(ACCA) sepparated by dielectric layers span the width of this channnel. Holess through th he electrodees
a
allow the floow of electrrolyte. Eachh anode-cathhode pair iss an electrocchemical ceell, in which
h charges arre
trransferred between
b annode and caathode by ionsi in the electrolyte.. Traditionaally, standaard machineed
p
platinum (Ptt) mesh is used
u to makke the electrrodes, and plastic
p or ceramic gridds produce thet dielectric
innter-electroode spacers [19]. This approach
a is convenientt to quicklyy produce saamples, but significantlly
liimits the poossible rangge of the geeometrical parameters
p of the cell and conseqquently the optimizatioon
p
possibilities of the senssing cell. Meanwhile
M m
making sizees of elements of the ccell smallerr expands thhe
f
frequency range and decreases
d coonvection produced
p noise
n insidee the sensitiive cell. An
A alternativve
a
approach to build the sensing
s elem
ment is to use
u (MEMS S) techniquees, which haave reduced d the internal
d
dimensions of MET eleectrodes close to 1 µm, improved the t sensitiviity and reprroducibility [18].

Figure 1. Schematic
S o the basic MET sensinng element..
of

The sensiing mechannism of the MET is based on usin ng of the eleectrochemiccal cell wheere reversible
c
chemical reaactions transsfer charge between
b anoode and catthode via eleectrolyte ionns in solutio
on. Typicallyy,
M uses cooncentrated iodine-iodidde electrolyyte containin
MET ng potassium
m iodide (KII) or lithium m iodide (LiI)
a a small amount of elemental Iodine
and I (I2) [20].
[ In the presence off iodide, ioddine turns in nto a soluble
c
compound, t
tri-iodide, a follows [220]:
as
(11)
when the eleectrodes aree biased andd thus curreent passes th
w hrough the electrochem
mical cell, the
t followinng
r
reversible acctive electroochemical reactions occcur on the electrodes:
e
On cathodes, reductiion of tri-iodide:
2 3 (22)
The reverrse reactionn takes placee on anodess:
3 2 (33)
Sensors 2013, 13 4584

The electrical current through the solid/liquid interface becomes possible because of the tri-iodide
ions presence in the solution. That’s why this component of the solution is called active. According to
Equation (2), the interface charge transfer is associated with generation and absorption of the tri-iodide
ions on the electrode surface. So the electrical current through any electrode can be related to the flux
of active ions toward or backward of the electrode according to the following:
(4)
,

where D is the diffusion coefficient, с is the concentration of the active charge carriers, q is the charge
transferred across the interface in single electrochemical reaction (two times absolute values of the
electron charge in our case), n is a unit vector normal to the surface of the electrode, integration is
done over S, electrode surface area. Here only diffusion is considered as mechanism responsible for the
active ions transport in the electrolyte volume. The migration is not included due to the screening of
the electrical field in the highly concentrated electrolyte and the convection does not contribute to
charge transfer through the electrode surface due to zero-velocity condition on the solid surface.
The operation principle of MET can be described as follows: when electric voltage is applied to the
system, electrochemical current (background current) appears, regardless of the presence of
mechanical motion of the electrolyte. As the inter-electrode voltage is increased, the reaction rates on
the electrodes increase too. Finally, in the situation when any tri-iodide ion arrives the cathode
immediately participates in the electrochemical reaction in Equation (2), further increase of the voltage
does not change the current and the saturation regime occurs. In this regime the cathode current is sensitive
to variation of volumetric transport of tri-iodide ions. Anode current variations follow the cathode ones,
keeping the electrolyte uncharged. In the presence of a mechanical motion input, electrolyte starts
moving due to inertia, and convective transport of ions changes the electrode current according to the
mechanism described above. The difference of the cathode currents in two anode-cathode pairs is
employed as the output signal for a MET motion sensor. Although each electrode current is non-linear
with respect to fluid velocity, the combined output of both cathodes is linear for a very wide range of
fluid velocities. Mathematically, the sensor’s output current is given by [20]:
(5)
, ,

where , are the currents through the surface of the corresponding cathodes, , are the
surface areas of the corresponding cathodes.

2.1. Transfer Function

A signal conversion in MET motion sensor can be considered as a superposition of two processes:
first, input motion is converted to fluid motion of the electrolyte by mechanical system. Next, the
electrolyte’s velocity is measured by the electrochemical system, resulting in an output current of the
sensor. Therefore, the frequency-dependent transfer function of the entire device can be written as:
(6)
Sensors 2013, 13 4585

where describes the mechanical response of the fluidic system, which in the case of linear
MET sensor, is analogous to a solid-state damped, driven harmonic oscillator. In this sense, the
restoring force to the liquid inertial mass is provided by the rubber membranes which are located at the
two ends of the channel to seal the electrolyte, and the damping force is caused by hydrodynamic
resistance of the electrolyte as it flows through the sensing element. The equation that governs motion
of the electrolyte can therefore be expressed as:

(7)

where V is the volume of fluid passing through the channel, a is the external acceleration, Rh is the
hydrodynamic resistance which is solely determined by the channel geometry in laminar flow
condition, k is the coefficient of volume stiffness and depends only on the characteristics of the
membrane, is the density of the electrolyte, Sch is the cross-section area of the channel and L
represents the length of the channel, filled with electrolyte. By transforming Equation (7) to the
frequency domain, the magnitude of the transfer function of the fluid mechanical motion in frequency
domain can be obtained as follows:

| |
(8)

where is the volumetric flow rate.

in Equation (4) describes the ability of the electrochemical system to detect electrolyte
motion as a function of frequency. A simple model for was derived analytically by Larcam [7],
having a form of:

| |
(9)
1

where C (A/(m3/s)) is the conversion factor of the electrochemical cell, / is the diffusion
frequency and d is the inter-electrode distance. In-depth characterization of the above electrochemical
subsystem is based on the analytical and numerical solution of the following partial differential
equations [11,15,21–24]:

· (10.1)

· 0 (10.2)
· 0
(10.3)
Equation (10.1) is the time domain Navier-Stokes equation describing momentum balances, where
denotes the density / , is the velocity / , denotes dynamic viscosity · , and P
equals pressure . Equation (10.2) is the continuity equation for incompressible flow. The
boundary conditions include a known pressure difference which drives the flow through the channel
and the velocity is zero at the wall. Equation (10.3) is the diffusion-convection Nernst-Planck equation
Sensors 2013, 13 4586

showing the tri-iodide transport balance. Usually, on cathodes, zero concentration condition is used, which
corresponds to the saturation current regime described above. Dielectric surfaces are considered as non-
penetrable for the ions (zero flow condition). For anodes, fixed concentration condition is frequently
used, although this sort of condition can’t be considered as well-founded. The problem of the correct
formulation of the boundary condition on anode is discussed in [24]. Fortunately, the boundary condition
on anodes has little effect on cathodes currents difference which is used as the system output signal.
Several theoretical analyses have been conducted on the electrochemical system frequency response
at both low and high frequencies [11,21,22]. At low frequencies, where diffusion length ⁄
appears to be much higher than the characteristic dimension which is the inter-electrode spacing of
the four-electrode structure:

Q · const (11)

Therefore, using Equation (8):

· (12)

As frequency increases, the sensitivity starts to decay as becomes lower than the characteristic
dimension of the electrode system . For the mesh electrodes made with cylindrical wires, and
⁄ , the amplitude of cathode current yields:
⁄
· / (13)

Therefore, using Equation (8):

⁄
· / (14)

2.2. Feedback Subsystem

The feedback subsystem is added to allow an additional controllable force to act on the mechanical
system. The feedback system herein introduces a layer of complexity into the overall system’s transfer
function. The simplified block diagram shown as Figure 2 illustrates how the input signal is
manipulated by each subsystem. Text in blue describes the physical parameters being affected or
measured in each step. The input signal is processed by the mechanical and electrochemical systems in
order to produce the output signal, which is in turn used to adjust the mechanical system by the
feedback system. The feedback current IFB is simply the output current modified by a feedback
parameter F. Finally, the effect of the feedback is applying to the inertial mass with additional
counterforce, opposite to inertial force, and consequently effectively decreasing the input acceleration.
In the feedback system described above, a higher output signal gives negative feedback impeding fluid
flow, thereby increasing the dynamic range of measurable signals. Moreover, this could be used to
adjust sensitivity frequency response, allowing desired signals in certain bandwidth.
S
Sensors 2013, 13 45887

F
Figure 2. A block diagrram of MET
T motion sensor including feedbacck subsystem
m.

2 Noise
2.3.

The mainn contributoors to the seelf-noise of the MET sensor are thhe thermohyydrodynamic self-noise,
c
convection-i
induced selff-noise, andd geometry noise
n [10,25].

2
2.3.1. Therm
mohydrodynnamic Self-N
Noise

The thermmohydrodynnamic self-nnoise is gennerated from

m the fluctuations of thhe pressure difference
d o
on
b
both sides of
o the channnel of METT sensor. In units of in nput acceleration, this nnoise spectrral density is
f
frequency inndependent::
2 · ·
(155)

where is thhe absolute temperature

w t e, is Boltzzmann’s con
nstant, is the
t electrolyyte density, is the lengtth
o the transduucer channeel,
of is thee hydrodynaamic impedaance of the transducer which is givven by:
8
(166)

where r is thhe radius of the circullar cross-secction of thee channel annd is the vviscosity off the fluid. It
w
c be seen that
can t the lowwer , highher and reesult in loweer thermohyydrodynamicc self-noise. For a typical
M
MET sensorr with 5 10 · / a
and 5 10 , thhe theoretical thermohy
ydrodynamiic
s
self-noise is 3 /√ .

2
2.3.2. Conveection-Induced Self-Nooise

The convvection-induuced self-nooise is the result

r of thee natural connvection off the liquid [25]. Even if
thhe liquid ass a whole iss stable, sm
mall local vaariations of the liquid density
d prodduce flows in the liquidd.
T
These flowss generate ana additionaal noise in the MET cell c output current.
c Thiis process is
i difficult to
t
d
describe anaalytically. However,
H nuumerical moodeling and d experimenntal results ccan provide enough datta
too calculate the convecttion self-noise and can be used to predict its value
v for anny specific configuratio
c on
o the MET
of T cell. In geeneral, the value
v of thee convection n induced self-noise
s ddepends on the Rayleiggh
n
number (Raa) of the ellectrolyte. ItI has beenn proven both experim mentally annd numerically that thhe
c
convection self-noise decreases as Ra deccreases. Low wer Ra coorresponds to a loweer electrolytte
c
concentratio on and smaller typical dimensions
d of the elem
ments of the transducer cell. On thee other handd,
R cannot be
Ra b too smalll because thhat would coorrespond to t a very smmall size forr the transdu ucer cell annd
Sensors 2013, 13 4588

very small concentration of the electrolyte, which usually means high hydrodynamic impedance (high
thermohydrodynamic noise) and low sensitivity of the transducer, respectively.

2.3.3. Geometry Self-Noise

The geometry self-noise is proportional to the electrochemical part of the transfer function of MET
transducer. To calculate the geometry self-noise, the following equation can be used:
4
(17)

where is the averaged electrochemical part of the transfer function, is the transducer conversion
coefficient from output noise current to input noise velocities, and is an empirical coefficient. The
geometry noise is the dominant one at very low frequency.

2.3.4. Shot Noise

The shot noise of the MET sensor is given by the following formula:
2
(18)

where is the quiescent current passing through the cell, is the transducer conversion coefficient,
is the absolute value of the charge passing through the electrode boundary in the elementary chemical
reaction on the electrode. In the electrochemical system usually used in MET transducer, equals 2 ,
where is the absolute value of the charge of the electron.

2.3.5. Electronic Self-Noise

Electronic self-noise of MET sensor comes from the signal conditioning electronics, including the
current to voltage converter and filters. The filters have unity gain in the proposed instrument pass
band and do not contribute to the electronic self-noise. The equivalent input-referred electronic noise
can be written as follows:
1 1
| | (19)

where is the resistor in the feedback of the operational amplifier used to convert the current into
voltage, is the impedance of the MET cell, and are voltage and current noise spectrum density
of the operational amplifier, respectively. To decrease the electronic self-noise, the transducer should
have a high impedance and conversion factor. Also, low-noise operational amplifiers should be used.
In all MET devices, the electronic self-noise contributes at relatively high frequencies.
Figure 3 shows the measured noise spectrum from a high-performance MET seismometer
(CME-6211) in comparison with two high-performance, industry-grade broadband seismometers, the
Streckheisen STS-2 and Trillium T240 units. The experiment is conducted at the Incorporated
Research Institutions for Seismology (IRIS), Program for Array Seismic Studies of the Continental
Lithosphere (PASSCAL) instrument center and EarthScope USArray array operations facility. The
S
Sensors 2013, 13 45889

nnoise perforrmance of the MET sensor

s is very similarr with thosee high-perfformance devices up to
t
p
periods of 600 seconds, demonstrati
d ing the feasibility of MET
M technollogy for brooadband seissmology.

Figuree 3. Noise performance of a ME ET seismommeter (CME E-6211) in compariso

on with
high-pperformancee, industry-grade brooadband seeismometerss. Red: MMET seism mometer
CME-6211; Bluee: Streckheeisen STS-22 broadban nd seismommeter; Green: Trillium
m T240
broadbband seismoometer.

3 Differentt Types of MET

3. M Motioon Sensor

Motion sensors
s baseed on MET technologyy include lin near and anngular accellerometers, rate sensorrs,
g
gyroscopes and seismoometers. Linnear MET sensors can be configurred for bothh horizontall and vertical
s
sensing, as shown
s in Figure 4(a annd b).
In convenntional ME ET sensors, platinum meshm and po
orous dielecctric are adhhesively bon
nded to formm
thhe electroddes and inssulator layers illustraated in Fig gure 1. In addition too the sensiing elemennt,
h
highly-flexib ble rubber membranes
m s are manuaally assembled at the twwo ends off the channeel to seal thhe
e
electrolyte, providing restoring
r foorce for thee mechaniccal subsysteem. Exampples of the performancce
p
parameters f commerrcially available MET linear
for l seism
mic sensors are
a listed inn Table 1. MTSS
M 1001 is
a single-axis, ultra-com mpact, highh gain, 1-Hzz seismic sensor. MTS SS 2003 coontains three orthogonal
o
oriented MT TSS 1001 seismic
s sennsors. It is probably th
he world sm mallest tri-aaxis 1-Hz seismomete
s er.
C
CME-6211 is a high performance
p e MET seissmometer, which
w is mentioned
m abbove. EP-300 is a verry
b
broadband s
seismometer r by eentec [26].
S
Sensors 2013, 13 45990

Figure 4. Schhematic of different

d typ
pe of MET motion
m senssor.

Table 1. Current performance parameters

p for linear MET
M seismicc sensors.
Perform
mance MTSS 1001
1 MTSS 2003
3 EP-300 CME-6
6211
Frequenncy range 1–300 Hz
H 1–300 Hz 0..0167–50 Hzz 0.0167––50 Hz
Sensor Noise √
50 ng/√ 50 ng/√ 100 ng/√ 10 ng/√√
Converrsion Factor 300 V/(m/s) 300 V/(m/s)) 2,,000 V/(m/s)) 2,000 V/(m/s)
V
Dynammic Range >110 dBB >110 dB 1550 dB at 1 H
Hz 135 dB
Non-linnearity <0.2% <0.2% N specified
Not <1%
Orientaation Any dirrection Any directio
on Any direction
Weightt 180 g 250 g 9..5 kg 7.5 kg
Temperrature “−40–+65 °C” “−40 – +65 °C” −12 – +55 °C”” −40 – +75 °C”
Power <80 mWW <250 mW <2250 mW <250 mW W

Figure 5 shows an assembled

a 3
3-axis linear MET seissmometer inncluding twwo orthogon nal horizontal
s
sensors and one verticaal sensor, aloong with thee sensing ciircuits.
Figure 4((c) shows thhe schematic of configuuration of a rotary METT sensor. Iff an angularr acceleratioon
iss applied as
a shown byy rounded arrow, the electrolyte flows through the sennsitive elem ment. This is
a
another appplication off the technnology. Currrently the MET anggular sensorr is probab bly the onlly
teechnology which com mmercially delivers higghly sensitiive rotationnal seismom meters with h a very low w
s
self-noise. Table 2 shhows the key perforrmances off current commercial
c lly availablle rotational
s
seismometer rs (METR-003 and R2) [27,28].
Sensors 2013, 13 4591

Figure 5. An assembled 3-axis linear MET seismometer including two orthogonal

horizontal sensors and one vertical sensor, along with the sensing circuits.

Table 2. Current performance parameters for MET rotational seismometers.

Performance METR-03 R2
Output angular rate angular rate
Noise at 1 Hz 8 5 10 / /√
10 /
/√
Full Scale Range 0.1 / 0.3 /
Bandwidth 0.033-100 Hz 0.033 -50 Hz
Operating Temperature range −40 – +75 °C −40 – +75 °C

The key advantages of MET sensors from other inertial sensors include, but are not limited to:
(1) The inertial mass is a liquid (electrolyte solution flowing through the transducer) and no
moving mechanics subject to wear out and possible damage, which makes the performance
more reliable and enables inherent ability to withstand high shock forces.
(2) The sensitivity of this sensor does not depend on the direction of sensitivity axis in space.
(3) High sensitivity and low self-noise at low and ultra-low frequency ranges or even DC with the
liquid inertial mass, properly selected parameters of the transducer and the proper structural design.
Despite the rather high output parameters obtained, the conventional MET devices developed and
produced at present have a number of faults that ultimately limit their application range. The main ones
are as follows:
(1) High cost of transducer manufacturing;
(2) High scatter of parameters of manually produced transducers resulting in the necessity of
individual tuning of the corresponding electronics for each sensor, which also increases the cost
of the device;
(3) Early decrease in sensitivity of the sensor in the high–frequency range;
S
Sensors 2013, 13 45992

(4) Veryy limited poossibilities too optimize the

t cell paraameters to decrease
d thee instrumentt self-noise..
These shoortcomings can be overrcome by inntroducing MEMS M micrrofabricationn techniquess to build thhe
M
MET sensinng element, as shown in i Figure 6 [18], which h reduces thhe MET celll size and produced thhe
innternal dimeensions closse to 1 µm. The use off MEMS alllows batch fabrication,, thereby low wer the cosst,
immproves thee sensitivityy and reprodducibility off the devicee, and has reeached 1 miicro Gee noise level.
Figure 7 details
d the faabrication prrocesses. A silicon wafeer with a low w-pressure cchemical vap por-depositeed
(LPCVD) siilicon nitridde (SiN) laayer on top forms the substrate of o the deviice. Four electrodes
e arre
1 nm/200 nm
10 n Ti/Pt layyers depositeed by an E-bbeam evapo orator and foollowed by llift-off. Each h electrode is
s
separated byy a 1 µm pllasma-enhannced chemiccal vapor-deeposited (PE ECVD) SiN N layer and branches
b to a
s
separate conttact pad useed for electrical connectiion to the ex
xternal circuuitry. Finallyy, focused ion
n beam (FIB B)
a deep reaactive ion ettching (DRIIE) are usedd to drill thee holes and etch backsidde silicon reespectively to
and t
f
form the chaannels. Optiical and SEM M images ofo the fabriccated single--50 µm-diam meter-channnel device arre
s
shown in Figgure 6(c andd d).

Figuree 6. Fabriicated MEM MS MET seismometeer. (a) asssembled M MET sensing cell,
(b) sennsor circuitss and protecction cell, (c)
( unpackagged MEMS S MET coree, and (d) Zooom-in
view ofo the MET T sensing element
e froom scanning electron microscopyy (SEM) sh howing
stackeed alternatinng platinum (electrodess) and silicon nitride (diielectric spaacers).

The self--noise specttrum densitty achieves 1.0 10 / /√ at 1 Hz for the cu urrent device,
a the senssitivity achiieves 400 / /
and [
[18]. Figuree 8 shows thhe spectrum
m density off a fabricateed
s
single-50 μm m diameterr-channel MEMS
M MEET seismom meter in coomparison with a com mmercializeed
s
seismometer r CMG-40T T (Guralp Syystems Ltd..).
S
Sensors 2013, 13 45993

Figure 7. Fabricaation processs of the ME

EMS MET seismomete
s er core.

Figuree 8. Spectruum densityy of a fabriccated singlee-50 μm diaameter-channnel MEMS

S MET
seismoometer in comparisonn with a commercialized seism mometer CMG-40T (Guralp
(
System
ms Ltd.). Reed—CMG-440T, Green n—MEMS MET M seism mometer.

Figure 4((d) is an altternative sennsing elemeent configurration with planar elecctrodes. This structure is
s
studied in [220] and it is shown thaat it has proomising perrformance comparable
c e with ring electrodes in i
o
open holes when
w the deevice size iss reduced too micro-scale. This plaanar electrodde structuree is very eassy
too fabricatee with staandard phootolithograpphy techno ology, offeering outsttanding prrecision annd
r
repeatabilityy. Furthermore, on the micro-scalee, the relativ ve flat and smooth surrface will prroduce mucch
leess disturbaance to the flow when there is eleectrolyte liq quid motion under exterrnal accelerration than is
f
found with the ring eleectrode struucture. This configuration can be applied to both linearr and angulaar
M
MET motionn sensors. AnA electrolyyte droplet--based low frequency micro-acceelerometer based b on thhis
p
planar electrrode structuure has beenn developedd [29]. Inspiired by the promising aand unique performancce
Sensors 2013, 13 4594

of MET and advantages of silicon-based planar microfabrication technique with small size, low cost,
outstanding precision and repeatability, the device employs a sub-microliter electrolyte droplet
encapsulated in oil as the sensing body and MET electrodes as read-out mechanism. Figure 9 illustrates the
device consisting of a planar four-electrode (anode-cathode-cathode-anode) MET cell located in a solid
rectangular housing channel on a silicon substrate with LPCVD silicon nitride. Four electrodes made
of 10 nm/100 nm Ti/Pt are deposited by E-beam evaporation and patterned with lift-off process. Then,
surface modification is performed utilizing a hydrophobic thin film coating on top of the electrodes
with proper patterning (hydrophilic spots surrounded by the hydrophobic areas). These hydrophilic
spots anchor the droplets and the surrounding hydrophobic area acts as feedback system allowing the
droplets to stabilize in the center hydrophilic spot in the case of high external acceleration input.
Before final assembly of the device, a 0.8 µL concentrated iodine-iodide electrolyte droplet and a small
amount of mineral oil are sequentially dispensed by micropipettes in the hydrophilic area covering
platinum electrodes. The oil not only prevents the electrolyte droplet from evaporating (maintaining
constant ion concentration), but also works as the elastic diaphragm to provide a restoring force.
Finally, the glass housing channel is assembled. Figure 10 shows the measured sensitivity frequency
response of a droplet based MET accelerometer with electrode width of 100 μ , and spacing of
30 μm. The device achieves sensitivity of 10.8 / 9.81 / at 20 Hz with nearly flat
response over the frequency range of 1 40 Hz and a low noise floor of 100 μ /√ at 20 Hz.
Furthermore, the novel idea of using oil film as sealing diaphragm eliminates the complicated
three-dimensional (3D) packaging used in both conventional and MEMS based MET sensors.

Figure 9. Schematic of the droplet-based MET accelerometer, (a) An overview, (b) Optical
image of the accelerometer core, (c) Zoom-in view of the droplet-covered electrodes.
Sensors 2013, 13 4595

Figure 10. Measured sensitivity frequency response of the droplet-based micro MET
accelerometer.

100
Measurement, h=100μm, d=30μm

10
Sensitivity (V/G)

0.1

0.01
1 10 100 1000
Frequency (Hz)

4. Conclusions

In conclusion, molecular electronic transducers comprising a simple set of four-electrodes, a liquid-state

electrolyte as inertial mass and housing show excellent ability to be applied in motion sensors, including
linear/angular accelerometers, gyroscopes and seismometers. The combination of high sensitivity,
large dynamic range, wide pass band, and low self-noise distinguishes MET sensors from conventional
MEMS motion sensors. The unique principles behind MET sensors also contribute to small size,
simple and low-cost fabrication, low power consumption, high shock sustainability and independence
of installation angle. Deployment of MEMS microfabrication in building MET core sensing element,
resulting in 1 µm internal dimension, improves the sensitivity and reproducibility of the device. The
performance of MET sensors can be easily optimized for a variety of applications by adjusting the
geometry or configuration of the sensing element, especially implementing MEMS. The mechanism of
MET provides a new paradigm of next generation high performance liquid-state motion sensor.

Acknowledgments

Researches, described in the present publication, was majorly supported by NASA-PIDDP-

NNX10AL25G (USA) and partly supported by Russian Foundation of the Basic Researches
(grants NN 11-07-00663-a, 12-07-12057-ofi_m, 12-07-33101-mol-a-ved). We sincerely thank
Matthew Fouch, Edward Garnero, Shane Ingate, Jon Oiler, John West and The Incorporated Research
Institutions for Seismology (IRIS) Program for Array Seismic Studies of the Continental Lithosphere
(PASSCAL) Instrument Center for the support for peer testing of the MET Seismometer with other
traditional seismometers.
Sensors 2013, 13 4596

References

1. Pike, W.T.; Standley, I.M.; Karl, W.J.; Kumar, S.; Semple, T.; Vijendran, S.J.; Hopf, T. Design,
Fabrication and Testing of a Micromachined Seismometer with Nano-G Resolution. In Proceedings of
the International Solid-State Sensors, Actuators and Microsystems Conference and Transducers,
Denver, CO, USA, 21–25 June 2009.
2. Levinzon, F.A. Ultra-low-noise seismic piezoelectric accelerometer with integral FET amplifier.
IEEE Sens. J. 2012, 12, 2262–2268.
3. Bernstein, J.; Miller, R.; Kelley, W.; Ward, P. Low-noise MEMS vibration sensor for geophysical
applications. J. Microelectromechan. Syst. 1999, 8, 433–438.
4. Hurd, R.M.; Lane, R.N. Principles of very low power electrochemical control devices. J. Electrochem.
Soc. 1957, 104, 727–730.
5. Wittenborn, A.F. Analysis of a logarithmic solion pressure detector. J. Acoust. Soc. Am. 1958, 30,
683–683.
6. Collins, J.L.; Richie, W.C.; English, G.E. Solion infrasonic microphone. J. Acoust. Soc. Am. 1964,
36, 1283–1287.
7. Larcam, C.W. Theoretical analysis of the solion polarized cathode acoustic linear transducer.
J. Acoust. Soc. Am. 1965, 37, 664–678.
8. Lidorenko, N.S.; Ilin, B.I.; Zaidenman, I.A.; Sobol, V.V.; Shchigorev, I.G. An Introduction to
Molecular Electronics; Enegoatomizdat: Moscow, Russia, 1984.
9. Kozlov, V.A.; Korshak, A.N.; Petkin, N.V. Theory of diffusion-type transducers for ultralow
electrolyte flow-rates. Sov. Electrochem. 1991, 27, 16–21.
10. Kozlov, V.A.; Safonov, M.V. Self-noise of molecular electronic transducers. Tech. Phys. 2003,
48, 1579–1582.
11. Kozlov, V.A.; Terent’ev, D.A. Transfer function of a diffusion transducer at frequencies
exceeding the thermodynamic frequency. Russ. J. Electrochem. 2003, 39, 401–406.
12. Zakharov, I.S.; Kozlov, V.A. Stationary convective diffusion and nonlinear effects in an
electrochemical transducer. Russ. J. Electrochem. 2003, 39, 397–400.
13. Panferov, A.P.; Kharlamov, A.V. Theoretical and experimental study of an electrochemical
converter of a pulsing electrolyte flow. Russ. J. Electrochem. 2001, 37, 394–398.
14. Reznikova, L.A.; Morgunova, E.E.; Bograchev, D.A.; Grigin, A.P.; Davydov, A.D. Limiting
current in iodine–iodide system on vertical electrode under conditions of natural convection.
Russ. J. Electrochem. 2001, 37, 382–387.
15. Kozlov, V.A.; Terent’ev, D.A. Frequency characteristics of a spatially-confined electrochemical
cell under conditions of convective diffusion. Russ. J. Electrochem. 2002, 38, 992–999.
16. Leugoud, R.J. Seismic Detection Using Mini Seismometer. In Proceedings of the Monitoring
Research Review: Ground-Based Nuclear Explosion Monitoring Technologies, Portsmouth, Virginia,
23–25 September 2008.
17. Abramovich, I.A.; Zhu, T. Next Generation Robust Low Noise Seismometer for Nuclear
Monitoring. In Proceedings of the Monitoring Research Review: Ground-Based Nuclear Explosion
Monitoring Technologies, Portsmouth, Virginia, 23–25 September 2008.
Sensors 2013, 13 4597

18. Huang, H.; Carande, B.; Tang, R.; Oiler, J.; Zaitsev, D.; Agafonov, V.; Yu, H. Development of A
Micro Seismometer Based on Molecular Electronic Transducer Technology for Planetary
Exploration. In Proceedings of the 26th IEEE International Conference on Micro Electro
Mechanical Systems (MEMS 2013), Taipei, Taiwan, 20–24 January 2013.
19. Kozlov, V.A.; Agafonov, V.M.; Bindler, J.; Vishnyakov, A.V. Small, Low-Power, Low-Cost
Sensors for Personal Navigation and Stabilization Systems. In Proceedings of the 2006 National
Technical Meeting of The Institute of Navigation, Monterey, CA, USA, 18–20 January 2006;
pp. 650–655.
20. Krishtop, V.G.; Agafonov, V.M.; Bugaev, A.S. Technological principles of motion parameter
transducers based on mass and charge transport in electrochemical Microsystems. Russ. J.
Electrochem. 2012, 48, 746–755.
21. Agafonov, V.M.; Nesterov, A.S. Convective current in a four-electrode electrochemical cell at
various boundary conditions at anodes. Russ. J. Electrochem. 2005, 41, 880–884.
22. Agafonov, V.M.; Krishtop, V.G. Diffusion sensor of mechanical signals: Frequency response at
high frequencies. Russ. J. Electrochem. 2004, 40, 537–541.
23. Sun, Z.; Agafonov, V.M. Computational study of the pressure-driven flow in a four-electrode
rectangular micro-electrochemical accelerometer with an infinite aspect ratio. Electrochimica Acta
2010, 55, 2036–2043.
24. Sun, Z.; Egorov, E.; Agafonov, V.M. The influence of the boundary condition on anodes for
solution of convection- diffusion equation with the application to a four-electrode electrochemical
cell. J. Electroanal. Chem. 2011, 661, 157–161.
25. Agafonov, V.M.; Zaitsev, D.L. Convective noise in molecular electronic transducers of diffusion
type. Tech. Phys. 2010, 55, 130–136.
26. EENTEC. Available online: http://eentec.com/EP-300_data.htm (accessed on 2 April 2013).
27. R-sensors, LLC. Available online: http://www.r-sensors.ru/ (accessed on 2 April 2013).
28. Leugoud, R. J.; Kharlamov, A.V. Second generation of a rotational electrochemical seismometer
using magnetohydrodynamic technology. J Seismologi. Soc. 2012, 16, 587–593.
29. Huang, H.; Liang, M.; Tang, R.; Oiler, J.; Ma, T.; Yu, H. An Electrolyte Droplet-Based Low
Frequency Accelerometer Based on Molecular Electronic Transducer. In Proceedings of the 17th
International Solid-State Sensors, Actuators and Microsystems Conference and Transducers 2013,
Barcelona, Spain, 16–20 June 2013.

© 2013 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).