IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 69, NO.

4, APRIL 2020 1387

Characterization of a PiezoMUMPs Microsensor

for Contactless Measurements of
DC Electrical Current
Carlo Trigona , Member, IEEE, Valentina Sinatra, Giuseppa Crea, Bruno Andò, Senior Member, IEEE,
and Salvatore Baglio, Fellow, IEEE

Abstract— This paper presents an integrated sensor able to resistance (AMR) solutions, as in [10], where coreless mag-
perform contactless measurements of dc electrical current. The netoresistive current sensors are discussed. The integrated
microsensor conceived is based on the PiezoMUMPs technology solution is also insensitive to external fields with a sensitivity
which embeds a layer of piezoelectric material thus offering a
mechanoelectric transduction of the measurand: the dc elec- of about 1 mV/A. A sensor bridge is required to convert
trical current flowing into a wire external to the microsensor the resistive output of the sensor. Giant magneto resistance
is here converted into a voltage signal without the need of (GMR) can be also used for sensing electrical quantities [11];
auxiliary polarization signals or integrated inductors/coils as in in particular, in [12], a current sensor is proposed that shows
the case electrostatic and electromagnetic devices, respectively. a resolution of about 0.6 A. It is noteworthy to mention
A prototype of the MEMS device has been designed, modeled,
numerically simulated, and then fabricated. A measurement that several more recent solutions have been presented in the
campaign allowed validating the sensor and to characterize its literature, such as Wu [13] presents a linear current sensor
performances. based on a piezoelectric material. In particular, the sensor
Index Terms— Contactless measurements, dc sensing, conceived in [13] is composed of a magnetic core made
microsensor, PiezoMUMPs, self-generating MEMS. by a piezoelectric layer sandwiched between two magnetic
layers. It senses the current flowing into a wire located inside
I. I NTRODUCTION the magnetic ring with sensitivity and resolution of about
1.52 mVpp /Arms and 1.6 Arms , respectively. The advantage

I N RECENT years, many research efforts have been focused

on the development of sensors and novel transduction meth-
ods for performing current measurements both at macroscale
of this device consists in the high quality of the output
signal without the need of sophisticated conditioning circuits.
Ripka et al. [14] propose a macroscale solution to detect static
and microscale [1]. It is in fact important to monitor the dc current where a current feedback strategy is implemented to
electricity consumption in many commercial and industrial improve the linearity of the transducer. The device proposed
applications, including solar inverter, Internet of Things, auto- in [14] presents higher sensitivity compared with previous
motive, electricity management systems, and home automation cases (about 100 mV/A) but it requires high power budget
measurements [2]–[4]. Several approaches have been proposed (about 0.5 W). Wang et al. [15] present a millimeter scale,
to perform indirect and contactless current measurement [5]; passive, dc electric current sensor based on a cantilever beam
in particular, noncontact current sensors can be considered the with a permanent magnet placed on the tip of the microme-
new generation sensor technology for nonintrusive load moni- chanical structure. The dc current which flows through the
toring (NILM) systems in several relevant environments as for wire, located in the close proximity of the sensor, induces a
example maritime vessels [6]. Traditional current sensors are displacement of the beam that in turn is converted into the
based on the Hall element placed in the air gap of a magnetic output voltage by the piezoelectric layer of the cantilever. The
core, and, typically, in order to improve the linearity, suitable device presents good performance in terms of both sensitivity
compensation procedures are implemented [7], [8]. These (about 80 mV/A) and linearity. The footprint of the device is
solutions find applications in a wide range of measurement about 4 mm × 4 mm, and the fabrication process requires a
that targets very different current values [9]. However, some postprocessing step for the deposition of a permanent magnet
limitations can be listed as the lack of operative range tuning in the top of a multispring beam. A downscaled prototype
and the difficulties in realizing MEMS scale current sensors. (MEMS) is presented in [16]: a silicon-based piezoelectric
This latter limitation is overcome with anisotropic magneto current sensor has been developed in order to monitor the
Manuscript received November 27, 2018; revised February 28, 2019; energy usage in an electric building. The device consists into
accepted March 4, 2019. Date of publication April 15, 2019; date of current a cantilever beam with a permanent magnet, used both as
version March 10, 2020. The Associate Editor coordinating the review process inertial mass and as sensitive element. In the presence of a
was Wuqiang Yang. (Corresponding author: Carlo Trigona.)
The authors are with the Dipartimento di Ingegneria Elettrica Elettronica static or variable current, flowing into a wire in the close
e Informatica (DIEEI), University of Catania, 95125 Catania, Italy (e-mail: proximity of the sensor, the magnet couples to the magnetic
carlo.trigona@dieei.unict.it). field produced by the current resulting into a force applied
Color versions of one or more of the figures in this article are available
online at http://ieeexplore.ieee.org. to the cantilever that induces vibrations at the frequency of
Digital Object Identifier 10.1109/TIM.2019.2908510 the input signal. The main advantages of the proposed device
0018-9456 © 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.

Authorized licensed use limited to: University of Catania. Downloaded on April 05,2023 at 15:01:00 UTC from IEEE Xplore. Restrictions apply.
1388 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 69, NO. 4, APRIL 2020

concern its reduced dimension and the fabrication process moving particle, and this behavior is modeled by the following
that does not include active materials to convert the measured equation:
current into an output voltage. −
→
However, a postprocessing procedure has to be accounted to FL = q · ( E + v × B)
 (1)
deposit a magnetic layer to implement the coupling between where q is the charge of the particle, E is the electric field, v
the MEMS device and the external current. Sensitivity and is the velocity of the particle, and B is the magnetic induction.
resolution have been estimated to about 2 μm/A and about Here, we assume that E is absent and an electrical current I
0.5 A at 50 Hz, respectively. In [17], an integrated U-shaped is driven into an electrical wire. Taking also into account that
has been presented that detects constant electrical currents with current I can be described as
an optical readout. The sensor is able to measure current-flux
density produced by using the Lorentz force principle; in I = λ
v (2)
particular, a measurement range of about 2.2 mA·T has been
in which λ is a free charge having velocity v, in a section dl,
achieved by reading the peak light intensity. In order to convert
the total charge dq can be expressed as
the output of the sensor into a voltage signal, an external
active conditioning circuit is required. In [18], a U-shaped dq = λdl (3)
magnetic field sensor based on the CMOS technology has
been presented. The device can be also used to measure cur- and hence the expression of the Lorentz force becomes
rents, and it uses a resistive readout (integrated strain gages), d F = dq v × B = λdl v × B.
 (4)
as consequence an integrated conditioning circuit has been also
developed by the authors in order to convert the mechanical The net force can be found by integrating (4)

displacement into a voltage signal.
In this paper, we propose a novel integrated piezoelectric F = dl I × B = I L B (5)
microsensor based on the PiezoMUMPs technology [19],
designed and fabricated with applications to the field of where L is the length in which I and B are perpendicular.
resonant sensors for contactless measurements of electrical In detail, this sensor is based on the interaction between
currents. It should be noted that few studies have been found an unknown electrical current I that produces the magnetic
in the literature to deal with the development of microma- induction B and a known current Icant driven through a silicon
chined sensors based on the PiezoMUMPs technology. In this cantilever beam. The Lorentz force acting on the cantilever can
specific case, this paper builds up and extends the results be, therefore, written as follows:
  
presented in [20], where some preliminary results, concerning μ 
I
FL = ( Icant × B) · L = Icant × 0
a microsensor for the contactless measurement of currents, · ·L
2π r
have been presented. This paper reports the current stage of
μ0
an activity that has been developed in several years. The = · Icant · I · L · sin θ. (6)
major contribution to the community of instrumentation and 2πr
measurements consists in the demonstration, from a theoretical The wire with the unknown current is placed at a distance
and experimental point of view, of the possibility to use an r from the cantilever. The cantilever considered in this paper
integrated fabrication process to realize a microscale measure- consists in a U-shaped beam. The Lorentz force in (6) deforms
ment system able to detect dc current with several intriguing the microelectromechanical structure, and the resulting dis-
advantages, such as: tunable range; a simple readout (without placement of the cantilever tip will be a function of the current
the need of additional conditioning circuit); a mechanically to be measured.
robust structure (U-shaped device); contactless working mode It is interesting to observe that the oriented indica-
(the measurements can be performed at few millimeter of dis- tion of the unknown current and of the electrical current
tance from the electrical wire where the current flows); ability into the MEMS device directs the Lorentz force, as illus-
to generate an electrical voltage via the adoption of an embed- trated in Fig. 1(a). Moreover, the U-shaped beam cantilever
ded aluminum nitride (AlN) piezoelectric layer without the exhibits a dynamic behavior which can be depicted using the
need of postprocessing deposition. Interesting performances well-known equivalent mass-spring system equation [21]:
have been obtained in terms of operative range (adjustable), in detail, this system can be modeled with the second-order
noise level (∼22.6 μV), and resolution (≈0.3 A). The design differential equation
concepts of the microsensor are presented here together with
m z̈ + d ż + V (t) + kz = FL (t). (7)
the validation of the working principle through the analytical
model and FEM simulations. Moreover, several experimental It is essential to highlight here that the electrical current to
results and characterization diagrams are reported that, when be evaluated is converted in a force, then into a displacement
compared with the theoretical model, show the suitability of z of the cantilever beam free end, and finally into a voltage
the proposed approach. through the use of a piezoelectric material composing the
MEMS device [22] and whose behavior can be described as
II. W ORKING P RINCIPLE AND M ODELING follows:
  
The current sensor presented in this paper exploits the
Lorentz force exerted by magnetic induction on a charged V (t) = ż − V (t)/RC . (8)

Authorized licensed use limited to: University of Catania. Downloaded on April 05,2023 at 15:01:00 UTC from IEEE Xplore. Restrictions apply.
TRIGONA et al.: CHARACTERIZATION OF A PiezoMUMPs MICROSENSOR 1389

TABLE I
L IST OF U SED PARAMETERS

Fig. 1. (a) Working principle. (b) Focus on the geometrical parameters.

In addition, in order to complete the model and to match it with

the experimental results, another equation has been introduced

Ṽ (t) = V (t) + α (9)

in which an offset, indicated with α and expressed in volt, has

been added to the simulated output voltage, V (t), since the
model does not take into consideration the noise level.
In relation to (6), B is the magnetic induction generated
by a steady current I , that is the dc electrical current to be
measured, in according to the Bio–Savart law; r is the distance 2 E 2 (Te − Ti )
Icant = × = × CAl × m Al ×
between the electrical wire in which I is passing and the tip Rbeam t Rbeam t
sensor, μ0 is the vacuum permeability, Icant is the current (11)
injected into the MEMS device itself, and L is the top beam
where Rbeam = 0.87  is the input resistance, and CAl =
length [l4 shown in Fig. 1(b)]. θ is the angle between the
0.9 × 103 J/(kg·◦ C) and m Al = 2.56 × 10−9 kg are the specific
magnetic field vector and the beam length vector, here we
heat and the mass of the aluminum (Al) layer, respectively,
assume θ = 90◦ and, therefore, sin(θ ) = 1. The amplitude of
Te = 660 ◦ C and Ti = 25 ◦ C are the aluminum melting
the beam deflection z depends on the Lorentz force amplitude,
point and the room temperature, at last t has been fixed
which is proportional to the product between Icant and B.
at 15 s. It is invaluable to underline that t has been chosen
Furthermore, current Icant , generated into the U-shaped beam
in order to analyze the spectrum of the output signal to low
cantilever, can be either a dc or ac current: in this latter case,
frequency. The inertial mass has been determined considering
if the micromechanical structure is exposed to an excitation
the following equation:
source with a frequency equal to its resonant frequency, as a
consequence, the beam will have the maximum deflection m = ρSi × V = ρSi × tSibulk × l2 × l5 (12)
at its tip. For this reason, in general, an ac current, having
in which ρSi = 2329 kg/m3
is the silicon density, whereas the
a specific condition in terms of frequency and amplitude,
other parameters are indicated in Section III.
is adopted preferably in order to obtain a more sensitive output
In relation to the parameters determined experimentally,
response. In relation to (7), m is the inertial mass of the
output capacitor C has been calculated through an LCR
microelectromechanical structure, d represents the damping
meter, in particular Automatic LCR Meter 4225 Wayne Keer.
coefficient,  is the damping term caused by the piezoelectric
As regards the output resistance R, it has been approximated
transduction, and k is the mechanical stiffness. Furthermore,
at 1 M, resistance of the probe. The mechanical damping
FL (t) represents the Lorentz force, whereas z̈, ż, and z are
[23], d, and the piezoelectric damping [24], , have been
the acceleration, the velocity, and the displacement ofthe
determined by the literature.
cantilever tip, respectively. Eventually, in relation to (8), is
Finally, in relation to (9), α parameter has been estimated
the coupling constant and RC is the load of the piezoelectric
experimentally, and it corresponds with the noise level mea-
stack. In addition, V (t) and V̇ (t) represent the voltage and its
sured.
first derivative in the output of the piezoelectric layer. In detail,
All parameters listed in Table I allowed implementing the
all parameters describing the previous equations are listed
model described by (6)–(9), as shown in Fig. 2. It is important
below and their values, originated by the model, experiment,
to underline that the simulated output voltage values have
and literature, are indicated in Table I.
been carried out extracting the peak value of the fast Fourier
In order to compute the current into the cantilever, Icant ,
transform (FFT) pertinent to the sinusoidal output voltage in
the power expression has been considered, and in particular,
the range [0–8 A].
the assumption of equality between mechanical and thermal
powers has been imposed
III. D ESIGN AND FABRICATION
E 1 The device has been fabricated by using the PiezoMUMPs
P = = Rbeam Icant
2
(10)
t 2 technology [19]. In this process, a silicon-on-insulator (SOI)

Authorized licensed use limited to: University of Catania. Downloaded on April 05,2023 at 15:01:00 UTC from IEEE Xplore. Restrictions apply.
1390 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 69, NO. 4, APRIL 2020

TABLE II
D EVICE PARAMETERS

Fig. 2. Simulated behavior of the current sensor.

Fig. 3. Cross-sectional view of the PiezoMUMPs technology [19], [25] (not

in scale).

Fig. 5. (a) Geometry in Comsol Multiphysics. (b) Focus on the suspended

structure.

and widths have been designed, and some of them have already
been characterized [19].
In this paper, we have focused our attention to the device
circled in Fig. 4, whose parameters are indicated in Fig. 1(b)
Fig. 4. Layout of device. and listed in Table II. By using the equivalent section method
[26], [27], the elastic constant k and afterward the natural
frequency f n have been calculated as k = 5.74 N/m and f n =
wafer is used as substrate with specific characteristics in 1.39 kHz, respectively. In addition, the theoretical resistance
terms of thickness. After having doped the top surface of beam, Rbeam of 0.87 , has been estimated taking account
the Si layer and grown a thermal oxide with a thickness into R = 0.055 /Sq.
of 200 nm, a piezoelectric layer in AlN of 0.5 μm is deposited.
At last, a metal stack composed of 20-nm chrome and
IV. N UMERICAL A NALYSIS
1000-nm aluminum is set down. The etching process is
performed with deep reactive ion etch (RIE) and described In this section, finite-element model simulations imple-
in Fig. 3. mented in Comsol Multiphysics environment (5.1 Version)
The AlN layer is used to generate an electric output: this and related to the U-shaped beam cantilever are presented in
aspect represents the most important advantage in this process, order to obtain a closer view of the device behaviors in terms
because an electric signal is directly available in the output of of the displacement and the stress. In Fig. 5(a), the device
the sensor. In this manner, the sensor output is not correlated geometry is illustrated where layers of thin silicon, AlN and
with the variation of resistance (resistive output) or capacitance aluminum for the suspend structure [shown in Fig. 5(b)], and
(capacitive output). thick silicon for the inertial mass are considered. In order to
The layout of the sensor has been realized by using proceed with the FEM analysis, an appropriate mesh has been
MEMSPro as the CAD design tool. The layout of the whole realized [see Fig. 6(a)] contemplating different dimensions of
die is shown in Fig. 4. Several devices having different lengths the layers and of the inertial mass, and especially it has been

Authorized licensed use limited to: University of Catania. Downloaded on April 05,2023 at 15:01:00 UTC from IEEE Xplore. Restrictions apply.
TRIGONA et al.: CHARACTERIZATION OF A PiezoMUMPs MICROSENSOR 1391

Fig. 6. (a) Mesh of the designed geometry in Comsol Multiphysics.

(b) Focus on the critical points. Fig. 8. Cantilever tip displacement as a function of f 1 = 1.38 kHz.

Fig. 7. Displacement as a function of the frequency. Fig. 9. (a) von Mises stress as a function of f 1 = 1.38 kHz. (b) Focus on
the most stressed points.

tightened in correspondence with the critical points which are

the corners, as depicted in Fig. 6(b).
In order to reduce the computational burden, the whole
geometry has been divided in two domains: the first includes
the three layers composing the U-shaped beam cantilever and
the second one is related only to the inertial mass. An opti-
mized mesh has been obtained, indeed its implementation has
been densed along the layers (in this case, a fitted mesh
has been considered), whereas it has been approximate in
correspondence with the inertial mass. (In this case, a normal
mesh has been imposed.) In detail, a free mesh has been Fig. 10. Cantilever tip displacement as a function of f 3 = 11.806 kHz.
chosen, and it consists of 44 519 elements of which the
tetrahedral (tets) mesh elements are 28 844, the triangular mesh In particular, the first, third, and fifth oscillation modes
elements are 14 163, the mesh edges are 1471, and eventually, have been estimated, and for each value of frequency, the
the mesh vertices are 41. displacement and the von Mises stress, expressed in [m]
Afterward, an analysis in the frequency domain has been and in [N/m2 ], respectively, have been plotted. The eigen
selected in order to estimate the natural frequency and to frequency analysis has returned the following frequencies:
confirm the value previously indicated by the model. In Fig. 7, f 1 = 1.38 kHz for the first mode, f 3 = 11.8 kHz for the third
the cantilever tip displacement as a function of the frequency mode, and at last f 5 = 43 kHz for the fifth mode. In Fig. 8,
is displayed; it is worth to underline that the two natural the cantilever tip displacement in correspondence with f1 is
frequencies, the first determined with the equivalent section shown, whereas the von Mises stress and a focus on the most
method and the second simulated in Comsol Multiphysics, stressed points are illustrated in Fig. 9(a) and (b).
are corresponding. The analysis in the frequency domain has The largest displacement is observable on the cantilever
been carried out adopting a Lorentz force, having an amplitude tip with an amplitude of 281 nm in the z-direction, and the
of 30 nN, applied to the cantilever tip in the z-direction; this maximum von Mises stress is obtained in correspondence with
Lorentz force value has been imposed using parameters values the anchors and it has a value of 410 kN/m2 .
indicated earlier. In relation to the third oscillation mode with f 3 = 11.8 kHz,
In addition, an eigen frequency analysis, considering the the highest displacement is present on the cantilever tip with a
first six frequencies, has been realized in order to evaluate value of 32.8 nm (as illustrated in Fig. 10), whereas in Fig. 11,
the oscillation modes of the U-shaped beam cantilever. This the von Mises stress with a maximum value of 1.96 MN/m2
analysis acquires a relevant significance since the performance is shown.
of a piezoelectric material is correlated with the velocity, In this case, the micromechanical structure is char-
as indicated previously in (8). Symmetry in the structure has acterized by the highest displacement in correspondence
been exploited such to avoid even-order oscillation modes. with the inertial mass; however, a minor contribution

Authorized licensed use limited to: University of Catania. Downloaded on April 05,2023 at 15:01:00 UTC from IEEE Xplore. Restrictions apply.
1392 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 69, NO. 4, APRIL 2020

Fig. 11. von Mises stress as a function of f 3 = 11.806 kHz.

Fig. 14. von Mises stress as a function of f 5 = 43 kHz.

Fig. 12. (a) Critical section in relation to the von Mises stress at f 3 = Fig. 15. Die and details on the U-shaped beam cantilever.
11.806 kHz. (b) Zoomed-in view of the critical points.

mode; nevertheless, the maximum value is lower and equal to

102 kN/m2 .
It is extremely fascinating to note that the whole dis-
placement ranges from [0.77–561 nm] considering the three
oscillation modes. However, taking into account that the output
voltage in a piezoelectric material is correlated with the
frequency, a tradeoff with the displacement is essential; for
this reason, judging the displacement in the fifth oscillation
mode comparable with a noise level, the contribution of
the fifth mode has been neglected, and the characterization
of the microelectromechanical structure has been performed
considering only the first mode at frequency f 1 = 1.38 kHz
Fig. 13. Cantilever tip displacement as a function of f 5 = 43 kHz.
and the third mode at frequency f3 = 11.806 kHz.

is observable long the arms too. This is because the oscil- V. E XPERIMENTAL S ETUP
lation mode is changed involving, in a curve, the central An appropriate experimental setup has been implemented
section of the arms (shown in Fig. 11). In relation to the in order to characterize the Lorentz force sensor. (The die
von Mises stress, it is appropriate to observe that the critical is shown in Fig. 15.) The die has dimensions 11.15 mm ×
points are changed due to the variation of the oscillation mode 11.15 mm; however, the active area, in which all devices have
and the overall value of the von Mises stress is increased. been realized, is equal to 9.65 mm × 9.65 mm. Furthermore,
Because of this, a focus about it has been considered necessary the PCB board, in which the die is packaged and bonded, has
and illustrated in Fig. 12(a) and (b). dimension 50 mm × 48 mm.
Finally, the fifth oscillation mode, having a frequency of The study of the U-shaped beam cantilever is focused on
f 5 = 43 kHz, has been considered and the following two the characterization as current sensor; therefore, an electrical
trends (see Figs. 13 and 14) in terms of the displacement wire, having a diameter of 0.15 mm, has been placed at the
and the von Mises stress have been determined. In detail, distance of r = 1 mm from the device, and the dc current to be
in Fig. 13, the displacement affects essentially the central measured is passed through it (Fig. 16) by using a controlled
section of the U-shaped beam cantilever and it presents a reference voltage source. A signal generator is used to drive
maximum value of 0.77 nm. Furthermore, the curve of arms the current through the MEMS sensor to obtain the Lorentz
is similar to the previous case (as illustrated in Fig. 14); force. The whole experimental setup is presented in Fig. 17,
however, the variation of the position in the inertial mass is where an Agilent 33220A function waveform is used to drive
neglectful because a lower displacement is impressed to arms. a sinusoidal current into the device and an Infiniium
As regards the fifth oscillation mode, the von Mises stress MSO9064A oscilloscope is needed to monitor the output
exhibits the same critical points in comparison with the third signal (Vout ). A dc current has to be measured; hence,

Authorized licensed use limited to: University of Catania. Downloaded on April 05,2023 at 15:01:00 UTC from IEEE Xplore. Restrictions apply.
TRIGONA et al.: CHARACTERIZATION OF A PiezoMUMPs MICROSENSOR 1393

Fig. 19. Experimental comparison between the first and the third oscillation
Fig. 16. Packaged and bonded die with the electrical wire located at the modes.
distance of 1 mm from the U-shaped beam cantilever.

Fig. 20. Output signal (Vout ) in logarithm as a function of the logarithm of

the frequency.

ISO-TECH IDM 101. Afterward, the experimental mechan-

ical frequencies of the U-shaped beam cantilever have been
Fig. 17. Experimental setup. estimated: in detail, this has been done by evaluating the
root-mean-square (rms) value of the output voltage signal for
different frequency values oscillating around the theoretical
values. The two experimental mechanical frequencies have
been identified at f 1 = 1.4 kHz for the first mode and at
f 3 = 10.3 kHz for the third mode. The next step is concerned
about the experimental comparison between two oscillation
modes evaluating the voltage in the output of the piezoelectric
layer; for this reason, the same operative conditions have been
Fig. 18. Focus on the input sinusoidal current into the MEMS device. repeated, or rather, a sinusoidal current driven into the MEMS
device having an amplitude of 30 mAPP and a dc electrical
a power supply, specifically K.E.R.T. System 420 Professional, current into the electrical wire with a range [0–8 A]. In Fig. 19,
is employed to force a known dc current into an electrical wire, the output voltage as a function of the current passing through
located at 1 mm of distance from the sensor. the electrical wire and varying into the range [0–8 A] is shown
Moreover, in order to validate the amplitude of the sinu- for each excitation frequency: it is intriguing to observe that
soidal current impressed to the prototype for preserving the the third oscillation mode (in green) exhibits a better response
MEMS device, a digital multimeter ISO-TECH IDM 101 is in terms of sensitivity than the first one (in red).
used. Indeed, a theoretical calculation has been carried out and These results allowed explaining and to verify experimen-
a maximum current value of 15 mAP has been determined. For tally why the authors have decided to analyze both modes:
this reason, the function waveform has been set in order to indeed, although the first oscillation mode produces the highest
generate a sinusoidal voltage having an amplitude of 20 VPP , displacement, the output voltage of the piezoelectric stack
and a resistor having a value of 680  has been connected is correlated with the velocity too [as described in (8)];
in series with the input device (see Fig. 18): in this way, therefore, a tradeoff between two parameters, displacement
a sinusoidal current having an amplitude of 30 mAPP is driven and frequency, has to be found in order to achieve the
into the U-shaped beam cantilever. best performance. This condition is fulfilled with the third
oscillation mode, and for this reason, the authors have chosen
VI. R ESULTS to continue the sensor characterization examining in depth the
In this section, the results achieved analyzing the first and third oscillation mode exclusively.
the third oscillation modes are illustrated. In particular, to be thorough, in relation to the experimental
At the beginning, the experimental resistance beam, Rbeam mechanical frequency in the third oscillation mode, some of
of 2.3 , has been estimated through the digital multimeter the experimentally measured data are reported in Fig. 20,

Authorized licensed use limited to: University of Catania. Downloaded on April 05,2023 at 15:01:00 UTC from IEEE Xplore. Restrictions apply.
1394 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 69, NO. 4, APRIL 2020

follows:
 

10
Mean Value = X n /10 (13)
n=1
where X n represent the measurements for each value of
current.
The sensor current presented in this paper is a tunable
device: it is sufficient to change the amplitude of the sinusoidal
current driven into the U-shaped beam cantilever (fixed in this
Fig. 21. Output voltage of the microsensor with all ten measurements and paper case to 30 mAPP ) to modify the operative range; indeed,
their dispersion for each dc current. in according to (6), a variation of the current into the MEMS
introduces a shift of the operating region in order to keep
constant the Lorentz force, FL . This characteristic represents
one of the most remarkable advantages of this sensor, making
it opposite to other sensor topologies, for instance, the Hall
effect sensors, where the operative range is established and
intrinsic in the structure.
In order to conclude the results section, it can be appreciable
to note the following.
1) The natural frequency, f n , has been estimated in
three different environments verifying a match among
Fig. 22. Transduction diagram of the microsensor.
the modeling ( f n = 1.39 kHz), the numerical
( f n = 1.38 kHz), and the experimental ( f n ≈ 1.4 kHz).
where the output signal (Vout ), expressed in logarithm, is
2) The frequency of the third oscillation mode has been
shown as a function of the frequency of the bias current
determined numerically ( f 3 = 11.806 kHz) and con-
driven through the U-shaped beam cantilever. The amplitude
firmed experimentally ( f 3 = 10.3 kHz).
of the bias current is kept constant to 30 mAPP , whereas
3) The mechanical stiffness has been calculated in all
the external dc current in the electrical wire (I ) is fixed to
three approaches finding k = 5.74 N/m in the model,
4 A. Therefore, in order to characterize the working region
k = 5.60 N/m in the FEM analysis, and k = 5.766 N/m
of the device with the highest responsivity of the MEMS
in the experiment.
sensor, a sinusoidal current, with an amplitude of 30 mAPP
and a frequency of 10.3 kHz, is applied to the U-shaped beam
cantilever, while a dc current in [0–8 A] range is imposed in VII. C ONCLUSION
the electrical wire (I ). It is interesting to highlight that for each An integrated microsensor for measuring dc electrical cur-
dc current value, ten measurements are collected, and, later, rent based on the Lorentz force has been presented in this
the mean value is determined. In Fig. 21, the output voltage paper. The PiezoMUMPs technology has been used for the
of the microsensor is illustrated having a quasi-linear trend in fabrication of a U-shaped beam microcantilever. MEMSPro
all the range [0–8 A]. has been the CAD tool used for the device layout design.
The same graph is also shown including the uncer- The most interesting advantage is represented by the pos-
tainty band of 3σ shown in Fig. 22, where the σ value sibility, offered by the piezoelectric material embedded in
of 42.9 × 10−6 V has been estimated and a responsivity of the PiezoMUMPs technology, to convert the measurand in an
80.8 × 10−6 V/A has been evaluated. In accordance with electric signal straight available in the output of the sensor. For
the regulation UNI CEI ENV 13005, the type-A uncertainty this reason, since the output voltage in a piezoelectric material
has been estimated experimentally, and a maximum value of is correlated with the frequency, an FEM analysis has been
8.3 × 10−6 V has been evaluated. carried out in order to analyze the first three odd oscillation
In order to characterize the device in terms of resolution a modes: the first and the third modes have produced promising
noise level, specified in standard deviation and expressed in results because a tradeoff with the displacement is invaluable,
volt has been estimated experimentally and a value of 22.6 μV whereas the fifth mode has been rejected seeing as how the
has been carried out. Therefore, it has been possible to deter- displacement comparable with a noise level. Initially, a com-
mine the resolution in ampere considering the ratio between parison between the two frequencies evaluated experimentally
the noise level [V] and the responsivity [V/A] and obtaining and estimated at f 1 = 1.4 kHz and f 3 = 10.3 kHz in terms of
a resolution value of 0.28 A. sensor response allowed selecting the third oscillation mode
A good matching between the model and the experimental in order to obtain the best performance.
results has been obtained; in fact, the simulated output voltage The current sensor presented here is able to adapt the opera-
values have been overlapped to the transduction diagram tive range of the dc electrical current to be measured changing
shown in Fig. 22, thus verifying the validation of the model. the amplitude of the sinusoidal current driven into the micro-
Mean value represents the average value of the ten measure- mechanical structure. Considering an amplitude of 30 mAPP
ments collected for each dc current values and expressed as related to the current into the U-shaped beam cantilever an

Authorized licensed use limited to: University of Catania. Downloaded on April 05,2023 at 15:01:00 UTC from IEEE Xplore. Restrictions apply.
TRIGONA et al.: CHARACTERIZATION OF A PiezoMUMPs MICROSENSOR 1395

operative range [0–8 A] has been identified. Furthermore, [19] C. Trigona, A. Algozino, F. Maiorca, B. Andò, and S. Baglio, “Design
the working principle, the model, and the transduction mech- and characterization of PiezoMUMPs microsensors with applications
to environmental monitoring of aromatic compounds via selective
anism have been validated, and the results obtained encourage supramolecular receptors,” Procedia Eng., vol. 87, pp. 1190–1193,
pursuing efforts to optimize the device design in order to Jan. 2014.
improve the performances. Further research efforts are ongoing [20] C. Trigona, V. Sinatra, G. Crea, R. Nania, B. Andò, and S. Baglio,
“PiezoMUMPs microsensor for contactless measurements of DC-
in order to develop an optimized MEMS current sensor with electrical current,” in Proc. IEEE Int. Instrum. Meas. Technol. Conf.
the piezoelectric transduction mechanism for measuring both (I2MTC), May 2018, pp. 1–5.
ac and dc current flows. [21] J.-S. Wu and D.-W. Chen, “Dynamic analysis of a uniform cantilever
beam carrying a number of elastically mounted point masses with
dampers,” J. Sound Vib., vol. 229, no. 3, pp. 549–578, Jan. 2000.
[22] F. Cottone, H. Vocca, and L. Gammaitoni, “Nonlinear energy harvest-
ACKNOWLEDGMENT ing,” Phys. Rev. Lett., vol. 102, Feb. 2009, Art. no. 080601.
[23] B. Andò, S. Baglio, C. Trigona, N. Dumas, L. Latorre, and
The authors would like to thank A. Russo, who collaborates P. Nouet, “Nonlinear mechanism in MEMS devices for energy harvesting
with the National Institute of Nuclear Physics (INFN), Catania, applications,” J. Micromech. Microeng., vol. 20, no. 12, Nov. 2010,
Art. no. 125020.
Italy, in setting up the finite-element analysis. [24] B. Andó, S. Baglio, F. Maiorca, and C. Trigona, “Analysis of two dimen-
sional, wide-band, bistable vibration energy harvester,” Sens. Actuators
A, Phys., vol. 202, pp. 176–182, Nov. 2013.
R EFERENCES [25] MEMS Process. Accessed: Jul. 20, 2018. [Online]. Available:
http://www.memscap.com/products/mumps/piezomumps
[1] K. Iniewski, Smart Sensors for Industrial Applications. Boca Raton, FL, [26] X.-T. He, S.-L. Chen, and J.-Y. Sun, “Applying the equivalent
USA: CRC Press, ch. 21, 2016. section method to solve beam subjected to lateral force and bending-
[2] G. Geng, J. Wang, K.-L. Chen, and W. Xu, “Contactless current compression column with different moduli,” Int. J. Mech. Sci., vol. 49,
measurement for enclosed multiconductor systems based on sensor no. 7, pp. 919–924, Sep. 2007.
array,” IEEE Trans. Instrum. Meas., vol. 66, no. 10, pp. 2627–2637, [27] B. Andó, S. Baglio, G. L’Episcopo, V. Marletta, N. Savalli, and
Oct. 2017. C. Trigona, “A BE-SOI MEMS for inertial measurement in geo-
physical applications,” IEEE Trans. Instrum. Meas., vol. 60, no. 5,
[3] L. Grno, “Magnetic fiber integrator and its application in precision
pp. 1901–1908, May 2011.
noninvasive DC current sensor,” IEEE Trans. Instrum. Meas., vol. 66,
no. 8, pp. 2012–2021, Aug. 2017.
[4] J. S. Donnal and S. B. Leeb, “Noncontact power meter,” IEEE Sensors J.,
vol. 15, no. 2, pp. 1161–1169, Feb. 2015.
[5] D. Lawrence, J. S. Donnal, S. Leeb, and Y. He, “Non-contact measure-
ment of line voltage,” IEEE Sensors J., vol. 16, no. 24, pp. 8990–8997,
Dec. 2016.
[6] P. Lindahl, G. Bredariol, J. Donnal, and S. Leeb, “Noncontact electrical Carlo Trigona (M’16) received the M.S. degree in
system monitoring on a U.S. Coast guard cutter,” IEEE Instrum. Meas. automation engineering and control of complex sys-
Mag., vol. 20, no. 4, pp. 11–20, Aug. 2017. tems and the Ph.D. degree in electronic, automation
[7] P. Ripka, Magnetic Sensors and Magnetometers. London, U.K.: Artech and control of complex systems from the University
House, 2001. of Catania, Catania, Italy, in 2006 and 2009, respec-
[8] M. Crescentini, M. Marchesi, A. Romani, M. Tartagni, and tively.
P. A. Traverso, “A broadband, on-chip sensor based on Hall effect for From 2010 to 2011, he was a Post-Doctoral
current measurements in smart power circuits,” IEEE Trans. Instrum. Researcher and a Lecturer with the Laboratoire
Meas., vol. 67, no. 6, pp. 1470–1485, Jun. 2018. d’Informatique, de Robotique et de Microélec-
[9] S. D. T. Dewi, C. Panatarani, and I. M. Joni, “Design and development tronique de Montpellier, University Montpellier II,
of DC high current sensor using Hall-effect method,” in Proc. AIP Conf. Montpellier, France. From 2011 to 2017, he was a
Proc., 2016, vol. 1712, no. 1, Art. no. 030006. Post-Doctoral Researcher and a Lecturer with the Dipartimento di Ingegneria
[10] H. Beltran, C. Reig, V. Fuster, D. Ramírez, and M. D. Cubells-BeltrÁn, Elettrica Elettronica e Informatica (DIEEI), University of Catania. From
“Modeling of magnetoresistive-based electrical current sensors: A tech- 2017 to 2018, he was a Post-Doctoral Researcher with the Chemnitz Univer-
nological approach,” IEEE Sensors J., vol. 7, no. 11, pp. 1532–1537, sity of Technology, Chemnitz, Germany. He is currently an Assistant Professor
Nov. 2007. of electronic instrumentation and measurements with DIEEI. He has coau-
[11] F. Xie, R. Weiss, and R. Weigel, “Giant-magnetoresistance-based galvan- thored more than 130 scientific publications, which include chapters in books,
ically isolated voltage and current measurements,” IEEE Trans. Instrum. papers in international journals, and proceedings of international conferences.
Meas., vol. 64, no. 8, pp. 2048–2054, Aug. 2015. His current research interests include sensors, transducers, microsystems,
[12] O. Yong, J. He, J. Hu, and S. X. Wang, “A current sensor based microsensors, fluxgate magnetometers, and energy harvesting.
on the giant magnetoresistance effect: Design and potential smart grid Dr. Trigona received the National Scientific Habilitation to be an Associate
applications,” Sensors, vol. 12, no. 11, pp. 15520–15541, 2012. Professor in 2017.
[13] Z. Wu, “A wide linearity range current sensor based on piezoelectric
effect,” IEEE Sensors J., vol. 17, no. 11, pp. 3298–3301, Jun. 2017.
[14] P. Ripka, J. Kubik, M. Duffy, W. G. Hurley, and S. O’Reilly, “Current
sensor in PCB technology,” IEEE Sensors J., vol. 5, no. 3, pp. 433–438,
Jun. 2005.
[15] D. F. Wang et al., “Passive MEMS DC electric current sensor:
Part I—Theoretical considerations,” IEEE Sensors J., vol. 17, no. 5, Valentina Sinatra was born in Catania, Italy, in
pp. 1230–1237, Mar. 2017. 1982. She received the M.S. degree in microelec-
[16] O. Z. Olszewski et al., “A MEMS silicon-based piezoelectric AC current tronic engineering from the University of Catania,
sensor,” Procedia Eng., vol. 87, pp. 1457–1460, Jan. 2014. Catania, in 2013, where she is currently pursuing
[17] F. Keplinger, S. Kvasnica, A. Jachimowicz, F. Kohl, J. Steurer, and the Ph.D. degree with a focus on magnetic sensors.
H. Hauser, “Lorentz force based magnetic field sensor with optical Ms. Sinatra received the title of Specialist in Micro
readout,” Sens. Actuators A, Phys., vol. 110, no. 1, pp. 112–118, and Nano Technologies for Integrated Biosensors in
Feb. 2004. Diagnostic Applications in 2015.
[18] S. Baglio, L. Latorre, and P. Nouet, “Resonant magnetic field microsen-
sors in standard CMOS technology,” in Proc. 16th IEEE Instrum. Meas.
Technol. Conf., vol. 1, May 1999, pp. 452–457.

Authorized licensed use limited to: University of Catania. Downloaded on April 05,2023 at 15:01:00 UTC from IEEE Xplore. Restrictions apply.
1396 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 69, NO. 4, APRIL 2020

Giuseppa Crea was born in Messina, Italy, in 1990. Salvatore Baglio (F’13) received the Laurea degree
She received the B.S. in electronic engineering from in electronic engineering and the Ph.D. degree in
the University of Messina, Messina, in 2015. She electrical engineering from the University of Cata-
is currently pursuing the master’s degree with the nia, Catania, Italy, in 1900 and 1994, respectively.
University of Catania, Catania, Italy, with a focus He is currently a Full Professor of instrumentation
on experimental thesis in the design and characteri- and measurements with the Dipartimento di Ingegne-
zation of microsensors. ria Elettrica Elettronica e Informatica (DIEEI), Uni-
versity of Catania. He teaches courses in electronic
measurements systems and micro- and nanosensors.
He is a Principal Investigator in several scientific
research projects dealing with the development of
innovative sensor systems, granted by private companies and by different
national and international institutions. He is also a member of the Board
of Ph.D. course in electronic and automation engineering with the University
Bruno Andò (SM’13) received the M.S. degree of Catania. He has authored more than 350 scientific publications including
in electronic engineering and the Ph.D. degree in international journals, books, conferences, and patents. His current research
electrical engineering from the Università di Catania, interests include the development of measurement methodologies and systems
Catania, Italy, in 1994 and 1999, respectively. based on nonlinear dynamics, development of micro- and nanosensors and
From 1999 to 2001, he was a Researcher with transducers, exploitation of material properties toward integrated transducers,
the Electrical and Electronic Measurement Group, and nonlinear methodologies and integrated devices for energy harvesting.
Dipartimento Elettrico Elettronico e Sistemistico, Dr. Baglio is a member of the I&M Society AdCom where he served for VP
University of Catania, where he became an Assis- Education from 2016 to 2017 and as an Executive for VP in 2018. He is an
tant Professor in 2002. He has coauthored several Editor-in-Chief of the I&M Video Tutorials and an Associate Editor in Chief
scientific papers presented in international confer- of the IEEE T RANSACTIONS ON I NSTRUMENTATION AND M EASUREMENTs.
ences and published in international journals and
books. His current research interests include sensor design and optimiza-
tion, advanced multisensor architecture for ambient-assisted living, sensor
networks, characterization of new materials for sensors, nonlinear techniques
for signal processing with particular interest in stochastic resonance, dithering
applications, and distributed measurement systems.

Authorized licensed use limited to: University of Catania. Downloaded on April 05,2023 at 15:01:00 UTC from IEEE Xplore. Restrictions apply.