sensors

Article
The Effect of Changes in Magnetic Field and Frequency on the
Vibration of a Thin Magnetostrictive Patch as a Tool for
Generating Guided Ultrasonic Waves
Akram Zitoun 1 , Steven Dixon 2,3, *, Mihalis Kazilas 1 and David Hutchins 2

1 Brunel Composites Centre, College of Engineering, Design and Physical Sciences, Brunel University London,
London UB8 3PH, UK; akram.zitoun@brunel.ac.uk (A.Z.); mihalis.kazilas@brunel.ac.uk (M.K.)
2 School of Engineering, The University of Warwick, Coventry CV4 7AL, UK; D.A.Hutchins@warwick.ac.uk
3 Department of Physics, The University of Warwick, Coventry CV4 7AL, UK
* Correspondence: s.m.dixon@warwick.ac.uk

Abstract: A set of experiments was designed and conducted to investigate the vibrational ultrasonic
response of a thin magnetostrictive patch bonded to a glass plate, with changes in static and dynamic
magnetic fields applied to the patch. Such arrangements are often used as a means of generating
guided waves in pipes or plates, by attaching a patch to a sample’s surface. The effect of varying
the applied static and dynamic magnetic field’s amplitudes and directions and the frequency of the
dynamic magnetic field was studied. It was demonstrated that the vibration of the magnetostrictive
patch could be controlled and enhanced by optimizing the magnetic fields. It was also shown that for
low-amplitude dynamic magnetic fields, Lorentz forces generated within the patch and the resonant
frequency of the patch could also contribute to the enhancement of the vibration of the patch for
low-amplitude fields. For high-amplitude dynamic magnetic fields, the magnetostriction effect



can be the main transduction mechanism, which can be optimized for non-destructive testing and


inspection purposes.
Citation: Zitoun, A.; Dixon, S.;
Kazilas, M.; Hutchins, D. The Effect
Keywords: magnetostriction; magnetic field; magnetic domains; vibration
of Changes in Magnetic Field and
Frequency on the Vibration of a Thin
Magnetostrictive Patch as a Tool for
Generating Guided Ultrasonic Waves.
Sensors 2022, 22, 766. https:// 1. Introduction
doi.org/10.3390/s22030766 Magnetostriction is a physical phenomenon that occurs in almost all ferromagnetic
Academic Editor: Seung-bok Choi
materials. It was first reported by James Joule in 1842 [1], and thereafter was investigated
by other authors [2,3]. The term magnetostriction refers essentially to the coupling between
Received: 6 December 2021 the applied magnetic field and the strain or stress generated due to such a field. The direct
Accepted: 17 January 2022 effect (also called the Joule effect) occurs when the dimension of the sample changes due to
Published: 20 January 2022
the application of an external magnetic field. The inverse effect (the Villari effect) occurs
Publisher’s Note: MDPI stays neutral when the magnetization of the ferromagnetic sample changes due to strain [4–6]. Other
with regard to jurisdictional claims in coupling concepts have also been investigated, including the Wiedemann effect [7], the
published maps and institutional affil- Matteucci effect [8] and the magnetoimpedance stress dependent effect [9]. Magnetostric-
iations. tion is commonly observed when an external magnetic field is applied to a ferromagnetic
material [10].
The magnetostrictive coefficient is usually written as λ and is defined as the ratio of
the fractional change in length δl over the initial length l0 , as shown in Equation (1):
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland. δl
λ= (1)
This article is an open access article l0
distributed under the terms and
conditions of the Creative Commons Atomic magnetic dipoles are associated with atoms located at specific locations within
Attribution (CC BY) license (https:// the crystal lattice of the ferromagnetic sample. These magnetic dipole moments tend to
creativecommons.org/licenses/by/ align in the same direction within a magnetic domain spontaneously, as this situation is
4.0/). energetically favorable [11]. In the absence of an external magnetic field, the magnetic

Sensors 2022, 22, 766. https://doi.org/10.3390/s22030766 https://www.mdpi.com/journal/sensors

 align with the direction of the external applied magnetic field. This g
the material due to the magnetic anisotropy in the crystal structure
Sensors 2022, 22, 766
induced depends on the specific sample properties, including 2 of 15
mat
microstructure and the shape of the sample [15].
For materials with polycrystalline structures, the magnetostri
domains within a ferromagnetic sample are oriented in a distribution that minimizes
tained using
the external the following
magnetic equation:
field [12–14]. When an external magnetic field is applied to the
ferromagnetic material, the magnetic domains and the magnetic dipole moments start to
align with the direction of the external
2 2 applied32 2magnetic 2 field. This generates strain within
the material due to the 100
magnetic
1 1=λ
anisotropyλ (α β + α β + α β ) + 3λ (α α β β + α α β
2 2in the
2
crystal
3 3 structure.
111The 1degree
2 1of 2strain 2 3 2
2
induced depends on the specific sample properties, including material composition and
microstructure and the shape of the sample [15].
where Forλmaterials
100 andwith λ111polycrystalline
are the magnetostriction coefficients
structures, the magnetostriction ratio canalong
be ob- the <
tained using the following equation:
tions, respectively, and αi (i = 1, 2, 3) and βi (i = 1, 2, 3) are, respectively
relativeλ = to32 the
λ100 (αfield
1 β 1 + αdirection
2 2
2 β 2 + α3 β 3 ) + overlapping
2 2 2 2
3λ111 (α1 α2 β 1 β 2 + αwith
2 α3 β 2 β 3the
+ α3 αaxis
1 β 3 β 1 )where
(2) the
rated and the direction overlapping the axis following which th
where λ100 and λ111 are the magnetostriction coefficients along the <100> and <111>
measured [16,17].and
directions, respectively, Asαican (i = 1,be
2, 3)seen
and βi from
(i = 1, 2, Equation (2), the
3) are, respectively, there is a non
direction

tween the external applied magnetic field and the resultant strain
cosines relative to the field direction overlapping with the axis where the magnetization
is saturated and the direction overlapping the axis following which the magnetostriction
taken into consideration
is measured [16,17]. As can be seen when designing
from Equation (2), therea magnetostriction-based
is a nonlinear relationship

relationship between whenthe applied field (H) andsystem. the magnetostriction

between the external applied magnetic field and the resultant strain, and this needs to be
taken into consideration designing a magnetostriction-based The nonlinear
cobalt alloy is illustrated in Figure 1.
relationship between the applied field (H) and the magnetostriction ratio (λ) for an iron–
cobalt alloy is illustrated in Figure 1.
80

Magnetostriction ratio (ppm)

70

60
(ppm)

50

40
Magnetostriction ratio

30

20

10

0
-1500 -1000 -500 0 500 1000

Magnetic field H (A/m)

Figure 1. Magnetostriction curve for an iron–cobalt alloy obtained through modeling in Comsol.
Figure 1. Magnetostriction curve for an iron–cobalt alloy obtained through
The magnetostriction effect observed and the behavior of the material when a field
is applied differs from one material to another [18–21]. As shown in Figure 1, the mag-
The magnetostriction
netostriction effect
curve is symmetric in shape observed
[22,23] and the
and is obtained behavior
by slowly of the m
varying an
applied differs from one material to another [18–21]. As shown in F
externally applied magnetic field to a ferromagnetic specimen and capturing the change in
dimensions. The X-axis can be represented by the magnetic flux density B, magnetic field
striction
strength Hcurve is symmetric
or magnetization M. in shape [22,23] and is obtained by slo
nally applied magnetic field to a ferromagnetic specimen and captu
The relationship linking the magnetic properties and the stress-related response of the
magnetostrictive material, as reported in [23–27], can be written as follows:
mensions. The X-axis can be represented by the magnetic flux den
ε = s( H ) σ + dH
strength H or magnetization M.
(3)

The relationship linking the− d−magnetic properties and the

(4) stre
1
σ = cε H

the magnetostrictive material,

B = d T as
σ + reported
µ(σ) H in [23–27], can be(5)written
𝜀𝜀 = 𝑠𝑠 (𝐻𝐻) 𝜎𝜎 + 𝑑𝑑𝑑𝑑
𝜎𝜎 = 𝑐𝑐𝑐𝑐 − 𝑑𝑑−1 𝐻𝐻
Sensors 2022, 22, 766 3 of 15

where ε, σ and B are strain, stress and magnetic field density matrices, respectively, and s(H) ,
c and d are the elastic compliance at a specific value of H, the stiffness and the piezomagnetic
matrices, respectively. dT is the transpose matrix of d, d−1 is the inverse matrix of d and µ(σ)
is the permeability of the material measured at a specific value of the stress σ. The variables
s(H) , d and µσ are nonlinear, as they depend on the stress and the magnetic field. These
equations can be used to design magnetostrictive thin films as either sensing or actuating
elements. Equation (3) establishes the magnetostrictive element as an actuation system via
the Joule effect, whereas Equation (5) establishes the magnetostrictive material as a sensing
element via the Villari effect.
The magnetostriction phenomenon is widely used for different applications, such
as the active element in ultrasonic transducers used to generate elastic waves for nonde-
structive testing and inspection [28–31], microelectromechanical systems or MEMS [29],
positioning sensors [32], level-gauging sensors [33] and accurate actuation mechanisms [34].
Despite the fact that magnetostriction is a well-known physical property of ferromag-
netic materials and that it has numerous applications within industry, some aspects of
magnetostrictive behavior do not seem to have been reported fully, such as the effect of
the variation of the excitation frequency and the direction and amplitude of the externally
applied magnetic fields. The dynamic magnetic field driving frequency has been considered
critical in designing the magnetostrictive patch [35], as ultrasonic guided wave application
often requires the generation of a specific frequency of wave with a specific displacement
of wave at the sample surface. Numerical modeling has established that different guided
wave modes can be optimally selected and generated by modifying the coil geometry that
generates the dynamic magnetic field and by optimizing the permanent or static magnetic
field that magnetically biases the patch [36].
This experimental paper helps to explain and demonstrate the response of a thin patch
due to magnetostriction from different orientations and relative amplitudes and excitation
frequencies of dynamic and static magnetic fields. The experimental sample used here is a
patch bonded to a large glass plate to generate ultrasonic guided waves in the plate [29],
but the general findings of the forces generated at the patch can be applied to other types
of guided waves where a patch may be applied to a pipe or a bar, for example.

2. Experimental Work
2.1. Electromagnet and RF Coil Design Generating the Different Magnetic Fields
An electromagnet was used to provide a bias magnetic field for the magnetostrictive
patch, that was effectively a pseudostatic field compared to the dynamic magnetic field
obtained from a different coil. The electromagnet used in these experiments contained
an E-shaped ferrite core with dimensions 65 mm (length) × 32.8 mm (height) × 27.4 mm
(width), as shown in Figure 2a,b. This shape enabled two directions of static magnetic field
(Bs ) to be applied to the thin magnetostrictive patch that was bonded to the sample being
tested. Placing the patch in the region under the central pole of the electromagnet allowed
the static magnetic field to be applied perpendicularly to the surface, whilst an in-plane
field would be generated within the pole gaps.
A COMSOL finite element model was developed to investigate the magnitude of
the magnetic flux density generated by the electromagnet. The results are presented in
Figure 2c for the area within the central pole region (where the field was predominantly
perpendicular to the face of the electromagnet) and in Figure 2d for the pole gap region
(where it was mainly parallel).
Two hundred turns of wire were used for the electromagnet coil so as to generate a
magnetic flux density between the poles of up to 1 T. Modeling indicated that a current
of 25 amps would be needed in the presence of a magnetostrictive patch, which altered
the field from the free space value. The same current would produce a magnetic flux
density of approximately 0.5 T if the patch was not present. This was because the patch
had a high value of magnetic permeability, which would concentrate magnetic field lines
and thus increase the flux density. The source of the dynamic magnetic field (Bd ) was a
Sensors 2022, 22, 766 4 of 15

racetrack coil of 10 turns and a 0.5 mm spacing between turns, on a printed circuit board
(PCB), and is shown in Figure 3a. The COMSOL model results shown in Figure 3b provide
an understanding of the characteristics of the magnetic field generated by the rf coil and
helped in the transducer design to optimize field generation in the required direction. The
magnetostrictive patch had to be accommodated under this coil so that Bd was always
applied in the direction parallel to the surface of the patch. The magnetostrictive patch
used in the experiments was fabricated from a VACOFLUX 48 iron–cobalt alloy supplied
by ® VACUUMSCHMELZE GmbH & Co., Hanau, Germany, with the dimensions and
properties given in Table 1. These patches exhibit isotropic behavior when different static
Sensors 2022, 22, x FOR PEER REVIEW
and dynamic fields are applied [29], as demonstrated by placing the patch under a coil and
4 of 15
permanent and then rotate the system by 90◦ while monitoring the vibration behavior. It
was shown that rotation had minimal effect on the vibration of the specimen.

65 mm
32.8 mm

Central Pole
pole gap

(a) (b)
Magnetic flux density across the EM central pole (Patch absent) 0.5
0.4

0.4

0.3 0.3 electromagnet

0.2
pole gaps
0.2
0.1
Magnetic flux density (T)

0
Magnetic flux density (T)

0.1

-0.1

0 -0.2

-0.3

-0.1
electromagnet -0.4

-0.2
central pole -0.5
0 50 100 150 200 250 300 350 400 450
0 10 20 30 40 50 60
Distance across the electromagnet (mm)
Distance across the electrmagnet

(c) (d)
Figure
Figure2.2.(a)
(a)The
Theferrite
ferrite core used for
core used for the
theelectromagnet;
electromagnet;(b)
(b)photograph
photographofof
thethe assembled
assembled electro-
electromag-
magnet. Results of Comsol modeling for the magnetic flux density generated by the electromagnet
net. Results of Comsol modeling for the magnetic flux density generated by the electromagnet are
are shown across (c) the central pole region and (d) in the two pole gaps.
shown across (c) the central pole region and (d) in the two pole gaps.
TableA1.COMSOL finite
Patch design element model was developed to investigate the magnitude of the
parameters.
magnetic flux density generated by the electromagnet. The results are presented in Figure
Patch Material Iron–Cobalt Alloy
2c for the area within the central pole region (where the field was predominantly perpen-
dicular to the face ofShape
the electromagnet) and in Figure 2d for thesquare
pole gap region (where
it was mainly parallel).
Dimensions 20 mm × 20 mm × 0.1 mm
Young’s Modulus 200 GPa
Two hundred turns of wire were used for the electromagnet coil so as to generate a
Poisson Ratio 0.29
magnetic flux density between the poles of up to 1 T. Modeling8.12
Density indicated
g/cm3 that a current of
25 amps would be needed in the
Electrical Resistivity presence of a magnetostrictive0.42 µΩmwhich altered the
patch,
field from the freePermeability
space value. The same current would produce a magnetic
18,000 N A−2 flux density
of approximately 0.5Magnetostriction
Saturation T if the patch was not present. This was because the patch had a high
70 ppm
value of magnetic permeability,
Saturation Magnetisation which would concentrate magnetic field lines and thus
2.35 T
increase the flux density. The source of the dynamic magnetic field (Bd) was a racetrack
coil of 10 turns and a 0.5 mm spacing between turns, on a printed circuit board (PCB), and
is shown in Figure 3a. The COMSOL model results shown in Figure 3b provide an under-
standing of the characteristics of the magnetic field generated by the rf coil and helped in
Sensors2022,
Sensors 22,x766
2022,22, FOR PEER REVIEW 5 5ofof15
15

Relative posi-
tion of the patch

Magnetic flux density (T)

under the RF coil

Distance (mm)
Y

Distance (mm)

(a) (b)
Figure
Figure3.3.(a)
(a)The
Thedimensions
dimensionsof
ofthe
thecoil
coilused
usedto
togenerate
generatethe
thedynamic
dynamicmagnetic
magneticfield.
field.(b)
(b)COMSOL
COMSOL
model results for the dynamic magnetic field (Bd) and the relative position of the magnetostrictive
model results for the dynamic magnetic field (Bd ) and the relative position of the magnetostric-
patch.
tive patch.
Table 1. Patch design
2.2. Experimental parameters.
Apparatus
The experiments Patchwere conducted to investigate the change
Material in vibrational
Iron–Cobalt Alloy properties of
the magnetostrictive patch Shape in the three orthogonal (X, Y, Z) directions for
square variations in both
the static magneticDimensions
field from the electromagnet (Bs ) and20the mm RF×coil
20 mm(resulting
× 0.1 mmin changes
in the dynamicYoung’sfield Bd ).Modulus 200 GPa
A schematic diagram of the apparatus is shown in Figure 4a. The custom-designed
Poisson Ratio 0.29
Innerspec pulsing unit contained two subsystems, providing different 3outputs: an RF
Density 8.12 g/cm
current pulsing unit connected to the RF PCB coil, which provided a nominal 5-cycle tone-
Electrical Resistivity 0.42 µΩm
burst pulse over an operating frequency range from 100 to 250 kHz, and an electromagnet
Permeability
power unit, as shown in Figure 4a. The electromagnet used to provide 18,000 N Athe
−2
pseudostatic
Saturation Magnetostriction 70
magnetic field (Figure 4a) was in the form of the E-shaped ferrite core, as shown earlier ppm
in Figure 2a. Saturation
It offeredMagnetisation
the capability of generating both an out-of-plane 2.35 T quasi-static field
under the central pole region, and an in-plane field between the central and side poles.
2.2.
TheExperimental
measurements Apparatus
indicated that the generated quasi-static magnetic field ranged from
0 to The
0.5 Texperiments
under the central pole (out-of-plane),
were conducted to investigate andthefrom 0 to in
change 1.1vibrational
T betweenproperties
the poles
(in-plane), i.e., close to that expected from the COMSOL model. The
of the magnetostrictive patch in the three orthogonal (X, Y, Z) directions for variations racetrack coil shown in
both the static magnetic field from the electromagnet (Bs) and the RF coil (resultingtwo
earlier in Figure 3 was connecting to a high-power current pulse generator, delivering in
levels ofin
changes power at 450 Wfield
the dynamic andB1.8
d). kW to generate the higher frequency dynamic field that
generated the ultrasonic
A schematic diagram waves.
of theTypical outputs
apparatus from these
is shown two pulsing
in Figure 4a. Thesystems are shown
custom-designed
in Figure 4b, and were those provided by the control system of
Innerspec pulsing unit contained two subsystems, providing different outputs: an RF the Innerspec unit, noting
cur-
that the real current signals may have been different to this. Note
rent pulsing unit connected to the RF PCB coil, which provided a nominal 5-cycle tone- that the frequency of the
tone-burst
burst pulse input
over an into the rf coil
operating could be range
frequency adjustedfromto 100
examine
to 250the
kHz,frequency
and an response of the
electromagnet
patch under various static field conditions.
power unit, as shown in Figure 4a. The electromagnet used to provide the pseudostatic
magnetic field (Figure 4a) was in the form of the E-shaped ferrite core, as shown earlier in
Figure 2a. It offered the capability of generating both an out-of-plane quasi-static field
under the central pole region, and an in-plane field between the central and side poles.
The measurements indicated that the generated quasi-static magnetic field ranged from 0
to 0.5 T under the central pole (out-of-plane), and from 0 to 1.1 T between the poles (in-
plane), i.e., close to that expected from the COMSOL model. The racetrack coil shown
earlier in Figure 3 was connecting to a high-power current pulse generator, delivering two
levels of power at 450 W and 1.8 kW to generate the higher frequency dynamic field that
 generated the ultrasonic waves. Typical outputs from these two pulsing systems are
shown in Figure 4b, and were those provided by the control system of the Innerspec unit,
noting that the real current signals may have been different to this. Note that the frequency
Sensors 2022, 22, 766
of the tone-burst input into the rf coil could be adjusted to examine the frequency response
6 of 15

of the patch under various static field conditions.

(a)
6

3
Current amplitude (Amps)

2
Electromagnet input
Electromagnet input current
current waveform
waveform

RF coil
Coil input
input current
current waveform
waveform

-1
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Time (s) -5
10

(b)
Figure
Figure4. (a)
4. Schematic diagram
(a) Schematic of theofapparatus.
diagram (b) The(b)
the apparatus. nominal output from
The nominal the from
output dual electromag-
the dual elec-
net/rf
tromagnet/rf coil driving system, showing both the electromagnet drive current andtone-burst
coil driving system, showing both the electromagnet drive current and the 5-cycle the 5-cycle
used to exciteused
tone-burst the rf
to pancake
excite thecoil. Note that
rf pancake theNote
coil. rf coil
thatwas
theexcited onceexcited
rf coil was the current to the
once the electro-
current to the
magnet had stabilized.
electromagnet had stabilized.

The dynamic
The dynamicdisplacement
displacementof of
thethe
magnetostrictive
magnetostrictive patch
patchwas
wasmeasured
measured using a scan-
using a scan-
ning vibrometer Polytec CLV 3000 3D (Polytec GmbH, Waldbronn,
ning vibrometer Polytec CLV 3000 3D (Polytec GmbH, Waldbronn, Germany), which Germany), which
could
could be used
be used to detect
to detect bothboth in-plane
in-plane andand out-of-plane
out-of-plane motion.
motion. TheThe laser
laser system
system included
included three
three separate
separate Doppler
Doppler vibrometer
vibrometer heads,
heads, detecting
detecting ultrasonic
ultrasonic motion
motion inin the
the three
three orthog-
orthogonal
onal axes
axes (X,(X, Y and
Y and Z, Z,
asas shown
shown inin Figure5)5)from
Figure fromthe
thelaser
laserlight
lightreflected
reflectedfrom
fromthethesurface
surface of
of the
the patch.
patch.
 Sensors2022,
Sensors 22,x766
2022,22, FOR PEER REVIEW 7 7ofof15
15

(a)

(b)
Figure
Figure5.5.(a)(a)
Relative positions
Relative of the
positions magnetostrictive
of the patch,
magnetostrictive the coil
patch, the and
coil the
andelectromagnet to gen-to
the electromagnet
erate the first set of experiments where the static magnetic field was out-of-plane while the
generate the first set of experiments where the static magnetic field was out-of-plane while the dynamic
magnetic
dynamic field was in
magnetic plane.
field was(b) Relative
in plane. (b)positions of the magnetostrictive
Relative positions patch, thepatch,
of the magnetostrictive coil and
the the
coil
electromagnet to generate the second set of experiments where both the static magnetic field and
and the electromagnet to generate the second set of experiments where both the static magnetic field
the dynamic magnetic field were in-plane.
and the dynamic magnetic field were in-plane.
Note
Notethat
thatthe
theoutput
outputofofthethevibrometer
vibrometerinineach
eachcase
casewas
wasaawaveform
waveformofofthe theultrasonic
ultrasonic
particle
particle velocity at each point of measurement. The 5-cycle tone-burst shownininFigure
velocity at each point of measurement. The 5-cycle tone-burst shown Figure4b 4b
was delayed via the triggering gate system to coincide with the time at
was delayed via the triggering gate system to coincide with the time at which the fieldwhich the field
from
from the electromagnet
the electromagnet reached
reached a constant
a constant and maximum
and maximum value,
value, as as indicated
indicated in Figurein 4b.
Figure
This
4b. This ensured that the field from the electromagnet could be considered to be
ensured that the field from the electromagnet could be considered to be static, i.e., constantstatic, i.e.,
constant
for each for each measurement.
measurement. The patch The patch
was wasunder
placed placedthe
under thesection
linear linear of
section of the race-
the racetrack coil,
track coil, as shown earlier
as shown earlier in Figure 3. in Figure 3.
Sensors 2022, 22, 766 8 of 15

3. Results
3.1. Static Magnetic Field (Bs ) in the Out-of-Plane Direction
The experiment arrangement described above allowed the response of the magne-
tostrictive patch, as measured by the vibrometer as a function of frequency (f) and magnetic
field direction. The first set of tests had the static field Bs in the out-of-plane (Z) direction
while the dynamic field Bd was in-plane and predominantly in the X direction, as shown in
Figure 5. The amplitude of the static field Bs and the dynamic coil excitation frequency f
were then changed, and the ultrasonic vibration signal of the patch surface was recorded for
the three different directions of motion (X, Y and Z) using the vibrometer. The amplitude
of the dynamic field Bd was kept constant during these experiments. The frequency f of
the current through the rf coil was varied from 100 to 240 kHz in 10 kHz steps, while the
amplitude of Bs was varied from 0.1 to 0.5 T, with an increasing step of 0.05 T.
The results of these experiments are presented in Figure 6a–c, which shows changes in
the vibrational (particle velocity) amplitudes in the three orthogonal directions (X, Y, Z),
respectively, as detected by the laser vibrometer. The resultant magnitude of the velocity,
shown in Figure 6d, was calculated using the equation below:
q
Magnitude MagV = Vx 2 + Vy 2 + Vz 2 , (6)

where Vx , Vy and Vz are the particle velocity amplitudes in the directions X, Y and Z,
respectively. Consider first the X direction (Figure 6a), where it can be seen that there
was an abrupt reduction in vibration amplitude at a frequency of about 180 kHz. This
change occurred at a lower frequency for smaller values of Bs . Similar trends were seen
in the Z direction (Figure 6c), but now a maximum in vibration amplitude was observed
at Bs values above 0.4 T in the frequency range f = 140–180 kHz. In the Y direction
(Figure 6b), a reasonably high amplitude existed for f = 140–200 kHz, but again only at
higher Bs values. The overall magnitude of the vibration demonstrated that the most
effective frequencies to use were those in the f = 140–200 kHz range, although relatively
high vibrational levels could be obtained even for low frequencies with static magnetic field
values below Bs = 0.2 T. Hence, an appropriate selection of the correct frequency and static
field combination was required in order to generate a high amplitude ultrasonic signal,
and this choice would vary depending on the direction of interest (e.g., in attempting to
generate a particular ultrasonic wave mode in a sample to which the magnetostrictive
material may be attached).
This configuration led to the generation of Lorentz forces in the patch, mainly in the X
direction. In addition, magnetostrictive forces were primarily expected to be generated in
the X direction. From the figures below, it can be seen that both forces were constructively
interfering. In fact, in Figure 6a, the amplitude of velocity was higher than the velocity
amplitude recorded in the Y and Z directions. The amplitude of velocity in the X direction
was double that compared to the velocity in the Y direction, and it was much higher
than the amplitude of particle velocity vibrating in the Z direction. It can also be noticed
that the vibrations were greater in the same direction as the dynamic magnetic field (Bd ).
Preliminary modeling of the vibrating natural frequencies of the patch geometry identified
that a vibrational mode of the patch geometry was oscillating at 100 kHz, which is in
agreement with the results below. As a matter of fact, the variation of the frequency and
the static magnetic field amplitude had minimal effect of the amplitude of the particle
velocity. This was due to the fact that the patch was vibrating within the range of its natural
frequency, which contributed to the overall vibration of the patch, compared to the Lorentz
forces and the magnetostrictive forces. The mechanism had an effect on the generation of
the guided waves, as their excitability was highly dependent on the driving frequency. In
this case, the vibrations generating the guided waves were being induced mainly by the
natural frequencies of the patch, rather than the effect of magnetostriction or Lorentz forces.
 to use were those in the f = 140–200 kHz range, although relatively high vibrational levels
could be obtained even for low frequencies with static magnetic field values below Bs =
0.2 T. Hence, an appropriate selection of the correct frequency and static field combination
was required in order to generate a high amplitude ultrasonic signal, and this choice
would vary depending on the direction of interest (e.g., in attempting to generate a par-
Sensors 2022, 22, 766 ticular ultrasonic wave mode in a sample to which the magnetostrictive material may 9 of 15
be
attached).

Sensors 2022, 22, x FOR PEER REVIEW 9 of 15

(a) (b)

(c) (d)
Figure6.6.Particle
Figure Particlevelocity
velocityamplitude
amplitudemapping
mappingwhile
whilevarying
varyingthethestatic
staticmagnetic
magneticfield
fieldamplitude
amplitudeand
and
the frequency of the coil excitation in (a) the X direction, (b) the Y direction and (c) the Z direction.
the frequency of the coil excitation in (a) the X direction, (b) the Y direction and (c) the Z direction.
(d) The resultant magnitude was calculated using Equation (6).
(d) The resultant magnitude was calculated using Equation (6).

3.2. BThis
s and configuration led toto
Bd In-Plane, Parallel thethegeneration of Lorentz
Surface of the forces inPatch
Magnetostrcitve the patch, mainly intothe
and Orthogonal
Each Other In addition, magnetostrictive forces were primarily expected to be generated
X direction.
in the
In X direction.
this From the figures
set of experiments, the Bdbelow,
and Bsitfields
can be seeninthat
were the both
X and forces were construc-
Y direction, respec-
tively interfering. In fact, in Figure 6a, the amplitude of velocity
tively (i.e., orthogonal to each other). Again, the resultant particle velocity amplitudes was higher than the ve-
locityrecorded
were amplitude in recorded
the X, Y and in theZY and Z directions.
directions as a functionThe amplitude
of f and Bsof velocity
. In in the
addition, twoX
direction was
amplitudes double
of the inputthat
powercompared
were used to the
to velocity
generatein thethe Y direction,
dynamic field Band
d toit was much
investigate
higher
the effectthan the amplitude
of varying of particle
this parameter forvelocity
a constantvibrating
Bs . In in theconfiguration,
this Z direction. Itthe canLorentz
also be
noticed
force that the would
generation vibrations were greater
be expected to beinzero.
the same direction
The results as the
for the casedynamic
where the magnetic
input
field (B ). Preliminary modeling of the vibrating natural frequencies
power used to generate the dynamic field was 450 W are shown in Figure 7. Significant
d of the patch geometry
identified that
differences canabe vibrational
seen relativemode toof the patch
Figure geometry
6, where was was
the field oscillating
out of at 100 kHz,
plane. Now, which
the
is in agreement
highest amplitudes within the results
all three below. As
directions (X,aYmatter
and Z)ofoccurred
fact, theforvariation of the frequency
lower frequencies of up
and
to 40 the
kHz, static
andmagnetic
there wasfield
littleamplitude had minimal
signal at higher effectThe
frequencies. of the amplitude
magnitude of (Figure
plot the particle
7d)
reflected
velocity.these changes
This was due into the three different
fact that directions,
the patch in that the
was vibrating highest
within amplitude
the range of itswas
nat-
achieved for specific
ural frequency, whichcombinations
contributed ofto
frequency
the overallandvibration
applied static
of themagnetic field amplitude.
patch, compared to the
The
Lorentz forces and the magnetostrictive forces. The mechanism had an effect on thefrom
magnitude was greatest at frequencies in the 105–140 kHz range, as expected gen-
Figure
eration7a–c.
of the guided waves, as their excitability was highly dependent on the driving
frequency. In this case, the vibrations generating the guided waves were being induced
mainly by the natural frequencies of the patch, rather than the effect of magnetostriction
or Lorentz forces.

3.2. Bs and Bd In-Plane, Parallel to the Surface of the Magnetostrcitve Patch and Orthogonal to
Each Other
Sensors 2022, 22, x FOR PEER REVIEW 10 of 15

Sensors 2022, 22, 766 amplitude. The magnitude was greatest at frequencies in the 105–140 kHz range, 10
asofex-
15

pected from Figure 7a–c.

(a) (b)

(c) (d)
Figure 7. Velocity mapping while varying the static magnetic field Bs and the frequency in the (a)
Figure 7. Velocity mapping while varying the static magnetic field Bs and the frequency in the (a) X,
X, (b) Y and (c) Z directions, respectively. (d) The resultant magnitude of the particle velocity. The
(b) Y and (c) Z directions, respectively. (d) The resultant magnitude of the particle velocity. The input
input power used to generate the dynamic field was 450 W throughout.
power used to generate the dynamic field was 450 W throughout.
Results for the higher dynamic field power input of 1.8 kW are shown in Figure 8. It
Results for the higher dynamic field power input of 1.8 kW are shown in Figure 8.
can be seen now that the distribution in amplitudes was spread over a wider range of
It can be seen now that the distribution in amplitudes was spread over a wider range of
frequencies than at the lower power level (Figure 7), with the maximum amplitudes being
frequencies than at the lower power level (Figure 7), with the maximum amplitudes being
reached over the 120–160 kHz frequency range in the case of the Y direction. However,
reached over the 120–160 kHz frequency range in the case of the Y direction. However, the
the maximum
maximum amplitudes
amplitudes in theinZthe Z direction
direction werelocated
were now now located
as beforeas before at frequencies
at frequencies below
below 120 kHz. As both magnetic fields were located parallel to the surface of the magne-
120 kHz. As both magnetic fields were located parallel to the surface of the magnetostrictive
tostrictive
patch, patch, theinvibration
the vibration in the was
the Z direction Z direction
mainly awas mainly
result of theavibration
result of in
theboth
vibration
X and inY
both X and
directions, Y to
due directions,
a Poissondue to a Poisson
ratio-type ratio-type
effect. Note effect.
also that Note also
the velocity that therecorded
amplitude velocity
amplitude
in recorded
the X direction hadindoubled
the X direction
from 0.02hadtodoubled
0.04 m/s, from 0.02
while thetovelocity
0.04 m/s,captured
while theinveloc-
the Y
ity captured in the Y direction increased by
direction increased by 50% (from 0.02 to 0.03 m/s). 50% (from 0.02 to 0.03 m/s).
From Figure 7, it can still be seen that the vibration velocity amplitude remained
relatively high at around 100 kHz, indicating that by adopting this design, the vibration
due to the natural frequency was still significant and could influence the effects of the
Lorentz forces and the magnetostrictive forces. It is worth mentioning that the vibrations
due to the natural frequencies of the patch could be selected, as in an NDT application
the patch will be glued, which leads to a variation of the overall natural frequency of
the system.
Sensors 2022, 22, x FOR PEER REVIEW 11 of 15
Sensors 2022, 22, 766 11 of 15

(a) (b)

(c) (d)
Figure 8. Velocity mapping while varying the static magnetic field Bs and the frequency in the (a)
Figure 8. Velocity mapping while varying the static magnetic field Bs and the frequency in the (a) X,
X, (b) Y and (c) Z directions, respectively. (d) A map of the resultant magnitude of the particle ve-
(b) Y and (c) Z directions, respectively. (d) A map of the resultant magnitude of the particle velocity.
locity. The input power used to generate the dynamic field was 1.8 kW throughout.
The input power used to generate the dynamic field was 1.8 kW throughout.
From Figure 7, it can still be seen that the vibration velocity amplitude remained rel-
Another important observation can be obtained by comparing the results from the
atively high at around 100 kHz, indicating that by adopting this design, the vibration due
situation where the external static field was applied either out-of-plane (Figure 6) or
to the natural
parallel frequency
to the surface was
of the still significant patch
magnetostrictive and could influence
(Figures theIn
7 and 8). effects of the
fact, the Lorentz
magnitude
of the vibration had increased by almost an order of magnitude (from 2.5 × 10due
forces and the magnetostrictive forces. It is worth mentioning that the vibrations to
−4 to
the natural frequencies of the patch could be selected, as in an NDT
2.2 × 10−3 m/s), and the working region had increased to provide more combinations application the patch
will
of thebefrequency/static
glued, which leads to a variation
magnetic of the overall
field amplitudes, at natural
which afrequency
reasonable of response
the system.was
Another important observation can be obtained by comparing
observed. Although the boundary conditions of the patch had changed between the results fromboth
the
situation where the external static field was applied either out-of-plane
experiments, when the magnetostrictive patches were used to generate guided waves for (Figure 6) or par-
allel to the purposes,
inspection surface of the
the magnetostrictive patch (Figures
variation of the magnetic 7 and 8). In
field directions andfact, the magnitude
amplitude could
of the vibration had increased by almost an order of
enhance the signal generation and increase detectability of defects. magnitude (from 2.5 × 10−4 to 2.2 ×
10−3 m/s), and thetoworking
Compared regionexperiments,
the previous had increased bytousing
providehighmore combinations
power with both of the fre-
magnetic
quency/static
fields magneticthe
placed in-plane, field amplitudes,
vibration of theatpatch
whichmight
a reasonable response
be expected to be was observed.
larger. In fact,
Although the boundary conditions of the patch had changed between
the patch vibration pattern was similar to the magnetostriction curve of the ferromagneticboth experiments,
when theThe
material. magnetostrictive patchesincreased
vibration amplitude were used to generate
when guided
a high static wavesfield
magnetic for amplitude
inspection
purposes, the variation of the magnetic field directions and amplitude
was applied. The vibration velocity was inversely proportional to the frequency, and the could enhance the
signalwas
patch generation
vibrating and
at increase detectability
high amplitudes within of defects.
the low-frequency range.
 Compared to the previous experiments, by using high power with both magnetic
fields placed in-plane, the vibration of the patch might be expected to be larger. In fact,
the patch vibration pattern was similar to the magnetostriction curve of the ferromagnetic
material. The vibration amplitude increased when a high static magnetic field amplitude
Sensors 2022, 22, 766 was applied. The vibration velocity was inversely proportional to the frequency, 12 and the
of 15
patch was vibrating at high amplitudes within the low-frequency range.
Another set of experiments examined the relative amplitudes of the static (Bs) and
dynamic (Bdset
Another ) fields for particular
of experiments fixed values
examined the of dynamic
relative rf excitation
amplitudes frequency
of the static (B(f). Here,
s ) and
the amplitudes
dynamic (Bd ) fieldsoffor
vibration in the
particular X and
fixed valuesY directions
of dynamic were capturedfrequency
rf excitation while varying Bs for
(f). Here,
thea fixed dynamic
amplitudes field amplitude
of vibration in the XBdand
of 1.8 kW. The two
Y directions werefields were orthogonal
captured while varyingto each
Bs forother
a
to minimize
fixed dynamic Lorentz force generation,
field amplitude as the
Bd of 1.8 kW. study
The two on magnetostriction
fields were orthogonal generation was of
to each other
tointerest.
minimizeThe resultsforce
Lorentz are shown in Figure
generation, as the9a–d for on
study frequencies of 100, 120,
magnetostriction 200 and 220
generation waskHz,of
respectively.
interest. The results are shown in Figure 9a–d for frequencies of 100, 120, 200 and 220 kHz,
respectively.
0.02 0.022

Vibration in X direction Vibration in X direction

0.018 Vibration in Y direction Vibration in Y direction

0.02

0.016
0.018

0.014
0.016

0.012

Velocity (m/s)
Velocity (m/s)

0.014

0.01

0.012

0.008

0.01
0.006

0.008
0.004

0.002 0.006

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Static magnetic field amplitude (T) Static magnetic field amplitude (T)

(a) (b)
-3 -3
10 10
13 11

Vinration in X direction Vibration in X direction

12 Vibration in Y direction 10 Vibration in Y direction

11
9

10
8

9
7
Velocity (m/s)
Velocity (m/s)

6
7

5
6

4
5

4 3

3 2
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Static magnetic field amplitude (T) Static Magnetic field amplitude (T)

(c) (d)
Figure
Figure 9. 9. Magnetostrictive
Magnetostrictive patch
patch vibration
vibration in in XX
thethe and
and Y directions
Y directions asas a function
a function of of static
static magnetic
magnetic
field amplitude for (a) 100, (b) 120, (c) 200 and (d) 220 kHz.
field amplitude for (a) 100, (b) 120, (c) 200 and (d) 220 kHz.

AsAscancanbebeseen
seenininFigure
Figure 9, 9, the
the amplitudes
amplitudes in in both
bothdirections
directionsshowed
showednonlinear
nonlinear be-
havior with
behavior withincreasing
increasingstatic
staticmagnetic
magnetic field amplitude,
field amplitude, and this
and confirmed
this confirmed thethe
results from
results
the the
from previous set set
previous of experiments.
of experiments. ForFor
relatively low
relatively lowfrequencies,
frequencies,it was observed
it was observed that
thatthe
amplitude of the velocity in the Y direction saturated when applying
the amplitude of the velocity in the Y direction saturated when applying a static magnetic a static magnetic
field
field inin the
the range
range from
from 0.40.4
toto
0.60.6
T, T, with
with thethe amplitude
amplitude ofof the
the velocity
velocity decreasing
decreasing at at higher
higher
values. Conversely, when high frequencies were selected to drive the dynamic magnetic
field applied to the patch, the velocity in both X and Y directions had similar trends, as
the amplitude increased when the applied static magnetic field was increased. This trend
was similar to the magnetostrictive curve measured by applying a static magnetic field to
establish the properties of the ferromagnetic material. It is also important to mention that
while increasing the static magnetic field, the amplitude of the velocity in the Y direction
surpassed the amplitude of the velocity in the X direction. The “switching” point value
Sensors 2022, 22, 766 13 of 15

(circled in Figure 9 in each case) decreased as the frequency increased. In fact, the amplitude
in the Y direction became higher than in the X direction more quickly when increasing the
oscillation frequency.

4. Discussion
It was evident that the relative values of Bd and Bs had a significant effect on the
generation of magnetostriction within a thin magnetostrictive patch. First, the magnetic
field generated through the electromagnet was applied in the out-of-plane direction com-
pared to the magnetostrictive patch surface, while the dynamic magnetic field direction
was in-plane. The results showed that by modifying the frequency and the amplitude
of the static magnetic field, the vibration generated within the patch could be controlled
and enhanced. The analysis showed a complex link between rf coil frequency and the
static magnetic field amplitude. For the case where the frequency was around 100 kHz,
the variation of the applied magnetic field had a small effect on the vibration amplitude in
all directions. At higher frequencies, careful selection of the frequency and the static field
amplitude was required for this configuration, as the enhancement of the signal was not
achieved by simply increasing the static magnetic field amplitude. This could be seen for
frequencies such as 130 kHz and at frequencies above 200 kHz.
The second set of experiments consisted of placing both Bd and Bs in-plane, with the
two fields orthogonal to each other, so as to minimize the generation of Lorentz forces in the
patch, allowing magnetostriction effects to be studied in isolation. The power energizing
the driving coil was also set to two different values (450 W and 1.8 kW). In all cases, a
more complex interaction was observed. With a driving power of 450 W into the dynamic
coil, the performance of the patch was optimum only over certain frequency and static
magnetic field ranges. This could be due to the fact that the energy levels used to generate
the magnetostrictively induced vibrations were relatively low. The static magnetic field (Bs )
acted to align the magnetic domains in the direction of Bs . The modulation provided by Bd
was superimposed on this, and when the power induced was relatively low, the effect on the
rotation of the magnetic domains was very limited. It should be noted that the magnitude
of vibration in all directions was enhanced and, in some cases doubled, compared to the
out-of-plane configuration. This was due to the fact that thin magnetostrictive films were
used. Increasing the power energizing the dynamic coil to 1.8 kW showed a more consistent
trend. Increasing the amplitude of the dynamic field enhanced velocity amplitudes in the X
direction and the Y direction. It was also observed that higher velocities could be achieved
at higher drive frequencies, in this case at 240 kHz. In fact, by increasing the input power,
the energy applied to the magnetostrictive patch was increasing, which could facilitate a
wider operational range. For higher levels of Bd , the patch responded by generating higher
acoustic particle velocity amplitudes at higher frequencies.
In addition, by placing both magnetic fields so that the generation of the Lorentz
forces was minimized, a switching point was observed when varying the amplitudes of
the magnetic fields (Figure 9). For the low-frequency range (from 100 to 220 kHz), a higher
level of Bs was required so that the dominant vibrations were aligned with the Y direction.
The switching point between the magnitudes of vibration in the X and Y directions moved
to a lower frequency when the frequency was increasing. In fact, for higher frequencies,
the energy induced by the dynamic magnetic field was decreasing, which was leading to
the situation where the vibration of the magnetic domains was dominated by the static
magnetic field applied in the Y direction.

5. Conclusions
By using the laser vibrometer system, the vibration of a thin magnetostrictive patch
when subject to variable experimental parameters was analyzed and studied. Interesting
observations were identified by varying the static magnetic field direction and amplitude,
the dynamic magnetic field or the excitation frequency. It was demonstrated that by using
the same patch and modifying the excitation parameters, the signal generated from the
Sensors 2022, 22, 766 14 of 15

patch through the magnetostriction phenomena could be enhanced, and the patch was able
to generate higher vibration velocities and ultrasonic wave amplitudes. The amplitude of
the vibration could be altered by simply modifying the direction or amplitude of Bd and Bs .
In previous work, the frequency effect on magnetostriction was not well-established, as the
effect was only studied by applying a static magnetic field, without considering the effect
of the application of an oscillating magnetic field. It was shown that by placing the static
magnetic field in the out-of-plane direction and the selection of low operating frequencies,
the variation of the static magnetic field was minimal, but that the effect was significant
when the same field was in-plane. By further increasing the amplitude of the dynamic field
as well as the static magnetic field, the performance of the thin magnetostrictive patch could
be enhanced and controlled. A more global understanding and investigation of the effect
of these parameters will be of interest to the NDE research community, as such patches
can be widely used to overcome multiple limitations imposed by the use of piezoelectric
transducers for applications such as guided-wave inspection.

Author Contributions: Conceptualization, A.Z., D.H. and S.D; methodology, A.Z.; software, A.Z.;
validation, A.Z., D.H. and S.D.; formal analysis, A.Z.; investigation, A.Z.; resources, A.Z. and M.K.;
original draft preparation, A.Z.; writing—review and editing, A.Z., S.D., D.H. and M.K.; supervision,
D.H., S.D. and M.K. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded through the “NDTonAIR” Marie Skłodowska Curie Training
Network in Non-Destructive Testing and Structural Health Monitoring of Aircraft structures (MSCA-
ITN) under the action H2020-MSCA-ITN-2016- under Grant number 722134.
Institutional Review Board Statement: Not Applicable.
Informed Consent Statement: Not Applicable.
Data Availability Statement: Not Applicable.
Conflicts of Interest: The authors declare no conflict of interests.

References
1. Zhu, W.; Bian, L.X.; Cheng, L.; Rui, X.T. Non-linear compensation and displacement control of the bias-rate-dependent hysteresis
of a magnetostrictive actuator. Precis. Eng. 2017, 50, 107–113. [CrossRef]
2. Bieńkowski, A.; Kulikowski, J. The magneto-elastic Villari effect in ferrites. J. Magn. Magn. Mater. 1980, 19, 120–122. [CrossRef]
3. Witthauer, A.; Kim, G.; Faidley, L.; Zou, Q.; Wang, Z. Design and characterization of a flextensional stage based on Terfenol-D
actuator. Int. J. Precis. Eng. Manuf. 2014, 15, 135–141. [CrossRef]
4. Li, M.; Li, J.; Bao, X.; Mu, X.; Gao, X. Magnetostrictive Fe82 Ga13.5 Al4 .5 wires with large Wiedemann twist over wide temperature
range. Mater. Des. 2017, 135, 197–203. [CrossRef]
5. Chang, H.; Liao, S.; Hsieh, H.; Wen, J.; Lai, C.; Fang, W. Magnetostrictive type inductive sensing pressure sensor. Sens. Actuators A
Phys. 2016, 238, 25–36. [CrossRef]
6. Bieńkowski, A.; Szewczyk, R. Magnetostrictive Properties of Mn0.70 Zn0.24 Fe2.06 O4 Ferrite. Materials 2018, 11, 1894. [CrossRef]
7. Zhukov, A.; Ipatov, M.; Churyukanova, M.; Talaat, A.; Blanco, J.M.; Zhukova, V. Trends in optimization of giant magne-
toimpedance effect in amorphous and nanocrystalline materials. J. Alloys Compd. 2017, 727, 887–901. [CrossRef]
8. Ghosh, D. Structural Health Monitoring of Composite Structures Using Magnetostrictive Sensors and Actuators. Ph.D. Thesis,
Indian Institute of Science, Bengaluru, India, 2006.
9. Joule, J.P. On the effects of magnetism upon the dimensions of iron and steel bars. Phil. Mag. 1847, 111, 30–76.
10. Hirao, M.; Ogi, H. EMATs for Science and Industry: Noncontacting Ultrasonic Measurements; Springer Science & Business Media:
Berlin/Heidelberg, Germany, 2003.
11. Jiles, D. Introduction to Magnetism and Magnetic Materials; CRC Press: Boca Raton, FL, USA, 2015.
12. Seco, F.; Martín, J.M.; Jiménez, A.R. Improving the accuracy of magnetostrictive linear position sensors. IEEE Trans. Instrum.
Meas. 2008, 58, 722–729. [CrossRef]
13. Li, Y.; Sun, L.; Jin, S.; Sun, L.B. Development of magnetostriction sensor for on-line liquid level and density measurement. In
Proceedings of the 2006 6th World Congress on Intelligent Control and Automation, Dalian, China, 21–23 June 2006; pp. 5162–5166.
14. Gao, X.; Pei, Y.; Fang, D. Magnetomechanical behaviors of giant magnetostrictive materials. Acta Mech. Solida Sin. 2008, 21, 15–18.
[CrossRef]
15. Sablik, M.J.; Jiles, D.C. A model for hysteresis in magnetostriction. J. Appl. Phys. 1988, 64, 5402–5404. [CrossRef]
16. Butler, J.L. Application Manual for the Design of Terfenol-D Magnetostrictive Transducers; Edge Technologies Inc.: Ames, IA, USA, 1988.
Sensors 2022, 22, 766 15 of 15

17. Liang, X.; Dong, C.; Chen, H.; Wang, J.; Wei, Y.; Zaeimbashi, M.; He, Y.; Matyushov, A.; Sun, C.; Sun, N. A review of thin-film
magnetoelastic materials for magnetoelectric applications. Sensors 2020, 20, 1532. [CrossRef] [PubMed]
18. Hall, D.L.; Flatau, A.B. One-dimensional analytical constant parameter linear electromagnetic-magnetomechanical models of a
cylindrical magnetostrictive (Terfenol-D) transducer. J. Intell. Mater. Syst. Struct. 1995, 6, 315–328. [CrossRef]
19. Li, P.; Liu, Q.; Li, S.; Wang, Q.; Zhang, D.; Li, Y. Design and numerical simulation of novel giant magnetostrictive ultrasonic
transducer. Results Phys. 2017, 7, 3946–3954. [CrossRef]
20. Zhai, G.; Jiang, T.; Kang, L.; Wang, S. Minimizing influence of multi-modes and dispersion of electromagnetic ultrasonic Lamb
waves. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2010, 57, 2725–2733. [CrossRef] [PubMed]
21. Moffett, M.B.; Clark, A.E.; Wun-Fogle, M.; Linberg, J.; Teter, J.P.; McLaughlin, E.A. Characterization of Terfenol-D for magne-
tostrictive transducers. J. Acoust. Soc. Am. 1991, 89, 1448–1455. [CrossRef]
22. Gińko, O.; Juś, A.; Szewczyk, R. Test Stand for Measuring Magnetostriction Phenomena under External Mechanical Stress with
Foil Strain Gauges. In International Conference on Automation; Springer: Berlin/Heidelberg, Germany, 2016; pp. 843–853.
23. Weiss, P. L’hypothèse du champ moléculaire et la propriété ferromagnétique. J. Phys. Theor. Appl. 1907, 6, 661–690. [CrossRef]
24. Kim, Y.Y.; Kwon, Y.E. Review of magnetostrictive patch transducers and applications in ultrasonic nondestructive testing of
waveguides. Ultrasonics 2015, 62, 3–19. [CrossRef]
25. Dapino, M.J. On magnetostrictive materials and their use in adaptive structures. Struct. Eng. Mech. 2004, 17, 303–330. [CrossRef]
26. Joule, J.P. On a new class of magnetic forces. Ann. Electr. Magn. Chem. 1842, 8, 219–224.
27. Dimitropoulos, P.D.; Avaritsiotis, J.N. A micro-fluxgate sensor based on the Matteucci effect of amorphous magnetic fibers. Sens.
Actuators A Phys. 2001, 94, 165–176. [CrossRef]
28. Villari, E. Ueber die Aenderungen des magnetischen Moments, welche der Zug und das Hindurchleiten eines galvanischen
Stroms in einem Stabe von Stahl oder Eisen hervorbringen. Ann. Phys. 1865, 202, 87–122. [CrossRef]
29. Zitoun, A.; Dixon, S.; Edwards, G.; Hutchins, D. Experimental Study of the Guided Wave Directivity Patterns of Thin Removable
Magnetostrictive Patches. Sensors 2020, 20, 7189. [CrossRef]
30. Laguerre, L.; Aime, J.; Brissaud, M. Magnetostrictive pulse-echo device for non-destructive evaluation of cylindrical steel materials
using longitudinal guided waves. Ultrasonics 2002, 39, 503–514. [CrossRef]
31. Kim, Y.; Moon, H.; Park, K.; Lee, J. Generating and detecting torsional guided waves using magnetostrictive sensors of crossed
coils. NDT E Int. 2011, 44, 145–151. [CrossRef]
32. Juś, A.; Nowak, P.; Gińko, O. Assessment of the magnetostrictive properties of the selected construction steel. Acta Phys. Pol. A
2017, 131, 1084–1086. [CrossRef]
33. Wu, J.; Tang, Z.; Yang, K.; Lv, F. Signal strength enhancement of magnetostrictive patch transducers for guided wave inspection
by magnetic circuit optimization. Appl. Sci. 2019, 9, 1477. [CrossRef]
34. Lee, E.W. Magnetostriction curves of polycrystalline ferromagnetics. Proc. Phys. Soc. 1958, 72, 249. [CrossRef]
35. Hu, C.; Xu, J. Center frequency shift in pipe inspection using magnetostrictive guided waves. Sens. Actuators A Phys. 2019, 298,
111583. [CrossRef]
36. Kim, H.J.; Lee, J.S.; Kim, H.W.; Lee, H.S.; Kim, Y.Y. Numerical simulation of guided waves using equivalent source model of
magnetostrictive patch transducers. Smart Mater. Struct. 2014, 24, 015006.