Article

Investigation of the Flexural Behavior and Damage Mechanisms

of Flax/Cork Sandwich Panels Manufactured by Liquid
Thermoplastic Resin
Anas Ait Talaoul 1, * , Mustapha Assarar 1 , Wajdi Zouari 1 , Rezak Ayad 1 , Brahim Mazian 2 and Karim Behlouli 3

1 Institute of Thermics, Mechanics and Materials (ITheMM EA 7548), Université de Reims

Champagne-Ardenne, 51097 Reims, France; [email protected] (M.A.);
[email protected] (W.Z.); [email protected] (R.A.)
2 Centre Européen de Biotechnologie Et de Bioéconomie (CEBB), CentraleSupélec, Laboratoire de Génie Des
Procédés Et Matériaux, Université Paris-Saclay, 3 rue des Rouges Terres, 51110 Pomacle, France;
[email protected]
3 NatUp Fibres (Eco-Technilin, La French Filature, Lemaitre Demeestere), RD6015, 76190 Valliquerville, France;
[email protected]
* Correspondence: [email protected]

Abstract: This study investigates the flexural behavior of three sandwich panels composed of
an agglomerated cork core and skins made up of cross-ply [0,90]2 flax or glass layers with areal
densities of 100 and 300 g/m2 . They are designated by SF100, SF300, and SG300, where S, F, and G
stand for sandwich material, flax fiber, and glass fiber, respectively. The three sandwich materials
were fabricated in a single step using vacuum infusion with the liquid thermoplastic resin Elium® .
Specimens of these sandwich materials were subjected to three-point bending tests at five span
lengths (80, 100, 150, 200, and 250 mm). Each specimen was equipped with two piezoelectric sensors
to record acoustic activity during the bending, facilitating the identification of the main damage
mechanisms leading to flexural failure. The acoustic signals were analyzed to first track the initiation
and propagation of damage and, second, to correlate these signals with the mechanical behavior of
the sandwich materials. The obtained results indicate that SF300 exhibits 60% and 49% higher flexural
Citation: Ait Talaoul, A.; Assarar, M.;
Zouari, W.; Ayad, R.; Mazian, B.;
and shear stiffness, respectively, than SG300. Moreover, a comparison of the specific mechanical
Behlouli, K. Investigation of the properties reveals that SF300 offers the best compromise in terms of the flexural properties. Moreover,
Flexural Behavior and Damage the acoustic emission (AE) analysis allowed the identification of the main damage mechanisms,
Mechanisms of Flax/Cork Sandwich including matrix cracking, fiber failure, fiber/matrix, and core/skin debonding, as well as their
Panels Manufactured by Liquid chronology during the flexural tests. Three-dimensional micro-tomography reconstructions and
Thermoplastic Resin. J. Compos. Sci. scanning electron microscope (SEM) observations were performed to confirm the identified damage
2024, 8, 539. https://doi.org/ mechanisms. Finally, a correlation between these observations and the AE signals is proposed to
10.3390/jcs8120539 classify the damage mechanisms according to their corresponding amplitude ranges.
Academic Editor: Ahmed Koubaa
Keywords: bio-based sandwich panels; flax fiber; liquid thermoplastic resin; flexural behavior;
Received: 19 November 2024
damage investigation; acoustic emission
Revised: 6 December 2024
Accepted: 11 December 2024
Published: 17 December 2024

1. Introduction
Sandwich composites are essential in various industrial sectors, including aeronautics,
Copyright: © 2024 by the authors. shipbuilding, and transportation. Their low weight combined with good flexural rigidity
Licensee MDPI, Basel, Switzerland. allows for the design of robust, lightweight structures. Additionally, these materials are
This article is an open access article valued for their durability in harsh marine conditions, helping to reduce maintenance
distributed under the terms and
costs [1]. Traditionally, synthetic fibers like glass and carbon fiber [2–4] have been used
conditions of the Creative Commons
to reinforce sandwich structures, especially their skins. However, these synthetic fibers
Attribution (CC BY) license (https://
pose significant environmental drawbacks due to their non-recyclable nature and energy-
creativecommons.org/licenses/by/
intensive production. With growing environmental awareness, natural fibers are emerging
4.0/).

J. Compos. Sci. 2024, 8, 539. https://doi.org/10.3390/jcs8120539 https://www.mdpi.com/journal/jcs

J. Compos. Sci. 2024, 8, 539 2 of 21

as a sustainable alternative. They not only lower the carbon footprint of composites but
also offer promising mechanical properties, potentially replacing synthetic options like
glass fibers [5].
Flax fibers are renewable and recyclable and hence contribute to the development of
more environmentally friendly composites while maintaining specific competitive mechan-
ical properties compared with glass fiber-based composites [6,7]. Additionally, flax fibers
present excellent vibration damping and deformability while preserving high mechanical
performance, making them an ideal candidate to replace glass fibers in sandwich compos-
ites [8]. Consequently, flax fibers were selected to reinforce the skins of the sandwich panels
studied in this work.
Often used as a core material in sandwich structures, agglomerated cork is produced
from granules or the residue of cork, a natural material removed from the bark of cork
oaks [9,10]. NL10 cork, used in this work as the core in the studied sandwich panels, is a
specific variety of agglomerated cork supplied by Sicomin (Châteauneuf-les-Martigues,
France), and designed for technical applications [11]. This material is distinguished by its
unique mechanical properties, including its isotropic behavior under stress and its ability to
absorb shocks without undergoing significant plastic deformation [12]. Unlike conventional
materials, such as polyvinyl chloride (PVC) foams, cork is a renewable resource that also
provides excellent thermal and acoustic insulation properties.
The resin employed in this study is Elium® supplied by Arkema (Lacq, France), a
thermoplastic resin that stands out from traditional thermosetting resins like epoxy due to
its recyclability at room temperature and its rapid polymerization process [13]. In terms
of mechanical performance, it has been found that Elium® provides superior flexural and
tensile strength to epoxy while significantly reducing processing times—an essential ad-
vantage in industrial production settings, where efficiency is critical [14]. Moreover, Elium®
allows for repairs through reshaping at room temperature, a feature not achievable with
thermosetting resins such as epoxy. These characteristics make Elium® an environmentally
friendly and technically viable alternative for sandwich composites, particularly in marine
applications, where both durability and high performance are required [15].
Several research studies have focused on eco-friendly sandwich panels and their
applications. For example, the study by Jiang et al. [16] demonstrated, through flexural
mechanical characterization, that sandwich materials with flax/polypropylene (PP) skins
and polyethylene terephthalate (PET) foam cores can be used as side panels for trucks. Mak
et al. [8] compared the flexural behavior of two types of sandwich structures reinforced
with flax and glass fibers. Their results demonstrated that sandwich structures reinforced
with flax fibers exhibited strength and stiffness comparable to those reinforced with glass
fibers while providing superior deformability. Additionally, Castegnaro et al. [17] produced
a racing sailboat using biocomposite materials made from flax fibers and epoxy resin,
demonstrating the effectiveness of natural fibers in marine applications. Recently, Robin
et al. [18] developed alternative sandwich materials with flax composite skins and PET foam
cores for marine structures. They compared their mechanical performance to sandwich
materials with glass/polyester skins and PVC foam cores, which are already used in
marine construction. Their results showed lower flexural stiffness than the reference glass
fiber-based sandwich but promising levels of supported impact energy. In another study,
Sarasini et al. [19] evaluated the impact response of a sandwich panel with flax/epoxy
skins and an agglomerated cork core. Their results highlighted the important role of
cork in absorbing energy from low-velocity impacts, compared to conventional synthetic
foams, by reducing the extent of damage through its thickness. In another work, Mancuso
et al. [20] examined the flexural mechanical behavior of a sandwich composite made with
flax/epoxy skins and an NL10 cork core. Their results enabled the integration of this
material into the hull of a sailboat. To date, few studies have explored the use of Elium®
thermoplastic resin in the manufacturing of sandwich panels, unlike laminates, for which
many more studies are available in the literature [1,21–23]. Moreover, to the best of the
J. Compos. Sci. 2024, 8, 539 3 of 21

authors’ knowledge, no study has focused on sandwich materials combining flax/Elium®

skins with an agglomerated cork core.
Hence, our work aims to evaluate the mechanical performance of sandwich materials
reinforced with flax fibers, with the goal of proposing a sustainable alternative to synthetic
sandwich materials, particularly those reinforced with glass fibers. To this end, in this
paper, we study sandwich materials made with unidirectional (UD) flax/Elium® skins
and an NL10 agglomerated cork core. The flax fibers were provided by Eco-Technilin
(Valliquerville, France) and the NL10 cork was supplied by Sicomin (Châteauneuf-les-
Martigues, France). The [0,90]2 stacking sequence chosen for the skins is already considered
in the fabrication of commercial ships using glass fibers. Therefore, the main objective of
this study is to test the feasibility of replacing glass fibers with flax fibers while achieving
comparable flexural properties. Three sandwich materials were fabricated: SF100 and
SF300, both reinforced with UD flax fibers, with areal mass densities of 100 and 300 g/m2 ,
respectively, and SG300, reinforced with glass fibers, with an areal mass density of 300 g/m2 .
SF300 was selected to allow comparison with SG300, as both reinforcements have the same
areal mass density.
The three sandwich panels were fabricated using the “one-shot” vacuum resin infusion
technique. This method significantly enhances the quality of composite materials by
ensuring complete fiber impregnation and eliminating air bubbles that could compromise
mechanical properties. The mechanical performance of SF100, SF300, and SG300 was
assessed through tensile and three-point bending tests, along with acoustic emission (AE)
analysis to monitor damage progression during testing. In addition, tomographic and
microscopic observations were conducted to identify the primary damage mechanisms
appearing in the three sandwich materials. A correlation between these observations and
the acoustic emission data enabled the classification of the main damage modes based on
their amplitude ranges.

2. Experimental Procedure
2.1. Material and Manufacturing Process
In this study, three UD fabrics were selected as reinforcements for the skins of the
sandwich panels under investigation. The flax fabrics, designated as FlaxTape 100 and
FlaxTape 300, with an areal density of 100 g/m2 and 300 g/m2 , respectively, were provided
by Eco-Technilin (Valliquerville, France). Additionally, the 300 g/m2 glass fiber fabric
was supplied by Sicomin (Châteauneuf-les-Martigues, France). The fiber densities were
1450 kg/m3 for flax and 2450 kg/m3 for glass. The three fabrics will be referred to hereafter
as F100 for FlaxTape 100, F300 for FlaxTape 300, and G300 for glass 300. The core material in
the fabrication of the three sandwich materials consisted of agglomerated cork NL10, also
supplied by Sicomin (Châteauneuf-les-Martigues, France). Its thickness and density were
5 mm and 170 kg/m3 , respectively. The matrix used in this work was the thermoplastic
resin Elium® 151 XO/SA, with a density of 1010 kg/m3 , provided by Arkema (Lacq,
France). Table 1 summarizes the key properties of this resin.

Table 1. Properties of the thermoplastic resin Elium® 151 XO/SA, supplied by Arkema.

Property Units Value

Tensile Strength MPa 47.1
Tensile Modulus GPa 2.68
Tensile Elongation at Break - 4.50%
Flexural Strength MPa 80.51
Flexural Modulus GPa 2.77
Flexural Elongation at MaxStrength - 4.42

To characterize the flax/Elium® and glass/Elium® skins, [0]4 and [0,90]2 laminates
were fabricated using the vacuum infusion technique; Table 2 summarizes their properties.
The three sandwich materials are designated by SF100, SF300, and SG300 in this paper
J. Compos. Sci. 2024, 8, 539 4 of 21

(SF100 is the sandwich panel with skins reinforced by the F100 fabric). Figure 1 illustrates
the layer-stacking sequence of the three sandwich composites. The thicknesses of SF100,
SF300, and SG300 were 6.83 ± 0.07, 10.06 ± 0.1, and 6.87 ± 0.05 mm, respectively.

Table 2. Properties of the flax/Elium® and glass/Elium® laminates (Vf and Vp are the fiber volume
fraction and the porosity content, respectively).

Layers Laminates Thickness (mm) Vf (%) Vp (%)

[0]4 UD—F100 1.49 31 4.02
UD—F300 2.67 37 2.55
UD—G300 1.10 50 1.50
[0,90]2 Skin—F100 1.90 32 4.02
J. Compos. Sci. 2024, 8, x FOR PEER REVIEW 5 of 21
Skin—F300 2.81 38 2.55
Skin—G300 0.95 52 1.50

Release film

Flow Media

Peel ply

Fiber 0°
Fiber 90°

Fiber 0°

Fiber 90°

Cork NL 10

Fiber 90°

Fiber 0°
Fiber 90°
Fiber 0°

Peel ply

Flow Media

Glass mold

Figure1.1.Layer
Figure Layerstacking
stackingofofthe
thethree
threesandwich
sandwichpanels
panelsstudied
studiedininthis
thiswork.
work.

2.2. Sandwich
Static Tensile Testswith dimensions of 400 mm × 400 mm were fabricated. The pro-
panels
cess began
Tensilebytests
preparing the reinforcements:
were performed on the UD samples were
and [0,90] cut from fabric rolls (glass or
2 laminate samples at a constant
flax) and stacked to form a four-layer preform
crosshead speed of 2 mm/min using an Instron universal for the UD laminates
testingand [0,90]2 (Instron,
machine for the
sandwich skins. For the sandwich panels (Figure 1), the preform of the lower
Élancourt, France) with a load cell of 100 kN. The machine was equipped with two exten- skin was
placed first, followed by an NL10 cork sheet, and then the preform of the upper
someters to measure both longitudinal and transverse strains (Figure 2). For each config- skin was
positioned on top, allowing for a one-shot infusion process. The principle
uration, the mean value and standard deviation presented in this paper were calculated remains the
same: the dry constituents (fibers
from at least five tested specimens. for the skins and the cork core) are layered and then
placed under vacuum. Subsequently, the preform was impregnated with Elium® resin
via vacuum infusion. The decision to infuse the sandwich materials in a one-shot process
minimizes interface issues between the core and the skins and prevents the formation of
air bubbles [24]. Indeed, the assembly process requires the surface treatment of both the
core and the skins to enhance adhesion as well as the careful selection of the adhesive’s
viscosity and compatibility. Managing these parameters increases the likelihood of error.
For this reason, the decision was made to manufacture the three sandwich structures of this
study using a single, integral process. As already reported in the literature, this approach
improves cohesion between the skins and the core [25], enhances the mechanical properties
 Cork NL 10

J. Compos. Sci. 2024, 8, 539 Fiber 90° 5 of 21

Fiber 0°
Fiber 90°
compared to bonded sandwich panels [26], and significantly reduces both the production
time and cost [27]. Fiber 0°
A vacuum pressure of 0.8 bar was applied at room temperature to facilitate resin
Peel ply
infusion into the preform. This vacuum was used to advance the resin and impregnate the
components. A preliminary degassing of the resin was necessary to eliminate air trapped in
Flow Media
it. The polymerization phase was carried out at room temperature
Glass mold
for at least 8 h to allow
the proper formation of the composite skins and their adhesion to the cork core. Finally,
following the supplier’s recommendations, the sandwich panels were placed in an oven at
80 ◦ C for an additional hour to ensure maximum performance.
Tensile samples of the UD and [0,90]2 laminates were prepared in accordance with the
Figure
ISO 527-51.and
Layer
ISOstacking of the three
527-4 standards, sandwich
respectively. For panels studied
the flexural tests,inthe
this work.
SF100, SF300,
and SG300 samples were prepared in accordance with the NF T54-606 standard. All the
2.2. Static
samples wereTensile Tests
cut using a CNC digital cutting machine.

Tensile
2.2. Static Tensile tests
Tests were performed on the UD and [0,90]2 laminate samples a
crosshead speed
Tensile tests were of 2 mm/min
performed on theusing
UD and an[0,90]
Instron universal
2 laminate samples testing
at a con-machin
stant crosshead speed of 2 mm/min using an Instron universal testing
Élancourt, France) with a load cell of 100 kN. The machine was equipped machine (Instron, with
Élancourt, France) with a load cell of 100 kN. The machine was equipped with two ex-
someters to measure both longitudinal and transverse strains (Figure 2). For e
tensometers to measure both longitudinal and transverse strains (Figure 2). For each
uration, thethe
configuration, mean
mean value and
value and standard
standard deviation
deviation presentedpresented in were
in this paper this calcu-
paper wer
fromfrom
lated at least
at leastfive
five tested specimens.
tested specimens.

Figure
Figure 2. UD-F100
2. UD-F100 laminate
laminate in the
in the tensile tensile
test. test.
2.3. Three-Point Bending Tests
2.3.Three-point
Three-Point Bending
bending Testsperformed on the sandwich specimens of SF100, SF300,
tests were
and SG300, measuring 300
Three-point mm × 30
bending mm, were
tests in accordance with theon
performed NF T54-606 standard, at
the sandwich a
specimen
constant speed of 2 mm/min. The tests were performed using a load cell with a maximum
SF300, and SG300, measuring 300 mm × 30 mm, in accordance with the NF T54
capacity of 100 kN (Instron, Élancourt, France). Three support span lengths were tested:
ard,150,
100, atand
a constant speed3).ofFor2 each
250 mm (Figure mm/min. The testsmaterial,
type of sandwich were performed using a load
at least five samples
maximum
were capacity
tested through of 100 kN (Instron, Élancourt, France). Three support sp
bending.
J. Compos. Sci. 2024, 8, x FOR PEER REVIEW 6 of 2

J. Compos. Sci. 2024, 8, 539 were tested: 100, 150, and 250 mm (Figure 3). For each type of sandwich material,
6 of 21 at leas
five samples were tested through bending.

Figure3.3.SG300
Figure SG300 sample
sample in the
in the three-point
three-point bending
bending test. test.

2.4. Acoustic Emission Acquisition System

2.4. Acoustic Emission Acquisition System
Monitoring and identifying the primary damage mechanisms in composite materials
Monitoring and identifying the primary damage mechanisms in composite material
under mechanical loading is crucial for evaluating their reliability and durability during
under mechanical
service. Mechanical loading is crucial
stress causes for evaluating
localized damage thattheir reliability
generates and durability
detectable transient during
elastic waves that vary in nature and frequency. In this context, the acoustic emission (AE)transien
service. Mechanical stress causes localized damage that generates detectable
elastic waves
technique that vary
was chosen in nature
to monitor andand frequency.
identify In this
the primary context,
damage the acoustic
mechanisms emission (AE
appearing
in the sandwich
technique was specimens
chosen toduring
monitor theand
bending tests.the primary damage mechanisms appear
identify
Thethe
ing in recording
sandwich of acoustic signals
specimens from the
during the sandwich
bending specimens
tests. was carried out using
two PAC Micro-80 piezoelectric sensors, positioned on the upper
The recording of acoustic signals from the sandwich specimens surfaces of was
the specimens
carried out using
with silicone grease (Figure 3). These sensors resonate within a frequency range of 100 kHz
two PAC Micro-80 piezoelectric sensors, positioned on the upper surfaces of the speci
to 1 MHz, with a resonance peak around 300 kHz. An EA PCI-2 acquisition system
mens with silicone grease (Figure 3). These sensors resonate within a frequency range o
developed by Mistras Group (Sucy-en-Brie, France) was used to convert the electrical
100 kHz
signals toacoustic
into 1 MHz, signals.
with a resonance peak around
This two-channel system 300 kHz.
offers An EA PCI-2
a sampling rate ofacquisition
5 MHz sys
tem developed by Mistras Group (Sucy-en-Brie, France) was used to
and a preamplification of 40 dB. The data acquisition system was calibrated using the convert the electrica
signals into
pencil-lead acoustic
break signals. This two-channel system offers a sampling rate of 5 MHz and
procedure.
a preamplification of 40 dB. The data acquisition system was calibrated using the pencil
2.5. Micro-Tomographic and SEM Observations
lead break procedure.
To thoroughly analyze the damage mechanisms in the materials, tomographic and
microscopic observationsand
2.5. Micro-Tomographic were conducted
SEM on the sandwich samples following the three-
Observations
point bending tests. For the scanning electron microscope (SEM) observations, a JSM-7900F
To Field
Schottky thoroughly
Emissionanalyze the damage by
SEM, manufactured mechanisms
JEOL-Europe in SAS
the materials, tomographic
(Croissy, France) was and
microscopic
used. observations
Small samples were10
measuring conducted
mm × 10on mm thewere
sandwich
carefullysamples
cut fromfollowing
the largerthe three
point bending
specimens, dried,tests. For the
metallized, andscanning
placed in electron
a vacuummicroscope (SEM)high-resolution
chamber to ensure observations, a JSM
imaging and enhanced
7900F Schottky Field SEM observation
Emission quality. These preparation
SEM, manufactured steps were
by JEOL-Europe SAS essential to France
(Croissy,
capture
was used.the fine details
Small of the measuring
samples fracture surfaces
10 mm and×microstructural damage. cut from the large
10 mm were carefully
specimens, dried, metallized, and placed in a vacuum chamber to by
For the micro-tomography analysis, an EasyTom XL Ultra 150–160, RX Solutions
ensure high-resolution
(Chavanod, France), was used. A microfocal source (150 kV, 75 W, 5 µm) and an amorphous
imaging and enhanced SEM observation quality. These preparation steps were essentia
silicon array sensor were selected to provide a balance between the sample size and the
to capture the fine details of the fracture surfaces and microstructural damage.
spatial resolution. The acquisition protocol consisted of 1568 projections at a voltage of
40 kV,For the micro-tomography
a current analysis,
of 500 µA, and an exposure an of
time EasyTom XL Ultra
5 s. The total 150–160,
scanning by RX Solution
time exceeded
2(Chavanod, France),
h. The resolution of was used. A microfocal
the reconstructed 3D imagessourcewas(150
5 µm. kV,The
75 W, 5 μm) and
projections werean amor
reconstructed into image stacks using X-Act software (RX Solutions, Chavanod, France) size and
phous silicon array sensor were selected to provide a balance between the sample
and
the further
spatialprocessed
resolution.with
TheAvizo 2019 (Thermo
acquisition Scientific,
protocol Waltham,
consisted MA,projections
of 1568 USA). at a voltage
of 40 kV, a current of 500 μA, and an exposure time of 5 s. The total scanning time ex
3. Results and Discussion
ceeded 2 h. The resolution of the reconstructed 3D images was 5 μm. The projections wer
3.1. Tensile Properties of the Laminates
reconstructed into image stacks using X-Act software (RX Solutions, Chavanod, France
To analyze the mechanical behavior of the UD and [0,90]2 laminate samples, Figure 4
and further processed with Avizo 2019 (Thermo Scientific, Waltham, MA, USA).
presents their stress–strain curves obtained from the tensile tests. The typical behavior of
 3. Results and Discussion

J. Compos. Sci. 2024, 8, 539

3.1. Tensile Properties of the Laminates
7 of 21
To analyze the mechanical behavior of the UD and [0,90]2 laminate samples, Figure 4
presents their stress–strain curves obtained from the tensile tests. The typical behavior of
the composites
the reinforced with
composites reinforced with glass
glass and
and flax
flax fibers
fibers can
can be
be observed
observed [28–30].
[28–30]. The
The UD-G300
UD-G300
laminate exhibits an almost linear response up until brittle failure. In contrast,
laminate exhibits an almost linear response up until brittle failure. In contrast, the UD-F100 the UD-
and UD-F300 laminates show two distinct linear regions (Figure 4a). The first region,
F100 and UD-F300 laminates show two distinct linear regions (Figure 4a). The first region,
purely elastic,
purely elastic, involves
involves strains
strains not
not exceeding
exceeding 0.2%,0.2%, allowing
allowing for the measurement
for the measurement of of the
the
longitudinal modulus.
longitudinal modulus. The The second
second linear
linear region
region extends
extends upup to
to failure, representing over
failure, representing over
70% ofof the
thetotal
totalbehavior
behaviorofofthese
these two
two laminates.
laminates. Bensadoun
Bensadoun et [29]
et al. al. [29]
andand Charlet
Charlet et al.et[31]
al.
[31] attributed this change in slope for the UD-F100 and UD-F300
attributed this change in slope for the UD-F100 and UD-F300 laminates to specific mecha- laminates to specific
mechanisms
nisms occurringoccurring at the microscopic
at the microscopic level individual
level within within individual flaxPhenomena
flax fibers. fibers. Phenomena
such as
suchalignment
the as the alignment of microfibrils
of microfibrils withaxis
with the fiber theduring
fiber axis
axialduring
loading,axial
theloading, the rear-
rearrangement of
rangement of
amorphous amorphous
regions within regions withinorthe
the cell walls, thecell walls,
elastic or the elastic
response of alignedresponse of aligned
microfibrils may
microfibrils
explain thesemay
twoexplain these two
linear phases [31].linear phases [31].

700 400

UD Glass 300 [0,90]2

Glass 300
525 300

Stress (MPa)
Stress (MPa)

350 200

Flax 300 Flax 300

175 100
Flax 100 Flax 100

0 0
0,0 0,5 1,0 1,5 2,0 2,5 0,0 0,5 1,0 1,5 2,0 2,5
Strain (%) Strain (%)

(a) (b)
Figure 4.
Figure 4. Tensile curves of
Tensile curves of (a)
(a) the
the UD
UD laminates
laminates and
and (b)
(b) the
the [0,90]
[0,90]2 sandwich
sandwich skins.
skins.
2

mechanicalproperties
The mechanical properties obtained
obtained fromfrom these
these curves
curves are summarized
are summarized in Tables
in Tables 3 and 4.3
and 4.show
They Theythatshow thethat the longitudinal
longitudinal modulus modulus
of UD-G300of UD-G300
is 30% and is 30%
37%and 37%than
higher higher than
those of
those of UD-F300
UD-F300 and UD-F100,
and UD-F100, respectively.
respectively. It is also It is also
noted noted
that that the longitudinal
the longitudinal modulusmodu-of the
lus offiber
glass the glass
skins fiber
is 48%skins is 48%
to 67% to 67%
higher thanhigher
those of than
thethose
F300 andof theF100F300 andrespectively.
skins, F100 skins,
The tensile strengths
respectively. of UD-G300
The tensile strengthsand Skin-G300
of UD-G300 andareSkin-G300
also 60% to are70%
alsohigher
60% to than
70%those of
higher
the
thanflax fiber
those of laminates. This
the flax fiber is attributed
laminates. Thisto is the inherently
attributed to thesuperior mechanical
inherently superiorproperties
mechan-
of
icalglass fibers compared
properties to flax
of glass fibers fibers, as to
compared wellflaxasfibers,
to the asdifferences
well as tointhe
fiber volume fractions
differences in fiber
between the flax/Elium ® and glass/Elium ® laminates (32%, 38%, and 50% for UD-F100,
volume fractions between the flax/Elium and glass/Elium laminates (32%, 38%, and 50%
® ®

UD-F300,
for UD-F100, andUD-F300,
UD-G300, andrespectively). It is important
UD-G300, respectively). It istoimportant
note that to thenote
timethat
required
the timeto
impregnate a stack of four
required to impregnate glass
a stack layers
of four is twice
glass layersasisshort
twiceas as that
shortforas flax
that layers.
for flax This is
layers.
because the manufacturing
This is because process process
the manufacturing of glass of fibers
glassinvolves a sizing step,
fibers involves which
a sizing creates
step, which a
physico-chemical bond between
creates a physico-chemical bondthe fibers and
between the resin,
the fibers and improving the qualitythe
the resin, improving andquality
speed
of
andimpregnation, as well as protecting
speed of impregnation, as well as the fiber from
protecting theits environment.
fiber from its environment.

Table 3. Tensile
Table 3. Tensile properties
propertiesof
ofthe
theUD
UDlaminates.
laminates.

Laminates
Laminates 𝐄 (GPa)
EL𝑳 (GPa) 𝐄E𝑻T (GPa)
(GPa) 𝐆
G𝑳𝑻 (GPa)
LT (GPa) 𝛔σ𝑹R (MPa)
(MPa) 𝛆ε𝑹R (%)
(%)
UD-G300
UD-G300 32.2
32.2 ±±1.7
1.7 12.2
12.2 ± ±0.1
0.1 4.3 ± 0.3
4.3 ± 0.3 578 ± 57
578 ± 57 2.1
2.1 ±± 0.1
0.1
UD-F300
UD-F300 22.1 ±
22.1 ±1.7
1.7 1.7 ±
1.7 ±0.1
0.1 3.1 ±
3.1 ± 0.4
0.4 212 ±
212 ± 18
18 1.2 ±
1.2 ± 0.2
0.2
UD-F100
UD-F100 20.2 ± 1.8
20.2 ± 1.8 2.7 ±
2.7 ±0.1
0.1 1.6 ± 0.2
1.6 ± 0.2 211 ± 27
211 ± 27 1.3 ± 0.1
1.3 ± 0.1

Table 4. Tensile properties of the [0,90]2 skin laminates.

Laminates EL =ET (GPa) σR (MPa) εR (%)

Skin—G300 24.6 ± 0.9 353 ± 41 1.8 ± 0.3
Skin—F300 12.6 ± 0.9 108 ± 3 1.3 ± 0.1
Skin—F100 8.1 ± 0.2 101 ± 13 1.6 ± 0.3
J. Compos. Sci. 2024, 8, 539 8 of 21

3.2. Bending Properties of SF100, SF300, and SG300

Figure 5 shows the force–deflection curves from the three-point bending tests per-
formed on the SF100, SF300, and SG300 samples for the three support span lengths of 100,
150, and 250 mm. The analysis of these curves reveals the typical behavior for UD fiber-
reinforced sandwich structures, which is initially linear elastic, followed by a non-linear
behavior until failure [32]. This non-linearity is more pronounced in SF300, especially for
the 150 mm and 250 mm support spans. Additionally, the stiffness and maximum load of
the three sandwich materials increase as the support span decreases, while the deflection
at failure increases with the support span. These bending curves enable the estimation of
stiffness, maximum load, and maximum deflection of each sandwich panel, as summarized
in Table 5. The obtained results show that SF300 outperforms the two other sandwich
materials across the three support spans, with its flexural stiffness being 26% to 50% higher
than that of SG300 and 66% to 77% higher than that of SF100. Furthermore, the maximum
load and deflection at failure of SF300 exceed those of SF100 and SG300. The superior
mechanical performance of SF300 is largely due to its skins being 33% thicker than those of
SF100 and SG300. Furthermore, although the skins of SF100 and SG300 are nearly identical
in thickness, the performance of SG300 is 30% to 65% superior to that of SF100, particularly
in terms of flexural stiffness and maximum load. This difference can be attributed to the
superior
J. Compos. Sci. 2024, 8, x FOR PEER REVIEW mechanical properties of glass fibers compared to flax fibers, as demonstrated
9 of 21 in
Table 3 through the tensile characterization of the skins.

750 750
SF100 L = 100 SG300
600 600
Load (N)
Load (N)

450 450 L = 150

L = 100
300 300
L = 150
L = 250
150 L = 250 150

0 0
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
Displacement (mm) Displacement (mm)
(a) (b)
750
SF300
600 L = 100
Load (N)

450

300 L = 150

150 L = 250

0
0 5 10 15 20 25 30 35
Displacement (mm)
(c)

Figure
Figure 5. 5.Load–deflection curvesofof
Load–deflection curves (a)(a) SF100,
SF100, (b) SG300,
(b) SG300, and
and (c) (c) SF300
SF300 from
from the the three-point
three-point bending bend-
ingtests.
tests.

We now determine the equivalent flexural and shear stiffnesses of the three sandwich
materials based on the classical theory of sandwich beams. In the linear domain of the
load–deflection curve, the relationship between the applied load 𝑃 and the maximum
deflection at the midpoint of the sandwich specimen, denoted as 𝑤𝑐 , for a support span
𝐿, is given by the following equation [33]:
𝑤𝑐 𝐿² 1
= + (1)
𝑃𝐿 48𝐷 4𝑁
J. Compos. Sci. 2024, 8, 539 9 of 21

Table 5. Results from the bending curves of SF100, SF300, and SG300 (wc max represents the maxi-
mum deflection).

Span Length (mm) Sandwich Stiffness (N/mm) Fmax (N) wc max (mm)
L = 100 SG300 169 ± 20 607 ± 81 9.1 ± 1.3
SF300 360 ± 29 882 ± 148 14.1 ± 0.3
SF100 121 ± 8 313 ± 28 12.9 ± 0.8
L = 150 SG300 92 ± 9 505 ± 33 12.1 ± 2.5
SF300 159 ± 66 760 ± 38 17.9 ± 3.5
SF100 42 ± 1 223 ± 1 17.1 ± 1.4
L = 250 SG300 27 ± 2 365 ± 60 16.8 ± 1.7
SF300 56 ± 9 562 ± 20 29.1 ± 4.4
SF100 13 ± 1 130 ± 4 26.7 ± 2.8

We now determine the equivalent flexural and shear stiffnesses of the three sandwich
materials based on the classical theory of sandwich beams. In the linear domain of the
load–deflection curve, the relationship between the applied load P and the maximum
deflection at the midpoint of the sandwich specimen, denoted as wc , for a support span L,
is given by the following equation [33]:

wc L² 1
= + (1)
PL 48D 4N

D, in N.mm2 , represents the equivalent flexural stiffness of the sandwich, while N denotes
its equivalent shear stiffness. Figure 6 shows the evolution of wc /PL as a function of L2
(the square of the span length) for the three sandwich composites. It should be noted that
two additional support spans (80 and 220 mm) were added to refine the determination of
these two properties According to Equation (1), the slopes and y-intercepts of these curves
allow for the determination of the equivalent stiffnesses, D and N, and their values are
summarized in Table 6. We also determined the equivalent flexural modulus Ex = D/I
and shear modulus Gxz = N/S, where I and S denote the second moment of area and
the total cross-sectional area of the sandwich specimen, respectively. As shown in Table 6,
the obtained results highlight the superior performance of SF300 in terms of equivalent
flexural and shear stiffness. Table 6 also indicates that SG300 presents an equivalent flexural
modulus that is 18% higher than that of SF300 and 50% higher than that of SF100. However,
SF300 exhibits a shear modulus that is 43% higher than that of SF100 and 29% higher than
that of SG300. This result can be primarily explained by the thickness of the SF300 skins,
which is 30% greater than that of SF100 and SG300. Indeed, an increase in skin thickness
leads to greater flexural stiffness, which imposes significant shear stress on the core. This
observation is further supported by the fact that SG300 and SF100 have almost identical
thicknesses, resulting in a comparable shear modulus.

Table 6. Equivalent flexural and shear stiffnesses of SF100, SF300, and SG300.

Sandwich D (106 N.mm²) N (kN) Ex (GPa) Gxz (MPa)

SF100 5.2 5.1 6.5 24
SF300 26.1 12.5 10.5 42
SG300 10.4 6.3 12.8 30

To better compare the flexural performance of SG300, SF300, and SF100, Figure 7
presents the radar diagrams. These diagrams illustrate the variation in the five parameters
normalized by density, based on the theoretical expression of the performance index
including the square of the density [33]. The parameters considered in this comparison
are the flexural modulus Ex , the shear modulus Gxz , the maximum deflection wc max , the
maximum load Fmax, and the flexural stiffness (defined as the linear slope of the load–
J. Compos. Sci. 2024, 8, 539 10 of 21

J. Compos. Sci. 2024, 8, x FOR PEER deflection

REVIEW curve). These diagrams reveal that the normalized maximum deflection10ofofSF100 21
is 50% and 75% higher than those of SF300 and SG300, respectively. Moreover, the radar
diagram of the SF300 sandwich highlights its notable specific mechanical performance,
particularly
modulus that in terms
is 18%of flexural
higher thanmodulus and stiffness.
that of SF300 and 50% To identify
higher than the
thatsandwich material
of SF100. How-
that offers the best compromise among the five selected normalized characteristics,
ever, SF300 exhibits a shear modulus that is 43% higher than that of SF100 and 29% higher the area
ofthan
the radar diagram for each sandwich was calculated. The results indicate
that of SG300. This result can be primarily explained by the thickness of the SF300 that the radar
diagram area of
skins, which is SF300 is 19%than
30% greater larger
thatthan that ofand
of SF100 SF100. Additionally,
SG300. Indeed, an the areasinofskin
increase SF300
and SF100 are
thickness leads48% and 36%
to greater largerstiffness,
flexural than thatwhich
of SG300, respectively.
imposes significantThese
shear results
stress onsuggest
the
core.
that This
flax observation
fiber-based is further
sandwich supported
materials by the
could fact thataSG300
constitute and SF100
promising have almost
alternative to SG300
identical
based thicknesses,
on the resulting
five selected in a comparable shear modulus.
parameters.
35 16

30 SF100 14 SG300

12
25
-1)

-1)
10
wc / PL (105.N

wc / PL (105.N
20
8
15
6

10
4

5 2

0 0
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70

L 2 ( 10 3 mm 2 ) L 2 ( 10 3 mm 2 )
(a) (b)
8

SF300

6
-1)
wc / PL (105.N

0
0 10 20 30 40 50 60 70

L 2 ( 10 3 mm 2 )
(c)

Figure6.6.Determination
Figure Determination of
of the
the equivalent flexuraland
equivalent flexural andshear
shearstiffnesses
stiffnesses
ofof SF100
SF100 (a),(a), SG300
SG300 (b),(b), and
SF300 (c). (c).
and SF300

Table 6. Equivalent flexural and shear stiffnesses of SF100, SF300, and SG300.

Sandwich D (𝟏𝟎𝟔 N.mm²) N (kN) 𝑬𝒙 (GPa) 𝑮𝒙𝒛 (MPa)

SF100 5.2 5.1 6.5 24
SF300 26.1 12.5 10.5 42
SG300 10.4 6.3 12.8 30

To better compare the flexural performance of SG300, SF300, and SF100, Figure 7 pre-
sents the radar diagrams. These diagrams illustrate the variation in the five parameters
 tics, the area of the radar diagram for each sandwich was calculated. The results indi
that the radar diagram area of SF300 is 19% larger than that of SF100. Additionally,
areas of SF300 and SF100 are 48% and 36% larger than that of SG300, respectively. Th
results suggest that flax fiber-based sandwich materials could constitute a promising
J. Compos. Sci. 2024, 8, 539 ternative to SG300 based on the five selected parameters. 11 of 21

Ex / r2
(10-4 N.kg-2.m4) G300
2,0 F300
F100
1,5

wc max / r2 1,0
Slope / r²
4 2,0
8 -2 7
(10 kg .m ) 3
2 0,5
1,5 (102N.kg-2.m5)
1,0
1 0,5

2
2
4
4
6
6
Fmax / r2 8 Gxz / r2
4 -2 6 8
(10 N.kg .m ) (10-1N.kg-2.m4)
Figure
Figure 7. Radar 7. Radar
diagram diagram
of SF100, of SF100,
SF300, SF300,
and SG300, asand SG300,
tested as tested
through through
three-point three-point bending.
bending.

3.3. Acoustic 3.3.

Emission Analysis
Acoustic Emission Analysis
This sectionThis aimssection
to identify
aims tothe main the
identify damage
main mechanisms in the three
damage mechanisms in thestudied
three studied sa
sandwich materials that occurred during the three-point bending tests, such as matrix
wich materials that occurred during the three-point bending tests, such as matrix
microcracking, fiber/matrix debonding, fiber pull-out, and fiber breakage. The acoustic
crocracking, fiber/matrix debonding, fiber pull-out, and fiber breakage. The acoustic
signals recorded during the bending tests were analyzed using NOESIS software [34] to
nals recorded during the bending tests were analyzed using NOESIS software [34] to id
identify the main damage mechanisms and their evolution throughout the tests.
tify the main damage mechanisms and their evolution throughout the tests.
Figure 8 illustrates the evolution of the cumulative acoustic activity and applied load
Figure 8 illustrates the evolution of the cumulative acoustic activity and applied l
over time during the bending tests for the three sandwich materials, SF100, SF300, and
over time during the bending tests for the three sandwich materials, SF100, SF300,
SG300, and the two support spans of 100 mm and 250 mm. These graphs can be globally
SG300, and the two support spans of 100 mm and 250 mm. These graphs can be glob
divided into two or three distinct phases. The first one is characterized by minimal acoustic
divided into two or three distinct phases. The first one is characterized by minimal aco
activity, reflecting the linear elastic behavior of SF300 and SG300. However, in the case of
tic activity, reflecting the linear elastic behavior of SF300 and SG300. However, in the
SF100, this phase is prolonged to encompass the advent of its non-linear behavior. The
of SF100, this phase is prolonged to encompass the advent of its non-linear behavior.
second phase shows an increase in acoustic activity, indicating the onset of the non-linear
second
flexural behavior ofphase
SF300shows an increase
and SG300. Duringin this
acoustic activity,
phase, indicating
the damage the onset
is more of the non-lin
extensive,
flexural behavior of SF300 and SG300. During this phase, the damage
and the acoustic signals are likely to be due to damage of the upper skin in the proximity is more extens
of the loading zone. The third phase is marked by a sharp increase in cumulative hits, a
stabilization of the applied load for SF300, and rigidity loss for SF100 and SG300. This
can be attributed to various damage mechanisms, such as fiber breakage and fiber/matrix
interface failure, occurring during the final stages of the bending tests. Furthermore, SF100
and SG300 exhibit distinct acoustic activity for the support span 250 mm, principally
showing a two-phase evolution. In both cases, the second phase is characterized by a
progressive increase in the number of hits leading up to failure. However, the onset of
acoustic activity in SF100 occurs significantly later than that in SG300, specifically at a load
of 200 N for L = 100 mm and 60 N for L = 250 mm. This delay could be attributed to the
flexural behavior of SG300, which is influenced by the brittle behavior of its skins.
 and SG300 exhibit distinct acoustic activity for the support span 250 mm, principally
showing a two-phase evolution. In both cases, the second phase is characterized by a pro-
gressive increase in the number of hits leading up to failure. However, the onset of acous-
tic activity in SF100 occurs significantly later than that in SG300, specifically at a load of
J. Compos. Sci. 2024, 8, 539 200 N for L=100 mm and 60 N for L=250 mm. This delay could be attributed to the flexural 12 of 21
behavior of SG300, which is influenced by the brittle behavior of its skins.

EA Load
4000 EA Load 250 400 140
3500 350 120
200
3000 300
100
Number of Hits

Number of Hits
2500 250

Load (N)
150 80

Load (N)
2000 200
60
1500 100 150
40
1000 100
50 20
50
500
0 0
0 0
0 250 500 750 1000 1250
0 100 200 300 400 500 600 700
Time (s)
Time (s)

(a) (b)
EA Load
18000 EA Load 800
2500 500
16000 700
450
14000 600 2000 400
Number of Hits

Number of Hits
12000 350
500
Load (N)

Load (N)
10000 1500 300
400 250
8000
300 1000 200
6000
150
4000 200
500 100
2000 100 50
0 0 0 0
0 50 100 150 200 250 0 50 100 150 200 250 300 350 400
Time (s) Time (s)

(c) (d)
EA Load
14000 EA Load 700
7000 400
12000 600 350
6000
10000 500 300
Number of Hits

Number of Hits

5000
250

Load (N)
Load (N)

8000 400 4000

200
6000 300 3000
150
4000 200 2000
100
2000 100 1000 50

0 0 0 0
0 50 100 150 200 250 300 350 0 100 200 300 400 500

Time (s) Time (s)

(e) (f)
Figure 8.
Figure Acousticevents
8. Acoustic eventsofof SF100
SF100 (a,b),
(a,b), SG300
SG300 (c,d),
(c,d), andand SF300
SF300 (e,f)(e,f) in synchronization
in synchronization withwith
the
the applied
applied load load during
during the three-point
the three-point bendingbending tests,support
tests, with with support
spans ofspans
100 mmof 100 mm
(a,c,e) (a,c,e)
and and
250 mm
250 mm (b,d,f).
(b,d,f).

To identify,
To identify, separate,
separate, and classify the
and classify the signals
signals recorded
recorded during
during the
the bending
bending tests,
tests, the
the
data were represented in a multidimensional space using their frequency and
data were represented in a multidimensional space using their frequency and time pa-time parame-
ters. TheseThese
rameters. data were then normalized
data were to ensure
then normalized an equal
to ensure an statistical weightweight
equal statistical for eachfor
feature.
each
Signal classification was performed using the K-means algorithm, which minimizes the
Euclidean distance between signals within the same group. To this end, five parameters
(duration, energy, counts to peak, amplitude, and counts) were chosen to classify the
recorded acoustic signals [35].
Figure 9 presents the classification results for the three sandwich materials based on
three classes. The evolution of these classes can be divided into two or three phases, similar
to those observed in Figure 8. The results in Figure 9 indicate that the proportion of signals
in each class varies depending on the material and the fixed support span. The first class is
the most prevalent, with an activity rate of approximately 60% to 90%, depending on the
sandwich material and the span length. The second class exhibits a ratio ranging from 10%
to 30%, depending on the material and the span length. Finally, the third class, representing
less than 10% of the total acoustic activity, primarily emerges during the final phase of the
bending tests.
 signals in each class varies depending on the material and the fixed support span. The
first class is the most prevalent, with an activity rate of approximately 60% to 90%, de-
pending on the sandwich material and the span length. The second class exhibits a ratio
ranging from 10% to 30%, depending on the material and the span length. Finally, the
J. Compos. Sci. 2024, 8, 539 third class, representing less than 10% of the total acoustic activity, primarily 13
emerges
of 21
during the final phase of the bending tests.

SF100 SF100 Class 1

Class 1
L=250mm L=100mm

Class 2

Class 2
Class 3

Class 3

SG300 SG300
Class 1
L=250mm L=100mm Class 1

Class 2
Class 2

Class 3 Class 3

SF300 Class 1
SF300 Class 1
L=250mm L=100mm
Class 2

Class 2
Class 3

Class 3

Figure
Figure9.9.Evolution
Evolutionof
ofthe
the cumulative numberof
cumulative number ofhits
hitsduring
duringthe
thethree-point
three-point bending
bending tests.
tests.

ToToassociate
associatethese
thesethree
three classes
classes with
with thethe corresponding
corresponding damage
damage mechanisms,
mechanisms, an an
anal-
analysis of the signal parameters, particularly focusing on their amplitudes, was conducted.
ysis of the signal parameters, particularly focusing on their amplitudes, was conducted.
In this context, Figure 10 presents histograms showing the distribution of the number of
hits as a function of the amplitude of the acoustic signals recorded during the three-point
bending tests on SF100, SF300, and SG300. Microscopic observations were also performed
on several samples from these three sandwich materials to further support this association.
 In this context, Figure 10 presents histograms showing the distribution of the number of
hits as a function of the amplitude of the acoustic signals recorded during the three-point
bending tests on SF100, SF300, and SG300. Microscopic observations were also performed
J. Compos. Sci. 2024, 8, 539 14 of 21
on several samples from these three sandwich materials to further support this associa-
tion.

Class 1
SF100 Class 1
SF100
L=250mm L=100mm

Class 2
Class 2

Class 3 Class 3

SG300 SG300
Class 1 L=250mm L=100mm
Class 1

Class 2
Class 2
Class 3 Class 3

SF300 SF300
Class 1
L=250mm L=100mm
Class 1

Class 2
Class 2

Class 3
Class 3

Figure 10.
Figure 10. Acoustic
Acoustic signal
signal classification
classification of
of SF100,
SF100, SF300,
SF300, and
and SG300.
SG300.

Figure
Figure 10
10 shows
shows that
that the
the amplitude
amplitude range
range ofof the
the signals
signals for
for the
the first
first class falls between
between
32
32 dB and
and 42
42dB,
dB,aarange
rangeoften
oftenassociated
associated in in
thethe literature
literature with
with fiber/matrix
fiber/matrix interface
interface fric-
friction and matrix cracking [21,35,36]. These signals first appear at the beginning
tion and matrix cracking [21,35,36]. These signals first appear at the beginning of the sec- of the
second phase, as shown in Figure 9, and their occurrence significantly
ond phase, as shown in Figure 9, and their occurrence significantly increases increases in the third
phase.
phase. This observation is consistent with the results of Chilali et al. [21], [21], who
who studied
studied aa
flax ®
flax fiber-reinforced
fiber-reinforcedlaminate
laminatewithwithananareal
arealdensity
densityof of330
330g/m²
g/m²impregnated
impregnated with with Elium
Elium®
resin. Their results
resin. Their results indicate
indicate that
thatacoustic
acousticsignals
signalslinked
linkedtotomatrix
matrixcracking,
cracking,considered
considereda
alow-amplitude
low-amplitude damage mechanism, occur within the 32–45 dB range.
damage mechanism, occur within the 32–45 dB range. Therefore, we pro- Therefore, we
propose to associate the first-class signals in SF100, SF300, and SG300 with
pose to associate the first-class signals in SF100, SF300, and SG300 with matrix cracking. matrix cracking.
This
This hypothesis
hypothesis isis further
further supported
supported by by the
the microscopic
microscopic observations
observations in in Figure
Figure 11,
11, which
which
reveal
reveal several matrix cracks in both the upper and lower skins of SF300 and SG300, as
several matrix cracks in both the upper and lower skins of SF300 and SG300, as well
well
as in the cavities of the agglomerated cork core filled with Elium® resin. Additionally, these
matrix cracks indicate that the applied mechanical stress exceeds the matrix’s elastic limit
(Table 1), leading to rapid crack propagation.
 J. Compos. Sci. 2024, 8, x FOR PEER REVIEW 15 of 21
J. Compos. Sci. 2024, 8, x FOR PEER REVIEW 15 of 21

as in the cavities of the agglomerated cork core filled with Elium®® resin. Additionally,
J. Compos. Sci. 2024, 8, 539 as in the cavities of the agglomerated cork core filled with Elium resin. Additionally,
15 of 21
these matrix cracks indicate that the applied mechanical stress exceeds the matrix’s elastic
these matrix cracks indicate that the applied mechanical stress exceeds the matrix’s elastic
limit (Table 1), leading to rapid crack propagation.
limit (Table 1), leading to rapid crack propagation.
Bottom side Bottom side Top side
Bottom side Bottom side Top side
SF300 SF100 SG300
SF300 SF100 SG300

Matrix cracking
Matrix cracking
Matrix cracking
Matrix cracking Matrix cracking
Matrix cracking

SF300 SG300
SF300 SG300 Cork cracking
Cork cracking

Cork cracking
Cork cracking

Zoom in
Zoom in

Figure 11. SEM observations of matrix and core cracking in SF100, SF300, and SG300.
Figure11.
Figure 11. SEM
SEM observations
observations of matrix and core
core cracking
cracking in
in SF100,
SF100,SF300,
SF300,and
andSG300.
SG300.

Theincrease
The increase in the second-class signals is primarily observed during the second
The increaseininthethesecond-class
second-class signals is primarily
signals is primarilyobserved
observedduring the second
during phase,
the second
phase,
accounting accounting for approximately
for approximately 30% of the30% of the
recorded recorded signals
signals insignals
this zonein this zone
for SF300 for
andSF300
SF100
phase, accounting for approximately 30% of the recorded in this zone for SF300
and15%
and SF100 forand 15%Infor
SG300. SG300. In
addition, theaddition,
significanttheincrease
significant
in theincrease
signalsin of the
thissignals of this
class coincides
and SF100 and 15% for SG300. In addition, the significant increase in the signals of this
classacoincides
with load range with a loadfrom
varying range varying
310 from
to 650 N, 310 to 650
indicating N,these
that indicating
signalsthat these
result signals
from other
class coincides with a load range varying from 310 to 650 N, indicating that these signals
result
damage from other
mechanisms damage
beyond mechanisms
those beyond
associated those
with associated
the first with
class. the first class.
result from other damage mechanisms beyond those associated with the first class.
In the
In the case of SF300,
SF300, their
theirsignificant
significantoccurrence
occurrencecorresponds to to
a load of of
310 N for 𝐿
In the case of SF300, their significant occurrence corresponds
corresponds a load
to a load of 310310 N for
N for 𝐿
L== =250
250mm.
250 mm.Furthermore,
mm. Furthermore,
Furthermore,
thethe
amplitude
amplitude
the amplitude
range
range
of these
range signals
of these
of these signals
corresponds
signals corresponds
corresponds
to values
to values
typi-
to values
typi-
cally associated
typically withwith
associated fiber/matrix interface
fiber/matrix debonding
interface [21]. This
debonding [21].hypothesis
This hypothesisis supported
is sup-
cally associated with fiber/matrix interface debonding [21]. This hypothesis is supported
by
ported the microscopic observations in Figure 12, which also reveal several
severalexamples
examplesof
by theby the microscopic
microscopic observations
observations in Figure
in Figure 12, which
12, which alsoalso reveal
reveal several examples ofof
debonding between
debonding between the core
core and
and the
the skins,
skins, particularly
particularly nearnear the
the central
centralsupport.
debonding between the the particularly near the central support.
support.
SF300 SG300 Skin/core Skin/core SF100
Fiber/matrix
Fiber/matrix SF300 SG300 Skin/core Skin/core SF100
debonding debonding debonding
debonding debonding debonding

Figure 12. SEM observations of damage to the skin/core interfaces.

Figure 12. SEM observations of damage to the skin/core interfaces.
Figure 12. SEM observations of damage to the skin/core interfaces.
The signals from the third class exhibit higher amplitudes (Figure 10) compared to
The signals
signals from
from the
the third
third class
class exhibit higher amplitudes (Figure 10)
10)compared to
thoseThe
from the first two classes, with aexhibit higher
significant amplitudes
increase (Figure
occurring primarilycompared
during theto
those from the first two classes, with a significant increase occurring primarily during the
those from the first two classes, with a significant increase occurring primarily during the
final phase. At this stage, the damage mechanisms associated with the first and second
classes already account for more than 60% of the overall acoustic activity. This suggests that
fiber damage is less frequent than matrix or interface damage, likely due to the superior
mechanical properties of the fibers. Therefore, we propose to attribute the signals of this
 The mechanical properties of fibers play a crucial role in the acoustic signature of
damage in composite materials. For instance, due to the high elastic properties of glass
fibers, the acoustic signals associated with their breakage exhibit higher amplitudes com-
pared to those from flax fiber breakage. Another notable observation concerns the number
J. Compos. Sci. 2024, 8, 539 of hits, with the SG300 graphs showing 40% more hits at the 250 mm span and 30% at 16the
of 21
100 mm span than those of SF100 and SF300. This difference may be attributed to the
structure of the flax fibers. Indeed, the cellulose microfibrils in flax fibers are aligned along
the longitudinal
third class to fiberaxis, offeringand
debonding strong
the tensile
variousresistance, whileto
modes related components
fiber breakagesuch(Figure
as lignin13).
and hemicellulose contribute to their flexibility, which is important for
Such damage typically presents as high amplitudes due to the greater elastic properties applications re-of
quiring
the fibers.deformation before failure
This interpretation seems[5]. Consequently,
in accordance flax
with thefibers help
results mitigate
reported bydamage
Saidane
within the sandwich structure, particularly matrix cracking. Indeed, the
et al. [35], who analyzed signals from materials reinforced with glass and flax fibers number of (with
hits
associated
areal withofmatrix
densities 300 and failure for SG300
350 g/cm², is twice those
respectively) of SF100 and
impregnated withSF300.
epoxy resin.

SG300 SF100 SF300

Fiber breakage
Fiber breakage Fiber breakage

SG300 SF100 Fiber pullout SF300

Fiber pullout
Fiber pullout

Fiber/matrix
debonding
Fiber
breakage

Figure13.
Figure 13. SEM
SEM observations
observations of
of fiber
fiber damage.
damage.

Fiber
The breakage in
mechanical the lowerofskins
properties fibersthat area subjected
play crucial roletoin
tensile stress issignature
the acoustic observedof indam-
the
central
age area of thematerials.
in composite three sandwich materials
For instance, duedue to the
to the highconcentration of stress
elastic properties in this
of glass re-
fibers,
gion. This type of damage can be attributed to the limited tensile strength
the acoustic signals associated with their breakage exhibit higher amplitudes compared to of the skins and
the concentration
those from flax fiberof shear stressAnother
breakage. at the fiber/matrix interfaces, concerns
notable observation a phenomenon that is of
the number espe-
hits,
cially pronounced with smaller support spans [37,38].
with the SG300 graphs showing 40% more hits at the 250 mm span and 30% at the 100 mm
Additionally,
span than poor adhesion
those of SF100 and SF300.inThis
critical areas accelerates
difference fiber degradation
may be attributed underofthe
to the structure the
combined
flax effect of the
fibers. Indeed, stress. The SEM
cellulose observations
microfibrils from
in flax Figure
fibers 14 reveal
are aligned damage
along within the
the longitudinal
fiberoffering
axis, unit structure, showing
strong tensile cellulose while
resistance, microfibrils being pulled
components such asout or broken.
lignin This inter-
and hemicellulose
contribute to their flexibility, which is important for applications requiring deformation
before failure [5]. Consequently, flax fibers help mitigate damage within the sandwich
structure, particularly matrix cracking. Indeed, the number of hits associated with matrix
failure for SG300 is twice those of SF100 and SF300.
Fiber breakage in the lower skins that are subjected to tensile stress is observed in the
central area of the three sandwich materials due to the concentration of stress in this region.
This type of damage can be attributed to the limited tensile strength of the skins and the
concentration of shear stress at the fiber/matrix interfaces, a phenomenon that is especially
pronounced with smaller support spans [37,38].
Additionally, poor adhesion in critical areas accelerates fiber degradation under the
combined effect of stress. The SEM observations from Figure 14 reveal damage within
the fiber unit structure, showing cellulose microfibrils being pulled out or broken. This
inter-structural deterioration is accentuated by shear stress, which increases the adhesion
defects at the fiber/matrix interface and creates weak points in the material [29].
 J. Compos. Sci. 2024, 8, x FOR PEER REVIEW 17 of 21

J. Compos. Sci. 2024, 8, 539 structural deterioration is accentuated by shear stress, which increases the adhesion17de-
of 21
fects at the fiber/matrix interface and creates weak points in the material [29].

SF300 SF300 SF300

Fiber pullout Microfibril

Microfibrils
Microfibril
Fiber pullout

Figure 14. SEM observations of damage inside fibers.

Figure 14. SEM observations of damage inside fibers.

Another noteworthy
Another noteworthy observation
observation isis related
relatedtotoclassifying
classifyingsignals
signalsassociated
associatedwithwithde-
de-
lamination and fiber/matrix interface failure, which is more challenging
lamination and fiber/matrix interface failure, which is more challenging than classifying than classifying
signals for
signals for fiber
fiber breakage,
breakage, fiber
fiber pull-out,
pull-out, and
and matrix
matrix failure.
failure.The
Thelatter
lattertwo
twodamage
damagemodes
modes
require an accurate classification, as they occur at medium and variable amplitudes,de-
require an accurate classification, as they occur at medium and variable amplitudes, de-
pending on
pending on the
the type
type ofof reinforcement.
reinforcement. FromFrom anan instrumentation
instrumentationstandpoint,
standpoint,monitoring
monitoring
the damage
the damage in in sandwich
sandwich structures
structures with
with piezoelectric
piezoelectricsensors
sensorsisismore
morecomplex
complexthan thaninin
laminates due to the significant thickness of the sandwich and the
laminates due to the significant thickness of the sandwich and the different damage modes different damage
modes
for eachfor each component
component (skins and(skins andAssarar
core). core). Assarar et al.
et al. [39] [39] the
varied varied the sensor
sensor placement
placement on their
on their sandwich structures to investigate differences in signal detection,
sandwich structures to investigate differences in signal detection, aiming to better aiming to better
under-
understand
stand the damage
the damage detectable
detectable withconfiguration.
with each each configuration.
Their Their
resultsresults
indicateindicate
that a that a
sensor
sensor placed on the upper skin can detect matrix cracking (40–65 dB), fiber
placed on the upper skin can detect matrix cracking (40–65 dB), fiber pull-out (40–75 dB), pull-out (40–
75 dB), delamination
delamination (50–85 dB),(50–85 dB), interface
interface failure
failure (50–90 (50–90
dB), dB), breakage
and fiber and fiber (80–100
breakage (80–100
dB), which
dB), which closely aligns with the damage types observed
closely aligns with the damage types observed in our sandwich materials. in our sandwich materials.

3.4. Damage
3.4. Damage Analysis Using Micro-Tomographic Observations Observations
In this
In this section, micro-tomographic
micro-tomographic observations
observationsof ofthe
theSF100,
SF100,SF300,
SF300,and andSG300
SG300sam- sam-
plesare
ples are analyzed
analyzed to to identify their primary damagedamage modes.
modes.The Theobservations
observationsofofthe theSF100
SF100
and SF300
and SF300 samples
samples (Figure 15) reveal notable
notable differences
differencesin indamage
damagebehavior
behaviorwhen whentested
tested
with aa 100
with 100 mmmm support
support span.
span. These differences
differences highlight
highlightthe thecomplexity
complexityand andvariability
variability
of damage
of damage modes,modes, largely influenced
influenced by by variations
variationsin inthe
thereinforcement
reinforcementdensitydensityand andcore
core
structure. In the case of SF100, failure predominantly occurs in the
structure. In the case of SF100, failure predominantly occurs in the lower skin (under lower skin (under ten-
sile stress),
tensile stress),with
withthethe
damage
damage propagating
propagating into
intothethe
core.
core.ThisThistype
typeofoffailure
failureisistypical
typicalofof
materials with
materials with limited tensile strength and and is
is intensified
intensifiedby bythe theporosity
porosityofofthethecore,
core,which
which
serves as
serves as aa crack
crack initiation
initiation site [40]. These
These porosities
porositiesare areclearly
clearlyvisible
visibleininFigure
Figure15a15aasas
black spots, with
black with diameters
diametersranging
rangingfromfrom0.50.5
to to
2 mm2 mmfor for
the the
larger pores.
larger In contrast,
pores. the
In contrast,
lower
the lowerskinskin
of SF300 is more
of SF300 resistant
is more to tensile
resistant stressstress
to tensile than that
thanof SF100
that (Figure
of SF100 15b). Ad-
(Figure 15b).
ditionally, SF300
Additionally, SF300shows
shows signs ofof
signs core
corecompression
compressionand anddelamination
delaminationbetweenbetweenthe the0°0◦and
and
◦
90° layers,
90 layers, caused by the shear stress. Failure in the 90°◦
transverse shear stress. Failure in the 90 layers is observedinin
transverse layers is observed
both sandwich
both sandwich materials.
materials.
Figure 16 shows some damage modes that appear in SF300 and SG300 after the
bending tests with a 250 mm support span. SF300 displays significant damage in its lower
skin, particularly in the outer 0◦ layer. This failure appears to be preceded by the formation
of transverse cracks in the 0◦ and 90◦ layers, followed by extensive delamination at the
interfaces between these layers. This delamination weakens the outer 0◦ layer, leading to its
shear failure. These observations suggest poor interlaminar adhesion between the layers of
the SF300 skins and a certain susceptibility to transverse shear stress, which characterizes
the composite materials reinforced with natural fibers like flax fiber [41].
 F rupture 90
Porosity 0 /90 interface
F

J. J.Compos.
Compos.Sci.
Sci.2024,
2024,8,8,539
x FOR PEER REVIEW 18 18
of of2121

Compressive 7000 µm 3000 µm Compressive

side Layer side
Layer breaking Delamination at
F Core breaking rupture 90
(90 ) Porosity 0 /90 interface

Skin breaking Tensile Tensile Cork crush F

side side zone

(a) (b)

Figure 15. Micro-tomographic observations of the (a) SF100 and (b) SF300 specimens near the load-
ing point for L = 100 mm.

Layer breaking Figure 16 shows some damage modes that appear in SF300 and SG300 after the bend-
Core breaking
(90 ) ing tests with a 250 mm support span. SF300 displays significant damage in its lower skin,
particularly in theTensile
Skin breaking outer 0° layer. This failure appearsCork
Tensile to be preceded by the formation of
crush
side side zone
transverse cracks in the 0° and 90° layers, followed by extensive delamination at the inter-
faces between these layers. This delamination weakens the outer 0° layer, leading to its
(a) (b)
shear failure. These observations suggest poor interlaminar adhesion between the layers
of the SF300
Figure
Figure 15. skins and a certain
15.Micro-tomographic
Micro-tomographic susceptibility
observations
observations of
ofthe toSF100
the(a)
(a) transverse
SF100and (b)shear
and(b) SF300 stress,
SF300specimenswhich
specimens nearcharacter-
near the
theloading
load-
izes
pointthe composite
for Lfor
ing point = 100 materials
mm.mm.
L = 100 reinforced with natural fibers like flax fiber [41].

Compressive Layer rupture Figure 16 shows some damage modes that appear in SF300 and SG300 after the bend-
Compressive
side 0 ,90 Fing tests with a 250 mm support span. Layer
SF300rupture
displays
Delamination at
side skin,
90 significant damage in its lower
0 /90 and
F be90preceded
particularly in the outer 0° layer. This failure appears to by the
/Corck interfaces formation of
transverse cracks in the 0° and 90° layers, followed by extensive delamination at the inter-
faces between these layers. This delamination weakens the outer 0° layer, leading to its
shear failure. These observations suggest poor interlaminar adhesion between the layers
of the SF300 skins and a certain susceptibility to transverse shear stress, which character-
izes the composite materials reinforced with natural fibers like flax fiber [41].
Layer rupture 90

Compressive Layer rupture Compressive

side 0 ,90 F Transversal Layer rupture 90
Delamination at
side
0 /90 and
craking of 90 F 90 /Corck interfaces
and 0 layers
Delamination at Layer shear
Tensile rupture 0 Delamination at Tensile
0 /90 interface
side 3000 µm 3000 µm 0 /90 interface side

(a) (b)
Layer rupture 90
16. Micro-tomographic
Figure 16. Micro-tomographicobservations
observationsofofthe (a)(a)
the SF300 andand
SF300 (b) (b)
SG300 specimens
SG300 for Lfor
specimens = 250
L =mm.
250
mm.
SG300 also shows significant delamination in the lower skin under tension (Figure 16b),
Transversal
though
SG300the also
shearcraking
failures
shows of 90are less pronounced
significant delaminationcompared
in the lowerwith SF300.
skin underThe failure(Figure
tension in the
SG300though
lower skin and 0 layers
appears to initiate withpronounced
fiber breakage ◦ layers (highlighted area)
in the 90with
16b), the shear failures are less compared SF300. The failure in
Delamination at Layer shear ◦ /90Tensile
◦ interfaces.
Tensile parallel to the point of load application, followed by delamination
Delamination at at the 0
0 /90 interfacethe rupture
SG3000lower skin appears to initiate with fiber breakage in the 90° layers (highlighted
side Unlike in SF300, 3000 µm
thepoint
damage 3000 µm
in SG300 is limited 0to
/90delamination
interface and does side
not progress
area) parallel to the of load application, followed by delamination at the 0°/90° in-
to shear failure in the skin. This difference can
(a)terfaces. Unlike in SF300, the damage in SG300 is limited(b)be attributed to the superior mechanical
to delamination and does not
properties of glass fibers, which offer better resistance to tensile and interlaminar stresses
compared to flax fibers [42]. observations of the (a) SF300 and (b) SG300 specimens for L = 250
Figure 16. Micro-tomographic
mm.
4. Conclusions
SG300
In also shows
this study, significant
laminated delamination
and sandwich in thematerials
composite lower skin under
were tensionusing
fabricated (Figure
flax
16b),glass
and though
UDthe shearan
fabrics, failures
Eliumarematrix,
® less pronounced
and a cork compared with the
core through SF300. The failure
one-shot vacuum in
the SG300
resin lower
infusion skin appears
method. to initiate
The feasibility of with fiber breakage
replacing in with
glass fibers the 90° layers
flax fibers(highlighted
in sandwich
area) parallel
materials was to the point ofAload
investigated. application,
comparison followed
of five specificby delamination
properties, at the 0°/90°
the flexural in-
modulus,
terfaces. Unlike in SF300, the damage in SG300 is limited to delamination
the shear modulus, the maximum deflection, the maximum load, and the flexural stiffness, and does not
in three-point bending tests shows that the flax fiber-based sandwich materials SF100 and
J. Compos. Sci. 2024, 8, 539 19 of 21

SF300 could constitute an alternative to glass fiber-based sandwich material SG300 for
applications requiring the maximization of these parameters.
Furthermore, the analysis of the acoustic signals recorded during the bending tests,
combined with the SEM and tomographic observations, enabled the identification of the
main damage modes in the studied sandwich panels, SF100, SF300, and SG300. Matrix
cracking emerged as the most prevalent damage mechanism, with an activity rate of
approximately 60% to 90%, occurring more frequently than fiber pull-out (10% to 30%) or
fiber breakage (less than 10%). The matrix cracking signals appear at low amplitude at the
start of the bending tests, while fiber pull-out and breakage occur later, depending on the
span of the support. We also found that the mechanical properties of the flax and glass
fibers play a crucial role in the acoustic signature of the identified damage. For instance,
due to the high elastic properties of glass fibers, the acoustic signals associated with their
breakage exhibit higher amplitudes compared to those from flax fiber breakage.

Author Contributions: Conceptualization, A.A.T. and M.A.; methodology, M.A. and W.Z.; software,
A.A.T.; validation, M.A., W.Z., R.A., B.M. and K.B.; formal analysis, A.A.T., M.A., W.Z. and B.M.;
writing—original draft preparation, A.A.T.; writing—review and editing, M.A., W.Z., R.A. and
B.M.; supervision, W.Z. and M.A.; All authors have read and agreed to the published version of
the manuscript.
Funding: This research received no external funding.
Data Availability Statement: The data that support the findings of this study are available from the
corresponding author, A.A.T., upon reasonable request.
Acknowledgments: The authors would like to extend their gratitude to Arkema France for providing
the resin used in this study. Special thanks also go to Pierre Gérard for his invaluable support and for
sharing his expertise during the fabrication phase of the composite materials.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Monti, A.; EL Mahi, A.; Jendli, Z.; Guillaumat, L. Quasi-static and fatigue properties of a balsa cored sandwich structure with
thermoplastic skins reinforced by flax fibres. J. Sandw. Struct. Mater. 2019, 21, 2358–2381. [CrossRef]
2. Dogan, A. Low-velocity impact, bending, and compression response of carbon fiber/epoxy-based sandwich composites with
different types of core materials. J. Sandw. Struct. Mater. 2021, 23, 1956–1971. [CrossRef]
3. Selver, E.; Kaya, G. Flexural properties of sandwich composite laminates reinforced with glass and carbon Z-pins. J. Compos.
Mater. 2019, 53, 1347–1359. [CrossRef]
4. Astasari, A.; Sutikno, S.; Wijanarko, W. Bending and Torsional Characteristics of Carbon Fiber and Balsa Wood Sandwich
Composite. IPTEK J. Proc. Ser. 2017, 3, 5–9. [CrossRef]
5. Baley, C. Analysis of the flax fibres tensile behaviour and analysis of the tensile stiffness increase. Compos. Part Appl. Sci. Manuf.
2002, 33, 939–948. [CrossRef]
6. Pickering, K.L.; Efendy, M.G.A.; Le, T.M. A review of recent developments in natural fibre composites and their mechanical
performance. Compos. Part Appl. Sci. Manuf. 2016, 83, 98–112. [CrossRef]
7. Dittenber, D.B.; GangaRao, H.V.S. Critical review of recent publications on use of natural composites in infrastructure. Compos.
Part Appl. Sci. Manuf. 2012, 43, 1419–1429. [CrossRef]
8. Mak, K.; Fam, A.; MacDougall, C. Flexural Behavior of Sandwich Panels with Bio-FRP Skins Made of Flax Fibers and Epoxidized
Pine-Oil Resin. J. Compos. Constr. 2015, 19, 04015005. [CrossRef]
9. Gil, L. Cork Composites: A Review. Materials 2009, 2, 776–789. [CrossRef]
10. Reis, L.; Silva, A. Mechanical Behavior of Sandwich Structures using Natural Cork Agglomerates as Core Materials. J. Sandw.
Struct. Mater. 2009, 11, 487–500. [CrossRef]
11. Urbaniak, M.; Goluch-Goreczna, R.; Bledzki, A.K. Natural Cork Agglomerate as an Ecological Alternative in Constructional
Sandwich Composites. BioResources 2017, 12, 5512–5524. [CrossRef]
12. Sergi, C.; Sarasini, F.; Russo, P.; Vitiello, L.; Barbero, E.; Sanchez-Saez, S.; Tirillò, J. Effect of temperature on the low-velocity impact
response of environmentally friendly cork sandwich structures. J. Sandw. Struct. Mater. 2022, 24, 1099–1121. [CrossRef]
J. Compos. Sci. 2024, 8, 539 20 of 21

13. Qin, Y.; Summerscales, J.; Graham-Jones, J.; Meng, M.; Pemberton, R. Monomer Selection for In Situ Polymerization Infusion
Manufacture of Natural-Fiber Reinforced Thermoplastic-Matrix Marine Composites. Polymers 2020, 12, 2928. [CrossRef]
14. Alshahrani, H.; Ahmed, A. Study on Flexural Behavior of Glass Fiber Reinforced Plastic Sandwich Composites Using Liquid
Thermoplastic Resin. Polymers 2022, 14, 4045. [CrossRef]
15. Gobikannan, T.; Portela, A.; Haldar, A.K.; Nash, N.H.; Bachour, C.; Manolakis, I.; Comer, A.J. Flexural properties and failure
mechanisms of infusible thermoplastic- and thermosetting based composite materials for marine applications. Compos. Struct.
2021, 273, 114276. [CrossRef]
16. Jiang, Q.; Chen, G.; Kumar, A.; Mills, A.; Jani, K.; Rajamohan, V.; Venugopal, B.; Rahatekar, S. Sustainable Sandwich Composites
Manufactured from Recycled Carbon Fibers, Flax Fibers/PP Skins, and Recycled PET Core. J. Compos. Sci. 2021, 5, 2. [CrossRef]
17. Castegnaro, S.; Gomiero, C.; Battisti, C.; Poli, M.; Basile, M.; Barucco, P.; Pizzarello, U.; Quaresimin, M.; Lazzaretto, A. A
bio-composite racing sailboat: Materials selection, design, manufacturing and sailing. Ocean Eng. 2017, 133, 142–150. [CrossRef]
18. Robin, A.; Davies, P.; Arhant, M.; Le Jeune, S.; Lacotte, N.; Morineau, E.; Ioos, F.; Dourlen, P.; Cairo, R. Mechanical performance of
sandwich materials with reduced environmental impact for marine structures. J. Sandw. Struct. Mater. 2024, 26, 99–113. [CrossRef]
19. Sarasini, F.; Tirillò, J.; Lampani, L.; Barbero, E.; Sanchez-Saez, S.; Valente, T.; Gaudenzi, P.; Scarponi, C. Impact behavior of
sandwich structures made of flax/epoxy face sheets and agglomerated cork. J. Nat. Fibers 2020, 17, 168–188. [CrossRef]
20. Mancuso, A.; Pitarresi, G.; Tumino, D. Mechanical Behaviour of a Green Sandwich Made of Flax Reinforced Polymer Facings and
Cork Core. Procedia Eng. 2015, 109, 144–153. [CrossRef]
21. Chilali, A.; Zouari, W.; Assarar, M.; Kebir, H.; Ayad, R. Analysis of the mechanical behaviour of flax and glass fabrics-reinforced
thermoplastic and thermoset resins. J. Reinf. Plast. Compos. 2016, 35, 1217–1232. [CrossRef]
22. Manaii, R.; Terekhina, S.; Chandreshah, L.; Shafiq, N.; Duriatti, D.; Guillaumat, L. Influence of fibre orientation, moisture
absorption and repairing performance on the low-velocity impact response of woven flax/Elium® thermoplastic bio-composites.
J. Compos. Mater. 2023, 57, 3601–3618. [CrossRef]
23. Essassi, K.; Rebiere, J.-L.; Mahi, A.E.; Amine Ben souf, M.; Bouguecha, A.; Haddar, M. Health monitoring of sandwich composites
with auxetic core subjected to indentation tests using acoustic emission. Struct. Health Monit. 2022, 21, 2264–2275. [CrossRef]
24. Afendi, M.; Banks, W.M.; Kirkwood, D. Bubble free resin for infusion process. Compos. Part Appl. Sci. Manuf. 2005, 36, 739–746.
[CrossRef]
25. Srinivasagupta, D.; Joseph, B.; Majumdar, P.; Mahfuz, H. Effect of processing conditions and material properties on the debond
fracture toughness of foam-core sandwich composites: Process model development. Compos. Part Appl. Sci. Manuf. 2003, 34,
1085–1095. [CrossRef]
26. Tao, J.; Li, F.; Zhao, Q.; Zhang, D.; Zhao, Z. In-plane compression properties of a novel foam core sandwich structure reinforced
by stiffeners. J. Reinf. Plast. Compos. 2018, 37, 134–144. [CrossRef]
27. Dai, J.; Pellaton, D.; Hahn, H.T. A comparative study of vacuum-assisted resin transfer molding (VARTM) for sandwich panels.
Polym. Compos. 2003, 24, 672–685. [CrossRef]
28. Assarar, M.; Scida, D.; El Mahi, A.; Poilâne, C.; Ayad, R. Influence of water ageing on mechanical properties and damage events
of two reinforced composite materials: Flax–fibres and glass–fibres. Mater. Des. 2011, 32, 788–795. [CrossRef]
29. Bensadoun, F.; Verpoest, I.; Baets, J.; Müssig, J.; Graupner, N.; Davies, P.; Gomina, M.; Kervoelen, A.; Baley, C. Impregnated fibre
bundle test for natural fibres used in composites. J. Reinf. Plast. Compos. 2017, 36, 942–957. [CrossRef]
30. Poilâne, C.; Cherif, Z.E.; Richard, F.; Vivet, A.; Ben Doudou, B.; Chen, J. Polymer reinforced by flax fibres as a viscoelastoplastic
material. Compos. Struct. 2014, 112, 100–112. [CrossRef]
31. Charlet, K.; Eve, S.; Jernot, J.P.; Gomina, M.; Breard, J. Tensile deformation of a flax fiber. Procedia Eng. 2009, 1, 233–236. [CrossRef]
32. Ben Ammar, I.; Karra, C.; El Mahi, A.; El Guerjouma, R.; Haddar, M. Mechanical behavior and acoustic emission technique for
detecting damage in sandwich structures. Appl. Acoust. 2014, 86, 106–117. [CrossRef]
33. Berthelot, J.M. Composite Materials: Mechanical Behavior and Structural Analysis; Springer: New York, NY, USA, 1999.
34. Kattis, S. Noesis: Advanced data analysis, pattern recognition & neural networks software for acoustic emission applications. In
Deutsche Gesellschaft für Zerstörungsfreie Prüfung; Mistras Group Hellas A.B.E.E.: Athens, Greece, 2017.
35. Saidane, E.H.; Scida, D.; Assarar, M.; Ayad, R. Damage mechanisms assessment of hybrid flax-glass fibre composites using
acoustic emission. Compos. Struct. 2017, 174, 1–11. [CrossRef]
36. Allagui, S.; El Mahi, A.; Rebiere, J.-L.; Bouguecha, A.; Haddar, M. In-situ health monitoring of thermoplastic bio-composites
using acoustic emission. J. Thermoplast. Compos. Mater. 2023, 36, 4296–4316. [CrossRef]
37. Carlsson, L.A.; Kardomateas, G.A. Structural and Failure Mechanics of Sandwich Composites; Springer Science & Business Media:
London, UK; New York, NY, USA, 2011; ISBN 978-1-4020-3225-7.
38. AFNOR. NF T54-606: Plastics-Based Sandwich Structures—Bend Test; Association Française de Normalisation: La Plaine Saint-Denis,
France, 1989.
39. Assarar, M.; Bentahar, M.; El Mahi, A.; El Guerjouma, R. Monitoring of damage mechanisms in sandwich composite materials
using acoustic emission. Int. J. Damage Mech. 2015, 24, 787–804. [CrossRef]
40. Kanny, K.; Mahfuz, H. Flexural fatigue characteristics of sandwich structures at different loading frequencies. Compos. Struct.
2005, 67, 403–410. [CrossRef]
J. Compos. Sci. 2024, 8, 539 21 of 21

41. Alkbir, M.F.M.; Sapuan, S.M.; Nuraini, A.A.; Ishak, M.R. Fibre properties and crashworthiness parameters of natural fibre-
reinforced composite structure: A literature review. Compos. Struct. 2016, 148, 59–73. [CrossRef]
42. Gibson, R.F. Principles of Composite Material Mechanics, 2nd ed.; CRC Press: Boca Raton, FL, USA, 2007; ISBN 978-0-429-19053-7.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.