polymers
Article
Impact of PP Impurities on ABS Tensile Properties:
Computational Mechanical Modelling Aspects
Charles Signoret 1 , Anne-Sophie Caro-Bretelle 2, *, José-Marie Lopez-Cuesta 1 , Patrick Ienny 2 and Didier Perrin 1






Citation: Signoret, C.; Caro-
Bretelle, A.-S.; Lopez-Cuesta, J.-M.;
Ienny, P.; Perrin, D. Impact of PP
Impurities on ABS Tensile Properties:
Computational Mechanical
Modelling Aspects. Polymers 2021, 13,
1647. https://doi.org/10.3390/
polym13101647
Academic Editors:
David Garcia-Sanoguera and
Giulia Fredi
Received: 18 April 2021
Accepted: 12 May 2021
Published: 19 May 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Polymers 2021, 13, 1647. https://doi.org/10.3390/polym13101647
1 Polymers Composites and Hybrids (PCH), IMT Mines Ales, 30100 Ales, France;
charles.van.signoret@orange.fr (C.S.); José-Marie.Lopez-Cuesta@mines-ales.fr (J.-M.L.-C.);
Didier.Perrin@mines-ales.fr (D.P.)
2 LMGC, IMT Mines Ales, Université Montpellier, CNRS, 30100 Ales, France; Patrick.Ienny@mines-ales.fr
* Correspondence: Anne-Sophie.Caro@mines-ales.fr; Tel.: +33-(0)-6-23-39-62-92
Abstract: Recycling of plastics is hindered by their important variety and strong incompatibility.
However, sorting technologies bear costs and meet limits. Very high purities (<2 wt%) are difficult
to reach. Yet, such rates may be detrimental to functional properties. In this work, an ABS matrix
(major plastic in Waste of Electrical and Electronic Equipments) was filled with 4 wt% of PP to mimic
impurities in ABS after recycling. PP-g-MA was introduced in the blend to improve the compatibility.
A finite element model was developed from the mechanical behavior of each component. ABS and PP
were individually characterized from tensile tests instrumented with photomechanics and their be-
haviors were modelled through a set of numerical parameters (elasto-visco-plasticity with a Gurson’s
criterion behavior). Comparison between the determinist model results and the experimental data
(strength, volumetric variation) shows that this type of modelling could be a predictive tool in order
to anticipate composite mechanical properties and to understand micromechanisms of deformation
(damage, cavitation). The main result is that PP introduced at 4 wt% into ABS does not alter the static
mechanical properties despite polymers incompatibility. The addition of PP-g-MA modifies the local
properties and possibly conduct to a premature breakage of the polymer blend.
Keywords: polymer recycling; WEEE; predictive modeling; volume change
1. Introduction
Waste of Electrical and Electronic Equipment (WEEE) Plastics (WEEP) are not currently
intensively regenerated through mechanical recycling for several different reasons [1,2].
Dismantling these devices is a complex subject [3,4]. Moreover, the rapid increase of
tonnages [5], the potential toxicity of these products [6], and the presence of valuable
materials as precious metals [7] strongly motivate the establishment of proper end-of-life
management scenarios. Strong occurrence of toxic and/or environmentally harmful and
now forbidden (or soon to be) brominated flame retardants within polymeric parts [8]
is also a serious barrier to waste management as it is unconceivable to reemploy these
contaminated materials [9]. High purity rates of sorted polymers are probably unattainable,
especially as purity and yield are often contradictory notions [10]. However, even a few
percent of impurities can be detrimental to mechanical properties [11] as most polymers
are incompatible. In the area of 3D printing, recycling plastics has become increasingly
important as this technology can produce large quantities of polymer wastes (support
structures, parts with defects) [12–14].
As one of the most present polymers within WEEE [15,16] and owing to its potential
added value because of its properties, ABS (acrylonitrile-butadiene-styrene) is here consid-
ered as the material to be valorized. As the most present plastic after styrenics (ABS, HIPS
and their blends, ABS/PC, HIPS/PPE . . . ) and bringer of strong incompatibility because
of its different nature, PP (polypropylene) was chosen for the role of contaminant. Some
studies were reported about ABS/PP blends, focusing on PP-rich systems or exploring
https://www.mdpi.com/journal/polymers
 Some studies were reported about ABS/PP blends, focusing on PP-rich systems o
ing the whole range by 10 or 20 wt% steps [17–25]. Apart from strain at break,
ABS properties are seriously impaired with PP incorporation. The major point is t
interaction between ABS and PP due to their chemical differences leading to p
Polymers 2021, 13, 1647 chanical properties. Several compatibilizers were tried but they improved 2 of 19 pr
mainly in PP-rich zones, performances still far below pure ABS [19–21,24].
The aim of this study is to improve the comprehension of the micromechani
the whole range by 10 or 20 wt% steps [17–25]. Apart from strain at break, most of ABS
formation involved during a tensile test with PP added to ABS at a low weight
properties are seriously impaired with PP incorporation. The major point is the weak inter-
with between
action or without
ABS PP-g-MA.
and PP due to Atheir
predictive
chemicalmodel associated
differences leading towith
poortensile test and op
mechanical
strumentation is introduced. From the identification of the mechanical
properties. Several compatibilizers were tried but they improved properties mainly in behavior
phase,zones,
PP-rich the Gurson-Tvergaard-Needleman
performances still far below pure ABS (GTN) model [26–28] has been used to
[19–21,24].
the elasto-visco-plastic and damageable behavior ofofABS
The aim of this study is to improve the comprehension the micromechanics
and PP matrices. of Know
deformation involved during a tensile test with PP added to ABS at a low weight fraction
microstructure, a predictive modeling based on finite element analysis (FEA) w
with or without PP-g-MA. A predictive model associated with tensile test and optical
ducted to understand
instrumentation theFrom
is introduced. localthe
phenomena
identificationleading to blends
of the mechanical breakage.
behavior of each
phase, The paper is structured as follows:
the Gurson-Tvergaard-Needleman (GTN)Section 2 presents
model [26–28] materials
has been and experime
used to describe
the elasto-visco-plastic
tion and damageable
3 highlights modelling; Section behavior of ABSabout
4 discusses and PP matrices.
results; andKnowing
Section 5 bring
the microstructure, a predictive modeling based on finite element analysis (FEA) was
clusion.
conducted to understand the local phenomena leading to blends breakage.
The paper is structured as follows: Section 2 presents materials and experiments;
2. Materials
Section and modelling;
3 highlights Methods Section 4 discusses about results; and Section 5 brings
a2.1.
conclusion.
Materials
Virgin
2. Materials unformulated
and Methods polymer references based on Terluran GP22 (Styroluti
2.1. Materials
and PPH 9020 (PP Homopolymer-Total) were used for this study. With the aim to
ing Virgin unformulated
interfacial polymer
interactions, references based
Bondyram 1001 on Terluran provided
PP-g-MA GP22 (Styrolution) ABS Plast
by Polyram
and PPH 9020 (PP Homopolymer-Total) were used for this study. With the aim to improving
tries (Ram-on, Israël) was used.
interfacial interactions, Bondyram 1001 PP-g-MA provided by Polyram Plastic Industries
(Ram-on,ABS is biphasic
Israël) was used.by nature. SAN, a statistical copolymer of styrene and acry
often at 25 wt%
ABS is biphasic by [29]), is its
nature. SAN,majority phase.
a statistical However,
copolymer thisand
of styrene amorphous
acrylonitrilepolyme
brittle
often and
at 25 wt%thus unsuitable
[29]), for many
is its majority phase. applications. Thus, polybutadiene
However, this amorphous (PB) rubb
polymer is very
brittle and thus unsuitable for many applications. Thus, polybutadiene
sent as a nodular minority phase, as shown by AFM cartography on Figure (PB) rubber is 1, to
present as a nodular minority phase, as shown by AFM cartography on Figure 1, to
its impact properties. For common applications, PB loading rates are reported bet
improve its impact properties. For common applications, PB loading rates are reported
and 30%
between 10 within
and 30%ABS within depending on ifon
ABS depending flow properties
if flow propertiesororimpact resistance
impact resistance is is desi
For emulsion
desired [29]. For grafting, the main
emulsion grafting, thesynthesis
main synthesis route for for
route commercial
commercialABS, size distrib
ABS, size
distribution is reported between 0.1 and 1.0 µm, but mainly with
reported between 0.1 and 1.0 µm, but mainly with a median diameter around 0.3 a median diameter
around 0.3 µm [29] (Figure 1). PP is also biphasic in another way as it is semi-crystalline,
(Figure 1). PP is also biphasic in another way as it is semi-crystalline, meaning t
meaning that crystallites, where chains are spatially organized, are dispersed within an
tallites, where
amorphous phase.chains are spatially organized, are dispersed within an amorphou

Figure
Figure 1. 1. Nodular
Nodular dispersion
dispersion of PBABS—AFM
of PB within within ABS—AFM picture
picture produced produced of
by combination bytopogra-
combination
raphy
phy and and modulus
modulus cartography
cartography owing to
owing to AM-FM AM-FM mode.
mode.
Polymers 2021, 13, 1647 3 of 19

ABS and PP were mixed at 4 wt% since previous works from our laboratory [30]
highlighted that properties loss began at this rate. PP-g-MA was tried at 3 phr (parts per
hundred of resin) to improve interfacial adhesion between phases.
ABS pellets were dried at 80 ◦ C for at least 16 h before processing. ABS and PP were
manually mixed at desired ratio before extrusion within a 900 mm Clextral twin-screw
extruder of type BC21 at 250 rpm, 220 ◦ C along the screw, with a 5 mm nozzle, at a 6 kg/h
feed speed owing to a K-Tron KQx-2 weighing feeder from Coperion. Extrudate was
then pelletized owing to an SGS-E50 from CF Scheer & Cie (Stuttgart, Allemagne). For
mechanical properties, Type A dogbones specimen corresponding to ISO 3167, was injected
using a Krauss-Maffei 180/CX 50 molding injection press at 230 ◦ C.

2.2. Methods
Uniaxial tensile tests were performed on an MTS criterion model 45 universal testing
machine (Eden Prairie, MN, USA) following ISO 527 standard on above-mentioned spec-
imen. Tests were conducted at 1 and 10 mm/min (corresponding respectively to strain
rates 10−4 /s, 10−3 /s) to highlight possible time dependences. Specimens were brought
to 1% strain, unloaded, and then brought to break. The software used is TestXpert® (Ars-
Laquenexy, France) and allows the recording of time, load, and elongation. During these
works, x is defined as the tensile direction, y perpendicular to x and in the planar surface,
and z the out of plane direction.
The axial nominal stress is given by the following expression (Equation (1)):

F
σx = , (1)
S0

where F is the recorded load and S0 the initial sample section surface.
The optical extensometer involves a high-resolution charge-coupled device (CCD)
camera (Redlake Megaplus II (Tuscon, AZ, USA), 1920 × 1080 contiguous and square pixels,
coded in 256 grey levels), set in front of the specimen, which records images during the test.
The optical axis of the camera remains perpendicular to the in-plane surface of the specimen
during the test (Figure 2). The images acquisition is commanded by the LabVIEW® software
(National Instruments, Austin, TX, USA) which allows the simultaneous acquisition of the
images and the data from the testing machine (such as load and crosshead displacement).
According to the test speed used, images are recorded every 0.5 s. The scale factor is fixed
to 42 µm per pixel.
The in-plane Green Lagrange strains Ex , Ey were deduced from digital image cor-
relation (DIC) following a method well described in previous works [31]. A transverse
isotropic assumption was considered and validated as in a previous study on SEBS speci-
men submitted to uniaxial tensile test [32].
The volumetric strain is defined using front-view data only and assuming isotropy:

∆V
= λ2x λy − 1 where λi = 2Ei + 1, i = x, y,
p
(2)
V0

V is the current volume, V 0 the original volume, and λ x and λy are the in-plane
principal stretch ratios in the x and y direction. The part of elastic volumetric strain can be
computed (Equation (3)):
∆V
 
σx
= (1 − 2ν), (3)
V0 elas E
where E and ν are the elastic material parameters (Young modulus and Poisson’s ratio).
Polymers 2021, 13, 1647 4 of 19
Polymers 2021, 13, x FOR PEER REVIEW 4 of 20

Figure 2. Testing apparatus.

Figure 2. Testing apparatus.

The in-plane Green Lagrange strains 𝐸 , 𝐸 were deduced from digital image corre-
3. Results
lation
3.1. (DIC) following
Mechanical a method well described in previous works [31]. A transverse iso-
Characterization
tropicMechanical
3.1.1. assumption was consideredof
Characterization and
ABSvalidated
and PP as in a previous study on SEBS specimen
submitted to uniaxial tensile test [32].
ABS and PP behave mechanically quite differently. The first polymer is brittle in
The volumetric
comparison strain
to the second is defined
one. During using front-view
the first data only and
loading/unloading cycleassuming
until 1% isotropy:
of strain,
∆ ν) of virgin polymers are computed for both imposed speeds
the elastic properties (E and = 𝜆 𝜆 − 1 where 𝜆 = 2𝐸 + 1, 𝑖 = 𝑥, 𝑦, (2)
(see average values with standard deviation in Table 1). Regarding elastic properties, PP
seemsVmore
is theimpacted by the speed
current volume, V0 therate, especially
original volume,with
and 𝜆 and
respect 𝜆 Poisson’s
to the ratio, with
are the in-plane prin-
an increase of 10% while increasing the speed from 1 to 10 mm/min.
cipal stretch ratios in the x and y direction. The part of elastic volumetric strain can be
computed (Equation (3)):
Table 1. Elastic properties of virgin polymers (ABS and PP) deduced from instrumented tensile tests
∆𝑉 𝜎
at 1 and 10 mm/min. = (1 − 2𝜈), (3)
𝑉 𝐸
Loading Speeds (mm/min) 1 1 10 10
where E and ν are the elastic material parameters (Young modulus and Poisson’s
Polymers ABS PP ABS PP
ratio).
E (MPa) 2337 ± 9 1236 ± 25 2349 ± 12 1193 ± 42
3. Results ν 0.36 ± 0.01 0.39 ± 0.01 0.37 ± 0.01 0.43 ± 0.01
3.1. Mechanical Characterization
3.1.1. Mechanical
Moreover, the Characterization
ABS matrix behaves of ABS and PP
as a damageable elastic material. Its global behavior
is time ABS and PP behave mechanically quite differently. The(deduced
independent (Figure 3) and its volumetric strain fromisEquation
first polymer (2))
brittle in com-
follows an elastic law (deduced from Table 1 and Equation (3)) until 1.7% of
parison to the second one. During the first loading/unloading cycle until 1% of strain, the imposed strain
(Figure 4). However,
elastic properties (E PP
andpolymer behavior
ν) of virgin is more
polymers aresensitive
computed to for
theboth
strainimposed
speed with regard
speeds (see
toaverage
the stress values.
values withMoreover, the volumetric
standard deviation strain
in Table deviates from
1). Regarding theproperties,
elastic elastic curve from
PP seems
2.3%
moreofimpacted
strain andbyisthetime independent.
speed For a same
rate, especially with level of strain,
respect to the the stress curve
Poisson’s is very
ratio, with an
different for both polymers whereas the volumetric
increase of 10% while increasing the speed from 1 to 10 mm/min.strain is almost similar.
 Moreover,
Moreover, the ABS matrixthe ABS matrix
behaves behaves aselastic
as a damageable a damageable
material.elastic material.
Its global behav-Its global behav-
ior is time independent (Figure 3) and its volumetric strain (deduced from Equation (2)) Equation (2))
ior is time independent (Figure 3) and its volumetric strain (deduced from
follows
follows an elastic law an elastic law
(deduced from(deduced fromEquation
Table 1 and Table 1 and Equation
(3)) until 1.7%(3)) until 1.7% of imposed
of imposed
strain (Figure 4). However, PP polymer behavior is more sensitive to the strain speedthe
strain (Figure 4). However, PP polymer behavior is more sensitive to strain speed with
with
Polymers 2021, 13, 1647 regard to the stress values. Moreover, the volumetric strain deviates
regard to the stress values. Moreover, the volumetric strain deviates from the elastic curve from the elastic5curve
of 19
from 2.3% of strain and is time independent. For a same level of
from 2.3% of strain and is time independent. For a same level of strain, the stress curve isstrain, the stress curve is
very different for both polymers whereas the volumetric
very different for both polymers whereas the volumetric strain is almost similar. strain is almost similar.

Figure
Figure 3. Stress Figure
vs. 3.3.Stress
Stressvs.
Lagrangian vs.Lagrangian
Lagrangian
strain strain
strain
in uniaxial ininuniaxial
tensile uniaxial
tests tensiletests
tensile
at several testsatatseveral
imposed several imposed
strainimposed
rates strainrates
for strain rates
forfor
ABS
ABS
ABS and PP. and PP. and PP.

Figure 4. Volumetric strain vs. Lagrangian strain in uniaxial tensile tests at several imposed strain rates for ABS and PP,
focus on strain until % to enhance the deviation point at 2.4% from elastic behavior for PP sample.

3.1.2. Mechanical Characterization of Composites

Once PP is added into ABS matrix at 4 wt%, the global shape of the stress/strain curve
is similar to virgin ABS (see for example Figure 5 for the higher speed test). As expected,
the Young’s modulus is lowered by the introduction of a less stiff polymer just like the
associated maximum stress. This phenomenon is enhanced by the introduction of PP-g-MA
(see Table 2). The volumetric strain (Figure 6) is almost unchanged as PP is introduced into
ABS with or without the presence of PP-gMA. The introduction of PP impurities into ABS
does not impact ABS volumetric strain, even if the stress/strain curve is slightly lowered
as PP-g-MA is introduced into ABS/PP.

Polymers 2021, 13, 1647
is similar to virgin ABS (see for example Figure 5 for the higher speed test). As expected,
the Young’s modulus is lowered by the introduction of a less stiff polymer just like the
associated maximum stress. This phenomenon is enhanced by the introduction of PP-g-
MA (see Table 2). The volumetric strain (Figure 6) is almost unchanged as PP is introduced
into ABS with or without the presence of PP-gMA. The introduction of PP impurities 6 ofinto
19
ABS does not impact ABS volumetric strain, even if the stress/strain curve is slightly low-
ered as PP-g-MA is introduced into ABS/PP.

Figure 5. Stress vs. Lagrangian strain in uniaxial tensile tests at 10 mm/min for ABS, PP, and com-
Figure 5. Stress vs. Lagrangian strain in uniaxial tensile tests at 10 mm/min for ABS, PP,
posites.
and composites.
Table 2. Elastic and ultimate properties deduced from experiments for ABS, PP, ABS/PP compo-
Table Elastic
2. 10
sites at and ultimate properties deduced from experiments for ABS, PP, ABS/PP composites
mm/min.
at 10 mm/min.
ABS PP ABS/PP ABS/PP/PP-g-MA
ABS PP ABS/PP ABS/PP/PP-g-MA
E (MPa)
E (MPa) 2349
2349±±1212 1193
1193 ±±4242 2329±±24
2329 24 2212±±4545
2212
ν 0.37 ± 0.01 0.43 ± 0.005 0.37 ± 0.04 0.37 ± 0.02
ν
Polymers 2021, 13, x FOR PEER REVIEW σmax (MPa)
0.3742± ±
0.01
5 0.43 ± 0.005
31.7 ±3 0.37 ± 40.04
41 ± 0.37±± 0.2
39.4 0.02 7 of 20
ε u (−) 0.02 ± 0.01 0.18 ± 0.01 0.02 ± 0.01 0.027 ± 0.01
𝜎 (MPa) 42 ± 5 31.7 ± 3 41 ± 4 39.4 ± 0.2

𝜀 (−) 0.02 ± 0.01 0.18 ± 0.01 0.02 ± 0.01 0.027 ± 0.01

Figure 6. Volumetric Strain vs. Lagrangian strain in uniaxial tensile tests at 10 mm/min for ABS,
Figure 6. Volumetric Strain vs. Lagrangian strain in uniaxial tensile tests at 10 mm/min for ABS, PP,
PP, and composites.
and composites.

3.2.Morphology
3.2. MorphologyCharacterization
Characterization
Morphologieswere
Morphologies were monitored
monitored owing owing to (scanning
to SEM SEM (scanning
electron electron microscopy)
microscopy) (Quanta
(Quanta
FEG 200, FEG 200, Thermo-Fisher
Thermo-Fisher Scientific,Scientific,
Waltham,Waltham,
MA, USA)MA, USA) on cryofractured
on cryofractured dogbones dog-
(or-
thogonally or 45◦ ) andorpost-mortem
bones (orthogonally 45°) and post-mortem Charpy
Charpy impact impact fragments.
fragments. To produceTocryofractured
produce cry-
ofracturedpolymer
dogbones dogbones polymer
samples samples
were were firstinto
first immersed immersed into liquid
liquid nitrogen for nitrogen for ap-
approximately
proximately 5 min and then taken out and struck with a hammer. AFM (Oxford Asylum
Research MFP3D, Santa Barbara, CA, USA) was also performed on samples prepared us-
ing a Leica EM UC7 ultracryomicrotome (Leica, Nanterre, France). AM-FM (amplitude
modulation-frequency modulation) mode of AFM was applied to enable mapping of local
Polymers 2021, 13, 1647 7 of 19

5 min and then taken out and struck with a hammer. AFM (Oxford Asylum Research
MFP3D, Santa Barbara, CA, USA) was also performed on samples prepared using a Leica
EM UC7 ultracryomicrotome (Leica, Nanterre, France). AM-FM (amplitude modulation-
frequency modulation) mode of AFM was applied to enable mapping of local pseudo-
moduli, enabling easier differentiation between SAN and PB as their respective moduli
are different.

3.2.1. Nodular Dispersion of PP within ABS

As expected, polypropylene forms a nodular dispersion within ABS as Figure 7 shows.
Cryofractures prove that PP nodules adopt a spherical shape and not an ellipsoidal or
cylindrical one. Moreover, PP nodule sizes are reduced with PP-g-MA addition. A possible
adhesive interaction is visible in Figure 8 as fibrils are seen between what could be a PP
nodule and the matrix. However, this micrograph corresponds to a very specific area of
Polymers 2021, 13, x FOR PEER REVIEW 8 of 20
the fracture and was not found elsewhere. Part 2.4.2 below presents micrographs more
representative of the post-mortem samples.

Figure
Figure 7.7.Nodular
Nodular morphology
morphology ofof
PPPP within
within ABS
ABS at at 4 wt%
4 wt% with
with and
and without
without PP-g-MA-
PP-g-MA- twotwo mag-
magnifi-
nifications.
cations.

Size distribution of nodules was determined by image analysis (AphelionTM 3.2

(ADCIS)). SEM micrographs were binarized and PP nodules were identified and filtered
by surface area (see Figure 9). SEM pictures on samples cut with an ultracryomicrotome
were preferred as micrographs issued from cryofractures were too uneven to enable image
analysis. AFM pictures were also more complex to analyze as both PP and PB were visible.
Each identified object is assumed to have a disk shape well described by its diameter.
ABS/PP and ABS/PP/PP-g-MA were analyzed owing to this procedure. Around 800 PP
Polymers 2021, 13, 1647 8 of 19

nodules
Figure 7.were identified
Nodular for both
morphology compositions
of PP within ABS atwith nodules
4 wt% diameters
with and from 0.2 to
without PP-g-MA- 1.7mag-
two µm
for ABS/PP blend and from 0.1 to 1.0 µm for ABS/PP/PP-g-MA blends (Figure 10).
nifications.

Polymers 2021, 13, x FOR PEER REVIEW 9 of 20

visible. Each identified object is assumed to have a disk shape well described by its diam-
eter. ABS/PP and ABS/PP/PP-g-MA were analyzed owing to this procedure. Around 800
Figure
PP 8. SEM picture
nodules of cross-sections ofcompositions
ABS-PP-PP-g-MA samples after tensile test—fibrils link-
Figure 8. SEM were identified
picture for both
of cross-sections of ABS-PP-PP-g-MAwith nodules
samples afterdiameters from 0.2
tensile test—fibrils to 1.7
linking
ing a potential PP nodule to the matrix.
µm for ABS/PP blend and from
a potential PP nodule to the matrix. 0.1 to 1.0 µm for ABS/PP/PP-g-MA blends (Figure 10).
Size distribution of nodules was determined by image analysis (AphelionTM 3.2
(ADCIS)). SEM micrographs were binarized and PP nodules were identified and filtered
by surface area (see Figure 9). SEM pictures on samples cut with an ultracryomicrotome
were preferred as micrographs issued from cryofractures were too uneven to enable im-
age analysis. AFM pictures were also more complex to analyze as both PP and PB were

Figure9.9. Binarization
Figure BinarizationofofSEM
SEMmicrographs
micrographsof of
cuts performed
cuts using
performed an ultracyromicrotome—ABS
using an ultracyromicrotome—
+ PP and ABS + PP + PP-g-MA systems.
ABS + PP and ABS + PP + PP-g-MA systems.

3.2.2. Cavitation Phenomenon

Figure 11 compares the formation of porosity or cavities on the broken cross section
within ABS and PP after tensile tests pursued until failure. These pictures show porosi-
ties/cavities areas which seem larger in the case of PP but this polymer came to break at
higher strains in comparison with ABS.

Figure 10. Frequencies (a) and cumulated frequencies (b) of diameters sizes distribution for
Polymers 2021, 13, 1647 9 of 19
Figure 9. Binarization of SEM micrographs of cuts performed using an ultracyromicrotome—ABS
+ PP and ABS + PP + PP-g-MA systems.

Polymers
FigureFigure
2021,
10.13, 10. Frequencies
x FOR PEER
Frequencies (a)REVIEW (a) andfrequencies
and cumulated cumulated frequencies
(b) of (b)distribution
diameters sizes of diameters sizes distribution
for ABS/PP for
and ABS/PP/PP-g-MA 10 of 20
ABS/PP
composites. and ABS/PP/PP-g-MA composites.

3.2.2. Cavitation Phenomenon

Figure 11 compares the formation of porosity or cavities on the broken cross section
within ABS and PP after tensile tests pursued until failure. These pictures show porosi-
ties/cavities areas which seem larger in the case of PP but this polymer came to break at
higher strains in comparison with ABS.

Figure 11. Comparison of cavitation phenomena within ABS and PP—SEM pictures on cryofrac-
Figure 11. Comparison of cavitation phenomena within ABS and PP—SEM pictures on cryofractured
tured dogbones and post-mortem samples from tensile tests at 10 mm/min.
dogbones and post-mortem samples from tensile tests at 10 mm/min.

Figure1212shows
Figure showsthat thatthe
thecavitation
cavitationprocess
processrelated
relatedtotoPBPBparticles
particlesisisvery
verysimilar
similar
within virgin ABS even with different strain speeds. This is in accordance
within virgin ABS even with different strain speeds. This is in accordance with the obser-with the obser-
vationsofofFigure
vations Figure4 4for
forwhich
whichvolumetric
volumetricstrains
strainsofofvirgin
virginpolymers
polymersareare insensitivetotothe
insensitive the
imposed speeds.
imposed speeds.
Figure 13 shows cryofractures and post-mortem samples of SAN, ABS, ABS/PP, and
ABS/PP/PP-g-MA. As a fragile material, SAN does not form cavities and fractures are
thus very similar from cryofracture to post-mortem. As reported in the literature [33], PB
generates cavitation within ABS, enabling an improvement toward fracture as it dissipates
energy. Smallest holes on the picture are about 0.2 µm in diameters and biggest ones at
1.4 whereas most of them are around 0.4–0.5 µm. PP presence worsens dramatically the
phenomenon as most holes seen here are more around 1.0–1.5 µm. With PP-g-MA the
 Figure 11. Comparison of cavitation phenomena within ABS and PP—SEM pictures on cryofrac-
tured dogbones and post-mortem samples from tensile tests at 10 mm/min.
Polymers 2021, 13, 1647 10 of 19

Figure 12 shows that the cavitation process related to PB particles is very similar
within virgin ABS even with different strain speeds. This is in accordance with the obser-
vations
hole sizeofis Figure
around4 0.5–1
for which volumetric
µm, closer strains
to pure ABS. of virgin
This polymers
is surely due toare
theinsensitive to the
morphological
imposed speeds.
refinement described in Figure 10.

Figure 12. ABS cavitation independence toward strain speed—SEM pictures of post-mortem sam-
Figure 12. ABS cavitation independence toward strain speed—SEM pictures of post-mortem samples.
ples.
3.3. Modelling
3.3.1. Figure 13 shows cryofractures and post-mortem samples of SAN, ABS, ABS/PP, and
Behavior
ABS/PP/PP-g-MA. As a fragile material, SAN does not form cavities and fractures are thus
The literature reports an important number of damage models, able to account for
very similardegradation
mechanical from cryofracture to post-mortem.
in materials. As reported
Possibly, these in the literature
models include new internal[33], PB gen-
variables
erates cavitation within ABS, enabling an improvement toward fracture as
and evolution laws coupled to plasticity problems [34]. The macroscopic measured volume it dissipates
energy.isSmallest
change assumedholes
to be on the picture
related are about
to cavitation 0.2 µm
occurring in diameters
within and phases
the polymers biggestby ones
voidat
1.4 whereas most of them are around 0.4–0.5 µm. PP presence worsens dramatically
nucleation and growth [35]. The present work focuses on a Gurson-type approach which is the
phenomenon
based as most
on mechanics ofholes seen
porous here are
media. Themore
damagearound 1.0–1.5related
indicator, µm. With PP-g-MA
to the the hole
degradation of
the material, is the void volume fraction (i.e., the porosity) which progressively downscales
the yield surface. In the early eighties, Tvergaard and Needleman [28,36] extended the Gur-
son approach, namely GTN model, including material hardening, multiple voids, and void
coalescence. This GTN model, available in numerous finite elements software, has been
already used by authors to describe the micromechanisms of deformation of a PP/Bakelite
blend [37] for which the PP matrix exhibits porosity following a Gurson-type model.
In this damage formulation, the yield criterion uses hydrostatic pressure and porosity:
the criterion of plasticity and the plastic potential, depend on a macroscopic stress σ∗ and
an effective porosity fraction ft in the following way (Equation (4)):
  
3J2 (eσ) q2 I1 (e
σ) 2 2

+ 2 f t q 1 cosh − 1 + q 1 f t = 0, (4)
σ∗2 2σ∗

where e σ is the stress tensor J2 = 12 dev(e σ), I1 = trace(e

σ ) : dev(e σ ) and q1 , q2 are model
parameters. The q1 parameter allows to adjust the influence of ft on the yield surface and
the q2 parameter is related to the pressure effect.
Effective stress σ∗ is implicitly evaluated from Equation (4).
The plastic flow potential is written as (Equation (5)):
 
f = σ∗ − R( p), R = R0 + Q1 1 − e−b1 p , (5)

where p is the effective plastic strain and R0 , Q1 , b1 are material parameters.

The evolution of viscoplastic strain is written through a Norton’s law:

.
 n
. ∂σ∗ . f
ε = (1 − p) pe
e e∗ =
n∗ , n , p= , (6)
∂e
σ K

where n, K are material parameters.

Polymers 2021, 13, 1647 11 of 19

The total porosity change rate can be split into two components, one attributed to the
. .
void growth f g and the other one to the nucleation f n :
Polymers 2021, 13, x FOR PEER REVIEW 11 of 20
. . .
f t = f g + f n, (7)
.
size is around 0.5–1 µm,. closer
. to pure
. ABS. This is surely due to the morphological re-
where f g = (1 − f t )tre
ε and f n = A p, A is a constant.
finement described in Figure 10.

Figure 13. Cavitation phenomena observed on post-mortem tensile tests fragments through
Figure 13. Cavitation phenomena observed on post-mortem tensile tests fragments through SEM—
SEM—comparison to cryofractured samples—virgin SAN, virgin ABS, and ABS contaminated
comparison to cryofractured
with 4 wt% PP samples—virgin
and ABS contaminated SAN,
with 4 wt% PP virgin
with 3 ABS,
phr ofand ABS contaminated with 4 wt%
PP-g-MA.
PP and ABS contaminated with 4 wt% PP with 3 phr of PP-g-MA.
3.3. Modelling
3.3.2. Determination of ABS and PP Model Properties
3.3.1. Behavior
The behavior of PP and ABS can be well described through this modelling. The
The modulus
Young’s literature E,
reports an important
the Poisson’s rationumber of damage
ν and the models,
yield stress able determined
R0 were to account forfrom
mechanical degradation in materials. Possibly, these models include new internal
stress/strain data. The GTN model has been implemented using the Matlab software. varia-The
bles and evolution laws coupled to plasticity problems [34]. The macroscopic measured
volume change is assumed to be related to cavitation occurring within the polymers
phases by void nucleation and growth [35]. The present work focuses on a Gurson-type
Polymers 2021, 13, 1647 12 of 19

porosity evolution parameters q1 and q2 are arbitrarily fixed to 1, adjustments of these

values would necessitate some local data (as porosity distribution) [35]. The GTN model is
therefore reduced to the native Gurson model. The remaining material parameters were
identified by using a mean square method on stress/strain and volumetric strain/strain
Polymers 2021, 13, x FOR PEER REVIEW 13 of 20
data between experiments and numeric data at each timestep of the simulation. Parameters
Polymers 2021, 13, x FOR PEER REVIEW 13 of 20
present in Table 3 lead to a good accordance between experiments and modelling for both
stress (Figure 14) and volumetric strain (Figure 15). The contrast on the K value prevails
Table 3. Material model parameters for ABS and PP.
for the 3.
Table low time plasticity
Material dependency
model parameters forand
for ABS thePP.
ABS.
Material Properties ABS PP
Table 3. MaterialProperties
Material model parameters for ABS andABS
PP. PP
q1(−), q2() 1, 1 1, 1
Material Properties
q1(−), q2() ABS
1, 1 1,PP1
q1 (−A()
), q2 () 0.53
1, 1 0.23
1, 1
A()
A() 0.53
0.53 0.23
0.23
n(), K (MPa.s) 2.5, 410 2.4, 51.2
n(), K (MPa.s) 2.5, 410 2.4, 51.2
n(), K (MPa.s) 2.5, 410 2.4, 51.2
R MPa), 𝑄
𝑅0 ((MPa), Q1 (MPa),
(MPa), 𝑏b1(−)
(−) 27, 2,
27, 2, 210
210 11,17,
11, 17,58.3
58.3
𝑅 (MPa), 𝑄 (MPa), 𝑏 (−) 27, 2, 210 11, 17, 58.3

Figure 14. Modelling of ABS, stress and volumetric strain vs. Lagrangian strain.
Figure 14. Modelling of ABS, stress and volumetric strain vs. Lagrangian strain.
Figure 14. Modelling of ABS, stress and volumetric strain vs. Lagrangian strain.

Figure 15. Modelling of PP, stress, and volumetric strain vs. Lagrangian strain.
Figure15.
Figure 15.Modelling
Modellingof
ofPP,
PP,stress,
stress,and
andvolumetric
volumetricstrain
strainvs.
vs. Lagrangian
Lagrangian strain.
strain.
3.3.3.. Prediction of ABS/PP Blend Properties
3.3.3.For
. Prediction of simplicity,
the sake of ABS/PP Blend Properties
a plane stress assumption is made for the numerical com-
For the
putations evensake of simplicity,
if this a plane stress
strong hypothesis is far assumption
from reality.isThe
made for the between
interface numerical com-
matrix
putations even if this strong hypothesis is far from reality. The interface between
and fillers is assumed to be perfect (displacement continuity) and phase properties are matrix
and fillersthrough
described is assumed to be
the set perfect (displacement
of parameters of Table 3. continuity) and phase
This assumption properties
of perfect are
interface
 Polymers 2021, 13, 1647 13 of 19

3.3.3. Prediction of ABS/PP Blend Properties

13, x FOR PEER REVIEW For the sake of simplicity, a plane stress assumption is made for the14numerical of 20 com-
putations even if this strong hypothesis is far from reality. The interface between matrix
and fillers is assumed to be perfect (displacement continuity) and phase properties are
described through the set of parameters of Table 3. This assumption of perfect interface
We assume that the properties of ABS and PP in the blend are similar to those of the neat
will be discussed in next section as the compatibility of ABS and PP is quite questionable.
ABS and PP. We assume that the properties of ABS and PP in the blend are similar to those of the neat
From the particle
ABS andsizePP.distribution shown in Figure 10, two kinds of microstructures
were generated via DIGIMAT software
From the particle size(for both ABS/PP
distribution shown and ABS/PP/PP-g-MA
in Figure 10, two kinds blends).
of microstructures
Mesh microstructures based onviatriangular
were generated DIGIMAT elements
software (forand boundary
both ABS/PP and conditions are pre- blends).
ABS/PP/PP-g-MA
sented in Figure 16.
MeshThe RVE (representative
microstructures based on volume
triangularelement)
elements sizes were chosen
and boundary to stabi-
conditions are presented
in Figure 16. The RVE (representative volume element) sizes were
lize the macroscopic stress/strain responses. As local behavior is of interest for this study, chosen to stabilize the
the chosen mesh refinement leads to stabilize local blend response. We assume the struc-study, the
macroscopic stress/strain responses. As local behavior is of interest for this
ture symmetry. Achosen mesh refinement leads to stabilize local blend response. We assume the structure
prescribed displacement (corresponding to 10% of strain) was applied
symmetry. A prescribed displacement (corresponding to 10% of strain) was applied on
on the top of the sample. The non-linear numerical simulation was carried out using the
the top of the sample. The non-linear numerical simulation was carried out using the FE
FE software Zebulon, which
software was developed
Zebulon, at Mines at
which was developed Paristech [39]. This
Mines Paristech [38].RVE
Thismodel
RVE modelis is thus
thus useful to predict the mechanical properties of ABS/PP composites. The size of
useful to predict the mechanical properties of ABS/PP composites. The size of the PP the PP
nodules is much nodules
lower whenis mucha compatibilizer is added: thisisinduces
lower when a compatibilizer added: thisboth a reduction
induces in
both a reduction in
the RVE size andthe interparticular nodules distances
RVE size and interparticular nodulesfordistances
the ABS/PP/PP-g-MA blend (seeblend (see
for the ABS/PP/PP-g-MA
Figure 16). We canFigure 16). from
deduce We can deduce
these fromthat
results these results that
interface areasinterface
between areas
ABSbetween
and PPABSareand PP are
increased
increased since nodules aresince nodules are smaller.
smaller.

FigureFigure
16. Finite element
16. Finite modelling
element of ABS
modelling and
of ABS PPPPblend:
and blend:structure
structure and boundaryconditions.
and boundary conditions.

4. Discussion 4. Discussion
Comparison between Comparison between
numerical numerical
models models and experimental
and experimental data blends
data for ABS/PP for ABS/PP
is blends
is given in Figure 17. The simulated results for both systems
given in Figure 17. The simulated results for both systems (ABS/PP and ABS/PP/PP-g- (ABS/PP and ABS/PP/PP-
g-MA) coincide. As a first result, input materials parameters lead to good predictive
MA) coincide. As a first result, input materials parameters lead to good predictive mod-
modelling as “ABS/PP num” and “ABS/PP exp” are similar until 1.7% of strain. Though
elling as “ABS/PPunicity
num”of and “ABS/PP exp” are similar until 1.7% of strain. Though unicity
the solution to the model is not proven, the good accordance between experimen-
of the solution totalthe model
results andisthose
not predicted
proven, the good
by the accordance
model between
assures the validity experimental
of the set of parameters
results and those used
predicted by the model assures the validity of the set
here. After 1.7% of strain, in the experiments, the blendsofbroke
parameters
quickly. used
Obviously, this
here. After 1.7% of strain,
is not in the
the case in experiments, the blends broke
the numerical simulations becausequickly. Obviously,
no rupture thisbeen
criterion has is added.
not the case in the numerical simulations because no rupture criterion has been added.
 Polymers 2021, 13, 1647 14 of 19
Polymers
Polymers 2021,
2021, 13,13, x FOR
x FOR PEER
PEER REVIEW
REVIEW 1515
of of
2020

Figure 17.
Figure Stress vs. Lagrangian strain for ABS/PP and ABS/PP/PP-g-MA composites, experiments
Figure17.
17.Stress
Stressvs.
vs.Lagrangian
Lagrangianstrain
strainfor
forABS/PP
ABS/PPand
andABS/PP/PP-g-MA
ABS/PP/PP-g-MAcomposites, experiments
composites, experiments
and
and modelling.
andmodelling.
modelling.

Afirst
AA firstapproach
first approachtoto
approach tosimulate
simulatepolymer
simulate polymerblends
polymer blendsbreakdown
blends breakdownisisisdevoted
breakdown devotedtoto
devoted toevaluate
evaluatethe
evaluate the
the
equivalent
equivalent stress
equivalentstress (Von
stress(Von Mises)
(VonMises) in each
Mises)inineach element
eachelement of
elementofofthethe RVE.
theRVE. A stress
RVE.AAstress mapping
stressmapping at
mappingatat1.7% 1.7%
1.7%of of
of
macroscopic
macroscopic strain
macroscopicstrain is presented
strainisispresented in Figure
presentedininFigure
Figure18 18 for
18for both
forboth blends.
bothblends.
blends.

Figure
Figure
Figure 18.
18.
18. Von
Von
Von Mises
Mises
Mises stress
stress
stress cartography
cartography
cartography atatat 1.7%
1.7%
1.7% ofof
of macroscopic
macroscopic
macroscopic strain
strain
strain inin
in ABS/PP
ABS/PP
ABS/PP and
and
and ABS/PP/PP-
ABS/PP/PP-
ABS/PP/PP-
g-MA blends.
g-MAblends.
g-MA blends.

To Toquantify
To quantifythis
quantify thismapping,
this mapping,we
mapping, weweanalyzed
analyzedthe
analyzed theVon
the VonMises
Von Misesstress
Mises stressdistribution
stress distributionwithin
distribution withinthe
within the
the
ABS.As
ABS.
ABS. Aslocal
As localbehavior
local behaviorisisisnot
behavior notelement
not elementsize
element sizedependent,
size dependent,the
dependent, thecomparison
the comparisonbetween
comparison betweencompati-
between compati-
compati-
bilized
bilized and
and uncompatibilized
uncompatibilized ABS/PP
ABS/PP is possible even
even
bilized and uncompatibilized ABS/PP is possible even if RVE and elements sizes are if
if RVE
RVE and
and elements
elements sizes
sizes arenot
are not
not
equivalent.
equivalent. The
The Figure
Figure 19
19 reflects
reflects the
the distribution
distribution of
of Von
Von Mises
Mises
equivalent. The Figure 19 reflects the distribution of Von Mises stresses into ABS polymer stresses
stresses into
intoABS
ABS polymer
polymer
for
for several
forseveral macroscopic
severalmacroscopic strains
macroscopicstrains (0.3,
strains(0.3, 1.5,
(0.3,1.5, 1.9,
1.5,1.9, and
1.9,and
and2.1%)
2.1%)
2.1%) for
for ABS/PP
for ABS/PPand
ABS/PP and ABS/PP/PP-g-MA
andABS/PP/PP-g-MA
ABS/PP/PP-g-MA
composites.
composites. The
The ultimate
ultimate stress
stress ofof ABS
ABS polymer
polymer is is around
around
composites. The ultimate stress of ABS polymer is around 43 MPa, beyond which 43 43
MPa, MPa, beyond
beyond whichwhich
the
the the
sam-
sam-
sample
pleplebreaks.breaks.
breaks. The
The The
main
main main conclusion
conclusion
conclusion is is is
that
that that reducing
reducing
reducing the
the the
size
size size
ofof of
PPPP PP nodules
nodules
nodules byby by compatibi-
compatibiliza-
compatibiliza-
lization leads to a local increase of the level of equivalent stress
tion leads to a local increase of the level of equivalent stress by the diminution ofofthe
tion leads to a local increase of the level of equivalent stress by by
thethe diminution
diminution of the
thein-in-
interparticular
terparticular distancebetween
distance betweennodulesnodules(see (see Figure
Figure 16).16). The frequency curves curves are
are thus
thus
terparticular distance between nodules (see Figure 16). The frequency curves are thus
shifted
shifted to the right.
shifted toto the
the right.
right.
13, x FOR PEER REVIEW 16 of 20
Polymers 2021, 13, 1647 15 of 19

Figure
Figure 19. 19. Frequencies
Frequencies of elements
of elements versus equivalent
versus equivalent stressesmacroscopic
stresses for several for several strains
macroscopic
for ABSstrains for
into ABS/PP and
ABS into ABS/PP
ABS/PP/PP-g-MA blends. and ABS/PP/PP-g-MA blends.

The number
The number of elements of equivalent
with elements with equivalent
stresses stresses
higher thanhigher
43 MPa than 43 MPa (corresponding
(corresponding
to the value of ABS breakage) was computed from these data (see Table 4: Percentage
to the value of ABS breakage) was computed from these data (see Table 4: Percentage of
of elements in ABS). At 1.9% of macroscopic strain both blends show a large number of
elements in ABS). At 1.9%
broken of macroscopic
elements in ABS, 39.3%strain both blends
for ABS/PP showfor
and 71.8% a large number of bro-
ABS/PP/PP-g-MA. Even if the
ken elements in ABS, 39.3% for
macroscopic ABS/PPVon
simulated and 71.8%
Mises for ABS/PP/PP-g-MA.
stresses are similar for bothEven if the mac- the local
microstructures,
roscopic simulated Von
stress Mises stresses
repartition area similar
leads to for both microstructures,
more premature breakage in the case theoflocal stress
ABS/PP/PP-g-MA
repartition leadsblend.
to a more prematurecan
This observation breakage inresult
explain the the case of ABS/PP/PP-g-MA
presented blend.
in Figure 5, the number of elements
This observationthat
can have
explainreached the critical
the result breakage
presented value is
in Figure 5,higher in the presence
the number of elementsof PP-g-MA
that thus
inducing progressively a reduction of stress in experiments for
have reached the critical breakage value is higher in the presence of PP-g-MA thus induc-ABS/PP/PP-g-MA blend
ing progressivelyinacomparison
reduction of with ABS/PP.
stress in experiments for ABS/PP/PP-g-MA blend in com-
parison with ABS/PP.
Table 4. Percentage of elements in ABS with an equivalent stress higher than the ABS strength for
ABS/PP and ABS/PP/PP-g-MA blends.
Table 4. Percentage of elements in ABS with an equivalent stress higher than the ABS strength for
ABS/PP and ABS/PP/PP-g-MA blends. % of Elements in ABS with an Equivalent Stress Higher than 43 Mpa

% of Elements in ABS with In

anABS/PP In ABS/PP/PP-g-MA
Equivalent Stress Higher than 43 Mpa
εmacro = 0.3% 0.0 0.0
In ABS/PP In ABS/PP/PP-g-MA
εmacro = 0.3% εmacro = 1.5% 0.0 4.2 0.0 5.6

εmacro = 1.5% εmacro = 1.9% 4.2 39.3 5.6 71.8

εmacro = 1.9% εmacro = 2.1% 39.3 80.6 71.8 86.9

εmacro = 2.1% 80.6 86.9

To complete this analysis cartography of porosity is given for 1.7% of macroscopic
To complete this (Figure
strain analysis20). In the PP nodules
cartography the porosity
of porosity never
is given forexceeds
1.7% of1% meaning that the PP
macroscopic
never reaches the flow stress state. The porosity shows a concentration around the PP
strain (Figure 20). In the PP nodules the porosity never exceeds 1% meaning that the PP
nodules in a direction perpendicular to the tensile direction. Even if these cartographies
never reaches the areflow
quitestress state.
similar, The porosity
the distribution of shows
porositya around
concentration around
the nodules seemsthe PP
different. The
nodules in a direction
Figure perpendicular
21 regroups the to the tensile
number direction.
of elements versusEven
localifporosities
these cartographies
into ABS polymer for
are quite similar,several
the distribution
macroscopic ofstrains
porosity(1.5,around
1.9, andthe nodules
2.1%) seemsand
for ABS/PP different. The Fig- blends.
ABS/PP/PP-g-MA
ure 21 regroups the
Thenumber
more theof elements versus
macroscopic imposed local porosities
strain increases,into
the ABS
morepolymer for sev- ABS/PP
the shift between
eral macroscopic strains (1.5, 1.9, and 2.1%) for ABS/PP and ABS/PP/PP-g-MA blends. The ABS/PP
and ABS/PP/PP-g-MA is pronounced. The breakage is initiated at the interface
more the macroscopic imposed strain increases, the more the shift between ABS/PP and
ABS/PP/PP-g-MA is pronounced. The breakage is initiated at the interface ABS/PP by cav-
itation; this phenomenon is increased while reducing the size of nodules using a compat-
ibilizing agent for example.
Polymers 2021, 13, 1647 16 of 19

Polymers
Polymers 2021,
2021, 13,
13, xx FOR
FOR PEER by cavitation;
PEER REVIEW
REVIEW this phenomenon is increased while reducing the size of nodules using
17 ofa
17 of 20
20
compatibilizing agent for example.

Figure 20. Porosity cartography at 1.7% of macroscopic strain in ABS/PP and ABS/PP/PP-g-MA
Figure
Figure 20.
20. Porosity
Porosity cartography
cartography at
at 1.7%
1.7% of
of macroscopic
macroscopic strain
strain in
in ABS/PP
ABS/PP and
and ABS/PP/PP-g-MA
ABS/PP/PP-g-MA
composites.
composites.
composites.

Figure 21.
Figure21. Frequencies
21.Frequencies
Frequenciesof of elements
ofelements versus
elementsversus effective
effective porosity
versus effective porosity fraction
fraction (see
(see Equation
Equation (4)) for several
Figure porosity fraction (see Equation (4))
(4)) for
for several
several
macroscopic strains for ABS into ABS/PP and ABS/PP/PP-g-MA composites.
macroscopic strains for ABS into ABS/PP and ABS/PP/PP-g-MA composites.
macroscopic strains for ABS into ABS/PP and ABS/PP/PP-g-MA composites.
Regarding
Regardingthethe volumetric
thevolumetric strain,
volumetricstrain, once
strain,once again
onceagain the
againthe prediction
theprediction gives
predictiongives satisfactory
givessatisfactory results
satisfactoryresults
results
Regarding
(Figure
(Figure 22).
22). Nevertheless,
Nevertheless, a
a deviation
deviation is
is observed
observed from
from 1.7%
1.7% of
of strain:
strain: the
the model
model underes-
underes-
(Figure 22). Nevertheless, a deviation is observed from 1.7% of strain: the model underesti-
timates
timates the volume change. This result is
is in
in accordance with previous analysis, after 1.7%
mates thethe volume
volume change.
change. ThisThis result
result is in accordance
accordance with
with previous
previous analysis,
analysis, afterafter
1.7%1.7%
of
of
of macroscopic
macroscopic strain
strain a
a large
large number
number of
of ABS
ABS elements
elements have
have exceeded
exceeded
macroscopic strain a large number of ABS elements have exceeded the ABS stress breakage. the
the ABS
ABS stress
stress
breakage.
breakage.
The The
The comparison
comparison comparison between
between experiments
between experiments and numericand
experiments and numeric is
is therefore
numericinvalid.
is therefore therefore invalid.
invalid. Exper-
Experimentally, Exper-
an
imentally,
interfacial an interfacial
imentally,decohesion
an interfacial decohesion
decohesion
between PP andbetween
between
ABS occursPP and
PP and ABS
thusABS occurs thus
occurs thus
increasing increasing
volumetricthe
the increasing the vol-
vol-
strain
umetric
umetric strain
explaining strain explaining
explaining
the observed the
the observed
observed deviation.
deviation. deviation.
Polymers 2021, 13, 1647 17 of 19
Polymers 2021, 13, x FOR PEER REVIEW 18 of 20

Figure22.
Figure 22.Volumetric
Volumetric strain
strain vs.vs. Lagrangian
Lagrangian strain
strain for for ABS/PP
ABS/PP composite,
composite, experiments,
experiments, and and model-
modelling.
ling.
5. Conclusions
5. Conclusions
Polymer mechanical recycling is viewed nowadays as an efficient way to reduce
plasticPolymer
pollutionmechanical
and reduce recycling is viewed nowadays
the environmental footprint as an
of efficient way to from
its production reduce plas-
fossil
tic pollution
resources. andisreduce
Sorting primordialthe environmental
to achieve economic footprint of its production
competitiveness since mostfrompolymers
fossil re-
sources.
are Sorting is
incompatible. primordial
Residual to achieve
impurities economic
are however competitiveness
inevitable. since most
Understanding polymers
the potential
impact of these impurities
are incompatible. Residual is impurities
thus important to evaluate
are however recyclability
inevitable. of end-of-lifethe
Understanding plastics.
poten-
Atial
first step isof
impact tothese
work impurities
on a modelisblend. thus PP was incorporated
important to evaluate at recyclability
4% in weightof into ABS as
end-of-life
they are two
plastics. major
A first constituents
step is to workofon WEEE. PP-g-MA
a model blend.was PPadded to these blends
was incorporated withinthe
at 4% aim
weight
tointo
improve
ABS astheir
theymechanical
are two major performances.
constituents of Experiments
WEEE. PP-g-MA and predictive
was added modelling were
to these blends
both
withusedthe to
aimassess first the ability
to improve of this blend
their mechanical to be reused atExperiments
performances. least in staticand applications,
predictive
and second, the necessity of compatibilization.
modelling were both used to assess first the ability of this blend to be reused at least in
The
static main conclusion
applications, of this
and second, thepaper is that
necessity of static properties of ABS are almost not
compatibilization.
impaired The by
mainpossible PP contamination
conclusion of this paper isatthat 4 wt%staticfrom impact of
properties tests
ABSwith this impurity
are almost not im-
rate.
paired Hence, this recycled
by possible material canatbe
PP contamination used from
4 wt% as animpact
ABS polymer,
tests with in this
the case of static
impurity rate.
application. Contrary to expectations, the addition of PP-g-MA, which
Hence, this recycled material can be used as an ABS polymer, in the case of static applica- was expected to
improve ABS/PP adhesion and to reduce the size of the nodular
tion. Contrary to expectations, the addition of PP-g-MA, which was expected to improvephase, induced a negative
effect
ABS/PP on the ultimate
adhesion andstress.
to reduceThe the
main reason
size of theisnodular
that aroundphase,theinduced
interfacial area, which
a negative effect
increases as the nodules sizes decrease, the level of stress triaxiality in
on the ultimate stress. The main reason is that around the interfacial area, which increases ABS is much higher.
Locally, the ABSsizes
as the nodules around PP nodules
decrease, cavitates
the level much
of stress more when
triaxiality in ABS nodules
is much are even smaller.
higher. Locally,
Tothe ABS around PP nodules cavitates much more when nodules are even smaller.mimic
go further in this predictive analysis, cohesive model zones should be used to To go
various
furtherinterfacial compatibilities.
in this predictive analysis,An alternative
cohesive model could
zonesalso be the
should beuse
usedof to
anmimic
intermediate
various
layer between
interfacial phases with graduate
compatibilities. properties
An alternative could[39].
alsoThe situation
be the use ofmay be very different
an intermediate layer
for dynamic applications as softer inclusions (PP are
between phases with graduate properties [40]. The situation may be very softer inclusions in ABS) different
can help tofor
dissipate energy and accommodate stress.
dynamic applications as softer inclusions (PP are softer inclusions in ABS) can help to
dissipate energy and accommodate stress.
Author Contributions: Conceptualization, C.S., A.-S.C.-B., J.-M.L.-C., P.I. and D.P.; methodology,
A.-S.C.-B. and P.I.; software, A.-S.C.-B.; validation; formal analysis, A.-S.C.-B. and P.I.; investiga-
Author Contributions: Conceptualization, C.S., A.-S.C.-B., J.-M.L.-C., P.I., D.P.; methodology, A.-
tion, C.S.; writing—original draft preparation, A.-S.C.-B.; writing—review and editing, All authors;
S.C.-B. and P.I.; software, A.-S.C.-B.; validation; formal analysis, A.-S.C.-B. and P.I.; investigation,
project administration, D.P. and P.I. All authors have read and agreed to the published version of
C.S.; writing—original draft preparation, A.-S.C.B.; writing—review and editing, All authors; pro-
the manuscript.
ject administration, D.P. and P.I. All authors have read and agreed to the published version of the
manuscript.
Funding: This work was supported by BPI France via the FUI 20 (Fonds unique interministériel) grant.
Funding: This workThe
Acknowledgments: wasauthors
supported by BPI
would like France viaBenjamin
to thank the FUI 20 (FondsAlexandre
Gallard, unique interministériel)
Cheron and
grant.
Jean-Claude Roux for technical support, respectively for polymer processing, mechanical testing and
morphology characterization. Suez and Pellenc ST are gratefully acknowledged for partnership in
Acknowledgments: The authors would like to thank Benjamin Gallard, Alexandre Cheron and
this work.
Jean-Claude Roux for technical support, respectively for polymer processing, mechanical testing
Polymers 2021, 13, 1647 18 of 19

Conflicts of Interest: The authors declare no conflict of interest.

References
1. Wang, R.X.Z. Recycling of non-metallic fractions from waste electrical and electronic equipment (WEEE): A review. Waste Manag.
2014, 34, 1455–1469. [CrossRef] [PubMed]
2. Buekens, A.; Yang, J. Recycling of WEEE plastics: A review. J. Mater. Cycles Waste Manag. 2014, 16, 415–434. [CrossRef]
3. Vanegas, P.; Peeters, J.R.; Cattrysse, D.; Tecchio, P.; Ardente, F.; Mathieux, F.; Dewulf, W.; Duflou, J.R. Ease of disassembly of
products to support circular economy strategies. Resour. Conserv. Recycl. 2018, 135, 323–334. [CrossRef]
4. Movilla, N.A.; Zwolinski, P.; Dewulf, J.; Mathieux, F. A method for manual disassembly analysis to support the ecodesign of
electronic displays. Resour. Conserv. Recycl. 2016, 114, 42–58. [CrossRef]
5. Robinson, B.H. E-waste: An assessment of global production and environmental impacts. Sci. Total. Environ. 2009, 408, 183–191.
[CrossRef]
6. Turner, A. Black plastics: Linear and circular economies, hazardous additives and marine pollution. Environ. Int. 2018, 117,
308–318. [CrossRef]
7. Cucchiella, F.; D’Adamo, I.; Koh, S.L.; Rosa, P. Recycling of WEEEs: An economic assessment of present and future e-waste
streams. Renew. Sustain. Energy Rev. 2015, 51, 263–272. [CrossRef]
8. Hennebert, P.; Filella, M. WEEE plastic sorting for bromine essential to enforce EU regulation. Waste Manag. 2018, 71, 390–399.
[CrossRef]
9. Guzzonato, A.; Puype, F.; Harrad, S.J. Evidence of bad recycling practices: BFRs in children’s toys and food-contact articles.
Environ. Sci. Process. Impacts 2017, 19, 956–963. [CrossRef]
10. Beigbeder, J.; Perrin, D.; Mascaro, J.-F.; Lopez-Cuesta, J.-M. Study of the physico-chemical properties of recycled polymers from
waste electrical and electronic equipment (WEEE) sorted by high resolution near infrared devices. Resour. Conserv. Recycl. 2013,
78, 105–114. [CrossRef]
11. Perrin, D.; Mantaux, O.; Ienny, P.; Léger, R.; Dumon, M.; Lopez-Cuesta, J.-M. Influence of impurities on the performances of HIPS
recycled from Waste Electric and Electronic Equipment (WEEE). Waste Manag. 2016, 56, 438–445. [CrossRef]
12. Das, A.; Chatham, C.A.; Fallon, J.J.; Zawaski, C.E.; Gilmer, E.L.; Williams, C.B.; Bortner, M.J. Current understanding and
challenges in high temperature additive manufacturing of engineering thermoplastic polymers. Addit. Manuf. 2020, 34, 101218.
[CrossRef]
13. Goh, G.; Toh, W.; Yap, Y.; Ng, T.; Yeong, W. Additively manufactured continuous carbon fiber-reinforced thermoplastic for
topology optimized unmanned aerial vehicle structures. Compos. Part B Eng. 2021, 216, 108840. [CrossRef]
14. Banerjee, S.S.; Burbine, S.; Shivaprakash, N.K.; Mead, J. 3D-printable PP/SEBS thermoplastic elastomeric blends: Preparation and
properties. Polymers 2019, 11, 347. [CrossRef] [PubMed]
15. Martinho, G.; Pires, A.; Saraiva, L.; Ribeiro, R. Composition of plastics from waste electrical and electronic equipment (WEEE) by
direct sampling. Waste Manag. 2012, 32, 1213–1217. [CrossRef] [PubMed]
16. Maris, E.; Botané, P.; Wavrer, P.; Froelich, D. Characterizing plastics originating from WEEE: A case study in France. Miner. Eng.
2015, 76, 28–37. [CrossRef]
17. Bonda, S.; Mohanty, S.; Nayak, S.K. Influence of compatibilizer on mechanical, morphological and rheological properties of
PP/ABS blends. Iran. Polym. J. 2014, 23, 415–425. [CrossRef]
18. Tostar, S.; Stenvall, E.; Foreman, M.R.S.J.; Boldizar, A. The influence of compatibilizer addition and gamma irradiation on
mechanical and rheological properties of a recycled WEEE plastics blend. Recycling 2016, 1, 101–110. [CrossRef]
19. Patel, A.C.; Brahmbhatt, R.B.; Devi, S. Mechanical properties and morphology of PP/ABS blends compatibilized with PP-g-2-
HEMA. J. Appl. Polym. Sci. 2001, 88, 72–78. [CrossRef]
20. Patel, A.C.; Brahmbhatt, R.B.; Sarawade, B.D.; Devi, S. Morphological and mechanical properties of PP/ABS blends compatibilized
with PP-g-acrylic acid. J. Appl. Polym. Sci. 2001, 81, 1731–1741. [CrossRef]
21. Deng, Y.; Mao, X.; Lin, J.; Chen, Q. Compatibilization of polypropylene/Poly(acrylonitrile-butadiene-styrene) blends by
polypropylene-graft-cardanol. J. Appl. Polym. Sci. 2015, 132, 41315. [CrossRef]
22. Luo, Z.; Lu, Q.; Ma, F.; Jiang, Y. The effect of graft copolymers of maleic anhydride and epoxy resin on the mechanical properties
and morphology of PP/ABS blends. J. Appl. Polym. Sci. 2014, 131. [CrossRef]
23. IIbrahim, M.A.H.; Hassan, A.; Wahit, M.U.; Hasan, M.; Mokhtar, M. Mechanical properties and morphology of polypropylene/
poly(acrylonitrile–butadiene–styrene) nanocomposites. J. Elastom. Plast. 2017, 49, 209–225. [CrossRef]
24. Kum, C.K.; Sung, Y.-T.; Kim, Y.S.; Lee, H.G.; Kim, W.N.; Lee, H.S.; Yoon, H.G. Effects of compatibilizer on mechanical,
morphological, and rheological properties of polypropylene/poly(acrylonitrile-butadiene-styrene) blends. Macromol. Res.
2007, 15, 308–314. [CrossRef]
25. Lee, Y.K.; Lee, J.B.; Park, D.H.; Kim, W.N. Effects of accelerated aging and compatibilizers on the mechanical and morphological
properties of polypropylene and poly(acrylonitrile-butadiene-styrene) blends. J. Appl. Polym. Sci. 2013, 127, 1032–1037. [CrossRef]
26. Gurson, A.L. Continuum theory of ductile rupture by void nucleation and growth: Part I—Yield criteria and flow rules for porous
ductile media. J. Eng. Mater. Technol. 1977, 99, 2–15. [CrossRef]
27. Tvergaard, V. On localization in ductile materials containing spherical voids. Int. J. Fract. 1982, 18, 237–252.
28. Tvergaard, V.; Needleman, A. Analysis of the cup-cone fracture in a round tensile bar. Acta Met. 1984, 32, 157–169. [CrossRef]
Polymers 2021, 13, 1647 19 of 19

29. Moore, J. Acrylonitrile-butadiene-styrene (ABS)—A review. Composites 1973, 4, 118–130. [CrossRef]

30. Signoret, C.; Girard, P.; Guen, A.; Caro-Bretelle, A.-S.; Lopez-Cuesta, J.-M.; Ienny, P.; Perrin, D. Degradation of styrenic plastics
during recycling: Accommodation of PP within ABS after WEEE plastics imperfect sorting. Polymers 2021, 13, 1439. [CrossRef]
31. Christmann, A.; Ienny, P.; Quantin, J.; Caro-Bretelle, A.; Lopez-Cuesta, J. Mechanical behaviour at large strain of polycarbonate
nanocomposites during uniaxial tensile test. Polymer 2011, 52, 4033–4044. [CrossRef]
32. Caro-Bretelle, A.; Ienny, P.; Leger, R. Constitutive modeling of a SEBS cast-calender: Large strain, compressibility and anisotropic
damage induced by the process. Polymer 2013, 54, 4594–4603. [CrossRef]
33. Hall, R.A. Computer modelling of rubber-toughened plastics: Random placement of monosized core-shell particles in a polymer
matrix and interparticle distance calculations. J. Mater. Sci. 1991, 26, 5631–5636. [CrossRef]
34. Besson, J. Continuum models of ductile fracture: A review. Int. J. Damage Mech. 2010, 19, 3–52. [CrossRef]
35. Boisot, G.; Laiarinandrasana, L.; Besson, J.; Fond, C.; Hochstetter, G. Experimental investigations and modeling of volume change
induced by void growth in polyamide 11. Int. J. Solids Struct. 2011, 48, 2642–2654. [CrossRef]
36. Tvergaard, V. Influence of voids on shear band instabilities under plane strain conditions. Int. J. Fract. 1981, 17, 389–407.
[CrossRef]
37. Caro, A.; Bernardeau, F.; Perrin, D.; Leger, R.; Benezet, J.; Ienny, P. Computational modelling of void growth in Phenolic Molding
Compounds filled PolyPropylene from optical measurements. Polym. Test. 2018, 71, 209–216. [CrossRef]
38. Besson, J.; Leriche, R.; Foerch, R.; Cailletaud, G. Object-oriented programming applied to the finite element method Part II.
application to material behaviors. Rev. Eur. Élém. Finis 1998, 7, 567–588. [CrossRef]
39. Liu, J.L.; Xia, R.; Zhou, X.H. A new look on wetting models: Continuum analysis. Sci. China Phys. Mech. Astron. 2012, 55, 1–9.
[CrossRef]