Mechanical recycling of high density polyethylene/flax fiber composites

Thèse

Nathalie Benoit

Doctorat en génie chimique

Philosophiæ doctor (Ph. D.)

Québec, Canada

© Nathalie Benoit, 2017

Mechanical recycling of high density polyethylene/flax fiber composites

Thèse

Nathalie Benoit

Sous la direction de :

Denis Rodrigue, directeur de recherche Rubén González-Núñez, codirecteur de recherche

iii

Résumé Ce travail de doctorat est consacré à la production, au recyclage mécanique long-terme et à la caractérisation

de matériaux polymères et composites à base de polyéthylène haute densité (HDPE) et de fibre de lin. L’objectif

est de déterminer l’aptitude au recyclage long-terme de ces composites et de leur matrice, tout en évaluant la

perte de performance subie. Le recyclage est réalisé ici par une extrusion en boucle fermée, durant 50 cycles,

sans ajout intermédiaire de matières vierges et sans prise en compte de la détérioration et de la contamination

subies lors du cycle de vie des produits.

Dans la première partie, une revue de littérature présente l’état de l’art concernant le recyclage mécanique des

composites thermoplastiques. Les différents types de recyclage de composites sont présentés, ainsi que les

différents travaux réalisés sur le recyclage de composites thermoplastiques à base de fibres naturelles ou

inorganiques. Enfin, les différentes limitations rencontrées lors du recyclage de ces composites sont mises en

lumière et des solutions sont présentées. Au cours de cette revue, des lacunes importantes sur le recyclage

mécanique long-terme de ces composites sont observées.

Dans la seconde partie de ce travail, le polyéthylène haute densité est étudié et recyclé seul afin de connaître

ses propriétés et son comportement au recyclage, tout en servant de base de comparaison pour les composites

produits par la suite. L’étude des propriétés physique, thermique, moléculaire et mécanique permet d’analyser

les différents mécanismes de dégradation induits par le recyclage mécanique. Les résultats montrent une

diminution de la contrainte au seuil d’écoulement et une forte augmentation de l’élongation à la rupture avec le

recyclage, indiquant que des phénomènes de rupture de chaînes ont lieu dans le polymère. La plupart des

autres propriétés demeurent constantes et confirment le maintien des performances du polymère avec le

recyclage.

Dans la dernière partie de cette thèse, deux séries de composites sont produites à partir du polyéthylène haute

densité et de la fibre de lin (15% en masse), avec et sans polyéthylène greffé d’anhydride maléique (MAPE)

comme agent couplant. Toutes deux seront caractérisées similairement au polymère afin d’évaluer l’effet de la

présence de fibre dans le polymère. Une analyse de la distribution de fibres est aussi réalisée afin d’observer

l’effet du recyclage mécanique sur la taille des fibres. L’analyse mécanique révèle que la fibre fournit un renfort

efficace au polymère, en particulier avec l’agent couplant, mais les propriétés à la rupture diminuent. Cet effet

diminue avec le recyclage, alors que les propriétés à l’élongation augmentent, du fait de la réduction de longueur

des fibres. L’effet de l’agent couplant disparaît aussi au cours du recyclage. Toutefois, la majorité des

performances mécaniques après recyclage restent supérieures à celles du polymère.

iv

Abstract This thesis focuses on the production, the mechanical recycling and the characterization of polymers and

composites based on high density polyethylene (HDPE) and flax fibers. It aims to determine the materials

potential towards long-term recycling and to evaluate the resulting loss of performance. The recycling is realized

by closed-loop extrusion, and repeated up to 50 times, without any addition of new material, and without any

consideration of the possible degradation and contamination undergone during the life-cycle of the products.

In the first part, a literature review presents the state of the art concerning the mechanical recycling of

thermoplastic composites. The various types of composites recycling are introduced, as well as the various

works conducted on the recycling of thermoplastic composites reinforced with both natural and inorganic fillers.

Finally, the various limitations to the composites recycling are presented and some solutions are suggested.

During this review an important lack of knowledge on the long-term mechanical recycling of these composites is

observed.

In the second part of this work, the high density polyethylene is studied and recycled in order to know its

properties and its behavior towards recycling, as well as to be used as a comparison basis for the further parts.

The study of the mechanical, thermal, molecular and physical properties leads to the better understanding of the

various degradation mechanisms induced by mechanical recycling. The results show a decrease of the yield

stress and an important increase of the strain at break with recycling, indicating that chain scissions take place

in the polymer during recycling. Most of the other properties remained stable, and confirmed the conservation of

the polymer performances with recycling.

In the last part of this work, high density polyethylene is used to produce two series of composites with 15% wt.

of flax fiber, with and without maleic anhydride grafted polyethylene (MAPE) as a coupling agent. Similar

characterizations as for the matrix are conducted on both composites as to evaluate the effect of the fibers in

the polymer matrix. A complete analysis of the fiber distribution is also performed to observe the effect of

mechanical recycling on the fiber dimensions. The mechanical analysis reveals that the fibers provides an

efficient reinforcement to the matrix, and especially with coupling agent, but the properties at break decrease.

Nevertheless, this effect decreases with recycling, while the elongation properties increase due to the fiber size

reduction. The effect of the coupling agent disappears with recycling. However, most mechanical properties

remain higher for the composites after recycling than for the neat matrix.

v

Table of content

Résumé ........................................................................................................................................................ iii

Abstract ....................................................................................................................................................... iv

Table of content ............................................................................................................................................ v

List of Tables .............................................................................................................................................. viii

List of Figures .............................................................................................................................................. ix

Abbreviations ............................................................................................................................................... xi

Symbols ..................................................................................................................................................... xiii

Acknowledgments ....................................................................................................................................... xiv

Forewords ...................................................................................................................................................xv

Chapter I. Introduction .............................................................................................................................. 1

I.1 Natural fiber and thermoplastic based composites ................................................................................ 1

I.1.1 Basic notions about composites .................................................................................................... 1

I.1.2 Polymer matrices .......................................................................................................................... 2

I.1.3 Natural fibers ................................................................................................................................ 3

I.1.4 Natural fibers vs. synthetic fibers ................................................................................................... 5

I.1.5 Interface properties and coupling agent ......................................................................................... 6

I.1.6 Applications .................................................................................................................................. 7

I.2 Recycling ............................................................................................................................................. 9

I.2.1 Polymer recycling ......................................................................................................................... 9

I.2.2 Composite recycling .................................................................................................................... 10

I.3 Thesis objective and organization ....................................................................................................... 10

Chapter II. Mechanical recycling of thermoplastic composites .................................................................. 13

Résumé ...................................................................................................................................................... 13

Abstract ...................................................................................................................................................... 14

II.1 Introduction ....................................................................................................................................... 15

II.2 Recycling methods............................................................................................................................ 18

II.2.1 Thermal recycling ...................................................................................................................... 19

II.2.2 Chemical recycling ..................................................................................................................... 20

II.2.3 Mechanical recycling .................................................................................................................. 20

II.2.4 Conclusion................................................................................................................................. 21

II.3 Recycling of thermoplastic composites reinforced with organic fillers .................................................. 22

vi

II.3.1 Wood ........................................................................................................................................ 23

II.3.2 Cellulose ................................................................................................................................... 28

II.3.3 Flax ........................................................................................................................................... 29

II.3.4 Sisal and hemp .......................................................................................................................... 30

II.3.5 Rice hulls and kenaf ................................................................................................................... 31

II.3.6 Nettle ........................................................................................................................................ 32

II.3.7 Conclusion................................................................................................................................. 32

II.4 Recycling of thermoplastic composites reinforced with inorganic fillers ............................................... 35

II.4.1 Glass ......................................................................................................................................... 35

II.4.2 Carbon ...................................................................................................................................... 38

II.4.3 Talc ........................................................................................................................................... 39

II.4.4 Conclusion................................................................................................................................. 39

II.5 Limitations and solutions ................................................................................................................... 41

II.6 Conclusion........................................................................................................................................ 44

Acknowledgements ..................................................................................................................................... 46

Chapter III. Long-term recycling of high density polyethylene and characterization of its closed-loop degradation………………………………………………………………………………………………………………..47

Résumé ...................................................................................................................................................... 47

Abstract ...................................................................................................................................................... 48

III.1 Introduction ...................................................................................................................................... 49

III.2 Materials .......................................................................................................................................... 52

III.3 Experimental.................................................................................................................................... 52

III.3.1 Sample Production .................................................................................................................... 52

III.3.2 Physical Properties ................................................................................................................... 54

III.3.3 Thermal Properties ................................................................................................................... 54

III.3.4 Mechanical Properties ............................................................................................................... 55

III.4 Results ............................................................................................................................................ 55

III.4.1 Density ..................................................................................................................................... 55

III.4.2 Gel Permeation Chromatography .............................................................................................. 55

III.4.3 Melt Flow Index ......................................................................................................................... 59

III.4.4 Thermogravimetric Analysis ...................................................................................................... 59

III.4.5 Differential Scanning Calorimetry............................................................................................... 59

III.4.6 Tensile Properties ..................................................................................................................... 59

vii

III.4.7 Flexural Properties .................................................................................................................... 64

III.5 Conclusion ....................................................................................................................................... 64

Acknowledgements ..................................................................................................................................... 65

Chapter IV. Long-term closed-loop recycling of high density polyethylene/flax composites. .................... 66

Résumé ...................................................................................................................................................... 66

Abstract ...................................................................................................................................................... 67

IV.1 Introduction ..................................................................................................................................... 68

IV.2 Materials ......................................................................................................................................... 71

IV.3 Experimental ................................................................................................................................... 72

IV.3.1 Sample Production ................................................................................................................... 72

IV.3.2 Physical Properties ................................................................................................................... 74

IV.3.3 Thermal Properties ................................................................................................................... 75

IV.3.4 Mechanical Properties .............................................................................................................. 75

IV.3.5 Morphology .............................................................................................................................. 76

IV.4 Results and Discussion .................................................................................................................... 76

IV.4.1 Density ..................................................................................................................................... 76

IV.4.2 Gel Permeation Chromatography .............................................................................................. 76

IV.4.3 Thermogravimetric Analysis ...................................................................................................... 80

IV.4.4 Differential Scanning Calorimetry .............................................................................................. 83

IV.4.5 Morphology .............................................................................................................................. 85

IV.4.6 Tensile Properties ..................................................................................................................... 91

IV.4.7 Bending Properties ................................................................................................................... 96

IV.4.8 Impact Properties ..................................................................................................................... 96

IV.5 Conclusions .................................................................................................................................... 97

Acknowledgements ..................................................................................................................................... 99

Chapter V. Conclusions and recommendations ...................................................................................... 100

V.1 General conclusion ......................................................................................................................... 100

V.2 Perspectives .................................................................................................................................. 102

References ............................................................................................................................................... 104

viii

List of Tables

Table I-1: Classification of natural fibers according to their origin. ................................................................... 4

Table I-2: Comparison between the physical and mechanical properties of usual natural and synthetic fibers [26, 27, 28, 29].............................................................................................................................................. 6

Table I-3: Plastics codes in Canada [52]. ...................................................................................................... 9

Table I-4: Plastic waste recovered in Quebec during 2008 [52]. .................................................................... 10

Table II-1: Recycling methods for thermoplastics composites and their characteristics. ................................. 22

Table II-2: Overview of the investigations published on the mechanical recycling of natural organic fillers reinforced composites with their main parameters. ....................................................................................... 34

Table II-3: Overview of the works considering the mechanical recycling of inorganic fillers reinforced composites with their main characteristics. .................................................................................................. 40

Table II-4: Main limitations and solutions for the mechanical recycling of thermoplastic composites. .............. 44

Table III-1: Injection molding parameters. .................................................................................................... 53

Table IV-1: Injection molding parameters. .................................................................................................... 74

Table IV-2: Average molecular weights for the matrix and the composites. ................................................... 77

ix

List of Figures

Figure I-1: Examples of potential and current commercial applications for natural fiber based composites [28, 42, 43, 44, 45, 46, 47, 48, 49, 50]. ................................................................................................................. 8

Figure III-1: Virgin HDPE thermogram from DSC. ........................................................................................ 52

Figure III-2 : Molecular weight distribution for generations 0 (PG0) and 50 (PG50). ....................................... 56

Figure III-3: Number average molecular weight as a function of generation number. ..................................... 56

Figure III-4: Weight average molecular weight as a function of generation number. ....................................... 57

Figure III-5: Polydispersity index as a function of generation number. ........................................................... 57

Figure III-6: Intrinsic viscosity as a function of generation number................................................................. 58

Figure III-7: Typical tensile stress-strain curves for different generation. ....................................................... 60

Figure III-8: Stress at break as a function of generation number. .................................................................. 61

Figure III-9: Strain at break as a function of generation number. ................................................................... 61

Figure III-10: Yield stress as a function of generation number. ...................................................................... 62

Figure III-11: Yield strain as function of generation number. ......................................................................... 62

Figure III-12: Young's modulus as a function of generation number. ............................................................. 63

Figure III-13: Energy at break as a function of generation number. ............................................................... 63

Figure III-14: Flexural modulus as a function of generation number. ............................................................. 64

Figure IV-1: Initial fiber L/D aspect ratio range distribution. ........................................................................... 71

Figure IV-2: Initial fiber average length range distribution. ............................................................................ 72

Figure IV-3: Average molecular weights as a function of generation number for the CS composites. ............. 78

Figure IV-4: Average molecular weights as a function of generation number for CA composites. ................... 78

Figure IV-5: Number average molecular weight (Mn) as a function of generation number. ............................. 79

Figure IV-6: Weight average molecular weight (Mw) as a function of generation number. ............................... 79

Figure IV-7: Polydispersity index as a function of generation number. ........................................................... 80

Figure IV-8: TGA results for the matrix. ........................................................................................................ 81

Figure IV-9: TGA results for the composites. ................................................................................................ 82

Figure IV-10: Peak temperature obtained from the weight derivative curves for the matrix and the composites as a function of generation number.............................................................................................................. 82

Figure IV-11: Onset degradation temperature for the matrix and the composites as a function of generation number. ...................................................................................................................................................... 83

Figure IV-12: Melting point as a function of generation number. ................................................................... 84

Figure IV-13: Melting peak width as a function of generation number............................................................ 84

Figure IV-14: Crystallinity as a function of generation number. ..................................................................... 85

Figure IV-15: Average fiber length as a function of generation number. ........................................................ 86

Figure IV-16: Average fiber diameter as a function of generation number. .................................................... 86

Figure IV-17: Fiber length distribution for the CS composites. ....................................................................... 87

Figure IV-18: Fiber length distribution for the CA composites. ....................................................................... 87

Figure IV-19: Fiber L/D ratio distribution for the CS composites. ................................................................... 88

Figure IV-20: Fiber L/D ratio distribution for the CA composites. ................................................................... 88

Figure IV-21: Average L/D ratio as a function of generation number. ............................................................ 89

Figure IV-22: Typical tensile stress-strain curves for different samples.......................................................... 92

Figure IV-23: Stress at break as a function of generation number. ................................................................ 92

Figure IV-24: Strain at break as a function of generation number. ................................................................. 93

x

Figure IV-25: Yield stress as a function of generation number. ..................................................................... 93

Figure IV-26: Yield strain as a function of generation number. ...................................................................... 94

Figure IV-27: Young's modulus as a function of generation number. ............................................................. 94

Figure IV-28: Energy at break as a function of generation number. ............................................................... 95

Figure IV-29: Bending modulus as a function of generation number. ............................................................ 96

Figure IV-30: Impact strength as a function of generation number. ............................................................... 97

xi

Abbreviations

ABS Acrylonitrile butadiene styrene

AFM Atomic-force microscopy

ATR Attenuated total reflectance

BLS Bis(1-octyloxy-2,2,6,6-tetramethyl-4-piperidyl) sebacate

CMC Ceramic matrix composite

DSC Differential scanning calorimetry

ELV End-of-Life Vehicles

FBC Fluidized-bed combustion

FRTP Fiber reinforced thermoplastics

FTIR Fourier transform infrared spectroscopy

GMT Glass mat thermoplastic

GPC Gel permeation chromatography

HDPE High density polyethylene

HT-DTA High temperature dynamic-thermal analysis

HT-GPC High temperature gel permeation chromatography

LALS Low angle light scattering

LDPE Low density polyethylene

LLDPE Linear low density polyethylene

MA Maleic anhydride

MAPE Maleic anhydride-grafted polyethylene

MAPP Maleic anhydride-grafted polypropylene

MIR Mid infrared

MMC Metal matrix composite

MSW Municipal solid waste

NIR Near infrared

PA Polyamide

PA12 Polyamide 12

PA66 Polyamide 66

PC Polycarbonate

PE Polyethylene

PEEK Polyetheretherketone

PET Polyethylene terephthalate

xii

PLA Poly lactic acid

PLLA Poly-L-lactic acid

PMC Polymer matrix composite

PMPI Polymethylene polyphenyl isocyanate

PP Polypropylene

PS Polystyrene

PUR Polyurethane

PVC Polyvinyl chloride

RI Refractive Index

SEM Scanning electron microscope

TCB 1,2,4-trichlorobenzene

TDA Triple detector array

TGA Thermogravimetric analysis

UV Ultraviolet

WPC Wood plastic composites

xiii

Symbols

CA Composite with coupling agent

CAGN Composite with coupling agent at generation “N”

CS Composite without coupling agent

CSGN Composite without coupling agent at generation “N”

d Fiber diameter (mm)

E Young’s modulus (MPa)

Eb Bending modulus (MPa)

Fbk Impact strength (kJ/m²)

IV Intrinsic viscosity (dl/g)

L Fiber length (mm)

L/D Aspect ratio (-)

MFI Melt flow index (g/10 min)

MFR Melt flow rate (g/10 min)

Mn Number average molecular weight (kDa)

Mw Weight average molecular weight (kDa)

PDI Polydispersity index (-)

PGN Polymer at Generation “N”

TEB Tensile energy at break (J)

Tg Glass transition temperature (°C)

Tm Melting point (°C)

wt.% Percentage by mass (%)

Xcr Degree of crystallinity (%)

∆Hf Melting enthalpy (J/g)

∆H100 Heat of fusion of 100% crystalline material (J/g)

σb Stress at break (MPa)

σy Yield stress (MPa)

εb Strain at break (MPa)

εy Yield strain (MPa)

xiv

Acknowledgments First and foremost, I would like to thank my supervisor Professor Denis Rodrigue for his trust, his support and

help all along this project. I thank him for his advice and insight in this project, as well as for his availability. I am

also grateful for the opportunity to go to Guadalajara and Guelph. I also would like to thank Professor Rubén

González-Núñez, my codirector, for his help, his support, and for his kind welcome during my stay for the project

in Guadalajara.

All those years would not have been the same without the people I worked with during this time. First of all, I

would like to thank Yann Giroux for his training and help on different equipment, but above all, for his priceless

help and support all along this project. Thanks to him for being there for me day after day. Then, I would like to

thank all my colleagues and the staff in the Chemical Engineering Department of Université Laval. At last, I want

to thank all the group member of the Proyectos and the Ingeniería Química departments of the Centro

Universitario de Ciencias Exactas e Ingenierías for their help, their warm welcome and their support during my

stay in Guadalajara.

I acknowledge the financial support of the Natural Sciences and Engineering Research Council of Canada

(NSERC), the Centre Québécois sur les Matériaux Fonctionnels (CQMF), the Centre de Recherche sur les

Matériaux Renouvelables (CRMR) and the Centre de Recherche sur les Systèmes Polymères et Composites à

Haute Performance (CREPEC).

Last, but not least, I want to thank all my friends and family members, especially my parents, my boyfriend and

my best friends for their love, support, patience understanding and encouragement.

xv

Forewords This thesis is composed of five chapters. The first chapter is a general introduction on natural fiber reinforced

composites. It presents different aspects of these materials such as their properties, applications and recycling.

The constituents and their specificities are also presented, as well as general notions and statistics on their

recycling. The second chapter is a review on the mechanical recycling of thermoplastic composites published

as a book chapter. In this chapter, the different recycling methods for composites were presented briefly. Then,

the different works and studies on the mechanical recycling of organic and inorganic fiber reinforced

thermoplastics were reviewed. Finally, the limitations associated with composites recycling and their possible

solutions were presented in the last part of this work. This chapter is accepted for publication as:

Benoit, N., González-Núñez, R. and Rodrigue, D. Mechanical recycling of thermoplastic composites,

in Thermoplastic Composites: Emerging Technology, Uses and Prospect. E. Ritter Ed., Nova Science

Publishers, New York, Chapter 3, pp. 95-142, ISBN: 978-1-53610-727-2 (2017).

Chapters III and IV present experimental results in the form of published or submitted journal papers. In chapter

III, high density polyethylene was recycled up to 50 times by closed-loop extrusion cycles. The physical, thermal,

mechanical and molecular properties were analyzed to understand the effect of long-term recycling on the

material structure and performances. This paper is accepted as:

Benoit, N., González-Núñez, R. and Rodrigue, D. High density polyethylene degradation followed by

closed-loop recycling, Progress in Rubber, Plastics & Recycling Technology, 33, 17 (2017).

Chapter IV investigates the long-term recycling of composites. These composites were made from the same

high density polyethylene reinforced with 15% wt. of flax fiber, with and without the addition of maleic anhydride

grafted polyethylene as a coupling agent. Up to 50 closed-loop reprocessing cycles were conducted on the

composites to analyze their effect on such materials and their constituents. Thermal, physical, morphological,

mechanical and molecular characterizations were performed to evaluate the potential of such materials towards

recycling. This paper has been submitted as:

Benoit, N., González-Núñez, R. and Rodrigue, D. Long-term recycling of high density polyethylene/flax

fiber composites and characterization of its closed-loop degradation, Progress in Rubber, Plastics and Recycling

Technology, submitted, 2016.

Finally, the fifth chapter is a general conclusion about the above-mentioned works followed by recommendations

for future works. It should be mentioned that for all the publications, my contribution was performing the

xvi

experimental works, collecting and analyzing the data, writing the first draft of the manuscripts and making all

the corrections. The publications were then revised by all co-authors.

More results obtained from this work were also presented in the following conference presentation:

Nathalie Benoit, Denis Rodrigue, Effect of recycling and weld-lines on the properties of injection

molded high density polyethylene reinforced with flax fibers, 14th International Symposium on Bioplastics,

Biocomposites & Biorefining, May 31st - June 3rd 2016, Guelph, ON, Canada.

.

1

Chapter I. Introduction

I.1 Natural fiber and thermoplastic based composite s

I.1.1 Basic notions about composites

Composites are constituted of at least two phases, one being the reinforcement (the fibers or particles), and the

other one the binder (the matrix). The association of these two materials leads to interesting properties resulting

from a synergetic combination of both components properties. The matrix binds the fibers in order to distribute

the stresses undergone, protect the fibers and provide the cohesion, the shape and the non-structural properties

and of the composite material. The reinforcement increases the rigidity and the mechanical performances of the

matrix [1, 2]. It can be found in a wide diversity of shapes and configurations: powder, particles, fibers (log and

short), fabric, layers, etc. Composite materials exhibit some typical characteristics allowing them to be

distinguished from blends. First, the composite components should be immiscible and an interface should exists

between them. They generally have higher properties than the bulk matrix material. Depending on the matrix

nature, composites can be classified in three main categories [2]:

• Polymer matrix composites (PMC) are the most commonly used composites nowadays due to their

easy processing, as well as the availability and low cost of their components. The presence of the reinforcement

in the polymer matrix increases most of its mechanical properties, thus allowing them to be used in more

demanding applications. Polymer based composites can be processed through similar methods than neat

polymers, such as extrusion, injection, compression molding, rotomolding, etc. They can be divided in two main

categories: the common composites, which have low cost and moderate properties, and the high performances

ones, which are more expensive and essentially used in aeronautics and sports for more demanding

applications.

• Ceramic matrix composites (CMC) are mainly used for structural and non-structural applications at very

high temperatures and very high level of stresses. As ceramics exhibit a fragile behavior, especially towards

impact stresses, they are often reinforced with ductile material as to deviate the crack propagation. Due to their

matrix properties, they also exhibit high porosity, high chemical resistance and high rigidity. They are also lighter

than most of the metals used for similar applications. However, due to the materials used and the complexity of

processing, those materials are usually expensive and are mainly used for spatial and aeronautic applications.

• Metal matrix composites (MMC) are high performance materials. However, they are also still expensive,

thus limiting their potential applications. They are generally constituted of a light metal matrix, such as aluminum,

magnesium or titanium, reinforced with a ceramic or metal reinforcement (fibers, particles or powders). If short

ceramic fibers are used, then traditional metal processing methods can be used. On the other hand, if long

2

ceramic or metal fibers are used, the processing methods are more complex and expensive. Another major

drawback to this type of composite is the potential high reactivity of the components, which limits the combination

of matrix and reinforcements that can be used.

Natural fiber thermoplastic composites are formed by a thermoplastic phase as the matrix and natural fibers as

reinforcement. They are more and more used due to their wide range of properties and their low cost. This thesis

will focus on a thermoplastic matrix, and especially on high density polyethylene (HDPE).

I.1.2 Polymer matrices

Polymers can be classified in three main categories, defined by their thermomechanical properties:

thermoplastics, thermosets and elastomers [2, 3, 4]. In thermoplastic polymers, the molecular chains are

generally not crosslinked. These polymers often exhibit an elastic-plastic behavior, some ductility and are

thermoformable. They can be melted and then molded, formed or welded when submitted to high temperature.

This process is repeatable and can be done each time the polymer is submitted to high temperatures. They can

be used with simple processing methods and exhibit some flexibility. Before processing, they can also be stored

at ambient conditions for long time periods. They can be amorphous or semi-crystalline. Crystallinity can be up

to 80% and grants the polymer opacity, high thermal expansion coefficient and high wear and failure resistance,

as well as higher mechanical properties. On the contrary, the thermosets are constituted of a three dimensional

tight-meshed crosslinked network between the molecular chains. They are processed through a curing reactive

step and cannot be reshaped after hardening, even at high temperatures, making their recycling very difficult.

Before processing, they have to be stored at low temperatures, and for limited time periods, if they already

contained the hardener. Finally, elastomers are characterized by their ability to have important elasticity and

relaxation behaviors despite their wide-meshed crosslinking of the chains. They are malleable under stress, but,

as thermosets, they generally cannot be melted. Elastomers can be thermosets or thermoplastics depending on

their nature, but due to their special properties, they are often considered as a specific category of polymers [2,

3, 4]. Most of the polymers can be used as matrices for the production of natural fiber composites. However,

crosslinked elastomers and thermosets cannot be easily recycled.

Thermoplastics are often classified in three categories depending on their price and their volume of production

[5]. Commodity plastics are the polymers used in high volume with a wide range of applications. They have an

annual production of several hundred million tons. They have relatively good mechanical properties and their

cost is generally less than a few dollars per kg. PE, PP, PVC and PS are typical examples of such thermoplastics.

Technical thermoplastics are used for more technical applications. They present slightly higher mechanical

properties, but their cost is higher, reaching five to ten dollars per kg, and their annual production volume is

around ten million tons. ABS, PA, PC and PUR are examples of technical thermoplastics. Finally, specialty

3

thermoplastics are polymers that are used for very specific and demanding applications. They exhibit much

higher mechanical properties, but are also much more expensive as their price ranges from ten to several

hundreds of dollars per kg. Due to their high cost, their production is less than one million tons per year.

Examples of such polymers are PPS, PEEK, PEI and LCP.

All these types can be used for thermoplastic composites, but, due to their low cost and their availability,

commodity plastics are more often used. However, due to the temperature limitation induced by the presence of

natural fibers, not all thermoplastics can be used as the matrix. Considering their low melting point and their

good mechanical properties, polyolefins are often used to produce natural fiber based composites, contrarily to

PS and PVC, which both exhibit a low impact resistance at ambient temperature, and, in the case of PS, a higher

melting point [6]. Polyethylene is one of the most produced and used thermoplastics. As most synthetic polymers,

it is derived from petroleum. It has several advantages such as low price, simple structure, broad availability and

ability to be reprocessed various times and recycled. It is a semi-crystalline thermoplastic with linear or branched

structures. It is used in several applications in a wide range of fields such as cables, transportation, construction,

packaging, leisure, sports, electronics, etc. Among all the different grades of polyethylene, HDPE presents

higher mechanical properties, as well as higher density (around 0.965 g/cm3). However, it has the main drawback

of degrading slowly. Due to its synthetic origin, its non-renewability, its slow degradation and its high volume of

applications, its recycling became a major issue [4, 7].

I.1.3 Natural fibers

Natural fibers can be of different types, origins and natures. However, vegetal fibers are the most widely used

for the production of thermoplastic based composites, mainly due to their availability and lower cost [1, 8, 9, 10].

Their worldwide production is estimated at about 4 billion tons [11]. They are generally classified in three main

categories, as seen in Table I-1. First, the vegetal fibers include the fibers extracted from plant stems (rattan,

flax, jute, hemp, ramie, wood, etc.), the fibers extracted from the seeds (mainly cotton and kapok) and the hard

fibers extracted from the stalks (wheat, rice, barley, grass, bamboo, etc.), the fruits (coco), or the leaves (agave,

banana, sisal, etc.) of the plants. Then, the animal fibers, which come from animal fur, fleece and secretions

(silk). Finally, there are mineral fibers such as asbestos and basalt [1, 8, 9, 12]

4

Table I-1: Classification of natural fibers according to their origin.

Type Origin Examples

vegetal

stems flax, jute, hemp, ramie, rattan, wood

seeds cotton, kapok

stalks wheat, rice, barley, grass, bamboo

leaves agave, banana, sisal

fruits coco

animal

fleece wool

fur alpaca

secretions silk

mineral basalt, asbestos

Vegetal fibers are bio-composites mainly composed of cellulose, hemicellulose and lignin, and at much lower

extent secondary components such as extracts, water, sugars, starches, pectins, proteins, and inorganic

compounds [6, 12]. Most vegetal fibers can be assimilated to a natural composite constituted of a lignin matrix

reinforced with cellulose fibrils and hemicellulose acting as a as coupling agent [13]. Each constituent has a

specific function. The cellulose is mainly responsible for the fiber rigidity, resistance and structural stability, as

well as for the polar aspect. The lignin ensures cohesion and other properties, while the hemicellulose improves

the compatibility between lignin and cellulose. Each natural fiber presents unique, but variable properties

depending on many factors such as the species, age, geographical origin, size and shape, environment and

climate [6]. They are generally hydrophilic and can contain up to 13% wt. of humidity [14]. They tend to

agglomerate as the hydroxyl groups create hydrogen bonds with other cellulose molecules and with water

molecules [12, 14].

Besides wood, flax is one of the most common fiber used in North America and Europe. In 2015-2016, Canada

produced about 940,000 tons of flax seeds and this amount should increase every year with the increasing

interest for flax reinforced composites [15]. It is broadly available and is strongly used in the automobile sector

to make various parts and boards [16]. Flax fibers are extracted from the stems of Linum usitatissimum, an

annual plant from the Linaceae family that can be found in Asia and Europe [17]. They present high tenacity and

thermal conduction, but they also have low elasticity and high absorption [11]. They have higher yield and are

twice to three times more resistant than cotton [1, 6]. Depending on the applications, they can be of different

sizes and shapes, ranging from long fibers to powder, and can also be used woven. Long fibers are used to

create canvases, whereas shorter are used to create ropes and papers. All sizes and shapes can be used to

5

produce flax fiber based composites, depending on the properties needed, and to fit different applications, from

panels to construction materials [18, 19].

I.1.4 Natural fibers vs. synthetic fibers

Natural fibers (wood, flax, hemp, jute, kenaf, coco, etc.) are gaining interest over synthetic reinforcement (glass,

carbon, aramid, talc, etc.) due to their numerous advantages such as low cost, lower density, high specific

stiffness and strength, availability, as well as electrical and acoustic insulation properties [6, 10, 20]. They are

non-toxic, non-corrosive and are much less abrasive for tooling and equipment. They can be easily used with

traditional processing methods. They also have a good notoriety due to their natural source, as well as their

renewable, sustainable and degradable aspects [20, 21]. They appear as an interesting alternative to more

traditional reinforcements despite of their lower mechanical properties. However, they generally have lower

resistance, lower durability, poorer fire resistance and higher quality variability than synthetic fibers. Their use

also induces some processing issues due to their hydrophilic nature and their sensitivity to temperature, leading

to low interface bonding, parasite foaming and void creation [21, 22, 23, 24]. Volatiles and water are released

during the processing of natural fiber composites, to an extent depending on their preliminary drying, and leading

to irregularities in the composites structure. Natural fibers also tend to degrade at temperature above 230°C

[24]. Despite all these limitations, natural fibers are widely used in the composite industries and many of their

issues can be overcome by careful selection of the matrix, the processing parameters and by preliminary drying.

In order to find applications and to replace inorganic fiber composites in some applications, natural fibers should

exhibit interesting and competitive properties. Table I-2 compares the main properties of natural fibers and

synthetics ones. Except for the carbon fiber, which has very high rigidity, it appears that, when considering

specific properties, some natural fibers have properties close to most of the other synthetic fibers, and especially

to glass fibers which are the most used synthetic fibers. Considering that these fibers generally have lower

density and cost than inorganic fibers, they thus appear as an interesting alternative to more traditional

reinforcements such as glass fibers [25].

6

Table I-2: Comparison between the physical and mechanical properties of common natural and synthetic fibers [26, 27, 28, 29].

Fiber type Density

(g/cm3)

Elongation at

break (%)

Tensile strength

(MPa)

Elastic modulus

(GPa)

Cotton 1.5-1.6 3.0-10.0 290-600 3-14

Jute 1.3-1.5 1.5-3.0 390-800 10-50

Flax 1.4-1.5 1.5-4.0 345-1500 25-90

Hemp 1.5 1.6-4.0 400-900 25-70

Kenaf 1.4 1.6-2.1 220-940 25-55

Ramie 1.5 2.0-3.8 290-700 45-130

Sisal 1.3-1.5 2-14 175-593 9-38

Coir 1.2 15-30 15.0-30.0 4-6

Softwood Kraft pulp 1.5 4.4 1000 40

Bamboo 1.4 2.0 500-740 30-50

Coconut 1.2 20-40 1200-1800 4-6

Spider silk 1.3 28-30 1300-2000 30

E-glass 2.5 0.5-3.0 1200-3500 70-75

S-glass 2.5 2.8 2000-4570 86

Aramid 1.4 3.3-3.7 3000-3150 63-67

Carbon 1.4 1.4-2.0 4000 230-240

I.1.5 Interface properties and coupling agent

Composite properties essentially depend on the materials capacity to distribute and transfer loads and stresses

between both constituents [23, 24]. This is governed by the interface quality. However, for most natural fiber

composites, interface quality is known to be poor due to the incompatibility and difference in polarity between

the fiber and the matrix. Most natural fibers such as flax are hydrophilic, while polyolefins such as high density

polyethylene are hydrophobic. This leads to a low interfacial adhesion and to low fiber wetting by the matrix due

to water absorbed at their surface. Due to the poor wetting of the fibers by the polymer, they tend to agglomerate,

leading to an inhomogeneous distribution and to low mechanical properties. Moreover, due to the poor interface

properties, stress transfer is limited and stresses are not distributed ideally in the composite [23, 30, 31]. To limit

this effect and improve the interface quality, three main solutions exist. First, a coupling agent can be used to

improve the interface adhesion and the compatibility between the constituents. Coupling agents are polymers

grafted with functional groups that can react with both the matrix and the fiber. They have characteristics and

7

properties of both constituents and create a chemical bridge between them, thus increasing the interfacial

adhesion. The coupling agent has to be chosen depending on the matrix and fiber nature and properties, as to

have the maximum compatibility with both of them. For example, for natural fiber reinforced polyethylene, maleic

anhydride grafted polyethylene (MAPE) is often used as a coupling agent because the polyethylene part

entangles with the matrix and the maleic anhydride groups react and bond with the hydroxyl groups on the fiber

surface. Then, it is also possible to treat the fiber to improve its adhesion or limit the water absorption. Depending

on the type of treatment, different effects can be obtained such as the cleaning of the fiber, the modification of

the fiber surface or of its chemistry. Two main types of treatment exist: physical and chemical. The chemical

treatments modify the surface chemistry to improve the compatibility with the matrix, while physical treatments

modify the surface structure and properties by physical, mechanical, thermal or electromagnetic effects, and

influence bonding with the matrix. However, the last ones are complex, often harmful for the environment, and

thus still uncommon [22, 23, 24, 32, 33, 34]. The use of one of these solutions, which is preliminary drying,

substantially decreases the harmful effect of natural fibers making them suitable for the production of

thermoplastics based composites.

I.1.6 Applications

Over the last years, the interest for natural fiber composites has grown strongly with the development of new

materials and the optimization of the ones already on the market. These composites are used in a great diversity

of fields for both structural and non-structural applications such as [16, 24, 35, 36, 37, 38, 39]:

• Construction (floor, roof, panels, railing)

• Transportation (car parts, train parts)

• Furniture (chairs, tables, decks)

• Leisure (toys, CD, phone and computer cases)

• Sport (kayaks, boat parts, snowboards)

• Music (guitars, violins, harmonicas)

• Decoration and home (jars, clocks, scales)

Some examples are presented in Figure I-1. Natural fiber composites are widely used in the automotive and

construction industries because of the cost and weight reduction generated by the use of natural fibers instead

of more traditional fibers and materials [26, 27, 38, 39, 40]. Lately, natural fiber use increased by 20% per year

8

in the automotive sector [41]. Outdoors and sports applications also increased strongly in the last years, but

remain limited due to the moisture sensibility of such composites. In order to find new applications and markets,

these composites constantly need to be competitive and performant compared to more traditional materials such

as inorganic fiber reinforced composites and wood. Besides their environmentally friendly image, they should

have improved properties and performances, as well as durability in order to increase their market share [6].

Figure I-1: Examples of potential and current commercial applications for natural fiber based composites [28, 42, 43, 44, 45, 46, 47, 48, 49, 50].

9

I.2 Recycling

I.2.1 Polymer recycling

Recycling is associated to the collect, sorting and treatment of waste to provide these materials a new lifecycle

by their direct use or their reintroduction into the processing cycle of other products. It is applied for various

materials such as plastics, metals, cardboard and paper, wood, solvents and textile. The recycling methods

depend on the material considered. The main idea is to decrease the waste volumes while simultaneously

preserving the natural resources by reusing materials and limiting the greenhouse gases emitted when burning

the materials. Even if the amount of petroleum used for the production of plastics is currently relatively low (less

than 10%), this number should increase with time as plastic demand and waste constantly increases, mainly

due to the actual life and consumer styles [51]. Currently, almost 80% of the polymers are petroleum based. It

is thus important to limit the waste and the consumption of new materials at best.

Plastic recycling takes place in consecutive steps. First, the materials should be collected, gathered and

transported to the sorting center to be sorted out by resin type. To facilitate the sorting of polymers, code

numbers have been attributed to the main plastics families used in everyday life. These codes are given in Table

I-3. Polyethylene and polyethylene terephthalate (PET) are often sorted out, due to their high volume, while

others are gathered. Industrial wastes are often easier to recycle than post-consumer ones due to their high

volume and low contamination [52]. Then, the sorted polymers are conveyed to the various treatment centers.

Various separation methods have been developed, but they remain expensive [7, 53]. However, they appear

unavoidable as the inhomogeneity of polymer waste leads to poor properties of the recycled materials. Three

main types of plastic recycling methods exist: mechanical, thermal (energetic) and chemical. The different types

of recycling, their characteristics and specificities will be detailed in the next chapter. It is also possible to reuse

directly some products after cleaning. However, for sanitary reasons, this method remains quite limited. Table

I-4 summarizes the plastic waste recovered in Quebec in 2008.

Table I-3: Plastics codes in Canada [52].

Plastic code Type of polymer

1 Polyethylene Terephthalate (PET)

2 High Density Polyethylene (HDPE)

3 Polyvinyl Chloride (PVC)

4 Low Density Polyethylene (LDPE)

5 Polypropylene (PP)

6 Polystyrene (PS)

7 All others plastics

10

Table I-4: Plastic waste recovered in Quebec during 2008 [52].

Polymer category Total (in metric tons)

PET 27 976

HDPE 35 687

PVC 9 516

LDPE 19 048

PP 5 954

PS 655

Mixed plastics (#1-#7) 3 288

Mixed plastics (#2-#7) 4 974

Mixed plastics (#3-#7) 1 662

Others 11 289

Total 120 050

I.2.2 Composite recycling

Natural fiber composites are widely used nowadays and production/waste volumes constantly increase.

Moreover, according to Faruk et al., the production volume of natural fiber reinforced composites should increase

from 2.33 million tons to about 3.45 million tons between 2013 and 2020 [6]. The recycling of thermoplastic

based composites is quite similar to the recycling of thermoplastics, even if some specifics and limitations appear

due to the presence of fibers. This will be reviewed in the next chapter with a focus on the mechanical recycling

of thermoplastic based composites and will not be detailed further here.

I.3 Thesis objective and organization

Due to the very high volume of thermoplastic based composites used, their recycling becomes a major issue.

Although several studies have been done on the mechanical recycling of natural fiber reinforced thermoplastics,

there are still lacks of knowledge concerning the effect of such recycling on the performances and properties of

the final material, as most works reported contradictory trends. Moreover, there is even less information on the

long-term and intensive aspects of such recycling, as most of the studies considered less than ten processing

cycles. Thus, the main objective of this thesis is to study the recyclability of HDPE/flax composites during long-

term intensive (up to 50 cycles) mechanical recycling. Two composite series were realized, with and without

coupling agent, and the HDPE matrix alone was also studied to complete the knowledge on the intensive

mechanical recycling of thermoplastics. A complete characterization (in terms of physical, thermal, molecular,

11

morphological and mechanical properties) of the recycled samples was performed to evaluate the degradation

and the loss of performance undergone after long-term mechanical recycling.

To achieve the main objective, this thesis tries to understand the degradation undergone by the materials during

their long-term recycling through the following secondary objectives:

• Link all results to understand and quantify the degradation mechanisms taking place in the material and

its constituents during recycling.

• Comparison between the short- and long-term recycling behaviors.

• Comparison between the neat matrix and the composite behaviors and properties as to understand the

effect of fibers on the recycling behavior.

• Evaluation of the effect of coupling agent on the recycling behavior and the degradation undergone

during recycling.

• Evaluation of the recycling potential of both the neat matrix and the composites after 50 close-loop

reprocessing cycles.

Since the thesis is a paper-based document, it is composed of five main chapters:

In the first chapter, a brief introduction on natural fiber composites, their properties, their applications, their issues

and their recycling was presented. General notions and definitions on composites and their constituents were

also reported. The materials used in the following parts (high density polyethylene and flax fiber) were also

described briefly, with their main characteristics and applications. Finally, the context of this work was also

explained and general ideas on recycling were also presented.

The second chapter presents a literature review on the mechanical recycling of thermoplastic composites. The

various types of composites recycling are briefly introduced. Then, the various works conducted on the recycling

of natural fiber reinforced thermoplastic composites are presented, followed by the works on the mechanical

recycling of composites reinforced with inorganic fibers. Finally, the various limitations to the composites

mechanical recycling are considered and some possible solutions to these limitations are suggested. This review

confirms the lack of knowledge on the long-term mechanical recycling of thermoplastic polymers and

composites.

The third chapter is the first step to understand the behavior of the thermoplastic materials towards mechanical

recycling. In this part, the HDPE neat matrix was produced and recycled by extrusion up to 50 cycles. The effect

12

of intensive mechanical recycling on the polymer structure and properties (physical, thermal, molecular and

mechanical) is reported, thus giving information on the matrix behavior during recycling.

The fourth chapter is the second step to understand the thermoplastic composite behavior towards mechanical

recycling. It focuses on the intensive mechanical recycling of HDPE/flax composites, considering composites

with and without MAPE as a coupling agent. A complete characterization (physical, thermal, morphological,

molecular and mechanical) is also conducted and the processing parameters are chosen to be the same as for

the neat matrix study in the third chapter. For each composite formulation, a total of 50 closed-loop reprocessing

cycles are carried out and compared to the neat matrix to understand the effect of recycling on the composite

properties, as well as to evaluate the effect of fiber and coupling agent on the recycling behavior.

Finally, the fifth chapter is an overall conclusion briefly reviewing the main results, observations and conclusions

of the different chapters. It also presents several recommendations for future works.

13

Chapter II. Mechanical recycling of thermoplastic composites

Résumé

Les composites thermoplastiques ont progressivement fait leur place sur le marché ces dernières années. Leur

principal avantage, en comparaison avec les composites thermodurcissables, est leur possibilité d’être mis en

forme de manière répétée par de simples procédés de fusion et moulage. Ceci permet aux matériaux d’être

récupérés et remis en œuvre s’ils sont recalés au contrôle qualité (post-industriel) ou tout simplement à la fin de

leur cycle de vie (post-consommation). Cette caractéristique apparaît très intéressante considérant que le

recyclage est un des aspects primordiaux du développement durable, d’autant plus que la production et

l’utilisation de composites thermoplastiques augmentent constamment. Ces composites se retrouvent en effet

pour toutes sortes d’applications telles que l’emballage, l’automobile, l’aéronautique, l’ameublement, la

construction et le bâtiment, ou encore les sports et loisirs. La demande constamment croissante pour ce type

de produits engendre un très grand nombre de pièces produites, utilisées puis jetées. En raison de cet important

volume de composites produits et utilisés, de nouvelles lois et directives toujours plus strictes sont constamment

développées en faveur du développement durable et du recyclage. C’est pourquoi il est important d’étudier et

de comprendre l’état de tels matériaux en fin de vie, ainsi que de développer des applications et des méthodes

pour réintroduire ces grandes quantités de matières dans les lignes de production. Afin de limiter les quantités

de matière consommées et mises en décharge, les déchets de composites thermoplastiques doivent être

considérés comme une source de matières premières pour la production de nouveaux produits composites par

recyclage et par remise en œuvre. Toutefois, afin d’obtenir de bonnes propriétés finales, il est essentiel de

comprendre les mécanismes de dégradation et le comportement des matériaux lors leur cycle de vie et leur

recyclage. Ce chapitre est une revue de littérature sur les différentes possibilités de recyclage de composites

thermoplastiques, et notamment sur les techniques de recyclage mécaniques disponibles. Quelques chiffres sur

la production et le recyclage de composites seront présentés et discutés, puis une brève présentation des

différentes méthodes de recyclage disponibles pour les composites thermoplastiques, de leurs avantages et

limitations sera faite. Finalement, une vue d’ensemble des différents travaux sur le recyclage mécanique de

composites thermoplastiques à renforts organiques et inorganiques est présentée. Dans tous les cas, le

comportement du matériau face au recyclage et les propriétés résultantes sont considérés, ainsi que les

solutions développées pour améliorer les performances et la qualité des matériaux recyclés.

Mots-clés: Thermoplastiques, composites, recyclage, propriétés.

14

Abstract

Thermoplastic composites have found several applications over the years, but their main advantage when

compared with thermoset composites is their possibility to be reshaped after processing through further melting

and remolding processes. This allows the materials to be recovered and reprocessed if they fail quality control

(post-industrial) or after their end of life (post-consumer). This is very attractive as recycling is one of the main

principles of sustainable development. This is even more appropriate as thermoplastic composites are more and

more produced and consumed in a wide range of applications such as packaging, automotive and aeronautics,

furniture, building and construction, as well as sport and leisure goods. Thus increasing demands leads to a high

number of parts being produced and discarded. With this high volume of composites production and use, the

emergence of constantly new and stronger policies and laws towards sustainable development and recycling

were developed. This is why it becomes important to study and understand the conditions of the materials at

their end of life and develop applications like recycling to reintroduce these high amounts of materials into

production lines. To limit both the amount of material consumed and landfilled, thermoplastic composite waste

can be considered as a source of raw material for the manufacture of products through recycling and

reprocessing. However, to obtain good final properties, it is essential to understand the degradation processes

and the materials behavior during their life cycle and their recycling. In this chapter, a review of the different

possibility to recycle thermoplastics composites is presented, with a focus on the mechanical recycling

techniques available. Some figures about composites production and recycling are presented and discussed,

followed by a brief presentation of the different recycling methods available for thermoplastic composites with

their advantages and limitations. Then, an overview of mechanical recycling is made considering both organic

and inorganic fillers. For all cases, the material behavior towards recycling and the resulting properties are

considered, as well as the solutions developed to improve the performance and quality of the recycled materials.

Keywords: Thermoplastics, composites, recycling, foams, properties.

N. Benoit, R. González-Núñez, and D. Rodrigue. Mechanical recycling of thermoplastic composites, in

Thermoplastic Composites: Emerging Technology, Uses and Prospect. E. Ritter Ed., Nova Science Publishers,

New York, Chapter 3, pp. 95-142, ISBN: 978-1-53610-727-2 (2017).

15

II.1 Introduction

Composite materials are composed of two phases, one being the matrix, and the other one, the reinforcement

(fiber or particles). The matrix protects the reinforcement, provides cohesion to the material and binds the

reinforcing particles together, thus distributing the stresses under load. The reinforcement increases the rigidity

and the performances of the material. The association of both phases leads to interesting properties, combining

the properties of both components. The performances of the resulting composites depend on the capacity of the

material to transfer the stresses between both constituents, and thus on the interface quality. These materials

have been produced continuously for a few decades now. The high demand for composites created a new

category of plastic-based waste that has to be taken care of. Nowadays, 8.7 million tons of polymer composites

are produced each year, thus generating high volumes of waste [54]. For example, in France, 30,000 tons of

composites waste are collected every year. As they will reach their end of life soon and the quantities will

increase in the future, their recycling should be studied and understood as to recover the materials and their

value, as well as simultaneously avoid their accumulation in landfill [55]. Recycling of polymer based materials

is environmentally very interesting as it leads to the reduction of waste simultaneously with saving of virgin

resources, and especially petroleum [55]. However, due to the heterogeneous nature between the matrix and

the reinforcement, and to the lowering of the properties of the recycled materials, thermoplastic composites are

currently rarely recycled at the industrial scale, as the existing techniques are often expensive [55, 56, 57]. Thus,

most composites are currently converted to energy and fuel or fiber recovery. However, in the latter, the fibers

recovered show an important drop of performance (up to 90% for glass fibers) and suffer from a lack of cost

competitiveness, which usually make them unsuitable for most applications and markets [58, 59].

Thermoplastic composites are widely used nowadays. Their matrices are characterized by their aptitude to be

repeatedly melted by heating and solidified by cooling, and this in a reversible way. The reinforcement add

rigidity and strength to the matrix, thus making a material combining the properties of both constituents.

Thermoplastic composites have become an area of increased interest due to their relatively low cost, wide range

of mechanical properties, toughness, problem free storage, resistance to chemical attack, ease of processability,

and above all their better recyclability. This is related to their fundamental ability to be reprocessed and reshaped

when reheated leading to much simpler and direct recycling paths [54, 56, 60, 61, 62]. Because they have a

melting point, they can thus be remelted, reshaped and reharden multiple times. This is not the case for

thermosets having highly reticulated networks which cannot be melted after their first shaping [63, 64]. Although

thermosets are currently the most common composites representing about 90% of the composite market in

2014, thermoplastic composites are continuously increasing their market share due to their numerous benefits

[54, 56, 62]. Over the last decades, the demand for thermoplastic composites increased with the development

of new applications in several fields such as decking, outdoors equipment, household, leisure, construction,

16

transportation, packaging, sports, leisure, domestic applications, energy, automotive and aerospace [7, 54, 56,

62, 65]. This increased demand on composites produced and used led to an increase of waste generated, thus

increasing the pressure for the development of efficient and viable recycling methods for such materials [7, 32,

66]. It was projected that in the next few years, the total global production of composites will exceed 10 million

tons, thus occupying a volume of over 5 million cubic meters [59]. Among these composites, more than 90% are

glass reinforced (GR) composites, and about 40% of this volume is associated to thermoplastics composites.

With constantly new environmental directives and legislations being established, both producers and

manufacturers are asked to consider the environmental impact of their products. This is particularly true for

thermoplastics composites, which are gaining interest over thermosetting composites due to their several

advantages like design flexibility, easier processing and shorter processing time [32, 63, 64, 66]. Moreover, most

of them usually decompose slowly, thus requiring to be taken care of at the end of their life cycle.

In the last few years, the society has constantly increased its concern about sustainable development leading

the actual governments to consider new laws and directives towards resource management and waste

revalorization [67]. The plastic production continuously increase to fit the increasing demand for such materials.

Around the world, the general plastic production increased by 38% in ten years, reaching 311 million tons in

2014, with an important part of thermoplastics materials [4]. In Europe, thermoplastics materials represent more

than 72% of the total production [4, 67]. Thus, polymers represent an important part of the municipal solid waste

(MSW) in many countries such as the United States, Canada and in Europe. For example, they represent up to

20% of the MSW in the United States, and only a few percent were recycled [32, 68]. Moreover, thermoplastics

are generally the main constituents of the polymer fraction of MSW [69]. In Europe, 7.7 mT of plastic wastes

were recovered in 2015. Among this waste, only 30% of the total volume was recycled, 39% was energetically

revalorized, while the remaining 31% was simply landfilled [4]. In Canada, in the province of Quebec, plastics

waste constantly increased over the last 20 years. Their revalorization only accounts for about 5% of the total

revalorization for all materials, and represents only 16% of the total plastics waste collected from MSW [52].

These low recycling rates are essentially due to the lack of techniques available and problems associated with

the collect and sorting processes. This shows that more efforts must be done in these fields to increase the

revalorization rate of such materials. However, it should be noted that, over the last few years, both energy

recovery and recycling increased, while landfilling rate decreased [4]. For example, in Europe, between 2006

and 2014, for the 25 million tons of the global plastic waste produced, the landfilled fraction decreased by 38%,

while the energy recovery and recycling shares increased by 46% and 64% respectively, thus reaching 8 million

tons of polymers discarded to landfill, 10.2 million tons of polymers energetically recovered, and 7.7 million tons

of polymer waste recycled [4]. Thus, recycling of thermoplastics and their derivatives is fundamental in the actual

social and political context, especially since they have a slow decomposition rate, as well as a low density, but

represent very important volumes of materials [68]. Considering these facts, and especially the important volume

17

generated daily, some efforts still need to be done to manage resources, as well as to manage, treat and

revalorize this kind of waste. However, if the waste volumes and recycling rate of thermoplastics are usually

known, very few statistics can be obtained about thermoplastics composites [69]. Although some companies are

already using recycled thermoplastics to produce wood plastic composites (WPC), no use of recycled

composites was reported. In any case, the composition of the waste stream depends on the geographical

location, time of the year (season), current trends and several other parameters [69].

Recycling consists in the collect, the sorting and the treatment of waste. Materials sorting is an important issue

as the presence of impurities and contaminants can lead to unsuitable or poor properties of the recycled

materials. Three main approaches can be considered for the recycling of composite waste [70, 71]. The simplest

way is to dispose into landfill. However, this is also the worse scenario from an environmental and economical

point of view, as it can lead to space problems if the volume of waste is too important [57, 66]. The second option

is the incineration of the material, with possible energy recovery or heat generated, but this also lead to air

pollution and environmental problems. This option is also known as thermal recycling and is often considered

as a part of the third category, which is recycling. The last, but most important option, is recycling. It is based on

the partial or total recovery of the constituents of the global material. It is sustainable, economically viable and

appears to be the best environmental and technical choice [68, 72, 73]. It reduces the energy used and the

waste volume disposed to landfill, while simultaneously reducing the raw material consumption. However, it

induces to some extent some changes in the mechanical, physical and chemical properties of the final materials,

that may affect the material processing conditions and the quality of the end products [56, 66, 68, 74]. It often

gives, not always justified, a “low quality” image to the products, thus limiting its attractiveness. Compared to

virgin composites, recycled composites often meet problems to satisfy the quality and economical requirements,

thus limiting the choice of suitable market [57, 68]. Several works considered the partial addition of virgin

materials to counter this drawback. Coupling agents are also often used to improve the composites properties,

but their sensitivity towards recycling is not well known [55]. It is thus essential to know how the properties of

thermoplastics composites are modified by recycling, and especially how the reinforcement modify polymer

degradation [73]. This should be done by the study of the recycling effect on the materials properties, by the

development of new, cheaper and more efficient recycling methods, as well as better separation techniques. But

it should also be done by increasing the recyclability of the produced composites. Besides thermal recycling,

two other categories of recycling methods exist: chemical recycling and mechanical recycling [55, 56, 67, 73].

Chemical recycling leads to reusing the constituents for the production of new composites, while mechanical

recycling is often performed on the whole composite.

The recycling of thermoplastics is well known and was the focus of several authors. The extensive long-term

recycling of such polymers was studied by Oblak et al. [74], Jin et al. [36, 37] and Benoit et al. [75]. However,

18

the recycling of thermoplastics composites, and especially of their extensive or long-term recycling, is much

more limited. This chapter aims at reviewing the various works on the mechanical recycling of thermoplastics

composites to present the state of the art in this field. To do so, the main recycling methods for this type of

composites are presented first with their advantages and drawbacks. Then, the recycling of natural organic fillers

reinforced composites will be considered, followed by the study of the recycling of inorganic fillers reinforced

composites. Finally, the main limitations to the recycling of thermoplastics composites will be presented, as well

as their possible solution. Finally, this review will summarize and conclude about the actual state and viability of

the mechanical recycling of thermoplastics composites.

II.2 Recycling methods

Due to the lack of knowledge, the complexity of recycling and the predominance of thermosets until recently, the

majority of composites materials were disposed to landfill [76]. But the same situation is still going on. However,

due to the constantly increasing market of composites materials and the important volumes of waste generated,

as well as the high cost of the reinforcements and additives used in these materials, other alternatives must be

considered [64, 77]. Nowadays, due to their numerous advantages, thermoplastics composites are gaining

interest and gradually taking the place of thermosets composites in various applications, making their recycling

easier [77]. They present lower manufacturing time, lower sensitivity to impact damage, easier manufacturing,

and better recyclability due to their aptitude to be remelted [77]. These advantages allow to consider other ways

of managing composite waste.

Composting can be conducted for biodegradable materials, but in this case, both matrix and reinforcement

should be biodegradable to be sure not to contaminate the environment with permanent residues [78]. For very

specific cases, reuse of materials after cleaning or basic operations can be considered. If this method is

environmentally respectful, however, in the polymer and composite fields, reuse is quite limited due to health

and safety reasons [64, 67]. Because of their inhomogeneous nature, recycling is a great challenge for polymer

composites. The presence of reinforcement leads to technical challenges such as equipment wear,

incompatibility, restrictions on temperature and other processing parameters, special attention during processing

or other technical constraints. The waste stream has to be consistent and sufficient (in quality and quantity), and

the price low enough to make the recycling process viable and profitable [52, 78].

Various recycling methods exist for polymer composite materials. These methods can be divided in three main

categories: thermal recycling, chemical (or feedstock) recycling and mechanical (or physical) recycling [52, 56,

58, 67, 76, 77, 78, 79, 80, 81]. Each one has its own list of advantages and drawbacks [78, 82]. The choice of

the method depends on several parameters including the nature and quality of materials, their initial state, the

final state and quality desired for the materials, as well as their constituents, the nature of the composite

19

constituents, the desired post-recycling application combined with economic, technical and legislative

considerations [64]. The recycled composites need to be of good quality and relatively low price to be competitive

with virgin materials and thus place themselves in the current markets [79]. A complete overview on recycling

issues and technologies can be found in Henshaw et al. [79] and Goodship [83]. For polymer and composite

waste, regardless of the recycling method, two waste categories can be differentiated: industrial production

waste (from the production rejects and waste which has not been contaminated or used) and post-consumption

waste (subjected to contamination and degradation during its lifecycle) [67, 68]. The latter generally requires

some additional cleaning and sorting steps as to limit contamination, separate the metal, paper, adhesive and

other foreign parts and, if applicable, to sort the composites by polymer matrix and fiber types [58, 67].

II.2.1 Thermal recycling

In thermal recycling, the matrix and the fibers are separated by decomposition of the polymer at high

temperatures (between 300 and 1000°C). This can be done through various thermal processes such as

pyrolysis, gasification, fluidized-bed combustion (FBC), rotary kiln, fluidized bed and mass burn [79, 81, 82, 84].

All these processes are based on the decomposition of the product under external heating and its transformation

into energy or liquid fuels [67, 76]. The energy, heat or steam generated during the process is recovered.

Regardless of the reinforcement type, some degradation takes place to a varying extent during thermal

processing [56]. The heat generated during the process can also be transformed into electricity [81]. In pyrolysis,

organic materials are decomposed at high temperature, generally in the absence of oxygen [79]. In the fluidized-

bed combustion, the materials are decomposed under high temperatures (450-500°C) in a silica sand bed

thermally heated and fluidized by hot air. The organic part of the composite is volatilized by the hot air and

transported through the air stream before separation. The volatilized polymer can then be burned to produce

heat and energy [81]. In the case of natural fiber composites and other low degradation temperature fibers, this

process cannot be used to recover fibers due to the high sensitivity of natural fibers towards high temperatures,

leading to important degradation while submitted to high temperatures. Waste plastics and common

reinforcements have high calorific values, thus making them interesting for energy recovery. Moreover, this type

of recycling has also been shown to significantly reduce the volume occupied by this type of waste, but it also

has been noticed that this recycling method can lead to the emission of pollutants and noxious emanations, thus

limiting its applications and promoting the recyclers to find other ways of recycling these materials [64]. However,

it is efficient and viable when no other method can be applied [78]. It is especially suitable for mixed or

contaminated waste, as well as for composites with insufficient properties after extensive recycling or severe

weathering [56, 58, 68, 73, 79, 81].

20

II.2.2 Chemical recycling

Chemical recycling is the chemical de-polymerization and dissolution of the matrix by organic or inorganic

solvents through various advanced processes such as liquid-gas hydrogenation, gasification, viscosity breaking,

steam treatment, catalytic cracking, pyrolysis, glycolysis, alcoholysis, solvolysis and hydrolysis [64, 77, 81]. For

this type of recycling, both the fibers and the matrix (as polymers, monomers, fuels or chemicals products) can

often be recovered [77]. The polymer matrix is degraded into its monomers or basic chemical products

(oil/hydrocarbon), while the fibers are filtered to be recovered [7, 64, 67, 77, 81, 84]. The monomers and chemical

products can then be directly reused as raw materials to create new polymers or as fuels and composites

materials. This method is complex, but suitable for materials with contaminants or when different materials are

mixed to some extent [58, 64, 78]. In hydrolysis, the depolymerization is conducted with water and catalysts [79].

This method achieves high product yield and decreases the amount of waste [64]. However, this type of recycling

is quite controversial as it uses chemicals, such as solvents, bases, acids and washing fluids, and thus leads to

the production of important amounts of waste chemicals, thus in contradiction with the benefits of recycling. It is

also generally expensive, due to the cost of the chemicals used and the high energy consumption with low

flexibility [77, 78]. These two points reduce strongly its attractiveness, especially for thermoplastics [52, 67, 68,

73, 79, 81]. Moreover, before degradation, the material is often ground to increase its surface area, thus leading

to better diffusion processes, but also to some mechanical degradation in the fiber and matrix, but at a much

lower extent than for thermal and mechanical recycling [68].

II.2.3 Mechanical recycling

Mechanical (physical) recycling is the recovery of the composite’s material through various mechanical

processes including grinding/pelletizing, reprocessing and shaping [7, 56, 58, 64, 67, 68, 73, 77, 78, 84]. In the

case of thermoplastics, after the usual cleaning and sorting steps, the materials can directly be pelletized,

remelted and reshaped through extrusion, injection molding or compression molding. They can be processed in

any shape and are suitable for a wide range of applications [64, 76, 79, 81]. They usually have a fairly good

homogeneity [79]. On the contrary, for thermosets, direct remelting cannot be done and the ground materials

are often used as a new type of reinforcement [76, 79, 81]. This requires very high pressure and simple shapes.

The resulting materials can then be used for limited applications [79]. However, the high pressures,

temperatures, and shear stresses undergone during these three steps lead to irreversible changes in the

polymer structure and properties, as well as in the fibers [56, 67, 68, 73, 79]. This reduction of properties and

performances of the recycled materials is also reinforced by the loss of properties undergone by the material

during its lifecycle. This loss is due to their exposure to various long-term threats such as temperature, light

(especially UV), moisture and global use wear, and cannot be compensated during the reprocessing steps

leading to lower properties than virgin materials [67]. For demanding applications, partial addition of virgin

21

material (polymers and/or fibers) can be considered [79]. The heterogeneity and diversity of the matrix is another

factor substantially influencing the properties of the materials mechanically recycled.

Recycling of industrial production waste is called closed-loop recycling, while recycling of post-consumption

products is called open-loop recycling [68]. The former is obviously better for mechanical recycling due to its low

level of contamination and its homogeneity, but both are suitable for mechanical recycling. In mechanical

recycling, the purity of the waste stream is very important [64]. In the case of mixed waste recycling, due to the

incompatibility and immiscibility between most polymers, the properties of the recycled material are generally

lower than for separated homogeneous waste recycling [67, 79]. The more complex and contaminated the

composite waste is, the harder it is to recycle and to obtain interesting properties. Thus, in order to maximize

the properties and performances of the recycled composites, the waste stream should contain as few types of

polymer matrix as possible, and very few contaminants. Therefore, sorting is a critical step in the mechanical

recycling process. Nowadays, different separation techniques exist for plastics such as density segregation,

selective dissolution/precipitation, manual sorting, triboelectric, mid infrared (MIR), and near infrared (NIR)

techniques [64]. However, the presence of reinforcement in the composites make them difficult to apply, thus

adding another technical challenge. Some methods based on pulverization have been developed to limit the

negative effect of polymer mixed waste, but are not applicable to composites, leading to severe degradation of

the reinforcement [67]. Moreover, even for polymers, a complete separation is rarely conducted, due to its high

complexity to implement and perform [64].

II.2.4 Conclusion

All the methods presented above are summarized in Table II-1 for a better comparison and overview of each

process.

Both incineration and chemical recycling lead to the loss of the value added during polymerization and composite

manufacturing. Considering the current prices of polymers and composites, this appears as an important

drawback [30]. From an industrial point of view, mechanical recycling is very interesting as it uses conventional

processing methods while simultaneously limits the amount of virgin material and energy used while reducing

the emissions of greenhouse gases. It is also flexible, simple and inexpensive [64, 68, 84]. However, several

efforts still need to be done to improve the materials properties, as well as to develop more efficient recycling

and sorting methods [7, 68]. Some efforts should also be made to integrate the notions of recycling, dismounting,

separation and identification of the composite materials during the conception/design phase to make these steps

easier at their end-of-life [76, 84]. In various works, it was reported that both mechanical recycling and chemical

recycling are the most widely practiced methods [7, 67]. However, the mechanical recycling of thermoplastics

composites seems to be the most interesting technique in terms of sustainable development, as well as from an

22

industrial point of view, due to its low cost, easiness of application and reliability [7, 67, 78]. The technologies

and infrastructure already exist for the production and recycling of plastics, as well as for the production of

composites. They are widely available and can be easily adapted for composite recycling with limited efforts

[67]. This chapter will thus focus on the mechanical recycling of thermoplastics composites next.

Table II-1: Recycling methods for thermoplastics composites and their characteristics.

Type of

recycling Based on Examples Advantages Drawbacks

Products

obtained

Thermal

recycling

Decomposition

at high

temperatures

Pyrolysis,

gasification, fluidized-

bed combustion

Volume reduction,

ideal for mixed

fractions and poor

quality waste

Thermal

degradation,

gases and

pollutant

emanations

Energy,

heat, steam,

liquid fuels

Chemical

recycling

De-

polymerization

in solvents

Liquid-gas

hydrogenation

gasification, viscosity

breaking, steam

treatment, catalytic

cracking, pyrolysis,

glycolysis,

alcoholysis,

solvolysis, hydrolysis

High product yield

Production of

waste

chemicals,

expensive, low

flexibility,

degradation,

high energy

consumption

Fibers,

polymers,

monomers,

fuels,

chemicals

products

Mechanical

recycling

Reprocessing of

the composites

Extrusion, injection

molding, compression

molding

Flexible (shape

and process),

suitable for

several

applications, few

steps

Thermo-

mechanical

degradation,

requires initial

materials

separation

Composites

II.3 Recycling of thermoplastic composites reinforc ed with organic fillers

Due to an increasing awareness for sustainable development and recycling, organic fillers are strongly gaining

interest and consideration in the thermoplastic composite field. New environmental policies are constantly

23

implemented by more and more countries, thus increasing the pressure on manufacturers to consider the

environmental impact of their products [30]. Therefore, the interest of using natural fibers as a reinforcement in

thermoplastics composites is constantly growing. Moreover, most natural fiber reinforced composites are

produced with thermoplastics matrices. Due to their numerous advantages compared to more traditional fillers,

such as their renewable origin, biodegradability, non-abrasiveness, low production and manufacturing cost, low

density and high specific properties, they are present in a wide range of fields [40, 55, 72, 85, 86, 87].

Pervaiz and Sain [87] compared the energy consumption of glass and natural fibers and found that using vegetal

fibers instead of glass fibers could save 60% of energy per ton of product. Natural fibers are indeed often

biodegradable and produce less pollution because of their ability to burn without noxious gases or solid residues

[72, 86]. Compared to glass and carbon fibers, they also reduce respiratory and physical irritations during

handling [88]. Thus, they are progressively replacing inorganic fillers in some applications from common non-

structural applications to some specific structural applications [30, 72, 86, 88]. This is especially the case in the

automotive industry where, due to directives towards the reduction of energy consumption and the requirement

of lightweight materials, industrials are trying to progressively replace traditional materials by others with lower

cost and environmental impact [30, 72, 78, 86]. With this consideration, mechanical recycling of such composites

allows significant waste reduction, while conserving both materials and energy at the same time [89]. However,

they also present several disadvantages limiting their expansion such as poor resistance to moisture, sensitivity

to high temperatures, tendency to aggregate and incompatibility with hydrophobic materials [30, 88]. It is well

know that, due to their sensitivity to high temperatures, natural fibers are more prone to thermal degradation

during processing than neat thermoplastics, and that poor matrix/fiber interface quality leads to low stress

transfer resulting in lower mechanical properties.

II.3.1 Wood

Twite Kabamba et al. studied the mechanical recycling of composites made of low density polyethylene (LDPE)

and yellow birch fibers [40]. In their work, they performed up to 10 reprocessing cycles by extrusion under

constant conditions to simulate the closed-loop recycling of these composites with a reinforcement content of

15 wt.%. For each generation, thermal, molecular, rheological and morphological properties were studied, with

the final aim of determining the effect of wood fiber on polymer degradation and on the recycling process. The

fiber length significantly decreased with the number of cycles, reaching about the third of its initial value after 10

reprocessing cycles, while fiber diameter was not significantly affected. However, this decrease was not linear

and most fiber degradation took place during the first generation, while the value leveled off after the 6th

generation. The molecular weight distribution of the polymer showed that the number average molecular weight

Mn of the polymer matrix was more affected than the mass average molecular weight Mw, with respective

decreases of 30% and 4% between the first and the tenth regenerations, indicating that chain scission mainly

24

occurred in the longest molecules. As the fibers and the polymer chains broke, the composites crystallinity

increased with increasing number of generations. In the rheological tests, the zero shear viscosity decreased by

30% between the first and the tenth generations, while in elongational rheology the strain at break did not

change, but the maximum stress decreased by about 46% indicating that modifications in the polymer chains

essentially occurred in chain entanglement rather than in the main backbone. The comparison between the

polymer and the composite rheological behavior towards recycling indicated that the matrix behavior changed

from strain hardening to strain softening for the composite. The other parameters did not display significant

differences between each generation. Thus, the results showed that the effect of recycling is mainly affecting

the fiber size, the matrix molecular mass and crystallinity, as well as the shear and elongation viscosities.

However, the results also indicated that the degradation undergone was higher for the composites than for the

same polymer without reinforcement because the presence of wood fibers in the polymer matrix increased the

mechanical stresses (higher flow resistance) leading to more chain scission and the production of micro-radicals

during processing. The presence of fibers also limits chains mobility and crystallite growth leading to lower

properties after recycling. However, this difference seems to decrease with regenerations, probably due to fiber

length decrease.

Wang et al. studied the effect of processing parameters on the properties of wood/polyethylene composites

(70/30) with coupling agent [31]. In comparison with the previous work, they showed that composites processing

with a coupling agent limited the decrease of mechanical properties of the composites if suitable processing

parameters are selected [31, 40]. However, this work only considered one processing cycle. Thus, it gave some

indications about the recycling behavior of such composites considering that most authors noted that the first

cycle is the most critical one, but the study appears to be incomplete as some recycling effects could be occurring

or be significant for a higher number of regenerations [30, 40]. It should be extended to take into consideration

the long-term effect of recycling, as well as other types of materials properties. However, the results of this work

gave some directives and recommendations that should be considered and applied when recycling

wood/thermoplastics composites. The authors indicated that the final properties of the composites is a

combination of the complex interactions between the constituents’ nature, the equipment characteristics, the

processing conditions and the interfacial properties. On the one hand, high temperatures or stresses led to a

higher level of fiber degradation and to lower properties. On the other hand, it also increased the composite

mixing, homogeneity and fiber distribution, which led to better properties. Thus, the screw speed and the extruder

temperature profile should not be too high as to limit fiber degradation, but should be high enough to allow a

smooth processing. The position of the reinforcement feeder appeared to be important to limit mechanical and

thermal degradation of the materials. Thus, an appropriate choice and combination of parameters, including

screw configuration and rotational speed, throughput rate, barrel temperature profile and feeder position is

essential to limit the mechanical and thermal degradation (color change) of the wood fibers during processing.

25

Petchwattana et al. studied the closed-loop recycling of WPC partially made of waste poly(vinyl chloride) (PVC)

[70]. It is the only work considering the mechanical recycling of partial non virgin composites. The first part of

their work focused on the determination of the best ratio between virgin and waste PVC for such composites. A

content of 30% of industrial waste blended with 70% of virgin materials was found to be the best formulation,

while the wood flour content was set at 20%. In the second part of the work, the composites produced were

subjected up to seven reprocessing cycles. The results indicated that these composites could be recycled up to

seven times without any critical loss of mechanical properties. As the two previous studies [31, 40], the results

showed a molecular weight decrease due to chain scission induced by the shear stresses generated in the

material during processing. At the end of the seven reprocessing cycles, Mn decreased by 35% while Mw

decreased by 30%. Moreover, the flexural strength and the flexural modulus decreased by 13% and 23%

respectively, after the second cycle. Then, no significant change was observed for the following cycles until the

seventh reprocessing cycle where the properties started to decrease again. The results also showed no

significant effect of recycling on the impact properties. The authors concluded that, besides the first drop, these

composites could be mechanically recycled without critically affecting their properties and performances.

The effect of the initial wood particle size and of the addition of coupling agent on the degradation of mechanically

recycled composites was reported by Soccalingame et al. [90]. The mechanical recycling was simulated by

multiple grinding and injection molding cycles. The polypropylene/spruce wood flour composites was previously

compounded on a twin-screw extruder to ensure homogeneity. Seven successive reprocessing cycles were then

conducted. The coupling agent used was polypropylene grafted with maleic anhydride (MAPP) and the wood

content was set to 20 wt.%. Two particle sizes were used. The first one, called G1, was mainly composed of fine

particles of different sizes, while the second one (G2) was made of coarse particles added with fine ones. Several

properties of the composites and their constituents were considered: morphological, molecular, mechanical,

rheological and thermal. Once again, the results showed that the mechanical properties did not substantially

decreased after the seven reprocessing cycles despite degradation of both constituents. A very good mechanical

stability was observed during the first five cycles, then a slow but limited loss of properties was observed. During

reprocessing, chain scission and fiber breakup were reported, but they did not seem to significantly modify the

mechanical properties of the recycled composites. The strength at break remained stable over the seven cycles,

while the elongation at break gradually increased. The authors also observed the presence of two simultaneous

phenomena. During the first five cycles, a slight decrease of Young’s modulus and strength at break was

observed, indicating chain scission. However, for the last two cycles, both increased suggesting that a

crosslinking mechanism became dominant. Newtonian viscosities also decreased progressively with recycling

due to chain scission and thermo-oxidation of the matrix induced by the high shear stresses generated during

processing. Similar to other studies, fiber dispersion in the matrix was improved during the first cycles, while

fiber breakup essentially occurred during the first cycle and for larger wood particles [70, 31, 40]. Moreover, the

26

fiber length was the most affected parameter by breakup, while fiber diameter only slightly decreased with

recycling. The number average molecular weight decreased by 34% for the larger fibers and by 16% for the

smaller ones showing an effect of the fiber dimensions on the thermo-mechanical degradation sustained by the

polymer matrix. This was attributed to the synergistic effect of wood fibers on the shear stresses generated in

the polymer matrix during processing. On the contrary, the weight average molecular weight remained stable for

both fiber sizes for all the generations considered indicating that chain scission mainly occurred at the extremities

of the polymer chains, rather than randomly.

Beg and Pickering observed the effect of thermo-mechanical degradation on the physical and mechanical

properties of radiata pine/polypropylene composites during reprocessing [88]. The wood fiber content was set

to 40 and 50 wt.% fiber with 4 wt.% of maleated polypropylene as a coupling agent. Eight reprocessing cycles

were conducted. As for Walz et al., they reported a slight reduction of impact strength, tensile modulus and

strength with recycling [91]. In agreement with several works, they also reported an increase of the tensile

elongation at break and crystallinity [30, 40, 90]. At the end of the recycling process, a decrease of tensile

strength (25%) was observed for the composites at 40 wt.% of fiber, while the Young’s modulus decreased by

17% after eight reprocessing cycles. For the composite at 50 wt.% of fiber, both parameters increased (up to

14% for tensile strength and up to 33% for Young’s modulus over the first two cycles) due to better fiber mixing

and dispersion during reprocessing, but then decreased after the third cycle due to chain scission and fiber

degradation. During reprocessing, both improvement of fiber dispersion and fiber degradation occurred

simultaneously. For lower recycling cycles, the dispersion effect was predominant leading to an increase of the

properties, while for higher cycle numbers, fiber degradation prevailed and the performances decreased. At 40

wt.% of fiber, flexural strength, flexural modulus and impact strength were found to decrease by 30%, 21% and

48% respectively, after eight cycles. Interfacial bonding was found to improve with recycling, while the average

fiber length was reduced by 84% after eight reprocessing cycles. Moreover, the fiber length distribution became

narrower and was shifted towards shorter fiber lengths, confirming that fiber breakup occurred. Contrary to most

previous studies, significant degradation of the properties was observed in this study [31, 40, 72, 90].

Augier et al. studied the influence of wood fiber addition on the mechanical and thermal properties of

mechanically recycled PVC/wood composites [92]. Twenty extrusion-milling cycles were performed on industrial

waste and the results were quite different from the trends seen before. Up to the fifth cycle, the mechanical and

thermal properties of the composites remained constant, but after 10 cycles, a degradation of the molecular

structure was noticed, while the other mechanical properties did not change significantly. However, the flexural

strength increased due to better fiber dispersion in the matrix. For the PVC alone, an important increase of the

impact strength was observed after 20 cycles, as well as no degradation of the molecular structure. The different

behaviors between the composite and the neat matrix was explained by the effect of the fibers which accelerated

27

the matrix degradation, as well as crosslinking reactions. However, the presence of fibers also increased the

composite stability and the mechanical properties of the composite. A significant decrease in the fiber

dimensions and a widening of the size distribution were also noted between the first and the fifth cycles, thus

confirming fiber degradation during mechanical recycling. After five cycles, no further degradation was noticed

and no change was observed for the aspect ratio over the whole recycling process. As for the molecular

distribution, a broadening and a clear shift of the peak to higher molecular weights was observed for the

composite after 20 cycles. This is in contradiction with most studies [30, 40, 70, 90, 93] and was attributed to

dehydrochlorination, formation of unsaturation and crosslinking processes. The authors concluded that no

significant change occurred in the first five cycles, thus making possible to recycle these composites up to five

times, and even more if the addition of virgin material was considered.

Kurniawana et al. evaluated the recyclability of wood/polypropylene composites through three consecutive

extrusion/injection molding cycles [93]. Different wood content up to 50 wt.% were tested and 3 wt.% of maleic

anhydride polypropylene was added to the formulation as a coupling agent. They noticed that recycling did not

decrease the properties of the composites, but increased them instead. Reprocessing appeared to increase the

composite strength until the third cycle, and then remained stable for the composites without coupling agent, but

continued to increase for the composites containing MAPP. Mechanical recycling also led to an increase of the

tensile modulus and elongation at yield for all composites, but composites with coupling agent presented higher

values. As for Twite Kabamba et al., they found that the melting point did not change, while the crystallinity

increased with recycling [40]. Those trends were attributed to chain scission and to the nucleating ability of wood

flour. However, the authors expected this trend to reach a saturation after a number of cycles, as reported by

Beg and Pickering [88]. Finally, this study showed quite unexpected results compared to other studies [30, 40,

72, 90, 93], but led the authors to the same conclusion that, up to five cycles, mechanical recycling did not

substantially decreased the performances and properties of the recycled composites.

Maldas et al. studied the mechanical and dimensional stability of polystyrene/aspen composites after three

mechanical recycling under extreme conditions (exposure to very high and low temperatures, to water and

boiling water [55]. Up to date, this is the only work considering these conditions. They also considered the

influence of coupling agent addition (3% polymethylene polyphenyl isocyanate or PMPI), as well as the use of

fiber coating with a polymer (10 wt.%) and isocynate (8 wt.%). They noticed that the mechanical properties and

dimensional stability of the recycled composites did not substantially change, even after exposure to extreme

conditions. Moreover, they noted that the composites with fiber treatment presented better properties than the

original composites, except after the treatment in boiling and very cold water. Tensile strength decreased during

the first two cycles, and then, slightly increased. On the contrary, elongation and energy at break increased

during the first two cycles, and then decreased. The modulus did not show any significant variation with recycling.

28

Åkesson et al. studied the mechanical recycling of high-density polyethylene (HDPE)/wood fiber composites

[94]. The composites were subjected to seven consecutive injection molding cycles, and some of them were

also submitted to accelerated thermo-oxidative ageing. The results, including tensile, flexural and impact

properties, thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), FTIR spectroscopy and

morphological measurements, showed that the performances were not much affected by the ageing conditions

or the repeated injection molding cycles. Tensile strength showed a slight continuous decrease with recycling,

dropping by 9% between the first and the seventh cycle, and flexural strength followed a very similar trend with

a total decrease of 7%. No significant change was observed for the elastic modulus, as well as for flexural

modulus. On the contrary, the elongation at break sharply increased with recycling, from 4.5% for the original

composite up to 7.8% after the seventh reprocessing cycle (73% increase). This was attributed to lower fiber

length which dropped by 64% during the whole recycling process. The results of the aged sample were very

similar to the unaged sample with less than 5% difference for all the parameters studied. FTIR spectra did not

show any significant degradation. These results indicated that these composites were not much affected by

accelerated thermo-oxidative degradation and could be recycled quite well.

II.3.2 Cellulose

Graupner et al. focused on the behavior of polylactide (PLA) composites during mechanical recycling by injection

molding [95]. The PLA was reinforced with 30 wt.% of regenerated cellulose fibers (lyocell) of variable fineness

and was subjected to three reprocessing cycles. The first reprocessing cycle showed some degradation as the

tensile strength decreased by 28%. This degradation continued through the other cycles as the composite lost

again 22% of its tensile strength during the second cycle, and 5% more during the third cycle, thus totalizing a

global decrease of 46% after three reprocessing cycles. The fiber aspect ratio also decreased during

reprocessing, leading to lower properties. It was also shown that fibers with larger diameters were more sensitive

to breakup than finer fibers, thus leading to a lower decrease in tensile strength for composites reinforced with

finer fibers. It was then concluded that the fiber dimensions, and especially their aspect ratio, strongly affect the

loss of mechanical properties during reprocessing. The properties of a virgin composite with the same fiber

length distribution were also measured and compared to those of the recycled composites. The results indicated

that the fibers and the matrix were both damaged during reprocessing.

Fonseca-Valero studied the mechanical recycling of high density polyethylene/hardwood cellulose composites

by injection molding [89]. Various fiber contents were tested between 10 and 48 wt.%, as well as two different

MAPE coupling agents with different molecular weights and maleic anhydride (MA) contents. Five successive

injection molding cycles were conducted with no addition of virgin materials. The results showed that tensile

strength increased with recycling and, as this parameter appeared to be the most sensitive to fiber breakup, this

effect was higher for lower fiber concentrations. The coupling agent with the higher Mw and MA contents was

29

found to be the most stable towards mechanical recycling, while all the others seemed to lose their

compatibilizing effect after a few cycles due to the high degradation undergone during reprocessing. Recycled

composites also presented a higher relative melt flow index (MFI) compared to virgin ones, and this effect

increased as fiber content increased. ATR-FTIR, SEM and AFM analyses of both composites showed that

important fiber breakup took place in the composite during recycling, thus explaining the properties variations

and the mechanical behavior observed. Moreover, the carbonyl index revealed that the fibers underwent more

thermo-oxidation than the matrix. Due to this important degradation, the beneficial effect of increasing the fiber

content on the tensile strength of the composites was much lower when increasing the number of reprocessing

cycle, as this parameter appeared as the most sensitive to fiber geometry. Elongation at break also increased

substantially with recycling, showing a relative increase of 258%, 128% and 7% for the composites containing

10, 25 and 40 wt.% of cellulose fibers, respectively. With coupling agents, these relative gains were 216%, 33%

and 2% for the low Mw and MA content coupling agent, while being 221%, 59% and 15% for the higher Mw and

MA content coupling agent.

II.3.3 Flax

Moran et al. studied the effect of multiple recycling on the mechanical properties of polypropylene/flax

composites without coupling agent [30]. The flax concentration was set at 20 wt.% and five reprocessing cycle

under constant conditions were carried out to determine the thermo-mechanical degradation undergone by the

constituents during extrusion cycles, and to evaluate the feasibility and viability of recycling these composites.

The mechanical properties were found to increase during the first three cycles due to better mixing and

dispersion of the fiber. But decreases were observed for the next two cycles due to chain scissions and fiber

degradation leading also to darkening. Similarly to Twite Kabamba et al., they also observed that higher

molecular weight chains were more affected by chain scission [40]. They also noted that most of the fiber

degradation occurred during the first reprocessing cycle. No further significant degradation was seen during the

next cycles. However, contrary to Twite Kabamba et al., they did not observe a modification of the polymer

crystallinity over the whole recycling process. The recycled composites presented good impact performances

and high elastic modulus in both flexural and tensile tests. The tensile and flexural strengths were lower than

the matrix, but this difference decreased with recycling. The authors concluded that, even if the recycled material

suffered from chain scission, fiber breakup and global degradation, the mechanical properties were not

significantly affected, and the material could still be used for industrial applications. For this reason, they

concluded that the main parameters affecting the mechanical properties were the mixing quality and the fiber

modulus, instead of fiber geometry. Arbelaiz et al. also studied the mechanical properties of polypropylene/flax

composites after recycling [96]. The results revealed that the tensile properties were stable for the first three

cycles before showing a slight decrease after the fourth cycle, with and without coupling agent. This led the

30

authors to the conclusion that PP/flax composites could be mechanical recycled up to five cycles with no critical

loss of properties.

Le Duigou et al. studied the mechanical recycling of poly(L-lactide) (PLLA)/flax composites, with 20 and 30 wt.%

of fibers [97]. The composites were subjected to six successive injection molding cycles after an initial

compounding in a single-screw extruder. Young’s modulus appeared to remain stable until the third cycle without

any addition of virgin material. However, reprocessing by injection molding seemed to lower the polymer

molecular weight and glass transition temperature, as well as the fiber length and aspect ratio (L/D). It also

induced fiber bundle breakup. On the contrary, the degree of crystallinity showed a slight increase with recycling.

These variations also seemed to improve at higher fiber content, indicating that the latter significantly influenced

the material degradation and the resulting properties by increasing the strain concentrations and the shear rate

during processing. Fiber concentration also strongly influenced the polymer degradation as the average

molecular weight decreased by 83% and 94% at 20 and 30 wt.% of fibers respectively over the whole recycling

process. The glass transition temperature Tg decreased by 6°C at 20 wt.% of fiber and by 20°C at 30 wt.% after

seven injection molding cycles. All these changes were attributed to a higher mobility of the polymer chains due

to both chain scission in the polymer and reduction of fiber dimensions. Once again, fiber breakup essentially

occurred during the first recycling cycles and was attributed to the shear stresses generated in the processing

equipment, which is a function of temperature, pressure and screw speed. The fibers sustained a decrease in

length of about 20% during the extrusion step, and about 70% after both the extrusion and injection molding

steps. Contrary to most studies, after the third cycle and regardless of the fiber content, the authors reported

that both stress and strain at break decreased with reprocessing [30, 55, 88, 89, 90, 93, 94]. The Newtonian

viscosity plateau also significantly decreased with increasing cycle number, indicating an important degradation

of the PLLA matrix, thus confirming that chain scission and fiber breakup occurred.

II.3.4 Sisal and hemp

In several works, Bourmaud and Baley studied the effect of reprocessing on the properties of

polypropylene/vegetal fiber composites [72, 86]. First, they observed the effect of reprocessing on the rigidity of

polypropylene/sisal and polypropylene/hemp composites at 30 wt.% [86]. Once again, the results indicated that

the tensile modulus of the composites was not much affected by mechanical recycling. A slow decrease (10%)

of the modulus was observed for the PP/sisal composites, while the modulus of PP/hemp composites did not

change significantly. This difference was attributed to the different aspect ratio of both fibers which decrease

differently with reprocessing. They finally concluded that these composites exhibited good recyclability. In their

other work, they compared the thermal and mechanical properties of recycled polypropylene composites

reinforced with hemp, sisal and glass fibers [72]. The results showed that after seven reprocessing cycles,

Young’s modulus decreased by less than 1% and by 10% for PP/hemp and PP/sisal composites respectively,

31

while decreasing by 40% for the glass fibers. Tensile strength showed no significant variation over the whole

recycling process for PP/hemp composite, while it decreased by 17% in seven cycles for the PP/sisal and by

52% for PP/glass fiber composites. On the contrary, the elongation at break increased with recycling, reaching

22% for polypropylene/hemp composites, 28% for polypropylene/sisal composites, and up to 34% for glass fiber

composites. This parameter was strongly linked to interfacial bonding. Similarly to most studies on this subject,

the fiber length was shown to significantly decrease during the injection molding process and that the natural

fiber bundles were separated by mechanical recycling [30, 40, 97]. However, the glass fiber only shortened in

length, while the natural fibers were affected in both length and diameter, thus lowering the aspect ratio

reduction. The decrease of fiber length was also higher for natural fiber composites (72% and 77% for hemp

and sisal respectively) than for glass fiber (57%). This could explained the different mechanical behavior between

glass fiber- and natural fiber-reinforced composites, as both fiber length and aspect ratio were considered as

key factors for the rigidity and resistance of composites. Moreover, the adhesion between the fibers and the

matrix seemed to weaken with recycling. An important decrease of the interface adhesion was noted with

recycling as fiber debonding and pulling out were observed after the seven cycles, indicating that the fiber

surface treatment was no longer active after a few reprocessing cycles. The analyses also revealed chain

scission leading to higher ductility, as well as a reduction of viscosity and impact properties. The results indicated

that chain scission preferentially occurred in the macromolecule center of the highest molecular weight chains.

The Newtonian viscosity plateau also decreased by 57% for the glass fiber composites and by 65% for the

natural fibers composites after six reprocessing cycles. This drop was related to chain scission lowering the

molecular weight, as well as fiber size reduction. The crystallinity and crystallization temperature Tc of the natural

fibers composites increased with recycling. This was attributed to the nucleating effect of natural fibers coupled

with the higher mobility of the shortened fibers. The authors finally concluded that PP/vegetal fibers could be

mechanically recycled without any significant properties loss, while PP/glass fiber composites presented a

significant loss of interfacial adhesion and were not well suited for mechanical recycling without any preliminary

interfacial treatment. The use of a coupling agent was shown to improve interfacial adhesion and the mechanical

properties were more stable, but this effect tended to disappear after a few cycles.

II.3.5 Rice hulls and kenaf

Srebrenkoska et al. investigated the mechanical recycling of polypropylene/rice hulls and polypropylene/kenaf

composites [98]. The reinforcement content was set to 30 wt.% and maleic anhydride-grafted PP was used as

a coupling agent to improve the interfacial adhesion in both composites. The authors reported that the recycling

process did not significantly influenced the mechanical and thermal properties of both composites, but the

composites reinforced with rice hulls were more sensitive to recycling than the composites reinforced with kenaf

fibers: the flexural strength decreased by 5% for PP/kenaf composites and by 10% for PP/rice hulls composites.

32

The flexural modulus increased by 20% after recycling for PP/kenaf composites, while it remained stable for

PP/rice hulls composites. The thermal stability of both composites was only slightly affected by recycling, and

the matrix/fiber interface appeared to be slightly improved. They noted that the composites properties were very

similar to the virgin composites and concluded that these recycled composites could find suitable applications

in the current market.

II.3.6 Nettle

Steuernagel studied the mechanical recycling of nettle reinforced polypropylene and compared the results with

those obtained with glass fiber composites [80]. A focus was made on the evolution of stiffness and strength, as

well as fiber dimensions variations. The compounds were produced by a twin-screw extruder with and without

MAPP as a coupling agent. The resulting composites were then pelletized to be processed by injection molding

up to three times to simulate the mechanical recycling. The results showed a loss of stiffness with recycling for

both glass fiber composites and nettle fiber composites without coupling agent. The tensile strength of natural

fibers composites decreased slightly, with a maximum drop of 13% for the material without coupling agent, while

it decreased more strongly (about 25%) for the glass fiber composite. The composite with MAPP showed

relatively stable performances towards recycling. The melt viscosity decreased with recycling for all composites

studied.

II.3.7 Conclusion

Although natural fiber composites production is very well known, the recycling of these composites is much less

investigated [31, 40, 85, 87, 88, 89, 90, 91, 96, 98]. The recycling of this material appears more complex than

neat polymer recycling as both the fiber and matrix can degrade during the recycling process [94]. Moreover,

fiber addition modifies the properties and behavior of the polymer by creating two new interactions: polymer-filler

and filler-filler interactions compared to polymer-polymer interactions in the neat matrix. In general, fiber addition

decreases the elongation properties while increasing brittleness [89]. However, with successive recycling cycles,

the fibers are shortened and lose their reinforcing effect, leading to a lower influence on the composite elongation

[94].

Paukszta and Borysiak published a review considering the influence of processing conditions on the structure

and properties of natural fibers composites [99]. They used the methods available for recycling thermoplastics

composites reinforced with renewable lignocellulosic materials and noted that mechanical recycling modified the

components properties, affected the coupling agent when used, and changed the interactions between the

components. They also reported that the mechanisms taking place during composite recycling are similar to

those taking place during polymer recycling. They also reported that, during grinding and reprocessing, the fibers

were often shortened and underwent internal and external fibrillation. Simultaneously, their dispersion was

33

improved due to a better mixing of the components after the first few recycling cycles. The interface between

the fibers and the matrix also appeared to degrade with recycling, while the coupling agent seemed to lose its

efficiency after a few cycles.

Soroudi and Jakubowicz presented a review on the mechanical recycling of bioplastics, their blends and

biocomposites [78]. However, due to their similarity to petroleum-based polyolefins, bio-derived polyolefins were

not considered in their work which focused on the different mechanical recycling methods, their advantages and

drawbacks, as well as the potential of recycled materials. The authors made the assumption that biocomposites

have a lower recyclability than neat biopolymers due to the higher sensitivity of their reinforcements to thermo-

mechanical degradation, making composting the best recycling method. However, the different studies on

mechanical recycling of such composites showed that these materials could usually be mechanically recycled

for several cycles without any critical loss of performances. Their mechanical properties often remained stable

during those cycles, and the fiber/matrix interface was sometimes improved, even if some degradation occurred

to their components such as fiber length and molecular weight.

Table II-2 summarizes the main results of the studies reported in this chapter. These works seem to confirm the

conclusions reported by Soroudii et al., even if some additional observations can be made [78]. First, it was

expected that fiber incorporation increases the material viscosity and the shear stresses generated in the

extruder, leading to higher mechanical degradation, as well as higher chain scission and fiber breakup, and thus

to lower properties. Then, most of them concluded that fiber length decreases mainly during the first processing

cycle, right after the addition of fibers to the molten matrix, and to an extent that depends on the processing

conditions [40, 85, 100, 101]. Finally, they all agree that, under specific conditions, the recycling of thermoplastics

composites based on natural fibers can be done up to five or ten cycles without any critical loss of the final

properties and performances. However, the effect of recycling on the mechanical, molecular and thermal

properties of thermoplastics composites based on natural fibers remains quite uncertain as various results and

conclusions, sometime contradictory, were reported by the different studies. Depending on the types and

concentrations of the components used, the presence of a coupling agent, the processing parameters or the

environmental conditions, several studies reported the stability of mechanical and thermal properties of their

composites during recycling, while others reported positive or negative variations of some properties like tensile

modulus and strength during reprocessing. In this case, more work should be done to evaluate more precisely

the effect of these factors on the composite behavior towards recycling.

Moreover, even if various studies exist on the subject, several aspects still remain unknown. Some works

investigated the mechanical and morphological properties of natural fiber composites subjected to five or ten

recycling cycles [30, 31, 40, 72, 85, 86, 87, 100, 101], but none of them consider more intensive (higher number

34

of cycles) or open-loop recycling, and not all of them agree on the behavior of the composites during recycling

[30, 31, 40, 72, 85, 86, 87, 100, 101]. Furthermore, they do not consider the effect of ageing and contamination

during its lifetime, as well as the cumulative degradation effect of multiple recycling [94]. Furthermore, very few

studies considered high fiber contents (above 30 wt.%) [94]. A single study considered the effect of recycling

under extreme conditions, subjecting the composite to different temperature and moisture conditions. None of

the works available focused on foamed composites and some work should be done on these aspects.

Table II-2: Overview of the investigations published on the mechanical recycling of natural organic fillers reinforced composites with their main parameters.

Authors Matrix Reinforcement Fiber content

(wt.%)

Coupling

agent

Total number

of cycles

Twite Kabamba et al. [40] LDPE Wood 15 - 10

Wang et al. [31] PE Wood 30 HDPE-AA 1

Petchwattana et al. [70] PVC Wood 20 - 7

Soccalingame [90] PP Wood 20 MAPP 7

Beg and Pickering [88] PP Wood 40-50 MAPP 8

Augier et al. [92] PVC Wood 40 - 20

Kurniawana et al. [93] PP Wood ≤ 50 MAPP 3

Maldas et al. [55] PS Wood 5-35 PMPI 3

Åkesson et al. [94] HDPE Wood 10 - 7

Graupner et al. [95] PLA Cellulose 30 - 3

Fonsecca-Valero [89] HDPE Cellulose 10-48 MAPE 5

Moran et al. [30] PP Flax 20 - 5

Arbelaiz et al. [96] PP Flax 1-20 MAPP 5

Le Duigou et al. [97] PLLA Flax 20-30 - 6

Bourmaud and Baley [86] PP Sisal, hemp 30 - 7

Bourmaud and Baley [72] PP Sisal, hemp, glass 30 MAPP 7

Srebrenkoska et al. [98] PP Rice hulls, kenaf 30 MAPP 2

Steuernagel [80] PP Nettle N/A MAPP 3

35

II.4 Recycling of thermoplastic composites reinforc ed with inorganic fillers

Due to their high specific properties, carbon and glass fibers are among the most widely used reinforcement in

composites, not only for thermoplastics. This is associated to their wide range of applications in structural and

semi-structural applications [102]. These composites represent important volumes of material and thus need to

be recycled. In the automotive field, these fibers are still widely used due to their high specific properties.

However, in Europe, the End-of-Life Vehicles (ELV) directive 2000/53/EC requires that the producers should

reuse and recover a minimum of 95% of a vehicle by weight, and a minimum of 85% of it should be reused or

recycled, thus limiting to 5% the amount of material that can be discarded [102, 103]. This directive led to

numerous studies on the recycling of such composites, in order to develop efficient and economical recycling

technologies. Moreover, as far as recycling is concerned, thermoplastics seem to be the best matrix choice due

to their ability to be remelted and reshaped. The resulting composites present good chemical and impact

resistances, as well as design flexibility, good reprocessability and recyclability [102]. Their recycling methods

have thus been widely studied as they are gaining more and more interest with new environmental regulations

[103]. The fiber length and concentration, interfacial adhesion and polymer matrix properties are the key factors

for the composites final properties. Thus, mechanical recycling methods should be optimized to best preserve

these properties during reprocessing [102].

II.4.1 Glass

Ville et al. studied the mechanical recycling of glass fiber reinforced polyamide 12 (PA12) in both laboratory and

industrial scale twin-screw extruders over a range of processing parameters [85]. Contrary to most studies on

this subject, which considered the recycled materials after its reprocessing, they took samples at various

locations along the extruder screw. The results showed that the fiber length decreased with recycling, leading to

a significant modification of the fiber length distribution. Moreover, an important part of this breakup occurred at

the early stages of mixing, right after the introduction of the fibers in the extruder, near the feeder. However,

some size reductions also occurred further down the screw. After the first breakup right after the fiber

introduction, the fiber length was also reduced after the first mixing element, and then continuously decreased

until the end of the screw. Finally, a last sharp fiber size reduction was observed when the composite came

through the die. Then, when the fiber reached a critical length, breakup was more difficult and the influence of

the processing parameters was more limited. The size distributions shifted towards smaller values and became

narrower. The extent of degradation varied significantly with the processing parameters and equipment used. It

increased with screw speed and residence time, but decreased with feed rate. Surprisingly, it was higher in the

small laboratory scale extruder than in the industrial one. The authors reported that various mechanisms took

place during fiber breakup such as individual breakage, bundle breakup into individual fibers by rupture and

36

erosion, by buckling and by fiber-fiber interactions. They finally concluded that, due to their important influence

on the composites properties, reprocessing conditions should be very carefully chosen to limit the loss of

performances of recycled composites.

Corvaglia et al. proposed a procedure for the secondary mechanical recycling of polypropylene (PP)-based

sandwich panels into short-fiber composites and studied the influence of processing parameters and fiber

content on the mechanical and physical properties of the recycled material [103]. The material was made of a

polypropylene foam core with polypropylene/continuous glass fibers laminate skins. This type of recycling was

qualified as “secondary” as the continuous glass fibers were transformed into short glass fibers during

reprocessing. To be recycled, this material was ground in a mill and then remelted by melt mixing or single-

screw extrusion. The samples for the mechanical and morphological characterizations were then made by

injection molding. The results showed that the tensile strength was more sensitive to fiber length reduction than

tensile modulus. As in Ville et al., the residence time seemed to strongly influence the thermo-mechanical

degradation undergone by the composite due to the high temperatures and mechanical stresses generated in

the processing equipment [85]. The reduction of the molecular weight and fiber size induced by this degradation

increased with residence time, depending on the screw speed. However, if the residence time should be as low

as possible to limit degradation, it should also be high enough to get good fiber dispersion and proper fiber

wetting by the matrix. According to the mechanical results, a maximum screw speed of 40 rpm was proposed

for the processing of such materials. Similarly, the processing temperature should be set very carefully. It should

be low enough as to limit thermo-degradation of the composite, but high enough to limit the matrix viscosity

which is known to induce high shear stresses and fiber degradation. This was confirmed by morphological

analyses showing a higher amount of short fibers at lower temperature. The authors also reported that the

residence time should be kept under 15 minutes and processing temperatures should be set slightly higher than

200°C to avoid excessive degradation for both components, and preserve the mechanical properties of the

recycled material. They also indicated that, according to their mechanical characterizations, the internal mixer

led to lower matrix degradation compared to the extruder and should be favored for the reprocessing of these

composites.

Cornier-Ríos et al. studied the mechanical recycling of polyethylene terephthalate (PET)/glass composites [104].

The glass filler content was set to 15 wt.% and six reprocessing cycles were performed for each material and

four recycled content were considered (0, 25, 50, and 100%) in the virgin material. The thermal properties

appeared to remain stable regardless of the number of generation and recycle ratio, whereas the mechanical

properties slightly decreased (maximum 6%) with recycling and recycle ratio due to fiber length reduction. The

tensile strength showed a reduction of about 3% per cycle for the totally recycled material, and also decreased

with increasing recycling ratio and recycling generations for samples with mixed generations. The decrease was

37

however lower for lower recycle ratio as the addition of virgin materials limited the loss of properties because of

its lower degradation state. On the contrary, elongation at break and elastic modulus did not show any particular

trend with mechanical recycling or recycle ratio. Moreover, crystallinity remained stable regardless of the number

of cycle or the recycled material content. As the materials were easily melt processed at standard temperatures,

it was concluded that no crosslinking occurred during recycling. All these results indicated that the recycling of

glass fiber reinforced PET could be done without any significant loss of properties.

Kemmochi et al. focused on the effect of closed-loop recycling on the properties of fiber reinforced thermoplastics

(FRTP), and especially on the fiber length inside the composites [105]. Different types of glass reinforcement

were considered: powder, short, long and continuous fibers. They simulated the lifecycle by artificial accelerated

ageing coupled with natural weathering. They reported that the FRTP could be recycled in four main steps, from

the continuous fiber to the powder and then to energy recovery. First, the continuous fiber reinforced

thermoplastics became a long fiber reinforced thermoplastics after one recycling cycle. Then, due to fiber

degradation with reprocessing, those long FRTP became short fiber reinforced thermoplastics. After another

cycle, they can finally be recycled as glass powder reinforced thermoplastics. In the end, as the reinforcement

effect lose its effect, they can be energetically recycled or added to virgin materials. They indicated that each

step was equivalent to five to twenty years of normal product use. This work also revealed that, in the case of

short fiber composites, for a defined fiber length, tensile and bending strengths were higher for recycled materials

than for virgin ones with similar fiber distribution. This was attributed to better fiber orientation in the thermoplastic

matrix. The bending rigidity also showed no significant influence of the recycling for each category. On the

contrary, the mechanical strengths of these composites were only affected by fiber length and orientation.

A general review was also done on the different ways of mechanically recycle glass mat reinforced

thermoplastics (GMT) [102]. Thermoplastics composites were shown to be easier to recycle due to their ability

to be remelted and reshaped multiple times. However, due to the necessary milling step of mechanical recycling,

they cannot be used as GMT and present a certain loss of properties after this type of recycling. Glass mat

reinforced thermoplastics are long glass fibers composites usually made of two layers of glass mat wetted by a

polymer matrix and mostly supplied in sheet form as semi-finished material. They are shaped by a subsequent

molding process and present high properties. After the milling process, the fibers become shorter and cannot

be used as long fibers anymore. The authors reported that three ways exist to reuse this type of material. First,

small parts can be remelted and molded into a larger part. However, this method is difficult to implement as it is

hard to find enough parts of the same kind to be recycled. Otherwise, ground GMT could be used as raw material

chips to produce semi-finished sheets of materials with the same processes again. Finally, they could also be

used as short glass fiber reinforcement in extruded or injection molded thermoplastics composites.

38

II.4.2 Carbon

Colucci et al. studied the mechanical recycling of polyamide 66 (PA66)/short carbon fibers by injection molding

[106]. The fiber content was set to 30 wt.%. As these composites were intended for automotive applications, the

materials were submitted to some artificial aging by temperature, moisture and ultraviolet (UV) treatments before

recycling. These factors were chosen to simulate common outdoor conditions (moisture, rain and sunlight) that

the material will sustain in combination to mechanical stresses during its lifecycle, and thus evaluate their effect

on the composite properties. It was found that aging led to lower tensile strength and Young’s modulus, while

the elongation properties slightly increased. These trends were attributed to the photo- and thermo-oxidative

degradations generated in the polymer matrix by these three factors. Rheological tests revealed that the complex

viscosity decreased by 71% with both aging and recycling, while unaged samples showed a much lower

decrease (12%). This was attributed to chain scission in the polymer matrix and confirmed the important effect

of ageing on the resulting properties of the recycled composites. The weight average molecular weight (Mw) of

the aged composites increased with recycling, while the number average molecular weight (Mn) did not

significantly changed. As most of the properties loss occurred after the artificial ageing, and that the mechanical,

morphological and thermal properties of the unaged composites remained constant with recycling, the authors

concluded that mechanical recycling did not have a significant effect on the composite properties so it can be

used for such composites without any significant loss of performances.

Sarasua and Pouyet investigated the effect of mechanical recycling on the mechanical and morphological

properties of poly-ether-ether-ketone (PEEK)/carbon composites [107]. The recycling process was performed

by injection molding and the short carbon fibers content was set at 10% and 30%. It was found that both fibers

and matrix degraded during recycling, while no significant modification was observed at the fiber-matrix interface.

The morphological analyses revealed that the fiber length distribution was narrowed and that the average fiber

length decreased with recycling. Moreover, the polymer molecular weight decreased by 28% over the ten

recycling cycles. The results obtained were also in agreement with the trends reported in most previous studies

[85, 103, 104]. Young’s modulus and tensile strength decreased with recycling, showing a Young’s modulus

reduction of 10% and 30% at 10% and 30% fibers respectively, and about 20% for their tensile strength. Impact

strength also decreased with recycling, showing a loss of 39% in ten cycles for the composite with 10% fibers,

and of 24% in ten cycles at higher fiber content (30%). On the contrary, ductility increased with reprocessing as

elongation at break increased by 225% for the 10% carbon fiber composites, and by 42% for the 30% carbon

fiber reinforced PEEK. The authors also reported that the failure mode was highly function of the fiber and that

recycling did not influence significantly the high deformation properties of the composites. They concluded that

important fiber damage occurred during processing, but the properties were well enough conserved during the

recycling process to use these materials for specific applications.

39

Schinner et al. presented two methods for the mechanical recycling of thermoplastics/carbon fiber composites

[108]. The first one considered the grinding of the thermoplastic composite waste and its incorporation as

reinforcement into compounds made of virgin materials. The results showed that the properties evaluated were

similar to those of the totally new materials, thus leading the authors to conclude to the viability of this type of

recycling. The second method was based on the reprocessing and reshaping of the used thermoplastics/carbon

fiber composites. This second approached also revealed that the properties were not significantly altered, thus

confirming the viability of this second method as a suitable mechanical recycling method for

thermoplastics/carbon fiber composites.

II.4.3 Talc

Wang et al. studied the behaviors of polypropylene/talc composites subjected to mechanical recycling [109].

This study revealed that the mechanical properties of these composites depended on the number of recycling

cycle, the reinforcement content and the reprocessing temperature. Both the tensile modulus and the tensile

strength at yield increased with recycling due to lower fiber dimensions and increase of their aspect ratio. On

the contrary, the compression modulus decreased continuously with increasing reprocessing. As for previous

studies, the authors found that mechanical recycling led to a reduction of the matrix molecular weights due to

chain scission and a crystallinity increase due to higher mobility of the shortened fibers [85, 103, 106]. They also

reported that the decomposition temperature remained stable during recycling, allowing the reprocessing at

similar temperatures than for virgin materials.

II.4.4 Conclusion

In their review on mechanical recycling of polyolefin waste, Yin et al. investigated inorganic filler reinforced

composites [67]. They reported that this type of filler generally improves the mechanical properties of materials,

even at small content and that the final properties of inorganic fillers reinforced composites depends on the filler

geometry and distribution, as well as on the interfacial adhesion between the fillers and the matrix. Besides the

commonly known effects, they noted that every inorganic filler has its own specific effect on the composite

properties: clay was shown to improve thermal stability and fire resistance, titanium dioxide had good

antimicrobial properties as well as ultra-violet (UV) protection, talc produced an effective nucleating effect, while

glass fiber led to high specific mechanical properties, low cost, lightweight and resistance to chemicals. They

also confirmed the trends seen in the previous studies; i.e. an increase of both Young’s modulus and tensile

strength with recycling, fiber breakup, and lower elongation at break. Most of the studies also concluded that the

modifications of the properties during reprocessing did not significantly changed the final performances of the

recycled materials and these materials could be successfully mechanically recycled without any important loss

of properties. All the works reported in this chapter are summarized in Table II-3.

40

As for natural organic reinforcement, even if several works were published on this subject, different aspects still

remain unknown. Once again, mechanical properties have been widely studied, as well as fiber length and

molecular weight distributions [85, 103, 104, 105, 106, 107, 109]. Thermal properties were also frequently

studied [103, 104, 109]. However, very few of them focused on the rheological and physical properties.

Moreover, all of them considered less than ten recycling cycles, as well as closed-loop recycling. Thus, long-

term recycling and open-loop recycling were not studied, as well as material inhomogeneity, ageing during its

life cycle and possible contamination were also left out of this investigation. The same comment applies to the

cumulative long-term effect of mechanical recycling occurring after a high number of reprocessing cycles. Once

again, recycling under extreme conditions, high reinforcement content, and recycling of foamed composites were

also left out. In order to completely understand the behavior under recycling, additional work must be done on

these different aspects.

Table II-3: Overview of the works considering the mechanical recycling of inorganic fillers reinforced composites with their main characteristics.

Authors Matrix Reinforcement

Fiber

content

(wt.%)

Number of

cycles

Ville et al. [85] PA12 Glass fiber 30 1

Corvaglia et al. [103] PP Glass fiber 40 1

Cornier-Ríos et al. [104] PET Glass fiber 15 5

Kemmochi et al. [105] Nylon 6 Glass - 4

Colucci et al. [106] PA66 Carbon 30 1

Sarasua and Pouyet [107] PEEK Carbon 10-30 10

Wang et al. [109] PP Talc 10-20 6

41

II.5 Limitations and solutions

As seen in the previous parts, for both organic and inorganic fillers, some problems are limiting the use of

mechanical recycling for thermoplastic composites. First, for natural fiber reinforced thermoplastics, the main

limitation to mechanical recycling is the sensitivity of these fibers towards high temperatures. As seen previously,

natural fibers often present degradation temperature around 150 to 200°C, which is lower than most of the

thermoplastics meting points, and below usual processing temperature [31, 110, 111]. According to Prasanth et

al., natural fibers are usually composed of cellulose, hemicellulose, and lignin, and each of these components

presents its own behavior and limits with respect to thermal degradation, which usually occurs in two consecutive

steps [111]. First, lignin degrades at low temperature (80-180°C), and then cellulose degrades at much higher

temperatures (280-380°C). In order to produce and recycle such composites, processing temperatures should

be higher than the melting point of the polymer matrix, and are thus higher than the onset thermal degradation

temperature of most natural fibers. This leads to some additional degradation and darkening of the fiber during

reprocessing (mechanical recycling), leading to properties losses, as well as to volatile emissions producing

undesired foaming [111]. As this additional thermal degradation is added to the thermo-mechanical degradation

generated by the shear stresses during processing, once the components of the composites chosen, the only

remaining way to limit the properties losses is to limit the thermo-mechanical degradation induced by the shear

stresses. This can be done by an optimization of the processing parameters. The processing temperatures and

profiles should be high enough to favor good reinforcement dispersion and to allow smooth processing, but not

too high to limit thermal degradation. Similarly, screw speed should be set high enough to maximize

homogenization and throughput and to minimize residence time, but not too high as to limit the shear rate

undergone by the composite during processing [111, 112, 113]. Winandy at al. also reported that using waste

newsprints instead of natural fibers could lead to lower loss of performance due to their lower degradation and

higher stability under high temperature [114]. The use of counter-rotating screws also seems to decrease this

loss of performance [114, 115]. Prasanth et al. noted that some chemical fiber modifications could improve the

thermal stability of these fibers, and thus inhibit the performances losses undergone by these composites [111].

Another problem encountered for natural organic fiber reinforced thermoplastics is the loss of performance

induced by moisture absorption. As most natural organic fibers are hydrophilic, they absorb moisture during their

lifecycle and their reprocessing leading to swelling and to foaming of the material during reprocessing, if no

previous drying treatment was applied [111]. Moisture is also absorbed in the fibers by capillary forces [111].

This modifies the fiber dimensions and density, affecting the physical, chemical, mechanical and morphological

properties of the recycled composites [115]. But foaming can be avoided by preliminary fiber drying before

recycling, but swelling will still occur during the use phase. The absorption of moisture by natural fibers is usually

reduced by chemical modification of the fiber surface [111]. The use of coupling agents, such as polyolefin

42

grafted with maleic anhydride, seems to counter this effect as the anhydride groups react with the hydroxyl

groups of natural fibers decreasing the number of sites where water molecules can bond [115].

The second and most important challenge for mechanical recycling of thermoplastics composites is

inhomogeneity of the waste stream. This appears to limit the final performances of the recycled materials by

many aspects. First, most thermoplastics are immiscible and incompatible, forming different phases when mixed

and leading to lower mechanical properties of the final recycled composites [32, 69, 116, 117]. This problem

requires the use of stabilizers and compatibilizers at high content to improve the cohesion between the phases,

which is hardly economically feasible, or the use of mechano-chemical methods leading to substantial

degradation of the composite, and especially of the reinforcement [32, 67]. Mechano-chemical processes are

simple, fast, economic and ecofriendly. However, important degradation of the composite’s reinforcement limits

its use for this type of materials. Mechano-chemical methods are based on the pulverization of the polymer

materials to obtain reactive blending without using any additives [67, 69, 115, 117, 118]. They use a wide range

of reactions based on various solicitations such as impact, shear, elongation, fracture and compression,

generally generating chains scission in the polymer matrix, but also induce the formation of free radicals which

are from different molecular species and react with each other to induce coupling and crosslinking [67].

Mechanical milling is the copulverization of materials with a high-energy ball milling process to get the materials

in a powder form. This increases the blending degree, but also leads to a major reduction of the material size

and, in the case of composite materials, to major fiber breakup. Some methods also use liquid and cryogenic

gases to make the pulverization step easier [67, 115, 117, 118]. The use of stabilizers and compatibilizers is

possible, but adds extra costs to the recycling process because these agents have to be renewed regularly as

they seem to lose their efficiencies after a few regenerations [69]. However, the addition of such agents leads to

better properties, and slightly decreases the need for sorting before reprocessing. Another way of improving the

interfacial adhesion is via irradiation with electrons to generate free radicals in the materials increasing reactivity

and adhesion between the phases [78].

Due to a wide variety of composite waste sources, diversity in grade and composition even after sorting, and

inconsistencies of the waste stream lead to a significant variability of the properties of the final materials [81,

115, 118]. The best solution would be to implement better sorting methods, but this is not economically and

technically feasible to reach a perfect sorting of each composite by matrix, fiber and additive types, especially

considering that composites are multi-components materials. Some differences in the composition will always

remain, especially regarding the additives contents in materials. Due to this diversity of sources, but also in life

conditions, the composites present different properties and behaviors, depending on their nature, composition,

and degradation extent [69]. The most important problem is the melting point. If it seem very uncommon that

recycling modify the meting point of the polymer matrix, as no significant difference could be noticed in most

43

recycling studies, in some specific cases, if the material contains some impurities and additives, the melting

point can be changed [69]. In this case, the blends will have a distribution of melting temperatures instead of a

unique melting point, which is an issue for recycling as the matrices with low melting point will flow faster than

the ones with high melting point, leading to a heterogeneous material induced by partial melting or thermal

degradation.

Postconsumer materials also usually contain contaminants that could further decrease the final properties of the

recycled composites, or induce toxicity, unpleasant odors or discoloration, which highly reduce the attractiveness

of the final recycled products [69, 71, 73, 81, 115, 116, 118, 119]. Contamination can be associated to the

environment of the product, or to its assembly with other parts. Common sources of contamination are inserts,

screws, bolts, staples, nails, metals parts, labels, adhesives, paints, varnish and other finish products, printing

products and residues [113, 115, 116]. Once again, compatibilizers can be used, but only for small volumes of

contaminants [71, 73]. The best solution remains to improve the waste collection, separation and cleaning steps,

but this adds extra costs and leads to several technical challenges. Still today, these steps remain quite limited

for composites [56, 68, 81]. However, these considerations should become easier in the future as a new trend

toward considering the dismounting, separation and recycling of the materials progressively settles into the

developers’ minds [68, 78, 81]. Deconstruction, separation of the various components and recycling are more

and more considered during the research and development phase of the composites, as to modify and adjust

the composite design to facilitate these steps at the end of the composite lifecycle.

Another limitation for the use of mechanical recycling for composites is the degradation undergone during its

lifecycle, which is not taken into account in most studies. This is also expected to decrease the stability and

performances of the final recycled material [67, 78, 116]. Various types of degradation can occur during a

composite lifecycle such as mechanical (shear and elongation) stresses, oxidation, burning, chemical attacks,

ultra-violet (UV), biological and moisture degradations [67, 78, 111, 114, 116]. Besides decreasing the final

properties and performances, these degradation processes can lead to partial reticulation of the polymer matrix,

thus to higher viscosity and the creation of stress concentration points making more difficult the recycling due to

the high pressure, energy and cost required, and the lower final performances [56, 67, 71]. This degradation is

unavoidable and is the focus of most actual studies on thermoplastics composites recycling.

Finally, the lack of markets available for recycled composites also appears as a limitation to the implementation

and the spreading of the mechanical recycled composites, as well as high costs [56]. If these composites are

more expensive than virgin ones, lower properties could be of interest for less demanding applications, but the

attractiveness of such materials is very limited [56]. This is why more research should be done to improve these

44

aspects as to meet the quality requirements at low cost, as well as being able to fit into the same applications

than for virgin composites. All these limitations and their potential solutions are summarized in Table II-4.

Table II-4: Main limitations and solutions for the mechanical recycling of thermoplastic composites.

Limitations Possible solutions

Sensitivity of natural fibers to high temperatures Optimization of the processing parameters and

material components

Sensitivity of natural fibers to humidity Fiber drying and surface treatment

Inhomogeneity of the waste streams Improved sorting methods, use of compatibilizers,

mechano-chemistry

Contamination of the waste Improved sorting methods, design modifications

for better sorting and recycling, use of

compatibilizers

Degradation during recycling Optimization of the processing parameters

Lack of market Final properties improvement, lowering of the

costs

II.6 Conclusion

Although thermoplastics composites processing is very well known, the behavior of these materials towards

mechanical recycling is much less understood. The recycling of these materials appears to be more complex

than neat polymer recycling as both the fiber and the matrix can degrade during recycling. Furthermore, the

presence of reinforcing particles modifies the polymer behavior and properties (morphology, mechanics and

rheology), while simultaneously creating new interactions between the filler and the matrix, as well as between

the particles of the matrix [94]. However, most degradation mechanisms are similar to those occurring during

the recycling of polymers [99]. The number average molecular weight Mn decreases with recycling, while the

weight average Mw remains mostly constant [40, 72, 90, 97]. Crystallinity increases with the number of cycles

[40, 72, 88, 93, 97, 109] and chain scission mostly occurs around the mid-point of longer molecules [30, 40, 72],

leading to a narrower molecular weight distribution.

However, most of the materials properties differ from the neat polymer case: fiber addition leads to more

degradation [30, 40, 92]. This degradation depends on the filler type and size [72, 80, 86, 90, 95], as larger

particles seem to lead to higher degradation [90, 95] and more rigid fibers like glass fibers seem more affected

by recycling [72, 80]. Viscosity increases with filler addition, but decreases faster due to particle breakup during

recycling [40, 86, 90, 97, 106]. Similarly, elongation at break decreases sharply with filler addition, but was found

45

to increase with recycling as filler dimensions decreases with successive recycling losing their reinforcing effect

[72, 86, 89, 90, 93, 94, 106, 107]. Several works also pointed out that fiber breakup mainly occurs during the

first cycle, and that its extent depends on the processing conditions [30, 40, 85, 97]. Thermo-mechanical

degradation is generated in the material by high temperature, high pressure and high shear stress. This

combination was found to depend on screw speed, residence time and feed rate [30, 31, 85, 89, 103, 107].

These parameters should be carefully chosen to get the best compromise between good homogeneity and low

degradation. If the temperature is too high, thermal degradation will be important and homogeneity will be good

due to the low viscosity of the melt, but if temperature is too low, poor homogeneity (filler distribution) occurs

and mechanical degradation is important due to high viscosity (stresses on the particles) [103].

Several aspects related to composite recycling remain unknown. The presence of crosslinking remains unclear

[40, 90, 104]. Furthermore, if all the authors agree to say that fiber length decreases due to fiber breakup, not

all of them agree about changes in the diameter and aspect ratio [30, 40, 85, 88, 89, 90, 94, 95, 97, 103]. Some

works reported that fiber diameter remained constant leading to a decreasing aspect ratio (L/D) [40, 72, 95, 97],

while others observed that diameter decreases and that the L/D was almost constant [30, 72, 92]. Mechanical

properties changes is also not clear. The majority of works indicates that the tensile and flexural strengths slightly

decreases with recycling [72, 80, 88, 94, 105, 107], while others reported that these parameters increase [89,

92, 93, 104, 105, 106] or even remained constant [86, 90]. Similar results were obtained for impact strength [72,

107, 120]. Also, the modulus was found to increase in some works [93, 98, 104], while most studies reported no

significant variation [55, 70, 86, 94, 97, 105]. Nevertheless, all the literature agrees to say that these variations

were not important and that no significant loss of performance resulted from the mechanical recycling of

thermoplastics composites [30, 31, 55, 70, 72, 86, 94, 96, 98, 106].

The behavior and the performances of thermoplastics composites have been the subject of several studies and

are well known. As the mechanical properties do not vary significantly, it can be concluded that composites

present great aptitudes to short-term mechanical recycling. However, the influence of recycling on the

mechanical, molecular and thermal properties of thermoplastics composites remains quite uncertain as various

results and conclusions were reported by the different studies. This is why more work should be done as to

evaluate more precisely the effect of these factors on the composite behavior towards recycling. Moreover, most

of these studies considered a very low number of cycles, generally less than 10. Extensive long-term recycling

of thermoplastics composites has not been studied, and additional work should be conducted on these aspects

to fill this lack of knowledge. Moreover, very few studies took into consideration extreme conditions like moisture

and temperature [55, 89] or very high filler contents (above 30 wt.%), and none considered open-loop recycling.

Most of them are not taking into consideration the ageing and contamination of the material during its lifetime,

the inhomogeneity of the waste stream, as well as the cumulative degradation effect of recycling occurring after

46

a higher number of reprocessing cycles. Finally, no work considered the mechanical recycling of foamed

thermoplastics composites as these materials are becoming very popular in view of material savings and weight

reduction. These aspects should be the focus of more research.

Mechanical recycling of thermoplastics composites also reached some limitations affecting their performances

and their viability. The inhomogeneity of the waste, as well as the possible contamination of the waste stream

lead to lower properties. Compatibilizers and filler surface treatments could be used to limit the properties losses,

but they were shown to lose their performances with recycling [72, 89]. The sensitivity of natural fibers

composites to high temperatures and humidity also decrease their performances. Furthermore, the thermo-

mechanical degradation undergone by the composites and their constituents during recycling also limit their

performances and their possible applications, leading to a lack of market for such composites. Partial solutions

exist and should be considered, but more research must be conducted in this field to overcome these limitations

and increase the performances of recycled composites, as well as to increase the number of possible

applications for such composites.

Acknowledgements

Financial support from the Natural Sciences and Engineering Research Council of Canada (NSERC, Canada),

the Research Center for High Performance Polymer and Composite Systems (CREPEC, Quebec), the Quebec

Research Funds – Nature and Technologies (FQRNT, Quebec), the Quebec Center for Functional Materials

(CQMF, Quebec), the Research Center for Advanced Materials (CERMA, Quebec), and the Renewable

Materials Research Centre (CRMR, Quebec) is acknowledged for this work.

47

Chapter III. Long-term recycling of high density polyethylene and characterization of its closed-loop degradation.

Résumé

L’objectif de ce travail est d’étudier la perte de propriétés subie par un polymère lors d’un procédé de recyclage

long-terme en boucle fermée. Pour cela, du polyéthylène haute densité est soumis à 50 cycles d’extrusion sous

conditions constantes. L’effet du recyclage est alors déterminé en observant l’évolution de la dégradation avec

le nombre de cycles considéré. Après certains cycles, le matériau est caractérisé en termes de propriétés

physiques (masse volumique, chromatographie à perméation de gel), thermiques (calorimétrie différentielle à

balayage, analyse thermogravimétrique) et mécaniques (traction, flexion). Aucun changement significatif de

masse volumique et de propriétés thermiques n’est observé. Toutefois, lors de la chromatographie à perméation

de gel, la masse moléculaire moyenne en poids (Mw) diminue avec le recyclage, alors que la masse moléculaire

moyenne en nombre (Mn) ne présente aucun changement significatif, induisant ainsi une diminution de l’indice

de polydispersité. La viscosité intrinsèque diminue aussi, alors que l’indice de fluidité à chaud augmente. D’après

les courbes contrainte-déformation obtenues en traction, le recyclage ne semble pas avoir d’effet sur le module

d’Young (Ey), alors que la déformation au seuil d’élasticité augmente modérément puis diminue légèrement, et

que la contrainte au seuil d’élasticité diminue. La contrainte et l’énergie à la rupture augmentent

significativement. Finalement, les tests de flexion trois points indiquent une diminution du module en flexion (Eb)

avec le recyclage. Globalement, le recyclage mène à une modification importante des propriétés mécaniques

du polymère en raison des ruptures de chaînes provoquées lors de la remise en œuvre.

Mots-clés: Recycling, extrusion, HDPE, molecular properties, mechanical properties, thermal properties.

48

Abstract

The objective of this work is to investigate the loss of performance undergone by a polymer during a long-term

closed-loop recycling process. High density polyethylene (HDPE) is subjected up to 50 extrusion cycles under

constant processing conditions. The effect of recycling is then determined by following its degradation with

increasing number of generation. After selected cycles, the material is characterized in terms of physical (density,

GPC), thermal (DSC, TGA), and mechanical (tension, flexion) properties. No significant change are observed in

density, DSC, and TGA tests. But for GPC, the weight average molecular weight (Mw) is found to decrease while

the number average molecular weight (Mn) do not change significantly, thus leading to a decreasing

polydispersity index. Intrinsic viscosity also decreases, while melt flow index (MFI) increases. From the tensile

stress-strain curves, recycling seems to have no significant effect on Young’s modulus (E), but a moderate

increase of the strain at yield is observed followed by a slight decrease, while the stress at yield decreases. For

the break-up conditions, stress and energy at break are found to increase significantly. Finally, three-point

bending tests show that the flexural modulus (Eb) decreases with recycling. Overall, the recycling process leads

to an important modification of the polymer’s mechanical properties mainly due to chain scission.

Key Words: Recycling, extrusion, HDPE, molecular properties, mechanical properties, thermal properties.

N. Benoit, R. González-Núñez, and D. Rodrigue. High Density Polyethylene degradation followed by closed-

loop recycling, Progress in Rubber, Plastics & Recycling Technology, 33, 17 (2017).

49

III.1 Introduction

Due to their numerous benefits such as low cost, wide variety, and flexibility, plastics are omnipresent in the

modern society. They have yet replaced several traditional materials in various fields such as construction,

packaging, transports, agriculture, electronics, chemistry, medicine, and household appliances. New polymers

and applications are continuously developed to meet the increasing demand for new materials and technologies.

But this diversity of applications and growing production lead to the creation of important volumes of domestic

and industrial waste. These materials must be collected, stored, and treated after their end of life which implies

logistic concerns. Still today, it is quite complex to sort all this material considering the diversity of waste sources,

polymer types, and products [7, 36, 37, 121].

With increasing number of new directives and policies towards sustainability, alternative energies, recycling, and

other environmentally friendly processes, lifecycle analysis and end-of-life considerations must be accounted for

at the same time as the development of newer, simpler, more efficient, and less expensive recycling methods.

But the latter are complex to develop as they need to integrate a wide diversity of economic and technical criteria

[121]. Sustainability requires financial investment that only contribute to better health and life quality. This is

generally the opposite of economic interests controlling the actual society and can only be achieved with

regulations and policies [122]. Due to both consumer and governmental demands, more and more recycling

methods are developed [118, 123]. Moreover, plastics can partly recover their value by recycling [58]. This allows

to preserve natural non-renewable resources while limiting the waste going to landfills [7, 36, 37].

There are several sources and types of materials. The higher the diversity of materials, the more difficult the

recycling and the more dispersion on final properties. For recycled materials, specifications cannot be too high

considering the broadening of the final properties distribution. Plastic waste is generated all along the lifecycle

of a product, including manufacturing. Even the most optimized operation has some material loss at each step.

DeRudder et al. defined two types of recycling [124]. If the material is recycled directly after its manufacturing,

this recycling is known as closed-loop, but if it goes through any part of its intended lifecycle, it is called open-

loop [123, 124].

There are four main options for recycling. In some specific cases, simple reuse of the product can be considered

after cleaning, but those cases are limited for health and safety reasons [7]. In mechanical recycling, the polymer

is collected and grinded into small pieces. They are then cleaned, separated into different grades, and

reprocessed. This type of recycling is widely practiced as it is quite easy, reliable, and economic as long as the

plastic is clean and not too diversified. The chemical structure remains almost unchanged, but properties often

decrease leading to the need of blending recycled and virgin polymers to keep the properties up [121]. In

feedstock recycling, also called chemical recycling, the polymer is turned back into its chemical components,

50

which can be used to produce new polymers. Several technologies are available, including pyrolysis,

gasification, depolymerization, solvolysis, blast furnace or smelter operations. Finally, energy recovery consists

of polymer combustion while recovering the inherent energy released to generate heat, steam or electricity for

other processes, thus conserving resources. This technique is essentially used for contaminated or highly mixed

fractions [7, 58, 36, 37, 125, 126].

Sorting is usually done manually by plastic code or type of container. Automated sorting equipment exists, but

are expensive and not totally reliable. Various separation techniques have been developed, but still remain

expensive or use hazardous solvents [7, 53, 95, 117, 127, 128, 129, 130]. The use of plastic blends decreases

the performances because of incompatibility between most polymers, and often needs the addition of

coupling/compatibilizing agents. These agents are usually block copolymers or functionalized ones making

bridges between all the phases (physical agent) or react with them (chemical agent) [121]. Impurities such as

paper, metal, and glass have also a negative impact on the final properties and must be eliminated. Moreover,

some waste such as PVC needs to be treated first to inhibit any negative effect on the process or environment

[7]. Finally, materials such as coextruded or multi-material products cannot be separated and have to be recycled

as a blend or by energy recovery [121].

Polyethylene (PE) is one of the most common polymers because of its good properties such as high flexibility

and tenacity, wide availability, and a large range of different grades. Nevertheless, PE presents one major

drawback which is degrading very slowly in nature [36, 37]. To limit its environmental impact, it is necessary to

recycle polyethylene. Several works studied PE recycling and highlighted that its degradation mainly depends

on the processing conditions more than reprocessing unless a very high number of cycles is imposed [7, 74,

121, 123]. The short-term recycling of HDPE showed that, if suitable processing conditions are used, the

properties of the recycled materials stay close to the virgin material [131].

Twite Kabamba et al. showed that polyethylene degradation mainly depends on processing conditions. They

reported a reduction in molecular weights and lower chain entanglement leading to lower elastic properties of

the polymer. They observed that both chain scission and branching occurred due to the creation of macro-

radicals during processing. Branching is favored by low chain mobility, while chain scission is enhanced by high

chain mobility and mechanical stresses [132]. Similar trends were observed by Wagner and Scaffaro [133, 134].

Recently, Oblak et al. studied the effect of mechanical recycling on processability and mechanical properties of

high density polyethylene (HDPE) [74]. They found that increasing the number of reprocessing cycles induces

significant structural changes in the material. Those were seen through MFI measurements and mainly took

place during the first 30 extrusion cycles. Hardness and modulus were measured using nano-indentation and

showed deterioration of the material mechanical properties after ten reprocessing cycles. The authors suggested

51

that chain branching is the main mechanism during the first 30 extrusion cycles, and then chain scission is taking

over. Finally, they indicated the presence of crosslinking after 60 cycles.

Jin et al. studied the long-term closed loop recycling of low density polyethylene (LDPE) [36, 37]. They evaluated

the effect of one hundred successive extrusion cycles on the material flow properties to optimize the processing

conditions and the final performances. The results showed the presence of thermal degradation and gelation

due to simultaneous chain scission, reticulation, and branching confirming the trends seen in previous works.

For linear polymers such as HDPE, chain scission appears to be the dominant mechanism, while for branched

polymers such as LDPE, branching and reticulation are the main processes [7, 36, 37, 135, 136]. Variation in

the degree of crystallinity can also indicate the formation of structural irregularities with crosslinking [40, 137].

Finally, variations in melt flow index, relaxation time, and insoluble (gel) content can be related to changes in the

polymer structure [80].

The work of Kostadinova Loultcheva et al. studied the recycling of industrial post-consumer HDPE containers

[131]. It is one of the very few studies dealing with open-loop recycling. On the other hand, Yarahmadi et al.

studied the mechanical recyclability of PVC in building floor applications [128]. In the work of Kostadinova

Loultcheva et al., they observed the effect of the processing parameters (residence time, mono- or twin-screw

extruder, and presence of additives) and number of cycle on the chemical structure and morphology, as well as

mechanical and rheological properties. Degradation appeared to be strongly dependent on the polymer structure

and processing conditions due to a competition between chain scission, branching and/or crosslinking. Thermal

degradation seemed to promote chain scission, while the presence of shear stress at low temperature induced

branching. So depending on the processing conditions selected, one of these two phenomena is dominant [131].

Prediction of the degradation level is difficult because of the high number of parameters involved and possible

interactions between them. The final properties are directly related to the nature of the polymer, the type of

process, the thermomechanical history, and the environment [132]. So far, investigations in this field were mostly

devoted to understand the loss of performance [138, 139, 140]. This is why several works studied different types

of polymer, their behavior, the phenomena occurring while reprocessing, and the resulting loss of properties.

However, long-term recycling has rarely been studied and remains quite uncertain. Moreover, several studies

only considered the mechanical properties of the polymer, thus a lack of knowledge about other properties exists.

This work was undertaken to better understand these aspects. The behavior of high density polyethylene is

followed as a function closed-loop mechanical recycling (direct reprocessing) to determine the loss of properties.

The properties are evaluated for a high number of cycles (50) to evaluate the long-term recycling aspect.

52

III.2 Materials

The polymer used is a high density polyethylene (HDPE) supplied by Petromont (Canada). Its melting point is

126°C (see Figure III-1), its initial crystallinity is about 44%, and its density is 0.930 g/cm3. Its initial number

average molecular weight (Mn) id 14 kDa, with a weight average molecular weight (Mw) of 157 kDa and a melt

flow index (MFI) of 0.4 g/10 min (230°C, 2.16 kg).

Figure III-1: Virgin HDPE thermogram from DSC.

III.3 Experimental

III.3.1 Sample Production

The samples are produced by extrusion followed by injection molding. The polymer pellets are fed into a twin-

screw extruder and this step is considered as the first generation to simulate the general processing of a polymer

material. The material is then pelletized and extruded again to obtain each recycled generation. All the extrusions

are produced with the same processing conditions to make the comparison easier.

The extrusion is performed on a co-rotating 27 mm twin-screw extruder ZSE-27 (Leistritz, Germany). The

extruder has a L/D ratio of 40, ten zones from the feeder to the die, and is equipped with a circular die of 3 mm

in diameter. After preliminary trials, the total flow rate is fixed at 6 kg/h and the optimum operating conditions

found are a screw speed of 125 rpm and a temperature profile of 195°C for the first eight zones (from the feeder

53

to the die) and 200°C for the 9th zone and the die (zone 10). The extruded material is then cooled in a water

bath and finally fed into a granulator model 304 (Conair, Germany). The granulator speed was controlled to get

a final pellet diameter of 3 mm. For each generation (up to 50 here), the pellets are randomly collected (about 1

kg) to perform all the characterizations.

To limit the amount of work, only selected generations are fully characterized, including density,

thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), and gel permeation chromatography

(GPC) directly on the pellets. For mechanical characterization (tension and flexion) the samples are produced

by injection molding on a Nissei model PS60E9ASE. The parameters, determined by preliminary trials, are kept

constant for all generations. For each injection cycle, two type IV dog bones (ASTM D638) and two rectangular

bars (one short (12.40 x 3.05 x 78 mm3) and one long (12.40 x 3.05 x 125 mm3)) are obtained. The parameters

used for injection molding are presented in Table III-1, with the temperature given from the nozzle to the hopper.

Table III-1: Injection molding parameters.

Parameters Units Values

Temperature

1

°C

200

2 200

3 195

4 195

Mold 30

Pressure

1

%

80

2 80

3 50

Time Injection

s 4

Curing 25

Speeds

1

%

21

2 18

3 7

4 5

Shot size mm 35

Screw speed rpm 40

Back pressure % 10

54

For comparison, the initial (virgin) HDPE referred as “Generation 0” and the first extruded generation are

analyzed. Then, all even generation between 2 and 10 are characterized, and every 5th generation between 10

and 50 is analyzed giving a total number of 15 generations analyzed.

III.3.2 Physical Properties

Density is measured using a gas (nitrogen) pycnometer model Ultrapyc 1200e (Quantachrome Instruments,

USA) with a micro-cell and samples between 0.5 and 1.5 g of material. Density tests are conducted directly on

the pellets and on injected samples, with three repetitions for each type of material and sample.

Gel permeation chromatography (GPC) tests are conducted on a HT-GPC model 350 high temperature triple

detector (Array HT-DTA) from Viscotek (Malvern, UK). This model is composed of a differential refractive index

(RI) detector, a four-capillary differential viscometer, and a low angle light scattering (LALS) detector. Three PL

gel (mixed B LS) columns with a pore size of 10 µm and an individual column dimension of 300 mm by 7.5 mm

is used to perform the trials using 1,2,4-trichlorobenzene (TCB) at 140°C with at a flow rate of 1 mL/min and a

polymer concentration of about 4 mg/mL. The tests are repeated three times to get a better precision. Calibration

was performed with PS99K and PS235K with the same range of concentration and reference values directly

provided by the company. For each sample, three parameters are measured: the number average molecular

weight (Mn), the weight average molecular weight (Mw), and the intrinsic viscosity (IV). Considering the

correlation between those parameters, the polydispersity index (PDI = Mw/Mn) can also be deducted. The general

distribution shape and width are also analyzed.

Melt Flow Index (MFI) measurements are performed according to ASTM D1238 and ISO 1133 standards on an

automatic Flow Rate Timer. Trials are conducted at 190 and 230°C under a weight of 2.16 kg for generations 0

and 50, with three repetitions.

III.3.3 Thermal Properties

Differential scanning calorimetry (DSC) is conducted on pellets using a model DSC 7 (Perkin Elmer, USA) with

5 to 15 mg of material. The chamber is purged with N2 gas to provide a stable atmosphere. The material is

subjected to three ramps at 10°C/min. The first ramp (heating) allows to erase the thermal history of the material.

First, the polymer is heated from 50 to 200°C and maintained at 200°C for 10 min. Then, cooling is done from

200 to 50°C with a holding period at 50°C for 10 min. Finally, the polymer is reheated from 50 to 200°C. The

analysis is realized on the second heat-up thermogram. For each test, the temperatures of the various peaks,

their widths, and the crystallinity are reported.

Thermogravimetric analysis (TGA) tests are performed on a model Q5000 from TA Instruments on the pellets

(about 30 to 40 mg). The runs are performed under a nitrogen atmosphere (flow rate of 25 mL/min) using HT-

55

platinum pans. Each sample was subjected to a temperature ramp from 50 to 850°C at 10°C/min. Trials are

only performed on generations 0, 1, 10, 30, and 50 with three repetitions for each one. For each curve, four

parameters are evaluated: the final ratio of degraded mass, the temperature and the width of the mass derivative

peak, and finally the temperature of initial degradation, corresponding to the onset temperature of the derivative

peak.

III.3.4 Mechanical Properties

Tensile tests are carried out using a universal mechanical tester model 5565 (Instron, USA) with a 500 N load

cell. The geometry is a type IV dog bone samples (injection molded directly according to ASTM D638). The

measurements are carried out at a rate of 10 mm/min. For each sample, at least ten specimens are used. Six

parameters are analyzed: yield stress (σy), yield strain (εy), stress at break (σb), strain at break (εb), tensile energy

at break (TEB), and Young’s modulus (E).

Flexural tests are conducted on the same universal mechanical tester equipped with a three point bending fixture

(span of 60 mm). The crosshead speed is 2 mm/min with rectangular samples (12.40 x 3.05 x 78 mm3) according

to ASTM D790. For each generation, at least five specimens are characterized. All tensile and flexural tests are

carried at room temperature.

III.4 Results

III.4.1 Density

Samples densities are measured to check any possible modification in the compaction (packing) of the polymer’s

chains. As expected, for both pellets and injected samples, no significant change during recycling is observed.

In all cases, the density remains constant and similar to the density of the virgin polymer (0.930 g/cm3).

III.4.2 Gel Permeation Chromatography

GPC results are indicative of any changes in the molecular weight distribution due the thermo-mechanical history

of the polymer. Figure III-2 shows a decrease of high molecular weights with an increase of medium ones,

suggesting that chains scission occurs mainly in long chains, but was limited since no significant increase in

small chains (low molecular weight) is observed for the range of conditions tested. As seen in Figure III-3 and

Figure III-4, the weight average molecular weight (Mw) decreases significantly with recycling showing a loss of

24%, while the number average molecular weight (Mn) does not show any trend (almost constant within

experimental uncertainty). These results indicate again that chain scission mainly occurs in the longest chains

to produce medium chains. As a result, this leads to a reduction of the polydispersity index by about 20% as

reported in Figure III-5. Figure III-6 shows that the intrinsic viscosity decreases steadily with recycling, thus

56

indicating higher chain mobility and confirming that chain scission occurs during processing. The same trends

were observed by Duigou et al. [97] and Oblak et al. [74].

Figure III-2 : Molecular weight distribution for generations 0 (PG0) and 50 (PG50).

Figure III-3: Number average molecular weight as a function of generation number.

57

Figure III-4: Weight average molecular weight as a function of generation number.

Figure III-5: Polydispersity index as a function of generation number.

58

Figure III-6: Intrinsic viscosity as a function of generation number.

Most of the works available in the literature reported a decrease of molecular weight with recycling, but not all of

them agree on which molecular weight [36, 37, 92, 131, 141, 142, 143, 144]. While some works do not specify

which molecular weight is analyzed, both Jin et al. and Twite Kabamba et al. reported a significant decrease of

the number average molecular weight [36, 37, 132], while here the decrease is mainly observed for the weight

average molecular weight, in agreement with the trend reported by Parmar et al. [145] and Oblak et al. [74].

However, all agree that a peak shift towards lower values is due to chains scissions related to mechanical and

thermo-oxidative degradation of the polymer during processing. On the other hand, Jin et al. and Shyichuk and

White reported a modification of the molecular weight distribution in both shape and width for low density

polyethylene (LDPE) and polystyrene (PS), respectively [36, 37, 146]. This modification is attributed to

crosslinking and chain scission taking place simultaneously in the polymer. Here, no clear change can be

detected in the distribution and both parameters do not seem to be significantly affected by recycling. This can

be associated to the different polymer used as HDPE shows a lower tendency to form crosslink and branching

than LDPE and PS, thus presenting mainly chain scission without crosslinking. Nevertheless, there is also the

effect of different processing parameters used, especially temperature which is much lower here (200°C) than

in the work of Jin et al. (240°C), thus producing less thermo-oxidative degradation. This was confirmed by FTIR

tests (results not presented) revealing no significant oxidation in the HDPE. The idea that mainly chain scission

occurs is also supported by the lack of insoluble fraction in the solutions prepared for the GPC tests. Jin et al.

also observed that the fraction of insoluble material increased with increasing recycling cycle number, while no

59

insoluble fraction is found for all the solutions prepared here [36, 37]. This brings more evidences that

crosslinking does not significantly occurs during our recycling trials. However, the results have to be taken as

relative to the specific polymer, equipment (screw elements design, length, configuration), as well as processing

conditions (temperature profile, flow rate, die design, etc.) used.

III.4.3 Melt Flow Index

Between generation 0 and 50, the melt flow index increases with recycling from 0.2 g/10 min to 0.5 g/10 min at

190°C, and from 0.4 g/10 min to 1.0 g/10 min at 230°C. This is consistent with most studies reported in the

literature indicating that MFI increases with recycling which represents a viscosity decrease related to chain

scission [123, 147, 148, 149, 150, 151]. This also supports the fact that no crosslinking occurs in the polymer

since Jin et al. and Oblak et al. indicated that MFI decreases when crosslinking takes place due to the partial

formation of a crosslinked structure [36, 37, 74].

III.4.4 Thermogravimetric Analysis

For these trials, none of the four parameters evaluated appear to be significantly modified by recycling, as they

present a maximum variation of 0.4% to 6.8%, which is within experimental uncertainty and thus represents

negligible variations.

III.4.5 Differential Scanning Calorimetry

The degree of crystallinity Xcr (%) is estimated using equation (1) where ∆H100 = 293 J/g is the heat of fusion

of the 100% crystalline material (HDPE), and ∆Hf is the measured enthalpy value [36]:

Xcr (%) = ∆Hf /∆H100 x 100% (1)

The DSC curve shows no significant difference in crystallinity nor crystallisation and melting points with recycling.

For all 50 cycles, the crystallinity is around 44% and the melting point is almost constant at 126°C, with a peak

width (Tonset - Tend) of about 50°C. All these parameters present a maximum variation of 4.0% to 7.8%, which is

again within the experimental uncertainty, and thus do not represent significant variation. This trend is quite

similar to the one observed by Jin et al. [36, 37]. Nevertheless, some studies reported an increase of crystallinity,

melting, and crystallisation points with recycling [131, 97, 147, 151]. However, this increase is generally attributed

to the presence of crosslinking during processing. Here, no significant crosslinking is observed during recycling

and confirms the negligible variation observed.

III.4.6 Tensile Properties

Recycling appears to have an important effect on the tensile properties and Figure III-7 presents typical stress-

strain curves for selected generations. All the parameters extracted for these curves are affected by

60

reprocessing. However, the properties at break are the most affected. As seen in Figure III-8, stress at break

(σb) increases by 90% after 25 recycling cycles, reaching 16 MPa, before showing a slight decrease and then a

stabilization around 14 MPa for higher generations. However, more cycles would be necessary to determine if

the stress at break remains constant or continue decreasing with further recycling. The strain at break (εb) is the

most affected parameter, with a sharp increase of 500% after 50 cycles, i.e. a final value six times higher than

the initial value as reported in Figure III-9. In Figure III-10, yield stress (σy) shows a decrease of 29% between

the first and the 50th generation. On the contrary, yield strain (εy) does not change significantly as it slightly

increases from 10% up to the 25th generation, before decreasing slowly from 8%, getting very close to its original

value (Figure III-11). Once again, more generations should be conducted to improve the analysis of these trends.

In Figure III-12, the Young’s modulus (E) does not show any clear trend, even if a slight decrease of about 5%

over the 50 generations can be observed, which is within the experimental uncertainty again. Finally, as seen in

Figure III-13, energy at break increases sharply (240%) from 8.3 J for neat HDPE to 28.3 J for the 50th generation.

These variations can be attributed to the weakening of the entanglement network and higher chains mobility

associated with chain scission, leading to a loss of rigidity and a gain of elasticity (elongation). Finally, the

recycled polymers present a more ductile behavior than the virgin material, but lower global performances with

lower strength and rigidity.

Figure III-7: Typical tensile stress-strain curves for different generations.

61

Figure III-8: Stress at break as a function of generation number.

Figure III-9: Strain at break as a function of generation number.

62

Figure III-10: Yield stress as a function of generation number.

Figure III-11: Yield strain as function of generation number.

63

Figure III-12: Young's modulus as a function of generation number.

Figure III-13: Energy at break as a function of generation number.

Most studies on the mechanical properties of recycled polymers reported opposite trends with increasing stress

at yield and Young’s modulus with decreasing yield strain and break properties with recycling [132, 97, 141, 144,

147]. Generally, these trends can be attributed to crystallinity and possible crosslinking. Nevertheless, the HDPE

64

here does not show any increase of crystallinity nor crosslinking during recycling, but only chains scission. This

leads to higher chain mobility and to lower degree of entanglement. In this case, ductility increases while the

other properties associated with rigidity decreases.

III.4.7 Flexural Properties

Flexural modulus (Eb) is less influenced by recycling than tensile modulus. However, it slightly decreases with

recycling. As seen in Figure III-14, the flexural modulus gradually decreases after 50 generations (12% of the

neat HDPE value). This loss of performance is again associated with chain scission in the polymer, leading to

fewer entanglement with recycling. This is consistent with the observations made above.

Figure III-14: Flexural modulus as a function of generation number.

III.5 Conclusion

In this work, high density polyethylene (HDPE) was processed and recycled multiple times by extrusion to

simulate a long-term closed-loop recycling process by direct reprocessing. The polymer was subjected up to 50

extrusion cycles under constant processing conditions. After each cycle, some material was collected and

conditioned for various characterizations. Polymer degradation was then characterized in terms of physical,

thermal, and mechanical properties. From the results obtained, several conclusions can be made.

65

Firstly, recycling did not seem to significantly affect the physical and thermal properties of the material as

negligible changes were observed in DSC, TGA or density. On the other hand, chains scissions led to a reduction

of the chain length and thus to higher chain mobility. This induced lower weight average molecular mass and

melt viscosity, leading to a higher MFI. Finally, a general decrease was observed in mechanical properties, with

the exception of tensile properties at break. At the end of the 50 cycles, the stress at break (σb) increased by

75% while the strain at break (εb) increased accordingly reaching a value six time higher than its initial one. The

yield strain decreased by 30% and the energy at break increased by 240%.

All these variations can be attributed to the higher mobility of polymer chains, as well as lower entanglement and

lower molecular weight due to chain scission taking place in the polymer during processing. No evidence of

crosslinking or branching were found in this work. For some mechanical characterizations (tension and flexion),

more trials should be conducted as to reduce dispersion while more generations should be realized to confirm

the trends. Moreover, prediction of the polymer degradation during processing is difficult because of the number

of parameters involved including: type of polymer, processing conditions, environment characteristics, and

thermal properties. The results obtained depend strongly on the selected processing conditions and can be

improved by the addition of a stabilizer to avoid chain scission and the lowering of the molecular weight, thus

limiting the resulting loss of performances and degradation. This loss can also be limited by using proper

conditions including lower residence times and temperatures. Moreover, it is possible to promote some

properties at the cost of others. Nevertheless, the results presented and the conclusions drawn upon them have

to be taken with care since they are only valid for the processing conditions used as well as for the polymer

selected.

Acknowledgements

Financial support from the Natural Sciences and Engineering Research Council of Canada (NSERC, Canada),

the Research Center for High Performance Polymer and Composite Systems (CREPEC, Quebec), the Quebec

Research Funds – Nature and Technologies (FQRNT, Québec), the Quebec Center for Functional Materials

(CQMF, Quebec), the Research Center for Advanced Materials (CERMA, Quebec), and the Renewable

Materials Research Centre (CRMR, Quebec) is acknowledged for this work. The technical help of Mr. Yann

Giroux is highly appreciated for the experimental work.

66

Chapter IV. Long-term closed-loop recycling of high density polyethylene/flax composites.

Résumé

Ce travail investigue la perte de performances et l’aptitude au recyclage de composites à fibres naturelle soumis

à un procédé de recyclage long terme en boucle fermée. Les composites sont produits à partir de fibre de lin et

de polyéthylène haute densité, et, pour la série avec agent couplant, avec polyéthylène greffé d’anhydride

maléique. Ces composites sont soumis à 50 cycles d’extrusion sous des conditions de mise en œuvre

constantes. Les résultats indiquent que l’ajout de fibres augmente la rigidité mais diminue les propriétés en

élongation. Le premier cycle de mise en œuvre entraîne une importante diminution de la longueur des fibres et

une modification des distributions de poids moléculaires, indiquant que l’ajout de fibres favorise les ruptures de

chaînes moléculaires et que les ruptures de fibres ont essentiellement lieu lors de la mise en œuvre initiale.

L’effet du recyclage sur les performances du composite est beaucoup moins significatif, sauf dans le cas des

propriétés mécaniques. Aucune variation significative n’est en effet observée lors des mesures de masse

volumique et de résistance à l’impact, ainsi que lors des essais en calorimétrie différentielle à balayage, en

analyse thermogravimétrique et en chromatographie à perméation de gel. Toutefois, les propriétés mécaniques

sont fortement affectées par la remise en œuvre et la plupart d’entre elles augmentent avec le recyclage. L’ajout

d’agent couplant améliore les propriétés des composites, mais cet effet tend à disparaître avec le recyclage.

Ces tendances sont associées à l’équilibre entre les ruptures de fibres et de chaînes macromoléculaires se

produisant dans le matériau lors du recyclage, et à la meilleure homogénéité de matériau qui en résulte

(meilleure distribution des fibres). Les résultats indiquent que le recyclage long-terme des composites est

possible puisque leurs performances globales demeurent acceptables.

Mots-clés: Recyclage, composite, extrusion, polyéthylène, lin, morphologie, propriétés mécaniques.

67

Abstract

This work investigates the loss of performance and the recyclability of natural fiber composites for a long-term

closed-loop process. Composites based on flax fibers and high density polyethylene (HDPE) are subjected up

to 50 extrusion cycles under constant processing conditions with or without maleic anhydride grafted

polyethylene (MAPE) as a coupling agent. The results show that the addition of fiber increases the rigidity but

decreases the elongation properties. The initial processing cycle leads to an important decrease of the fiber

length, and modification of the molecular weight distributions, thus indicating that the addition of fiber enhance

chain scission and that fiber breakup mainly happens during the initial processing. The effect of recycling is

much less significant, except for the mechanical properties. Negligible variations are observed for density,

differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), gel permeation chromatography

(GPC) and impact results. On the contrary, the mechanical properties are strongly affected by recycling as most

of them increase with recycling. The addition of a coupling agent improves the composite properties, but this

effect disappears with recycling. These trends are associated to a balance between fiber breakup and

macromolecular chain scission compared to more homogeneous materials (better fiber distribution) taking place

in the materials during recycling. The results show that long-term recycling of composites is possible as their

overall performances remain acceptable.

Keywords: Recycling, composite, extrusion, polyethylene, flax, morphology, mechanical properties.

N. Benoit, R. González-Núñez, and D. Rodrigue. Long-term recycling of high density polyethylene/flax fiber

composites and characterization of its closed-loop degradation, Progress in Rubber, Plastics and Recycling

Technology, submitted, 2016.

68

IV.1 Introduction

Polyethylene is used in a wide range of applications such as transportation, packaging, sports, leisure and

domestic appliances due to its main advantages such as a simple structure, a relatively low cost, interesting

mechanical properties, long term stability, good recycling behavior, but also presents a slow decomposition rate

[4, 7, 152]. It is thus essential to recycle the material at the end of its life. Currently, only 30% of the total volume

of plastic waste is recycled, 39% is energetically revalorized, while the remaining 31% is simply landfilled [4, 7,

152, 153, 154]. Considering these values, some efforts are necessary to revalorize these materials. Due to their

biodegradable and bio-sourced origin, natural fibers have been of high interest. They appear to be an interesting

alternative to neat plastics, solid woods and inorganic composites such as glass, aramid and carbon in several

fields such as sports, thermal and acoustic isolation, energy, furniture and domestic goods, transportation,

automotive, navigation and aeronautics [16, 23, 24, 56, 155, 156, 157, 158, 159, 160]. They are widely available

with high specific mechanical properties [31, 40, 80, 110, 132, 156]. They are also lightweight compared to

inorganic composites, thus lowering transportation costs, and less abrasive and damageable for tooling and

equipment [80, 155, 159]. However, natural fibers have some drawbacks reducing their attractiveness and

limiting their use such as high variability, poor resistance to moisture and fire, lower durability, limited processing

temperatures, tendency to agglomerate during processing and incompatibility with most thermoplastics [71, 80,

131, 132, 155, 156, 159, 161, 162, 163]. This is why coupling agents are added to improve adhesion within the

composite and to preserve the overall quality of the materials [132, 155, 132].

Recycling of composites is fundamental in the actual social and political context as it allows to reduce the amount

of waste generated while limiting the amount of new material consumed. If recycling is the most advantageous

method from an environmental point of view, it is however often neglected as it often requires several steps

[152]. Materials sorting as well as the presence of impurities and contaminants are the most important issues

leading to lower properties of the recycled materials [79, 131, 153, 164]. But when plastics wastes are collected,

different options exist. In some specific cases, the product can be directly reused or recycled. When none of

these options are possible, it can be landfilled even if this option is more and more restricted by law [71]. This is

why different recycling techniques are available which are divided in three categories: thermal recycling,

chemical recycling and mechanical recycling [79]. The best method depends on the material and its state.

Thermal recycling is based on the decomposition of the material at high temperature with energy recovery.

Chemical recycling is usually done by depolymerization or dissolution of the matrix using solvents. The by-

products are monomers or chemical products with fibers. However, this method leads to the production of

chemical wastes reducing its attractiveness [56, 79, 73]. In mechanical recycling, the materials are grinded,

remelted and reprocessed. The high pressure, temperature, and stresses (shear and elongation) undergone

69

during these steps lead to irreversible changes in the polymer structure and properties. A complete overview on

recycling issues and technologies can be found in Henshaw et al. [79].

Polymer recycling has been studied by several authors and all of them reported a relation between the

processing conditions and material degradation [36, 37, 131, 161]. The results also showed that recycled plastics

maintain the majority of their properties, except in the case of intensive or long-term recycling. For example, the

short-term recycling of polyethylene was shown to produce a material having properties very similar to the virgin

resins [36, 37, 131, 161]. But two works considered some aspects that are often left out in most studies and

should be taken into account. Kostadinova Loultcheva et al. studied the open-loop recycling by extrusion of

postconsumer high density polyethylene industrial containers [131]. They reported that a strong relation exists

between processing conditions, thermomechanical degradation and final performances of the materials.

Recycling led to higher crystallinity, rigidity, melting point and enthalpy, while molecular weight and elongation

at break decreased. All these variations were linked to two phenomena occurring simultaneously in the polymer:

chain scission and branching. The final properties of the materials depend on the relative extent of these

phenomena. Jin et al. focused on the intensive mechanical recycling of high density polyethylene (HDPE) by

conducting up to 100 successive extrusion cycles [36, 37]. The results showed a slight decrease of the final

properties. They also indicated the existence of chain scission, but instead of branching, the authors assumed

that crosslinking occurred, especially after 40 cycles since the presence of an insoluble fraction (gel) was

detected in the polymer. However, the recyclability of low density polyethylene (LDPE) seems to be maintained

up to 40 cycles. However, since these trends are different and not on the same polymer, more work must be

performed to better understand the phenomena controlling the recycling behavior of polymers.

Although natural fiber composites production is very well known, the recycling of such composites is much less

investigated [38, 39, 40, 80, 155, 156, 157, 158, 159, 160]. The actual recycling techniques for such materials

are often expensive and the recycled material is of poor quality, making it hard to find any suitable market for

them [56]. Most composites are currently burnt for energy (fuel) or fiber recovery. It is thus essential to know

how the properties of such composites are modified by recycling, and how the presence of fibers modifies the

matrix degradation to favor mechanical recycling. This should be done through two different approaches. First,

with the development of novel, more efficient and lower cost recycling methods, as well as better separation

technologies. Then, by the development of new and better recyclable composite materials [56]. Several studies

on the properties of mechanically recycled natural fiber composites can be found [40, 159], but much less

information is available on their open-loop or intensive recycling. However, all these works concluded that fiber

dimensions decrease in the first extrusion step and mainly immediately after their introduction in the molten

matrix. As expected, the extent of degradation depends on the processing conditions [40, 85, 100, 101, 146].

For example, Twite Kabamba et al. produced up to 10 cycles of closed-loop mechanical recycling of LDPE/birch

70

fiber composites by extrusion under constant conditions [40]. They showed that the effect of recycling is mainly

on parameters such as fiber dimensions, as well as matrix number average molecular mass (Mn), crystallinity

and viscosity. Wang et al. studied the effect of processing parameters on the properties of wood/high density

polyethylene composites with coupling agent [31]. They concluded that the final properties of the composites

are a combination of complex interactions related to the equipment characteristics, processing conditions,

constituents and the interfacial properties. They also suggested that an adequate choice of parameters, including

screw rotational speed, throughput rate and barrel temperature profile is essential to limit the thermal degradation

and color change (darkening) of the wood during processing. Moran et al. focused on the effect of multiple

recycling (up to five cycles) under constant processing conditions to analyze the mechanical properties of

polypropylene (PP)/flax fiber composites [30]. The two main parameters influencing the composite properties

were the elastic modulus and the mixing quality. They also concluded that even if the recycled material suffered

from chain scission, fiber breakup and global degradation, the mechanical properties were negligibly affected

and the material can still be used for industrial applications.

The behavior and the performances of natural fiber composites have been the subject of several studies, and

are thus partially known [16, 38, 39, 40, 155, 156, 157, 158, 159, 160]. However, most of the studies on

composites recycling focused on chemical recycling through matrix dissolution and fiber recovery processes [30,

31, 40, 80, 110, 132, 163]. So investigations on the mechanical recycling of composites are still limited. Most of

the works available in the literature reported only low properties variations and most authors concluded that

natural fiber composites with coupling agent have great aptitudes toward recycling. However, none of them

considered a number of cycles higher than 10 [40, 159, 163], nor open-loop recycling. Some efforts still need to

be done on these two aspects. In particular, it would be of great interest to confirm the long term recycling trends

for composites or to determine any specific behavior for the recycling of such materials. This work aims to fill

this by investigating the closed-loop production and recycling of flax fibers/HDPE composites up to 50 times

under constant extrusion conditions. The main objective is to determine the effect of coupling agent addition on

the long-term recycling of these composites. This is done by following different properties (morphological,

physical, thermal and mechanical) as a function of recycling cycle number.

71

IV.2 Materials

The polymer matrix is high density polyethylene (HDPE) grade C-7525 supplied by Petromont (Canada), while

the flax fibers are provided by Biolin Research Inc. (Saskatoon, Canada). Maleic anhydride grafted polyethylene

(MAPE) Epolene C-26 (Westlake Chemical Corporation, USA) is used as a coupling agent.

As reported in our previous work, the HDPE has a melting point of about 126°C and a density of 0.93 g/cm3 [75].

It also has a melt flow index (MFI) of 0.4 g/10 min at 230°C, a number average molecular weight (Mn) of 14 kDa,

and a weight average molecular weight (Mw) of 157 kDa. According to the provider, the MAPE used has a melt

flow rate (MFR) of 8.0 g/10 min and a molecular weight of 65 kDa. The flax fiber was grinded and sieved to get

a range of particle sizes between 250 µm and 1 mm. The initial aspect ratio (L/D) and length (L) distributions

were measured by optical microscopy for a total of 200 fibers and the results are presented in Figure IV-1 and

Figure IV-2 with an average L/D of 3.98 and an average length of 2.19 mm. The residual humidity of the fiber

was measured through a thermogravimetric analysis. Two types of temperature cycles were performed: one

isotherm at 100°C and one ramp (from ambient to 100°C, during 10 min) followed by a 5 min isotherm at 100°C.

An average residual humidity value of 7% was found. Finally, the density of the dried fiber was around 1.34

g/cm3.

Figure IV-1: Initial fiber L/D aspect ratio range distribution.

72

Figure IV-2: Initial fiber average length range distribution.

IV.3 Experimental

IV.3.1 Sample Production

Two series of composites were produced, with or without MAPE to evaluate the effect of coupling agent addition

on the degradation and the recyclability of these composites. The composites with coupling agent are noted as

CA, while the composites without are noted as CS.

The samples are produced in two steps: a compounding step by extrusion followed by a processing step by

injection molding. The extrusion is done on a co-rotating 27 mm twin-screw extruder ZSE-27 (Leistritz, Germany)

with a L/D ratio of 40 and ten zones from the feeder to the die. The extruder is equipped with a circular die of 3

mm of diameter. For comparison purposes, the conditions used are the same as in our previous work on HDPE

recycling (matrix alone): screw speeds (extruder and side stuffer) of 125 rpm and temperature profile set to

195°C for the first eight zones (from the feeder to the die) and to 200°C for the 9th zone and 10th zone (die) [75].

73

For the first compound, the polymer pellets are fed in the extruder through the main feeder (zone 1), while the

flax fibers are fed through the side stuffer (zone 4). The fibers are previously dried overnight at 80°C to eliminate

humidity and some volatiles. For the composite with MAPE, the coupling agent is directly blending with the

polymer (before feeding), at a concentration of 1.5% by weight. The composite formulation is obtained by

adjusting the outputs of both fiber and polymer (or polymer/MAPE) to get a fiber concentration of 15% by weight.

The global output is set to 6 kg/h. The compound is then cooled in a water bath and fed to a Conair 304 pelletizer

(Conair, Germany) with a controlled speed to get pellets having a 3 mm diameter. To produce the next

generations (recycling), the material is dried again and extruded several times to get the different cycles. Each

new cycle of extrusion, drying and pelletizing is considered as a generation. To compare the results of this work

to our previous work on the matrix recycling (neat polymer) [75], all processing and characterization parameters

were the same. In total, up to 50 generations were produced.

For time and cost purposes, only selected generations are characterized. As a basis of comparison, the virgin

matrix, referred here as “Generation 0” is studied, as well as both composites after their first extrusion cycle

(here called “Generation 1”). Then, every 2nd generation between 2 and 10 are characterized. Finally, every 5th

generation between 10 and 50 are studied giving a total number of 15 samples. For each material and generation

studied, one kilogram of pellets is randomly and progressively collected in small quantities at different production

times to get a random sampling. These samples are all taken after the stabilization of the processing conditions

(after start-up).

After their extrusion, the recycled materials are then shaped into samples by injection and compression molding.

Although some characterization such as density, thermogravimetric analysis (TGA), differential scanning

calorimetry (DSC) and gel permeation chromatography (GPC) can be conducted directly on the pellets, others

like tension, bending and impact need specific sample geometries. These samples are molded with an injection

molding machine Nissei model PS60E9ASE (60 tons). For each injection cycle, two type IV dog bone (ASTM

D638) and two rectangular parts (one short (78 x 12.7 x 3.17 mm3) and one long (127 x 12.7 x 3.17 mm3)) are

obtained. The parameters used for injection molding are presented in Table IV-1, with the temperature given

form the nozzle to the hopper.

74

Table IV-1: Injection molding parameters.

Parameters Units Values

Temperature

1

°C

195

2 200

3 195

4 190

Mold 30

Pressure

1

%

80

2 80

3 50

Time Injection

s 4

Curing 25

Speed

1

%

21

2 18

3 7

4 5

Shot size mm 35

Screw speed rpm 40

Back pressure % 10

IV.3.2 Physical Properties

The density of the materials is measured using a gas (nitrogen) pycnometer model Ultrapyc 1200e

(Quantachrome Instruments, USA) with a micro-cell. For the composites, the samples are previously dried

before testing. To get more information, density tests are conducted on the pellets and injected molded samples

with 3 repetitions.

Gel permeation chromatography (GPC) tests are conducted at 140°C on a HT-GPC model 350 high temperature

triple detector (Array HT-DTA) from Viscotek (Malvern, UK). This model is composed of a differential refractive

index (RI) detector, a four-capillary differential viscometer and a low angle light scattering (LALS) detector. Three

PL gel (mixed B LS) columns with a pore size of 10 µm and an individual column dimension of 300 mm by 7.5

mm is used to perform the trials using 1,2,4-trichlorobenzene (TCB) at 140°C with a flow rate of 1 mL/min and

a HDPE concentration of about 4 mg/mL Calibration of the equipment is done with PS 99K and PS 235K within

the same range of concentrations using the reference values directly provided by the producer. Three

75

replications are done for better precision. For each test, three parameters are measured: the number average

molecular weight (Mn), the weight average molecular weight (Mw), and the intrinsic viscosity (IV). Considering

the correlation between the first two parameters, the polydispersity index (PDI = Mw/ Mn) can also be reported.

The distribution curve (width) is also analyzed. The composites were dissolved in TCB and then filtered to

eliminate the fiber as to not clog the columns.

IV.3.3 Thermal Properties

Differential scanning calorimetry (DSC) is conducted directly on pellets using a Perkin Elmer DSC7 using

between 5 to 10 mg of material under a N2 atmosphere. The sample is subjected to three temperature ramps

with a heating/cooling speed of 10°C/min. First, the sample is submitted to a heat-up ramp from 50 to 200°C,

and then maintained at 200°C for 10 min. Then, a cool-down ramp from 200 to 50°C is applied, before a 10 min

isotherm at 50°C. Finally, the material is reheated from 50 to 200°C. The analysis was performed on the second

heat-up thermogram. For each test, the temperature of the various peaks with their width at mid height are

reported. Three repetitions are conducted for each test. The degree of crystallinity (Xcr) was estimated using

equation (1) with ∆H100 (%) = 288 J/g being the melting enthalpy of 100% crystalline polyethylene [40] and x is

the weight fraction of flax fibers (15%):

Xcr = ∆Hf (%) / [(1-x)∆H100 (%)] (2)

Thermogravimetric analysis (TGA) is performed on a model Q5000 IR (TA Instruments, USA). The samples (30-

40 mg) are taken from the pellets and the composites are previously dried before testing. The tests are performed

under a nitrogen atmosphere with a flowrate of 25 mL/min and HT-platinum pans. Each sample was subjected

to a temperature ramp from 50 to 850°C at a heating speed of 10°C/min. For time and cost considerations, due

to the very low variations of the parameters with recycling, trials are only performed on a limited amount of

generation: 0, 1, 10, 30 and 50. For each of these generations and materials, three repetitions are conducted to

get the peak temperatures from the derivative weight signal and the initial (onset) degradation temperature

determined as the 5% weight loss.

IV.3.4 Mechanical Properties

Mechanical tests are carried out using a universal mechanical tester model 5565 (Instron, USA). All mechanical

tests are performed at room temperature.

For tensile properties, a 500 N load cell was used for type IV dog bone samples (ASTM D638). The

measurements are carried out at 10 mm/min. For each generation, at least ten specimens are used to get the

yield stress (σy), yield strain (εy), stress at break (σb), strain at break (εb), tensile energy at break (TEB) and

Young’s modulus (E).

76

Flexural tests are conducted on the same universal mechanical tester equipped with a three point bending fixture

(60 mm span) and a 50 N load cell. Rectangular samples (78 x 12.7 x 3.17 mm3) produced by injection molding

(ASTM D790) were used. The crosshead rate is 2 mm/min and only the bending modulus (Eb) is considered.

For each generation analyzed, a minimum of five specimens are tested.

The impact properties are evaluated through Charpy impact trials (ASTM D6110). These tests are carried out

on a pendulum tester model Impact 104 (Tinius Olsen, USA). The pendulum is able to supply an energy of 15 J

on rectangular samples having dimensions of 12.4 x 3.05 x 33 mm3. For each material, at least 10 samples were

tested and V-notched on an automatic notcher model ASN 120m (Dynisco, USA). All the trials were done at

room temperature to report on the impact strength (Fbk).

IV.3.5 Morphology

The morphology of the fibers is analysed by optical microscopy (stereo-microscope SZ-PT 126x, Olympus). The

average length (L), diameter (D) and the aspect ratio (L/D) distributions are measured. In order to get a good

estimation of these parameters, a minimum of 200 fibers were measured. This observation is performed on the

fibers obtained after matrix dissolution like for the GPC samples. After filtration, the solution is used to produce

GPC sample, while the fibers collected on the filter are dried and used for morphological observation.

The fracture feature and composite morphology are also observed with a scanning electron microscope (SEM)

JSM-840A (JEOL, Japan). To get a better (brittle) fracture surface, the samples are first immersed in liquid

nitrogen before breaking. The exposed surfaces are then coated with a thin layer of gold using a sputter coater

and the microstructure of the exposed surfaces is examined in terms of fiber size, distribution repartition and

orientation. Any possible coupling will also be studied.

IV.4 Results and Discussion

IV.4.1 Density

Samples densities are measured to see the effect of fiber addition and to observe any possible compaction in

the composites. The initial density of the composite is around 1.00 g/cm3. As expected, this value ranges

between the values of the matrix (0.93 g/cm3) and the fibers (1.34 g/cm3). For both pellets and injection molded

samples, no significant change in density was observed during the recycling as the density varied between 0.964

and 1.032 g/cm3 which is close to the density of the virgin composite (1.00 g/cm3).

IV.4.2 Gel Permeation Chromatography

First, an important decrease of both composites weight average molecular weight (Mw) can be noted after fiber

addition as reported in Table IV-2. For the CS composites, Mw shows a 25% decrease between the CSG1

77

composite and the virgin matrix (Table IV-2), while the number average molecular weight (Mn) is almost constant.

This indicates an important degradation of the polymer chains during the first processing of the CS composites.

On the contrary, the results show that the CA composites present an increased Mw compared to the neat matrix

with a similar Mn. This behavior can be attributed to the addition of the coupling agent in the CA composites.

Then, it can also be noted that no clear trend can be observed for both composites towards recycling. As seen

in Figure IV-3 and Figure IV-4, Mn and Mw do not show any particular trend along further recycling. No significant

changes are also observed for the polydispersity index. This confirms that chain scission happens mainly during

the first recycling cycle and differs from the trend observed in our previous study on the neat matrix recycling

where Mn continuously decreased with recycling indicating that chain scission took place during recycling [75].

When comparing the molecular properties for the virgin matrix and the first generation of the matrix and the

composites (Figure IV-3 to Figure IV-7), it can be seen that, except for Mn, all the other properties are higher for

the matrix than for the composites. However, this difference is only significant for the first cycle, thus confirming

that chain scission mainly happened during the initial (first generation) processing of the composites. This trend

is similar as observed for wood fiber composites [40, 85, 100, 101], and confirms that the first cycle is the most

significant one.

Table IV-2: Average molecular weights for the matrix and the composites.

Sample Mn (kDa) Mw (kDa)

PG0 14 157

CSG1 14 118

CAG1 13 176

78

Figure IV-3: Average molecular weights as a function of generation number for the CS composites.

Figure IV-4: Average molecular weights as a function of generation number for CA composites.

79

Figure IV-5: Number average molecular weight (Mn) as a function of generation number.

Figure IV-6: Weight average molecular weight (Mw) as a function of generation number.

80

Figure IV-7: Polydispersity index as a function of generation number.

Most of the works available in the literature on composite recycling rarely investigated molecular properties. In

two of their works, Twite Kabamba et al. studied the effect of recycling on the matrix molecular weight and

branching [40, 80]. They found that Mn and Mw decreased by 30 and 17%, respectively. Soccaligame et al.

reported that mechanical recycling by injection molding of wood/polypropylene composites led to a decrease of

Mn, while Mw remained constant [90]. No clear modification of the shape or width of the molecular weight

distribution can be observed with recycling, thus confirming that scission is more important than crosslinking and

branching, contrary to Jin et al. [36, 37] and Shyichuk and White in their work on linear low density polyethylene

(LLDPE) and polystyrene (PS) [146]. This is similar to the trend observed for the matrix [75]. As explained in our

previous work, this can be attributed to the different polymers used or to different processing conditions and

equipment configurations. In each case, these results have to be considered carefully as they mainly apply for

the processes and materials used in the study.

IV.4.3 Thermogravimetric Analysis

Figure IV-8 and Figure IV-9 present typical TGA curves obtained for the matrix and the composites, respectively.

The derivative weight signal shows only one peak for the matrix (Figure IV-8), but two peaks can be noted for

the composites due to the presence of flax fiber (Figure IV-9). As seen in Figure IV-10 no significant differences

are observed with recycling as the maximum variation of the derivative weight peak temperature is 5°C for both

81

CS and CA composites, which is within experimental error. Similarly, as seen in Figure IV-11, no significant

modification of the onset degradation temperature can be observed as the maximum variations are close to the

experimental error (7°C and 18°C for the CS and the CA composites, respectively).

Figure IV-8: TGA results for the matrix.

82

Figure IV-9: TGA results for the composites.

Figure IV-10: Peak temperature obtained from the weight derivative curves for the matrix

and the composites as a function of generation number.

83

Figure IV-11: Onset degradation temperature for the matrix and the composites as a function of generation number.

IV.4.4 Differential Scanning Calorimetry

The DSC curves show no significant differences in the crystallinity nor the crystallisation or melting points with

recycling. For all the 50 cycles, the melting point is around 126°C (Figure IV-12) with a melting peak width at

mid-height of about 8°C (Figure IV-13), and the crystallinity is around 45% (Figure IV-14: Crystallinity as a

function of generation number.Figure IV-14). These results represent maximum variations of 0.1°C, 0.3°C and

1.5% for the CS composites, and 1.3°C, 1.2°C and 2.0% for the CA composites, which are all within the

experimental error. These results are similar as reported by both Moran et al. [30] and Srebrenkoska et al. [98].

Most of the studies indicate a constant melting point but an increasing crystallinity with recycling [163].

Nevertheless, this increase is generally attributed to possible crosslinking during processing, as well as to

increased chain mobility, leading to polymer structure reorganization. Here, the constant crystallisation

parameters are additional information to support the conclusion that no significant crosslinking occurred during

recycling.

84

Figure IV-12: Melting point as a function of generation number.

Figure IV-13: Melting peak width as a function of generation number

85

Figure IV-14: Crystallinity as a function of generation number.

IV.4.5 Morphology

As expected, fiber length decreases with recycling. As seen in Figure IV-15, fiber breakup mainly occurs during

the first processing cycle, but breakup also occurs during further cycles. The fiber length decreases by 77%

during the first processing of the CS composites, while the value is 69% for CA composites. Similarly, the

average diameter also decreases with recycling, but at a greater extent during the first processing cycle. The

average diameter is reduced by 58% for the CS composites and by 43% for CA composites during the first

processing cycle, and by 95% and 96% for the CS and CA composites respectively, along the 50 reprocessing

cycles (Figure IV-16). The standard deviation decreases with recycling, indicating that the distribution tends to

get narrower with recycling. Moreover, the presence of a coupling agent in the composites do not seem to

significantly influence the fiber breakup, especially after a few (about 2 recycling cycles). If the fiber length and

diameter distribution and the average length are not exactly the same for both composite series, they are

however very similar (Figure IV-17 and Figure IV-18). On the contrary, the aspect ratio L/D is much less

influenced by recycling as seen in Figure IV-19 Figure IV-21. No significant variation can be observed for the

average value or the L/D distribution, as seen in Figure IV-19 and Figure IV-20, except for very high L/D ratios

which tend to disappear after the first cycles. These trends are similar as observed by Moran et al., Ville et al.,

Sarasua and Pouyet, Augier et al., and Bourmaud and Baley [30, 72, 85, 86, 92, 107]. They are also consistent

with the observation that fiber breakup affects both dimensions; i.e. fiber length and diameter decrease

86

simultaneously.

Figure IV-15: Average fiber length as a function of generation number.

Figure IV-16: Average fiber diameter as a function of generation number.

87

Figure IV-17: Fiber length distribution for the CS composites.

Figure IV-18: Fiber length distribution for the CA composites.

88

Figure IV-19: Fiber L/D ratio distribution for the CS composites.

Figure IV-20: Fiber L/D ratio distribution for the CA composites.

89

Figure IV-21: Average L/D ratio as a function of generation number.

SEM micrographs of selected samples are presented in Figure 22 and these images confirm the results

presented so far. First, there is a clear reduction of fiber length and diameter between the 1st, 10th and 50th

generation for both composite series (with and without coupling agent), but the main difference occurs between

the first and the tenth generations. Fiber distribution also seems to change with recycling as a wider fiber size

distribution can be observed for the first generations of both composite series. The distribution also shifts towards

smaller values, and thus confirms the trends observed in fiber size analysis (Erreur ! Source du renvoi

introuvable. Figure IV-21) showing that fiber breakup mainly affects longer fibers. Moreover, it can also be

noticed that the fiber dispersion, and thus the composite homogeneity, improves with recycling. This can be

seen by a change of the texture on the surface (cross-section) of the broken samples. Finally, a small effect of

the coupling agent can be noted, especially for the first generation. It can be seen that CSG1 shows evidence

of fiber pull-out (holes) as well as interfacial gaps between the fibers and the matrix, thus indicating poor

compatibility and adhesion between both components. For the composite with coupling agent, no such voids

can be seen as the fibers are well embedded in the matrix [ [165].

90

Figure 22: SEM micrographs of selected samples: (a) CAG1 (x25), (b) CAG1 (x150), (c) CSG1 (x25), (d) CSG1 (x150), (e) CAG10 (x25), (f) CAG10 (x150), (g) CSG10 (x25), (h) CSG10 (x150), (i) CAG50 (x25), (j) CAG50 (x150), (k) CSG50

(x25) and (l) CSG50 (x150).

91

IV.4.6 Tensile Properties

Recycling appears to have an important effect on the tensile properties (Figure IV-23) as all the parameters are

affected by reprocessing. However, the results indicate that the most affected parameters by recycling are the

tensile properties at break (Figure IV-24 and Figure IV-25). Both stress and strain at break are increasing with

recycling, but the variations are much smaller than for the matrix. Figure IV-24 shows that for CS composites,

the stress at break (σb) increases by 252% after 30 recycling cycles, reaching 5.2 MPa, before slightly decreasing

to 4.2 MPa for the 50th generation. This trend is similar as observed for the matrix in our previous work, but to a

higher level [75]. On the contrary, for CA composites, the stress at break (σb) increases by 573% over the whole

50 cycles of recycling indicating that the coupling agent used has an effect on the composite behavior. The strain

at break (εb) is also strongly affected with a sharp increase of 141% after 50 cycles for the CS composites, and

199% for the CA composites. This confirms again the positive effect of the coupling agent on the composite

behavior towards recycling as seen on Figure IV-25. This trend is similar as for the matrix, but to a much lower

extent in this case. Figure IV-26 shows that the yield stress (σy) is higher for the composites than for the matrix.

It also shows a 14% increase over the two first generations before decreasing by 21% over the remaining cycles

for the CS composites. On the other hand, the value steadily decreases by 26% over the 50 cycles for the CA

composites, similar as for the neat matrix. On the contrary, the yield strain (εy) is lower for the composites than

for the matrix. Moreover, while it did not change significantly for the matrix in our previous study [41], it increases

by 47% for CS composites and 48% for CA composites, as seen in Figure IV-27. In Figure IV-28, the Young’s

modulus (E) is higher for the composites than for the matrix. But the trends are more complex as it decreases

by 17% for CS composites and by 30% for the CA composites. This trend is similar to that observed for the

matrix to a higher extent, but still remains within the experimental error. Finally, as seen in Figure IV-29, the

energy at break is lower for the composites than for the matrix. It also increases much slower than for the matrix

as it only increases by 179% for the CS composites and by 166% for the CA composites, while it increases by

240% for the matrix. These variations can be attributed to two phenomena. Firstly, there is a lower entanglement

level and higher chain mobility due to chain scission, leading to a loss of rigidity and a gain of mobility

(elongation/elasticity). Secondly, it can be associated to a decrease in fiber length, thus increasing even more

the chain mobility. Finally, similar to the results obtained on the neat matrix, the composites have a more ductile

behavior with recycling, but lower overall performances with lower strength and rigidity as observed in Figure

IV-23. It can also be noted that the addition of the coupling agent leads to better properties for the first cycles,

but this effect tends to decrease with recycling as no difference is observed between both composites after 30

cycles. It also seems that the presence of a coupling agent slightly lowers the performances for a higher number

of cycles, but a higher number of recycling cycles would be needed to confirm this trend.

92

Figure IV-23: Typical tensile stress-strain curves for different samples.

Figure IV-24: Stress at break as a function of generation number.

93

Figure IV-25: Strain at break as a function of generation number.

Figure IV-26: Yield stress as a function of generation number.

94

Figure IV-27: Yield strain as a function of generation number.

Figure IV-28: Young's modulus as a function of generation number.

95

Figure IV-29: Energy at break as a function of generation number.

Most of studies on the mechanical properties of recycled composites with natural fibers do not agree on the

recycling behavior, as they mainly report opposite trends [72, 75, 98, 86, 88, 90, 92, 94, 107, 163]. Twite

Kabamba et al. also observed that the yield stress decreases with recycling, but they also indicated that strain

at break remained constant [163]. On the contrary, Soccalingame et al. reported an increase of the strain at

break, but a constant stress at break [90]. Bourmaud and Baley, Beg and Pickering, Akesson et al. and Maldas

et al. all reported an increase of the strain at break and a decrease of the yield stress [55, 72, 86, 88, 94], while

Le Duigou et al. noted that both stress and strain at break decreased with recycling [97]. Finally, Fonsecca-

Valero observed an increasing tensile strength and elongation at break [89]. These different behaviors could be

attributed to the nature of the matrix/fibers used, as well as other factors like thermo-mechanical history. The

increase of yield stress with recycling can be attributed to better fiber dispersion in the matrix. However, as fiber

length decreases, this effect is counterbalanced by the loss of reinforcing effect of the fibers. Other trends such

as increasing yield stress or modulus can be attributed to higher crystallinity level or matrix crosslinking, while

others trends such as increasing elongation at break could be the result of chain scission. The results reported

here do not show any gel behavior or any crystallinity increase during recycling, but are consistent with the chain

scission and fiber breakup observed.

96

IV.4.7 Bending Properties

Bending properties are much less influenced by recycling than tensile ones. As expected, the bending modulus

is higher for the composites than for the matrix, but it also decreases with recycling and this decrease is more

important. Figure IV-30 reports a 16% decrease for the CS composites, while the value is 30% for the CA

composites. These values are much higher than the 12% for the neat matrix [75]. This loss of performance is

due to chain scission and fiber breakup in the composites when recycling, leading to fewer entanglement and

higher chain mobility, producing a rigidity loss. This is in agreement with the tensile modulus results and most

studies on the mechanical recycling of composites [88, 98].

Figure IV-30: Bending modulus as a function of generation number.

IV.4.8 Impact Properties

As expected, Figure IV-31 shows that the addition of fibers in a polymer substantially decreases the impact

properties due to the creation of stress concentration points at the fiber surface. While the neat matrix did not

break under notched Charpy tests, both CS and CA composites broke. It can be seen that for the first processing

cycle, the impact strength is much higher for the CS composites than for the CA composites. This indicates that

the coupling agent has a negative effect on the impact properties, probably due to better stress transfer to the

fragile fibers (low elongation at break). For the CS composites, the impact strength (Fbk) decreases from 29.5 to

18.9 kJ/m², a drop of 36%. After the second cycle, no significant further effect can be observe and the impact

97

strength remains constant around 18.0 kJ/m². On the contrary, for the CA composites, the impact strength

remains around 18.0 kJ/m² during the 50 recycling cycles. In their works, Beg and Pickering [88] and Sarasua

and Pouyet [107] also observed a decrease of the impact properties of composites with recycling. So the results

here are consistent with these works. Lower impact properties can be attributed to chain scission (loss of

entanglement) in the matrix during the first processing cycle, while the stabilization of the properties for further

cycles can be explained with respect to the molecular properties remaining constant with generation number

(see Figure III-3 Figure IV-6). As fiber breakup mainly occurred during the first processing cycle, negligible effect

is observed for these composites.

Figure IV-31: Impact strength as a function of generation number.

IV.5 Conclusions

In this work, flax fiber/high density polyethylene (HDPE) composites were processed and recycled multiple times

by extrusion to simulate a long-term closed-loop recycling direct reprocessing. Maleic anhydride grafted

polyethylene was used as a coupling agent in one series to determine the effect of interfacial stress transfer.

Both composite series were subjected to up to 50 extrusion cycles under constant processing conditions. After

these cycles, the samples for characterization were produced by injection and compression molding. The

degradation of the composites and their constituents during the various extrusion cycles were characterized in

terms of physical, thermal and mechanical properties.

98

From the results obtained, some conclusions can be made. Firstly, the addition of fiber leads to a decrease of

the elongation properties and an increase of the composites rigidity. It also induced modification of the chain

length of the polymer matrix and a reduction of the fiber length by fiber breakup and chain scission taking place

during the initial processing cycle. The fiber length decreased by 77% during the initial processing of the CS

composites, and by 69% for CA composites. Then, the recycling did not seem to significantly affect the density

of the composite materials. Differential scanning calorimetry (DSC) results also indicated that, for both series,

the melting and crystallization parameters were not affected by recycling and remained stable for the 50 cycles.

Thermogravimetric analysis (TGA) also showed no changes in the degradation process, as all the studied

parameters showed no significant variation. All these observations were similar to those made for the matrix. On

the contrary, both number and weight average molecular mass (Mn and Mw) showed no significant variation with

recycling for both composites, and as a result, neither did the polydispersity index. Fiber length strongly

decreased with recycling, showing a loss of 95% and 96% (for respectively CS and CA composites) after 50

recycling cycles while the aspect ratio L/D remained stable indicating a simultaneous decrease of the diameter.

Mechanical properties were the second most affected parameters with recycling, except for the impact properties

which remained constant. The most affected parameters were the properties at break; i.e. the stress (σb) and

strain at break (εb) which increased with recycling for both composites. After 50 generations, the stress at break

(σb) increased respectively by 252% and 573% for the CS and CA composites. Similarly, the strain at break (εb)

increased respectively by 141% and 199% for the CS and CA series. The yield stress (σy) increased by 14% for

the first 8th generations, and then decreased by 21% during the remaining cycles while the yield strain (εy)

increased by 47% for CS composites and by 48% for CA composites over the whole recycling process. Most of

these trends were similar to the neat matrix as reported in our previous study. Young’s modulus (E) did not seem

much influenced by recycling, partially due to the high experimental error on this parameter. Finally, the flexion

modulus (Eb) decreased by 16% and 30% for the CS ad CA series, respectively. All these results confirmed the

effect of recycling on the composite behaviors. However, for most parameters, this effect remained quite limited.

All the variations observed can be attributed to chain scission and fiber breakup taking place in the materials.

These phenomena induced higher chain mobility due to lower entanglement and lower molecular weight.

Similarly to our first work on the mechanical recycling of the neat matrix, no evidence of crosslinking or branching

was observed. However, for some characterizations, more trials should be conducted to lower experimental

error and to confirm the trends. Moreover, more generations should be realized to confirm the trends at longer

terms. The results obtained here depends strongly on the material constituents, the materials history, the

processing conditions and equipment. They could however be improved by the addition of a stabilizer before

processing, such as anti-UV and anti-oxidation additives, thus avoiding chain scission and lowering of the

molecular weight limiting the loss of performances and matrix degradation. This loss can also be limited by using

proper conditions including low residence time and temperature. However, prediction of the composites

99

degradation and their constituents during processing remains difficult due to the high number of parameters

involved. Depending on these conditions, it would be possible to improve some properties, but this is generally

at the cost of others. Thus, as for the polymer matrix, the results and conclusions have to be taken in the general

context of composites recycling, but are mainly valid for the materials and processing conditions used in this

work.

Acknowledgements

Financial support from the Natural Sciences and Engineering Research Council of Canada (NSERC,

Canada), the Research Center for High Performance Polymer and Composite Systems (CREPEC, Quebec),

the Quebec Research Funds – Nature and Technologies (FQRNT, Québec), the Quebec Center for Functional

Materials (CQMF, Quebec), the Research Center for Advanced Materials (CERMA, Quebec), and the

Renewable Materials Research Centre (CRMR, Quebec) is acknowledged for this work. The technical help of

Mr. Yann Giroux is highly appreciated for the experimental work.

100

Chapter V. Conclusions and recommendations

V.1 General conclusion

Over the last decades, thermoplastic composites were used at a great extent in a wide range of applications.

Due to the increasing volume of such materials and the resulting environmental effects, their recycling became

of great interest. The increasing public concern about their environmental impact led to the instauration of

constantly new laws and directives and, as a result, to the development of several studies and works on the

subject. Recycling of these materials would allow the reduction of waste landfilled simultaneously to the

conservation of raw materials. However, even if their processing is yet well known, the recycling behavior of

these materials is much less understood. Fibers addition in the polymer matrix modifies the materials behavior,

and both the fibers and the matrix can degrade during recycling. This creates new challenges for their recycling.

However, several aspects related to composite recycling remain unknown and most studies focused on short

term recycling, with less than 10 reprocessing cycles. Thus the objective of this thesis was to fill this lack of

information by studying the long-term mechanical recycling of high density polyethylene/flax composites. In

particular, this work aimed to better understand the degradation and loss of performance undergone by a

polymer and their composites. This was done by studying the materials behavior towards recycling and their link

to the materials structure and properties at three scales: macroscopic, microscopic and molecular.

In the first part of this project, high density polyethylene was reprocessed by extrusion up to 50 times to simulate

a long-term closed-loop recycling process under constant processing conditions. Physical, thermal, mechanical

and molecular analyses were conducted to evaluate the degradation undergone and the effect of intensive

recycling on the structure and performances of the polymer. The results showed that recycling did not seem to

significantly affect the physical and thermal properties of the material, but some molecular and mechanical

modifications could be noted. A general decrease of the mechanical properties was also observed, except for

tensile properties at break which increased significantly with recycling. Indeed, at the end of the 50 cycles, the

stress at break (σb) increased by 75%, the strain at break (εb) by 500%, and the energy at break (TEB) by 240%,

while the yield strain decreased by 29%. These variations were attributed to the higher mobility and lower

entanglement of polymer chains induced by chains scission taking place in the polymer during recycling. These

chain length reductions were confirmed by both the molecular characterizations and by the viscosity decrease

observed via melt flow rate (MFR) tests, as the average weight molecular weight decreased by 24% and the

MFR increased by almost 100%. No evidence of crosslinking or branching was found during the whole recycling

process.

As a second step of this project, two flax fiber/high density polyethylene composites were produced and recycled

up to 50 times by closed-loop extrusion under constant processing conditions. Maleic anhydride grafted

101

polyethylene (MAPE) was used as a coupling agent in one of the series. The degradation of the composites and

their constituents during the various extrusion cycles were characterized in terms of physical, thermal, molecular,

mechanical and morphological properties. The results showed that the addition of fibers in the polymer led to a

higher rigidity (+92% for the CS composites and +112% for the CA composites) and tensile strength (+13% for

the CA composites), but to lower elongation properties (-82% for σb and -68% for εb for the CS composites and

(-89% for σb and -75% for εb for the CA composites). As for the matrix, physical and thermal properties were not

affected by recycling and remained stable for the 50 cycles. However, contrarily to the matrix recycling, the

average molecular masses (Mn and Mw) showed no significant variation with recycling for both series of

composites, and neither did the polydispersity index and the intrinsic viscosity, except for the first processing

cycle. Fiber length and diameter were the most affected parameters, showing respectively a strong decrease of

77% and 95% for the CS composites, and of 69% and 96% for the CA composites, with a significant drop in the

first cycle. As a result, the aspect ratio (L/D) remained almost constant over the recycling process. Most of the

mechanical properties also changed with reprocessing, except for Young’s modulus (E) and impact strength (Fb)

which remained constant. The most affected parameters were the stress at break (σb), the strain at break (εb)

and the energy at break (TEB) which all increased with recycling. After 50 generations, for the CS and CA

composites respectively, the stress at break increased by 252% and 573%, the strain at break by 141% and

199%,, and the energy at break by 179% and 166%. The yield stress decreased when increasing the number of

cycle while yield strain increased. The bending modulus also decreased with recycling but as for most

parameters, this effect remained quite limited. These modifications were attributed to the higher chain mobility

and lower fiber length induced by chain scissions and fiber breakup taking place in the composite. Similarly to

the first part on the matrix recycling, no evidence of crosslinking or branching was observed. The coupling agent

seemed to slightly improve the composites properties, except for the impact strength (Fb), but this effect almost

disappeared after a few recycling cycles. However, more generations should be performed to confirm these

trends at longer term and to reduce the experimental errors.

In both parts of this work, the results obtained were strongly function of the materials, their thermo-mechanical

history, processing conditions and equipment which could be improved by an adequate choice of processing

parameters such as low residence time and temperature, as well as the addition of stabilizers during processing,

such as anti-UV and anti-oxidants, limiting chain scission and matrix degradation, thus the resulting loss of

performances. However, the prediction of the degradation of composites and their constituents during

processing remains difficult due to the important number of parameters involved. Thus, for both the neat matrix

and the composites, these results and conclusions have to be taken in the general context of composites

recycling, as they are mainly valid for the materials and processing conditions used in this work.

102

V.2 Perspectives

This thesis represents a first step to study the long-term mechanical recycling of thermoplastic matrices and their

composites. It showed that these materials could be recycled under proper conditions and maintained their

properties to some extent. However, some aspects have not been studied both by choice and/or by lack of time,

and some others remained uncertain. As they present a scientific interest, additional work should be conducted

on these aspects to fill this lack of knowledge. The following aspects should be considered for further studies:

• In this thesis, virgin high density polyethylene was used as a starting matrix. It could be interesting to

change the nature and grade of the polymer matrix, as to determine its effect on the composite behavior towards

recycling. To do so, common thermoplastics such as polypropylene (PP), polystyrene (PS), polyvinyl chloride

(PVC), as well as other grades of polyethylene (linear low density polyethylene (LLDPE), low density

polyethylene (LDPE), medium density polyethylene (MDPE) could be used.

• The composites were produced with flax fibers presenting a range of particle sizes between 250 µm

and 1 mm. In order to better understand the effect of the fiber nature and dimensions on the composite

mechanical recycling, future works should investigate different geometries (shape, size, distribution, etc.) and

natures (types and species of wood, vegetal, natural, synthetic, mixtures, etc.) of reinforcement. As some surface

modification treatment of the reinforcement were shown to improve the composite properties, these treatments

should also be considered.

• The coupling agent used show a slight improvement of the composite properties for the first cycles, but

this effect disappeared after a few cycles. Some other coupling agents of various nature and grades should be

used to better understand their effect on the mechanical recycling process and the resulting properties, as well

as to know if the different types of agents have an effect on the materials final properties.

• Since the content of each constituent (high density polyethylene, flax fiber and coupling agent) was

fixed in this study, other formulations should be tested to get more insight on the effect of composition on the

composite behavior towards recycling. Due to the lack of studies on the mechanical recycling of composites at

high filler contents (above 30% wt.), such formulations should be of great interest for future works.

• Moreover, a coupling agent was the only additive considered here. Based on the degradation

undergone during mechanical recycling, further works should consider other additives such as anti-oxidants and

anti-UV to improve the recycled composites final properties.

• Future work should also change the processing conditions (temperature, screw speed, screw

configuration, flow rate, residence time, die geometry and size, cooling, etc.) and methods (compression

103

molding, injection molding, rotomolding, etc.) as to understand the effect of these parameters on the materials

behavior towards recycling and the resulting final properties. Due to the high number of parameters involved

and to possible interactions between them, more investigations are required for a better understanding of the

degradation process involved to predict the effect of these parameters on the materials properties and to predict

and optimize the composites final performances for different processing conditions. More cycles could also be

needed to clarify the effect of some parameters, such as the properties at break, the yield strain and the flexural

modulus, in order to determine if the trends observed continue or not.

• In this work, rheological and interfacial properties of the composites were not considered. As they both

could lead to a better understanding of the materials behavior towards recycling, and thus to improve the final

properties, they should be of interest of a future work. These characterizations could be used to verify the

hypothesis made in this work.

• Another important issue to consider for future studies is the materials lifecycle. Closed-loop recycling

from virgin materials was considered here. However, this does not take into consideration the possible

degradation and contamination of the materials during their lifecycle. Open-loop recycling should be considered

in future works in various ways. First, recycled materials could be considered as raw materials for similar studies,

as well as the possibility of partial addition of virgin materials during the recycling process. Then, the deterioration

and ageing undergone during the materials lifecycle between each recycling cycle should be taken into

consideration and simulated for a better representation of the real recycling process. However, this is a variable

and complex phenomenon to model and it will require application studies to implement it. Finally, the

inhomogeneity of the waste should also be considered in future works, and the development of new sorting

methods for such composites should also be of interest.

• As this work took place at a laboratory scale, validation of the observations and conclusions made here

should be done at larger scales to verify the feasibility of the industrial implementation of this recycling process.

Economic and market analyses should also be conducted to verify the viability of this method as well as the

possible applications for these recycled composites.

• Finally, the mechanical recycling of foamed thermoplastic composites has not been studied yet. It could

be interesting to study the short- and long-term mechanical recycling of thermoplastic composites, by variation

of both constituents and foaming processes.

104

References

[1] C. Baley, "Fibres naturelles de renfort pour matériaux composites," 10 01 2013. [Online]. Available: http://www.techniques-ingenieur.fr. [Accessed 07 12 2016].

[2] L. Gendre, "Les grandes familles de matériaux composites," 04 05 2011. [Online]. Available: http://eduscol.education.fr. [Accessed 09 12 2016].

[3] Ensinger, "Plastics classification," [Online]. Available: http://www.ensinger-online.com/en/materials/basics-of-plastics/plastics-classification/. [Accessed 08 12 2016].

[4] PlasticsEurope, "Plastics - the Facts 2015: An analysis of European plastics production, demand and waste data," 2015. [Online]. Available: http://www.plasticseurope.org. [Accessed 4 September 2016].

[5] M. Biron, Thermoplastics and Thermoplastic Composites, 2nd ed., Oxford: Elsevier Ltd, 2013.

[6] O. Faruk, A. K. Bledzki, H. P. Fink and M. Sain, "Biocomposites reinforced with natural fibers: 2000-2010," Progress in Polymer Science, vol. 37, no. 11, pp. 1552-1596, 2012.

[7] K. Hamad, M. Kaseem and F. Deri, "Recycling of waste from polymer materials: An overview of the recent works," Polymer Degradation and Stability, vol. 98, no. 12, pp. 2801-2812, 2013.

[8] S. Mishra, A. K. Mohanty, L. T. Drzal, M. Misra and G. Hinrichsen, "A review on pineapple leaf fibers, sisal fibers and their biocomposites," Macromolecular Materials and Engineering, vol. 289, pp. 955-974, 2004.

[9] A. K. Mohanty, M. Misra and L. T. Drzal, "ustainable bio-composites from renewable resources: opportunities and challenges in the green materials world," Journal of Polymers and the Environment,

vol. 10, no. 1-2, pp. 19-26, 2002.

[10] J. Müssig, Industrial Applications of Natural Fibres: Structure, Properties and Technical Applications, West Sussex: John Wiley & Sons, 2010.

[11] K. L. Pickering, M. G. A. Efendy and T. M. Le, "A review of recent developments in natural fibre composites and their mechanical performance," Composites Part A, vol. 83, pp. 98-112, 2016.

[12] A. K. Bledzki and J. M. Gassan, "Composites reinforced with cellulose based fibres," Progress in

Polymer Science, vol. 24, no. 2, pp. 221-274, 1999.

[13] A. K. Mohanty, M. Misra and G. Hinrichsen, "Biofibres, biodegradable polymers and biocomposites: An overview," Macromolecular Materials and Engineering, Vols. 276-277, no. 1, pp. 1-24, 2000.

[14] A. K. Mohanty, M. Misra and L. T. Drzal, "Surface modifications of natural fibers and performance of the resulting biocomposites: An overview," Composite Interfaces, vol. 8, no. 5, pp. 313-343, 2001.

[15] Flax Council of Canada, "Canada - A Flax Leader," 2016. [Online]. Available: http://flaxcouncil.ca. [Accessed 09 12 2016].

[16] A. K. Mohanty, M. Misra and L. T. Drzal, Natural Fibres, Biopolymers and Biocomposites, Boca Raton: CRC Press, 2005.

[17] B. S. Panigrahy, A. Rana, P. R. Chang and S. Panigrahi, "Overview of Flax Fiber Reinforced Thermoplastic Composite," in CSBE/SCGAB Annual General Meeting, Edmonton, 2006.

[18] D. E. Akin, R. B. Dodd and J. A. Foulk, "Pilot plant for processing flax fiber," Industrial Crops and

Products, vol. 21, pp. 369-378, 2005.

105

[19] X. Li, S. Panigrahi and L. G. Tabil, "A Study on Flax Fiber-Reinforced Polyethylene Biocomposites," Applied Engineering in Agriculture, vol. 25, no. 4, pp. 525-531, 2009.

[20] M. J. John and S. Thomas, "Biofibres and biocomposites," Carbohydrate Polymers, vol. 71, pp. 343-364, 2008.

[21] F. Yao, Q. Wu, Y. Lei, W. Guo and Y. Xu, "Thermal decomposition kinetics of natural fibers: Activation energy with dynamic thermogravimetric analysis," Polymer Degradation and Stability, vol. 93, no. 1, pp. 90-98, 2008.

[22] A. M. Mohd Edeerozey, H. . M. D. Akil, A. B. Azhar and M. I. Zainal Ariffin, "Chemical modification of kenaf fibers," Materials Letters, vol. 61, no. 10, pp. 2023-2025, 2007.

[23] M. J. John and R. D. Anandjiwala, "Recent developments in chemical modification and characterization of natural fiber-reinforced composites," Polymer Composites, vol. 29, no. 2, p. 187–207, 2008.

[24] S. Mukhopadhyay and R. Fangueiro, "Physical modification of natural fibers and thermoplastic films for composites - A review," Journal of Thermoplastic Composite Materials, vol. 22, no. 2, pp. 135-162, 2009.

[25] S. V. Joshi, L. T. Drzal, A. K. Mohanty and S. Arora, "Are natural fiber composites environmentally superior to glass fiber reinforced composites?," Composites Part A, vol. 35, no. 3, pp. 371-376, 2004.

[26] X. Li, L. G. Tabil and S. Panigrahi, "Chemical Treatments of Natural Fiber for Use in Natural Fiber-Reinforced Composites: A Review," Journal of Polymers and the Environment, vol. 15, no. 1, pp. 25-33, 2007.

[27] H. Ku, H. Wang, N. Pattarachaiyakoop and M. Trada, "Composites Part B," A review on the tensile

properties of natural fiber reinforced polymer composites, vol. 42, no. 4, pp. 856-873, 2011.

[28] R. W. D. Brouwer, "Nature Fiber composites in structural components: alternative applications for sisal ?," in Food and Agriculture Organization of the UN (FAO) and the Common Fund for Commodities

(CFC) Proceedings, Italy, 2000.

[29] S. Taj, M. A. Munawar and S. U. Khan, "Natural fiber-reinforced polymer composites," Proceedings of

the Pakistan Academy of Sciences, vol. 44, no. 2, pp. 129-144, 2007.

[30] J. Morán, V. Alvarez, R. Petrucci, J. Kenny and A. Vazquez, "Mechanical properties of polypropylene composites based on natural fibers subjected to multiple extrusion cycles," Journal of Applied Polymer

Science, vol. 103, no. 1, pp. 228-237, 2007.

[31] Y. Wang, H. C. Chan, S. M. Lai, H. F. Shen and Y. K. Hsiao, "Effect of Compounding Conditions on the Wood Flour/Polyethylene Composites in Twin-Screw Extruders," in Proceedings of the 59th Annual

Technical Conference, Dallas, 2001.

[32] N. Lu, R. H. Swan and I. Ferguson, "Composition, structure, and mechanical properties of hemp fiber reinforced composite with recycled high-density polyethylene matrix," Journal of Composite Materials,

vol. 46, no. 16, pp. 1915-1924, 2012.

[33] T. J. Keener, R. K. Stuart and T. K. Brown, "Maleated coupling agents for natural fibre composites," Composites Part A, vol. 35, no. 3, pp. 357-362, 2004.

[34] S. Kalia, B. S. Kaith and I. Kaur, "Pretreatments of natural fibres and their application as reinforcing material in polymer composites - a review," Polymer Engineering & Science, vol. 49, no. 7, pp. 1253-1272, 2009.

106

[35] E. Twite Kabamba and D. Rodrigue, "Effect of recycling on polymer foaming," in 8th International

Blowing and Foaming Processes Conference, Munich, 2006.

[36] H. Jin, J. Gonzalez-Gutierrez, P. Oblak, B. Zupancic and I. Emri, "Effect of extensive recycling on flow properties of LDPE.," Polymer Degradation and Stability, vol. 97, pp. 2262-2272, 2012.

[37] H. Jin, J. Gonzalez-Gutierrez, P. Oblak, B. Zupancic and I. Emri, "Effect of Extensive Recycling on Flow Properties of LDPE," in Proceedings of the 71st Annual Technical Conference & Exhibition, Cincinnati, 2013.

[38] N. Benoit and D. Rodrigue, "Extrusion-Calendering of Foamed and Unfoamed Wood/Plastic Composites," in Proceedings of the 70th Annual Technical Conference, Orlando, 2012.

[39] N. Benoit and D. Rodrigue, "Mechanical properties of foamed and unfoamed Wood/plastic Composites," in Proceedings of the 71th Annual Technical Conference, Cincinnati, 2013.

[40] E. Twite-Kabamba, Z. Fassi Fehri and D. Rodrigue, "Properties of Recycled LDPE/Birch Fibre Composites," Progress in Rubber, Plastics and Recycling Technology, vol. 27, no. 1, pp. 1-20, 2011.

[41] F. Ahmad, H. S. Choi and M. K. Park, "A Review: Natural Fiber Composites Selection in View of Mechanical, Light Weight, and Economic Properties," Macromolecular Materials and Engineering, vol. 300, no. 1, pp. 10-24, 2015.

[42] B. Coxworth, "Motive Industries unveils Kestrel EV," 15 09 2010. [Online]. Available: http://newatlas.com/kestrel-ev-pictures-released/16380. [Accessed 09 12 16].

[43] Kupilka, "Kupilka cutlery set," 2014. [Online]. Available: http://kupilka.fi/en/material. [Accessed 09 12 2016].

[44] B. Gavelda, "XBoards Introduces First Flax Fiber Snowboard," 06 02 2012. [Online]. Available: http://snowboarding.transworld.net/news/xboards-introduces-first-flax-fiber-snowboard/#hACEWsih3OHZ6gWO.97. [Accessed 09 12 2016].

[45] Jack Davies Design, "Sustainable laptop case," 2013. [Online]. Available: http://cargocollective.com/jackdaviesdesign/Sustainable-laptop-casing. [Accessed 09 12 2016].

[46] Canadian Hemp Guitars, "The Gallery," 2016. [Online]. Available: http://www.canadianhempguitars.com/. [Accessed 09 12 2016].

[47] Greenline, "Special cases and industrial cases made from retainable natural fiber reinforced plastics," 2016. [Online]. Available: http://www.naturfaserverbundwerkstoffe.de/en/product-examples/carrying-cases/natural-fiber-special-case.php. [Accessed 09 12 2016].

[48] Hemp Eyewear, "Our story," 28 06 2016. [Online]. Available: http://www.hempeyewear.com. [Accessed 09 12 2016].

[49] Green Bay Decking, "Dekk," 2016. [Online]. Available: http://www.greenbaydecking.com/idekk. [Accessed 09 12 2016].

[50] High Street Amsterdam Holland, "Scale hemp plastic," 2016. [Online]. Available: http://www.highstreet.nl/amsterdam/scale-hemp-plastic.html. [Accessed 09 12 2016].

[51] Association canadienne de l’industrie des plastiques, "Top 10 Plastics Facts," 2016. [Online]. Available: https://www.plastics.ca/articles_merge?node=plasticfacts. [Accessed 07 09 2016].

[52] Recyc-Québec, "Plastiques: fiches informatives," 03 2010. [Online]. Available: http://www.recyc-quebec.gouv.qc.ca. [Accessed 07 09 2016].

107

[53] G. Grause, A. Buekens, Y. Sakata, A. Okuwaki and T. Yoshioka, "Feedstock recycling of waste polymeric material," Journal of Material Cycles and Waste Management, vol. 13, no. 4, pp. 265-282, 2011.

[54] F. Ruch, "Composites et recyclage : état de l’art et perspectives," 17 March 2016. [Online]. Available: http://recyclage-plastique.com. [Accessed 4 September 2016].

[55] D. Maldas and B. V. Kokta, "Effect of recycling on the mechanical properties of wood fiber–polystyrene composites. Part I: Chemithermomechanical pulp as a reinforcing filler," Polymer Composites, vol. 11, no. 2, p. 77–83, 1990.

[56] Y. Yang, R. Boom, B. Irion, D.-J. Van Heerden, P. Kuiper and . H. De Wit, "Recycling of composite materials," Chemical Engineering and Processing: Process Intensification, vol. 51, pp. 53-68, 2012.

[57] J. Rybicka, A. Tiwari and G. A. Leeke, "Technology readiness level assessment of composites recycling technologies," Journal of Cleaner Production, vol. 112, no. 1, pp. 1001-1012, 2016.

[58] PlasticsEurope, "Plastics: Too valuable to be thrown away.," 2009. [Online]. Available: http://www.plasticseurope.org. [Accessed 4 September 2016].

[59] J. L. Thomason, L. Yang and R. Meier, "The properties of glass fibres after conditioning at composite recycling temperatures," Composites: Part A, vol. 61, pp. 201-208, 2014.

[60] J. Li and F.-F. Sun, "The Effect of Maleic Anhydride Graft on the Interfacial Adhesion of Carbon Fiber Reinforced Thermoplastic Polystyrene Composite," Journal of Composite Materials, vol. 43, no. 23, pp. 2717-2725, 2009.

[61] A. Mahfoudh, "Étude de la production et de la caractérisation de composites bois-plastiques," M.Sc. thesis, Université Laval, Quebec, 2013.

[62] E. Witten, T. Kraus and M. Kühnel, "Composites Market Report 2015," 21 September 2015. [Online]. Available: http://www.eucia.eu. [Accessed 4 September 2016].

[63] S. Migneault, A. Koubaa, F. Erchiqui, A. Chaala, K. Englund and M. P. Wolcott, "Composition, structure, and mechanical properties of hemp fiber reinforced composite with recycled high-density polyethylene matrix," Journal of Composite Materials, vol. 40, no. 1, pp. 1915-1924, 2011.

[64] Y. Kazemi, "Mechanical and morphological characterization of wood plastic," M.Sc. thesis, Université Laval, Quebec, 2013.

[65] M. Buggy, L. Farragher and W. Madden, "Recycling of composite materials," Journal of Materials

Processing Technology, vol. 55, no. 3-4, pp. 448-456, 1995.

[66] G. Oliveux, L. O. Dandy and G. A. Leeke, "Current Status of Recycling of Fibre Reinforced Polymers: review of technologies, reuse and resulting properties," Progress in Materials Science, vol. 72, pp. 61-99, 2015.

[67] S. Yin, R. Tuladhar, F. Shi, R. A. Shanks, M. Combe and T. Collister, "Mechanical reprocessing of polyolefin waste: A review," Polymer Engineering & Science, vol. 55, no. 12, p. 2899-2909, 2015.

[68] L. Verdolotti, F. Lucolano, I. Capasso, M. Lavorgna, S. Iannace and B. Liguori, "Recycling and recovery of PE-PP-PET-based fiber polymeric wastes as aggregate replacement in lightweight mortar: Evaluation of environmental friendly application," Environmental Progress & Sustainable Energy,, vol. 33, no. 4, pp. 1445-1451, 2014.

[69] S. Kazemi Najafi, "Use of recycled plastics in wood plastic composites - A review," Waste Management,

vol. 33, no. 9, p. 1898-1905, 2013.

108

[70] N. Petchwattana, S. Covavisaruch and J. Sanetuntikul, "Recycling of wood-plastic composites prepared from poly(vinyl chloride) and wood flour," Construction and Building Materials, vol. 28, no. 1, pp. 557-560, 2012.

[71] M. Sjöqvist and A. Boldizar, "Molecular Modification and Compatibilization of Collected Polyethylene," Journal of Polymers and the Environment, vol. 19, no. 2, p. 335-340, 2011.

[72] A. Bourmaud and C. Baley, "Investigations on the recycling of hemp and sisal fibre reinforced polypropylene composites," Polymer Degradation and Stability, vol. 92, no. 6, pp. 1034-1045, 2007.

[73] E. Strömberg and S. Karlsson, "The design of a test protocol to model the degradation of polyolefins during recycling and service life," Journal of Applied Polymer Science, vol. 112, no. 35, pp. 1835-1844, 2009.

[74] P. Oblak, J. Gonzalez-Gutierrez, B. Zupančič, A. Aulova and I. Emri, "Processability and mechanical properties of extensively recycled high density polyethylene," Polymer Degradation and Stability, vol. 114, pp. 133-145, 2015.

[75] N. Benoit, R. González-Núñez and D. Rodrigue, "High density polyethylene degradation followed by closed-loop recycling," Progress in Rubber, Plastics & Recycling Technology, vol. 33, no. 1, pp.17-37 (2017).

[76] A. Conroy, S. Halliwell and T. Reynolds, "Composite recycling in the construction industry," Composites

Part A, vol. 37, no. 8, p. 1216-1222, 2006.

[77] C. D. Papaspyrides, J. G. Poulakis and C. D. Arvanitopoulos, "Recycling of glass fiber reinforced thermo-plastic composites. I. Ionomer and low density polyethylene based composites," Resources

Conservation and Recycling, vol. 14, no. 2, pp. 91-101, 1995.

[78] A. Soroudi and I. Jakubowicz, "Recycling of bioplastics, their blends and biocomposites: A review," European Polymer Journal, vol. 49, pp. 2839-2858, 2013.

[79] J. M. Henshaw, W. Han and A. D. Owens, "An Overview of Recycling Issues for Composite Materials," Journal of Thermoplastic Composite Materials, vol. 9, no. 1, pp. 4-20, 1996.

[80] L. Steuernagel, "Recycling behaviour of natural fiber Reinforced Thermoplastics," in Proceedings of the

Polymer Processing Society 29th Annual Meeting, Nuremberg, 2013.

[81] E. Asmatulu, J. Twomey and M. Overcash, "Recycling of fiber-reinforced composites and direct structural composite recycling concept," Journal of Composite Materials, vol. 48, no. 5, pp. 593-608, 2014.

[82] J. G. Poulakis, P. C. Varelidis and C. D. Papaspyrides, "Recycling of polypropylene-based composites," Advances in Polymer Technology, vol. 16, no. 4, p. 313-322, 1997.

[83] V. Goodship, Management, Recycling and Reuse of Waste Composites, Cambridge: WP and CRC Press, 2009.

[84] F. Vilaplana and S. Karlsson, "Quality Concepts for the Improved Use of Recycled Polymeric Materials: A Review," Macromolecular Materials and Engineering, vol. 293, no. 4, pp. 274-297, 2008.

[85] J. Ville, F. Inceoglu, N. Ghamri, J. L. Pradel, A. Durin, R. Valette and B. Vergnes, "Influence of Extrusion Conditions on Fiber Breakage along the Screw Profile during Twin Screw Compounding of Glass Fiber-reinforced PA," International Polymer Processing, vol. 28, no. 1, pp. 49-57, 2013.

[86] A. Bourmaud and C. Baley, "Rigidity analysis of polypropylene/vegetal fibre composites after recycling," Polymer Degradation and Stability, vol. 94, no. 3, pp. 297-305, 2009.

109

[87] M. Pervaiz and M. M. Sain, "Carbon storage potential in natural fibre composites," Resources

Conservation and Recycling, vol. 39, no. 4, pp. 325-340, 2003.

[88] M. D. H. Beg and K. L. Pickering, "Reprocessing of wood fibre reinforced polypropylene composites. Part I: Effects on physical and mechanical properties," Composites Part A, vol. 39, no. 7, pp. 1091-1100, 2008.

[89] C. Fonseca-Valero, O. Almudena, J. Arranz-Andrés and C. González-Sánchez, "Mechanical recycling and composition effects on the properties and structure of hardwood cellulose-reinforced high density polyethylene eco-composites," Composites Part A, vol. 69, pp. 94-104, 2015.

[90] L. Soccalingame, A. Bourmaud, D. Perrin, J. C. Bénézet and A. Bergeret, "Reprocessing of wood flour reinforced polypropylene composites," Reprocessing of wood flour reinforced polypropylene

composites: Impact of particle size and coupling agent on composite and particle properties, vol. 113, p. 72–85, 2015.

[91] K. Walz, R. Jacobson and A. R. Sanadi, "Effect of reprocessing/recycling on the mechnical properties of kenaf-PP composites," Wisconsin, 1994.

[92] L. Augier, G. Sperone, C. Vaca-Garcia and M.-E. Borredon, "Influence of the wood fibre filler on the internal recycling of PVC-based composites," Polymer Degradation and Stability, vol. 92, no. 7, pp. 1169-1176, 2007.

[93] D. Kurniawan, B. S. Kim, H. Y. Lee and J. Y. Lim, "Effects of repetitive processing, wood content, and coupling agent on the mechanical, thermal, and water absorption properties of wood/polypropylene green composites," Journal of Adhesion Science and Technology, vol. 27, no. 12, pp. 1301-1312, 2013.

[94] D. Åkesson, T. Fuchs, M. Stöss, A. Root, E. Stenvall and M. Skrifvars, "Recycling of wood fiber-reinforced HDPE by multiple reprocessing," Journal of Applied Polymer Science, vol. 133, pp. 43877-43884, 2016.

[95] N. Graupner, K. Albrecht, G. Ziegmann, H. Enzler and J. Muessig, "Influence of reprocessing on fibre length distribution, tensile strength and impact strength of injection moulded cellulose fibre-reinforced polylactide (PLA) composites," EXPRESS Polymer Letters, vol. 10, no. 8, pp. 647-663, 2016.

[96] A. Arbelaiz, B. Fernández, J. A. Ramos and I. Mondragon, "Mechanical properties of short flax fibre bundle/polypropylene composites: Influence of matrix/fibre modification, fibre content, water uptake and recycling," Composites Science and Technology, vol. 65, no. 10, pp. 1582-1592, 2005.

[97] A. Le Duigou, I. Pillin, A. Bourmaud, P. Davies and C. Baley, "Effect of recycling on mechanical behaviour of biocompostable flax/poly(L-lactide) composites," Composites Part A, vol. 39, no. 9, pp. 1471-1478, 2008.

[98] V. Srebrenkoska, G. B. Gaceva, M. Avella, M. E. Errico and G. Gentile, "Recycling of polypropylene-based eco-composites," Polymer International, vol. 57, no. 11, p. 1252-1257, 2008.

[99] D. Paukszta and S. Borysiak, "The Influence of Processing and the Polymorphism of Lignocellulosic Fillers on the Structure and Properties of Composite Materials-A Review," Materials, vol. 6, no. 7, pp. 2747-2767, 2013.

[100] R. Von Turkovich and L. Erwin, "Fiber fracture in reinforced thermoplastic processing," Polymer

Engineering & Science, vol. 23, no. 13, pp. 743-749, 1983.

110

[101] F. Inceoglu, J. Ville, N. Ghamri, J. L. Pradel, A. Durin, R. Valette and B. Vergnes, "Correlation between processing conditions and fiber breakage during compounding of glass fiber-reinforced polyamide," Polymer Composites, vol. 32, no. 11, pp. 1842-1850, 2011.

[102] "Recycling: the key to future success," Reinforced Plastics, vol. 35, no. 7-8, pp. 20-23, 1991.

[103] P. Corvaglia, A. Passaro, O. Manni, L. Barone and A. Maffezzoli, "Recycling of PP-based Sandwich Panels with Continuous Fiber Composite Skins," Journal of Thermoplastic Composite Materials, vol. 19, no. 6, pp. 731-745, 2006.

[104] H. Cornier-Ríos, P. A. Sundaram and J. T. Celorie, "Effect of Recycling on Material Properties of Glass-filled Polyethylene Terephthalate," Journal of Polymers and the Environment, vol. 15, no. 1, pp. 51-56, 2007.

[105] K. Kemmochi, H. Takayanagi, C. Nagasawa, J. Takahashi and R. Hayashi, "Possibility of closed loop material recycling for fiber reinforced thermoplastic composites," Advanced Performance Materials, vol. 2, no. 4, pp. 385-394, 1995.

[106] G. Colucci, O. Ostrovskaya, A. Frache, B. Martorana and C. Badini, "The effect of mechanical recycling on the microstructure and properties of PA66 composites reinforced with carbon fibers," Journal of

Applied Polymer Science, vol. 132, no. 29, pp. 42275-42284, 2015.

[107] J.-R. Sarasua and J. Pouyet, "Recycling effects on microstructure and mechanical behaviour of PEEK short carbon-fibre composites," Journal of Materials Science, vol. 32, no. 2, pp. 533-536, 1997.

[108] G. Schinner, J. Brandt and H. Richter, "Recycling Carbon-Fiber-Reinforced Thermoplastic Composites," Journal of Thermoplastic Composite Materials, vol. 9, no. 3, pp. 239-245, 1996.

[109] K. Wang, N. Bahlouli, F. Addiego and S. Ahzi, "Elastic and yield behaviors of recycled polypropylene-based composites: Experimental and modeling study," Composites Part B, vol. 99, pp. 132-153, 2016.

[110] B. Xie, Y.-H. Cui, J. Xu, X.-X. Wang, H.-H. Zhang and Z.-H. Zhang, "Investigation on microstructure and properties of foamed (wood fiber)/(recycled polypropylene) composites," Journal of Vinyl and Additive

Technology, vol. 18, no. 2, pp. 105-112, 2012.

[111] R. Prasanth, R. Shankar, A. Dilfi, V. K. Thakur and J.-H. Ahn, "Eco-Friendly Fiber-Reinforced Natural Rubber Green Composites," in Green Composites from Natural Resources, Boca Raton, CRC Press, 2013, pp. 205-254.

[112] C. Borsoi, C. Cavinato, L. C. Scienza and A. J. Zattera, "Processing of recycled expanded polystyrene and waste newsprint fibers composites," in Polymer Processing Society 29th Annual Meeting, Nuremberg, 2013.

[113] B. A. Consuegra Siaotong, "Effects of fiber content and extrusion parameters on the properties of flax fiber-polyethylene composites," Saskatoon, 2006.

[114] J. E. Winandy, N. M. Stark and C. M. Clemons, "Considerations in recycling of Wood-Plastic Composites," in 5th Global Wood and Natural Fibre Composites Symposium, Kassel, 2004.

[115] S.-K. Yeh and R. K. Gupta, "Improved wood–plastic composites through better processing," Composites Part A, vol. 39, no. 11, pp. 1694-1699, 2008.

[116] L. Scelsi, . A. Hodzic, C. Soutis, S. A. Hayes, S. Rajendran, M. A. AlMa'adeed and . R. Kahraman, "A review on composite materials based on recycled thermoplastics and glass fibres," Plastics Rubber and

Composites, vol. 40, no. 1, pp. 1-10, 2011.

111

[117] F. Cavalieri and F. Padella, "Development of composite materials by mechanochemical treatment of post-consumer plastic waste," Waste Management, vol. 22, no. 8, pp. 913-916, 2002.

[118] K. Khait, E. G. Riddick and J. M. Torkelson, "Recovery of post-consumer plastic waste via Solid State Mechanochemistry," in Proceedings of the 59th Annual Technical, Dallas, 2001.

[119] E. Twite-Kabamba, "Study of the recycling of foamed PE on an extrusion line," Ph.D. thesis, Université Laval, Quebec, 2006.

[120] M. E. Otheguy, A. G. Gibson, E. Findon, R. M. Cripps, . A. Ochoa Mendoza and M. T. Aguinaco Castro, "Recycling of end-of-life thermoplastic composite boats," Plastics, Rubber and Composites, vol. 38, no. 9-10, pp. 406-411, 2009.

[121] J.-J. Robin, "Recyclage des thermoplastiques," Techniques de l’Ingénieur, vol. 3832, pp. 1-16, 2003.

[122] H.-T. Joung, S.-J. Cho, Y.-C. Seo and W.-H. Kim, "Status of recycling end-of-life vehicles and efforts to reduce automobile shredder residues in Korea," Journal of Material Cycles and Waste Management,

vol. 9, no. 2, pp. 159-166, 2007.

[123] S. Kahlen and M. Gahleitner, "Multiple extrusion effects on polypropylene impact- and random copolymers - A recycling study," in Proceedings of the Polymer Processing Society 29th Annual

Meeting, Nuremberg, 2013.

[124] J. DeRudder, Y. Yang, Y. He, W. Zhao and A. Flores, "The Effect of Recycle on the Properties of Engineering Thermoplastics," in Proceedings of the 70th Annual Technical Conference, Orlando, 2012.

[125] J. Cui and E. Forssberg, "Mechanical recycling of waste electric and electronic equipment: a review," Journal of Hazardous Materials, vol. 99, no. 3, pp. 243-263, 2003.

[126] A. G. Buekens and J. G. Schoeters, "Technical methods in plastics pyrolysis," Macromolecular

Symposia, vol. 135, no. 1, pp. 63-81, 1998.

[127] N. Furgiuele, A. H. Lebovitz, K. Khait and J. M. Torkelson, "Novel strategy for polymer blend compatibilization: Solid-state shear pulverization," Macromolecules, vol. 33, no. 2, pp. 225-228, 2000.

[128] N. Yarahmadi, I. Jakubowicz and L. Martinsson, "PVC floorings as post-consumer products for mechanical recycling and energy recovery," Polymer Degradation and Stability, vol. 79, no. 3, pp. 439-448, 2003.

[129] I. Pillin, N. Montrelay, A. Bourmaud and Y. Grohens, "Effect of thermo-mechanical cycles on the physico-chemical properties of poly (lactic acid)," Polymer Degradation and Stability, vol. 93, no. 2, pp. 321-328, 2008.

[130] J.-K. Lee and J.-H. Shin, "Triboelectrostatic separation of pvc materials from mixed plastics for waste plastic recycling," Korean Journal of Chemical Engineering, vol. 19, no. 2, pp. 267-272, 2002.

[131] M. Kostadinova Loultcheva, M. Proietto, N. Jilov and F. P. La Mantia, "Recycling of high density polyethylene containers," Polymer Degradation and Stability, vol. 57, no. 1, pp. 77-81, 1997.

[132] E. Twite Kabamba and D. Rodrigue, "The effect of recycling on LDPE foam properties," Progress in

Rubber, Plastics and Recycling Technology, vol. 24, no. 1, pp. 1-17, 2008.

[133] R. Scaffaro, F. P. La Mantia, L. Botta, M. Morreale, N. T. Dintcheva and P. Mariani, "Competition between chain scission and branching formation in the processing of high-density polyethylene: Effect of processing parameters and of stabilizers," Polymer Engineering and Science, vol. 49, no. 7, pp. 1316-1325, 2009.

112

[134] M. H. Wagner, "The Rheology of Linear and Long-chain Branched Polymer Melts," Macromolecular

Symposia, vol. 236, no. 1, pp. 219-227, 2006.

[135] V. Goodship, Introduction to plastics recycling, Shawbury: Smithers Rapra Press, 2007.

[136] J. Dostál, V. Kašpárková, M. Zatloukal, J. Muras and L. Šimek, "Influence of the repeated extrusion on the degradation of polyethylene. Structural changes in low density polyethylene," European Polymer

Journal, vol. 44, no. 8, pp. 2652-2658, 2008.

[137] G. Teteris, "Degradation of polyolefines during various recovery processes," Macromolecular Symposia,

vol. 144, no. 1, pp. 471-479, 1999.

[138] U. W. Gedde, J. Viebke, H. Leijström and M. Ifwarson, "Long‐term properties of hot‐water polyolefin pipes-a review," Polymer Engineering & Science, vol. 34, no. 24, pp. 1773-1787, 1994.

[139] G. G. Cameron and F. Davie, "The thermal degradation of poly(methyl acrylate) V. Relative rates of crosslinking and chain scission," Die Makromolekulare Chemie, vol. 149, no. 129, pp. 169-175, 1971.

[140] F. Berzin, B. Vergnes and L. Delamare, "Rheological behavior of controlled-rheology polypropylenes obtained by peroxide-promoted degradation during extrusion: Comparison between homopolymer and copolymer," Journal of Applied Polymer Science, vol. 80, no. 8, pp. 1243-1252, 2001.

[141] L. B. Brennan, D. H. Isaac and J. C. Arnold, "Recycling of acrylonitrile–butadiene–styrene and high-impact polystyrene from waste computer equipment," Journal of Applied Polymer Science, vol. 86, no. 3, p. 572–578, 2002.

[142] K. Hamad, M. Kaseem and F. Deri, "Effect of recycling on rheological and mechanical properties of poly (lactic acid)/polystyrene polymer blend," Journal of Materials Science, vol. 46, no. 9, pp. 3013-3019, 2011.

[143] Y. Tsuneizumi, M. Kuwahara, O. Okamoto and S. Matsumura, "Chemical recycling of poly(lactic acid)-based polymer blends using environmentally benign catalysts," Polymer Degradation and Stability, vol. 95, pp. 1387-1393, 2010.

[144] K.-H. Su, J.-H. Lin and C.-C. Lin, "Influence of reprocessing on the mechanical properties and structure of polyamide 6," Journal of Materials Processing Technology, Vol. 192-193, pp. 532-538, 2007.

[145] H. B. Parmar, R. K. Gupta and S. N. Bhattacharya, "Rheological and molecular properties of organic peroxide induced long chain branching of recycled and virgin high density polyethylene resin," Polymer

Engineering & Science, vol. 49, no. 9, pp. 1806-1813, 2009.

[146] A. V. Shyichuk and J. R. White, "Analysis of chain-scission and crosslinking rates in the photo-oxidation of polystyrene," Journal of Applied Polymer Science, vol. 77, no. 13, p. 3015-3023, 2000.

[147] J. Aurrekoetxea, M. A. Sarrionandia, I. Urrutibeascoa and M. L. I. Maspoch, "Effects of recycling on the microstructure and the mechanical properties of isotactic polypropylene," Journal of Materials Science,

vol. 36, no. 11, p. 2607-2613, 2001.

[148] J. Z. Liang and W. Peng, "Melt viscosity of PP and FEP/PP blends at low shear rates," Polymer Testing,

vol. 28, no. 4, pp. 386-391, 2009.

[149] Z. Gao, T. Kaneko, I. Amasaki and M. Nakada, "A kinetic study of thermal degradation of polypropylene," Polymer Degradation and Stability, vol. 80, no. 2, pp. 269-274, 2003.

[150] V. A. González-González, G. Neira-Velázquez and J. L. Angulo-Sánchez, "Polypropylene chain scissions and molecular weight changes in multiple extrusion," Polymer Degradation and Stability, vol. 60, no. 1, pp. 33-42, 1998.

113

[151] M. A. Silva Spinacé and M. A. De Paoli, "Characterization of poly(ethylene terephtalate) after multiple processing cycles," Journal of Applied Polymer Science, vol. 80, no. 1, pp. 20-25, 2001.

[152] Canton de Genève et Vaud, "Guide des achats professionnels responsables," State of Vaud Sustainable Development Unit, Geneva, 2015.

[153] N. K. Madi, "Thermal and mechanical properties of injection molded recycled high density polyethylene blends with virgin isotactic polypropylene," Materials & Design, vol. 46, pp. 435-441, 2013.

[154] Recyc-Québec, "Bilan 2010-2011 de la gestion des matières résiduelles au Québec," Quebec, 2012.

[155] M. Taşdemır, H. Biltekin and G. T. Caneba, "Preparation and characterization of LDPE and PP-Wood fiber composites," Journal of Applied Polymer Science, vol. 112, no. 5, p. 3095-3102, 2009.

[156] C. Tissandier, R. C. Vazquez Fletes, R. González-Núñez and D. Rodrigue, "Microcellular Agave Fibre-High Density Polyethylene Composites Produced by Injection Molding," Journal of Materials Science

and Engineering A, vol. 2, no. 11, pp. 677-692, 2012.

[157] A. N. Netravali and S. Chabba, "Composites get greener," Materials Today, vol. 6, no. 4, p. 22-29, 2003.

[158] U. Riedel and J. Nickel, "Natural fibre-reinforced biopolymers as construction materials - new discoveries," Angewandte Makromolekulare Chemie, vol. 272, no. 1, pp. 34-40, 1999.

[159] R. Gosselin, D. Rodrigue and B. Riedl, "Injection Molding of Postconsumer Wood–Plastic Composites I: Morphology," Journal of Thermoplastic Composite Materials, vol. 19, no. 6, pp. 639-657, 2006.

[160] J. J. Schemenauer, T. A. Osswald, A. R. Sanadi and D. F. Caulfield, "Melt Rheological Properties of Natural Fiber-Reinforced Polypropylene," in Proceedings of the 58th Annual Technical Conference, Orlando, 2000.

[161] A. Boldizar, A. Jansson, T. Gevert and K. Moller, "Simulated recycling of post-consumer high density polyethylene material," Polymer Degradation and Stability, vol. 68, no. 3, pp. 317-319, 2000.

[162] H. Fang, Y. Zhang, J. Deng and D. Rodrigue, "Effect of Fiber Treatment on the Water Absorption and Mechanical Properties of Hemp Fiber/Polyethylene Composite," Journal of Applied Polymer Science,

vol. 127, pp. 942-949, 2013.

[163] E. Twite-Kabamba, A. Mechraoui and D. Rodrigue, "Rheological properties of polypropylene/hemp fiber composites," Polymer Composites, vol. 30, no. 10, pp. 1401-1407, 2009.

[164] A. Boldizar, T. Gevert and M. Markinger, "Quality Assessment of Polymeric Recyclates Part 1:Repeated Processing and Thermo-Oxidative Ageing of LDPE," Polymer Recycling, vol. 1, pp. 249-255, 1995.

[165] B. Madsen, M. Aslan and H. Lilholt, "Fractographic observations of the microstructural characteristics of flax fibre composites," Composites Science and Technology, vol. 123, pp. 151-162, 2016.