Sustainable pathway towards large scale melt processing of the new generation of renewable cellulose–polyamide composites

Modern society's growing demands for accountable high-performance and more environmentally friendly materials is leading to increased interest and fast development of sustainable polymeric composite materials. New generations of “greener” products originating from renewable resources fulfil emerging requirements of low environmental and health & safety impacts and contribute to diminishing global dependence on fossil feedstock. The preparation of sustainable polymeric composites via reliable and reproducible melt-compounding methods is still challenging but has the potential to yield applicable and market competitive products. This literature survey reviews the current state of research involving the use of cellulosic materials, as bio-sourced and sustainable reinforcement in melt-processed polyamides and focuses on the main hurdles that prevent their successful large-scale melt-compounding. Particular emphasis is dedicated to emerging bio-sourced polyamides fitting the performance of engineering materials and at the same time offering additional interesting properties for advanced applications such as piezoelectricity for transducers, sensors, actuators and energy harvesters.


Introduction
Research in the development of sustainable polymeric composite materials has received increased academic and industrial interest in recent decades. This is due to modern society's growing demands for accountable high-performance materials as well as more environmentally conscious consumers, industries and governments. Products, in general, Currently, she is developing the GREENPEHS project as a Marie Curie Fellow at the University of Alcalá (SUSCATCOM group). Her research contributes to the smart energy area creating innovative strategies and producing, by an energy, cost-efficient and scalable process, new advanced biobased piezoelectric polymers capable to generate renewable energy.
Professor Boldizar's research concerns the understanding of useful relations between the structure, the manufacturing processing and the functional properties of polymeric materials. A special interest is devoted to the structuring and shaping melt processing of renewable polymeric materials, as this have shown to be a signicant challenge towards making a better use of sustainable polymeric materials and polymer composites. Current work deals with the ageing and durability of polymeric materials in connection to recycling of polymeric materials, the properties of polymers based on renewable resources and how to make better use of polymer composites reinforced with cellulose bre or brils. are sought to be "greener" or in other words originate from renewable resources, have a low environmental impact and prompt low health & safety concerns. These requirements trail sustainability concerns as well as diminishing global supplies of fossil feedstock resources such as crude oil. 1,2 A straightforward research approach to produce polymeric composite materials that full these requirements is the use of feedstock from sustainable natural resources. The challenge in its essence being their processing via reliable and reproducible methods, yielding applicable and market competitive products. Currently, conventional melt processing techniques are assumed to continue being key processing methods for the large-scale production of thermoplastic composites. Extrusion and injection moulding are well established industrial-scale facilities considered sustainable because are inexpensive, fast, and organic solvent-free techniques.
Sustainability aiming thermoplastic composite products can regard cellulosic materials as ideal reinforcements. 3 Being biosourced, they are abundant as well as biodegradable, and in many instances exhibit favourable mechanical properties when compared to many synthetic counterparts with the added benet of a lower cost and light weight. 4,5 Recent scientic advances in the production of cellulose nanomaterials such as cellulose nanocrystals (CNC) and cellulose nanobrils (CNF) have shown numerous new possibilities for wide range of applications. 2,6,7 Polyamides (PAs), commonly referred to as nylons, are of interest as polymer matrices due to their excellent mechanical and thermal properties as well as their relative ease of processing. [8][9][10] Application wise, many polyamides and composites thereof are recognized as engineering grade materials which can in many instances even replace metal parts.
Extruded and moulded polyamide composites are currently found in a wide range of technical applications e.g. in automotive parts, electrical components, and food packaging. The increased availability of bio-sourced polyamides revamps them as viable sustainable materials today. 11 The research interest in cellulose-polyamide composites started in earnest in the 1980s. 12 Aer a decline of interest mainly due to the temperature sensitivity of cellulose setting challenges for its melt processing with polyamide, research publications in this topic has intensied in the last ve years (Fig. 1).
This survey on the current state-of-the-art on cellulosepolyamide composites focuses on the main hurdles that prevent their successful large-scale melt-compounding. Different research approaches and challenges are presented to support the development of a new generation of environmental-friendly materials bases on sustainable processing and feedstocks with advanced properties.

Polyamides and biosourced polyamides
Polyamides are semi-crystalline thermoplastics characterized by a repeating polar amide group, (-CONH-). Since its rst introduction in the market as a moulding material in 1940s, polyamides have been improved, being nowadays rather common engineering thermoplastics. Synthetic polyamides (e.g. PA6, PA66, PA46 and PA612) have been of interest as polymer matrices for composites with wide array of llers and reinforcements due to their excellent mechanical and thermal properties, relative ease of melt processing (albeit at temperatures above 240 C), relatively good adhesion to reinforcements, resistance to oils and corrosive chemicals, and attractive surface appearance. [8][9][10] Polyamides and composites thereof have been used in a variety of packaging, engineering electrical, textile medical and auto applications including more demanding applications such as gas pipes and offshore oileld. Properties and wide application range of high performance class nanoller reinforced polyamides has been reviewed recently by Francisco et al. 14 The most commonly traded grades of polyamides are PA6.6, PA6, PA66, Kevlar and other PAs such as the biosourced PA11, and their market is expected to register a compound annual growth rate (CAGR) of over 4% during the period 2016-2024. 15 Biosourced polyamides (e.g. PA11, PA1010, PA410 and, to a lesser extent, PA610), as the name implies, are sourced from fully or partially derived renewable feedstock. These materials inherit the characteristic PAs properties, exhibiting high mechanical strength and thermal performance with added processing advantages consequently opening up new opportunities in their future market. 11, 16 The global biosourced PA market size was valued at USD 110 million in 2016 and it is expected to reach USD 220 million by 2022, according to a new study by Grand View Research, Inc. These values represent a predicted CAGR of 12.2% from 2015 to 2022, supporting the future increasing of their demand (Fig. 1c). The global production capacity of bio-PA is similar to that of the common bioplastics such as biobased PE, PLA and PBAT, being around 12% of the 2.11 million tonnes of bioplastics produced in 2019 (one percent of the more than 359 million tonnes of plastic produced annually). 13 In general, biosourced PAs have Giada Lo Re, Doctor Europeaus PhD in Chemical and Materials Engineering, serves as Associate Professor at Chalmers University of Technology. Her research covers macromolecular chemistry and engineering of biopolymeric and bio-nanocomposite materials, towards large-scale meltprocessing. Specic interests focus on understanding how the structure-modication of lignocelluloses affects their viscoelastic properties and processability, including bottom-up approaches from the molecular design towards new routes for sustainable reactive processing, aiming at the control of the material nanostructure. She serves as expert evaluator of research grants on behalf of the Natural Sciences and Engineering Research Council of Canada (NSERC) and of the European Commission.
somewhat lower melting temperatures, density, ductility, and moisture absorption than widely used synthetic nylons e.g. PA6 and PA66. Thanks to their renewability, recyclability, light weight, inexpensive nature, electromechanical resistivity, ductility, and creep resistance, biosourced PAs have been attracting attention from various end-use sectors and they are used in a wide variety of applications (Fig. 1d). Textiles emerged as the second-largest end-use segment in 2016, aer automotive applications, while the electrical & electronics sector is projected to emerge as one of the fastest growing end-use sectors in the next few years. 13 However, the main disadvantage to these biosourced materials is that they are currently more expensive compared to traditional nylons. The common biosourced PA11 is a castor oilbased biopolymer and is a semi crystalline polymer that exhibits six different crystalline phases. 17 Its degree of crystallinity and phase composition can have a signicant inuence on its exhibited mechanical properties, giving it a wide selection of adjustable usages. 18 In their study, Zhang et al. 19 showed that there is an optimum annealing temperature for PA11, around 165 C, when crystallinity can be maximized. Although PA11 represents only a minor portion of global nylon production, the demand for PA11 and biosourced polyamides is expected to grow (Fig. 1c), making them of interest for research within both academia and industry. In fact, the rises in oil price and the increasing environmental awareness as well as the stricter environmental policies will make fossil-based polymers more expensive, favouring the production of bio-based alternatives. The production of traditional PA has higher environmental impact than those of biobased PA contributing the potential global worming (z7 Kg CO 2 eq. per kg PA 12 against z4 Kg CO 2 eq. per kg bio-PA 12). 11

Cellulose as reinforcement
Cellulose is a natural biopolymer and is a main constituent of plant cell walls, tunicates, and many species of bacteria where it serves as the fundamental structural reinforcement. The combined global annual production of cellulose by plants is estimated to be 1.5 Â 10 12 tons, making cellulose the most abundant polymer on earth. 3 It can therefore be regarded as a rich source of materials for numerous applications and as such is currently receiving signicant attention in the context of sustainability. 20,21 Cellulose and its numerous derivatives have been extensively studied, focusing on their biological, chemical, as well as mechanical properties. [22][23][24] Cellulose polymeric molecule is recognized as a homopolysaccharide composed of D-anhydroglucopyranose repeating units linked by glycosidic bonds, produced via condensation polymerization of glucose. 3,22,25 Its linearity and many hydroxyl groups enable the formation of ordered crystalline structures which provide the unique mechanical properties to cellulose bres (CF). Different levels of the hierarchical structure in cellulose bres found in a typical plant source is shown in Fig. 2.
The polymers combine to form long and continuous hydrogen bond stabilized microbrils, and in turn several microbrils self-assemble in macrobrils, which can be found oriented in varying directions. 28 Combinations of three main polymers compose the different layers of the plant cell wall: cellulose, hemicellulose, and lignin. The cells are bound together by a middle lamella consisting of the polysaccharide pectin. 2,25 Pulp and its various forms 29,30 e.g. thermomechanical pulp and Kra pulp are generally derived by the mechanical and/or chemical separation of individual CF found in cellulose rich sources such as wood or cotton. The chosen production methodology yields bres with different compositions (e.g. cellulose, hemicellulose and lignin ratios), surface chemistries and physical properties. In the mechanical pulping process, temperature, humidity and mechanical forces are used to soen the bres followed by their separation by shearing force with minimal loss of bres. In the chemical pulping process, caustic alkaline chemicals are used to solubilize the lignin and most carbohydrates in the middle lamella which hold the bres together, resulting in a lower pulp yield but in general liberating stiffer bres. These different features affect their thermal stability and dispersibility in a polymer melt, and in turn their reinforcement effect. Further chemical and/or mechanical deconstruction at the microbril level results in nanocelluloses which are dened as cellulosic materials that possess at least one dimension in the nanometric scale such as "microbrillated cellulose" (MFC) (also referred as microcrystalline cellulose, MCC), cellulose nanobrils (CNFs), and cellulose nanocrystals (CNCs). 3,23,31 Pristine morphologies of the different cellulosic materials are showed in Fig. 3, and their main properties reported in Table 1. MFC are characterized by a non-homogeneous size distribution, and only some of the microbrillated brils are on the nanoscale level. As a consequence, the MFC average aspect ratio is in the range of 1 to 2, limiting their reinforcement effect in composites, while preserving some of the features of the nanocellulose, i.e. early gelation in water dispersion and horn-ication upon drying (as described further below). 32 CNCs and CNFs are typically distinguishable by the way they are obtained and resultant amorphous material content. CNFs are obtained by mechanical disintegration, or debrillation, oen facilitated by chemical or enzymatic pre-treatment and they are characterized by a residual intra-crystallin amorphous content. 20,25,37 CNCs are typically obtained by acid hydrolysis treatment leading to the elimination of most of the amorphous regions and leaving spindle shaped structures.
The properties of the nanocellulose are strongly determined by their surface features which depend mainly on the raw material used and its preparation process. For example, it is worth noting that hydrophobic compounds are usually still present on the surface of CNFs and that a surface charge may be xed by a pre-treatment step. Another example is that sulfuric acid is classically used for CNC production because it promotes the formation of negatively charged sulphate groups at the surface of the released crystals, resulting in very stable aqueous dispersions. However, hydrolysis with sulfuric acid causes the introduction of a considerable amount of negatively charged sulphate half-ester groups on the CNC surfaces which catalyse the thermal degradation during melt processing, especially for polyesters. 39 Hydrolysis of CNCs with hydrochloric or phosphoric acid instead may introduce phosphorylated CNCs with a much lower surface charge density and higher thermal stability. [39][40][41] Other strong acids have been used to produce CNCs with other surface moieties that exhibit different chemical qualities, such as hydrobromic and phosphotungstic or other organic acids. 22,24,39 Cellulose nanomaterials are oen obtained as very dilute suspensions (typically <2 wt%), usually in water because it is a convenient, no toxic and inexpensive polar liquid medium. If the concentration increases with only a few percent solid content, the viscosity of the dispersion increases sharply resulting in gelling. Upon drying, the nanoparticles aggregate through a to substantial and irreversible bre and bril hydrogen bonding (also called hornication, 42,43 see section below), and the nanoscale is irreversibly lost. Freeze-drying, and other similar methods, have been used to circumvent the aggregation of nanomaterials to preserve their individualized state. 44,45 The use of nanocellulose materials as biosourced green reinforcements in plastic composites has garnered particularly signicant research interest during the last 10 years. 2,7,16,31,37,[46][47][48][49][50][51] The intrinsic benets of nanocellulose which are oen mentioned in the literature include, such as type variability, low density, very high specic strength and modulus (rivalling that of steel 52 ) and a high capacity for surface modication and energy generation by burning aer usage. 21 The size reduction in cellulosic nanomaterials retains and amplify most of these properties e.g. mechanical stiffness and the specic surface area, leading to a substantial increase of the available surface hydroxyl groups. This increased hydrophilic character further limits their adhesion and dispersion in non-polar matrices. Nanocelluloses are more thermally stable than wood bers (WF) or our because they are puried from most of the lignin, hemicellulose or other chemically and thermally liable constituents. However, their thermal stability is still low in relation to typical polymer melt processing temperatures, as for polyamides. Nanocelluloses are currently on their way to becoming a fully commercial and widely available products with a multitude of applications. 22 An in-depth review of the large-scale production of nanocellulose and its economic aspects can be found in the review by de Assis et al. 20,25 Melt processing of cellulose reinforced polyamides While several laboratory-scale methods have been used to produce polyamide-based composites reinforced with cellulosics having attractive mechanical properties, e.g. via solution casting, 53-55 few of these processes can be considered industrially viable. The direct melt compounding of polymers and cellulose is in general considered more economical and formulation exible. They involve established industrial-scale  facilities with standard approaches such as extrusion and injection-moulding which are inexpensive, fast, and organic solvent-free techniques. Tables 2-4 include relevant studies on cellulose reinforced polyamides. The Ashby plot reported in Fig. 4 compares the specic Young's modulus as a function of the elongation of the cellulose/polyamide composites with the main engineering materials in use. In the plots emerging bio-sourced polyamides are also included, for the sake of highlighting their properties in comparison with non-biosourced polyamides.
The localization in the plots of the neat biosourced polyamide suggest their possible higher deformability in comparison with the traditional PAs. It is worth notice that biosourced PAs composites reinforced with cellulosic llers show values of Young modulus in very similar range to those reported for PAs composites reinforced with glass bres. Interestingly, the composites based on nanocellulose achieve relatively high mechanical performance at low CNC or CNF content, indicating a good reinforcement effect. Some of these composites maintain good deformability, particularly important for allow their melt processing for some application, e.g. lm blowing for packaging. The general overview offered by the Ashby plot underline that there is still room for improvement for interface design of nanocellulose/biosourced polyamide with improved performance, beyond their potential for secondary recycling (remelt processability) in line with the circularity of their sustainability.
Melt processes are either batch or continuous. Lab-scale polyamide-cellulose composites via batch processing is oen carried out in melt compounder (e.g. Brabender), 41,56-60 or micro-compounder and micro-injection moulders which allow for grams scale of materials. 56,[60][61][62][63] The continuous methods, with continuous feeding, require larger amounts of materials and are less common in cellulose nanocomposites research, most likely due to limited availability of nanocelluloses and cost and time increase for the experiments. However, many studies opted for continuous processes such as twin-screw extrusion, 61,62,64-72 which were ultimately preferred for scaling up, generally provide better component mixing and exible modular designs (e.g. screw congurations, pressure control and gas/steam venting options) compared to batch-wise equipment. 31,48 The main hurdles that prevent a successful meltcompounding of cellulosic materials with polymers in general have been summarized in a wide number of reviews to date. 2,7,21,71 In terms of a typical melt process progression, the issues are: (1) The irreversible hornication of cellulose materials upon drying (prior to melt processing).
(2) The non-uniform distribution/dispersion of cellulose materials in the polymer matrix.
(3) The cellulose materials thermal stability and degradation at elevated temperatures.
(4) The cellulose materials structural integrity (brillation) and shortening upon mechanical shearing during melt processing. Table 3 Summary of strategies and parameters of wet feeding approaches during melt processing of cellulose reinforced polyamides

Cellulose aggregation upon drying
Cellulose materials and in particular nanocelluloses spontaneously form tight, hydrogen-bonded networks during drying in an irreversible aggregation process oen referred to as horn-ication, 6,[42][43][44]47,73,74 that occurs at temperatures as low as 40 C. 75 To retain the benecial qualities of nanocellulose reinforcement, drying techniques aim to produce particulate solid materials, retaining as far as possible the nanosized structure; especially if feeding of the materials during melt processing is done by conventional means via a metering hopper and uniform material free-ow dosing is required.
Different drying methods like oven drying, spray drying, freeze drying and supercritical drying of nanocellulose suspensions have been compared in previous studies 44 and their advantages/ disadvantages have been compiled by Ng et al. 49 The inuence of different drying methods on CNF was also investigated by Peng et al. 75 and it was shown that the morphology as well as the surface energy can differ signicantly between different drying procedures. Spray drying is generally proposed as the most technically suitable and scalable process to dry the suspensions because stable particle sizes in the nano-to micrometre scale are obtained. However, another study showed that conventional spray drying produces a compact solid structure with very low porosity compared to spray freeze drying, although the latter is more expensive. 45 Peng et al. 56 prepared composites from PA6 and different nanocelluloses (MCC, CNF and CNC) in which they characterized the size-distribution of the materials upon spray drying prior to compounding.

Wet feeding of nanocelluloses
An approach to circumvent the self-aggregation of cellulose and nanocellulose materials upon drying is water-assisted melt processing using never-dried wet materials. Table 2 includes relevant studies on wet feeding approach during melt processing of cellulose reinforced polyamides. Typically, in such processes the aqueous ller suspension is directly fed and the water acts as a compatibilizer and plasticizer, until its evaporation during the melt process. The benets of water-assisted production of thermoplastic nanocomposites are listed by Karger-Kocsis et al. 76 This wet feeding of nanomaterials into an extrusion process leads to several advantages: (1) improved dispersion, (2) minimal degradation of cellulose, (3) surface modications may be avoided, and (4) reduced health risks because the nanomaterials are in a slurry. 48 Certain polyamides dissolutive behaviour in water has proven to be useful when a water-assisted melt processing route was opted for. The incorporated water in polyamides acts as a plasticizer with several advantageous exhibited material phenomena for the melt process.
These include a reduction of the polyamide glass transition temperature (T g ), a lowered melt viscosity at a constant temperature, and a melting temperature and crystallization suppression due to a phenomenon known as the cryoscopic effect. 69,76 These properties proved to be useful in water-assisted processing with a thermally sensitive and hydrophilic cellulose component, as was shown by Clemons et al. 60 who successfully produced CNC reinforced PA6 composites by water-assisted melt-compounding and noted an improvement in CNC dispersion (Fig. 6a-c). They also noted a melt temperature reduction up to 45 C (from 230 C to 185 C) by exceeding the 30 wt% water content which prevent the thermal degradation of the cellulosic materials (Fig. 5).
Peng et al. 69 also performed water-assisted extrusion compounding of PA6 with never-dried CNCs with effective pressure and screw design (Fig. 6d-f).

Distribution and dispersion
One of the most pervasive challenges in the preparation of composites, independent of the processing method, is the achievement of a good ller distribution and dispersion within the polymer matrix. Good distribution is dened as the llers being distributed uniformly throughout the polymer matrix which means at any chosen volume, the amount of the llers is the same. Good dispersion is characterized as low ller aggregation level and not signicant reduction of the pristine ller aspect ratio. 56 Composites with good distribution and dispersion of reinforcement expected to have superior nal   mechanical properties. When dealing with nanocellulose in particular, physical, and chemical characteristics, e.g. the high surface energy and hydrophilicity, promote an intrinsic tendency to aggregate. The physical challenge is the miscibility of the two polymeric materials: cellulose in the solid state into the melted polymer matrix. However, the term compatibility, which has not a dened physical meaning, is oen used referring to the quality of the interactions at the two phases interface.
Several articles argue that polyamides should pair well with neat cellulose materials because of their polar nature, as well as exhibiting hydrogen bonding in their molecular structures which can lead to better compatibility between a cellulose ller and the matrix. 7,16,64,78 It has been speculated that the cellulose with its abundant hydroxyl surface groups could form hydrogen bonds with the amines, resulting in good interfacial adhesion between cellulosic materials and polyamide matrices. 53,72 This hydrogen bonding between the cellulose materials and the polyamides may even facilitate nucleation and promote mechanically benecial crystallization in the polyamide matrix. 58,79 However, despite relevant advances in the past decades, overcoming cellulose agglomeration in relatively more hydrophobic polyamides remains difficult. 50,51,80 The quality of the distribution/dispersion is also difficult to evaluate. The inspection of cryofracture and microtome cut surfaces 81 via polarized light microscopy, elective dissolution of the polymer and the use of Raman imaging analysis have been proposed as tools to quantify the dispersion of nanocelluloses in thermoplastics. 82 Instrumental methods such as AFM 41 (Fig. 6g-i), and, in more recent studies, X-ray computed tomography 77 (Fig. 6l-n) have also been used to evaluate the extent of nanocellulose dispersion in a polyamide. Interfacial interaction between CNF and PA11 can be also successfully assessed via rheological studies, in which a good interaction would be exhibited by a high melt viscosity, as was also surmised by Semba et al. 77 Matching the surface chemistry of the cellulose to the polymer is a strategy commonly applied for numerous polymers via e.g. surface functionalization, coupling agents, non-covalent surfactant or covalently graed hydrophobic and/or stearic moieties. 83,84 Surface functionalization of the cellulose materials may include acetylation, esterication, silanization, silylation and glyoxalization to mention a few. 48,49 A study of the melt processing of acetylated MFC in PA410 has been conducted by Leszczyńska et al. 63 yielded better dispersion in addition to an improved thermal stability of the MFC. Another study graed CNC via fatty side chain esterication to improve the CNC interfacial interaction and dispersibility in PA11. 85 Silyl coupling agents, employed in both gas and liquid phase, were used to surface functionalize CF and CNC with subsequent melt processing in polyamides with promising results. 62,68 Surfactants have been used in extrusion melt processes to improve CNC dispersion in many other polymeric matrices before, 85 and a study by Peng et al. used a methyl laurate surfactant as a plasticizer for the processing of CNC in PA11. A treatment with cation reagents can change the cellulose surface charge from negative to positive, and is a strategy which was used in the melt processing of CNF in PA11 and PA12. 66,77,86 An example of coupling agents are silanes, which are commonly employed in the composite industry, 87,88 and an aminosilane has been used on cellulose bres for melt processing with PA12, 62 and on CNC in PA6. 68,89 Cellulose materials can also be coated with a hydrophobic polymer to avoid their aggregation during drying and to improve their distribution/dispersion in the polymer matrix, as was done by Corrêa et al. who coated CNC in PA6 and Zarges et al. who coated CF in PVA prior to melt compounding in polyamide matrices. 71,72,90,91 Thermal degradation of cellulose Depending on their source and processing, the decomposition of cellulosic bres is typically assessed by thermogravimetrical analysis (TGA, under nitrogen atmosphere at heating rate of 10 C min À1 ) and occurs in the range of 150-450 C. This corresponds to the decomposition of glycosylic units leading to a breakdown of the structure and formation of low molecular weight gaseous products like H 2 O, CO 2 , alkanes, and other hydrocarbon derivatives. 92 The thermal degradation of celluloses is a three step process: the rst step being the elimination of water; the secondand central stepbegins from approximately 250 C and is the advanced depolymerization of the cellulose resultant from the dehydration and decomposition of glycosyl to form char; and a third step, above 425 C, assigned to further degradation of charred residue into gaseous products. 93 The choice of cellulose material and acknowledging its thermal stability is of importance, particularly for melt-compounding processes exceeding 200 C, as in the case of polyamides. Sulfuric acid hydrolyzed CNCs are particularly sensitive because the production process incorporates sulphate groups on the cellulose crystal surface which are thermally unstable. 41,93 Hydrochloric acid hydrolysis may be used instead but resultant CNCs tend to aggregate more easily in the aqueous state, due to different surface charge characteristics, and are oen difficult to re-disperse. On the other hand, CNCs prepared by phosphoric acid, which are less common, were found to exhibit acceptable dispersion in polymers and a much higher thermal stability than sulfuric acid prepared CNCs. 94 This was the case for a study focused on PA6/CNC melt processing in which the authors advocated the use of phosphoric acid prepared CNCs as pertinent to their good results. 41 Another approach are the various cellulose surface modications (mentioned in the previous section) which can serve the added purpose of improving cellulose thermal stability. It is possible to coat the cellulose material via either chemical or physical wrapping with a macromolecule or surfactant, which may impart improved thermal stability to the cellulose. This is exemplied in studies in which modied CNFs with ionically adsorbed quaternary ammonium salts bearing long alkyl chains, through simple aqueous adsorption, resulted in improved thermal stability of nanocelluloses processed with PA12. 77,95 Studies involving the coating of CNC with dissolved PA6 (followed by drying the materials and then melt processing), 72 and the coating of CF with PVA provide cellulose with increased thermal stability. 71 A more intricate attempt used a multi-step process to incorporate CNCs in PA6 consisted of an in situ polymerization of caprolactam monomer (to yield PA6) in the presence of silane surface modied CNCs followed by melt extrusion. 68 The use of various melt-temperature prole control and processing techniques to lower the overall processing temperature is a route that has been explored for several polyamides. 70,71 Additives such as plasticizers, ceramic powders and inorganic halide salts have also been used separately and in combination as attempts to decrease the processing temperature and control melt viscosity for polyamides. 65 Recently, the addition of LiCl has been conrmed to be a way to suppress the melting point of polyamides. 96,97 A number of studies have also motivated water-assisted or liquid-mediated techniques as a means to circumvent cellulose thermal degradation issues in melt processes. 60,69,76 Cellulose structural integrity upon mechanical shearing Melt processes that involve particularly high shear melt mixing rates can have an adverse effect on the structural integrity of cellulosic materials, however this aspect is oen overlooked in the research. Evaluation of the shear melt mixing inuence can be determined through various imaging observations, e.g. the length and aspect ratios of nanocelluloses before and aer processing. The mechanical degradation is exhibited by the reduction in the nanocellulose length, affecting the overall mechanical and stress-transfer properties of the reinforcement. 21,51,71 A selective dissolution of the polymer matrix (by using an Soxhlet apparatus) allows to recover the cellulose materials from the composites (aer melt processing) for further morphological analysis. Blends of formic acid (FA) and methylene chloride (DCM) have been used successfully to dissolve polyamide composites to obtain reinforcing cellulose bers for morphological characterization. 98 Feldmann et al. 99 reported that glass bers break more oen when preparing the specimen using the injection molding process than the cellulose bres because of their higher stiffness and lower elongation at break (Fig. 7, bottom image). Moreover, they showed that the processing steps and the amount of cellulose incorporated into the polymeric matrix affect the structural integrity of the cellulose bres. In fact, composites processed by the single step injection moulding and thus, avoiding the previous extrusion process, showed less damaged bres, resulting longer than those submitted to the two-step process. Furthermore, increasing the amount of bres in the polymeric matrix the damage of the bres is higher, resulting in shorter bres (Fig. 7, bottom image). This behaviour is reected in the mechanical properties of the composites as it is possible to observe in Fig. 7 (top image). Composites obtained with lower thermal and mechanical stressed bres both due to the processing method or thanks to the amount of ller used as well, show high mechanical performance. 99 Reactive melt processing An approach relatively unexplored is reactive melt processing. Reactive melt-processes introduce a reactive agent(s) at chosen points during a conventional melt-process. The reaction initiation can occur while homogenizing the materials in the melt and is usually allowed sufficient time for completion. 100 Such chemical reactions may include e.g. polymerizations, graing, branching, controlled cross-linking, coupling and functionalization of the processed materials. 101 In a typical reactive extrusion process, the reactants can be fed to the extruder through normal means via the feed hopper or injected into specic points of the barrel. This allows the use of various liquid or gaseous reactants and the tailoring of intricate reaction sequences. An early attempt in 1985 utilized a two liquid component reaction injection moulding process to produce cellulose bre reinforced in situ polymerized PA6. 102 An interesting review focused on the reactive extrusion of biodegradable polymers has been conducted by Raquez et al. 101 Piezoelectricity as potential emerging application of PAs Thanks to its interesting properties, bio-PA are used in many eld of application such as automotive, construction, electronics, food and medical industries ranging from cable ties, fuel line applications for the automotive, metal coatings, advanced medical materials for prosthetic devices, 3D printing, additive manufacturing, laments such as bristles for toothbrushes, shoe soles for high-tech sport, cable housings, etc. 11 In this section a potential emerging application will be discussed.
The continuous growing demand of renewable energy solutions for portable smart electronic devices, experimented in the last two decades, has stimulated the development of advanced energy harvesting technologies from wasted energy sources. It could represent a promising alternative to fossil fuel. 103 In this context, devices based on piezoelectric materials can be a challenging alternative to convert mechanical energies into electricity for energetically autonomous wireless and electronic devices. 104 Indeed, piezoelectric materials are able to respond to both mechanical and electrical stimuli by producing energy or deforming mechanically depending on the nature of the stimulus. It is for this reason that these materials are of great interest and already have many applications in several areas such as piezoelectric transducers, sensors, actuators and energy harvesters. 105,106 Piezoelectric energy harvesters (PEHs) are regarded as promising independent renewable power sources for low-power electronic devices such as wireless sensors, portable devices, and medical implants. 107 Currently, commercial PEHs use piezoelectric ceramics with high piezoelectric coefficients providing excellent energy performance. However, these ceramics are still expensive and not easy scalable for industrial production of PEHs devices. 108 Moreover, the increasing demand for exible, translucent or transparent smart devices opens up for piezoelectric polymers which have lower piezoelectric coefficients, but they show advantages in terms of cost and large-scale processability. 109 The tensile piezoelectricity in stretched and poled lms of polyvinylidene uoride (PVDF) was rst demonstrated by Kawai in 1969. 110 This discovery triggered widely spread investigations on the pyro-, piezo-, and ferroelectricity of PVDF, its copolymers, nylons, and other polymers for subsequent years. 111,112 PAs are characterized by a repeating unit of [HN-(CH 2 ) x -CO] n , with odd or even number of carbons atoms between the amide groups. Odd numbered PAs contain pairs of -NH and -C]O groups aligned in the same direction that are able to form stable dipole moments, which leads to a polar structure that exhibit their observed ferroelectric, piezoelectric and pyroelectric behaviour. 113,114 In comparison, these groups in even numbered PAs are aligned in an alternating way, leading to the net cancelation of the dipoles along the polymeric chains. 115 In 1991, Takase et al. 116 studied the variation of the piezoelectric response of PA11 and PA7 with temperature, compared to that of PVDF. They observed the highest values of piezoelectric strain constant (d 31 ¼ 14 pC N À1 and 17 pC N À1 for PA11 and for PA7, respectively), stress constant (e 31 ¼ 21 mC m À2 and 27 mC m À2 for PA11 and for PA7, respectively) and electromechanical coupling coefficient (k 31 ¼ 0.054 for PA11 and 0.049 for PA7) of PA11 and PA7 lms at temperature ranges from 100 to 200 C. 103 This behaviour suggested that the orientation of dipoles depends on the temperature, so that above glass transition temperature (T g ), the mobility of the polymeric chains is high, and they are more sensitive to orientation induced by the electric eld applied. In the same work, Takase et al. 116 observed that the cooling rate have also a notable inuence on the piezoelectric constant, since the crystal structures depends on the processing condition. Using the same polarization method, if the material is quenched, the formation of the g phase is favoured while a slow cooling rate leads to the formation of the a phase.
Although all PA11 have a polar crystal structure due to its dipole's orientations, specic crystalline phase types can maximize the electric polarization. 114 It is known that g crystalline phase has the best piezoelectric response in PA11 even if in literature is reported that d 0 phase may also contribute. 117,118 In 1993, B. Z. Mei et al. found that the remanent polarization and the coercive eld of odd PAs series (PA5, PA7, PA9 and PA11) increase linearly with dipole density as does the melting point. It is important to specify that, decreasing the number of CH 2 groups in PAs chains, the concentration of amide groups increases as well as the dipole density. Their work showed also that both uniaxially oriented and unoriented odd-numbered PAs exhibit ferroelectric hysteresis behaviour but those uniaxial oriented show higher ferroelectric response. [119][120][121] Between odd PAs, PA11 can be synthesized by renewable resources (castor oil), thus contributing to have a much smaller carbon footprint and consequently decreasing the environmental impact of PEHs devices. 122 Apart from the excellent properties such as high exibility, optical transparency, relative cheapness, similarly to other biosourced polymers, PA11 has also other advantages, non-toxicity and biocompatibility. In this context, PA11 could be a good alternative for implants and biomedical health monitoring systems, which require devices to be biocompatible. 123 With the aim to increase the piezoelectric response of PA11, in literature are reported several strategies, such as different methods to process the material, i.e. the fabrication of selfpoled nanowires (Fig. 8(1) and (3)), which are highly exible, lightweight and sensitive to small vibrations, as well as mechanical stretching and annealing to increase the dipoles orientation and the degree of crystallinity of the piezoelectric crystalline phase of PA. 104,124 Another method reported in literature to improve the piezoelectric properties of PA11, is the incorporation of nano-llers such as piezoelectric lead zirconate titanate, 127 barium titanate, 128 layered silicates 129,130 and CNCs. 126 Cellulose is classied as natural piezoelectric material and its piezoelectric response have been studied in the last years, becoming a suitable material for energy harvesting device as well as storage application. The shear piezoelectricity of cellulose-based biopolymers such as wood, amylase, chitin, and starch can be comparable to that of quartz crystal. 131 In 2012, Csoka et al. 132 reported the piezoelectric response of ultrathin lms of aligned CNCs. Their piezoelectric properties were ascribed to the collective contribution of the asymmetric crystalline structure of the cellulose crystals, showing a shear piezoelectric constant value comparable to that of a reference piezoelectric metal oxide. Some year aer, Rajala et al. 133,134 prepared piezoelectric sensors from CNF lm and investigated their piezoelectric response. They found sensitivity values between 4.7 and 6.4 pC N À1 in ambient conditions. Aer comparing these results with PVDF-based sensor devices, the authors suggested CNF-based materials as a suitable precursor material for disposable piezoelectric sensors, actuators and energy generators with potential applications in the elds of electronics, sensors, and biomedical diagnostics. 103 More recently, F. Ram et al. 126 reported the fabrication of a exible piezoelectric energy harvester based on PA11 and CNCs by a solution casting process. They showed for the rst time in literature the possibility to enhance the electroactive g phase in PA11 using CNCs, resulting in the increase of the piezoelectric performance of the device (Fig. 8(2)). In fact, the incorporation of low concentration of CNCs, i.e. 2-5 wt% in the PA11 matrix resulted in the almost complete transition of aphase to the polar g-phase. As a result of the best formulation, energy harvesting devices made from PA11 reinforced with 5 wt% of CNCs showed about 2.6 times higher output voltage as compared to the devices composed by the neat matrix under similar impact conditions, and the effect was durable over 800 cycles. 126 Considering the high piezoelectric performance of PA11 and CNCs, together with their low cost, exibility, durability, and the possibility to process this material by an environmentally-friendly, sustainable and scalable method such as the reactive extrusion lead to consider them as promising materials for potential advanced applications in self powering sensors and powering of other small electronics.

Conclusions and perspectives
Biosourced polyamides are promising engineering thermoplastics fully or partially derived from renewable feedstock. As traditional PAs, they exhibit high mechanical strength and thermal performance showing added processing advantages and better properties. The general main challenges of incorporating cellulosic materials in polymeric matrices have largely impeded the successful preparation of nanocellulose-polyamide composites via large-scale melt processing methods. To our knowledge, the number of reported studies involving the manufacture of cellulose-polyamide composites at a large-scale is still limited but has gained interest in the past ve years being a cost efficient, sustainable, and organic solvent-free method. Wet feeding has been introduced as a means to overcome dispersion and nanocellulose agglomeration issues as well as its thermal degradation thanks to the plasticizer effect of water which is able to reduce PAs melting temperature. Moreover, water assisted compounding might enhance the degree of crystallinity which yielded further mechanical property benets. Furthermore, as discussed previously, biosourced PAs are good candidates for the development of advanced energy harvesting devices based on piezoelectric materials, being a challenging alternative to convert mechanical wasted energies into electricity as a promising alternative to fossil fuel. Considering the high piezoelectric performance of PAs based composites reinforced with cellulose, together with their low cost, excellent mechanical properties, durability, and the possibility to process this material by a solvent-free, sustainable and scalable method such as reactive extrusion and wet feeding, this materials are considered as promising materials for potential advanced applications in self powering sensors and powering of other small electronics.

Conflicts of interest
There are no conicts to declare.