Composite Cellulose Nanofibrils Filaments

IN

DEGREE PROJECT TECHNOLOGY, FIRST CYCLE, 15 CREDITS

STOCKHOLM SWEDEN 2020 ,

Composite Cellulose Nanofibrils Filaments

MARIUS DE MOURGUES

KTH ROYAL INSTITUTE OF TECHNOLOGY

SCHOOL OF INDUSTRIAL ENGINEERING AND MANAGEMENT

A BSTRACT

Biodegradable polymers are emerging as a new solution to satisfy the increasing demand of green environmentally friendly material. At the same time, the interest for lighter and stronger structures never stops growing. In this paper, we report the production steps to achieve cellulose nanofibrils (CNF) composite filaments via a new green synthesis route known as wet spinning. This new technique avoids the traditional harmful viscose process and produces biodegradable CNF filaments with interesting mechanical properties. This approach is then applied to produce never seen before composite CNF filaments using a three-layered head extruder. In order to obtain conductive filaments, PEDOT/PPS is successfully mixed with CNF to produce in-situ composite filaments. Scanning electron microscopy (SEM), atomic force measurements and tensile tests are employed to characterize the properties of the filaments.

S AMMANFATTNING

Biologiskt nedbrytbara polymerer börjar framträda som en lösning för det ökade behovet av

miljövänliga material. Samtidigt så växer intresset för lättare och starkare strukturer. I denna rapport

tar vi upp produktionsstegen för att uppnå nanofibril komposit cellulosa fibrer (CNF), med hjälp av

en ny grön polymerization mest känd som ”wet spinning”. Med denna nya teknik så behövs inte

dem traditionella miljöfarliga viskosprocesserna och man producerar biologiskt nedbrytbara CNF

filaments med intressanta mekaniska egenskaper. Denna metod appliceras sen för att producera en

komposit som aldrig setts innan. CNF fibrer som består av tre lager ”head-extruder”. För att få

fibrer med ledningsförmåga så mixas PEDOT/PPS med CNF för att producera ”in-situ komposit

fibrer”. Svepelektronmikroskop (SEM), atomkraftsmikroskopi och töjningstester används för att

karaktärisera egenskaperna av fibrerna.

T ABLE OF CONTENT

1- Introduction ... - 1 -

1.1 Purpose ... - 1 -

1.2 Goal ... - 1 -

1.3 Background ... - 1 -

2- Formation of CNF filaments ... - 2 -

2.1 Nanocellulose ... - 2 -

2.1.1 Cellulose Nanocrystals ... - 4 -

2.1.2 Cellulose Nanofibrils ... - 4 -

2.1.3 Nanocellulose Applications ... - 5 -

2.2 Cellulose Filaments ... - 6 -

2.3 Cellulose Nanofibrils Filaments ... - 7 -

2.3.1 Spinning processes ... - 7 -

2.3.2 Spinning parameters ... - 8 -

2.3.3 Post treatment ... - 9 -

2.3.4 Mechanical properties ... - 9 -

3- Laboratory work and Experimental Procedure ... - 10 -

3.1 Motivation ... - 10 -

3.2 Experimental Plan ... - 10 -

3.2.1 CNF filaments ... - 10 -

3.2.2 Hollow and Composite functionalized filaments ... - 11 -

3.2.3 SEM, AFM and Tensile tests settings ... - 11 -

4- Results and Discussion ... - 12 -

4.1 Native coaxial filaments ... - 12 -

4.2 Triaxial Filaments ... - 14 -

4.2.1 Hollow CNF filaments ... - 14 -

4.2.2 PEDOT/PSS CNF filaments ... - 15 -

4.2.3 PEDOT/PSS composite filaments application ... - 17 -

4.3 Social, environmental and ethical aspects ... - 18 -

5- Conclusion and new perspectives ... - 19 -

5.1 Conclusion ... - 19 -

5.2 Future work ... - 19 -

6- Acknowledgement ... - 20 -

7- References ... - 21 -

1- I NTRODUCTION 1.1 P URPOSE

In a world where the demand for eco-friendly and green material does not cease to increase, composites produced from renewable resources might become the new standard in a near future.

Lighter, stiffer and stronger; composite material’s market share grows day by day while fossil resources are inexorably diminishing. The search for biodegradable composites produced from inexhaustible resources is obviously set. Economic viability, processing ease, mechanical stiffness and biodegradability are setting natural fibres like cellulose nanofibrils filaments composites as potential candidates to satisfy this increasing demand.

1.2 G OAL

The aim of this project is to create composite cellulose filaments in situ via wet spinning, for potential applications including advanced textiles, biomedical devices and electronics. Composite filaments will take advantage of the excellent mechanical properties of cellulose nanofibrils (CNFs), in addition to other additives (polymers, nanoparticles) to target specific properties, such as conductivity, stretch- ability and wet-strength.

1.3 B ACKGROUND

Nanocellulose has generated great interest over the last decade thanks to its wide range of applications

and the numerous ways to produce it easily from natural resources.[1] This recent breakthrough was

enabled thanks to the discovery of chemical pre-treatment methods facilitating the extraction of

cellulose from sources like wood, cotton or plants. After being extracted, cellulose is used as fibril

shape or crystal form depending on their function. This project targets cellulose nanofibrils (CNF)

and the process of creating filaments from CNF. To produce composite filaments via wet-spinning, a

co-extruder is employed to include the polymer within the CNF and increase specific mechanical

properties. Once the background knowledge is settled, the experimental procedure to achieve two-

layer and three-layer CNFs based filaments will be investigated. The result’s section includes the

characterization of regular CNF filaments by scanning electron microscopy, it allows us to appreciate

the surface aspect of both CNF filaments and composite filaments. Tensile tests are on the other hand

delivering information on the intrinsic mechanical properties of CNF filaments. Afterward, results

provided by previous experiments will be thoroughly analysed to relate the potential outcomes of

these never seen before CNF composite filaments. A final section will be dedicated to the composite

filaments combinations that have not been reviewed in this work as well as the general improvements

to be made if this project was to be continued.

2- F ORMATION OF CNF FILAMENTS

2.1 N ANOCELLULOSE

While cellulose in the form of fibres and dissolved polymer have been used for centuries to create paper, textiles, recently cellulose nanomaterials (10 – 1000 nm) have become of great interest for advanced applications due to its superior mechanical properties. Cellulose is the most abundant natural polymer on earth, representing 35-50% of the Earth’s plant-based biomass.[2] Consisting of β 1-4 linked glucose units, cellulose is a linear polysaccharide that forms the main structural component in the plant cell wall and can be extracted from plants, bacteria and animals. Cellulose is mainly extracted from wood. In trees, cellulose forms a hierarchical structure, whereby cellulose fibres (20 µm – 10 mm) includes smaller elementary fibrils (10 nm) that are composed of linearly arranged cellulose chains. This scaling breakdown is presented in figure 1[3].

Figure 1 – Cellulose origin scaled down to the molecular structure[3]

Cellulose microfibrils consists of amorphous (disordered) and crystalline (ordered) regions of

individual cellulose chains.[4] The flexibility and plasticity of the cellulose chains is partially due to

the disordered regions while ordered regions contribute to the stiffness and the elasticity of the

molecule. The crystallinity and the degree of polymerization (DP) of the microfibrils depends on the

cellulose source, for example wood or cotton. The hydroxyl groups bounded to the main carbon

chains provide a variety of functions to cellulose molecules including hydrophilicity and insolubility

in most solvents. Each end of the cellulose molecule are chemically different and the polar groups

being not symmetrical provide hydrophilicity.[5] In addition to Van der Waals forces, the hydroxyl

groups allow for hydrogen bonds between the hydroxyl groups of the chains, leading to parallel

stacking of chains to create microfibrils. This function is generally ensured by the C6 hydroxyl group

(figure 2,[4]). The freedom of C6 rotation gives rise to several internal and external interactions as

well as different conformation configuration, each of them associated to a cellulose subcategory like

Cellulose I or cellulose II.

Figure 2 - Cellulose molecular structure[4]

In general, cellulose nanomaterials can be classified as cellulose nanocrystals (CNCs) or cellulose nanofibrils (CNFs). Prior to the chemical-mechanical process to extract CNC or CNF, the cellulose source (wood, cotton, plants) is initially broken down to the microscale mechanically. This mechanical process is then followed by the isolation step to reach the nanoscale.

“Isolation” designates the processing step when the microfibril cellulose is scattered into smaller particles. The first isolation stage consists in a purification and a pre-treatment process to facilitate the separation of cellulose molecular chains. Several techniques are employed to reach the molecular scale during the second step, the mechanical process and the acid hydrolysis are the most popular.

Among the numerous breakthroughs in the past years, the pre-treatment process is undoubtedly an innovation that allowed CNC and CNF to find feasible applications. Before that, the mechanical processes were too energy-consuming to really expect outcomes from CNF or CNC. In addition to CNC and CNF from plat sources, cellulose fibres are secreted by some bacteria. These bacterial cellulose (BC) fibres are characterized by a higher aspect ratio than CNF or CNC. BC is commonly used in biomedical applications such as drug delivery because it is non-toxic and can easily degrade in the human body. Recent studies investigated the possibility of prototype blood vessels made of BC fibres.[6] These fibres are also extensively used in food packaging and textile application.

The figure 3 allows us to evaluate the structural differences between the CNC and the CNF thanks to an AFM device. The a) picture shows an even distribution of CNC. The aspect ratio of CNC is drastically lower than on the b) image representing cellulose nanofibrils. The CNF have smaller diameters but longer cellulose chains.

Figure 3 – CNC (left) and CNF (right) images obtained with AFM

The contrast highlights several nanofibrils entanglements. One could also notice that some fibrils have a larger diameter than others, but the thickest threads are nothing else than fibrils bonded together attesting to its strong hydrogen bonds network.

a) b)

2.1.1 C ELLULOSE N ANOCRYSTALS

Nanocellulose offers a high specific mechanical strength and stiffness, is thermally stable with a low thermal expansion coefficient.[7] It is said to replace composites like glass fibre in the future thanks to its biodegradability. However, challenges remain like achieving long fibres of cellulose or reaching the same mechanical properties as glass fibres for example.

Cellulose nanocrystals (CNC) are extracted via sulfuric acid hydrolysis. Degradation by sulfuric acid occurs specifically in the amorphous regions breaking the cellulose chains apart and leaving the crystalline regions intact just like the scheme showed in the figure 4 [3].

Figure 4 – Sulfuric acid hydrolysis degrading the amorphous areas [3]

Needle-shape crystal particles commonly produced have a radius of 3-35nm and a length of 200- 500nm.[4] The DP decrease is directly correlated to the cellulose source (wood, cotton). Recent studies have been investigating a degradation process including enzymatic hydrolysis combined with mechanical shearing to obtain CNC.[8] Cellulose nanocrystals have noticeably been employed as particulate reinforcement fillers in polymer matrix like PLA to improve the mechanical properties.

Elastic modulus of CNC was found to reach 150 GPa and the theoretical tensile strength measured was 7.5-7.7 GPA.[9] Such mechanical properties open new horizons of applications if CNC is used as a reinforcement agent. CNC hydrogels-based also generated interest from the biomedical department due to their high specific area and open pores.

2.1.2 C ELLULOSE N ANOFIBRILS

Cellulose nanofibrils (CNF) are extracted by chemical-mechanical processes, whereby pre-treated macroscopic cellulose fibres undergo intense mechanical shear via homogenization or grinding.

CNFs are chemically treated such that anionic carboxyl groups are grafted to the surface of the fibril, in order to reduce the energy required for separation during mechanical processing. Specifically, repulsion between the charged groups on the cellulose fibre help liberate individual microfibrils.

Commonly, this is achieved by oxidation with 2,2,6,6-tetra-methylpiperidinyl-1-oxyl (TEMPO).

Using this method, the hydroxyl group at the C6 position (figure 2) of the cellulose chain is replaced by a carboxyl group.[10] In addition to reducing energy costs, charged groups on the CNF surface provide electrostatic stabilization, allowing CNFs to form stable colloidal dispersions. Wood fibres pre-treated with TEMPO oxidation method can reach an aspect ratio of 300, 4 times more than CNC’s aspect ratio. However their high specific surface area and surface charge make them more sensitive to water because more hydroxyl groups are prone to interact with water molecules via hydrogen bonds.[11] Another effect of pre-treatment is to enhance fibril orientation during wet-spinning resulting in a higher aspect ratio and a larger surface charge.

After this pre-treatment stage mechanical processes are used to separate the fibres from the microfibril state. These processes typically include high pressure homogenizers and grinders/refiners.[3]

Homogenization is a process in which cellulose fibres are forced through small holes with a specific

geometry. High pressure is applied and the high shear along the longitudinal axis of the microfibrils enables the extraction of long fibres. The high energy consumption of this method was the main challenge for commercialization until pre-treatment methods appeared on the market and facilitated the process. Nanofibrils of 20-100nm in diameter and 10-30m in length are extracted by this method.

Increasing the number of homogenizations further reduces CNFs size. Grinding is another technique used to extract CNF where slurry cellulose is sheared between one static and one rotating grinding disk. The shearing force generated leads to fibres extruded from the outlet suspension. Many other methods exist but grinding and homogenization are the most commonly used. Some post-treatments are then applied on the fibre to ensure good compatibility with the matrix or chemical modification to enhance the surface properties.[4] In general CNF is dispersed and stored in water once extracted because dry wood pulp (extracted cellulose) usually leads to scattered particles instead of long fibre with a higher crystallinity and aspect ratio. These two parameters are essential for the mechanical properties further obtained.

2.1.3 N ANOCELLULOSE A PPLICATIONS

Once produced, CNF suspensions can be used to create different structures such as films, aerogels or hydrogels. The suspensions properties highly depend on the processing method. Yet, independently from the production technique, suspensions carry a gel-like and thixotropic behaviour ie. the longer shear stress is applied, the lower the viscosity gets. This characteristic is also called shear-thinning.

Their complicated rheological behaviour is justified by their inner flow differences because of the partially amorphous structure.[4] While CNF suspensions are dried to produce films with a high level of transparency, cellulose filaments are extruded from hydrogel solutions. Hydrogel consist of CNF suspensions at very low concentration (1-2%) in an aqueous solution. Due to the negative charges, the repulsive interactions enable homogeneous CNF suspensions to be stable in gel solution.

Moreover their high aspect ratio increases the solution’s viscosity without decreasing the CNF concentration and thus create hydrogel with no fluidity.[10]

Figure 5 – Nanocellulose applications at different scale [12] [13] [14]

Many applications arose with biodegradable films, aerogels and hydrogels in several fields such as tissue engineering or drug delivery. The figure 5 shows different applications of nanocellulose from the nanoscale to the macroscale. At the nanoscale the frame is showing chiral plasmonics particles using twisting along cellulose nanocrystals as a template for gold particles. The a and b figures are presenting cellulose nanocrystals which thanks to their amphiphilic properties are stabilizing oil/water interfaces. At the macroscale, applications like 3D bioprinting of human cartilage tissue

J. Majoinen, et al., Adv. Mater. 2016, 28, 5262.

Nano Micro Macro

I. Kalashnikova et al. Biomacromolecules 2012, 13, 267−275

H. Rosilo, et al. Nanoscale 2014, 6, 11871.

R. T. Olsson, et al. Nat.

Nanotechnol. 2010, 5, 584 M. S. Reid, et al., ACS Macro Lett. 2019, 8, 1334.

K. Markstedt, et al. Biomacromolecules 2015, 16, 1489

Tardy et al. Adv. Funct. Mater. 2019, 29, 1808518

with nanocellulose can be observed (D, E and F images). Also at the macroscale, a study investigated the interfacial properties of cellulose nanocrystals in order to predict the interactions with a polymer matrix and act as reinforcement agents.[15] Finally Tardy et Al studied the tessellation of chiral- nematic film by micro templating to improve the mechanical properties. The tessellation is performed by controlling the capillary stress during the assembly of CNC films.

Studies also revealed that CNF suspensions could indeed be used as high gas barrier films on polylactide(PLA) surface thanks to their efficient oxygen impermeability.[16] Films have also been extensively investigated in food packaging applications where the CNF surface was silylated with hexamethyldisiloxane (HMDSO) via plasma deposition to ensure a hydrophobic film.[17] Another common application of CNF suspensions is the paper industry. It has been shown that CNF coated via a roll-to-roll process on papers improved drastically the grease and oxygen resistance.[18] On the other hand aerogels were studied to replace regular polystyrene foam. They are produced from CNF suspensions just like hydrogels, but the liquid medium is replaced by air. Somehow, hydrogels remain the most common form of nanocellulose. CNF filaments are produced from hydrogel with very low concentration of cellulose.

2.2 C ELLULOSE F ILAMENTS

Traditional cellulose filaments are made from dissolved cellulose whereby cellulose is broken down into individual cellulose chains via chemical modifications and solvents. This is largely achieved at the industrial scale via the viscose process. Viscose method can process cellulose from wood pulp whereas other treatments can only produce lignin-based cellulose. In this case cellulose reacts with sodium hydroxide to form alkali cellulose and is later converted into cellulose xanthate. The solution is then spinned and the cellulose chains aggregate into an amorphous fibre of cellulose. The main con of the viscose process is the pollution generated by carbon disulphide. Theoretically, this solvent is recovered during the manufacturing process however, incidents leading to poisoning occurred in the past which led to legal cases.[19] In addition to its environment impact, this process employs cheap raw materials, but the high energy and water consumption still make it expensive.

Regenerated cellulose filaments made by this process are also called rayon and are the oldest cellulose filaments ever made.[20] Its nature has been widely debated whether it should be considered as a synthetic or a natural fibre. Viscose rayon is also known as artificial silk or regular rayon. Other sorts of rayon fibres can be found on the market like acetate cellulose or Bemberg rayon fibres. These fibres presented interesting properties to replace regular cotton and silk in textiles and woven products. However, the low mechanical performances reduce their application range. Even if rayon has a lower environmental impact than cotton in terms of degradability, the energy intensive consumption and the harmful solvents are two reasons why this synthesis route is slowly abandoned.

Recent studies also stated that viscose fibres often end up in the ocean in the form of micro- plastics.[21] An alternative to the viscose process is the lyocell technique relying on cellulose dissolution and dry-spinning.[20]

In the end cellulose filaments present poorer mechanical properties and a higher environmental

impact compared to cellulose nanofibrils filaments.

2.3 C ELLULOSE N ANOFIBRILS F ILAMENTS

Compared to regular cellulose filaments, cellulose nanofibrils (CNF) filaments have a highly ordered aligned structure of cellulose chains providing greater stiffness. Cellulose filaments have been produced for many years by the viscose process which can have negative health and environment impact. However, cellulose dissolution and regeneration included in this process are no longer needed as the spinning technique, a revolutionizing solvent-based process offering a good renewable and sustainable alternative to produce filaments has been discovered. Spinning technique are either performed by wet spinning or dry spinning depending on the melting point of the polymer. Many parameters are influencing the quality of the fibres extruded. The shearing rate directly impacts the alignment of CNF and so does the pressure in the cellulose chamber or the diameter of the extruder.

Parallel disposition of CNF facilitates strong hydrogen interactions in between the fibres and thus improved mechanical properties.

2.3.1 S PINNING PROCESSES

Two spinning mechanisms are commonly used to produce CNF filaments. Alternatively, melt spinning is often employed with polymers that can melt. It is not the case for CNF because its degradation temperature is above its melting point. It would burn before melting.

Dry spinning technique is based on the dissolution of the CNF via a solvent. After purification, the solution is extruded by a spinneret into a warm air chamber where the solvent is evaporated. The dry solid CNF remains and is post treated to improve its mechanical properties. The main drawback of this methods is the hazardous effect of solvents.[22]

Wet-spinning process is also a solvent-based technique. In this work CNF filaments will be produced via wet-spinning of CNF dispersions using a co-axial extruder. This process includes numerous steps such as: coagulation, pre-drying and washing. Under continuous flow, CNFs and HCl are co- extruded, whereby the low pH environment protonates the CNF carboxyl groups, eliminating electrostatic double layer forces between the COO - groups and causing the fibrils to aggregate into an aligned cellulose filament (Figure 6). CNF are then dried either at room temperature like in our experiment or in an oven. This pre-drying step facilitates the formation of hydrogen bonds along the filaments.[23] In the wet state, hydroxyl groups on the cellulose chains interact with water via hydrogen bonds rather than creating inner interactions. As the filaments dry, the water content decreases and the hydroxyl groups are forced to interact between each other resulting in a stiffer filament. The washing step is not mandatory but it can help to remove unwanted residues like chloride ions bonded to the surface.

Figure 6 – CNF filaments extruded by wet spinning

Additional techniques have been developed to increase the aspect ratio of the filaments produced via wet spinning by reducing their diameter. A stretching process was especially introduced to enhance

HCl

HCl

CNF

H ⁺

H ⁺

H ⁺

H ⁺

H ⁺

the alignment. While the filaments are coagulating in the low pH bath, a stretching device bonded to the filaments end ensures a continuous flow. Filaments are generally extended by 10 to 20%.

Promising results were obtained with this stretching method including a higher tensile strength and stiffness compared to non-stretched filaments.[23] Magnetic field or electric field have also been employed in electro spinning taking advantage of the surface charges of CNF to orient fibres parallelly.

2.3.2 S PINNING PARAMETERS

Many parameters can influence the CNF filaments properties during the spinning process. To begin with, the solid volume fraction of CNF in hydrogels is obviously affecting the solution viscosity. A low CNF content in dispersion can lead to void formation and thus discontinuities in the filament formation. CNF concentration can be lowered down to 0,4% for non-pre-treated filaments, under such concentration, the risk of having discontinuities increases seriously. TEMPO oxidation can possibly enable lower CNF content hydrogels to be extruded. Generally, a lower free volume in the CNF content seems to improve the fibres alignment. On the other hand, a too high CNF content will surpass the spinnability limit and won’t be sheared correctly. The HCl concentration also influences the packing factor of the fibre when coagulation occurs. The higher the acid concentration is, the lower the repulsive interactions intensity between microfibrils will be and therefore filaments aggregation will be stronger.

The shear rate of the spinneret is the main factor responsible for fibre alignment. A perfect theoretical shearing rate is hard to predict as the CNF shape is never constant, but studies demonstrated that a 200 s -1 shearing rate was particularly effective to maximize mechanical properties. Below this value the fibre alignment is less effective (see figure 7).[24]

Figure 7 – CNF Filament Young’s modulus against spinning shear rate[25]

The figure above shows a constant increase of the stiffness with the shearing rate whereas the tensile strength is reaching a plateau at 200 s -1 . It is proposed that a higher shearing rate would generate voids and decrease the aspect ratio.

Finally, the drying step is crucial. Filaments should be dried under a continuous tensile stress along the fibre direction to align microfibrils.[26] Several minutes are needed for the filaments to coagulate.

Some studies have been advising a drying time of up to 5 hours. The impact of drying filaments in an

oven has not demonstrated impressive enhancement in the mechanical properties compared to the

energy consumption it demands.[10]

2.3.3 P OST TREATMENT

Like pre-treatment, the filaments can also undergo a post treatment after the drying procedure.

However, instead of reducing the energy demand to break the fibres like the TEMPO oxidation method, post treatment provides new characteristics, especially on the surface of the filaments by affecting the specific surface area (SSA), the volume, the size and the distributions of the surface pores. Such characteristics can be obtained from low temperature nitrogen absorption. Chemical modification can occur such as physical adsorption, polymer grafting or molecular grafting.[4]

Studies have investigated the effect of acidic treatment. The SSA is increasing as the acid content increases. The most significant results were obtained with nitric acid (HNO 3 ) treatment in a boiling bath at 60 wt%. The SSA increased from 314 m 2 /gr to 349 m 2 /gr.[27] In general a higher SSA allows better interactions between fillers additives and resin matrix; and thus a good load transfer when stress is applied.

2.3.4 M ECHANICAL PROPERTIES

CNF turned out having interesting properties. Indeed, the alignment of the nanofibrils, the parallel

stacking of cellulose chains makes this defect-free molecular structure much stronger than regular

cellulose filaments. Some experiments even led to an 86 GPa Young’s Modulus which is higher than

a regular fibre glass modulus (70 MPa) and a tensile strength of 1,57 GPa for the filaments have been

reported in the literature.[28] The mechanical properties are affected by a number of factors including

the pre-treatment method, the aspect ratio which is directly related to the degree of polymerization,

the additives and other factors.

3- L ABORATORY WORK AND E XPERIMENTAL P ROCEDURE 3.1 M OTIVATION

Achieving regular CNF filaments has already been widely discussed in the literature. Therefore, the aim of this report is to compare classic CNF filaments with CNF filaments produced via wet spinning using a three-layered extruding head (Figure 8). Using this set-up, we have the potential to create multifunctional materials using scalable methods. This device could open a new area of research for cellulose-filaments-based composites. The goal is to target a specific property like conductivity or stretch ability and try to improve CNF filament current characteristics towards this objective. CNF will be co-extruded with another polymer and a solvent. Replacing the viscose process by this wet spinning technique enables a greener product of the CNF filaments without using harmful solvents.

By doing an in-situ extraction, post treatments are no longer necessary as the two different materials are co-extruded at the same which translates in costs reduction.

The first goal of the experiment was to achieve regular CNF filaments via a co-axial wet spinning set up (figure 8). Hollow CNF filaments were then performed using the three-layer head extruder and finally PEDOT/PSS composite filaments were extruded.

3.2 E XPERIMENTAL P LAN

3.2.1 CNF FILAMENTS

As you can observe in figure 6, the core spinning solution is CNF and the sheath spinning solution is HCl. The CNF solution is prepared using a two-barrel nozzle with a 1wt% CNF dispersion flowing at a rate of 10mL/hour along the inner channel. The 0,01 M HCl solution flows at 50 mL/hour through the syringe along the outer shell. The setup for CNF filaments production is showed in figure 8.

Before starting the experiment, the CNF pipe must be carefully inspected to avoid any bubbles. If so, pressure is added in the syringe until bubbles collapse or disappear. Bubbles directly impact the filaments properties by creating discontinuities. The two pipes are connected to the co axial extruder by two outputs respectively supplying the shells. The filaments are injected in a coagulation bath where water is mixed with an HCl beaker to reach pH 2.

Figure 8 – Co-axial wet spinning setup[29]

Once the setup is finished, the experiment starts, pressure is progressively applied on each syringe

according to the flow’s limit. The first filament might contain a few impurities that can be removed

before collection. Only a few seconds are required for the filaments to aggregate when they enter in

the acid bath. When a nice flow is observed trough the bath, the filaments can be collected after

aggregation either by a collector spinning (figure 8) or with a tweezer if it is just to make some measurements and long filaments are not required. 30 cm sections of the filament were removed from the bath and hung vertically between two support beams. Filaments are then dried at room temperature for 24 hours to increase the internal hydrogen bonds. Some filaments might break during the drying procedure because of impurities or pores.

3.2.2 H OLLOW AND C OMPOSITE FUNCTIONALIZED FILAMENTS

Before extruding the three-layered filaments, hollow filaments were produced with a three-barrel nozzle using a 1 wt% CNF dispersion flowing at a rate of 10 mL/h in the second layer of the nozzle.

HCl at a concentration of 0.01 M was flowed through the innermost and outermost channels at a rate of 5 mL/h and 50 mL/h, respectively.

Finally, in order to achieve conductive filaments a 1 wt% PEDOT/PSS dispersion, at pH 2, was flowed through the innermost channel at a rate of 5 mL/h; CNF serves as the coating sheaths and HCl is covering the CNF as the third layer to ensure aggregation. To visualize the wet structure of the hollow CNF filaments, wet filaments were immersed in liquid nitrogen and subsequently freeze dried.

The scheme of the procedure is described in figure 9. The HCl plays the same role as in the regular CNF filaments by protonating the COO- groups into COOH groups. Thus, causing aggregation of CNF around the PEDOT/PSS polymer.

Figure 9 – Three layered head with PEDOT/PSS core

3.2.3 SEM, AFM AND T ENSILE TESTS SETTINGS

Native coaxial filaments could be observed with a scanning electron microscope (SEM) in order to obtain a good overview of the surface. Solid and hollow CNF filaments were imaged using a S-4800 field emission scanning electron microscope (SEM). Samples were coated with Pt/Pd in a Cressington 208 HR sputter coater for 20s to reduce charging during imaging. The images were then collected with an accelerating voltage and emission current of 1.0 kV and 10 µA respectively.

Atomic Force Microscopy (AFM) images of dried CNF filaments were collected under ambient conditions using a MultiMode 8 (Bruker, Santa Barbara, CA) in TappingMode with RTESP-150 cantilevers having a nominal spring constant and resonant frequency 5 N/m and 150 kHz respectively.

The tensile properties of solid CNF filaments were measured using an Instron 5944, equipped with a 500 N load cell. Filaments were mounted in paper supports which were cut prior to measurement.

The Young’s modulus and the ultimate strength were averaged from a minimum of 10 filaments.

4- R ESULTS AND D ISCUSSION 4.1 N ATIVE COAXIAL FILAMENTS

The first part of the results consists of regular CNF images obtained by scanning electron

microscopy. Thanks to these SEM images displayed in figure 10, it is easy to observe the nanofibril alignment along the filament direction at the micro-scale.

Figure 10 - Regular CNF filaments pictures performed by SEM

The surface appears rather rough but fibres are homogeneously distributed. The edges on the c) image are fractured but still uniform. Large wrinkles are the result of the drying process. As water evaporates from the filament, the diameter of the filament significantly reduces, leading to buckling and wrinkling of the filament surface. The diameter of the filaments according to the scale is 40 m which sets the CNF filament’s diameter just above a glass fibre diameter measuring from 5 to 24 m.[30]

The nanofibrils appear densely packed and bonded by the hydrogen interactions. No pores or voids can be observed.

Figure 11 is a 3D AFM images allowing us to observe the wrinkles mentioned from the SEM images.

Wrinkles are in the range of 100s of nm in height. The amplitude error image shows distinctly the CNF filaments aligned along the axis of the flow.

10 µm 10 µm

5 µm 1 µm

a) b)

c) d)

Figure 11 - CNF filaments images performed by AFM

The second part of the results targets the mechanical properties of the CNF filaments. Figure 12 is providing the results from a several tensile tests.

Figure 12 – Stress strain curve of regular CNF filaments

From a mechanical point of view tensile tests revealed an ultimate strength of 249 MPa with a variance of 60MPa. Young Modulus was measured from the curve’s slope and attained 18 GPa (+- 3GPa). The maximum strain obtained reaches approximately 4%. Such a low deformation sets CNF polymer in the very brittle category. As an order of magnitude, elastomers can reach a deformation of 700%. Regular plastics are mostly around 300%.[31] Compared to glass fibres reaching a stiffness of 70 GPa, CNF filaments have a consequently lower Young modulus. Yet, if we compare them against regular polymers like polyethylene or polyamide that respectively have a Young modulus of 0,3 and 2,5 GPa[32]; then CNF filaments are much stiffer. It is interesting to notice that CNF filaments behave similarly to metals. An elastic deformation region is clearly identified from 0 to 1% of deformation. The plastic deformation region can then be observed from 1% to 4% of strain deformation. The plastic deformation of CNF filaments must involve secondary interactions after cellulose chains are broken. One could suppose that this plastic deformation region is enabled by the amorphous regions in the cellulose molecule. Crystalline regions might be preserved by the rearrangement of these amorphous parts. On the other hand, glass fibre typically breaks suddenly with nearly no plastic deformation around 5% strain.

0 50 100 150 200 250 300

0 1 2 3 4 5

Stress (MPa)

Strain (%)

A recent study demonstrated that CNF filaments could reach a Young’s modulus of 86 GPa and a tensile strength of 1,57 GPa.[28] Such results were achieved thanks to a nearly perfect alignment of the microfibrils and a very good packing factor due to crosslinks between the chains. The technique employed was flow-assisted assembly with a two-layer extruder. CNF is flowing in the core while a first sheath flow of deionized water is added to ensure repulsive interactions, to prevent glass transition of the CNF if it touches the channel walls and to enhance the fibril’s alignment. A second flow of HCl is then covering the CNF to reduce the interactions. The deionized flow could be one of the reasons why this study obtained better results than our experiments. The presence of pores and defects could also have highly affected our results.

Somehow all these results already prove the potential of CNF filaments in many applications like medical biodegradable stiches able to withstand the stress, the textile industry where CNF filaments could replace cotton or polyesters polymer or even in composites if CNF is mixed with a thermoplastic matrix. Indeed, thanks to its high aspect ratio and large specific surface area, CNF can be used as fibre reinforcements where the filaments can be combined to form a tow. Studies including CNF filaments in epoxy matrix showed convincing results thanks to the good adhesion between the polymer matrix and oxygen atoms grafted to the CNF filament to increase the surface free energy of the nanofibers.[33]

4.2 T RIAXIAL F ILAMENTS

4.2.1 H OLLOW CNF FILAMENTS

Additional pictures were obtained by SEM. The first images from figure 13 are presenting hollow filaments without PEDOT/PSS extracted by the three-layered head with the CNF. The CNF was injected in the second shell and HCl was flowing in the core and in the 3 rd layer.

Figure 13 – Freeze- and air-dried hollow CNF filaments produced by the three-layered extruder The a) picture is presenting a freeze-dried hollow filament and the b) image shows a freeze-dried hollow filament between two air-dried hollow filaments. The freeze-dried filaments were directly stored in liquid nitrogen after aggregation took place in the acid bath. This process reduces ice crystal formation and better preserves the structure in the liquid state. After freezing the filaments were imaged via SEM. The b) SEM picture allows us to compare air-dried and freeze-dried filaments. Air dried filaments are drastically shrinking. This important decrease in size indicates how the filaments are still full of water when picked from the bath. The freeze filaments have a diameter of 1 mm whereas the air-dried filament’s diameter is only 0,1 to 0,2mm. In other words, the filaments are shrinking by 10 times when air dried. If one was considering the volume reduction it would mean that the air-dried filaments represent 1% of the filament volume in freeze state. This reduction in size

1 mm 1 mm

a) b)

can be related to the wrinkles that we see in the 2 axial filaments. This dramatic reduction in diameter causes folding and buckling of the filament surface. Indeed, it is interesting to observe that the core of the air-dried CNF collapsed and the filament is wrapped on itself. Additionally, it appears that the CNFs collapse into sheet-like structures. However, when looking at higher magnification images (figure 14) one could observe individual CNFs forming a web-like structure across some of the pores (picture c)). This web-like structure would not be possible with CNCs as they are too short to entangle.

Overall there is still an alignment of the fibres along the axis of the filament.

It is interesting to notice that the inner surface of the CNF filament appears smooth and defect free whereas the outer surface looks much more porous and rougher. Figure 14 is obviously exposing the contrast between this porous outer surface and the inner surface of the filaments. There is no clear explanation to why the surfaces have different aspects. One could however suppose that the HCl flowing through the core could not be compressed and thus the pressure in the core developed a force counter acting the aggregation mechanism resulting in this smooth surface. On the outer surface where the pressure is lighter, the filaments could aggregate and arrange themselves without the pressure’s constrain.

Figure 14 – Freeze- and air-dried hollow CNF filament surface image obtained by SEM

4.2.2 PEDOT/PSS CNF FILAMENTS

Conducting polymers gained interest in the recent years as they are expected to meet the increasing energy demand by storing and conducting energy. Electrical storage could be performed in larger electrochemical cells capable to be discharged through high transistor circuits. However, these objectives remain unattained because material with ion and electronic conductivity are rare. Some material like ceramics have a high ion conductivity. On the other hand, organic polymers present excellent electronic conductivity but none of them are gathering successfully both properties to form a mixed-ionic electronic conductor (MIEC). In this search for the relevant material, CNF composites with PEDOT/PSS could possibly fulfil these expectations while remaining a green material.

Following the same procedure as the hollow filaments, three-layer filaments with PEDOT/PSS in the core were produced. Poly(3,4-ethylenedioxythiophene)/poly(styrenesulfonate) (PEDOT/PSS) is a mix of two polymers generating a great interest in the industry thanks to its lightweight, processability ease, low costs and conductivity. Styrene sulfonate (PSS) is deprotonated and carries a negative charge whereas PEDOT is a conjugated ionomer carrying a positive charge. This charge difference between PEDOT and PSS clusters governs the conductivity and drives the charge transport (figure 15). Thus, the electronic conductivity of this co-polymer is highly dependent upon the phase separation.[34] To hinder this dependence, additional high melting point solvent treatments using N-

10 um

c)

50 um

d)

methyl pyrrolidone or Di-methylsulfoxyde (DMSO) are performed on PEDOT/PSS to increase their conductivity by several orders of magnitude.[35] Indeed DMSO impacts the nanostructure of PEDOT/PSS by increasing the crystallinity and the phase separation of PSS.[36]

Figure 15 – PEDOT/PSS molecular structure[37]

In solution state, PEDOT/PSS particles are segregated between PEDOT rich region and PSS rich region (figure 15 – (b)). In films, PEDOT-clusters are surrounded by PSS-rich region. The adsorption of PEDOT/PSS on cellulose film is highly dependent on the pH of the solution. It has been shown that a low pH (around 2) was favouring the adsorption thanks to the low pKa of the carboxylic groups on the cellulose surface.[38] At this pH, the negatively charged sulfonate groups bonded to the PEDOT/PSS surface and the negatively charged microfibrils are not anymore repelling each other due to electrostatic repulsions. The protonation of cellulose facilitates the adsorption. Additional work has studied the geometrical influence of the cellulose on the adsorption, but nothing was proved except the influence of the substitution degree. The degree of substitution determines the number of hydroxyl group per monomer. For an optimal adsorption of PEDOT/PSS, this number should not overwhelm 0,35[34]. Above this limit the hydroxyl concentration is so high that CNF is dissolved in a polyelectrolyte form (CMC). In conductive paper the percentage of PEDOT represents 20 wt % while the remaining part is a non-conductive cellulose/PSS component.

The final mixture is a conductive, transparent organic material used in several applications such as paper reinforcement via spin coating, emulsion gels, organic transistors, fuel cells or supercapacitor electrodes.[39]

Figure 16 – Extraction of PEDOT/PSS/CNF filaments in the acidic bath

Figure 16 shows a picture of how the composite filaments were extracted in the acid bath. The dark

blue component is the PEDOT/PSS mixture while the transparent layer is CNF. PEDOT/PSS carries

the advantage of being extruded in situ, leading to more possibilities of functional fibres.

Figure 17 – Freeze and air dried PEDOT/PSS/CNF filaments

This SEM acquisition shown in figure 17 allows us to analyse the surface aspect of the three-barrel filament. The diameter is similar to the hollow filament’s diameter. The PEDOT/PSS located in the core appears very porous, its spongy structure looks almost like straws that are perfectly aligned along the microfibrils direction. After being air-dried the filament collapses and the diameter shrinks again by 10 times from 1mm to 0,1mm. The evaporation of the solvents finished, -stacking of the PEDOT bonds builds a crystalline network. The PEDOT/PSS is fully encapsulated in the CNF shell and no leaks or large cracks were reported. It suggests that the wet strength and structure of the hollow CNF filament is good. Conductivity tests would be needed to certify that the filaments are indeed conductive, however this SEM characterization allows us to be optimist on the results of such tests.

4.2.3 PEDOT/PSS COMPOSITE FILAMENTS APPLICATION

Obtaining conductive yet biodegradable filaments gives rise to new opportunities like supercapacitor electrodes, conductive films also called “power paper”(figure 18) or simply conductive wires. This application section is mainly relying on two studies respectively investigating power electronic applications[40] and power papers[41]. It is however important to notice that the filaments manufactured in these studies are still different from our work as you can notice on figure 18. Indeed, PEDOT/PSS is covering the CNF located in the core. Such studies report a decrease in the mechanical properties of the composite film with a Young modulus of 660 MPa (50) and a tensile strength of 13,5 MPa (). This decrease is coherent with our expectations as PEDOT:PSS possesses poor mechanical properties. Even if the films are weaker than regular CNF filaments they can still be handle and/or processed as required.

Figure 18 – Power paper produced from PEDOT/PSS/CNF filaments[34]

1 mm

1 mm 500 um

e) f)

The conductivity of the filaments was tested with four-probe electrical measurement. Conductivity appeared to be highly dependent on the degree of substitution (hydroxyl groups on the cellulose surface). Two regimes could be observed. In the low range with few carboxymethyl charges the conductivity attains 400S/cm. When the degree of substitution is superior to 0,4, conductivity decreases drastically to 150S/cm. One could suppose that hydroxyl groups bonded to the PEDOT/PSS are lowering the charge transport as the bond direction is perpendicular to the fibre axis. Such crosslinks could possibly slow down the electron between PEDOT and PSS clusters.

Youngsang K and Jeonghun K [41] revealed that power papers post treated with Di-methylsulfoxyde (DMSO) were obtaining conductivity values more than 100 times higher than the PEDOT/PSS/CNF non-treated power papers.

Few studies are reporting the use of PEDOT/PSS composite CNF filaments in supercapacitors. Yet, this field will necessarily be reviewed in a near future.

4.3 S OCIAL , ENVIRONMENTAL AND ETHICAL ASPECTS

Today’s engineers mission is to ensure a sustainable future while improving the properties of material to satisfy the increasing demand. While this work is exploratory in nature it aims to meet the general increasing demand for sustainable products from companies that are seeking to lower their carbon footprint. Using naturally derived cellulose-based composite materials is one potential solution to tackle current environmental challenges. Cellulose nanofibrils filaments are delivering new opportunities to develop light weight, strong, yet biodegradable structures. On the other hand, PEDOT:PSS/CNF filaments proved to have interesting properties while remaining biodegradable. In addition to being degradable both CNF filaments are synthesized by in situ polymerization. The absence of environmentally-harmful solvents in this process enables a promising and greener route.

While this work is only an overview of what could be produced on a larger scale, it reveals the

potential for a sustainable way to replace regular materials like glass fibre or conductive wires.

5- C ONCLUSION AND NEW PERSPECTIVES 5.1 C ONCLUSION

By carefully following the experimental procedure, the wet spinning technique was successfully employed to extract two and three layer filaments. CNF filaments proved to have very interesting mechanical properties due to their alignment ability conferring strong chains interactions. The surface aspect characterized by SEM and AFM images highlighted the good packing factor of the microfibrils with a defect free surface. Indeed, high quality SEM images were captured to show how the freeze- drying technique was maintaining the fibres in wet-state without altering their surface. This nitrogen- based process gave an opportunity to take never seen before SEM pictures of the filaments. In the future, building more knowledge on the wet structure will allow improvements on the shape and the design of the filaments.

However, the main achievement of this report relies in the production of three layers composite CNF.

The first ever production of hollow CNF filaments was carried out using the very same technique as the two-layer extraction but this time adding a third layer. The extraction of CNF composites with PEDOT/PSS in the core is the first demonstration of in-situ production. A cleaner and greener process than the regular one used until now for CNF applications. In-situ extraction offers a lot of opportunities for new advanced materials such as processing time gain, stronger interactions between materials and environmentally friendly solvents. This report shows already the whole potential of CNF based composite filaments in many fields like medicine, textiles, electronic and packaging industries.

5.2 F UTURE WORK

To follow the thread of this report, one should consider testing the conductivity of the filaments in order to appreciate the application potential in electronic and energy storage application. However even if PEDOT/PSS filaments were the only composite achieved with the three-layered head extruder; many options remain available for the polymer choice. After conductivity, the second property targeted was the stretch ability of the filaments. Like previously stated, CNF filaments have poor deformation properties. Rubbers appear like a good category to be co-extruded with CNF and improve this characteristic. Elastomers like cis-polyisoprene (natural rubber, NR), cis-polybutadiene (butadiene rubber, BR) are potential polymers to add. Using the three-layer nozzle, the cis- polyisoprene can be flown in the second shell while CNF should be flown in the core to provide bending flexibility without compromising the rigidity along the filament axis.

Among the other possible composites that were supposed to be tested, polylactide (PLA) is one of

them. This biodegradable polymer is easily processed and covers a wide range of economically viable

applications. CNF filaments could be used as fibres reinforcements to boost the poor mechanical

properties of PLA. Finally, polyamides fibres could be a suitable solution to enhance the thermal

properties of CNF filaments. Their cationic charge would perfectly interact with the anionic CNF via

Van der Waals interactions and thus contribute to the aggregation and the cohesion of both polymers.

6- A CKNOWLEDGEMENT

I would like to thank my project manager Michael Reid from the Fibre and Polymer Technology department at KTH who helped me relentlessly through the report by answering every single question and providing invaluable knowledge and feedbacks.

I would also like to thank my two supervisors Anders Tilliander (Head of unit of Processes)

and Anders Eliasson (director of studies) from the Material science and Engineering department at

KTH for their help and support in this project.

7- R EFERENCES

[1] “Nanocellulose Market worth $783 million by 2025.”

https://www.marketsandmarkets.com/PressReleases/nanocellulose.asp (accessed Apr. 22, 2020).

[2] H. V. Lee, S. B. A. Hamid, and S. K. Zain, “Conversion of Lignocellulosic Biomass to Nanocellulose: Structure and Chemical Process,” The Scientific World Journal, 2014.

https://www.hindawi.com/journals/tswj/2014/631013/ (accessed Apr. 16, 2020).

[3] R. J. Moon, A. Martini, J. Nairn, J. Simonsen, and J. Youngblood, “Cellulose nanomaterials review: structure, properties and nanocomposites,” p. 54, 2011.

[4] O. Nechyporchuk, N. Belgacem, and J. Bras, “Production of cellulose nanofibrils: A review of recent advances,” Ind. Crops Prod., vol. 93, pp. 2–25, Dec. 2016, doi:

10.1016/j.indcrop.2016.02.016.

[5] “Comprehensive Cellulose Chemistry. Volume 1. Fundamentals and Analytical Methods By D. Klemm, B. Philipp, T. Heinze, U. Heinze, and W. Wagenknecht. Wiley:  Weinheim, Germany.

1998. 260 pp. $236.25. ISBN 3-527-29413-9.,” J. Am. Chem. Soc., vol. 121, no. 37, pp. 8677–8677, Sep. 1999, doi: 10.1021/ja9857514.

[6] I. F. Almeida et al., “Bacterial cellulose membranes as drug delivery systems: An in vivo skin compatibility study,” Eur. J. Pharm. Biopharm., vol. 86, no. 3, pp. 332–336, Apr. 2014, doi:

10.1016/j.ejpb.2013.08.008.

[7] J.-H. Kim et al., “Review of nanocellulose for sustainable future materials,” Int. J. Precis.

Eng. Manuf.-Green Technol., vol. 2, no. 2, pp. 197–213, Apr. 2015, doi: 10.1007/s40684-015-0024- 9.

[8] J. George and S. Sabapathi, “Cellulose nanocrystals: synthesis, functional properties, and applications,” Nanotechnol. Sci. Appl., vol. 8, pp. 45–54, Nov. 2015, doi: 10.2147/NSA.S64386.

[9] “Elastic Modulus of Single Cellulose Microfibrils from Tunicate Measured by Atomic Force Microscopy | Biomacromolecules.” https://pubs.acs.org/doi/10.1021/bm900520n (accessed Apr. 28, 2020).

[10] S. Iwamoto, A. Isogai, and T. Iwata, “Structure and Mechanical Properties of Wet-Spun Fibers Made from Natural Cellulose Nanofibers,” Biomacromolecules, vol. 12, no. 3, pp. 831–836, Mar. 2011, doi: 10.1021/bm101510r.

[11] M. J. Lundahl et al., “Strength and Water Interactions of Cellulose I Filaments Wet-Spun from Cellulose Nanofibril Hydrogels,” Sci. Rep., vol. 6, no. 1, pp. 1–13, Jul. 2016, doi:

10.1038/srep30695.

[12] I. Kalashnikova, H. Bizot, B. Cathala, and I. Capron, “Modulation of cellulose nanocrystals amphiphilic properties to stabilize oil/water interface,” Biomacromolecules, vol. 13, no. 1, pp. 267–

275, Jan. 2012, doi: 10.1021/bm201599j.

[13] K. Markstedt, A. Mantas, I. Tournier, H. Martínez Ávila, D. Hägg, and P. Gatenholm, “3D Bioprinting Human Chondrocytes with Nanocellulose-Alginate Bioink for Cartilage Tissue Engineering Applications,” Biomacromolecules, vol. 16, no. 5, pp. 1489–1496, May 2015, doi:

10.1021/acs.biomac.5b00188.

[14] B. L. Tardy, B. D. Mattos, L. G. Greca, T. Kämäräinen, K. W. Klockars, and O. J. Rojas,

“Tessellation of Chiral-Nematic Cellulose Nanocrystal Films by Microtemplating,” Adv. Funct.

Mater., vol. 29, no. 25, p. 1808518, 2019, doi: 10.1002/adfm.201808518.

[15] M. S. Reid, J. Erlandsson, and L. Wågberg, “Interfacial Polymerization of Cellulose Nanocrystal Polyamide Janus Nanocomposites with Controlled Architectures,” ACS Macro Lett., vol. 8, no. 10, pp. 1334–1340, Oct. 2019, doi: 10.1021/acsmacrolett.9b00692.

[16] H. Fukuzumi, T. Saito, T. Iwata, Y. Kumamoto, and A. Isogai, “Transparent and High Gas Barrier Films of Cellulose Nanofibers Prepared by TEMPO-Mediated Oxidation,”

Biomacromolecules, vol. 10, no. 1, pp. 162–165, Jan. 2009, doi: 10.1021/bm801065u.

[17] J. Vartiainen and T. Malm, “Surface hydrophobization of CNF films by roll-to-roll HMDSO

plasma deposition,” J. Coat. Technol. Res., vol. 13, no. 6, pp. 1145–1149, Nov. 2016, doi:

10.1007/s11998-016-9833-1.

[18] V. Kumar, A. Elfving, H. Koivula, D. Bousfield, and M. Toivakka, “Roll-to-Roll Processed Cellulose Nanofiber Coatings,” Ind. Eng. Chem. Res., vol. 55, no. 12, pp. 3603–3613, Mar. 2016, doi: 10.1021/acs.iecr.6b00417.

[19] “FindLaw’s Supreme Court of Alabama case and opinions.,” Findlaw.

https://caselaw.findlaw.com/al-supreme-court/1023084.html (accessed Apr. 23, 2020).

[20] “Rayon Fibers.” https://polymerdatabase.com/Fibers/Rayon.html (accessed Apr. 23, 2020).

[21] I. R. Comnea-Stancu, K. Wieland, G. Ramer, A. Schwaighofer, and B. Lendl, “On the Identification of Rayon/Viscose as a Major Fraction of Microplastics in the Marine Environment:

Discrimination between Natural and Manmade Cellulosic Fibers Using Fourier Transform Infrared Spectroscopy,” Appl. Spectrosc., vol. 71, no. 5, pp. 939–950, May 2017, doi:

10.1177/0003702816660725.

[22] “Melt Spinning , Dry spinning and Wet Spinning Method | Textile Study Center.”

https://textilestudycenter.com/melt-spinning-dry-spinning-and-wet-spinning-method/ (accessed Apr. 15, 2020).

[23] H. C. Kim, D. Kim, J. Y. Lee, L. Zhai, and J. Kim, “Effect of Wet Spinning and Stretching to Enhance Mechanical Properties of Cellulose Nanofiber Filament,” Int. J. Precis. Eng. Manuf.- Green Technol., vol. 6, no. 3, pp. 567–575, Jul. 2019, doi: 10.1007/s40684-019-00070-z.

[24] M. J. Lundahl, V. Klar, L. Wang, M. Ago, and O. J. Rojas, “Spinning of Cellulose

Nanofibrils into Filaments: A Review,” Ind. Eng. Chem. Res., vol. 56, no. 1, pp. 8–19, Jan. 2017, doi: 10.1021/acs.iecr.6b04010.

[25] “Figure 4. Young’s modulus (a), tensile strength (b) and breaking...,” ResearchGate.

https://www.researchgate.net/figure/Youngs-modulus-a-tensile-strength-b-and-breaking-elongation- c-as-a-function-of_fig3_311362102 (accessed Apr. 16, 2020).

[26] K. M. O. Håkansson et al., “Hydrodynamic alignment and assembly of nanofibrils resulting in strong cellulose filaments,” Nat. Commun., vol. 5, no. 1, pp. 1–10, Jun. 2014, doi:

10.1038/ncomms5018.

[27] S. Yaragalla, R. Mishra, S. Thomas, N. Kalarikkal, and H. Maria, Carbon-Based Nanofillers and Their Rubber Nanocomposites Fundamentals and Applications Edited by Carbon-Based

Nanofillers and Their Rubber Nanocomposites Fundamentals and Applications Edited by. 2019.

[28] N. Mittal et al., “Multiscale Control of Nanocellulose Assembly: Transferring Remarkable Nanoscale Fibril Mechanics to Macroscale Fibers,” ACS Nano, vol. 12, no. 7, pp. 6378–6388, Jul.

2018, doi: 10.1021/acsnano.8b01084.

[29] “Figure 4. Coaxial wet-spinning setup.,” ResearchGate.

https://www.researchgate.net/figure/Coaxial-wet-spinning-setup_fig3_327110292 (accessed Apr.

27, 2020).

[30] “Glass Fibre - an overview | ScienceDirect Topics.”

https://www.sciencedirect.com/topics/engineering/glass-fibre (accessed Apr. 28, 2020).

[31] “Mechanical Behavior of Polymers | MATSE 81: Materials In Today’s World.”

https://www.e-education.psu.edu/matse81/node/2109 (accessed May 04, 2020).

[32] “Young’s Modulus - Tensile and Yield Strength for common Materials.”

https://www.engineeringtoolbox.com/young-modulus-d_417.html (accessed May 20, 2020).

[33] M. Rahaman, D. Khastgir, and A. Aldalbahi, Carbon-Containing Polymer Composites.

2019.

[34] D. Belaineh et al., “Controlling the Organization of PEDOT:PSS on Cellulose Structures,”

ACS Appl. Polym. Mater., vol. 1, no. 9, pp. 2342–2351, Sep. 2019, doi: 10.1021/acsapm.9b00444.

[35] S.-P. Rwei, Y.-H. Lee, J.-W. Shiu, R. Sasikumar, and U.-T. Shyr, “Characterization of Solvent-Treated PEDOT:PSS Thin Films with Enhanced Conductivities,” Polymers, vol. 11, no. 1, Jan. 2019, doi: 10.3390/polym11010134.

[36] D. Hohnholz, H. Okuzaki, and A. G. MacDiarmid, “Plastic Electronic Devices Through

Line Patterning of Conducting Polymers,” Adv. Funct. Mater., vol. 15, no. 1, pp. 51–56, 2005, doi:

10.1002/adfm.200400241.

[37] Z. Zhu, C. Liu, F. Jiang, J. Xu, and E. Liu, “Effective treatment methods on PEDOT:PSS to enhance its thermoelectric performance,” Synth. Met., vol. 225, pp. 31–40, Mar. 2017, doi:

10.1016/j.synthmet.2016.11.011.

[38] E. Montibon, L. Järnström, and M. Lestelius, “Characterization of poly(3,4-

ethylenedioxythiophene)/poly(styrene sulfonate) (PEDOT:PSS) adsorption on cellulosic materials,”

Cellulose, vol. 16, no. 5, pp. 807–815, Oct. 2009, doi: 10.1007/s10570-009-9303-3.

[39] A. Y. Mehandzhiyski and I. Zozoulenko, “Computational Microscopy of

PEDOT:PSS/Cellulose Composite Paper,” ACS Appl. Energy Mater., vol. 2, no. 5, pp. 3568–3577, May 2019, doi: 10.1021/acsaem.9b00307.

[40] A. Malti et al., “An Organic Mixed Ion–Electron Conductor for Power Electronics,” Adv.

Sci., vol. 3, no. 2, p. 1500305, 2016, doi: 10.1002/advs.201500305.

[41] “(PDF) Fabrication of Highly Conductive Porous Cellulose/PEDOT:PSS Nanocomposite Paper via Post-Treatment,” ResearchGate.

https://www.researchgate.net/publication/332433537_Fabrication_of_Highly_Conductive_Porous_

CellulosePEDOTPSS_Nanocomposite_Paper_via_Post-Treatment (accessed May 18, 2020).

TRITA ITM-EX 2020:319

www.kth.se