BIOCOMPOSITES BASED ON

CORE-SHELL CELLULOSE

NANOFIBERS

Preparation, Structure and Properties

Kasinee Prakobna

Doctoral Thesis

Wallenberg Wood Science Center

Department of Fiber and Polymer Technology

School of Chemical Science and Engineering

KTH Royal Institute of Technology

Stockholm, 2015

Supervisor

Professor Lars A. Berglund

Copyright © 2015 Kasinee Prakobna

All Rights Reserved.

ISBN 978-91-7595-513-1

ISSN 1654-1081

TRITA-CHE Report 2015:18

Tryck: US-AB, Stockholm 2015

AKADEMISK AVHANDLING

Som med tillstånd av Kungliga Tekniska Högskolan i Stockholm framlägges till

offentlig granskning för avläggande av teknologie doktorsexamen den 26 maj

2015, kl. 10.00 i sal F3, Lindstedtsvägen 26, KTH, Stockholm. Fakultetsopponent:

Dr. David Plackett, The University of British Columbia, Canada. Avhandlingen

försvaras på engelska.

“Strive not to be a success, but rather to be of value.”

Albert Einstein

ABSTRACT

Cellulose nanofibers (CNFs) are of interest as load-bearing components for

polymer matrix nanocomposites. A wide range of nanostructured materials

including nanopaper/films, foams, aerogels, and hydrogels can be prepared from

CNFs. The material performance can be fine-tuned when CNFs are combined

with different polymer matrices. The main idea of the present study is to test the

hypothesis that the concept of Core-shell CNFs can provide processing and

performance advantages for nanocomposites through improved nanostructural

control.

The problems of matrix distribution and interface structure at nanoscale are

investigated. The first part of this thesis (Paper I-III) describes an alternative

preparation procedure for biocomposites based on the Core-shell concept.

Inspired by the structural framework and mechanical function of the primary cell

wall in plants, Core-shell CNFs are formed by coating wood CNFs with a

polysaccharide matrix, and are subsequently used for fabrication of biocomposite

films. The nanostructure of Core-shell CNFs and their nanocomposites is

characterized. Mechanical properties of the biocomposites at various hydration

conditions are investigated. A study on molecular water mobility and moisture

stability of the materials is conducted. In the present thesis, three different

biological polysaccharides including amylopectin, xyloglucan and

galactoglucomannan have been used for man-made nanocomposites based on

Core-shell nanofibers.

The later sections of the thesis (Paper IV-V) describe an alternative method to

disintegrate holocellulose nanofibers from wood chips. These CNFs based on

holocellulose possess Core-shell structure with native hemicelluloses as “shell”.

Peracetic acid pretreatment is used for the preparation of holocellulose. This

procedure is a promising single-step pulping as preparation for nanofibrillation.

The nanostructure of the holocellulose nanofibers are characterized, and

compared with CNFs prepared from enzymatic pretreatment. The holocellulose

nanofibers are used for preparation of films and porous materials. The influence

of nanostructural characteristics of two different nanofibers (CNFs from

enzymatic pretreatment vs holocellulose nanofibers from peracetic acid

pretreatment) on final properties of the nanocomposites is clarified.

Favorable characteristics of Core-shell fibrils are reported in terms of colloidal

stability, controlled matrix distribution, improved interface characteristics, and

improved hygromechanical properties of the biocomposites. In both a scientific

and industrial context, the Core-shell nanofiber concept thus offers great potential

for the materials design of new cellulose nanomaterials with unique

characteristics.

Keywords: biocomposites, cellulose nanofibers, polysaccharide, holocellulose,

nanostructure, dispersion, mechanical properties, hygromechanical properties,

moisture stability, film, porous material

SAMMANFATTNING

Nanofibrer från cellulosa (CNF) är av stort intresse som lastbärande

komponent i nanokompositer och ett flertal nanostrukturerade material såsom

nanopapper, nanoskum, aerogeler eller hydrogeler kan framställas från CNF.

Materialegenskaperna kan finjusteras då CNF tillsätts till en polymermatris.

Avhandlingen handlar om ett koncept för att kontrollera matrisfördelning och

gränsyta mellan CNF och polymermatrisen, i form av sk ”Core-shell”-CNF. Första

delen (artikel I-III) handlar om ”Core-shell”-CNF framställda genom att adsorbera

en polysackarid på ytan av CNF. Nanostrukturen hos ”Core-shell”-CNF och dess

kompositmaterial undersöks, liksom mekaniska egenskaper vid olika

hydreringstillstånd. Molekylrörlighet hos vatten studeras med NMR. Tre olika

polysackarider undersöks; amylopektin, xyloglukan och galaktoglukomannan.

Avhandlingens andra del (artikel IV och V) handlar om CNF från holocellulosa,

med vedflis som råvara. Holocellulosa innebär att huvudsakligen lignin har

extraherats från veden, medan cellulosa och hemicellulosa är välbevarade. Denna

CNF har också ”Core-shell”-struktur med hemicellulosa som ytskikt. Perättiksyra

används som en mild framställningsmetod. Det innebär en process i ett enda steg

för att framställa massa med bevarad cellulosastruktur. Nanostruktur och kemisk

sammansättning karakteriseras och jämförs med enzymatiskt framställd CNF

från sulfitmassa. CNF från båda typerna av nanocellulosa (holocellulosa och

enzymatisk CNF) används för att framställa kompositer och porösa material

(honeycombs och skum). Sambanden mellan struktur och materialegenskaper

klarläggs, liksom effekterna från olika typer av CNF. Avhandlingen visar att

”Core-shell”-CNF är ett fungerande koncept för nanostrukturerade kompositer

genom påvisande av ett flertal lovande och fördelaktiga egenskaper såsom

kolloidal stabilitet, kontrollerad matrisfördelning och förbättrade gränsyte- samt

hygromekaniska egenskaper. Såväl vetenskapligt som industriellt kan ”Core-

shell”-CNF bidra till utvecklingen av nya nanomaterial baserade på cellulosa.

LIST OF PAPERS

This thesis is based on the following five appended papers:

High-Performance and Moisture-Stable Cellulose-Starch Nanocomposites I.

Based on Bioinspired Core-Shell Nanofibers

Kasinee Prakobna, Sylvain Galland, and Lars A. Berglund (2015),

Biomacromolecules, 16, pp 904-912.

Core-Shell Cellulose Nanofibers for Biocomposites – Nanostructural II.

Effects in Hydrated State

Kasinee Prakobna, Camilla Terenzi, Qi Zhou, István Furó, and Lars A.

Berglund (2015), Carbohydrate Polymers, 125, pp 92-102.

Strong Effects from Galactoglucomannan Hemicelluloses on Mechanical III.

Behavior of Wet Cellulose Nanofiber Gels

Kasinee Prakobna, Victor Kisonen, Chunlin Xu and Lars A. Berglund,

Submitted manuscript.

Holocellulose Nanofibers of High Molar Mass and Small Diameter for IV.

High-Strength Nanopaper

Sylvain Galland, Fredrik Berthold, Kasinee Prakobna, and Lars A. Berglund,

Submitted manuscript.

Mechanical Performance and Architecture of Biocomposite Honeycombs V.

and Foams from Core-Shell Holocellulose Nanofibers

Kasinee Prakobna, Fredrik Berthold, and Lars A. Berglund, Manuscript.

The author’s contributions to the appended papers are as follows:

I. All of the experimental work and data analysis except TEM observation,

and all of the manuscript preparation.

II. All of the experimental work and data analysis except 2H NMR relaxometry,

and most of the manuscript preparation.

III. All of the experimental work and data analysis except extraction and molar

mass analysis of GGM, and all of the manuscript preparation.

IV. Part of the experimental work and data analysis include preparation of

holocellulose nanofibers (peracetic acid pretreatment), observation on

nanostructure and light transmittance of the materials, and most of the

manuscript preparation.

V. All of the experimental work and data analysis, and all of the manuscript

preparation.

Other relevant papers not included in this thesis:

Composite Films of Nanofibrillated Cellulose and O-Acetyl-VI.

Galactoglucomannan (GGM) Coated with Succinic Esters of GGM

Showing Potential as Barrier Material in Food Packaging

Victor Kisonen, Kasinee Prakobna, Chunlin Xu, Arto Salminen, Kirsi S.

Mikkonen, Dimitar Valtakari, Patrik Eklund, Jukka Seppälä, Maija

Tenkanen, and Stefan Willför (2015), Journal of Materials Science, 50, pp 3189-

3199.

Nanostructural Effects on Polymer and Water Dynamics in Cellulose VII.

Biocomposites – 2H and 13C NMR Relaxometry Camilla Terenzi, Kasinee Prakobna, Lars A. Berglund, and István Furó

(2015), Biomacromolucules, DOI: 10.1021/acs.biomac.5b00330.

LIST OF ABBREVIATIONS

AFM Atomic force microscopy

AP Amylopectin

CNF(s) Cellulose nanofiber(s)

DP Degree of polymerization

DLS Dynamic light scattering

DMAc Dimethyl acetamide

DSP Digital speckle photography

DVS Dynamic vapor sorption

E Young’s modulus

Es Young’s modulus of the solid cell wall material

EIC Ethyl isocyanate

FE-SEM Field-emission scanning electron microscopy

GGM O-acetyl-galactoglucomannan

HPAEC High performance anion exchange chromatography

MALLS Multi angle laser light scattering

Mn Number average molar mass

Mw Weight average molar mass

NMR Nuclear magnetic resonance

PAD Pulsed amperometric detector

PDI Polydispersity index

PTFE Polytetrafluoroethylene

SEC Size exclusion chromatography

SC-CO2 Supercritical carbondioxide

TEM Transmission electron microscopy

TEMPO Tetramethylpiperidine-1-oxyl

TOCNs TEMPO-oxidized cellulose nanofibers

XG Xyloglucan

CONTENTS

INTRODUCTION .................................................................................................. 1 1

1.1 Objective .......................................................................................................... 2

BACKGROUND .................................................................................................... 3 2

2.1 Wood cell walls .............................................................................................. 3

2.1.1 Structural biopolymers .................................................................... 3

2.1.2 Ultrastructure of the cell walls ........................................................ 5

2.2 Cellulose nanofibers from wood .................................................................. 7

2.2.1 Pretreatment approaches ................................................................. 8

2.2.2 Reinforcement materials .................................................................. 9

2.3 Biocomposite films based on cellulose nanofibers .................................. 10

2.4 Core-shell cellulose nanofibers and their biocomposite films ............... 13

2.5 Holocellulose nanofibers and their biocomposite films ......................... 15

2.6 Application of nanofibers in porous materials ........................................ 16

EXPERIMENTAL................................................................................................. 19 3

3.1 Material components ................................................................................... 19

3.1.1 Cellulose nanofibers (CNFs) prepared by enzymatic pretreatment

(Paper I-V) ......................................................................................19

3.1.2 Holocellulose nanofibers prepared by peracetic acid pretreatment

(Paper IV-V) ...................................................................................19

3.1.3 Polysaccharide matrices (Paper I-III) .............................................20

3.2 Preparation of biocomposite materials ..................................................... 21

3.2.1 Cellulose and holocellulose nanopaper (Paper I-IV) ....................... 21

3.2.2 Neat polysaccharides (AP and XG) films (Paper I-II) ................... 21

3.2.3 Adsorption experiment and preparation of Core-shell nanofibers

(Paper I-III) .................................................................................... 21

3.2.4 Nanocomposite films (Paper I-III).................................................. 22

3.2.5 Porous materials based on nanofibers (Paper V) ............................ 23

3.3 Characterization methods ........................................................................... 25

3.3.1 Determination of matrix contents by sugar analysis (Paper I-III). 25

3.3.2 Colloidal behavior by dynamic light scattering (DLS) (Paper I-III)25

3.3.3 Total charge determination (Paper IV) .......................................... 25

3.3.4 Nanostructure of nanofibers and their networks (Paper I-V) ........ 26

3.3.5 Density and porosity (Paper I-V) ................................................... 26

3.3.6 Tensile mechanical properties (Paper I-IV) .................................... 27

3.3.7 Mechanical properties in compression (Paper V) ........................... 27

3.3.8 Dynamic vapor sorption (DVS) (Paper I-III)................................. 27

3.3.9 Molecular moisture dynamic by 2H solid-state NMR (Paper II) ... 28

3.3.10 Molar mass analysis of nanofibers by SEC (Paper IV) .................. 28

3.3.11 Molar mass analysis of matrices by SEC-MALLS (Paper I-II) ...... 29

3.3.12 Light transmittance (Paper IV) ...................................................... 29

RESULTS AND DISCUSSION ........................................................................... 31 4

4.1 Core-shell nanofibers and their biocomposite films (Paper I-III) ........... 31

4.1.1 Preparation of Core-shell nanofibers with polysaccharide matrices31

4.1.2 Nanostructure of the nanofibers and their networks ..................... 33

4.1.3 Hygromechanical properties of Core-shell vs “Mixed”

nanocomposites ...............................................................................41

4.1.4 Moisture sensitivity of Core-shell vs “Mixed” nanocomposites ....49

4.2 Conclusions (from Paper I-III) .................................................................... 54

4.3 Holocellulose nanofibers and their biocomposite films (Paper IV-V) .. 56

4.3.1 Holocellulose nanofibers .................................................................56

4.3.2 Films from holocellulose nanofibers ................................................60

4.3.3 Porous materials from holocellulose nanofibers ..............................62

4.4 Conclusions (from Paper IV-V) .................................................................. 68

FUTURE WORK .................................................................................................. 71 5

ACKNOWLEDGEMENTS ................................................................................. 73 6

REFERENCES ...................................................................................................... 75 7

INTRODUCTION 1

INTRODUCTION 1

Wood fibers, which are a widely available bioresource, have been exploited

for sustainable resource products such as paper packaging, textiles and wood

composites. To extend the traditional use of the wood fibers towards high-

performance materials, it is essential to utilize the full potential of their nanoscale

building blocks. Interest in cellulose nanofibers (CNFs) has increased as is

apparent both in academic and industrial research.1, 2 The growing interest in

CNFs is primarily due to their small dimensions, high aspect ratio and ability to

form entangled network in addition to outstanding mechanical properties. The

potential of wood CNFs has therefore been considered for advanced load-bearing

applications. Various approaches to disintegrate CNFs from wood fibers have

been explored.3-7 However, special attention to CNF characteristics and their

reinforcement potential is required. Inherent properties and the nanostructure of

disintegrated CNFs will decisively influence the final properties of

nanocomposites.

Nanocomposites of CNF reinforced natural biopolymers such as starch and

hemicelluloses have been progressively developed,8-13 primarily to overcome

poor mechanical behavior of the neat biopolymers (brittleness), low moisture

stability and inferior film-forming properties. Potential application of these

polysaccharide-based nanocomposites includes packaging materials.8, 11, 14-17

Favorable interfacial interactions between CNFs and the polysaccharide matrix

could be expected due to chemical similarities of the polymers. Since CNFs can

potentially offer large specific interface area in the nanocomposites, the

molecular and nanoscale control of interfacial phenomena between components

is necessary. In biological materials, plant cell walls can be viewed as models for

polysaccharide nanocomposites. The combination of mechanical stiffness,

strength, toughness and interfacial stability despite a highly hydrated state is

achieved within their nanostructural network. As compared with the natural

plant cell walls, it is important to note that typical man-made nanocomposites

possess poor nanostructural control of matrix distribution. In the present thesis,

an attempt has been made to coat wood CNFs with the polysaccharide matrix,

forming “Core-shell” nanofibers. Therefore, better control of the matrix

2 INTRODUCTION

distribution at nanoscale, as in the plant cell walls, is expected in the resulting

nanocomposites. The reinforcement effect and moisture stability of the Core-shell

nanocomposites have been compared with “Mixed” nanocomposites, prepared by

conventional mixing, and with a polymer matrix less favorably distributed.

1.1 Objective

The scientific objective of the present thesis is to develop preparation methods

for Core-shell nanofibers and characterize their structure. In addition, the objective

is to investigate processing-structure-property relationships for biocomposite

materials based on the Core-shell CNFs, with emphasis on nanostructural aspects.

In the first section (Paper I-III), the main purpose is to address the problem of

how nanostructural features of polymer matrix distribution and

interface/interphase is influencing hygromechanical properties and moisture

stability for the Core-shell nanocomposites.

The major focus in the second section (Paper IV-V) is related to wood

holocellulose nanofibers, which possess native Core-shell structure. Peracetic acid

pretreatment as well as key characteristics of the obtained holocellulose

nanofibers is of particular interest. In addition, the potential of how these

holocellulose nanofibers can be utilized for high-performance film and porous

materials is investigated.

BACKGROUND 3

BACKGROUND 2

2.1 Wood cell walls

Wood is of interest for many branches of industrial applications. The

molecular interactions among wood components and their arrangement on the

nanolevel have important implications and contribute to the remarkable

mechanical properties of wood.18, 19 The wood cell walls can be considered as a

naturally occurring biocomposite materials, possessing exceptional mechanical

properties of high modulus and strength combined with toughness and

flexibility. It can be transformed into nanofibers and that serve as building blocks

for nanomaterials. The architectural details of their structural biopolymer

organization and the nanostructural design resulting from biosynthesis are truly

remarkable and can inspire the development of man-made nanocomposites.

2.1.1 Structural biopolymers

Polysaccharides are the major biopolymeric components in the wood cell

walls. Cellulose, which is the primary load-bearing biopolymer in plants, is a

major structural component of the cell wall. Cellulose is a linear chain of many

thousand glucose units. They are covalently linked by β-1,4-glycosidic bonds,

and their repeat unit is cellobiose consisting of two glucose rings (Figure 1(a)).

The degree of polymerization (DP) for cellulose from wood can be as high as

10,000.20 The cellulose chains are arranged in parallel with an extended-chain

conformation, and are co-crystallized, forming microfibrils.21 The diameter of an

individual microfibril is about 3-4 nm, and the fibrils form microfibril aggregates

with diameters of approximately 20 nm.22, 23 Extensive hydrogen bonding, which

is formed within and between the cellulose chains of microfibrils, contributes to

their unique properties of mechanical strength and chemical stability. Specific

tensile strength of the microfibrils surpasses that of many natural and man-made

polymeric fibers.24 Thus, the microfibrils, which can be extracted from the cell

wall in the form of CNFs, offer the possibility to improve the mechanical

properties of a wide range of polymeric materials.

4 BACKGROUND

Figure 1. Chemical structure of (a) cellulose, (b) xyloglucan (XG) and (c) O-acetyl-

galactoglucomannan (GGM). They are the principal structural polysaccharides in

spruce cell walls.

The second most abundant group of polysaccharides in wood cell walls is

hemicelluloses. The content of hemicelluloses in wood varies in the range of 25-

40 wt%.25 In contrast to cellulose, hemicelluloses are hetero-polysaccharides,

consisting of linear homo- or copolymer backbone with varying degrees of

branching.26 They can consist of different monosaccharides such as glucose,

mannose, xylose, galactose and arabinose. The degree of polymerization (DP) for

hemicelluloses in woods probably varies between 50-200,20 which is much lower

than for cellulose. This is consistent with their function to physically link

cellulose microfibrils in hydrated state.

Type and content of hemicelluloses vary depending on the plant. For

softwoods such as spruce, xyloglucan (XG) and galactoglucommannan (GGM)

are the major hemicelluloses in primary and secondary cell walls, respectively.

XG has a backbone resembling cellulose with the exception that it is regularly

substituted with α-D-xylopyranose branches at the O-6 position. A minority of

this branched moiety is further substituted with β-D-galactopyranose.27 An

BACKGROUND 5

example of chemical structure for XG is presented in Figure 1(b). The ratio of

these monosaccharide constituents and their pattern of substitution varies

depending on plant origin of XG.28 Apart from XG in the primary cell wall,

storage XG (for instance present in tamarind seeds) is available in larger quantity

and shows potential for industrial applications. XG from tamarind seed can be

employed as an alternative high performance biopolymer.29

In the secondary cell wall, galactoglucomannan (GGM) is a typical

hemicellulose for softwoods. In the case of spruce, the GGM content ranges

between 20-25 wt%.20, 25 The main chain of GGM is built from randomly

distributed β-1,4-D-mannopyranosyl and β-1,4-D-glucopyranosyl units. The

hydroxyl group at C-2 or C-3 position of the mannose unit is partly substituted

by acetyl groups. Some of the mannose units are further substituted by α-1,6-D-

galactopyranosyl units (Figure 1(c)). The ratio of theses monosaccharide

constituents (mannose:glucose:galactose) varies in the range of 3.5-4.5:1:0.5-1 for

spruce GGM.20, 30, 31 Large fractions of water-soluble GGM can be extracted from

spruce wood.32, 33 There is potential to use GGM in packaging films due to its low

oxygen34 and water permeability.35

2.1.2 Ultrastructure of the cell walls

Wood cell walls have a hierarchical structural network consisting of several

layers and different constituents. The exceptional mechanical properties of the

cell walls are governed by the structural arrangement at multiple scales.18 The

ultrastructure on a nanoscopic to microscopic scale is suggested in Figure 2. The

cell walls provide a supportive framework assembled from reinforcing semi-

crystalline cellulose microfibrils. They are embedded within an amorphous and

hydrated matrix of hemicelluloses and lignin. These polymers are intimately

mixed on a molecular and nanostructural scale. The principal framework is

analogous to man-made composites such as glass fibers in a plastic matrix,36

although the molecular and nanostructural design features are much more

sophisticated in the biological composites.

6 BACKGROUND

Figure 2. Schematic illustration of the hierarchical structure in wood cell walls and in

the primary plant cell walls (adapted from Carpita and McCann 2000, and Gibson

2012).18, 36

The chemical composition and nanostructure of the cell walls vary in different

wood species and cell wall types. Primary cell walls consist of a thin layer of

randomly arranged microfibrils in a network, which provides greater ductility to

the cell walls. The structural network of primary cell walls may contain

approximately equal contents of cellulose and xyloglucan (XG), embedded in an

independent network of pectin.36 XG coats and provides physical cross-links to

cellulose microfibrils. There are three distinct domains of XG in the cell walls.37 It

corresponds to the domain where XG forms the cross-links between cellulose

microfibrils. In addition, XG is intimately associated on the surface, and may

even be entrapped within the microfibrils. The arrangement of microfibrils and

XG in the primary cell walls is depicted in Figure 2. The supramolecular

organization of XG plays an important role in regulating mechanical properties

of the primary cell walls.38 XG shows strong molecular interaction and adhesion

to cellulose even at hydrated conditions.39 The cell walls can become extensible

when XG is incorporated into the network.38, 40

Secondary cell walls are rigid and composed of three different layers (S1, S2,

S3). The mechanical properties of wood cell walls are largely governed by the

cellulose-rich middle secondary layer (S2), where the microfibrils are highly

oriented. Considering polymeric matrix in the secondary cell walls, a network of

glucomannan-lignin-xylan is formed in softwoods.41, 42 Hemicelluloses are,

however, closely associated with cellulose through hydrogen bonding.43 They

influence flexibility and rigidity of the structural network.44 Hemicelluloses may

BACKGROUND 7

also serve as a nanostructural interphase between cellulose and lignin. A

variation in chemical structure and arrangement of hemicelluloses is important

for tailoring physical properties of the cell walls.

The favorable interactions between cellulose and hemicelluloses at the

ultrastructural scale can also be considered in the context of cell wall assembly

and design. Cellulose chains are synthesized within rosettes at the plasma

membrane surface.45 The rosettes act as nanospinnerets extruding the cellulose

chains to form a self-assembling microfibril. Hemicelluloses are synthesized in

the Golgi apparatus, and diffuse towards the plasma membrane. After secretion,

the cellulose and hemicelluloses are associated into an integrated network of cell

walls.46 The coalescence and further aggregation of microfibrils is prevented by

the presence of hemicellulose as a coating layer,46 which also serves as a hydrated

lubricating boundary layer at the nanoscale.

The assembly of cell wall components takes place in an aqueous environment.

It is considered that the primary cell walls contain up to 75-80 wt% of water,

which is primarily located in the matrix region.46 This is advantageous from the

viewpoint of structural features. The hemicelluloses have enough mobility to

arrange for proper conformation.36 It is important for regulating their role in the

cell walls and also for the physical properties such as cell wall extension.47

2.2 Cellulose nanofibers from wood

Cellulose nanofibers (CNFs) are present in the structural network of wood

cell walls in the form of microfibrils. CNFs, which are classified within a family

of nanocellulose, are of interest as a strong reinforcement in nanocomposites. The

key characteristics of high aspect ratio,3 outstanding mechanical properties,48, 49

good network formation,4 high surface area50 and versatile chemical

functionality51 encourage the utilization of CNFs for novel applications. CNFs are

commonly disintegrated from wood fibers. The production of CNFs primarily

consists of a pretreatment step followed by mechanical disintegration to liberate

the microfibrils. Typical morphology of wood fibers and CNFs are shown in

Figure 3.

8 BACKGROUND

Figure 3. Light microscopy image of spruce wood fibers and atomic force microscopy

(AFM) image of the obtained CNFs via enzymatic approach.

2.2.1 Pretreatment approaches

Various pretreatment approaches have been developed for CNF production.

The general aim of the pretreatment is to improve nanofibrillation level and

reduce energy consumption during the subsequent mechanical disintegration.

Turbak et al.52 originally adopted a direct mechanical disintegration approach

without any pretreatment procedure. CNFs were liberated from a dilute

suspension of wood fibers. Non-uniform nanofibers having width in the range of

25 to 100 nm were obtained. To achieve sufficient degree of nanofibrillation, this

approach required a high number of passes through the mechanical

homogenizer. Consequently, the resulting nanofibers were shortened and

damaged, and high energy was consumed during mechanical disintegration.

Enzymatic pretreatment in combination with mechanical disintegration was

found to facilitate nanofibrillation by Henriksson et al.3 and Pääkkö et al.4 The

obtained CNFs have high average molar mass (DP ~ 1,000) and high aspect ratio

(typical diameter between 5-20 nm, and lengths of several micrometers). The

nanofibers are long, flexible and entangled, leading to formation of strong

networks and gels.4

Chemical pretreatment in combination with mechanical disintegration is

another route towards uniform CNFs. TEMPO-mediated oxidation was carried

out on the cellulose fiber surface, where primary hydroxyl groups at C-6 position

were converted to aldehyde and carboxylic groups.5, 53 After the chemical

BACKGROUND 9

pretreatment, mild mechanical treatment was adequate for nanofibrillation.

TEMPO-oxidized CNFs (TOCNs), bearing negative charge due to the introduced

carboxylic group, are individualized and have a uniform diameter of

approximately 3-4 nm. Depending on the oxidization level, nanofibers with low

molar mass (DP of 250-300) might be obtained. The length of TOCNs is about 0.5

µm.54 For TOCNs with high carboxylate content (~1.5 mmol/g), their suspension

is highly transparent.15 Similar concepts of surface modification were used to

facilitate nanofibrillation of wood fibers. For instance, positively charged CNFs

were prepared by surface quaternization.7 Another method of surface

modification using carboxymethylation was explored by Wågberg et al.6

Pulp fibers obtained after delignification with a considerable amount of native

hemicelluloses preserved are termed “holocellulose fibers”. To obtain nanofibers,

the holocellulose fibers could be prepared from chlorite delignification and

subjected directly to mechanical grinding.55 The high hemicellulose content is

positive for the ease of nanofibrillation. The holocellulose nanofibers contain up

to 20 wt% of hemicelluloses. The typical diameters of holocellulose nanofibers are

approximately 10-20 nm, which is related to the original size of fibril aggregates.

However, the wood cell walls contain individualized nanofibers (microfibrils)

with a diameter of approximately 3 nm.56 Therefore, it may be possible to convert

the holocellulose fibers into thinner nanofibers, as is the intention of the present

study.

2.2.2 Reinforcement materials

CNFs have received significant interest as high-performance reinforcement

materials. Their remarkable mechanical properties provide potential for load-

bearing applications. CNFs are also of obvious interest in the context of

sustainable development. The mechanical properties of CNFs are substantially

improved compared with wood fibers, having 10 microns scale diameter. The

strong effect of reinforcement for composites based on CNFs was first revealed

by Nakagaito and Yano.57 The strength of the CNF-based composite was

comparable to some metal alloys.

From the viewpoint of engineering design, a low-density material with

favorable mechanical properties is attractive. CNFs are likely to have advantages

10 BACKGROUND

over man-made fibers due to their intrinsic properties of low density, high

specific modulus and high specific strength, as presented in Table 1. It should be

noted that CNFs are assembled from semi-crystalline extended chains, and thus

the reinforcing effect of CNFs is expected to be somewhat lower than that of

cellulose crystals. The reinforcing capability of CNFs is also controlled by several

parameters such as their aspect ratio, degree of dispersion, orientation

distribution, the extent of mechanical damage and surface chemistry attributes.

The resulting nanostructural characteristics and mechanical properties of the

CNFs are of particular importance and efficient preparation routes for CNF

production are obviously also desirable.

Table 1 Mechanical properties of engineering fibers and cellulose-based

reinforcement materials.

Density

(g/cm3)

Young’s

Modulus

(GPa)

Tensile

Strength

(GPa)

Specific

Modulus

(GPa/gcm-3)

Specific

Strength

(GPa/gcm-3)

Glass fiber1 2.5 70 2.4 28 0.95

Carbon fiber58 1.8 150-500 1.5-5.5 80-280 0.83-3.1

Wood fiber58 1.5 14-27 0.3-1.4 9-18 0.20-0.93

Cellulose

nanofiber59-61 1.5 > 55 1.6-3.0 > 35 1.0-2.0

Crystalline

cellulose58, 62 1.6 130-150 7.5-7.7 80-95 4.7-4.8

2.3 Biocomposite films based on cellulose nanofibers

Long nanofibers and their entangled network allow the formation of flexible

CNF-based films. A rapid and convenient preparation for large CNF-based films

with high mechanical properties was reported by Sehaqui et al.63 The procedure

mainly involves water-based filtration and constrained drying. Upon drying,

strong interfibril interaction is developed.64 The obtained CNF film, referred to as

CNF nanopaper, shows high modulus (13.4 GPa) and strength (232 MPa)63 due to

good random-in-plane orientation of the nanofibers.48 The orientation can be

further improved by partial alignment of the nanofibers in the wet state by cold-

BACKGROUND 11

drawing.49 CNF-based films with improved modulus (> 30 GPa) and tensile

strength (575 MPa) were reported61 when the surface of nanofibers was grafted

with soft polymer to facilitate the alignment. The nanofibers have ability to form

a dense network, which is promising for barrier properties in packaging

application.14, 65 Owing to the unique characteristics of individualized CNFs in a

densely packed network, CNF-based films with optical transparency could also

be prepared,66 which is useful for electronic applications such as display screens,

etc.

In practice, CNFs, especially those with low surface charge, have a tendency

for self-aggregation due to strongly interacting surfaces.67 The state of CNFs

dispersion plays an important role in controlling the final structure of the

nanofibrillar network (nanopaper). Aqueous suspensions of CNFs can be

stabilized by electrostatic repulsion and steric hindrance.68 Good dispersion of

CNFs during processing is essential to achieve desirable mechanical

reinforcement.69

Polysaccharide biopolymers such as hemicelluloses and starch have been

widely investigated as a candidate component for biomass-based materials. To

transform these polysaccharides into green materials for wide-spread industrial

applications, several limitations on their inherent properties such as

hygromechanical properties and poor film formation need to be solved. Their

mechanical performance is very sensitive to moisture content, especially when

plasticizers are introduced. Substantial reduction in tensile modulus and strength

under high humidity conditions is a common problem.29, 70 The attractive

characteristics of CNFs are particularly interesting and may serve to overcome

these challenges. Strong molecular interactions between CNFs and other

polysaccharides at the interface as well as the semi-crystalline nature of CNFs are

advantages. For example, the addition of CNFs leads to strong reinforcement and

reduced water sensitivity for starch films.8, 9, 12, 71 There have been a considerable

number of studies devoted to hemicelluloses such as galactoglucomannan and

xylan,10, 11, 72, 73 and similar positive effects from CNF addition have been observed.

The neat films obtained from this class of polysaccharides are generally brittle

due to their insufficient chain length and high glass transition temperature.74 The

12 BACKGROUND

film forming capability of hemicelluloses and starch can be improved by the

flexible and entangled network of CNFs.13, 75

CNFs and other polysaccharides are usually compatible. In the context of

processing, they can be mixed in aqueous suspension, and easily shaped into the

final materials. In most studies, the effect of CNFs on mechanical properties of

the biocomposite films is studied at low CNFs content. By using filtration, well-

dispersed biocomposites with high CNFs content can successfully be prepared76

and strong reinforcement effects can be achieved.77

It has also been reported that various sources of CNFs such as tunicates and

bacteria can be used, although the reinforcement effect varies due to structural

organization and mixing effects.78, 79 The network formation is governed by the

geometry and stiffness of the nanofibers. For wood CNFs and polysaccharide

matrices, strong reinforcement is likely due to favorable interface bonding and

nanofiber entanglements.78 However, the reinforcing and film-forming potential

from CNFs could also vary due to the difference in nanostructural morphology,9

which is related to the extent of nanofibrillation, preparation and degree of

dispersion.

Biocomposite films based on CNFs and CNF/Amylopectin (AP) are shown in

Figure 4. CNFs are obtained as a gel-like suspension (~2 wt%) after enzymatic

pretreatment and disintegration. A translucent film of CNFs and CNF/AP can be

prepared by filtration of a hydrocolloidal suspension followed by drying. The

nanofibrillar network forms a lamellar structure at microscale. The mechanical

properties of CNFs are strongly influenced by the surrounding humidity due to

adsorbed moisture. The mechanical properties (modulus, strength, and ductility)

of CNF/AP at elevated humidity are improved compared with neat AP, but it is

still far below neat CNF. In the present study, we will investigate if the

mechanical properties of CNF/AP films can be improved by using a Core-shell

nanofiber approach.

BACKGROUND 13

Figure 4. Typical characteristics of biocomposite film based on enzyme-pretreated

CNFs, and the hygromechanical properties of CNF/Amylopectin biocomposite film.

2.4 Core-shell cellulose nanofibers and their biocomposite films

Fiber/matrix interfacial interaction is important for the reinforcement of the

composite materials.80 The localized region where the molecular conformation

and mobility of the matrix is influenced by the rigid surface of the fiber is defined

as an interphase.80 The molecular nature of the nanoscale interphase can affect the

final properties of the nanocomposites.81 Thus, it becomes important to have

better control of materials at the molecular and nanostructural scale.

The design of the cell walls has been studied at molecular level.36, 46 The simple

model of their organization for cellulose microfibrils (reinforcing element) and

hemicelluloses (polysaccharide matrix) is proposed by Fengel22 (Figure 5(a)). In

the primary cell walls, the surface of cellulose microfibrils is coated by XG.46

More specifically, the basic structural unit is assembled from the microfibrils

where the surface is coated by a monolayer of XG.82 An interesting consequence

is related to the role of interface hemicelluloses, which serve as a glue between

the microfibrils.83 The strong binding of hemicelluloses to cellulose is beneficial to

the molecular interaction and bond recovery during mechanical deformation of

the cell walls.84

14 BACKGROUND

Assemblies in the form of biomimetic composites, where CNFs and ductile

polymers serve as load-bearing and energy-dissipating components, have been

investigated.76, 85-87 The design principle for these materials is based on a

composite with a high content of CNFs, and the CNF surface is modified through

polymer adsorption. A study of bacterial cellulose synthesized in the presence of

hydroxyethyl cellulose (HEC), which is a hydrated amorphous polysaccharide,

were conducted.85 Biocomposites of cellulose fibrils coated by a monolayer of

HEC were obtained, where fibril dispersion and mechanical properties were

improved remarkably. A biomimetic composite with high toughness was

obtained in an analogous study on wood-based CNFs and HEC.76 The coating

layer of HEC on CNFs showed advantages for nanofiber alignment during tensile

stretching, resulting in biocomposites with high modulus and strength.88 The

effect of nanoscale interactions between the components on the mechanical

properties of biomimetic composites has been investigated.86 Adsorption of

polymers such as cellulose derivatives86 and xyloglucan89 can reduce friction on

cellulose surfaces, resulting in favorable nanofiber slippage and toughness

enhancement.90 The biocomposite films with small amount of these adsorbed

polysaccharides (approximately 2 wt%) become more ductile, indicated by

increased strain-to-failure and work of fracture.86, 87 This positive effect could also

be observed on biomimetic gels under wet conditions.87

Figure 5. Model of (a) ultrastructure organization of cellulose microfibrils within the

wood cell walls (Fengel 1971) and (b) ideal nanostructure of Core-shell cellulose

nanofibers.

BACKGROUND 15

To mimic the design strategies used by nature, Core-shell nanocomposites

with high CNF content and adsorbed polysaccharide matrix have been

developed.77, 91 The ideal model of Core-shell cellulose nanofibers is proposed in

Figure 5(b). The polysaccharide matrix is coated on the surface of CNFs, and

presented in a form of “shell” layer on “core” nanofiber surface. The unbound

matrix is removed from the mixed CNF/polysaccharide suspension, and Core-

shell nanocomposites are formed by filtration. Biomimetic approaches of nacre-

like composite prepared by polymer adsorption and filtration have been

developed for nanoclay-based materials.92, 93 In Core-shell nanocomposites, one

goal is to enhance and control the distribution of polymer matrix at

nanostructural level. The presence of the adsorbed matrix may inhibit self-

aggregation of CNFs in colloidal suspension. Steric stabilization of the

suspension is expected through polymer chain extension.94 The network with less

aggregated nanofibers is essential for ductility and strength of the final

nanocomposites. Moreover, the adsorbed matrix, which is confined on the CNF

surface, behaves differently from the bulk matrix. Strong interfacial interactions

between the components maximize the reinforcing effect in Core-shell

nanocomposites,77 especially under hydrated conditions.91 The hygromechanical

properties of the Core-shell nanocomposites are improved as a result of better

control of matrix distribution at the nanoscale.

2.5 Holocellulose nanofibers and their biocomposite films

Holocellulose is typically defined as a total polysaccharide fraction of

cellulose and hemicelluloses from wood cell walls. In fact, the native structural

organization of holocellulose, where the cellulose microfibrils are coated by

hemicellulose matrix,46 is the ideal structure for Core-shell nanofibers.

Holocellulose nanofibers with diameters of 10-20 nm have been isolated from

spruce wood powder via chlorite treatment and mechanical grinding by Iwamoto

et al.55 The high hemicellulose content is beneficial for nanofibrillation55 and the

coalescence of microfibrils into aggregates is inhibited.95 The physical properties

of holocellulose films were compared between once-dried and never-dried

nanofibers, and no significant difference was observed. The holocellulose films

showed tensile modulus of approximately 16 GPa and strength of 280 MPa,55

16 BACKGROUND

which is higher than that of enzyme-pretreated CNFs.63 The light transmittance

was slightly lower than for transparent acrylic resins, indicating good

transparency of the holocellulose film.55 Thus, holocellulose nanofibers are of

great interest for the preparation of biocomposites with good properties. Another

advantage of the holocellulose is the high yield of polysaccharide components

from starting material.

2.6 Application of nanofibers in porous materials

In the context of engineering design, porous materials can be classified into

two main categories; foams and honeycombs, depending on the architecture of

their solid cells.96 Foam is a cellular solid where the interconnected network of

the solid cells is built-up in three dimensions, while a honeycomb is the two-

dimensional arrays. The porous materials can possess low density; i. e., low

relative density – (ρ*/ρs), where ρ* is the density of the porous material and ρs is

the density of the bulk material.96 They have attracted attention for a wide range

of applications such as packaging,97, 98 insulating materials99 and biomedical

scaffolds.100, 101

There have been many studies on the potential use of CNFs as a reinforcing

component in porous materials, particularly for foams.98, 102-104 Since CNFs can

form stable suspensions in water, freeze-drying, also termed lyophilization,105 can

be utilized. The freeze-drying is a process, where a CNF suspension is subjected

to freezing prior to sublimation of the solidified phase under reduced pressure.101

The architecture of the porous materials such as pore morphology, pore size and

organization can be controlled during the freezing step, and depends on

nucleation and growth of ice-crystal templates.101, 103 Moreover, the cell wall

thickness and density can be altered by varying solid content and composition of

the cell wall components.103

The unique porous structure results in favorable mechanical response for the

foams. The CNF-based foams with a wide range of densities (7-103 kg/m3) and

mechanical properties can be tailored by varying the solid content of the mixture

suspension before freezing (Figure 6(a)).104 Mixtures of CNFs and water-soluble

polymers such as XG and starch can be prepared to improve mechanical

BACKGROUND 17

properties of the CNF foams. Ultra-high porosity CNF-based foams demonstrate

high mechanical properties and energy absorption capability, which is promising

for substitution of synthetic polymer-based polystyrene foams.102, 104

Compression tests are particularly interesting to investigate the mechanical

behavior of porous materials. Figure 6(b) shows a typical compressive stress-

strain curve of the CNF foams. There are three regions, which are correlated to

different deformation mechanisms, including (I) linear elasticity, (II) collapse

plateau, and (III) densification.96 Linear elasticity is controlled by cell wall

bending and stretching, and Young’s modulus (E*) is calculated as the initial

slope of the stress-strain curve. Collapse plateau is related to cell collapse, where

the strain is no longer recoverable. Based on the study on structure and

deformation mechanism of CNF foams by Ali and Gibson,106 the plateau is

associated with plastic deformation of the cell wall as well as some cracking. The

modulus and plateau stress of the foams are dependent on relative density as

described by Gibson and Ashby.96 Finally, densification occurs once the opposing

cell walls come into contact and completely collapse.

Figure 6. (a) Foam materials prepared from enzyme-pretreated CNFs (Sehaqui et

al. 2010)107 and (b) Compressive stress-strain curve for CNF foam (density 50

kg/m3), showing the three regimes of mechanical deformation.96

A honeycomb structure is a porous two-dimensional material formed with

well-defined pore connectivity and open porosity along one direction. The

18 BACKGROUND

cellular structure of wood is an example of honeycombs in nature (Figure 7(a)).

An ideal model of honeycomb structure is presented in Figure 7(b). The cell

structure of honeycombs is highly aligned, and thus offers good mechanical

response when loading along the cell, which is promising for load-bearing

applications. Both Young’s modulus (E*) and compressive strength (σ*) of

honeycombs vary linearly with relative density (ρ*/ρs):96

𝐸∗

𝐸𝑠=

𝜌∗

𝜌𝑠 (1)

𝜎∗

𝜎𝑠=

𝜌∗

𝜌𝑠 (2)

where Es and σs are the modulus and the compressive strength of the bulk

material. Relative density (ρ*/ρs) is the volume fraction of solid material.

One requirement for controlled fabrication of honeycomb-like structures is

associated with the control of cell wall alignment. From a practical point of view,

freeze-drying can be applied for the preparation of honeycombs. One way to

achieve CNF-based honeycombs is to control the direction of ice-crystal growth

during freezing. The concept is similar to freeze-casting technique used for

porous ceramics,100 where the material suspension is subjected to unidirectional

freezing before sublimation. As a consequence, the porous structure with

unidirectional channels,108 referred as honeycombs, is obtained.

Figure 7. SEM images of (a) cross-section and (b) longitudinal section of wood cell

wall structure (Gibson 2012)18 together with (c) ideal honeycomb structure.

EXPERIMENTAL 19

EXPERIMENTAL 3

The materials and overview of methods involved in the present thesis are

summarized below. More details are available in the appended papers.

3.1 Material components

3.1.1 Cellulose nanofibers (CNFs) prepared by enzymatic pretreatment

(Paper I-V)

To prepare enzyme-pretreated CNFs, never-dried spruce sulfite pulp (kindly

provided by Nordic Paper Seffle AB, Sweden) was treated with an endoglucanase

enzyme (Novozymes, Denmark), as previously reported by Henriksson et al.3, 48

For nanofibrillation, the enzyme-treated pulp was subjected to mechanical

disintegration using a microfluidizer (Microfluidics Ind., USA) for 3 passes

through big chambers (diameter of 400 and 200 µm), and 5 passes through small

chambers (diameter of 200 and 100 µm). A translucent suspension of enzyme-

pretreated CNFs (solid content ~ 2 wt%) was successfully prepared.

3.1.2 Holocellulose nanofibers prepared by peracetic acid pretreatment

(Paper IV-V)

To prepare holocellulose nanofibers, spruce wood chips were cut into fine

“match sticks”, and subjected to delignification using peracetic acid solution (32

wt% in acetic acid, Sigma-Aldrich) followed by mechanical disintegration. To

avoid decomposition of peracetic acid, metal ions in the wood sticks were

removed by treating with a DTPA/sodium sulfite solution before the

delignification step. The delignified wood fibers obtained from peracetic acid

pretreatment were slushed using a pulp slusher (10000 revolutions). For

nanofibrillation, the sample was subjected to mechanical disintegration using the

microfluidizer. It was sufficient to liberate the holocellulose nanofibers after

passing the sample for 1 pass through the big chambers and 3 passes through the

small chambers. A transparent suspension of holocellulose nanofibers (solid

content ~ 1 wt%) was successfully prepared. The holocellulose nanofibers are also

20 EXPERIMENTAL

referred to as Core-shell nanofibers (Paper IV-V), having the “shell” layer of native

hemicelluloses coating on the “core” nanofiber surface.

3.1.3 Polysaccharide matrices (Paper I-III)

Maize amylopectin (AP) was supplied from Sigma-Aldrich, Germany. Besides

amylopectin component, the purchased AP contains 7.4 wt% of amylose. An

aqueous suspension of AP (solid content ~ 1 wt%) was prepared with deionized

(DI) water. Several steps of heating (at 65 and 135 °C) were applied to achieve

solubilized AP. The obtained AP suspension was optically transparent. The

degree of polymerization (DP) for this AP is 1.03x106.

Tamarind seed xyloglucan powders (Glyloid 3S), which are a commercially

purified product, were provided by Dainippon, Japan. An aqueous suspension of

XG (solid content ~ 0.5 wt%) was prepared by dispersing the XG powders in DI

water, and heated at 50 °C overnight. To remove water-insoluble impurities, the

XG suspension was subsequently subjected to centrifugation at 4500 rpm for 30

min.29 The obtained XG suspension was freeze-dried and collected in a form of

dried XG for long time storage. The dried XG can be re-dispersed in DI water for

further experiments. The weight average molar mass (Mw) of XG is 493,000 Da,

and the number average molar mass (Mn) is 394,800 Da.

Spruce galactoglucomannan (GGM) can be isolated from process water of

thermomechanical pulping, as previously reported by Willför et al.32 and Xu et

al.109 The GGM powders used in the present work were provided by Willför

group. The spruce GGM from process water was purified by ethanol

precipitation. The obtained GGM has the weight average molar mass (Mw) of 39

Da. An aqueous suspension of spruce GGM (solid content ~ 0.1 wt%) was

prepared by dispersing the GGM powders in DI water, and heated at 90 °C for 4

h.

EXPERIMENTAL 21

3.2 Preparation of biocomposite materials

3.2.1 Cellulose and holocellulose nanopaper (Paper I-IV)

Cellulose and holocellulose nanopapers were prepared by a papermaking-like

process.63 A diluted suspension of nanofibers (solid content ~ 0.2 wt%) was mixed

using an Ultra Turrax mixer (IKA, T25 Digital) at 8000 rpm for 3 min.

Subsequently, the nanofibers suspension was degassed for 10 min. The network

of nanofibers was formed as a wet cake on a filter membrane (0.65 µm, DVPP

Millipore) during vacuum filtration, as described by Henriksson et al.48 To obtain

the nanopaper, the wet cake was subjected to vacuum drying using a Rapid

Köthen sheet dryer at 93 °C for 10 min.63

3.2.2 Neat polysaccharides (AP and XG) films (Paper I-II)

The preparation of Neat AP and Neat XG films was based on a general method

of film casting. Prior to film casting, aqueous suspensions of AP (solid content ~ 1

wt%) and XG (solid content ~ 0.5 wt%) were degassed. The samples were

carefully transferred to petri dishes coated with Teflon FEP film (Bytac®, USA)

and dried in an oven at 40 °C for a few days. A typical thickness of the resulting

films after drying was approximately 30-40 µm.

3.2.3 Adsorption experiment and preparation of Core-shell nanofibers

(Paper I-III)

For adsorption study, a series of CNF/AP, CNF/XG, and CNF/GGM mixtures

with varying matrix content were prepared from aqueous suspensions of CNFs

and the matrices. To allow adsorption, the mixed suspensions of CNF/AP and

CNF/XG (solid content ~ 0.2 wt% based on nanofibers) were thoroughly mixed

by magnetic stirring for 24 h at room temperature, while the mixing was

performed at 90 °C for 4 h in the case of CNF/GGM. After the adsorption period,

the excess/unbound matrix was removed from the mixture by centrifugation at a

speed of 4500 rpm for 10 min. The sediment was collected and washed several

times. The washing was performed by re-dispersing the sediment in DI water,

and repeating the centrifugation step in order to remove the unbound matrix.

22 EXPERIMENTAL

The obtained material contains CNFs coated with adsorbed polysaccharide

matrix, referred to as Core-shell nanofibers in Paper I-III.

3.2.4 Nanocomposite films (Paper I-III)

In the present thesis, nanocomposite films of CNFs with polysaccharide

matrices including AP, XG, and GGM were prepared by two different

approaches (Figure 8).

For Core-shell nanocomposites, diluted suspensions (solid content ~ 0.2 wt%)

of Core-shell nanofibers (CNF/AP, CNF/XG, CNF/GGM) with varying matrix

content were thoroughly dispersed using the Ultra Turrax, and then exposed to

degassing. Thereafter, the films were formed by vacuum filtration and drying,

according to the method for CNF nanopaper.

For “Mixed” nanocomposites, a series of aqueous mixtures for CNF/AP,

CNF/XG, and CNF/GGM suspensions (solid content ~ 0.2 wt%) with varying

matrix content were prepared. The thorough mixing was performed similar to

the procedure for adsorption experiment. The main difference between Core-shell

and “Mixed” nanocomposites is related to the unbound matrix. For “Mixed”

nanocomposites, the mixture suspensions of CNF/AP, CNF/XG, and CNF/GGM

were directly subjected to filtration without the removal of unbound matrix.

Other experimental details related to film formation were performed similar to

that for Core-shell nanocomposites.

EXPERIMENTAL 23

Figure 8. Schematic illustration of the difference between Core-shell and “Mixed”

approaches, used for the preparation of nanocomposite films. The figure is taken from

Paper I.

3.2.5 Porous materials based on nanofibers (Paper V)

Porous materials based on cellulose and holocellulose nanofibers were

prepared by freeze-drying, as previously reported by Sehaqui et al.104 In the

present thesis, both honeycombs and foams were developed by directional and

non-directional freezing, respectively. To obtain the porous materials with

varying densities, aqueous suspensions of cellulose and holocellulose nanofibers

with different concentrations were prepared. For the porous materials with low

density, the nanofiber suspensions were diluted with DI water. For the porous

materials with high density, the nanofiber suspensions were concentrated using

high-speed centrifugation (Avanti® J-E centrifuge, Beckman Coulter with a JLA-

10.500 rotor). By varying speed and time for the centrifugation, the nanofiber

suspensions with various concentrations were obtained. After the centrifugation,

the bottom phase, containing concentrated nanofibers, was collected. The

obtained nanofiber suspensions with different concentrations were carefully

transferred to aluminum molds (diameter of 50 mm) for freeze-drying.

24 EXPERIMENTAL

Prior to the freeze-drying, the samples were kept at 4 °C overnight in order to

avoid microscopic cracking during freezing. During freezing, the aluminum

molds were directly in contact with the liquid nitrogen (-196 °C), and thus the

samples were frozen. The preparation of foams and honeycombs are different

(Figure 9). For foams, the bottom and the side of the molds were directly in

contact with the liquid nitrogen, which allow non-directional freezing. In the case

of honeycombs, the samples were prepared by directional freezing. The side of

aluminum molds was covered with an insulating material, preventing ice-crystal

growth due to thermal gradient during the freezing. As a consequence, the ice-

crystal growth was allowed in only one direction, particularly from bottom to top

of the aluminum mold. Subsequently, the frozen samples were freeze-dried (3-4

days) in a benchtop freeze-dryer system (6 liter FreeZone, Labconco Corporation,

USA), as previously described by Sehaqui et al.104 Porous materials with different

densities varying between 6-55 kg/m3 were successfully prepared.

Figure 9. Schematic illustration of freezing step for the preparation of foams and

honeycombs. The figure is modified from Paper V.

EXPERIMENTAL 25

3.3 Characterization methods

3.3.1 Determination of matrix contents by sugar analysis (Paper I-III)

The content of polysaccharide matrix in the nanocomposites was determined

by sugar analysis, according to a standard method of SCAN-CM 71:09. Five main

sugar references of glucose, mannose, xylose, galactose, and arabinose were

selected for the analysis. To obtain dissolved monosaccharides for the analysis,

acid hydrolysis was performed on the freeze-dried sample of neat CNF and the

nanocomposites. Thereafter, the sugar composition was determined using high-

performance anion exchange chromatography system (Dionex, USA) equipped

with pulsed amperometric detection (HPAEC-PAD). Data processing was

performed with Chromeleon software.

3.3.2 Colloidal behavior by dynamic light scattering (DLS) (Paper I-III)

To characterize the aggregation of nanofibers in aqueous suspension, DLS

measurement was performed using a Zetasizer Nano ZEN3600 (Malvern

Instruments Ltd., UK). The light source was operated at 633 nm. The nanofiber

suspensions were adjusted to a final concentration of 100 mg/L in DI water. To

avoid interference from dust particles and large aggregates, the samples were

filtered through 5 µm membrane (Acrodisc) prior to the measurement. All

measurements were performed at 25 °C.

3.3.3 Total charge determination (Paper IV)

The total charge content of pretreated pulp fibers (enzymatic and peracetic

acid pretreatments) was determined by conductometric titration, according to a

standard method SCAN-CM 65:02. A diluted suspension of pretreated pulp

fibers was prepared in DI water. Prior to the titration, some NaCl was added in

the pulp suspension, and the pH was adjusted to 3 with 0.01 M HCl. During the

titration, a solution of 0.1 M NaOH was added and the conductivity of the pulp

suspension was followed. The total charge content was calculated based on the

sum of strong (sulphonic) and weak (carboxylic) acids on the pulp fibers.

26 EXPERIMENTAL

3.3.4 Nanostructure of nanofibers and their networks (Paper I-V)

Atomic force microscopy (AFM)

AFM was performed to analyze the nanostructure of nanofibers for Neat CNF

and Core-shell nanofibers. For all samples, a diluted suspension was spin-coated

on a mica disc. The samples were scanned using Nanoscope IIIa AFM (Veeco

Ltd., USA) in a tapping mode. The employed cantilevers had a tip radius of 8 nm

and spring constant of 5 N/m. Height images were recorded and used for the

analysis of nanofibers diameter.

Field-emission scanning electron microscopy (FE-SEM)

The morphology of nanofibrillar network and fracture surface of the

nanocomposite films in addition to the cellular structure of CNF-based porous

materials were observed using FE-SEM (Hitachi, S-4800, Japan). The sample was

placed on the surface of carbon tape mounted on a metal stub. Prior to

observation, the sample was coated with graphite and gold-palladium using

Cressington 208HR sputter coater for 5 and 30 s, respectively.

Transmission electron microscopy (TEM)

Structural information on the arrangement of nanofibers compared between

Core-shell and “Mixed” CNF/AP nanocomposites was provided by TEM

observation (FEI Tecnai Spirit, USA). The nanocomposites were embedded in

epoxy resin, and then cross-sectioned at room temperature using an ultra-

microtome (Leica EM UC7, Germany) equipped with a diamond knife (Diatome,

Switzerland). The obtained samples were stained by exposure for 1 h to RuO4

vapors.

3.3.5 Density and porosity (Paper I-V)

Density of the materials was determined based on their air-dried mass and

volume. The volume of each specimen was calculated by measuring thickness

and surface area of 3-5 set of data. At least five specimens were used for the

calculation of density for each sample.

EXPERIMENTAL 27

Porosity of the material was calculated based on the density according to the

following equation:

𝑃𝑜𝑟𝑜𝑠𝑖𝑡𝑦 = 1 − 𝜌𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙

𝜌𝑐𝑒𝑙𝑙𝑢𝑙𝑜𝑠𝑒 (3)

where the density of cellulose 𝜌𝑐𝑒𝑙𝑙𝑢𝑙𝑜𝑠𝑒 was considered to be 1500 kg/m3.48

3.3.6 Tensile mechanical properties (Paper I-IV)

Tensile measurements of Neat CNF and nanocomposite films were performed

using a universal testing machine Instron 5944 (Instron, USA) equipped with a

500 N load cell. The samples were conditioned at 23 °C under controlled relative

humidity of 50 or 85 RH% for at least one week prior to the tensile testing. The

samples were cut into thin rectangular strips (25x5 mm2). The gauge length was

set to 20 mm. The tensile behavior was recorded at a strain rate of 10 %/min. For

each sample, an average value of at least 5 specimens was reported for tensile

properties. Digital Speckle Photography (DSP) was performed to obtain an

accurate value of strain. A correlation of strains determined from the DSP and the

grip displacement during tensile testing was established. The reported modulus

was calculated based on the corrected value obtained from the DSP

measurement.

3.3.7 Mechanical properties in compression (Paper V)

The mechanical properties of porous materials (foams and honeycombs) were

measured by compression testing. Prior to the measurement, the sample was

conditioned at 23°C and 50 RH% for at least one week, and cut into 1 cm cubes.

The compression testing was performed using the Instron 5944 equipped with a

500 N load cell. The measurement was carried out on the samples both parallel

and perpendicular to the direction of ice-crystal growth. All specimens were

tested at a strain rate of 10%/min. For each sample, an average value from 3-5

specimens was reported for compressive properties.

3.3.8 Dynamic vapor sorption (DVS) (Paper I-III)

Moisture sorption isotherm of the materials was determined with a DVS

instrument (Surface measurement Systems Ltd, UK). Prior to DVS measurement,

28 EXPERIMENTAL

moisture was removed from the dried samples by additional drying in a vacuum

oven at 40 °C for 24 h. The DVS experiments were conducted in an adsorption

mode, starting from 0 to 90 RH%. Each RH condition was maintained for at least

400 min to ensure steady state. The moisture uptake at a particular RH% was

calculated according to the equation below:

𝑀𝑜𝑖𝑠𝑡𝑢𝑟𝑒 𝑢𝑝𝑡𝑎𝑘𝑒 = 100𝑥𝑊𝑚𝑜𝑖𝑠𝑡−𝑊𝑑𝑟𝑦

𝑊𝑑𝑟𝑦 (4)

where 𝑊𝑚𝑜𝑖𝑠𝑡 is the sample weight equilibrated at particular RH%, and 𝑊𝑑𝑟𝑦 is the

weight of the dried sample.

3.3.9 Molecular moisture dynamic by 2H solid-state NMR (Paper II)

2H solid-state NMR was used as a tool to investigate moisture-related

properties of Neat CNF, Neat XG and their nanocomposites at molecular level.

NMR relaxometry was performed to analyze molecular dynamics and molecular

distribution of moisture in the materials. Prior to the NMR experiment, all film

samples were conditioned at 92 RH% (saturated KNO3 solution in 2H2O) for 3

weeks. 2H2O was exploited in order to separate signals among the polymer and

water molecules. The NMR measurement was performed using a Bruker

Advance II spectrometer operating at a frequency of 46.1 MHz and a short (2.3

µs) 90° pulse. The distribution of 2H transverse relaxation time was obtained by

Inverse Laplace Transform (ILT).

3.3.10 Molar mass analysis of nanofibers by SEC (Paper IV)

Prior to molar mass analysis, the nanofibers were solvent-exchanged three

times with dry dimethyl acetamide (DMAc) at ambient temperature.

Subsequently, LiCl/DMAc and EIC reagents were added to dissolve the

nanofibers. The dissolved samples were treated in a Retsch vibratory mill type

MM-2 for 30 min followed by filtration through 0.45 µm PTFE-filter (Advantec,

MFS). The chromatographic system consists of a 2690 Separation Module (Waters

Corp.) equipped with a guard column (Mixed-A 20 µm 7.5x50 mm, Polymer

Laboratories) followed by four column (Mixed-A 20 µm 7.5x300 mm, Polymer

Laboratories) connected in series. The mobile phase of LiCl/DMAc was used. The

separations were performed at 80 °C. Molar mass calibration of SEC was

EXPERIMENTAL 29

performed using a linear pullulan standard set with Mw ranging from 738 to

1,660,000 Da (Polymer Laboratories). Detection was performed using a 2478 dual

wavelength absorbance detector (Water Corp.) at 295 nm coupled with a 410

differential refractometer (Waters Corp.).

3.3.11 Molar mass analysis of matrices by SEC-MALLS (Paper I-II)

Size exclusion chromatography coupled with a multi-angle laser light

scattering detector (SEC-MALLS) (PSS/Agilent SECurity GPC system, USA) was

employed to determine the molar mass of AP and XG matrices. The samples

were dissolved with a solvent system of DMSO/LiBr. Molar mass calibration of

SEC was performed using a linear pullulan standard set with Mw ranging from

342 to 805,000 Da (PSS, Germany). The solvent system of DMSO/LiBr was also

used for SEC analysis. Data processing was performed with PSS WinGPC

UniChrom software.

3.3.12 Light transmittance (Paper IV)

Light transmittance of nanopaper was measured using an UV spectrometer

(UV-2550, Shimadzu, Japan). The wavelengths ranging from 200 to 900 nm were

scanned. The transmittance spectra were recorded at room temperature, with air

as a reference.

RESULTS AND DISCUSSION 31

RESULTS AND DISCUSSION 4

The first section will describe Core-shell cellulose nanofibers with

polysaccharide matrices. The nanostructure, hygromechanical properties and

moisture sorption behavior of the obtained nanocomposite films will be

discussed. The second section will describe another type of Core-shell nanofibers,

referred to as native holocellulose nanofibers. Peracetic acid pretreatment will be

emphasized as an alternative method for preparing the holocellulose nanofibers.

The key characteristics of the holocellulose nanofibers as well as the materials

including film and honeycombs will be compared with those obtained from

enzyme-pretreated CNFs.

4.1 Core-shell nanofibers and their biocomposite films (Paper I-III)

4.1.1 Preparation of Core-shell nanofibers with polysaccharide matrices

To address the importance of matrix distribution at nanoscale, CNF-based

nanocomposites were prepared by two different methods, see Figure 8 and 10. In

the first approach, Core-shell cellulose nanofibers were prepared by coating the

polysaccharide matrix on the CNF surface. The bound matrix serves as a “shell”

layer for the “core” nanofiber. The excess matrix, termed “unbound” matrix, was

carefully removed from a colloidal suspension of Core-shell nanofibers prior to

film formation, as described in the experimental section. For “Mixed” method, it

is a traditional procedure where the colloidal suspension of CNFs and matrix is

subjected to simple mixing. The mixing step was allowed to continue until an

equilibrium state was reached. Thereafter, all nanocomposite films including

Core-shell and “Mixed” were formed by a papermaking-like process, where a wet

gel of nanofibers was obtained by filtration and then subjected to rapid water

evaporation.63 Strong secondary interactions including hydrogen bonding were

developed between the nanofibers after drying.

Different polysaccharide matrices including amylopectin (AP), xyloglucan

(XG), and galactoglucomannan (GGM) were exploited as the polymer matrix for

the CNF-based nanocomposites. The interactions between CNFs and these

32 RESULTS AND DISCUSSION

structurally different polysaccharides are responsible for their unique binding

capability.110 For Core-shell nanocomposites, the molecules of polysaccharide

matrix are tightly bound to the CNF surface. In contrast, “Mixed”

nanocomposites also have a fraction of unbound matrix. As shown in Figure 10,

the amount of remaining polymer matrix in Core-shell and “Mixed”

nanocomposites is plotted versus amount of polymer added to the suspension,

and the difference between two curves indicates the amount of unbound matrix.

A large fraction of unbound matrix is present in the case of AP matrix. The

difference in molecular structure among the matrix polysaccharides affects the

level of their interaction with cellulose. This also influences the fraction of bound

and unbound matrix in the final nanocomposites. Favorable association is

expected in both XG and GGM, which is related to the molecular rigidity of their

backbone.111-113 Also, adsorption to cellulose is part of their function in the plant

cell walls. XG can adsorb strongly to cellulose surfaces even in the hydrated

state.89 It should also be pointed out that molar mass can influence the observed

differences in the amount of remaining matrix. Higher molar mass may correlate

with a higher amount of adsorbed polymer.

RESULTS AND DISCUSSION 33

Figure 10. Composition of remaining matrix in the nanocomposites for (a) CNF/AP,

(b) CNF/XG, and (c) CNF/GGM. The data are for both “Mixed” and Core-shell

nanocomposites. The matrix contents in the obtained nanocomposite films are

presented next to the corresponding data points. The figure is a combination of data

from Paper I-III.

4.1.2 Nanostructure of the nanofibers and their networks

Nanofibers in aqueous suspension

It is essential to utilize the unique properties of individualized nanofibers in

nanocomposites. The preparation of CNF-based nanocomposites is typically

carried out in aqueous suspension. Therefore, the dispersion state of the colloidal

nanofibers is important to the nanostructure formation in the final

nanocomposites. This is eventually related to the material performance such as

mechanical properties.69 CNFs have a tendency to self-aggregate, which is an

unfavorable characteristic since agglomerates can act as stress concentrations and

cause premature nanocomposites fracture at small strain. To obtain a well-

34 RESULTS AND DISCUSSION

dispersed CNF suspension, electrostatic forces or steric hindrance can help to

separate the nanoparticles.114

Dynamic light scattering (DLS) data presents estimated hydrodynamic size of

nanofibers in aqueous suspension. As shown in Figure 11, distributions of the

size for Neat CNF and their Core-shell nanofibers with different polysaccharide

matrices are presented. Neat CNF exhibits a broad size distribution, with some

large size aggregates. In most cases, the size of large aggregates is reduced when

a polysaccharide matrix is adsorbed to the CNF surface. It is believed that the

presence of side chains or branches, for instance in AP, contributes to steric

stabilization68 of the CNFs in suspension.

It is important to note that DLS data is calculated based on diffusion

coefficients of spherical particle,115 which is not expected to correlate with the

nanofiber width. Hence, additional experiments were performed to observe

morphology and size distribution of the nanofibers. In the following section,

determination of the nanofiber width by atomic force microscopy (AFM) will be

presented.

Figure 11. Distribution of estimated hydrodynamic size of nanofibers in aqueous

suspension compared between Neat CNF and Core-shell nanofibers of (a) CNF/AP and

CNF/XG, and (b) CNF/GGM. The distributions are based on DLS data. The figure is a

combination of data from Paper I-III.

RESULTS AND DISCUSSION 35

Size distribution of nanofibers

The morphology and size distribution of nanofibers were observed from AFM

height images in order to avoid the tip broadening artifact. AFM section analysis

was pursued to determine the height of nanofibers, which in turn correlate to

their width. The histograms in Figure 12 show a comparison of height

distributions between Neat CNF and the Core-shell nanofibers with three different

polysaccharide matrices. The morphology of their nanofibers is also presented in

the inset. It is evident from the histograms that the diameter of nanofibers was in

the range of 2 to 20 nm. There are some differences between the Neat CNF used

for Paper I-II and III, perhaps due to the origin of the starting pulp fibers.

For all samples, the majority of nanofibers (80-95%) were in the range below

10 nm. The nanofiber aggregates with diameter above 10 nm were excluded in

part of the analysis. A summary of number average and weight average

diameters of all samples in addition to the percentage of nanofibers in the ranges

below 10 nm is listed in Table 2. Considering the values of number and weight

average diameters, Core-shell nanofibers tend to be wider in diameter, as

compared with their reference (Neat CNF). This is in line with the change in

percentage of nanofibers in the fine range below 10 nm. For Core-shell nanofibers,

the fraction of individual nanofibers with diameter of 3-5 nm decreased, while

the fraction of nanofibers with diameter of 5-8 nm and 8-10 nm increased

significantly as compared with Neat CNF. In this context, Core-shell CNF/GGM

shows the strongest effects.

The morphology of nanofibers for Core-shell CNF/GGM is depicted in Figure

12. FE-SEM images show the nanofibrillar network for Neat CNF and Core-shell

CNF/GGM hydrogels, which are dried from supercritical carbon dioxide (SC-

CO2). The nanofibers with larger diameter are clearly visible in Core-shell

CNF/GGM image. Since the GGM content is rather small (10.6 wt%), it is likely

that the GGM matrix is binding individual nanofibers together, which leads to

the increase in diameter of the observed nanofibers.

36 RESULTS AND DISCUSSION

Figure 12. Height distribution of nanofibers based on AFM measurement for (a) Neat

CNF, (b) Core-shell CNF/AP 23 wt%, (c) Core-shell CNF/XG 31 wt%, (d) Neat CNF for

GGM, and (d) Core-shell CNF/GGM 10.6 wt%. The light-colored region on the AFM

images represents the nanofibers with large diameter in the range of 10-14 nm. In the

case of CNF/GGM, nanofibers with increased diameters are also observable from the

FE-SEM images. The figure is a combination of data from Paper I-III.

RESULTS AND DISCUSSION 37

Table 2 Number average and weight average diameters (dn and dw), and the

percentage of nanofibers with focus on a small diameter fraction (< 10 nm). PDI is

polydispersity index dw/dn. The result is a combination of data from Paper I-III.

Nanofibrillar network in nanocomposites

In order to further study the difference in matrix distribution between Core-

shell and “Mixed” nanocomposites, the characteristics of their nanofibrillar

network were observed in dried hydrogels using FE-SEM. The CNF/XG

nanocomposite was selected to make a fair comparison between the two cases,

since their matrix content is comparable (XG ~ 31-32 wt%). As shown in Figure

13, there is a large difference in appearance among the nanofibrillar networks of

Neat CNF, Core-shell CNF/XG and “Mixed CNF/XG”. For Neat CNF, a highly

porous nanofibrillar network composed of uniformly sized nanofibers is

observed. For Core-shell CNF/XG, a network structure of nanofibers with larger

bundle size, as compared with Neat CNF, is noticeable. Based on molecular mass

analysis, the extended chain length of XG is as high as 800 nm in this study. It is

most likely that the assembly of nanofibers is formed via physical linkages of

adsorbed XG.116, 117 Instead, “Mixed” CNF/XG provides evidence for matrix-rich

regions in the presence of XG coated nanofibers. The XG matrix is also entrapped

within the pore space. As a consequence, an advantage of Core-shell over “Mixed”

nanocomposites is that the matrix distribution is more homogeneous and

controllable at nanoscale (Table 3).

Sample dn

(nm)

dw

(nm) PDI

% Fraction of all nanofibers

(unit: nm)

2.5-5.5 5.5-8.5 8.5-10.5 0-10

Neat CNF (Paper I-II) 5.7±2.2 6.3 1.1 39.2 35.1 10.8 86.5

Core-shell CNF/AP 23 wt% 6.1±2.0 6.6 1.1 30.0 44.7 11.3 87.3

Core-shell CNF/XG 31 wt% 6.3±2.2 6.7 1.1 27.1 38.2 14.7 80.0

Neat CNF (Paper III) 4.4±1.7 5.9 1.3 64.3 25.9 4.9 95.1

Core-shell CNF/GGM

10.6 wt% 5.4±1.7 7.5 1.2 35.6 45.3 8.7 89.6

38 RESULTS AND DISCUSSION

Figure 13. FE-SEM images of the nanofibrillar network for Neat CNF, and

nanocomposites of Core-shell CNF/XG 31 wt% and “Mixed” CNF/XG 32 wt%. Scale

bars are 300 nm. To preserve the nanofibrillar structure from aggregation during

typical drying, all samples were prepared from hydrogels (solid content ~ 15 wt%)

and subjected to SC-CO2 drying. The figure is taken from Paper II.

Table 3 Overview of preparation methods and characteristics of the nanofibrillar

network obtained from Core-shell vs “Mixed” nanocomposites

Core-shell nanocomposite “Mixed” nanocomposite

Preparation Polymer matrix is

adsorbed to CNF in

aqueous suspension.

Excess polymer is

removed prior to

filtration and drying.

Polymer and CNF are

mixed and then subjected

to filtration and drying.

Characteristic Polymer matrix is coated

on individual CNF.

Polymer matrix

distribution is more

homogeneous

CNF and polymer matrix

distributions are more

inhomogeneous.

RESULTS AND DISCUSSION 39

Fracture surface of nanocomposites

In addition to the observation of the nanofibrillar network, the differences in

fracture surface of the final films between Core-shell and “Mixed” nanocomposites

were also investigated. As shown in Figure 14, the film of CNF and all CNF/AP

nanocomposites (AP content ~ 23-25 wt%) showed lamellar structure.8, 48

However, there was a significant difference in the nanofibrillar nature among

them. For Core-shell CNF/AP (Figure 14(b)), the smooth surface of the nanofibers

indicates a homogeneous coating layer of AP. It is apparent that the network of

Core-shell CNF/AP is assembled from large bundles of the AP coated nanofibers,

instead of small nanofibril aggregates as can be observed in Neat CNF (Figure

14(a)). For the Mixed CNF/AP (Figure 14(c)), it can be seen that the lamellar

structure is inhomogeneous, apart from having smaller thickness for individual

layers as compared with Core-shell samples. This is most likely due to the

presence of unbound matrix in-between the CNF-rich laminae. Further

observation at fine scale was performed using TEM (Figure 14 (d,e)), and it

confirms the significant difference in nanostructure between Core-shell and

“Mixed” samples. Within a large thickness of individual lamina observed in

Figure 14(b), the layered structure of Core-shell nanocomposite is highly

organized at fine scale (Figure 14(d)). In contrast, larger agglomerated structures

are apparent in “Mixed” nanocomposites, see Figure 14(e).

The comparison of nanostructures between Core-shell and “Mixed” CNF/XG

nanocomposites with higher matrix content (31-32 wt%) is shown in Figure 15.

The fracture surfaces were observed after tensile testing. Features in line with

Core-shell nanocomposites, showing XG coated nanofibers, are noticeable; see

Figure 15(b). In Figure 15(c), “Mixed” CNF/XG also exhibits features of matrix-

rich regions within the fibrous network, which is in accordance with “Mixed”

CNF/AP nanocomposites.

40 RESULTS AND DISCUSSION

Figure 14. FE-SEM images of fracture surface for (a) Neat CNF, and nanocomposites of

(b) Core-shell CNF/AP 23 wt% and (c) “Mixed” CNF/AP 25 wt% films. These samples

were fractured under liquid N2. TEM images of ultra-thin sections for (d) the Core-

shell CNF/AP 23 wt% and (c) the “Mixed” CNF/AP 25 wt% also show the nanostructure

of the films at fine scale. The figure is modified from Paper I.

Figure 15. FE-SEM images of tensile fractured surface for (a) Neat CNF, (b) Core-shell

CNF/XG 31 wt% and (c) “Mixed” CNF/XG 32 wt% nanocomposite films. The figure is

taken from Paper II.

RESULTS AND DISCUSSION 41

4.1.3 Hygromechanical properties of Core-shell vs “Mixed” nanocomposites

The main focus in this section is related to the importance of matrix

distribution for hygromechanical properties of the biocomposites. The

mechanical properties of polysaccharide-based materials, including CNF

networks,69 are known to be very sensitive to moisture. The effect of matrix

distribution as well as interface structure in Core-shell and “Mixed”

nanocomposites is investigated with respect to mechanical properties, especially

at highly hydrated conditions.

Wet conditions

Let us consider the severe case when water is saturated in the system. The

comparison of Core-shell and “Mixed” nanocomposites was performed at two

different hydration states. First, wet gels of Neat CNF in addition to Core-shell and

“Mixed” CNF/XG were prepared and subjected to tensile testing (Figure 16(a) and

Table 4). The CNF/XG nanocomposites have common designs with the primary

cell walls. Data shows that XG provides physical linkages within the CNF

network,46 leading to an improvement in extensibility of the two CNF/XG

nanocomposites in wet condition.38 This is analogous to its biological function in

the wood cell walls.118 There is; however, a significant difference in tensile

properties in terms of Young’s modulus and strength for the two CNF/XG

nanocomposites. Depending on the preparation procedure, which is related to

the matrix distribution and interface structure, the mechanical properties of

CNF/XG wet gels are regulated in different manners. Core-shell CNF/XG

nanocomposites show unique improvements in mechanical properties.

Interestingly, the wet gel of Core-shell CNF/XG offers better performance in terms

of high modulus, strength, and extensibility compared with Neat CNF (at the

same dry content), despite XG matrix content up to 30 wt%. In addition to the

wet gels, soaked films of Core-shell and “Mixed” CNF/XG nanocomposites were

prepared and subjected to tensile testing (Figure 16(b) and Table 4). The effect of

different matrix distributions on the mechanical properties is very pronounced.

The soaked film of Core-shell CNF/XG shows better tensile properties than

“Mixed” nanocomposites. This is in good agreement with the study on wet gels.

Extensive cross-linking as well as effective matrix distribution at nanostructural

scale is an advantage for Core-shell CNF/XG nanocomposites.

42 RESULTS AND DISCUSSION

The study on mechanical properties in wet state of Core-shell nanocomposites

was further developed for the GGM matrix. This is mimicking the main

hemicellulose component of the secondary cell walls of spruce. As can be seen in

Figure 16(c), Core-shell CNF/GGM also showed significant improvement in tensile

properties as compared to Neat CNF. The corresponding tensile data of Core-shell

CNF/GGM is listed in Table 5. The development of the mechanical performance is

consistent with GGM content, which is an indirect indication of good dispersion

and homogeneous nanocomposite structure. The final GGM content was varied

in the narrow range between 5.6-13.4 wt% due to its low binding capability. The

adsorption results indicate that the majority of GGM exists in the form of bound

matrix (Figure 10(c)), presumably as a GGM coating of CNF. According to results

from AFM measurement, nanostructure of the nanofibers (Figure 12(d,e)) in

addition to considerations of the GGM content, it suggests that the presence of a

GGM coating facilitates physical linkages between adjacent nanofibers.113 The

GGM coating is closely integrated within the CNF network, resulting in Core-shell

CNF/GGM nanocomposites with superior mechanical properties under wet

conditions.

RESULTS AND DISCUSSION 43

Figure 16. Tensile stress-strain curves for CNF/XG and CNF/GGM nanocomposites

under wet conditions. The data are obtained from different testing conditions for (a)

wet gels of CNF/XG (after filtration without drying, see Table 4), (b) soaked films of

CNF/GGM (solid films are first dried and then soaked in water for 30 min, see Table

4) and (c) wet gels of CNF/GGM (after filtration without drying, see Table 5). The

figure is a combination of data from Paper II-III.

44 RESULTS AND DISCUSSION

Table 4 Tensile properties of Neat CNF, Core-shell CNF/XG 31 wt% and “Mixed”

CNF/XG 32 wt% nanocomposites as wet gels and soaked films. For the wet gel, the

suspension of nanofibers was filtered, and then the resulting material was subjected

to mechanical testing. When the wet gel was first dried and soaked in water for 30

minutes, this material is referred to as “Soaked films”. The data is extracted from

Paper II.

Wet gels (no drying) Soaked films (30 min)

Neat CNF Core-shell

CNF/XG

“Mixed”

CNF/XG

Core-shell

CNF/XG

“Mixed”

CNF/XG

Dry content (%) 22.7 24.0 21.7 23.7 22.7

Young’s modulus

(MPa)

9.5 ± 0.6 27.1 ± 4.8 3.3 ± 0.6 32.7 ± 8.1 7.2 ± 1.3

Tensile strength

(MPa)

0.4 ± 0.1 1.3 ± 0.4 0.2 ± 0.1 3.0 ± 0.8 1.1 ± 0.4

Elongation (%) 15.6 ± 3.8 41.9 ± 7.4 36.4 ± 7.0 32.2 ± 4.7 27.8 ± 4.8

Work of fracture

(kJ/m3)

34.8 ± 9.7 370.1 ± 93.9 47.5 ± 18.6 559.6 ± 112.3 148.1 ± 25.7

Table 5 Tensile properties for wet gels of Neat CNF and Core-shell CNF/GGM

nanocomposites with various GGM contents. The data is taken from Paper III.

Neat CNF

Core-shell CNF/GGM

(GGM content – wt%)

5.6 10.6 13.4

Solid content (%) 14.9 15.1 18.0 16.9

Young’s modulus (MPa) 4.1 ± 0.5 5.9 ± 2.2 21.3 ± 2.0 28.8 ± 6.6

Tensile strength (MPa) 0.11 ± 0.04 0.15 ± 0.03 0.33 ± 0.04 0.65 ± 0.16

Elongation (%) 9.2 ± 2.1 8.3 ± 1.0 15.8 ± 3.0 10.9 ± 1.3

Work of fracture (kJ/M3) 8.2 ± 3.0 7.8 ± 1.7 46.0 ± 10.1 50.8 ± 12.8

RESULTS AND DISCUSSION 45

Hydrated conditions (at relative humidity of 50 and 85 RH%)

There have been many studies on the improvement in mechanical properties

of CNF-based polymer films.48, 57, 76, 119, 120 The mechanical properties of

CNF/biopolymer films at high humidity are often critical for final engineering

applications. The aim of this experiment is to develop polysaccharide-based

biocomposite films with improved hygromechanical properties. The idea is

related to the outstanding mechanical behavior of Core-shell nanocomposites in

wet conditions, as shown in the previous section.

Starch is a polysaccharide, which has been widely investigated for potential

applications such as film and foam materials. There are, however, some

drawbacks with respect to the material performance and processing.

Amylopectin (AP) is the major constituent of starch. Previous studies in our

research group have demonstrated the potential of CNFs for plasticized AP. The

obtained CNF/plasticized AP films showed superior mechanical properties8 with

reduced moisture sensitivity,71 compared to the pure plasticized AP film. The

improvement of CNF/plasticized AP was primarily due to the inherent

characteristics of CNFs and their network. As the content of plasticized AP was

increased, the resulting materials showed strong reduction in mechanical

properties and moisture resistance.

In this experiment, Core-shell and “Mixed” CNF/AP nanocomposites were

prepared in the absence of plasticizer. Tensile behavior and corresponding

properties of the CNF/AP nanocomposites conditioned at 50 and 85 RH% are

presented in Figure 17 and Table 6. The difference in matrix distribution between

Core-shell and “Mixed” CNF/AP nanocomposites shows remarkable effects on

their hygromechanical properties. Outstanding mechanical properties can be

achieved in the case of Core-shell CNF/AP nanocomposites, which are comparable

to Neat CNF, despite variation in AP content. The tensile behavior of Neat CNF

and Core-shell CNF/AP in elastic and plastic regions is always consistent at

particular relative humidity (50 and 85 RH%). It is evident that the CNF network

completely dominates the properties of Core-shell CNF/AP nanocomposites, while

no such effect is observed for the “Mixed” CNF/AP. The reason is primarily due to

the control of bound matrix, which is coated on the nanofiber surface, in the case

of Core-shell nanocomposites.

46 RESULTS AND DISCUSSION

Figure 17. Tensile stress-strain curves for solid films of Neat CNF, Neat AP, Core-shell

CNF/AP 23 wt%, and “Mixed” CNF/AP 25 wt% nanocomposites at (a) 50 RH% and (b)

85 RH%. The figure is modified from Paper I.

RESULTS AND DISCUSSION 47

Table 6 Tensile properties for solid films of Neat CNF, Core-shell CNF/AP 23 wt% and

“Mixed” CNF/AP 25 wt% nanocomposites at different relative humidities (50 and 85

RH%). The values in parentheses are standard deviations. The data is extracted from

Paper I.

50 RH% 85 RH%

Neat

CNF

Neat

AP

Core-shell

CNF/AP

“Mixed”

CNF/AP

Neat

CNF

Neat

AP

Core-shell

CNF/AP

“Mixed”

CNF/AP

Density

(kg/m3)

1375 1445 1449 1415 - - - -

Porosity

(%)

8.5 NA 3.4 5.7 - - - -

Modulus

(GPa)

14.6

(0.7)

2.3

(0.2)

13.6

(0.2)

11.0

(0.7)

12.3

(0.8)

1.7

(0.3)

10.8

(0.5)

8.6

(0.4)

Tensile

strength

(MPa)

227

(26)

43

(4)

221

(17)

172

(20)

155

(26)

36

(3)

151

(12)

100

(12)

Yield

strength

(MPa)

121

(2)

46

(2)

117

(2)

90

(8)

60

(2)

33

(2)

69

(2)

50

(3)

Elongation

(%)

7.3

(1.5)

6.1

(1.8)

7.9

(1.5)

8.2

(1.3)

7.8

(1.5)

6.3

(0.6)

7.8

(0.6)

7.0

(1.7)

Work of

fracture

(MJ/m3)

11.4

(3.8)

2.4

(0.7)

12.1

(3.1)

9.7

(2.3)

8.0

(3.8)

1.9

(0.2)

8.0

(1.3)

4.9

(1.8)

In addition to AP, two structurally different hemicelluloses, XG and GGM,

were employed as a polysaccharide matrix for Core-shell nanocomposites. The

effects of matrix distribution and nanostructural interface on the

hygromechanical properties of Core-shell and “Mixed” CNF/XG conditioned at 50

and 85 RH% are demonstrated in Figure 18(a,b). Stress-strain curves of the two

CNF/XG nanocomposites in comparison with Neat CNF and Neat XG films are

provided. Compared with Neat CNF, the two CNF/XG nanocomposites showed

48 RESULTS AND DISCUSSION

lower stiffness, strength and elongation at 50 RH%. The result is in accordance

with the mechanical properties of biocomposites based on wet gel of bacterial

cellulose and XG.38 There is no significant difference between Core-shell and

“Mixed” CNF/XG at this condition (50 RH%). Perhaps strong interfacial

interaction121 between CNFs and XG in addition to the domains of entangled

and/or cross-linked fibrils38 are responsible for reduced ductility, reflecting

premature failure in the two CNF/XG films. At elevated humidity (85 RH%), the

two CNF/XG nanocomposites exhibited similar characteristics of lower stiffness

and strength than that of Neat CNF, whereas greater extensibility was observed in

the case of Core-shell CNF/XG film.

Tensile stress-strain curves for Core-shell CNF/GGM films, with varying GGM

contents in the range of 5.6-13.4 wt%, conditioned at 50 and 85 RH% are

presented in Figure 18(c,d). It is interesting to note that Neat CNF and Core-shell

CNF/GGM undergo similar elastic deformation. At 50 RH%, the modulus of Neat

CNF and Core-shell CNF/GGM films was varied within the same range of 13.5-13.8

GPa. When the humidity was elevated from 50 to 85 RH%, the moisture uptake of

Neat CNF and Core-shell CNF/GGM films was increased from approximately 7 to

13 wt%. Thus, the modulus was slightly decreased to the range of 12.0-12.8 GPa.

Strong strain-hardening48 was observed for Neat CNF and all Core-shell CNF/GGM

films, in spite of premature failure in the case of Core-shell CNF/GGM. The slope

observed during plastic deformation was comparable in all samples, and seems

insensitive to the GGM content. Strong interfacial adhesion between CNFs and

GGM is a likely explanation.

In brief, a significant difference in hygromechanical properties compared

between Core-shell and “Mixed” nanocomposite films is observed in the case of

the CNF/AP system. The evidence of nanostructual and interface effects is less

pronounced in CNF/hemicellulose films (CNF/XG and CNF/GGM). The reason is

primarily due to strong interfacial adhesion between CNFs and hemicelluloses.122,

123 The majority of hemicellulose matrix is intimately associated with the CNF

surface and confined within their nanofibrillar network. Only trace amount of

unbound hemicellulose is present in the CNF/hemicellulose nanocomposites.

Thus, the effect from “bulk” hemicellulose in “Mixed” nanocomposite films is

weak, despite strong effects in water saturated conditions.

RESULTS AND DISCUSSION 49

Figure 18. Tensile stress-strain curves for solid films of (a,b) CNF/XG and (c,d)

CNF/GGM nanocomposites at 50 and 85 RH%, respectively. The figure is a

combination of data from Paper II-III.

4.1.4 Moisture sensitivity of Core-shell vs “Mixed” nanocomposites

The inherent hygroscopic properties of polysaccharides are attributed to polar

hydroxyl groups and many polysaccharides have high solubility in water. CNFs,

however, possess attractive characteristics of semi-crystalline fibrillar structure.124

Moreover, they can form a dense nanofibrillar network,48 and have the ability to

provide strong molecular interaction via stabilized hydrogen bonding. Therefore,

there has been considerable interest to reduce moisture sensitivity of

polysaccharide-based materials by utilizing CNFs.10, 11, 70, 71 It has been suggested

that the existence of strong adhesion of the CNF/polysaccharide matrix interface

contributes to a reduction in moisture sensitivity.71, 125, 126 The aim of this study is

to investigate differences in interfacial effects at molecular and nanoscale.

50 RESULTS AND DISCUSSION

Moisture sorption isotherms compared between the nanocomposite films of

Core-shell CNF/XG 31 wt% and “Mixed” CNF/XG 32 wt% are plotted in Figure 19.

Prediction for moisture sorption of the CNF/XG nanocomposite based on rule of

mixtures; i.e. calculated on the basis of the moisture uptake contribution from the

individual components, is also displayed in the figure. It is apparent that the two

CNF/XG nanocomposites adsorb less moisture than expected, suggesting strong

molecular interactions between CNFs and XG in nature. Core-shell CNF/XG

nanocomposite exhibits the most promising result, where the moisture sorption

is almost identical to Neat CNF, despite an XG content of 31 wt%. This indicates

that the available sites for moisture sorption in the amorphous polymer are more

restricted in the case of Core-shell nanocomposites.

The moisture sorption data for Core-shell and “Mixed” CNF/AP

nanocomposite films are presented by showing the difference between

experimental data and predicted values based on the rule of mixtures, see Figure

20. Negative value of the differences means that the material adsorbs less

moisture than expected. The result consistently reveals that Core-shell CNF/AP

nanocomposite is less susceptible to moisture sorption than the “Mixed”

nanocomposite. This is in support of the difference in moisture sorption behavior

for the two CNF/XG nanocomposites. In Figure 20, data for Core-shell CNF/GGM

with varying GGM contents is also displayed. It is verified that Core-shell

nanocomposites generally adsorb less moisture than expected. This effect can be

correlated to the unique polymer matrix distribution for the Core-shell

nanocomposites. The more controlled nanostructural interphase is probably one

of the key reasons.

RESULTS AND DISCUSSION 51

Figure 19. Dynamic vapor sorption isotherms for solid films of Neat CNF, Neat XG,

“Mixed” CNF/XG 32 wt%, and Core-shell CNF/XG 31 wt%. Predictions for CNF/XG 31

wt% (MCNF/XG) is based on sorption in neat components and rule of mixtures; MCNF/XG =

MCNFWCNF + MXGWXG where M is measured moisture content, and W is weight

fraction of the component. The inset shows isotherms for CNF-based films in the

range 40-90% RH. The figure is taken from Paper II.

Figure 20. Comparison of moisture sorption between experimental data and

calculated values (based on rule of mixtures) for CNF/AP and CNF/GGM

nanocomposites. Negative values mean that the material adsorbs less moisture than

predicted from rule of mixtures. The figure is extracted from Paper I and II.

52 RESULTS AND DISCUSSION

To investigate the difference in material-moisture interaction between Core-

shell and “Mixed” CNF/XG nanocomposites at the molecular scale, measurement

of moisture mobility derived from 2H solid state NMR relaxation was performed.

The 2H NMR analysis of magnetization (M0r) and relaxation time (T2r) in addition

to moisture content (MCr) for Neat XG, Neat CNF, Core-shell CNF/XG 31 wt%, and

“Mixed” CNF/XG 32 wt% at 92 RH% are presented in Figure 21. In Figure 21(a),

the reported values are normalized to the values for Neat XG. M0r correlates with

the number of 2H atoms, which is related to the moisture content (MCr) in the

material. The moisture mobility in the Core-shell nanocomposites is significantly

reduced from that of “Mixed” nanocomposites, indicated by lower T2r in the Core-

shell CNF/XG film. In spite of differences in the water and matrix contents (Figure

21(b)) between Core-shell nanocomposite and Neat CNF, it is interesting to observe

that the water mobility in these materials is comparable. This indicates that water

clustering for Core-shell nanocomposites is localized at fibril surfaces, similar to

the Neat CNF.126 In “Mixed” nanocomposites, the water clustering results in faster

NMR-dynamics due to the presence of “unbound/bulk” matrix. This study

provides improved understanding of the mechanisms associated with the

nanostructural effects in the nanocomposites.

RESULTS AND DISCUSSION 53

Figure 21. (a) Ratios of gravimetric deuterated water content (MCr), NMR

magnetization (M0r) and relaxation time (T2r) for Neat XG, Neat CNF, “Mixed” CNF/XG

32 wt%, and Core-shell CNF/XG 31 wt% at 92 RH%. The reported values are

normalized to the values for Neat XG. The values in parentheses next to sample label

are the average moisture content from the NMR conditioning protocol. (b) A plot of 2H NMR T2 relaxation time as a function of gravimetric heavy water content for all

samples. Note that larger values for T2 correlate with increased water mobility. The

figure is taken from Paper II.

54 CONCLUSIONS

4.2 Conclusions (from Paper I-III)

We have investigated an alternative route for high-performance and moisture-

stable nanocomposites based on Core-shell concepts. Inspired by the cell wall

biosynthesis, Core-shell nanofibers have been prepared by coating wood CNFs

with a thin layer of polysaccharide matrix. This allows us to create nanofibers

with Core-shell structure, where the CNFs serve as the “core” reinforcing element

with the “shell” layer of polymer matrix. The preparation procedure in

combination with the data obtained from adsorption, colloidal particle size,

nanofiber diameter and high-resolution microscopy images are in support of the

Core-shell hypothesis. The nanocomposite network of Core-shell nanofibers has

been formed through a water-based papermaking-like process without polymer

loss. The Core-shell nanocomposites have advantages with respect to

nanostructural control of the polymer matrix distribution and interface structure

over a reference “Mixed” nanocomposites obtained from conventional mixing.

The nanostructural effects on mechanical and moisture-related properties for

the two nanocomposites have been investigated at different hydration states. The

Core-shell nanocomposite concept results in unique improvements in mechanical

properties and moisture stability compared with “Mixed” nanocomposites.

However, the extent of improvement in material performance for Core-shell

nanocomposites depends on the nature of the polymer matrix. For CNF/AP films,

the Core-shell nanocomposite conditioned at 50 and 85 RH% reveals strong

improvements in mechanical properties compared with the reference material.

This is due to nanostructural effects in terms of matrix distribution and interface

structure. For the CNF/hemicellulose samples (CNF/XG and CNF/GGM), the

Core-shell nanocomposites show a strong increase in mechanical properties

compared with “Mixed” nanocomposite and Neat CNF in wet conditions. The

enhancement arises from combined effects of favorable matrix distribution,

physical crosslinking, and strong interfacial adhesion between CNFs and

hemicelluloses. The results from 2H solid state NMR relaxation support the

difference in matrix distribution between Core-shell and “Mixed” nanocomposites

at the molecular level. In the Core-shell nanocomposites, strongly reduced water

mobility is obtained, possibly due to the confined XG matrix associated with the

CONCLUSIONS 55

CNF surface. Significant increase in water mobility is noticed for the “Mixed”

nanocomposites due to the presence of “bulk” or “unbound” matrix. The

increased moisture sorption in “Mixed” nanocomposites is in good agreement

with the NMR results. This can be explained in terms of the different states of

polymer matrix distribution in Core-shell and “Mixed” nanocomposites at

nanoscale. The Core-shell approach allows preparation of nanocomposite

materials with controlled nanostructure and improved hygromechanical

properties. Thus, the influence of polymer matrix distribution and interphase

structure has been clarified.

56 RESULTS AND DISCUSSION

4.3 Holocellulose nanofibers and their biocomposite films (Paper

IV-V)

4.3.1 Holocellulose nanofibers

Fundamental properties of nanofibers such as morphology, molecular weight,

surface chemistry and ease of nanofibrillation are importance for the preparation,

structure and properties of the nanocomposite materials. In addition to CNFs

obtained by enzymatic3 or chemical pretreatments,5 holocellulose nanofibers

containing high fraction of native hemicelluloses (up to 20 wt%) can be prepared.

Delignification processes such as chlorite treatment55 can be applied to obtain

holocellulose fibers prior to mechanical disintegration. Nevertheless, extensive

degradation of cellulose can be a drawback during chlorite delignification.127 It

suggests that the decrease in molecular weight of cellulose has a negative

influence on strength of the resulting material.48

The motivation of Paper IV is to introduce an alternative route for preparation

of holocellulose nanofibers. Peracetic acid pretreatment was performed directly

on wood chips prior to mechanical disintegration. In Figure 22, it is apparent that

the starting wood fibers are still intact after the peracetic acid pretreatment. The

fibers were progressively disintegrated into nanofibers by increasing the number

of passes through the microfluidizer. To evaluate the ease of disintegration for

nanofibers, the obtained materials are observed using FE-SEM. Sufficient

nanofibrillation of the holocellulose fibers was achieved after 1+5 passes (1 and 5

passes through the big and small chambers, see experimental section). Thus, the

disintegration of holocellulose nanofibers is more efficient in terms of nanofiber

structure and energy input than that required for the enzyme-pretreated CNFs

(3+5 passes). As shown in Figure 22, the obtained suspension of holocellulose

nanofibers shows good optical transparency, suggesting more uniform and

individualized nanofibers than the enzymatically pretreated CNFs commonly

used in our laboratory. The nanostructure observed by TEM and AFM in

addition to the distribution of holocellulose nanofiber diameters obtained by

AFM height analysis is presented in Figure 23. It is confirmed that the

holocellulose nanofibers are well-individualized. The diameter of holocellulose

RESULTS AND DISCUSSION 57

nanofibers is in the range of 2-11 nm. The majority of the holocellulose nanofibers

have diameters around 3-6 nm, which is closely related to the diameter of

individualized nanofibers (3-4 nm).5, 22 A significant decrease in large fibril

aggregates is observed, as compared to the reported data for CNFs in Paper I-II.

Figure 22. FE-SEM images of peracetic acid treated aspen wood after passing through

the microfluidizer with varying number of passes. The nomenclature of nanofibers is

referred to the number of pass through big + small chambers (1+5 means 1 and 5

passes through the big and small chambers, respectively). The photograph shows the

obtained suspension of holocellulose nanofibers compared with the reference

(enzyme-pretreated CNFs). The solid content of the suspensions is approximately 1-2

wt%. The figure is modified from Paper IV.

58 RESULTS AND DISCUSSION

Figure 23. Micrographs from (a) TEM and (b) AFM images of holocellulose nanofibers

obtained from aspen wood prepared by peracetic acid pretreatment and mechanical

distintegration. The sample was obtained after 6 passes (1+5) through the

microfluidizer. (c) Distribution of diameter for the holocellulose nanofibers based on

AFM height images. The figure is taken from Paper IV.

Two different sources of wood fibers, including aspen and spruce, were

subjected to peracetic acid pretreatment and nanofibrillation. Molar mass

analysis of these holocellulose nanofibers is compared with spruce enzyme-

pretreated CNFs in Table 7. The result reveals that holocellulose nanofibers of

high molar mass could be obtained from peracetic acid pretreatment. The

obtained nanofibers also show considerably lower values of polydispersity index

(PDI). It indicates a more narrow distribution of molar mass, which correlates

with nanofiber length.54 Compared with CNFs partially disintegrated with 1+0

passes, it is apparent that peracetic acid pretreatment shows promising result

with respect to high and preserved cellulose molar mass. Thus, the molar mass of

cellulose in the nanofibers is strongly influenced by the pretreatment approach.

The ease of nanofibrillation relies on chemical characteristics such as

hemicelluloses content55, 128 and surface charge.5 Table 8 summarizes these

characteristics of the nanofibers obtained from peracetic acid and enzymatic

pretreatments. It is evident that a considerable amount of native hemicelluloses

(up to 24 wt%) are preserved during the peracetic acid pretreatment. The charge

content of holocellulose nanofibers was approximately 200 µeq/g. Overall, the

hemicellulose and charge content of the holocellulose nanofibers is roughly twice

RESULTS AND DISCUSSION 59

as high as enzyme-pretreated CNFs. This plays an important role in the enhanced

nanofibrillation of the holocellulose nanofibers.

Table 7 Molar mass analyses of nanofibers obtained from peracetic acid and

enzymatic pretreatments determined using SEC. A different number of passes during

mechanical disintegration was applied to the nanofibers. The nomenclature of

nanofibers is referred to the number of passes through big + small chambers,

respectively. The data is taken from Paper IV.

Wood species Pretreatment Number of

passes

Molar mass

x 106 (g/mol)

Polydispersity

index (PDI)

Aspen Peracetic acid 1+0

1+1

1+3

1+5

1.3

1.3

0.83

0.83

19

15

11

7.0

Spruce Peracetic acid 1+0

1+1

1+3

1+5

1.1

0.83

0.62

0.54

12

9.4

6.5

6.3

Spruce Enzymatic 1+0

3+0

3+1

3+5

0.80

0.73

0.67

0.47

26

19

19

13

Table 8 Hemicellulose and charge contents of nanofibers obtained from peracetic acid

and enzymatic pretreatments. Different number of passes during mechanical

disintegration was applied to the nanofibers. The data is taken from Paper IV.

Wood species Pretreatment Number of

passes

Hemicellulose

content (wt%)

Charges

(µeq/g)

Aspen Peracetic acid 1+5 23.9 193

Spruce Peracetic acid 1+5 23.1 207

Spruce Enzymatic 3+5 12.3 101

60 RESULTS AND DISCUSSION

4.3.2 Films from holocellulose nanofibers

Optical transparency of holocellulose nanopaper film

Individualized nanofibers with uniform diameter is a key characteristic to

achieve high optical transparency.65 The light transmittance spectrum of the dried

films obtained from holocellulose nanofibers and enzyme-pretreated CNFs is

shown in Figure 24(a). The corresponding photograph of the obtained films with

a thickness of approximately 45 µm is presented in Figure 24(b). It is apparent

that the level of light transmittance for the holocellulose film was significantly

higher than that from the enzyme-pretreated CNFs, which can be explained in

terms of their differences in nanostructural characteristics. Interestingly, the

results are approaching the reported data for TEMPO-oxidized nanocellulose

films.65

Figure 24. (a) Light transmittance spectrum of spruce nanopaper prepared by

peracetic acid and enzymatic pretreatments and (b) photograph shows their light

transmittance behavior. The figure is taken from Paper IV.

Mechanical properties of holocellulose film

Tensile stress-strain curves of the holocellulose films prepared by peracetic

acid pretreatment are plotted in Figure 25, along with the corresponding tensile

properties in Table 9. The results are also compared with a typical film obtained

from enzyme-pretreated CNFs. An increase in mechanical properties in terms of

high Young’s modulus, tensile strength, and elongation was observed for the film

obtained from holocellulose nanofibers. The modulus was as high as 16.8 GPa

and 15.7 GPa for the holocellulose film from spruce and aspen, respectively. The

RESULTS AND DISCUSSION 61

fracture surface of holocellulose films exhibited a very dense structure. In

contrast, the film from enzyme-pretreated CNFs showed a typical layered

structure with porosity.48 Most likely, the unique architecture of holocellulose

films is influenced by the high fraction of native hemicelluloses. The tensile

strength of holocellulose films ranged from 290 to 327 MPa, depending on the

wood origin. Compared to the strength (243 MPa) and elongation (6.0%) of the

reference film (spruce enzyme-pretreat CNFs), the spruce holocellulose film

showed a strength of 290 MPa and elongation of 6.6%. These values reflect the

advantage of holocellulose nanofibers with high molar mass and uniform

diameter, obtained from peracetic acid pretreatment.

Figure 25. Comparison of tensile stress-strain curves and FE-SEM images of tensile

fracture surfaces for nanocellulose films prepared by peracetic acid and enzymatic

pretreatments. The figure is taken from Paper IV.

62 RESULTS AND DISCUSSION

Table 9 Tensile properties of nanocellulose films prepared by peracetic acid and

enzymatic pretreatments. The data is taken from Paper IV.

Wood

species

Pretreatment Number

of passes

Young’s

modulus

(GPa)

Tensile

strength

(MPa)

Elongation

(%)

Aspen Peracetic acid 1+5 15.7 ± 0.5 327 ± 12 7.2 ± 0.3

Spruce Peracetic acid 1+5 16.8 ± 1.4 290 ± 18 6.6 ± 0.4

Spruce Enzymatic 3+5 14.9 ± 1.5 243 ± 9 6.0 ± 0.4

4.3.3 Porous materials from holocellulose nanofibers

Honeycombs vs Foams

CNFs can serve as a reinforcement network for novel cellulosic materials such

as porous foams.102, 104 The nanoscale fibrils allow for tailoring cellular structures

of the porous materials at a fine scale. The purpose of this study is to prepare

CNF-based honeycombs, where the cellular structure of the porous materials is

better controlled. Honeycombs with different densities were prepared from

colloidal suspensions of holocellulose nanofibers with varying concentrations.

The relationship between the nanofiber concentration in suspension and density

of the resulting porous material is shown in Figure 26(a). Holocellulose

honeycombs with a limited range of densities (8-21 kg/m3) were obtained. It is

challenging to achieve high-density holocellulose honeycombs, due to the

difficulty to concentrate the suspension of holocellulose nanofibers. Compressive

stress-strain curves for the obtained holocellulose honeycombs are displayed in

Figure 26(b). The mechanical behavior of holocellulose honeycombs is strongly

dependent on the density. High density honeycombs resulted in improved

mechanical properties with respect to modulus, plateau stress and energy

absorption. Most likely, the cell wall thickness is increased resulting in increasing

density of the structural honeycombs.96

RESULTS AND DISCUSSION 63

Figure 26. (a) Concentration of nanofiber suspension before freezing and the resulting

density of honeycombs. (b) Longitudinal compressive stress-strain curves for

holocellulose honeycombs with varying densities. The figure is taken from Paper V.

Holocellulose honeycombs with highly aligned cell walls are expected due to

controlled directional freezing during preparation. Figure 27 illustrates the

hierarchical structure of the obtained holocellulose honeycombs. At

nanostructural level, the structural cell walls were built from a nanofibrillar

network of the holocellulose nanofibrils. At microstructure level, an elongated

and continuous porous network with well-defined pore connectivity was created.

It is evident that the cell geometry is highly anisotropic, and the alignment of the

cells is parallel to the freezing direction.

To investigate the relationship between cell wall structure and mechanical

behavior of the honeycombs, a comparison of Young’s modulus along

longitudinal and transverse directions is presented in Figure 28(a). There is a

substantial difference in modulus between the two directions. This unique

response confirms the anisotropic structure of the holocellulose honeycombs. In

contrast, for 3-dimensional cellular holocellulose foams (Figure 28(b)), only a

slight difference in modulus between the two directions was observed. This

reflects the low degree of anisotropy in cell shape of the foams. For the

honeycombs (Figure 28(a)), it is also interesting to observe a remarkable

improvement of the modulus in the longitudinal direction when the density is

increased. The modulus of the honeycombs was raised by two orders of

magnitude when the density was increased from 6 to 21 kg/m.3 This is

considerably improved compared to the reported data for CNF-based foams104

due to the anisotropy of cell geometry in the honeycombs.

64 RESULTS AND DISCUSSION

Figure 27. FE-SEM images of hierarchical structure for holocellulose honeycombs

with a density of 12 kg/m3. The yellow arrow indicates the freezing direction. The

figure is taken from Paper V.

Figure 28. Young’s modulus of holocellulose (a) 2-dimensional honeycombs and (b) 3-

dimensional foams compared in longitudinal and transverse loading directions. The

figure is taken from Paper V.

The micromechanics scaling models for honeycombs presented by Gibson

and Ashby96 allows us to predict the modulus of solid cell wall material (Es)

obtained from the holocellulose nanofibers. In Figure 29(a), the data for Young’s

modulus of the holocellulose honeycombs are plotted against relative density

(ρ*/ρs). By fitting the modulus data of holocellulose honeycombs to the equation

in Figure 29(a), a modulus of 133 MPa is estimated for the solid cell wall. This

value is two orders of magnitude lower than that for nanopaper structures of

holocellulose nanofibers (16.8 GPa), which should be representative for the solid

cell wall. There are several possible explanations. Most likely, the degree of

orientation of the cells is substantially off-axis. The relationship between cell axis

orientation and axial modulus is highly sensitive to orientation angle at small

RESULTS AND DISCUSSION 65

angles. In addition, defects in the form of large cell wall pores and smaller scale

porosity may also contribute to the reduced modulus.

To demonstrate the influence of microstructural control on mechanical

behavior of the porous materials, the reported data of modulus and energy

absorption for holocellulose foams in addition to nanocellulose foams from

previous study104 are also included in Figure 29(a,b). The modulus of the present

honeycombs is improved considerably compared to the foams, as expected from

theoretical predictions.96

Figure 29. (a) Young’s modulus and (b) energy absorption along longitudinal

direction of the honeycombs prepared from holocellulose nanofibers, plotted against

relative density. The result is presented with the data of holocellulose foams and

CNF-based foams from previous study by Sehaqui et al. (2010).104 The figure is taken

from Paper V.

Holocellulose vs Cellulose nanofibers

Inherent characteristics of the nanofibers, such as their mechanical properties

and ability to fuse and form a network are important for tailoring the porous

materials with high-performance and well-controlled structure. The focus of this

study is to compare the mechanical properties and cell wall structure of

honeycombs obtained from holocellulose versus cellulose nanofibers. The

holocellulose nanofibers were prepared from peracetic acid pretreatment,

whereas the cellulose nanofibers were prepared from enzymatic approach.

In Figure 30(a,b), the modulus and energy absorption of honeycombs

obtained from the two different nanofibers are plotted against relative density.

66 RESULTS AND DISCUSSION

The modulus of holocellulose honeycombs was significantly increased

(approximately 1 order of magnitude) from that of nanocellulose honeycombs,

although there was no difference in the energy absorption.

To learn about the details of the cellular structure, the microstructure of

honeycombs obtained from holocellulose and cellulose nanofibers was observed

using FE-SEM (Figure 31). The unique structure of honeycombs is apparent in

both cases. However, the cellular structure of holocellulose honeycombs is built

from a well-connected and more regular cell network. The microstructure of

holocellulose honeycombs partly explains the higher modulus. In contrast, the

honeycombs obtained from enzyme-pretreated CNFs (right image) contain large

pores, which explain lower honeycomb stiffness.

Figure 30. (a) Young’s modulus and (b) energy absorption along longitudinal

direction of the honeycombs prepared from holocellulose and cellulose nanofibers,

plotted against relative density. The figure is taken from Paper V.

RESULTS AND DISCUSSION 67

Figure 31. FE-SEM images of cell wall structure for honeycombs prepared from

(a) holocellulose nanofibers and (b) enzyme-pretreated CNFs. The materials have

similar density (approximately 11-12 kg/m3). The figure is taken from Paper V.

68 CONCLUSIONS

4.4 Conclusions (from Paper IV-V)

Holocellulose nanofibers are a promising nanocellulose reinforcement for

biocomposites, where the matrix is included as a CNF coating. The

corresponding nanocomposite films have improved mechanical properties

compared with previous CNF films. The holocellulose nanofibers can be

prepared directly from wood chips based on a combined process of peracetic acid

pretreatment and mechanical disintegration. The peracetic acid pretreatment

results in holocellulose nanofibers with high cellulose molar mass and uniform

CNF diameters. At the nanofibril level, the cellulose nanofiber “core” is coated by

the “shell” layer consisting of mixed native hemicelluloses. Nanofibrillation of the

pulp is facilitated, possibly due to the high content of hemicelluloses and the

presence of charged groups. The holocellulose nanofibers appear to mainly

consist of nanofibrils with a diameter as small as 3-6 nm.

By exploiting the unique characteristics of the obtained holocellulose

nanofibers, nanocellulose films with remarkable mechanical properties and

optical transparency were obtained. With respect to the mechanical properties,

the holocellulose films show Young’s modulus as high as 16.8 GPa, and tensile

strengths in the range of 290-327 MPa. The holocellulose films offer superior

mechanical properties over that obtained from enzyme-pretreated CNFs.

Moreover, the optical transparency of the holocellulose film is approaching that

obtained from TEMPO nanofibers (TOCNs). From the composite materials

perspective, structural control of nanofibers and polymer matrix distribution is

important in order to achieve desirable properties.

The potential to prepare CNF-based porous materials with improved

performance was investigated. Porous materials, including honeycombs and

foams, were prepared from nanofiber suspensions using freeze-drying

techniques. To prepare the honeycomb structure, a simple and versatile strategy

to control unidirectional freezing was applied for cell wall alignment. The

compressive modulus of the honeycombs is significantly improved compared

with foams, as expected from theoretical predictions. The effect is primarily

attributed to the cell alignment. A large difference of the modulus between

CONCLUSIONS 69

longitudinal and transverse directions is in support of a highly anisotropic

structure of the honeycombs, which is consistent with the microstructure

observed by FE-SEM. The predicted cell wall modulus was very low compared to

the modulus of holocellulose nanopaper. This indicates that improved cell

alignment and less cell wall defects would improve the properties even further.

The presence of native hemicelluloses facilitated the dispersion of holocellulose

nanofibers. Thus, the holocellulose honeycombs showed a more uniform cell

structure with less defects, as compared with honeycombs obtained from

enzyme-pretreated CNFs. This is also reflected in the increased out-of-plane

modulus for holocellulose honeycombs. The presence of native hemicelluloses in

the holocellulose nanofibers, acting as the “shell” layer, is of considerable

importance during preparation of the nanocomposite materials. In addition, the

interfiber bonding function of the hemicellulose “shell” has positive effects on the

final properties.

FUTURE WORK 71

FUTURE WORK 5

In order to enhance the mechanical properties of CNF-based films, there has

been an effort to align polymer coated88 or grafted61 nanofiber networks by

stretching. Better nanofibril alignment is achieved, resulting in CNF-based films

combining high modulus and strength. For Core-shell CNFs, the “shell” polymer is

promising and may provide nanoscale lubrication, facilitating nanofibril sliding

and reorientation in hydrated state. Wet gels of Core-shell nanocomposites can be

prepared and subjected to nanofibril alignment by stretching. As shown in the

present thesis, the wet gels of Core-shell nanofibers with hemicellulose matrices

(CNF/XG and CNF/GGM) exhibit improved stability compared with Neat CNF.

There may therefore be potential to align the wet gel of Core-shell

nanocomposites to a larger extent, and this would improve modulus and

strength.

This thesis has demonstrated that more inhomogeneous nanostructure (resin-

rich regions, CNF aggregates) has negative effects on the nanocomposite

properties. Improved hygromechanical properties and moisture stability of Core-

shell nanocomposite make these materials interesting as packaging materials.

Barrier properties of the Core-shell nanocomposite films need to be investigated. It

is possible that polymers strongly adsorbed to the cellulose nanofibers may show

improved gas barrier performance.

Core-shell nanofibers possess a large number of available hydroxyl groups

obtained from both CNFs and polysaccharide matrices. To extend the range of

properties for Core-shell nanocomposites, the surface hydroxyl groups can be

used as a molecular anchor for further functionalization or modification with

reactive chemical groups, for instance epoxides. The “shell” of neighboring CNFs

can be chemically cross-linked for improved hygromechanical stability. It may be

useful to achieve the materials with high-density of functional groups in

combination with improved moisture stability and reinforcing capability.

A successful development of CNF-based porous materials with highly

anisotropic structure (honeycombs) was demonstrated. From the viewpoint of

mechanical properties, honeycombs provide better out-of-plane performance

72 FUTURE WORK

than foams. However, the predicted modulus of the solid cell wall (Es) for

honeycombs is still far below that measured in tension for the nanopaper. It

suggests that the alignment of the structural cell wall is off-axis and that the cell

wall has defects. Therefore, better control of the cell wall alignment and cell

structure is necessary for improved mechanical properties.

ACKNOWLEDGEMENTS 73

ACKNOWLEDGEMENTS 6

First of all, I would like to express my sincere gratitude and appreciation to

my supervisor Prof. Lars Berglund for his encouragement and scientific guidance

throughout my study period. It has been a real pleasure working with him.

Special thanks are given to SCG Paper PCL (Thailand) in The Siam Cement

Group for the company’s scholarship and giving me the opportunity to

strengthen my research skills.

The Wallenberg Wood Science Center (WWSC) is gratefully acknowledged for

providing lab facilities.

I would like to acknowledge all my co-authors for their contributions to the

work. Former and present colleagues and friends in the Biocomposites group,

WWSC and the Department of Fiber and Polymer Technology are gratefully

acknowledged for making a pleasant working atmosphere. I would also like to

thank Assoc. Prof. Qi Zhou for his helpful advice. Aihua, Andong, Houssine,

Marielle and Sylvain for their kind support at the beginning of my study. Núria,

Michaela, Farhan, Gisela, Ivón, Surinthra (Janne), Anna, Mikaela, and Linn for

being really good friends and standing beside me. Ramiro is thanked for his

invaluable support in the lab. Malin and Jakob for their useful suggestions in the

group meetings. Seira, Ngesa, Joby, Qiliang, Nicholas, Andreas, Emma, and

Martin for their kinds and making the workplace more enjoyable. Special thanks

go to Shoaib for his useful suggestion for thesis preparation. Prof. Monica Ek is

acknowledged for giving support for quality check of the thesis.

I am also grateful to Khun Wichan Charoenkitsupat for giving me the

opportunity to become part of SCG Paper and always encouraging me to pursue

my ambitious goals. It meant a lot to me.

Finally, I am especially grateful to my dear family and boyfriend for offering

endless moral support and boundless love. All of their love, support and patience

give me the inner strength to go through all challenges. Thank you for always

being there for me.

REFERENCES 75

REFERENCES 7

1. Berglund, L. A.; Peijs, T., Cellulose biocomposites - From bulk moldings to

nanostructured systems. MRS Bulletin 2010, 35, 201-207.

2. Eichhorn, S. J.; Dufresne, A.; Aranguren, M.; Marcovich, N. E.; Capadona, J. R.;

Rowan, S. J.; Weder, C.; Thielemans, W.; Roman, M.; Renneckar, S.; Gindl, W.; Veigel,

S.; Keckes, J.; Yano, H.; Abe, K.; Nogi, M.; Nakagaito, A. N.; Mangalam, A.; Simonsen,

J.; Benight, A. S.; Bismarck, A.; Berglund, L. A.; Peijs, T., Review: current international

research into cellulose nanofibres and nanocomposites. Journal of Materials Science

2009, 45, 1-33.

3. Henriksson, M.; Henriksson, G.; Berglund, L. A.; Lindström, T., An

environmentally friendly method for enzyme-assisted preparation of microfibrillated

cellulose (MFC) nanofibers. European Polymer Journal 2007, 43, 3434-3441.

4. Pääkkö, M.; Ankerfors, M.; Kosonen, H.; Nykänen, A. I.; Ahola, S.; Österberg,

M.; Ruokolainen, J. T.; Laine, J.; Larsson, P. T.; Ikkala, O. T.; Lindström, T., Enzymatic

hydrolysis combined with mechanical shearing and high-pressure homogenization

for nanoscale cellulose fibrils and strong gels. Biomacromolecules 2007, 8, 1934-1941.

5. Saito, T.; Nishiyama, Y.; Putaux, J.-L.; Vignon, M.; Isogai, A., Homogeneous

suspensions of individualized microfibrils from TEMPO-catalyzed oxidation of native

cellulose. Biomacromolecules 2006, 7, 1687-1691.

6. Wågberg, L.; Decher, G.; Norgren, M.; Lindström, T.; Ankerfors, M.; Axnäs, K.,

The build-up of polyelectrolyte multilayers of microfibrillated cellulose and cationic

polyelectrolytes. Langmuir 2008, 27, 784-795.

7. Pei, A.; Butchosa, N.; Berglund, L. A.; Zhou, Q., Surface quaternized cellulose

nanofibrils with high water absorbency and adsorption capacity for anionic dyes. Soft

Matter 2013, 9, 2047-2055.

8. Svagan, A. J.; Azizi Samir, M. A. S.; Berglund, L. A., Biomimetic polysaccharide

nanocomposites of high cellulose content and high toughness. Biomacromolecules 2007,

8, 2556-2563.

9. Plackett, D.; Anturi, H.; Hedenqvist, M.; Ankerfors, M.; Gällstedt, M.;

Lindström, T.; Siró, I., Physical properties and morphology of films prepared from

microfibrillated cellulose and microfibrillated cellulose in combination with

amylopectin. Journal of Applied Polymer Science 2010, 3601-3609.

10. Hansen, N. M. L.; Blomfeldt, T. O. J.; Hedenqvist, M. S.; Plackett, D. V.,

Properties of plasticized composite films prepared from nanofibrillated cellulose and

birch wood xylan. Cellulose 2012, 19, 2015-2031.

11. Mikkonen, K. S.; Stevanic, J. S.; Joly, C.; Dole, P.; Pirkkalainen, K.; Serimaa, R.;

Salmén, L.; Tenkanen, M., Composite films from spruce galactoglucomannans with

microfibrillated spruce wood cellulose. Cellulose 2011, 18, 713-726.

12. Dufresne, A.; Vignon, M. R., Improvement of starch film performances using

76 REFERENCES

cellulose microfibrils. Macromolecules 1998, 31, 2693-2696.

13. López-Rubio, A.; Lagaron, J. M.; Ankerfors, M.; Lindström, T.; Nordqvist, D.;

Mattozzi, A.; Hedenqvist, M. S., Enhanced film forming and film properties of

amylopectin using micro-fibrillated cellulose. Carbohydrate Polymers 2007, 68, 718-727.

14. Lavoine, N.; Desloges, I.; Dufresne, A.; Bras, J., Microfibrillated cellulose - its

barrier properties and applications in cellulosic materials: a review. Carbohydrate

Polymers 2012, 90, 735-764.

15. Isogai, A.; Saito, T.; Fukuzumi, H., TEMPO-oxidized cellulose nanofibers.

Nanoscale 2011, 3, 71-85.

16. Siró, I.; Plackett, D., Microfibrillated cellulose and new nanocomposite

materials: a review. Cellulose 2010, 17, 459-494.

17. Mikkonen, K. S.; Tenkanen, M., Sustainable food-packaging materials based on

future biorefinery products: Xylans and mannans. Trends in Food Science & Technology

2012, 28, 90-102.

18. Gibson, L. J., The hierarchical structure and mechanics of plant materials. J R

Soc Interface 2012, 9, 2749-2766.

19. Neagu, C. R.; Gamstedt, K. E.; Bardage, S. L.; Lindström, M., Ultrastructural

features affecting mechanical properties of wood fibres. Wood Material Science and

Engineering 2006, 1, 146-170.

20. Sjöström, E., Wood chemistry: fundamentals and applications. Academic Press

Inc.: San Diego, 1993; pp 51-70.

21. Herth, W., Arrays of plasma-membrane ``rosettes´´ involved in cellulose

microfibril formation of Spirogyra. Planta 1983, 159, 347-356.

22. Fengel, D., Ideas on ultrastructural organization of cell-wall components. J.

Polymer Science: Part C 1971, 36, 383-392.

23. Terashima, N.; Fukushima, K.; He, L.-F.; Takabe, K., Comprehensive model of

the lignified plant cell wall. In Forage Cell Wall Structure and Digestibility, Jung, H. G.;

Buxton, D. R.; Hatfield, R. D.; Ralph, J., Eds. American Society of Agronomy-Crop

Science, Madison: Wisconsin, 1993; pp 247-270.

24. Niklas, K. J., Plant Biomechanics: An Engineering Approach to Plant Form and

Function. University of Chicago Press: Chicago, 1992.

25. Fengel, D.; Wegener, G., Wood: chemistry, ultrastructure, reactions. Walter de

Gruyter: Berlin, 1984.

26. Glasser, W. G., Cellulose and associated heteropolysaccharides. In Glycoscience,

Fraser-Reid, B.; Tatsuta, K.; Thiem, J., Eds. Springer Berlin Heidelberg: 2008; pp 1473-

1512.

27. Fry, S. C., The structure and functions of xyloglucan. J Exp Bot 1989, 40, 1-11.

28. Michael J. Gidley; Peter J. Lillford; David W. Rowlands; Peter Lang; Mariella

Dentini; Vittorio Crescenzi; Mary Edwards; Cristina Fanutti; Reid, J. S. G., Structure

and solution properties of tamarind-seed polysaccharide. Carbohydr Res 1991, 214,

229-314.

REFERENCES 77

29. Kochumalayil, J.; Sehaqui, H.; Zhou, Q.; Berglund, L. A., Tamarind seed

xyloglucan – a thermostable high-performance biopolymer from non-food feedstock.

Journal of Materials Chemistry 2010, 20, 4321-4327.

30. Willför, S.; Sjöholm, R.; Laine, C.; Roslund, M.; Hemming, J.; Holmbom, B.,

Characterisation of water-soluble galactoglucomannans from Norway spruce wood

and thermomechanical pulp. Cabohydrate Polymers 2003, 52, 175-187.

31. Willför, S.; Sundberg, A.; Hemming, J.; Holmbom, B., Polysaccharides in some

industrially important softwood species. Wood Science and Technology 2005, 39, 245-

257.

32. Willför, S.; Rehn, P.; Sundberg, A.; Sundberg, K.; Holmbom, B., Recovery of

water-soluble acetylgalactoglucomannans from mechanical pulp of spruce. Tappi

Journal 2003, 2, 27-32.

33. Song, T.; Pranovich, A.; Sumerskiy, I.; Holmbom, B., Extraction of

galactoglucomannan from spruce wood with pressurised hot water. Holzforschung

2008, 62, 659-666.

34. Hartman, J.; Albertsson, A.-C.; Lindblad, M. S.; Sjöberg, J., Oxygen barrier

materials from renewable sources: Material properties of softwood hemicellulose-

based films. Journal of Applied Polymer Science 2006, 100, 2985-2991.

35. Mikkonen, K. S.; Heikkilä, M. I.; Helén, H.; Hyvönen, L.; Tenkanen, M., Spruce

galactoglucomannan films show promising barrier properties. Carbohydrate Polymers

2010, 79, 1107-1112.

36. Carpita, N. C.; McCann, M. C., Chapter 2 ``The cell wall´´. In Biochemistry and

Molecular Biology of Plants, Buchanan, B. B.; Gruissem, W.; Jones, R., Eds. American

Society Plant Physiologists, Rockville, MD.: 2000; pp 52-108.

37. Markus Pauly; Peter Albersheim; Alan Darvill; York, W. S., Molecular domains

of the cellulose/xyloglucan network in the cell walls of higher plants. The Plant Journal

1999, 20, 629-639.

38. Whitney, S. E. C.; Gothard, M. G. E.; Mitchell, J. T.; Gidley, M. J., Roles of

cellulose and xyloglucan in determining the mechnaical properties of primary plant

cell walls. Plant Physiol 1999, 121, 657-663.

39. Zhang, Q.; Brumer, H.; Agren, H.; Tu, Y., The adsorption of xyloglucan on

cellulose: effects of explicit water and side chain variation. Carbohydr Res 2011, 346,

2595-2602.

40. Chanliaud, E.; Burrows, K. M.; Jeronimidis, G.; Gidley, M. J., Mechanical

properties of primary plant cell wall analogues. Planta 2002, 215, 989-996.

41. Lawoko, M.; Henriksson, G.; Gellerstedt, G., Structural differences between the

lignin-carbohydrate complexes present in wood and in chemical pulps.

Biomacromolecules 2005, 6, 3467-3473.

42. Lawoko, M.; Henriksson, G.; Gellerstedt, G., Characterization of lignin-

carbohydrate complexes from spruce sulfite pulp. Holzforschung 2006, 60, 162-165.

43. Aspinall, G. O., Chemistry of cell wall polysaccharides. In Carbohydrates:

78 REFERENCES

Structure and Function, 1980; pp 473-500.

44. Hayashi, T.; Kaida, R., Functions of xyloglucan in plant cells. Mol Plant 2011, 4,

17-24.

45. Montezinos, D., The role of the plasma membrane in cellulose microfibrils

assembly. In The cytoskeleton in plant growth and development, W., L. C., Ed. Academic

Press: London, 1982; pp 147-162.

46. Cosgrove, D. J., Assembly and enlargement of the primary cell wall in plants.

Annu. Rev. Cell Dev. Biol. 1997, 13, 171-201.

47. Edelmann, H. G., Water potential modulates extensibility of rye coleoptile cell

wall. Botanica Acta 1995, 108, 374-380.

48. Henriksson, M.; Berglund, L. A.; Isaksson, P.; Lindström, T.; Nishino, T.,

Cellulose nanopaper structure of high toughness. Biomacromolecules 2008, 9, 1579-1585.

49. Sehaqui, H.; Ezekiel Mushi, N.; Morimune, S.; Salajkova, M.; Nishino, T.;

Berglund, L. A., Cellulose nanofiber orientation in nanopaper and nanocomposites by

cold drawing. ACS Appl Mater Interfaces 2012, 4, 1043-1049.

50. Sehaqui, H.; Zhou, Q.; Ikkala, O.; Berglund, L. A., Strong and tough cellulose

nanopaper with high specific surface area and porosity. Biomacromolecules 2011, 12,

3638-3644.

51. Habibi, Y., Key advances in the chemical modification of nanocelluloses. Chem

Soc Rev 2014, 43, 1519-1542.

52. Turbak, A. F.; Snyder, F. W.; Sandberg, K. R., Microfibrillated cellulose, a new

cellulose product: properties, uses, and commercial potential. J. Appl. Polym. Sci.: Appl.

Polym. Symp. 1983, 37, 815-827.

53. Saito, T.; Hirota, M.; Tamura, N.; Kimura, S.; Fukuzumi, H.; Heux, L.; Isogai,

A., Individualization of nano-sized plant cellulose fibrils by direct surface

carboxylation Using TEMPO catalyst under neutral conditions. Biomacromolecules

2009, 10, 1992-1996.

54. Shinoda, R.; Saito, T.; Okita, Y.; Isogai, A., Relationship between length and

degree of polymerization of TEMPO-oxidized cellulose nanofibrils. Biomacromolecules

2012, 13, 842-849.

55. Iwamoto, S.; Abe, K.; Yano, H., The effect of hemicelluloses on wood pulp

nanofibrillation and nanofiber network characteristics. Biomacromolecules 2008, 9,

1022-1026.

56. Hult, E.-L.; Iversen, T.; Sugiyama, J., Characterization of the supermolecular

structure of cellulose in wood pulp fibres. Cellulose 2003, 10, 103-110.

57. Nakagaito, A. N.; Yano, H., The effect of morphological changes from pulp

fiber towards nano-scale fibrillated cellulose on the mechanical properties of high-

strength plant fiber based composites. Applied Physics A: Materials Science & Processing

2004, 78, 547-552.

58. Moon, R. J.; Martini, A.; Nairn, J.; Simonsen, J.; Youngblood, J., Cellulose

nanomaterials review: structure, properties and nanocomposites. Chem Soc Rev 2011,

REFERENCES 79

40, 3941-3994.

59. Nogi, M.; Iwamoto, S.; Nakagaito, A. N.; Yano, H., Optically transparent

nanofiber paper. Advanced Materials 2009, 21, 1595-1598.

60. Saito, T.; Kuramae, R.; Wohlert, J.; Berglund, L. A.; Isogai, A., An ultrastrong

nanofibrillar biomaterial: the strength of single cellulose nanofibrils revealed via

sonication-induced fragmentation. Biomacromolecules 2013, 14, 248-253.

61. Tang, H.; Butchosa, N.; Zhou, Q., A Transparent, hazy, and strong macroscopic

ribbon of oriented cellulose nanofibrils bearing poly(ethylene glycol). Advanced

Materials 2015, 27, 2070-2076.

62. Sakurada I., N. Y., Ito T., Experimental determination of the elastic modulus of

crystalline regions in oriented polymers. Journal of Polymer Science 1962, 57, 651-659.

63. Sehaqui, H.; Liu, A.; Zhou, Q.; Berglund, L. A., Fast preparation procedure for

large, flat cellulose and cellulose-inorganic nanopaper structures. Biomacromolecules

2010, 11, 2195-2198.

64. Battista, O. A., Microcystal Polymer Science McGraw Hill: New York, 1975.

65. Fukuzumi, H.; Saito, T.; Iwata, T.; Kumamoto, Y.; Isogai, A., Transparent and

high gas barrier films of cellulose nanofibers prepared by TEMPO-mediated

oxidation. Biomacromolecules 2009, 10, 162-165.

66. Nogi, M.; Handa, K.; Nakagaito, A. N.; Yano, H., Optically transparent

bionanofiber composites with low sensitivity to refractive index of the polymer

matrix. Applied Physics Letters 2005, 87, 243110/1-3.

67. Fall, A. B.; Lindstrom, S. B.; Sundman, O.; Odberg, L.; Wagberg, L., Colloidal

stability of aqueous nanofibrillated cellulose dispersions. Langmuir 2011, 27, 11332-

11338.

68. Hubbe, M. A.; Rojas, O. J., Colloidal stability and aggregation of lignocellulosic

material in aqueous: A review. BioResources 2008, 3, 1419-1491.

69. Benitez, A. J.; Torres-Rendon, J.; Poutanen, M.; Walther, A., Humidity and

multiscale structure govern mechanical properties and deformation modes in films of

native cellulose nanofibrils. Biomacromolecules 2013, 14, 4497-4506.

70. Dufresne, A.; Dupeyre, D.; Vignon, M. R., Cellulose microfibrils from potato

tuber cells: Processing and characterization of starch-cellulose microfibril xomposites.

Journal of Applied Polymer Science 2000, 76, 2080-2092.

71. Svagan, A. J.; Hedenqvist, M. S.; Berglund, L., Reduced water vapour sorption

in cellulose nanocomposites with starch matrix. Composites Science and Technology

2009, 69, 500-506.

72. Stevanic, J. S.; Mikkonen, K. S.; Xu, C.; Tenkanen, M.; Berglund, L.; Salmén, L.,

Wood cell wall mimicking for composite films of spruce nanofibrillated cellulose with

spruce galactoglucomannan and arabinoglucuronoxylan. Journal of Materials Science

2014, 49, 5043-5055.

73. Stevanic, J. S.; Bergström, E. M.; Gatenholm, P.; Berglund, L.; Salmén, L.,

Arabinoxylan/nanofibrillated cellulose composite films. Journal of Materials Science

80 REFERENCES

2012, 47, 6724-6732.

74. Gröndahl, M.; Eriksson, L.; Gatenholm, P., Material properties of plasticized

hardwood xylans for potential application as oxygen barrier films. Biomacromolecules

2004, 5, 1528-1535.

75. Peng, X. W.; Ren, J. L.; Zhong, L. X.; Sun, R. C., Nanocomposite films based on

xylan-rich hemicelluloses and cellulose nanofibers with enhanced mechanical

properties. Biomacromolecules 2011, 12, 3321-3329.

76. Sehaqui, H.; Zhou, Q.; Berglund, L. A., Nanostructured biocomposites of high

toughness—a wood cellulose nanofiber network in ductile hydroxyethylcellulose

matrix. Soft Matter 2011, 7, 7342-7350.

77. Prakobna, K.; Galland, S.; Berglund, L. A., High-performance and moisture-

stable cellulose-starch nanocomposites based on bioinspired core-shell nanofibers.

Biomacromolecules 2015, 16, 904-912.

78. Angles, M. N.; Dufresne, A., Plasticized starch/tunicin whiskers

nanocomposite material. 2. mechanical behavior. Macromolecules 2001, 34, 2921-2931.

79. Dammström, S.; Salmén, L.; Gatenholm, P., The effect of moisture on the

dynamical mechanical properties of bacterial cellulose/glucuronoxylan

nanocomposites. Polymer 2005, 46, 10364-10371.

80. Hull, D.; Clyne, T. W., An Introduction to Composite Materials. In Cambridge

Solid State Science Series, 2nd ed.; Cambridge University: 1996.

81. Dufresne, A., Nanocellulose: From nature to high performance tailored materials.

Walter de Gruyter: 2012.

82. Valent, B. S.; Albersheim, P., The structure of plant cell walls V. On the binding

of xyloglucan to cellulose fibers. Plant Physiol. 1974, 57, 105-108.

83. Fratzl, P.; Burgert, I.; Gupta, H. S., On the role of interface polymers for the

mechanics of natural polymeric composites. Physical Chemistry Chemical Physics 2004,

6, 5575-5579.

84. Keckes, J.; Burgert, I.; Fruhmann, K.; Muller, M.; Kolln, K.; Hamilton, M.;

Burghammer, M.; Roth, S. V.; Stanzl-Tschegg, S.; Fratzl, P., Cell-wall recovery after

irreversible deformation of wood. Nat Mater 2003, 2, 810-814.

85. Zhou, Q.; Malm, E.; Nilsson, H.; Larsson, P. T.; Iversen, T.; Berglund, L. A.;

Bulone, V., Nanostructured biocomposites based on bacterial cellulosic nanofibers

compartmentalized by a soft hydroxyethylcellulose matrix coating. Soft Matter 2009, 5,

4124-4130.

86. Olszewska, A.; Valle-Delgado, J. J.; Nikinmaa, M.; Laine, J.; Osterberg, M.,

Direct measurements of non-ionic attraction and nanoscaled lubrication in biomimetic

composites from nanofibrillated cellulose and modified carboxymethylated cellulose.

Nanoscale 2013, 5, 11837-11844.

87. Lucenius, J.; Parikka, K.; Österberg, M., Nanocomposite films based on

cellulose nanofibrils and water-soluble polysaccharides. Reactive and Functional

Polymers 2014, 85, 167-174.

REFERENCES 81

88. Sehaqui, H.; Morimune, S.; Nishino, T.; Berglund, L. A., Stretchable and strong

cellulose nanopaper structures based on polymer-coated nanofiber networks: an

alternative to nonwoven porous membranes from electrospinning. Biomacromolecules

2012, 13, 3661-3667.

89. Stiernstedt, J.; Brumer, H.; Zhou, Q.; Teeri, T. T.; Rutland, M. W., Friction

between cellulose surfaces and effect of xyloglucan adsorption. Biomacromolecules

2006, 7, 2147-2153.

90. Larson, M. C.; Miles, H. F., On the effects of friction, roughness and toughness

on interfacial sliding in brittle composites. Mechanics of Materials 1998, 27, 77-89.

91. Prakobna, K.; Terenzi, C.; Zhou, Q.; Furó, I.; Berglund, L. A., Core-shell

cellulose nanofibers for biocomposites - nanostructural effects in hydrated state.

Carbohydrate Polymers 2015, 125, 92-102.

92. Walther, A.; Bjurhager, I.; Malho, J. M.; Pere, J.; Ruokolainen, J.; Berglund, L.

A.; Ikkala, O., Large-area, lightweight and thick biomimetic composites with superior

material properties via fast, economic, and green pathways. Nano Lett 2010, 10, 2742-

2748.

93. Sehaqui, H.; Kochumalayil, J.; Liu, A.; Zimmermann, T.; Berglund, L. A.,

Multifunctional nanoclay hybrids of high toughness, thermal, and barrier

performances. ACS Appl Mater Interfaces 2013, 5, 7613-7620.

94. Olszewska, A.; Junka, K.; Nordgren, N.; Laine, J.; Rutland, M. W.; Österberg,

M., Non-ionic assembly of nanofibrillated cellulose and polyethylene glycol grafted

carboxymethyl cellulose and the effect of aqueous lubrication in nanocomposite

formation. Soft Matter 2013, 9, 7448-7457.

95. Duchesne, I.; Hult, E.-L.; Molin, U.; Daniel, G.; Iversen, T.; Lennholm, H., The

influence of hemicellulose on fibril aggregation of kraft pulp fibres as revealed by FE-

SEM and CP/MAS 13C-NMR. Cellulose 2001, 8, 103-111.

96. Gibson, L. J.; Ashby, M. F., Cellular solids - Structure and properties. 2 ed.;

Cambridge University Press: Cambridge, 1997.

97. Avella, M.; Cocca, M.; Errico, M.; Gentile, G., Biodegradable PVOH-based

foams for packaging applications. Journal of Cellular Plastics 2011, 47, 271-281.

98. Svagan, A. J.; Berglund, L. A.; Jensen, P., Cellulose nanocomposite biopolymer

foam-hierarchical structure effects on energy absorption. ACS Appl Mater Interfaces

2011, 3, 1411-1417.

99. Baetens, R.; Jelle, B. P.; Gustavsen, A., Aerogel insulation for building

applications: A state-of-the-art review. Energy and Buildings 2011, 43, 761-769.

100. Deville, S.; Saiz, E.; Tomsia, A. P., Freeze casting of hydroxyapatite scaffolds

for bone tissue engineering. Biomaterials 2006, 27, 5480-5489.

101. Deville, S., Freeze-casting of porous biomaterials: Structure, properties and

opportunities. Materials 2010, 3, 1913-1927.

102. Svagan, A. J.; Samir, M. A. S. A.; Berglund, L. A., Biomimetic foams of high

mechanical performance based on nanostructured cell walls reinforced by native

82 REFERENCES

cellulose nanofibrils. Advanced Materials 2008, 20, 1263-1269.

103. Svagan, A. J.; Jensen, P.; Dvinskikh, S. V.; Furó, I.; Berglund, L. A., Towards

tailored hierarchical structures in cellulose nanocomposite biofoams prepared by

freezing/freeze-drying. Journal of Materials Chemistry 2010, 20, 6646-6654.

104. Sehaqui, H.; Salajková, M.; Zhou, Q.; Berglund, L. A., Mechanical performance

tailoring of tough ultra-high porosity foams prepared from cellulose I nanofiber

suspensions. Soft Matter 2010, 6, 1824-1832.

105. Jennings, T. A., Lyophilization: Introduction and Basic Principles.

Interpharm/CRC: Boca Raton, 2002.

106. Ali, Z. M.; Gibson, L. J., The structure and mechanics of nanofibrillar cellulose

foams. Soft Matter 2013, 9, 1580.

107. Sehaqui, H.; Zhou, Q.; Berglund, L. A., High-porosity aerogels of high specific

surface area prepared from nanofibrillated cellulose (NFC). Composites Science and

Technology 2011, 71, 1593-1599.

108. Donius, A. E.; Liu, A.; Berglund, L. A.; Wegst, U. G., Superior mechanical

performance of highly porous, anisotropic nanocellulose-montmorillonite aerogels

prepared by freeze casting. J Mech Behav Biomed Mater 2014, 37, 88-99.

109. Xu, C.; Willför, S.; Holmlund, P.; Holmbom, B., Rheological properties of

water-soluble spruce O-acetyl galactoglucomannans. Carbohydrate Polymers 2009, 75,

498-504.

110. Eronen, P.; Österberg, M.; Heikkinen, S.; Tenkanen, M.; Laine, J., Interactions of

structurally different hemicelluloses with nanofibrillar cellulose. Carbohydrate

Polymers 2011, 86, 1281-1290.

111. Hayashi, T.; Ogawa, K.; Mitsuishi, Y., Characterization of the adsorption of XG

to cellulose. Plant Cell Physiol. 1994, 35, 1199-1205.

112. Kochumalayil, J. J.; Bergenstrahle-Wohlert, M.; Utsel, S.; Wagberg, L.; Zhou, Q.;

Berglund, L. A., Bioinspired and highly oriented clay nanocomposites with a

xyloglucan biopolymer matrix: extending the range of mechanical and barrier

properties. Biomacromolecules 2013, 14, 84-91.

113. Whitney, S. E. C.; Brigham, J. E.; Darke, A. H.; Reid, J. S. G.; Gidley, M. J.,

Structural aspects of the interaction of mannan-based polysaccharides with bacterial

cellulose. Cabohydrate Research 1998, 307, 299-309.

114. Evans, D. F.; Wennerström, H., The Colloidal Domain: Where Physics, Chemistry,

Biology, and Technology Meet (Advances in Interfacial Engineering) 2ed.; Wiley-VCH

Verlag GmbH: New York, 1999.

115. Pecora, R., Dynamic light scattering measurement of nanometer particles in

liquids. Journal of Nanoparticle Research 2000, 2, 123-131.

116. Hayashi, T., Xyloglucans in the primary cell wall. Annu. Rev. Plant Physiol.

Plant Mol. Biol. 1989, 40, 139-168.

117. Whitney, S. E. C.; Brigham, J. E.; Darke, A. H.; Reid, J. S. G.; Gidley, M. J., In

vitro assembly of cellulose/XG networks: ultrastructure and molecular. The Plant

REFERENCES 83

Journal 1995, 8, 491-504.

118. Carpita, N. C.; Gibeaut, D. M., Structural models of primary cell walls in

flowering plants: consistency of molecular structure with the physical properties of

the walls during growth. The Plant Journal 1993, 3, 1-30.

119. Iwamoto, S.; Nakagaito, A. N.; Yano, H., Nano-fibrillation of pulp fibers for the

processing of transparent nanocomposites. Applied Physics A 2007, 89, 461-466.

120. Svagan, A. J.; Samir, M. A. S. A.; Berglund, L. A., Biomimetic polysaccharide

nanocomposites of high cellulose content and high toughness. Biomacromolecules 2007,

8, 2556-2563.

121. Kelly, A., Interface effects and the work of fracture of a fibrous composite. Proc.

Roy. Soc. Lond. A. 1970, 319, 95-116.

122. Brumer, H.; Zhou, Q.; Baumann, M. J.; Carlsson, K.; Teeri, T. T., Activation of

crystalline cellulose surfaces through the chemoenzymatic modificatio of xyloglucan.

J Am Chem Soc 2004, 126, 5715-5721.

123. Akerholm, M.; Salmen, L., Interactions between wood polymers studied by

dynamic FT-IR spectroscopy. Polymer 2001, 42, 963-969.

124. Muller, M.; Czihak, C.; Schober, H.; Nishiyama, Y.; Vogl, G., All Disordered

Regions of Native Cellulose Show Common Low-Frequency Dynamics.

Macromolecules 2000, 33, 1834-1840.

125. Ma, X.; Yu, J.; Kennedy, J. F., Studies on the properties of natural fibers-

reinforced thermoplastic starch composites. Carbohydrate Polymers 2005, 62, 19-24.

126. Cunha, A. G.; Zhou, Q.; Larsson, P. T.; Berglund, L. A., Topochemical

acetylation of cellulose nanopaper structures for biocomposites: mechanisms for

reduced water vapour sorption. Cellulose 2014, 21, 2773-2787.

127. Hubbell, C. A.; Ragauskas, A. J., Effect of acid-chlorite delignification on

cellulose degree of polymerization. Bioresour Technol 2010, 101, 7410-7415.

128. Chaker, A.; Alila, S.; Mutjé, P.; Vilar, M. R.; Boufi, S., Key role of the

hemicellulose content and the cell morphology on the nanofibrillation effectiveness of

cellulose pulps. Cellulose 2013, 20, 2863-2875.