Dimensional Stability of Paper

Influence of fibre-fibre joints and fibre wall oxidation

PER LARSSON

Licenciate Thesis

Royal Institute of TechnologyDepartment of Fibre and Polymer Technology

Division of Fibre Technology

Stockholm, Sweden 2008

TRITA-CHE Report 2008-8ISSN 1654-1081ISBN 978-91-7178-862-7

KTH Fiber- och polymerteknologiSE-100 44 Stockholm

SVERIGE

Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan framlägges tilloffentlig granskning för avläggande av teknologie licentiatsexamen i pappers- och massa-teknologi, fredagen den 22 februari 2008 klockan 14.00 i STFI-salen, Drottning Kristinasväg 61, Stockholm.

© Per Larsson, januari 2008

Tryck: Universitetsservice US-AB

iii

Abstract

Paper is a very versatile material. Nevertheless, there are several factors limitingits usefulness, and one of the major issues is that cellulosic and ligno-cellulosic fibreslower their free energy by sorbing water and this water changes the dimensions of thepaper. This phenomenon is usually referred to as a lack of dimensional stability and isoften evident as misregister during multicolour printing or curl, cockle and wavy edgesduring printing, copying and storage or, with a wider definition, also as a shortenedlife-time of boxes during storage due to mechano-sorptive creep.

The work described in this thesis aims to study and quantify the importance of thedifferent mechanisms causing water-induced dimensional changes in a fibre networkand to investigate how to improve the dimensional stability of ligno-cellulosic mate-rials. This has been done both by altering the fiber properties such as the moisturesorptivity and by changing the adhesion and degree of contact within the fibre-fibrejoints. The properties of the fibre-fibre joints have been varied by drying laboratorysheets both under restraint and freely to minimise the generation of built-in stresses.

Bleached kraft fibres were treated using the polyelectrolyte multilayer (PEM) tech-nique to improve the adhesion between the fibres and to increase the molecular contactwithin the joints. In contrast, the degree of contact was impaired by hornifying thefibres before sheet preparation. For sheets allowed to dry freely, the PEM-treatmentincreased the hygroexpansion coefficient, i.e. the dimensional movement normalisedwith respect to the change in moisture content, when subjected to changes in relativehumidity whereas the hornification process resulted in a slightly lowered hygroexpan-sion coefficient. However, when the sheets were dried under restraint, the differentjoint and fibre modifications led to no difference in hygroexpansion coefficient. Thiswas interpreted as being a result of an increase in the total contact zone between thefibres when the sheets were dried under restraint, with a greater extension in the out-of-plain direction of the joint resulting in a transfer of a larger part of the transverseswelling to the in-plane expansion.

The sorptivity of the fibres was changed by oxidising the C2-C3 bond of the 1,4-glucans with periodate. This most likely created covalent cross-links in the fibre wallboth improved the integrity of the fibre wall by locking adjacent fibril lamellae to eachother and also removed possible sites for water sorption onto the cellulose surfaces.Periodate oxidation also led to a decrease in the crystallinity of the cellulose within thefibres, making more cellulose hydroxyl groups available for the adsorption of watermolecules. This means that the oxidation both decreased and increased the interactionbetween the fibre wall and moisture but, on two different structural levels. The cross-links significantly reduced the sorption rate when the papers was subjected to changesin relative humidity, as long as the fibres were not subjected to humidities close tosaturation. The smaller change in moisture content when the relative humidity waschanged between 20 and 85 % RH meant that the dimensional stability of the cross-linked sheets was increased. On the other hand, the hygroexpansion coefficient wasincreased in the case of papers made from fibres with the highest degree of oxidation,i.e. the sheets became more sensitive to absolute changes in moisture content.

iv

Sammanfattning

Papper är ett mycket mångsidigt material. Trots detta finns det ett flertal egenskapersom begränsar papperets användbarhet. Ett av de större problemen med cellulosa- ochlignocellulosafibrer är att de sänker sin fria energi genom att sorbera vatten, och dennasorption förändrar papperets dimensioner. Detta fenomen kallas vanligtvis för bristfäl-lig dimensionsstabilitet och uppträder i form av registerfel vid flerfärgstryck eller somkrullning, buckling och vågiga papperskanter vid utskrift, kopiering och lagring, ellermed en vidare definition som förkortad livslängd hos lådor på grund av mekanosorptivtkryp.

Avsikten med denna avhandling har varit att studera och kvantifiera vilka egenska-per som styr, och hur de påverkar, den vatteninducerade dimensionsförändringen somsker hos ett fibernätverk, samt hur dess dimensionsstabilitet kan förbättras. Detta harstuderats både genom att ändra fiberns fuktsorptionsegenskaper och genom att föränd-ra adhesionen och kontaktgraden mellan fibrerna i fiber-fiberfogarna. Fogegenskapernahar också varierats genom att tillverka laboratorieark torkade under inspänning samtark torkade fritt för att minimera mängden inbyggda spänningar i arket.

Blekt kraftmassa har behandlats med polyelektrolytmultilager (PEM) för att för-bättra adhesionen mellan fibrerna och för att öka kontaktgraden mellan fibrerna i fogen.Kontaktgraden har även minskats genom förhorning av fibrerna före arkformning. Förde ark som fick torka fritt gav PEM-behandlingen en ökad hygroexpansionskoefficient,det vill säga dimensionsförändringen normaliserad mot förändringen i fuktinnehåll, vidsamma förändring i relativ luftfuktighet medan förhorningen minskade hygroexpan-sionskoefficienten något. Om arken emellertid torkades under inspänning observeradesingen skillnad i hygroexpansionskoefficient mellan de olika fibermodifieringarna. Det-ta tolkades som ett resultat av en ökad kontaktzon och en större utbredning ut ur fogensplan, när arken torkades utan inspänning. En utbredning som medför att en större delav fiberns transversella expansion överförs som expansion i pappersplanet.

Fibrernas fuktsorptionsegenskaper förändrades genom natriumperjodatoxideringav 1,4-glukanernas C2-C3-bindning. Detta skapade sannolikt tvärbindningar i fiber-väggen som förbättrade fiberväggens tålighet både genom att låsa fibrillerna närmarevarandra och genom att ta bort potentiella adsorptionssäten som annars är tillgängligaför vattenadsorption. Perjodatoxidationen minskar också fibrernas kristallinitet och så-ledes frigjorde oxidationen hydroxylgrupper där vattenmolekyler kan adsorbera. Dettainnebar att oxidationen både minskade och ökade interaktionen mellan vatten och fiber-vägg, men dock på olika strukturell nivå. Tvärbindningarna visade sig också märkbartreducera sorptionshastigheten när arken utsattes för en förändrad luftfuktighet så längede inte tidigare utsatts för relativa luftfuktigheter nära mättnad. Som ett resultat av denlägre förändringen i fuktinnehåll vid en förändring i luftfuktighet från 20 till 85 % RFminskade dimensionsförändringens amplitud för de tvärbundna arken upp till 30 %.Emellertid uppvisade de tvärbundna arken en högre hygroexpansionskoefficeint, vilketinnebär att de blev mer känsliga för absoluta förändringar i fuktinnehåll.

List of Publications

This licentiate thesis is based on the following papers:

Paper I

Influence of Fibre-Fibre Joint Properties on the Dimensional Stabilityof Paper

Per A Larsson and Lars Wågberg, Manuscript accepted for publication in Cellulose

Paper II

The Influence of Periodate Oxidation on the Moisture Sorptivity and theDimensional Stability of Paper

Per A Larsson, Magnus Gimåker and Lars Wågberg, Manuscript submitted for pub-lication

v

Table of Contents

Introduction 1

Background 3Water uptake and swelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Sorption hysteresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4Dimensional stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5The fibre-fibre joint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Hornification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Polyelectrolyte multilayers . . . . . . . . . . . . . . . . . . . . . . . . . 9

Cross-linking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Experimental 11Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Fibres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11Chemicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11Hornification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11Polyelectrolyte multilayers . . . . . . . . . . . . . . . . . . . . . . . . . 12Fibre cross-linking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Carbonyl content determination . . . . . . . . . . . . . . . . . . 12Sheet preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Sheets dried under restraint . . . . . . . . . . . . . . . . . . . . . 13Freely dried sheets . . . . . . . . . . . . . . . . . . . . . . . . . 13Shrinkage measurements . . . . . . . . . . . . . . . . . . . . . . 14

Sheet testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14Mechanical testing . . . . . . . . . . . . . . . . . . . . . . . . . 14Dynamic mechanical analysis . . . . . . . . . . . . . . . . . . . 15Dimensional stability measurements . . . . . . . . . . . . . . . . 15Dynamic vapour sorption . . . . . . . . . . . . . . . . . . . . . . 15

vi

Table of Contents vii

Results and Discussion 17Joint influence on the dimensional stability . . . . . . . . . . . . . . . . . . . . 17Moisture sorptivity and fibre wall cross-linking . . . . . . . . . . . . . . . . . 22

Conclusions 29Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Acknowledgements 31

Bibliography 33

List of Figures

1 Schematic representation of a “bottleneck pore” . . . . . . . . . . . . . . . . 5

2 Schematic representation on how the periodate ion oxidises the C2-C3-bondof the cellulose into dialdehyde cellulose . . . . . . . . . . . . . . . . . . . . 13

3 Drying-frame equipped with two stretched PTFE wires to give as little restraintas possible during drying . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

4 Mechanical properties of restraint-dried and freely dried sheets. . . . . . . . . 175 Shrinkage and moisture content as a function of time for a freely dried sheet

and the shrinkage and density of freely dried and restraint-dried sheets. . . . . 186 Equilibrium hygroexpansional strain and equilibrium moisture content as a

function of relative humidity for restraint-dried and freely dried sheets. . . . . 197 Equilibrium hygroexpansional strain as a function of equilibrium moisture

content for restraint-dried and freely dried sheets. . . . . . . . . . . . . . . . 208 Kinetics of hygroexpansional strain and moisture content changes in restraint-

dried sheets and freely dried sheets. . . . . . . . . . . . . . . . . . . . . . . 209 Schematic representation of the fibre-fibre joint configuration depending on

drying strategy and whether or not PEMs are adsorbed onto the fibre surface. 2110 Hygroexpansion as a function of moisture content in restraint-dried sheets and

freely dried sheets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2211 Sorption isotherms for sheets from virgin, hornified and PEM-treated fibres . 2312 Sorption isotherms for sheets of untreated and periodate-oxidised fibres . . . 2313 Dry and wet tensile strength index and dry and wet strain at break as function

of carbonyl content in the periodate-oxidised fibres . . . . . . . . . . . . . . 2414 Sorption rate of oxidised fibres when subjected to a change in relative humidity

from 30 to 90 % RH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2415 Schematic rupture of a cross-linked fibre wall at low and high relative humidity 2516 Dynamic dimensional change when papers from oxidised fibres were subjected

to a change in relative humidity from 20 to 85 % RH . . . . . . . . . . . . . 2617 Equilibrium hygroexpansional strain and equilibrium moisture content as a

function of relative humidity and equilibrium hygroexpansional strain as afunction of equilibrium moisture content. . . . . . . . . . . . . . . . . . . . 27

viii

List of Tables

I Hygroexpansion coefficients and drying shrinkages for seven papers made ofdifferent pulps and fibre treatments. . . . . . . . . . . . . . . . . . . . . . . 7

II Amplitude of dimensional change, change in moisture content and hygroex-pansion coefficient after all built-in stresses were released – Restraint-driedand freely dried sheets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

III Amplitude of dimensional change, change in moisture content and hygroex-pansion coefficient after all built-in stresses are released – Oxidised fibres. . . 28

ix

Introduction

Dimensional stability is the ability of paper to withstand outer and inner perturbations suchas changes in temperature or moisture content without changing its dimensions. In print-ing, copying and converting operations, dimensional stability is often a critical parameter.Insufficient dimensional stability may cause misregister during multicolour printing, gen-erate curl, cockle and wavy edges during printing or copying, and in the case of packagingpapers, it may also lead to substantial problems in packaging machines and during storage.

During a four-colour offset printing operation 0.2− 0.3 g/m2 water i transferred tothe paper per colour (Thalén et al. 1994). For a paper with a grammage of 40 g/m2, thiscorresponds to a total increase in moisture content of 2− 3 percentage units, which formost pulp types corresponds to an expansion of about 1.5 mm/m1 between the first andthe last colour, i.e. an expansion clearly visible to the human eye. In addition, the increasein paper moisture content reduces the strength and stiffness of the paper, which increasesthe risk of mechanical failure of the paper web in the printing operation.

Although the effects of fibre swelling on fibre dimensions during papermaking havebeen extensively studied (e.g. Stone and Scallan 1967, Lindström and Carlsson 1978,Grignon and Scallan 1980, Laine et al. 2003), there is still only a limited understandingof the fundamental mechanism(s) controlling the dimensional changes of both individualfibres and entire fibre networks.

It is well known that the dimensional stability can be dramatically increased by cross-linking the fibre wall (Cohen et al. 1959). It is also a common belief that a smaller degree ofcontact in the fibre-fibre joint, decreased for example by hornification, results in a greaterdegree of dimensional stability of paper (Uesaka and Qi 1994) and the objective of thisthesis is to further explore how the fibre-fibre joint influences the dimensional stability andin what way fibre wall cross-linking increase the dimensional stability of paper.

1Calculation based on hygroexpansion data from Nanri and Uesaka (1993).

1

Background

Water uptake and swelling

When fibres absorb liquid water, hydrogen bonds within and between the fibres are brokenand the fibrils are separated from each other in the fibre wall, resulting in a swelling of thefibre wall, and this dimensional change is transferred throughout the whole fibre network.

When a single fibre shrinks it decreases its dimensions by only a couple of per centin length, while the transverse direction of the fibre can shrink between 10 and 30 percent (Tydeman et al. 1966). A major part of this shrinkage is recovered when the fibre isre-wetted or subjected to a high relative humidity. However, if inter-fibre interactions areallowed to force the fibres into greater shrinkage Page and Tydeman (1962) showed thatthis shrinkage can be about 12 per cent, which is equal to the mean shrinkage of the entirepaper.

From the theories considering gel swelling derived by Donnan and Proctor at the begin-ning of the 20th century, Grignon and Scallan (1980) developed a model for fibre swellingin solutions. The mechanisms proposed by Donnan and Proctor can be realised for fibresby virtually positioning a membrane just above the fibre surface. The acidic groups boundto the fibre surface cannot move through this virtual membrane, but water molecules andsimple ions can. This difference in mobility creates an osmotic pressure across the mem-brane. Based on this model, and by assuming that concentration ratios between ions outsideand inside the “gel-phase” can be used instead of activities, Grignon and Scallan (1980)showed that it was possible to estimate the swelling of fibres as a function of the concen-tration of simple salts and the dissociation constants of the acidic groups inside the fibrewall.

It has also been shown that a highly charged fibre swells more than a fibre with a lowercharge (e.g. Grignon and Scallan 1980). This was found to be true for bulk-charged fibres,and Laine et al. (2003) and Torgnysdotter and Wågberg (2003, 2004) have shown that anincrease in surface charge has no significant influence on the swelling properties of thefibre wall.

Paper is a hygroscopic material, i.e. the fibres spontaneously lower their free energy by

3

4 BACKGROUND

adsorbing water (cf. Hollenbeck et al. 1978), and a dry paper will readily adsorb moisturein a moisture-containing atmosphere and thus cause an expansion of the fibre wall since thedry fibre wall has been shown to contain no pores at all (Stone and Scallan 1967). This isusually referred to as hygroexpansion of the fibres. The water uptake from the surroundingair is much slower than that when the fibres are subjected to liquid water. It can take severalhours for paper to reach equilibrium moisture content, depending on the paper quality andon the amplitude of the relative humidity change (Jarrell 1927). Jarrell also found thatdesorption is more time-consuming than adsorption. Leisen et al. (2002) concluded thatmoisture adsorption is the rate-determining factor for the moisture uptake since the rate ofinter-fibre diffusion is much faster than the rate of the adsorption process. The coupling,i.e. the exchange of heat and moisture, between the paper and the surrounding air, also hasa significant influence on the rate of adsorption/desorption, and thus also on the rate ofdimensional change (Brecht and Hildebrand 1960).

There are several ways in which water can be absorbed into a fibrous material; eitheras free water, i.e. as inter-fibre water in the pores between the fibres or intra-fibre water inthe lumen of the fibre, or as bound water. And, using the definitions of Yoshida (1992),there are two types of bound water; freezing bound water and non-freezing bound water,i.e. water that shows no phase transition down to a temperature of 130 K. Berthold et al.(1994) argued that below complete saturation, i.e. at relative humidities below 100 %, thepresence of charged groups is a prerequisite for non-freezing bound water to occur.

Sorption hysteresis

Sorption of moisture from the air presents a significant hysteresis, i.e. the equilibriummoisture content is not a unique function of the relative humidity of the environment.There are several explanations of why paper presents a hysteresis effect.

Urquhart (1960) proposed that the hysteresis is due to a difference in carboxyl groupavailability during adsorption and desorption, while Barkas (1949) explained the hysteresiswith as being due to a plastic deformation of the cellulose gel.

An effect that contributes not only to the hysteresis effect but also to the rate of thehumidity-induced dimensional change is the heat/mass coupling, i.e. the exchange of heatand moisture, between the paper and the surrounding air (Brecht and Hildebrand 1960).Since adsorption is an exothermic process, the temperature at the fibre surface will increaseduring adsorption and thus lower the relative humidity and reduce the driving force forfurther adsorption unless the heat is transferred away from the surface. During desorptionthe effect will be the opposite.

Another contribution to the hysteresis of cellulosic and lignocellulosic materials is the“bottleneck” effect (schematically shown in Figure 1), which occurs in non-uniform cap-illaries with narrower and wider sections. According to the Kelvin equation (1), narrower

DIMENSIONAL STABILITY 5

water menisci

Figure 1: Schematic representation of a “bottleneck pore” during sorption where the water meniscushinders the moisture from diffusing further into, or out from, the pore.

capillaries will be filled with water at a lower relative humidity during adsorption andthus hinder the diffusion to, and the adsorption within, the wider sections of the capil-lary.During desorption, the narrower sections will prevent desorption away from the widersections which wish to desorb at higher vapour pressures and this will contribute to thehysteresis effect.

ln pp0

= −2γMrρRT

(1)

where

pp0

= The relative vapour pressureγ = Surface tension of waterM = Molecular mass of waterr = Radius of capillaryρ = Density of water

Everett and Whitton (1952) developed a theory of independent domains describing thesorption hysteresis in a general material. Later both Peralta (1995) and Chatterjee et al.(1997) applied this theory with good results to their experimental sorption isotherm datafor different wood and paper sources to describe the sorption behavior of the material.

Dimensional stability

The in-plane dimensional stability of paper is often quantified as the hygroexpansive strainof the paper upon a change in relative humidity:

εh =∆ll

(2)

a figure often normalised with respect ro the change in moisture content (∆MC) or relativehumidity (∆RH) as a hygroexpansion coefficient:

βMC =εh

∆MC(3)

6 BACKGROUND

The dimensional stability of paper is a history-dependent property due to the release ofbuilt-in stress induced during the drying process when the paper is re-wetted or subjectedto a change in relative humidity, but also due to the fact that paper exhibits a sorptionhysteresis (cf. Larocque 1936, Uesaka et al. 1989).

Beating of fibres increases the hygroexpansion in general and the hydroswelling in par-ticular, as well as the shrinkage of freely dried sheets (Brecht et al. 1956, Brecht 1957).Brecht (1957) also concluded that beating had a greater influence on the dimensionalchange than on the increase in moisture content, i.e. an increase in the βMC-value due tothe beating process. Nordman (1958) showed that there was a linear relationship betweenshrinkage during drying and hygroexpansion, while Salmén et al. (1987) later concludedthat this is true only when the sheets are wet-pressed to the same density. On the otherhand, at zero shrinkage, i.e. when dried under restraint, there is no or little correlation be-tween density and in-plane hygroexpansion (Salmén et al. 1987, Uesaka 1991, Lyne et al.1994).

Nanri and Uesaka (1993) measured the βMC-value of sheets from several differentpulps and at different shrinkages (Table I). They came to the conclusion that even thoughTMP and SGW pulps presented the lowest drying shrinkages, they did not present thelowest βMC-values and that the lower shrinkage was a result not of a lower hygroex-pansion coefficient but rather of a smaller change in moisture content from the onsetof shrinkage till dryness. Nanri and Uesaka (1993) also found a linear relationship be-tween the difference in βMC-values of freely dried sheets and sheets dried under restraint(βMC,free − βMC,restraint) and sheet shrinkage, where the mechanical pulps showed thelowest difference between the two drying strategies, which implies that drying under re-straint has a smaller effect on the βMC-values of mechanical pulps.

The out-of-plane hygroexpansion decreases with both increasing density and increas-ing shrinkage, i.e. when the shrinkage during drying is increased the in-plane hygroexpan-sion is increased at the same time as the out-of-plane expansion is decreased (Lyne et al.1994). The authors also concluded that the volume of the voids in the paper increased lessthan the paper as a whole when the paper was subjected to an increased relative humidity,i.e. that paper swells less than the individual fibres.

The hygroexpansion of paper is not only dependent on the drying mode, but dependson the degree of anisotropy, where the expansion of a anisotropic sheet compared withan isotropic sheet, is lower when more fibres are aligned in the direction of the measuredexpansion and greater in the cross-direction (Brecht and Hildebrand 1960). There is alsoa complex relation to the fibre morphology. Uesaka and Moss (1997) showed that theβMC-value of sheets dried under restraint decreased only slightly with decreasing fibrelength when the fibres were mechanically cut to different fibre lengths. When the pulp wasfractionated, however, it was found that the βMC-value was higher in the shorter lengthfractions. This was especially true for CTMP-fibres, where the βMC-value was signifi-

THE FIBRE-FIBRE JOINT 7

cantly higher in sheets made from fractions with an average fibre length of ca. 1.2 mm andshorter. Uesaka and Moss (1997) also found that the shorter fibre fractions had a higherextinction angle, suggesting that the cell-wall structure, e.g. the microfibril angle, affectsthe fibre hygroexpansion.

Table I: Hygroexpansion coefficients (βMC -value) and drying shrinkages for seven papers made ofdifferent pulps and fibre treatments (Nanri and Uesaka 1993).

Mechanical pulps Chemical pulpsSGW1 TMP2 CTMP3 CMP4 LYS5 LYS BKP6

(beaten) (beaten)

β-value (%/%)Restraint-dried 0.068 0.057 0.063 0.081 0.048 0.063 0.053Freely dried 0.086 0.074 0.087 0.125 0.075 0.123 0.092

Sheet shrinkage (%) 1.50 1.40 1.99 3.06 2.35 4.60 3.231Stone Grind Wood2Thermo Mechanical3Chemical Thermo Mechanical Pulp4Chemical Mechanical Pulp5Low Yield Sulphite6Bleached Kraft Pulp

The fibre-fibre joint

The specific forces holding the fibres together in a paper are still unknown. However, it isa well-known fact that the fibre-fibre joints (often referred to as fibre-fibre bonds) play amost important role. Since individual fibres can be extracted from a sheet without breaking,Davison (1972) concluded that the fibre-fibre joints are the weak links contributing to paperstrength. Davison also showed that the strength of a paper is significantly less than thestrength of the individual fibres. It was also shown theoretically by Page (1969) that thepaper strength increases asymptotically towards a strength limited by the strength of theindividual fibres with increasing joint strength and increasing fibre-to-fibre contact area.

According to Eklund (1969), the hygroexpansivity is reduced, i.e. the dimensional sta-bility is improved, after a couple of drying cycles or re-use of the pulp as the fibre ages andbecomes hornified. It is a common belief (Uesaka and Qi 1994) that the lower the degreeof fibre-to-fibre contact, the greater is the dimensional stability.

Torgnysdotter and Wågberg (2006) showed that the molecular contact between the twofibres in a fibre-fibre joint was significantly reduced when the fibres were dried, i.e. beinghornified, and that the degree of contact can be improved again by applying polyelectrolytemultilayers (PEMs) onto the fibres. In the case of cellulose fibres, the build-up of multi-layers consisting of polyallylamine and polyacrylic acid has been shown to significantly

8 BACKGROUND

increase both tensile strength and strain at break, due both to a higher degree of molecularcontact between the fibres and a better adhesion in the contact zone (Wågberg et al. 2002,Eriksson et al. 2005, Torgnysdotter and Wågberg 2006).

Hornification

Even though the term is commonly used in the pulp and paper industry, hornification is asomewhat arbitrary concept. It is used to describe all physical and chemical changes thatreduce the swelling, and thus the ability of the fibres to hold water, and the strength ofpapers made from these fibres. In 1944, Jayme quantified hornification as the loss in waterretention value (WRV), i.e.

Hornification = WRVref −WRVhWRVref

(4)

where WRVref is the initial water retention value and WRVh the value of the “hornified”sample. The hornification process has been found to occur below a certain critical watercontent, the fibre saturation point (Laivins and Scallan 1993).

Hornification is a complex process, probably consisting of several mechanisms andseveral theories have been proposed to explain the hornification process. Besides definingthe concept of hornification as in Equation 4, Jayme (1943) explained the phenomenonas an aggregation and tighter structure of the polysaccharide chains as water is removed.Lundberg and de Ruvo (1978) concluded that, since the equilibrium moisture content is notaffected by hornification, no drastic changes occur and that the most plausible explanationis closure of large pores, a result also found by Stone and Scallan (1966, 1968). It has alsobeen shown that the degree of hornification is increased with increasing degree of hydrogenbonding (Lindström and Carlsson 1982), which presumably induces a tighter structure ofthe cellulose fibrils in the fibre wall. Lindström and Carlsson (1982) also showed that ifthe fibres, on the other hand, were dried in their sodium form the hornification process wasmore or less stopped, provided the charge of the fibres was sufficiently high. This has lateralso been shown by Laivins and Scallan (1993).

It has also been proposed that hornification is to be due to both a generation of covalentcross-links and “irreversible hydrogen bonding”, i.e. the formation of hydrogen bonds thatdo not break when the fibres are re-wetted. Lactone formation has also been suggestedas a mechanism to explain hornification (cf. Samuelson and Törnell 1961, Lindström andCarlsson 1982, Fernandes Diniz et al. 2004) but, since the acid content of the pulps doesnot change during the hornification (Scallan and Tigerström 1992), the idea of lactonecross-links is questioned.

Yet another proposed mechanism is that, with the loss of water, the fibre crystallinityis increased, i.e. the cellulose chains are more ordered with a larger amount of hydrogen

CROSS-LINKING 9

bonds between the chains (cf. Weise 1998, Newman 2004).Since hornification is closely linked to the recyclability of pulp fibres, the literature

on hornification is very extensive and for more extensive reviews the reader is referred toe.g. Fernandes Diniz et al. (2004), Kato and Cameron (1999), Laivins and Scallan (1993),Nazhad and Paszner (1994) and Weise (1998).

Polyelectrolyte multilayers

The adsorption of polyelectrolytes onto a surface depends both on the charge of the surfaceand the charge of the polyelectrolyte, but it also depends on the conformation of the poly-electrolyte, which in turn depends on internal electrostatic repulsions within the polymerchain. This internal repulsion can be screened by adding salt or, in the case of weak poly-electrolytes, changed by changing the pH-value to achieve another degree of dissociationand thus a change in the charge density (Dautzenberg et al. 1994).

The concept of polyelectrolyte multilayers (PEMs) involves consecutively treating acharged surface with cationic and anionic polyelectrolytes, an idea first developed byDecher (1992). A polyelectrolyte is adsorbed onto a surface and thus recharges the surface,thereafter the surface is rinsed with ionised water to wash away any potentially unadsorbedpolyelectrolyte, followed by the adsorption of an oppositely charged polyelectrolyte andyet another rinsing step. This procedure is then repeated until the desired number of layershas been adsorbed onto the surface. Wågberg et al. (2002) used this technique with successon wood fibres to increase the sheet strength properties.

By using not only different polyelectrolytes but even nanoparticles, a great varietyof properties, such as increased strength, electric conductance and colour changes uponcertain stimuli, can be imparted to the substrate (Iler and Colloid 1966, Agarwal et al.2006, Zheng et al. 2006, Wistrand et al. 2007, Wågberg et al. 2008). By using weakpolyelectrolytes and varying the pH, i.e. changing the charge density of the polyelectrolyte,amount of the polyelectrolyte adsorbed, and hence the properties of the multilayer, can becontrolled (cf. Eriksson et al. 2005).

Cross-linking

It has long been known that the dimensional stability can be increased by cross-linking thefibre wall. Cohen et al. (1959) used formaldehyde to induce cross-links into the fibre walland Stamm (1959) showed that cross-links can be induced via a catalysed heat treatment.LeBel et al. (1968) soaked paper with polyfunctional chemicals, such as polyepoxidesand dialdehydes, with the ability to react with the hydroxyl groups of the cellulose andto form cross-links both within the fibre-wall and between the individual fibres, and thusincrease the dimensional stability of the paper, probably both by preventing the fibres from

10 BACKGROUND

swelling and by eliminating adsorption sites available to water molecules. Weatherwaxand Caulfield (1978) later showed that fibres cross-linked with formaldehyde containedsmaller pores, and probably an increased “bottleneck pore” formation (see Figure 1), anda lower moisture uptake.

By oxidising the C2-C3-bond of 1,4-glucans with periodate and consequently formingtwo reactive aldehyde groups, which can react with hydroxyl groups within the pulp fibreand form hemiacetal linkages (Ghosh and Dalal 1988, Zeronian et al. 1964), the fibre wallcan be cross-linked. Kim et al. (2000) showed that the oxidation also reduces the cellulosecrystallinity and that the oxidation is unevenly distributed over the fibre surface and createsmore flexible fibres.

Hou et al. (2007) used periodate to oxidise bleached kraft fibres to dialdehyde cellulose(DAC) and then formed sheets with many times greater dry and wet tensile strengths. Houet al. also sulfonated the DAC by reacting it with bisulfite. This resulted in even betterstrength properties up to oxidant dosages of 50 wt%. It was suggested that the increasein tensile strength of the sheets in this case was due to both changes in the fibre surfacemorphology and to an increased swelling of the fibres, since sulfonate groups are morehydrophilic than carboxyl groups.

Experimental

Materials

Fibres

Unbeaten, fully bleached, virgin softwood kraft pulp (SCA, Östrand Mill, Sweden) wasused throughout this work. Fines were removed by spray screening through a wire with amesh size of 75 µm in an equipment developed at STFI-Packforsk. The pulp was washedand the carboxyl groups of the fibres were converted to their sodium form (Wågberg andHägglund 2001).

Chemicals

Polyallylamine hydrochloride (PAH) with a molecular weight of 15, 000 Da and poly-acrylic acid (PAA) with a molecular weight of 7, 000 Da were used to build multilayers onthe fibres. The PAH was received as powder and dissolved in deionised water prior to use,and the PAA was delivered as a 50 % aqueous solution and diluted with deionised water tothe desired concentration before use.

Sodium metaperiodate for the oxidations and hydoxylamine hydrochloride for carbonyldetermination were delivered in solid form and both were used as received.

The hydrochloric acid, sodium hydroxide, sodium chloride and sodium bicarbonatewere all of analytical grade.

Methods

Hornification

One part of the pulp was hornified according to the following procedure. An aqueoussuspension of fibres was dewatered leaving a small amount of water at the surface of thefilter cake to minimise the effect of the capillary forces during pulp dewatering. Thereafter,the filter cake was left to dry at 30 ◦C in a fan dryer in order both to minimise cleavage

11

12 EXPERIMENTAL

of the cellulose chains during the subsequent temperature treatment (Kato and Cameron1999) and to minimise fibre-fibre joint formation during drying. After drying, the fibreswere cured for 24 hours at 105 ◦C.

Polyelectrolyte multilayers

The fibres were treated consecutively with PAH and PAA (Wågberg et al. 2002). Theadsorptions were performed in a 4 g/l fibre suspension with a background electrolyte con-centration of 0.01 M NaCl. Amounts of 30 mg/g fibre of the two polyelectrolytes wereadded separately and were allowed to adsorb for 20 minutes. Five layers of polyelec-trolytes were adsorbed prior to sheet forming. Between each adsorption step, the fibreswere thoroughly rinsed with deionised water. PAH and PAA were adsorbed at pH valuesof 7.5 and 3.5, respectively (Eriksson et al. 2005). After formation, the PAH/PAA-treatedsheets were heat-treated for 30 minutes at 160 ◦C to induce cross-links in the multilayerstructure, and thus to increase the tensile strength of the sheets (Eriksson et al. 2006)

Fibre cross-linking

Fibres were oxidised in a 6.3 g/l fibre suspension using sodium periodate according tothree oxidations strategies based on the theoretical amount of oxidisable C2-C3-bonds inpure cellulose and on time;

• the theoretical amount of NaIO4 (1.36 g/g) to oxidise all C2-C3-bonds for onehour (denoted sample PI),

• twice the theoretical amount of NaIO4 (2.72 g/g) for two hours (denoted samplePII), and

• four times the theoretical amount ofNaIO4 (5.43 g/g) for four hours (denoted sam-ple PIII).

The periodate oxidation mechanism and a proposed cross-linking mechanism are shownin Figure 2. The suspension was stirred all the time and kept dark in the to minimise sidereactions such as cellulose chain scission (Symons 1955). The oxidations were stopped bywashing the fibres thoroughly with deionised water, followed by immediate sheet forma-tion.

Carbonyl content determination

To assess the degree of oxidation, the amount of aldehyde formed was determined byaddition of hydroxylamine hydrochloride, which reacts quantitatively with all available

METHODS 13

O

OH

O

OHOH

O

IO4-

O

OH

O

OO

O

O

OH

O

OHOH

O

O

O

O

OHOH

O

O

OH

O

OOH

OProton Transfer

Figure 2: Schematic representation on how the periodate ion oxidises the C2-C3-bond of the celluloseinto dialdehyde cellulose (DAC) followed by a possible mechanism for the cross-linking reaction.

carbonyls while releasing protons. To 25 ml of a 0.25 M solution of hydroxylamine hy-drochloride, adjusted to pH 4, about 0.1 g of fibres were added and stirred for two hours.After this time, the fibres were filter off and dried, and the filtrate was titrated back to pH4 using sodium hydroxide. From the amount of sodium hydroxide required, the degree ofoxidation was calculated. A determination method similar to that used by e.g. Zhao andHeindel (1991).

Sheet preparation

Sheets dried under restraint

Sheets with an average grammage of 110 g/m2 were prepared using tap water in a “RapidKöthen” sheet former (Paper Testing Instruments, Pettenbach, Austria). The sheets weredried at 93 ◦C under a negative pressure of 95 kPa for 10− 15 minutes. The PAH/PAA-treated sheets were further heat-treated for 30 minutes at 160 ◦C to induce cross-linksin the multilayer structure (Eriksson et al. 2006). Since hygroexpansion is a history-dependent behaviour (Uesaka et al. 1989), the sheets were stored at 23 ◦C and 50 % RHuntil tested.

Freely dried sheets

Sheets with an average grammage (after shrinkage) of about 110 g/m2 were formed in the“Rapid Köthen” sheet former using tap water, but the sheets were dried for only 1 minuteat 93 ◦C under a negative pressure of 95 kPa. Instead of conventional blotting papers,two 105 µm mesh PTFE wires were used to ease the transfer to the specially designeddrying-frame shown in Figure 3. After one minute of pressurised drying, the sheet wastransferred to the drying-frame and placed between two stretched PTFE wires separatedby four 0.92 mm thick spacers to minimise the risk of curling and buckling. The sheetwas dried to equilibrium dryness at 23 ◦C and 50 % RH and kept there during storage.

14 EXPERIMENTAL

Figure 3: Drying-frame equipped with two stretched PTFE wires to give as little restraint as possibleduring drying. The sheet between the wires has a diameter of 20 cm.

The PAH/PAA-treated sheets were further heat-treated for 30 minutes at 160 ◦C to inducecross-links in the multilayer structure and possibly between the PAH and the fibres (Eriks-son et al. 2006) before storage.

Shrinkage measurements

Before the sheet was transferred to the drying frame, two pairs of holes were made per-pendicular to each other (140 mm apart within each pair) with a marking tool. After equi-libration at 23 ◦C and 50 % RH, the sheet was flattened by a PMMA-disc. The disc wasequipped with marks similar to those made by the marking tool and this set was used forcalibration. A digital camera was used in combination with image recognition in Matlabto measure the shrinkage of the sheet. The error of a single measurement was estimated tobe about 0.10 % over a width of about 140 mm, i.e. a value less than the natural scatter inshrinkage among freely died sheets.

Sheet testing

Mechanical testing

After conditioning at 23 ◦C and 50 % RH, sheet grammage was measured and both dryand one hour wet tensile testing was performed according the SCAN-test standards SCAN-P 67:93 and SCAN-P 20:95, respectively (only dry strength in Paper I). Sheet thicknessand density were evaluated by measuring the thickness according to a method developedat STFI-Packforsk (Schultz-Eklund et al. 1992).

METHODS 15

Dynamic mechanical analysis

The storage modulus was measured at ∼ 30 ◦C using a Perkin-Elmer Instrument Dynamicmechanical analyser (DMA) combined with a moisture generator. The samples were firstsubjected to 30 % RH for 90 minutes and the humidity was then increased stepwise to90 % at a rate of 0.5 units/minute.

Dimensional stability measurements

To study the dimensional movement and dimensional stability of the sheets, a dimensional-stability-meter developed at STFI-Packforsk was used. 15 mm wide paper strips weremounted horizontally between two clamps, where one of the clamps was free to move.This movement was monitored, in Paper I using a LVDT-sensor and in Paper II with alaser measuring device. The two measuring methods gave, for unknown reasons, differentresults when measuring the same sheets. However, this does not affect the interpretationwithin a given test series.

To adjust the relative humidity, a moisture generator mixing saturated and dry airstreams was connected to the dimensional-stability-meter-box. The moisture content ofthe strips inside the humidity-controlled box was estimated gravimetrically by placing testpieces of the papers to be tested within the chamber.

To ensure that equilibrium had been reached, the strips were equilibrated for at leastseven hours.

The hygroexpansion coefficient is here defined as the relative change in length dividedby the change in equilibrium moisture content ( ∆l

l · ∆MC ) after all built-in stresses had beenreleased, i.e. after several moisture cycles when the expansion reached reversibility.

Dynamic vapour sorption

A DVS, dynamic vapour sorption, equipment (Surface Measurement Systems Ltd.) wasused to obtain near equilibrium sorption isotherms at a temperature of∼ 33 ◦C. To achievethe desired relative humidity, dry and saturated air currents were mixed in the appropriateratio. A microbalance continuously measured the weight of the sample. The weight at0 % RH was used to calculate the moisture content of the sample.

Results and Discussion

Joint influence on the dimensional stability

For sheets dried both under restraint and with free shrinkage, the application of PEMsdramatically enhanced the mechanical properties, while the use of heavily hornified fibresimpaired the mechanical properties (Figure 4). This is in agreement with previous resultsthat have reported a significant difference in strength properties of the paper (Wågberg et al.2002, Eriksson et al. 2005) and hence in the fibre-fibre joint contact area (Torgnysdotterand Wågberg 2006) when PEMs are applied to wood fibres.

The shrinking process was fast as soon as a the moisture content dropped below a crit-ical level, in the case of the never-dried pulp a dry content below about 50 % (Figure 5a).The never-dried pulps shrank to a greater extent than the hornified ones when dried freelyand, when dried under restraint, the never-dried pulps became more consolidated, i.e. theydisplayed a higher density (Figure 5b). The PEM-treatment also had a slight effect on thesheet density, which was found to be somewhat higher for the PEM-treated sheets.

When the dimensional stability of the sheets was measured, by exposing them to hu-

0

50

100

150

200

250

300

Tensile strengthindex (kNm/kg)

Strain at break (%)

Tensile stiffnessindex (MNm/kg)

Rel

ativ

e di

ffere

nce

(%)

VirginHornifiedVirgin + PEMHornified + PEM

27.0 ± 0.58

2.35 ± 0.16

4.61 ± 0.08

11.7 ± 0.23

59.5 ± 1.88

30.9 ± 0.62 0.95 ± 0.03

4.65 ± 0.22 3.38 ± 0.16

2.37 ± 0.05

5.47 ± 0.11

3.73 ± 0.05

(a)

0

50

100

150

200

250

300

Tensile strengthindex (kNm/kg)

Strain at break (%)

Tensile stiffnessindex (MNm/kg)

Rel

ativ

e di

ffere

nce

(%)

VirginHornifiedVirgin + PEMHornified + PEM

8.78 ± 0.29

5.42 ± 0.55

1.02 ± 0.04

3.70 ± 0.29

2.05 ± 0.19

0.53 ± 0.06

21.91 ±

1.27

8.64 ± 0.43

1.51 ± 0.10

8.96 ± 0.79

3.98 ± 0.57

0.95 ± 0.08

(b)

Figure 4: Mechanical properties of (a) restraint-dried and (b) freely dried sheets. All differences arenormalised with respect to the virgin pulp and the bars indicate 95 % confidence limits. The absolutevalues for the different sheet properties are given above the respective bars.

17

18 RESULTS AND DISCUSSION

93

94

95

96

97

98

99

100

0 40 80 120 160 200 240 280 320 360 400 440 480Time (min)

She

et s

ize

(%)

0

10

20

30

40

50

60

70

80

90

Moi

stur

e co

nten

t (%

)

Sheet sizeMoisture content

(a)

-1

0

1

2

3

4

5

6

200 250 300 350 400 450 500 550 600 650 700Density (kg/m3)

Shr

inka

ge (%

)

VirginHornifiedVirgin +PEMHornified + PEM

Freely dried

Dried under restraint

(b)

Figure 5: (a) Shrinkage and moisture content as a function of time for a freely dried sheet made fromvirgin pulp when dried at 50 % RH and (b) the shrinkage and density of freely dried and restraint-dried sheets. The bars indicate 95 % confidence limits.

midity cycles between 20 and 85 % RH, no major difference in amplitude of either thedimensional changes or the moisture adsorbance was seen for the sheets dried under re-straint. However, in the studied humidity interval, the restraint-dried PEM-treated fibresseemed to adsorb somewhat more moisture (Figure 6a). The amplitude of the dimensionalchange was, as shown in Table II, somewhat lower for the hornified pulp than for boththe virgin and the multilayer-treated pulps when dried freely. These results agree with thelower shrinkage of the hornified pulps shown in Figure 5b (cf. Nordman 1958, Salménet al. 1987, Nanri and Uesaka 1993).

The restraint-dried sheets showed a significant permanent shrinkage after being sub-jected to the relative humidity cycles. This shrinkage was not as pronounced for the freelydried sheets, i.e. when the built-in stress was smaller. Due to the higher moisture con-tent after cycling, the multilayer-treated sheets presented a somewhat smaller permanentshrinkage (Figure 6).

Figure 7, where the values in Figure 6 have been replotted as dimensional changevs. moisture content, clearly shows that the hygroexpansional strain of the restraint-driedsheets followed a non-linear relationship during the first moisture cycle but that, as thebuilt-in stresses were released, the behaviour became linear. The freely dried sheets, on theother hand, showed a linear, and reversible, relationship for all cycles. This is in agreementwith earlier findings (cf. Larocque 1936, Uesaka et al. 1992).

Kinetic measurements of the hygroexpansional strain during adsorption (Figure 8)showed no significant difference in expansion rate between the different fibre treatmentsand the time to reach equilibrium dimensional change was about the same for sheets driedunder restraint and dried freely, i.e. due to the greater absolute expansion, the freely driedsheets expanded at a faster absolute rate than the restraint-dried sheets. The same behaviour

JOINT INFLUENCE ON THE DIMENSIONAL STABILITY 19

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

50 20 40 65 85 65 40 20 40 65 85 65 40 20 40 65 85 65 40 20 40 65 85 65 40 20 40 65 85 65 40 20 40 65 85 65 40 20 40 50

Relative humidity (%)

Dim

ensi

onal

cha

nge

(%)

0

4

8

12

16

20

24

28

Moi

stur

e co

nten

t (%

)

Virgin Hornified Virgin + PEM Hornified + PEM(a)

-1.2

-0.9

-0.6

-0.3

0

0.3

0.6

0.9

50 20 40 65 85 65 40 20 40 65 85 65 40 20 40 65 85 65 40 20 40 65 85 65 40 20 40 65 85 65 40 20 40 50

Relative humidity (%)

Dim

ensi

onal

cha

nge

(%)

0

4

8

12

16

20

24

28

Moi

stur

e co

nten

t (%

)

Virgin Hornified Virgin + PEM Hornified + PEM(b)

Figure 6: Equilibrium hygroexpansional strain and equilibrium moisture content as a function ofrelative humidity for (a) restraint-dried and (b) freely dried sheets. The bars indicate 95 % confidencelimits. Filled symbols indicate dimensional change and open symbols indicate moisture content.

20 RESULTS AND DISCUSSION

Table II: Amplitude of dimensional change, change in moisture content (∆MC) and hygroexpansioncoefficient (βMC ) for restraint-dried and freely dried sheets after all built-in stresses were released bymoisture cycling. The values are taken from the last desorption from 85 to 20 % RH in Figure 6, butcalculated as an expansion. The expansion values are given with 95 % confidence limits.

Restraint-dried Freely driedExpansion ∆MC βMC Expansion ∆MC βMC

(%) (%) (%/%) (%) (%) (%/%)

Virgin 0.469 ± 0.027 7.76 0.060 0.890 ± 0.062 8.57 0.104Hornified 0.503 ± 0.017 7.83 0.064 0.797 ± 0.020 8.33 0.096Virgin + PEM 0.528 ± 0.043 8.13 0.065 0.936 ± 0.047 7.41 0.126Hornified + PEM 0.466 ± 0.016 7.61 0.061 0.829 ± 0.027 7.45 0.111

-0.6

-0.4

-0.2

0

0.2

0.4

0 2 4 6 8 10 12 14 16Moisture content (%)

Dim

ensi

onal

cha

nge

(%)

VirginHornifiedVirgin + PEMHornified + PEM

(a)

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

0 2 4 6 8 10 12 14 16Moisture content (%)

Dim

ensi

onal

cha

nge

(%)

VirginHornifiedVirgin + PEMHornified + PEM

(b)

Figure 7: Equilibrium hygroexpansional strain as a function of equilibrium moisture content for (a)restraint-dried and (b) freely dried sheets. The figure shows the first three cycles of Figure 4. The barsindicate 95 % confidence limits.

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0 20 40 60 80 100 120 140 160 180 200Time (min)

Dim

ensi

onal

cha

nge

(%)

2

4

6

8

10

12

14

16

18

20

22

24

Moi

stur

e co

nten

t (%

)

VirginHornifiedVirgin + PEMHornified + PEM

(a)

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100 120 140 160 180 200Time (min)

Dim

ensi

onal

cha

nge

(%)

2

4

6

8

10

12

14

16

18

20

22

24

Moi

stur

e co

nten

t (%

)

VirginHornifiedVirgin + PEMHornified + PEM

(b)

Figure 8: Kinetics of hygroexpansional strain and moisture content changes in (a) restraint-driedsheets and (b) freely dried sheets subjected to a humidity change from 20 to 85 % RH after all built-in stresses were released. Every fourth value is given with 95 % confidence limits. Filled symbolsindicate dimensional change and open symbols indicate moisture content.

JOINT INFLUENCE ON THE DIMENSIONAL STABILITY 21

Figure 9: Schematic representation of the fibre-fibre joint configuration depending on drying strategyand whether or not PEMs are adsorbed onto the fibre surface.

could be seen during both adsorption (Figure 8) and desorption (see Paper I). On the otherhand, there was no clear difference in the relative rate of dimensional change (percentageof the equilibrium strain) between restraint-dried and freely dried sheets except that thepapers made from hornified fibres dried under restraint showed the slowest relative changewhen subjected to the humidity change (see Paper I).

There is significant difference in fibre-fibre contact area between never-dried fibres,dried fibres and PEM-treated fibres (Torgnysdotter and Wågberg 2006). Nevertheless, therewas no major difference in in-plane dimensional movement between sheets prepared withthe differently treated fibres (Figure 6). This implies, in agreement with Uesaka and Qi(1994), that neither the fibre-to-fibre contact area nor the adhesion properties, i.e. whetheror not there are covalent linkages, has any significant influence on the dimensional stabil-ity of paper as long as the sheet has been dried under restraint. For freely dried sheets,however, it is evident in Table II that the PEM-containing sheets seemed to show an ex-pansion equal to or greater than that of the non-PEM-treated sheets when exposed to thesame change in relative huidity, despite the fact that the change in moisture content for thegiven change in relative humidity was lower for the PEM-treated sheets, i.e. they presenteda larger βMC-value.

Since the density was significantly lower in the freely dried sheets (Figure 4), the fibre-fibre joints have a greater extension in the z-direction, schematically shown in Figure 9.The PEM-treatment also creates a larger contact zone, and not only a larger degree ofcontact, between the fibres when PEMs are applied on dried fibres2, as is also indicatedin Figure 9. Once the fibres expand due to moisture adsorption, the joints with a greater

2personal communication of data from Torgnysdotter and Wågberg (2006)

22 RESULTS AND DISCUSSION

0

0.1

0.2

0.3

0.4

0.5

0.6

2 4 6 8 10 12 14Moisture content (%)

Dim

ensi

onal

cha

nge

(%)

VirginHornifiedVirgin + PEMHornified + PEM

Time

(a)

0

0.2

0.4

0.6

0.8

1

2 4 6 8 10 12 14Moisture content (%)

Dim

ensi

onal

cha

nge

(%)

VirginHornifiedVirgin + PEMHornified + PEM

Time

(b)

Figure 10: Hygroexpansion as a function of moisture content in (a) restraint-dried sheets and (b)freely dried sheets when subjected to a humidity change from 20 to 85 % RH after all built-in stresseswere released. The bars indicate 95 % confidence limits. The broken line serves as a guide to the eye.

extension in the z-direction transfer a greater fraction of the transverse expansion into thein-plane direction. An idea mathematically demonstrated by Uesaka and Qi (1994).

The restraint-dried and freely dried sheets showed a major difference in dynamic hy-groexpansion when this was studied as function of sheet moisture content. In the first partof the adsorption process, both freely dried and restraint-dried sheets expand with a similarslope (Figure 10). At a certain point, sheets dried under restraint cease to expand whereas,at more or less the same moisture content around 10− 11 %, the freely dried sheets startto expand with a greater slope. This behaviour was not seen by Niskanen et al. (1997)when using their vertical experimental setup for hygroexpansion measurements. However,with a horizontal set-up, and assuming the same moisture sorption rate (moisture contentwas not monitored), their machine-direction samples ceased to expand much earlier thanthe cross-machine samples, i.e. a behaviour similar to the present results shown in Fig-ure 10. The explanation may be that, like the restraint-dried sheets, the freely dried sheetsshowed a linear in-plane expansion in the longitudinal direction of the fibres and later, athigher moisture contents, a straightening of microcompressions and kinks in the free sec-tions between the fibre-fibre joints. These compressions and kinks are not formed in themachine-direction, i.e. when sheets are dried under restraint (Page and Tydeman 1962).

Moisture sorptivity and fibre wall cross-linking

The hornification process and the use of PEMs had little effect, with the exception ofthe PEM-treated never-dried fibres, on the moisture sorptivity of the papers, as shown inFigure 11. The decrease in moisture content upon cycling (Figure 11a vs. 11b) is a resultof a recrystallisation at higher relative humidities of the cellulose crystallites, which have

MOISTURE SORPTIVITY AND FIBRE WALL CROSS-LINKING 23

0

2

4

6

8

10

12

14

16

0 10 20 30 40 50 60 70 80 90 100Relative humidity (%)

Moi

stur

e co

nten

t (%

)

VirginHornifiedVirgin + PEMHornified + PEM

(a)

Adsorption

Desorption

0

2

4

6

8

10

12

14

16

0 10 20 30 40 50 60 70 80 90 100Relative humidity (%)

Moi

stur

e co

nten

t (%

)

VirginHornifiedVirgin + PEMHornified + PEM

(b)

Desorption

Adsorption

Figure 11: Sorption isotherms for restraint-dried sheets, (a) before and (b) after all built-in stresseswere released.

0

2

4

6

8

10

12

14

16

18

0 10 20 30 40 50 60 70 80 90 100Relative humidity (%)

Moi

stur

e co

nten

t (%

)

RefPIPIIPIII

Figure 12: Sorption isotherms for sheets of untreated and periodate-oxidised fibres.

been partly decrystallised during the pulping process (Caulfield and Steffes 1969). As aconsequence, this will also result in a change in the cellulose free volume (Pekarovicovaet al. 1997). Interestingly, the already hornified fibres showed both the same sorptivity andthe same decrease in equilibrium moisture content as the virgin fibres after being subjectedto several moisture cycles (Figure 11); an observation in accordance with Lundberg andde Ruvo (1978), who argued that hornification does not affect the equilibrium moisturecontent under moist conditions. If the fibres, on the other hand, were cross-linked, thesorption behavior was, as shown in Figure 12, significantly changed due to a loweredcrystallinity and thus a changed availability for water-molecules to the sorption sites, suchas carboxyl groups, within the fibre wall. This is discussed further below.

It is difficult to show in a straightforward way that the fibres are cross-linked throughhemiacetal linkages, since these bonds already exist between the glucose units in cellulose.

24 RESULTS AND DISCUSSION

-10

0

10

20

30

40

50

60

70

80

-0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4Carbonyl content (mmol/g)

RefPIPIIPIII

(a)

Dry

and

wet

tens

ile s

treng

ht in

dex

(kN

m/k

g)

0

1

2

3

4

5

6

-0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4Carbonyl content (mmol/g)

Dry

and

wet

stra

in a

t bre

ak (%

)

RefPIPIIPIII

(b)

Figure 13: (a) Dry and wet tensile strength index and (b) dry and wet strain at break as functionof carbonyl content in the periodate-oxidised fibres. Filled symbols indicate dry properties and opensymbols indicate wet properties. The bars indicate 95 % confidence limits.

2

4

6

8

10

12

14

-20 0 20 40 60 80 100 120 140 160 180Time (min)

Moi

stur

e co

nten

t (%

)

0

10

20

30

40

50

60

70

80

90

100

Targ

et R

H (%

)

RefPIPIIPIIIRH

(a)

2

4

6

8

10

12

14

-20 0 20 40 60 80 100 120 140 160 180Time (min)

Moi

stur

e co

nten

t (%

)

0

10

20

30

40

50

60

70

80

90

100

Targ

et R

H (%

)

RefPIPIIPIIIRH

(b)

Figure 14: Sorption rate of oxidised fibres when subjected to a change in relative humidity from30 to 90 % RH (a) before and (b) after first time exposure to 98 % RH.

Indirectly, the amount of cross-linking can be quantified by measuring the amount of ox-idised hydroxyl groups as the carbonyl content of the fibres. However, since the strengthproperties, see Figure 13, in general, and the wet strength in particular, were significantlyincreased and in the case of wet strength up to about 40 % for the most oxidised fibres,it is very probable that covalent cross-links had been formed both within the fibre walland between individual fibres. It was also basically impossible to re-pulp the sheets indeionised water, but it was found to be readily possible under alkaline conditions. Thetensile stiffness of the sheets made from the cross-linked sheets was also increased and theDMA-measurements showed that these sheets had a greater ability to withstand changesin storage modulus when subjected to rapid changes in relative humidity (see Paper II),which is promising from a mechano-sorptive creep point of view (Byrd 1972a,b).

MOISTURE SORPTIVITY AND FIBRE WALL CROSS-LINKING 25

(a)

(b)

Figure 15: Schematic representation of the structure of (a) a cross-linked fibre wall during mois-ture sorption at lower moisture levels, and (b) how the cross-links eventually rupture as the moisturesorption is increased above a critical level corresponding to capillary condensation.

Not only was the equilibrium sorption altered, but the dynamics of both adsorption anddesorption were also significantly changed. When the sheets were subjected to a change inrelative humidity from 30 to 90 % RH during a time of two hours, the most cross-linkedfibres showed only a very limited moisture adsorption (Figure 14a).

A possible explanation of these detected effects of the periodate treatment (Figure 12and 14) could be that it has two opposite effects on the interaction between the fibres andmoisture. Firstly, the periodate creates covalent cross-links in the fibre wall and this bothimproves the integrity of the fibre wall and removes possible sites for water sorption ontothe cellulose surfaces. Secondly, the periodate decreases the crystallinity of the cellulosewithin the fibres (Kim et al. 2000) making the hydroxyl groups of the cellulose availablefor the sorption of water molecules. Assuming that these factors are of the same order ofmagnitude, the fibres should absorb similar amounts of water at low relative humidity andthis can indeed be seen in Figure 12. As the absorption continues, the available surfacesstart to become saturated with water molecules and, as been established earlier by Haselton(1954), a monolayer coverage (of N2-molecules) is achieved around a relative nitrogenpressure of 30 %. Above this level, a multilayer adsorption takes place and this will start toseparate the concentric lamellae of cellulose fibrils constituting the fibre wall. This will beeasier for the non-oxidised fibres than for the oxidised fibres containing cross-links withinthe fibre wall. This type of behaviour can also be detected in Figure 12 when the mostoxidised fibres and the non-oxidised sample are compared. For the non-oxidised samplethere is an increase in moisture content from 3.9 % to 7.7 % when the relative humidityis increased from 30 to 70 % RH whereas the increase for the oxidised sample is from4.5 % to 7.6 %.

On the other hand, after the fibres had been subjected to a relative humidity of 98 % RH(Figure 14b), there was no notable difference in sorption rate, indicating a breakage of the

26 RESULTS AND DISCUSSION

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

-10 0 10 20 30 40 50 60 70 80 90Time (min)

Dim

ensi

onal

cha

nge

(%)

0

10

20

30

40

50

60

70

80

90

100

Rel

ativ

e hu

mid

ity (%

)

Ref PI PII PIII RH(a)

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

-10 0 10 20 30 40 50 60 70 80 90Time (min)

Dim

ensi

onal

cha

nge

(%)

0

10

20

30

40

50

60

70

80

90

100

Rel

ativ

e hu

mid

ity (%

)

Ref PI PII PIII RH(b)

Figure 16: Dynamic dimensional change when papers from oxidised fibres were subjected to a changein relative humidity from 20 to 85 % RH (a) before and (b) after first time exposure to 98 % RH andall built-in stresses were released. The bars indicate 95 % confidence limits.

cross-links within the fibre wall. However, since there was still a significant wet strengthupon complete wetting, there was most probably only a limited breakage of bonds at higherhumidities where capillary condensation is noticeable. It can be suggested that this is dueto a slow breaking up of areas adjacent to covalent cross-links, as schematically demon-strated in Figure 15a. As the moisture level in the fibre wall is increased above the capillarycondensation limit, i.e. > 98 % RH (Stone and Scallan 1967), with liquid water present,the acetal linkages may start to hydrolyse and permit an expansion of the surfaces areawithin the fibre wall, as schematically shown in Figure 15b. It is naturally difficult to de-termine the exact point where capillary condensation starts to occur readily but, accordingto the Kelvin equation (1), all capillaries with a radius smaller than 10 nm are filled at90 % RH and capillaries smaller than 20 nm at 95 % RH. Andreasson et al. (2003) havemeasured the average pore size of a wet chemical pulp to be between 10 and 25 nm at pulpyields below 60 %. Since the papers from the differently treated fibres show no differencein moisture adsorption kinetics after they have been subjected to higher relative humidities(Figure 14b), this supports the hypothesis that partial breakage of acetal linkages occurswithin the fibre wall at higher moisture contents (Figure 15b).

The exposure of papers to almost saturated humidities will naturally affect the dimen-sional stability properties of the papers. Interestingly, as Figure 16b shows, a major part ofthe dimensional stabilisation effect remained even though the rate and amount of moistureadsorption were similar (Figure 14b), which also indicate that there are still considerableamounts of cross-links available. Due to the release of built-in stress, it is unfortunatelydifficult to study the kinetics of the dimensional change of non-cycled samples (Figure 16a)in the same way as the sorption kinetics (Figure 14a).

The oxidative treatment of the fibre wall also led to a significant increase in the equi-

MOISTURE SORPTIVITY AND FIBRE WALL CROSS-LINKING 27

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

50 20 40 65 85 65 40 20 40 65 85 65 40 20 40 65 85 65 40 20 40 65 85 65 40 20 40 65 85

Relative Humidity (%)

Dim

ensi

onal

cha

nge

(%)

0

4

8

12

16

20

24

28

Moi

stur

e co

nten

t (%

)

Ref PI PII PIII(a)

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0 2 4 6 8 10 12 14 16Moisture content (%)

Dim

ensi

onal

cha

nge

(%)

RefPIPIIPIII

(b)

Figure 17: (a) Equilibrium hygroexpansional strain and equilibrium moisture content as a function ofrelative humidity – filled symbols indicate dimensional change, and open symbols indicate moisturecontent; (b) Equilibrium hygroexpansional strain as a function of equilibrium moisture content. Thebars indicate 95 % confidence limits.

28 RESULTS AND DISCUSSION

Table III: Amplitude of dimensional change, change in moisture content (∆MC) and hygroexpan-sion coefficient (βMC ) after all built-in stresses were released. The values are taken from the lastdesorption from 85 to 20 % RH in Figure 17, but calculated as an expansion. The expansion valuesare given with 95 % confidence limits.

Expansion (%) ∆MC (%) βMC (%/%)

Ref 0.689 ± 0.027 7.83 0.088PI 0.615 ± 0.006 7.23 0.085PII 0.560 ± 0.009 6.24 0.090PIII 0.495 ± 0.009 4.66 0.106

librium dimensional stability when the paper was subjected to changes in relative humidity(Figure 17a). The amplitude of the equilibrium dimensional change after release of allbuilt-in stress, i.e. when reversible hygroexpansion has been achieved and there is no morepermanent shrinkage, when changing the relative humidity from 20 to 85 % RH, was re-duced by approximately 30 % for the most oxidised fibres, compared to the reference.Figure 17a shows that this was primarily due to a lower moisture uptake by the oxidisedfibres. The hygroexpansion coefficients, i.e. the slopes in Figure 17b, are similar for thereference and the two lower degrees of oxidation (PI and PII), but notably higher for themost highly oxidised fibres (PIII). This is further clarified in Table III. One of the moststriking differences, apart from the difference in degree of oxidation between the most ox-idised fibres and the other samples, is the increase in density (see Paper II), but this is not,however, a likely explanation since the density only slightly affects the hygroexpansion ofrestraint-dried sheets (Salmén et al. 1986). Hou et al. (2007) suggested that the increasein strength properties was due to an increase in fibre-fibre contact but, as shown in Pa-per I, the joint properties have no major impact on the hygroexpansion. On this basis, itis reasonable to suggest that the increase in the βMC-value is a result of a significantlyhigher moisture content at 20 % RH and a less porous fibre wall which is more sensitiveto changes in moisture content.

Conclusions

In the first part of this work, the influence of the fibre-fibre joint properties and the effectsof hornification of the fibre wall on the dimensional stability of paper were studied. Theresults showed that, when the degree of fibre-to-fibre contact was decreased through horni-fication or increased by adsorbing polyelectrolyte multilayers onto the fibres prior to sheetformation, there were no effect on the dimensional stability, when the relative humiditywas changed from 20 to 85 % RH, for sheets dried under restraint, but that in freely driedsheets an increase in contact area decreased the dimensional stability by means of a higherhygroexpansion coefficient. This was probably due to an increase in the out-of-plane con-formation of the fibre-fibre joint.

In the second part of the work, the effect of moisture sorptivity on the dimensionalstability of paper was studied. A decrease in moisture sorptivity was achieved by cross-linking the fibre wall through oxidation with sodium periodate to form dialdehyde cellu-lose, which can react into hemiacetal linkages with available hydroxyl groups. The resultsshowed that the dimensional movement was significantly lower for the cross-linked sheetsbecause of a lower, and slower, change in moisture content when the relative humidity waschanged from 20 to 85 % RH. However, the periodate cross-linking seemed to increasethe hygroexpansion coefficient of highly cross-linked sheets, i.e. made them more sensi-tive to absolute changes in moisture content. The cross-linked sheets also showed a slowerdegradation in storage modulus with changing relative humidity due to the slower moistureuptake by the sheets.

Future work

Since it is evident that the result, in terms of dimensional stability, depends on whether thisis expressed as βMC or as βRH values, it would be interesting to further examine the paperexpansion when exposed to liquid water, i.e. to measure the hydroexpansion. The dynamichydroexpansion is also of great interest for understanding the mechanisms of dimensionalstability during offset printing.

The βRH was dramatically decreased by treating the fibres with periodate and since

29

30 CONCLUSIONS

a lower βRH is assumed to result in improved mechano-sorptive creep properties (Byrd1972a,b), it would be of interest to examine the creep properties of sheets made fromperiodate oxidised fibres.

Finally, the true mechanisms behind the change in sorptivity of cross-linked fibres arestill unclear and further research is evidently needed. It would for example be interestingto study how the periodate cross-linking influences the pore-size distribution of the fibres,to see whether the expected decrease in average pore-size can be confirmed.

Acknowledgements

This is perhaps the most difficult part of a thesis to write and, if you do not find yourselfwithin these acknowledgements, it is not because I am not grateful, but is just due to thefact that there are so many to acknowledge. There are, however, some to whom I wish toextend some extra gratitude.

Firstly, I should like to express my deepest gratitude to my supervisor Professor LarsWågberg both for giving me the opportunity to be part of the Fibre Technology group andfor all the guidance and support during the work on this thesis.

I am also very grateful to BIM Kemi Sweden AB and the Knowledge Foundationthrough its graduate school YPK for financial support.

All members of the Department of Fibre and Polymer Technology and the YPK gradu-ate school are acknowledged for creating such a stimulating environment. My roommatesMagnus and Oskar are specially thanked for all the valuable and interesting discussions,and for all the laughs... and Brita for keeping track of all the “whats”, “hows” and “whens”!

STFI-Packforsk – Fellan, Sune, Ann, Anne-Mari and all other personnel, is acknowl-edged for granting access to their facilities and for all the valuable guidance with the equip-ment.

Jarmo Tulonen, TJT-Teknik AB, is acknowledged for substantially contributing to thedesign of the drying-frame.

Mats Rundlöf, AB Capisco, is acknowledged for the aid with the illustrational work inFigure 9.

Finally, I would like to send my gratitude to my friends and family for all the love andsupport during all years, especially to you Kaisa, for encouraging me to wake up and domy very best even on the darkest winter mornings – Olet niin ihana ja minä rakastan sinuakoko sydämestäni!

31

Bibliography

Agarwal, M., Lvov, Y., and Varahramyan, K. (2006). Conductive wood microfibres forsmart paper through layer-by-layer nanocoating. Nanotechnology, 17:5319–5325.

Andreasson, B., Forsström, J., and Wågberg, L. (2003). The porous structure of pulp fibreswith different yields and its influence on paper strength. Cellulose, 10:111–123.

Barkas, W. W. (1949). The Swelling of Wood Under Stress: A Discussion of its Hygro-scopic, Elastic and Plastic Properties. His Majesty’s Stationery Office, London.

Berthold, J., Desbrieres, J., Rinaudo, M., and Salmén, L. (1994). Types of adsorbed wa-ter in relation to the ionic groups and their counter-ions for some cellulose derivatives.Polymer, 35(26):5729–36.

Brecht, W. (1957). Beating and hygrostability of paper. In Fundamentals Paper-makingFibres, Trans. Symposium Cambridge, pages 241–62.

Brecht, W., Gerspach, A., and Hildenbrand, W. (1956). Tensions on drying and their effectson various paper properties. Das Papier, 10:454–8.

Brecht, W. and Hildebrand, W. (1960). Über die flächenbeständigkeit der papiere. DasPapier, 14(10A):610–624.

Byrd, V. L. (1972a). Effect of relative humidity changes during creep on handsheet paperproperties. Tappi., 55:247–252.

Byrd, V. L. (1972b). Effect of relative humidity changes on compressive creep response ofpaper. Tappi., 55:1612–1613.

Caulfield, D. F. and Steffes, R. A. (1969). Water-induced recrystallization of cellulose.Tappi, 52:1361–6.

Chatterjee, S. G., Ramarao, B. V., and Tien, C. (1997). Water-vapor sorption equilibriaof a bleached-kraft paperboard - a study of the hysteresis region. J. Pulp Pap. Sci.,23:J366–J373.

33

34 BIBLIOGRAPHY

Cohen, W. E., Stamm, A. J., and Fahey, D. J. (1959). Dimensional stabilization of paperby cross-linking with formaldehyde. Tappi, 42:934–49.

Dautzenberg, H., Jaeger, W., Kötz, J., Philipp, B., and Stscherbina, D. (1994). Polyelec-trolytes: Formation, Characterization and Application. Hanser, Munich, Germany.

Davison, R. W. (1972). The weak link in paper dry strength. Tappi, 55:567–73.

Decher, G. (1992). Single- and multilayer structures formed on substrates and methods forforming them. Ger. Offen., page 23 pp.

Eklund, D. (1969). Dimensional stability of paper from different types of pulp. PaperiPuu, 51:2:153–5, 158–61.

Eriksson, M., Notley, S. M., and Wågberg, L. (2005). The influence on paper strengthproperties when building multilayers of weak polyelectrolytes onto wood fibres. J. Col-loid Interface Sci., 292(1):38–45.

Eriksson, M., Torgnysdotter, A., and Wågberg, L. (2006). Surface modification of woodfibers using the polyelectrolyte multilayer technique: Effects on fiber joint and paperstrength properties. Ind. Eng. Chem. Res., 45:5279–5286.

Everett, D. H. and Whitton, W. I. (1952). General approach to hysteresis. Transactions ofthe Faraday Society, 48:749–57.

Fernandes Diniz, J. M. B., Gil, M. H., and Castro, J. A. A. M. (2004). Hornification-itsorigin and interpretation in wood pulps. Wood Sci. Technol., 37:489–494.

Ghosh, P. and Dalal, J. C. (1988). Studies on crosslinking of dialdehyde cellulose anddialdehyde starch using poly(vinyl alcohol). Journal of Polymer Materials, 5:241–7.

Grignon, J. and Scallan, A. M. (1980). Effect of ph and neutral salts upon the swelling ofcellulose gels. J. Appl. Polym. Sci., 25:2829–43.

Haselton, W. R. (1954). Gas adsorption by wood, pulp, and paper. i. the low-temperatureadsorption of nitrogen, butane, and carbon dioxide by sprucewood and its components.Tappi, 37:404–12.

Hollenbeck, R. G., Peck, G. E., and Kildsig, D. O. (1978). Application of immersionalcalorimetry to investigation of solid-liquid interactions: microcrystalline cellulose-watersystem. Journal of Pharmaceutical Sciences, 67:1599–606.

Hou, Q. X., Liu, W., Liu, Z. H., and Bai, L. L. (2007). Characteristics of wood cellulosefibers treated with periodate and bisulfite. Ind. Eng. Chem. Res., 46:7830–7837.

35

Iler, R. K. and Colloid, J. (1966). Multilayers of colloidal particles. Interface Sci., 21:569–94.

Jarrell, T. D. (1927). Effect of atmospheric humidity on the moisture content of paper.Paper Trade Journal, 85:47–51.

Jayme, G. (1943). Über die reaktionsfähigkeit von zellstoffen. Cellulosechemie, 21:4:73–86.

Jayme, G. (1944). Mikro-quellungsmessungen an zellstoffen. Der Papier-Fabrikant /Wochenblatt für Papierfabrikation, 6:187–194.

Kato, K. L. and Cameron, R. E. (1999). A review of the relationship between thermally-accelerated ageing of paper and hornification. Cellulose, 6:23–40.

Kim, U.-J., Kuga, S., Wada, M., Okano, T., and Kondo, T. (2000). Periodate oxidation ofcrystalline cellulose. Biomacromolecules, 1:488–492.

Laine, J., Lindström, T., Bremberg, C., and Glad-Nordmark, G. (2003). Studies ontopochemical modification of cellulosic fibers part 5. comparison of the effects of sur-face and bulk chemical modification and beating of pulp on paper properties. NordicPulp Paper Res. J., 18:325–332.

Laivins, G. V. and Scallan, A. M. (1993). The mechanism of hornification of wood pulps.In Baker, C. F., editor, Products of Papermaking: Transactions of the 10th FundamentalResearch Symposium, Oxford, volume 2, pages 1235–1260. Pira International, Leather-head, Surrey, UK.

Larocque, G. (1936). The extension of paper by adsorbed water vapor. Pulp Pap. Mag.Canada, 37:199–209.

LeBel, R. G., Schwartz, R. W., and Sepall, O. (1968). A novel approach to dimensionalstabilization of paper. Tappi, 51:2:79A–84A.

Leisen, J., Hojjatie, B., Coffin, D. W., Lavrykov, S. A., Ramarao, B. V., and Beckham,H. W. (2002). Through-plane diffusion of moisture in paper detected by magnetic reso-nance imaging. Ind. Eng. Chem. Res., 41:6555–6565.

Lindström, T. and Carlsson, G. (1978). The effect of chemical environment on fiberswelling. In Pap. - Eucepa Symp., pages 32–52. Stowarzyszenie Inz. Tech. Przem. Pa-pier., Lodz, Pol.

Lindström, T. and Carlsson, G. (1982). The effect of carboxyl groups and their ionic formduring drying on the hornification of cellulose fibers. Svensk Papperstidning, 85:R146–R151.

36 BIBLIOGRAPHY

Lundberg, R. and de Ruvo, A. (1978). The influence of drying conditions on the recoveryof swelling and strength of recycled fibers. Svensk Papperstidning, 81:355–8.

Lyne, Å. L., Fellers, C., and Kolseth, P. (1994). Three dimensional hygroexpansivity ofunfilled and filled laboratory made fine papers. Appita J., 47:380–6, 401.

Nanri, Y. and Uesaka, T. (1993). Dimensional stability of mechanical pulps - drying shrink-age and hygroexpansivity. Tappi J., 76(6):62–6.

Nazhad, M. M. and Paszner, L. (1994). Fundamentals of strength loss in recycled paper.Tappi J, 77:171–9.

Newman, R. H. (2004). Carbon-13 nmr evidence for cocrystallization of cellulose as amechanism for hornification of bleached kraft pulp. Cellulose, 11:45–52.

Niskanen, K. J., Kuskowski, S. J., and Bronkhorst, C. A. (1997). Dynamic hygroexpansionof paperboards. Nordic Pulp Paper Res. J., 12:103–110.

Nordman, L. (1958). Laboratory investigations into the dimensional stability of paper.Tappi, 41:23–30.

Page, D. H. (1969). Theory for the tensile strength of paper. Tappi, 52:674–81.

Page, D. H. and Tydeman, P. A. (1962). A new theory of the shrinkage, structure andproperties of paper. In Formation and Structure of Paper, pages 397–413.

Pekarovicova, A., Venditti, R. A., Cao, H., Lou, Y. M., and Jean, Y. C. (1997). Cellulosefree volume characterization using positron annihilation lifetime spectroscopy. J. PulpPap. Sci., 23:J101–J107.

Peralta, P. N. (1995). Modeling wood moisture sorption hysteresis using the independent-domain theory. Wood Fiber Sci., 27:250–7.

Salmén, L., Boman, R., Fellers, C., and Htun, M. (1987). The implications of fiber andsheet structure for the hygroexpansivity of paper. Nordic Pulp Paper Res. J., 2:127–31.

Salmén, L., Fellers, C., and Htun, M. (1986). The in-plane and out-of-plane hygroexpan-sional properties of paper. In Punton, V., editor, Papermaking Raw Materials, pages511–527. Mech. Engrg. Publishers, London.

Samuelson, O. and Törnell, B. (1961). Influence of lactones upon the determination ofcarboxyl groups in cellulose. Svensk Papperstidning, 64:198–203.

Scallan, A. M. and Tigerström, A. C. (1992). Swelling and elasticity of the cell walls ofpulp fibers. J. Pulp Pap. Sci., 18:188–93.

37

Schultz-Eklund, O., Fellers, C., and Johansson, P. A. (1992). Method for the local de-termination of the thickness and density of paper. Nordic Pulp Paper Res. J., 7:133–9,154.

Stamm, A. J. (1959). Dimensional stabilization of paper by catalyzed heat treatment andcross-linking with formaldehyde. Tappi, 42:44–50.

Stone, J. E. and Scallan, A. M. (1966). Influence of drying on the pore structure of the cellwall. In Bolam, F., editor, Consolidation of the Paper Web, volume 1, pages 145–174,Cambridge. B. P & B. M. A., London.

Stone, J. E. and Scallan, A. M. (1967). Effect of component removal upon the porousstructure of the cellwalls of wood. II. Swelling in water and the fiber saturation point.Tappi, 50(10):496–501.

Stone, J. E. and Scallan, A. M. (1968). Effect of component removal upon the porous struc-ture of the cell wall of wood. III. Comparison between the sulfite and kraft processes.Pulp Pap. Mag. Canada, 69:69–74.

Symons, M. C. R. (1955). Evidence for formation of free-radical intermediates in somereactions involving periodate. Journal of the Chemical Society, pages 2794–6.

Thalén, N., Kolmodin, H., and Glasare, K. (1994). Dimensional changes of newsprintin offset printing. In Fellers, C. and Laufenberg, T. L., editors, Moisture-Induces CreepBehaviour of Paper and Board, pages 149–165. Co-organized by STFI and USDA ForestService, Forest Products Laboratory.

Torgnysdotter, A. and Wågberg, L. (2003). Study of the joint strength between regener-ated cellulose fibers and its influence on the sheet strength. Nordic Pulp Paper Res. J.,18:455–459.

Torgnysdotter, A. and Wågberg, L. (2004). Influence of electrostatic interactions on fi-bre/fibre joint and paper strength. Nordic Pulp Paper Res. J., 19:440–447.

Torgnysdotter, A. and Wågberg, L. (2006). Tailoring of fibre/fibre joints in order to avoidthe negative impacts of drying on paper properties. Nordic Pulp Paper Res. J., 21:411–418.

Tydeman, P. A., Wembridge, D. R., and Page, D. H. (1966). Transverse shrinkage ofindividual fibers by microradiography. In Consol. Pap. Web, Trans. Symp., volume 1,pages 119–44.

Uesaka, T. (1991). Dimensional stability of paper. upgrading paper performance in enduse. J. Pulp Pap. Sci., 17:J39–J46.

38 BIBLIOGRAPHY

Uesaka, T.,