Influence of fibre modification on moisture sorption and the mechanical

properties of paper

MAGNUS GIMÅKER

Doctoral Thesis

KTH Royal Institute of Technology Department of Fibre and Polymer Technology

Division of Fibre Technology

Stockholm, Sweden 2010

TRITA-CHE Report 2010:11 ISSN 1654-1081 ISBN 978-91-7415-606-5 Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan framlägges till offentlig granskning för avläggande av teknologie doktorsexamen i pappers- och massateknologi fredagen den 23:e april 2010 klockan 10.00 i hörsal F3, Lindstedtsvägen 26, Stockholm. Avhandlingen försvaras på engelska. © Magnus Gimåker, mars 2010 Tryck: E-print AB, Stockholm

Abstract

Fibre modification might be a way to improve the performance of paper, to increase its cost competiveness and enable new paper-based products to be developed. Therefore, the influence of fibre modification (with polyelectrolytes or by fibre cross-linking) on the mechanical properties of special importance for packaging paper grades was studied. Creep deformation under varying humidity conditions (i.e. mechano-sorptive creep) is of outmost importance for the stacking life of paper-based boxes. The influence on creep behaviour of adsorbing polyallylamine (a cationic polyelectrolyte) to fibre surfaces or throughout the fibre walls was studied. Adsorption to fibre surfaces reduced the creep at constant humidity. The mechano-sorptive creep was not however influenced. The use of polyelectrolytes did not thus appear to be a feasible strategy for reducing mechano-sorptive creep. Polyelectrolytes can however be efficient in improving other mechanical properties. The use of multilayers consisting of polyallylamine (PAH) and polyacrylic acid (PAA) was for example shown to significantly increase the strength of paper with much less densification and build-up of residual stress than is obtained by beating. Cross-linking by oxidation with periodate radically decreased the mechano-sorptive creep of sheets made from the oxidised fibres. The basic mechanism behind the reduction in mechano-sorptive with cross-linking was found to be that the cross-linking slowed down the moisture sorption kinetics. A lower sorption rate led to smaller moisture content variations during the mechano-sorptive creep testing, and thus less sorption-induced swelling and stress concentrations at fibre/fibre joints. However, for cross-linking to be a practical way to reduce creep, the large problem of embrittlement must be solved. The shear strength of couched sheets was measured to study the interaction between the sheets at different solids content. The shear strength was low until a solids content of approximately 60−70% was reached, which suggests that interactions important for the strength at complete dryness start to develop at this solids content. The effect of different fibre modifications and additives on how the fibres interact during the consolidation process is not always well understood. The method of shear strength determination could in the future be applied to modified fibres to hopefully increase the understanding of how different modifications influence the fibre/fibre interactions. A deeper understanding might reduce the time for the development of new and improved fibre modifications.

Sammanfattning Fibermodifiering kan vara en möjlig väg att förbättra egenskaperna hos papper och på så sätt möjliggöra lansering av nya pappersbaserade produkter samt öka kostnadseffektiviteten hos redan existerande produkter. Därför studerades inverkan av fibermodifiering (genom polyelektrolytadsorption eller fiberväggs-tvärbindning) på mekaniska egenskaper som är av särskild stor betydelse för papper som används till förpackningar. Krypdeformation hos papper under varierande klimatbetingelser (s.k. mekano-sorptiv krypning) har stor inverkan på hur länge papplådor kan stå staplade utan att kollapsa. Därför adsorberades polyallylamin (en katjonisk polyelektrolyt) på fiberytor samt i fiberväggar för att undersöka effekten på krypbeteendet hos de resulterande pappersarken. Kryphastigheten under konstanta klimatbetingelser minskades signifikant när polyallylamin hade adsorberats till fiberytorna. Någon effekt på den mekanosorptiva krypningen kunde dock inte ses. Fibermodifiering med polyelektrolyter verkar därför inte vara ett lovande alternativ för att minska den mekanosorptiva krypningen. Användandet av polyelektrolyter kan emellertid vara väldigt effektivt för att förbättra andra mekaniska egenskaper. Tillsats av polyelektrolyter visade sig till exempel ge styrkeökning med mycket mindre densifiering och uppbyggnad av restspänningar än om fibrerna maldes. Fiberväggstvärbindning uppnådd genom oxidation med perjodat minskade den mekanosorptiva krypningen radikalt. Mekanismen bakom reduktionen i mekano-sorptiv krypning visade sig vara att tvärbindningen minskade fuktsorptions-hastigheten. En lägre sorptionshastighet resulterade i mindre fuktvariationer i pappret under den fuktcykling som användes vid krypprovningen och därmed mindre fibersvällning och stresskoncentrationer i fiberfogarna. För att tvär-bindning ska vara ett realistiskt alternativ för att minska den mekanosorptiva krypningen måste dock försprödningsproblematiken lösas. Mätning av skjuvstyrka hos guskade ark användes för att bestämma växelverkan mellan arken vid olika torrhalt. Skjuvstyrkan var låg fram till en torrhalt på ca 60−70%, vilket antyder att krafter som är viktiga för att hålla ihop fibrerna i det torra arket börjar utvecklas just vid denna torrhalt. Om denna metod för mätning av skjuvstyrka kombineras med olika fibermodifieringar kan förståelsen för hur dessa fibermodifieringar påverkar interaktionen mellan fibrer ökas. En ökad förståelse för hur olika fibermodifieringar påverkar fibrerna och pappret, kan förhoppningsvis minska den tid det tar att utveckla nya och förbättrade metoder för fibermodifiering.

List of publications This thesis is based on the following papers:

Paper I

Influence of polymeric additives on short-time creep of paper. Gimåker, M., Horvath, A. and Wågberg, L. (2007), Nordic Pulp & Paper Research Journal, 22(2): 217-227.

Paper II

Adsorption of polyallylamine to lignocellulosic fibres: effect of adsorption conditions on localisation of adsorbed polyelectrolyte and mechanical properties of resulting paper sheets. Gimåker, M. and Wågberg, L. (2009), Cellulose, 16(1): 87-101.

Paper III

On the mechanisms of mechano-sorptive creep reduction by chemical cross-linking. Gimåker, M., Olsson, A.-M., Salmén, L. and Wågberg, L. (2009), In Advances in Pulp and Paper Research, 14th Fundamental Research Symposium, Oxford, UK, Sept. 2009, 1001-1017.

Paper IV

The influence of periodate oxidation on the moisture sorptivity and dimensional stability of paper. Larsson, P. A., Gimåker, M. and Wågberg, L. (2008), Cellulose, 15(6): 837-847.

Paper V

Influence of beating and chemical additives on residual stresses in paper. Gimåker, M., Östlund, M., Östlund, S. and Wågberg, L. (2010), Manuscript

Paper VI

Shear strength development between couched papers during drying. Gimåker, M., Nygårds, M., Wågberg, L. and Östlund, S. (2010), Manuscript

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

Paper I

Principal author. Active part in outlining experiments and interpreting the results. Performed the experimental work, except fluorescent labelling which was performed by Andrew Horvath.

Paper II

Principal author and performed all experimental work. Large active part in outlining experiments and interpreting the results.

Paper III

Principal author. Large active part in outlining experiments and interpreting the results. Performed the experimental work, except creep testing of single fibres and paper sheets in the dynamical mechanical analyser which was performed by Anne-Mari Olsson.

Paper IV

Active part in outlining experiments and interpreting the results. Shared the experimental work with Per Larsson, who was the principal author.

Paper V

Principal author. Active part in outlining experiments and interpreting the results. Performed most experimental work with assistance from Felix Lindström. Determination of residual stresses was performed by Magnus Östlund.

Paper VI

Principal author. Large active part in outlining experiments and interpreting the results. Performed all the experimental work with assistance from Felix Lindström.

Table of contents

Introduction ...................................................................................................... 1

Objective of the work and outline of the thesis .......................................... 2

Background ....................................................................................................... 5

Paper strength and the formation of fibre/fibre joints .......................... 6 Mechanical behaviour of paper prior to failure ....................................... 8 Moisture sorption in fibres ....................................................................... 11 Mechano-sorptive creep ............................................................................ 12 Residual stresses in paper .......................................................................... 13 Fibre modification and strength additives .............................................. 14

Experimental ................................................................................................... 20

Materials....................................................................................................... 20

Fibres ............................................................................................................. 20 Chemicals...................................................................................................... 21

Methods ....................................................................................................... 21

Polyelectrolyte adsorption .......................................................................... 21 Visualising the location of adsorbed polyelectrolyte.............................. 22 Fibre cross-linking ....................................................................................... 22 Sheet preparation......................................................................................... 23 Paper testing ................................................................................................. 23 Creep testing ................................................................................................ 24 Moisture sorption measurements ............................................................. 26 Determination of residual stresses ............................................................ 27 Preparation of specimens for shear testing ............................................. 27

Results and discussion ................................................................................... 29

Fibre modification with polyelectrolytes ................................................ 29

Low ionic strength ...................................................................................... 29 High ionic strength ..................................................................................... 31 Kinetics of fibre wall penetration ............................................................. 32

Influence of polyelectrolytes on creep and mechano-sorptive creep..................................................................... 34

Adsorption to the exterior parts of the fibre wall .................................. 34 Adsorption throughout the fibre wall ...................................................... 40 Mechano-sorptive creep ............................................................................. 43

Influence of polyelectrolytes on residual stresses .................................. 46

Influence of fibre cross-linking on mechano-sorptive creep of hand-sheets ................................................. 51

Influence of fibre cross-linking on mechano-sorptive creep of single fibres ................................................. 56

Influence of cross-linking on moisture sorption ................................... 58

Development of shear strength during drying ....................................... 65

Concluding remarks ....................................................................................... 69

Acknowledgements ........................................................................................ 73

References ....................................................................................................... 74

1

Introduction Paper is a material with many uses. The main functions that paper provides is to serve as an information carrier (printing papers), to protect goods (packaging paper) and to serve as an absorbent (hygiene papers). From a sustainability viewpoint, paper has the advantage, compared to oil-based materials such as plastics, that it is produced from a renewable resource. Paper also has the advantage of being biodegradable and recyclable. Plastics can also be recycled but not all plastics are easily biodegradable. Cellulose is also the most common biopolymer on earth, which implies that there will probably be good access to wood and other plants at a rather inexpensive cost for many years to come. For materials based on oil, however, there is a risk that the easily available oil supplies will run dry, drastically increasing the cost for the production of such materials. In the future there may however be a keen competition for the earth’s biomass to produce bio-based energy, possibly jeopardising the inexpensive access to these resources for the production of biofibre-based materials. However, converting the biomass into highly processed products would create more added value and contribute more to the economic growth. Clearly, from as sustainability viewpoint, wood-fibre-based materials have many advantages, but in terms of product performance they also have many weaknesses. One of the most obvious is their sensitivity towards water and moisture. A second major disadvantage is that, compared to plastics, paper cannot easily be formed into complex geometrical shapes. A third major disadvantage is that paper materials have relatively low toughness and easily suffer permanent deformation, such as wrinkling, which limits their use for many applications. In the light of the environmental and sustainability benefits of paper, it would be advantageous to increase the use of paper and other biofibre-based materials, but for them to be able to compete with plastics and increase the use, thus assuring the cost competitiveness of the forest-based sector, it is absolutely necessary to improve the performance of paper and to create new biofibre-based materials. The modification of fibres with additives or by direct chemical reaction can be a way to improve the performance of paper, increasing it cost competiveness and opening the way to new paper-based products, and that is exactly what the work described in this thesis has explored.

2

Objective of the work and outline of the thesis The objective of the work described in this thesis was to explore the possibilities of using fibre modifications to improve the mechanical properties of paper and to seek to establish the mechanisms behind the improvements obtained. The work had a special focus on properties that are important for packaging grade papers. Creep, the time-dependent deformation of a sample held under a constant load, affects paper and many other materials. For paper-based boxes loaded in compression for a long time, the inherent creep of paper can eventually lead to the collapse of the boxes. Varying humidity accelerates the creep rate, so that the creep during cycling between low and high humidity exceeds the creep found at a high constant humidity. This phenomenon is usually referred to as mechano-sorptive creep or accelerated creep. Despite the importance of time-dependent mechanical behaviour such as creep for the performance of paper-based packaging, most previous studies on the influence of fibre modifications and chemical additives on the mechanical properties of paper are limited to studies of the elastic behaviour and ultimate strength. Accordingly, the possibilities of reducing creep by fibre modification were explored in Papers I, II and III. The fibre/fibre joints and the fibre walls are known to influence the mechanical properties of paper in different ways. It is generally considered that the properties of the fibre/fibre joints determine the ultimate strength, whereas the properties of the fibres themselves determine the viscoelastic behaviour of the fibre network up to failure. In most previous studies on the influence of chemical additives on the mechanical properties of paper, no direct evidence for the location of the added additives has been presented. Therefore, it was an explicit objective of the present study to consider the location of the added polyelectrolytes and to see whether there are differences if only the fibre surfaces are modified with polyelectrolyte (Paper I) and if the fibre walls are also modified either with polyelectrolyte (Paper II) or by cross-linking (Paper III).

3

Cross-linking has previously been shown to reduce mechano-sorptive creep in paper (Caulfield 1994), but no experimental support regarding the mechanisms by which cross-linking reduces mechano-sorptive creep was presented. Since mechano-sorptive creep is a direct consequence of the fact that paper easily sorbs moisture, the intention of this work was also to examine the influence of fibre wall cross-linking on the moisture sorption, in order to increase the understanding of the mechanism by which cross-linking reduces mechano-sorptive creep (Papers III, IV). Cross-linking is usually associated with severe embrittlement of the resulting paper, and might thus not be a practical way to reduce mechano-sorptive creep. However, an understanding of the mechanisms underlying the reduction in mechano-sorptive creep by cross-linking might in turn generate new ideas as to how to reduce mechano-sorptive creep without causing embrittlement. Other mechanical properties that are of special importance for packaging grades papers are residual stresses and shear strength. As described in the introduction, the ability to convert paper into different geometrical shapes, such as boxes, is very important. Paperboard is generally converted into boxes and other shapes by introducing creases, where the paperboard is subsequently folded and glued to its final shape. For successful creasing it is very important that the paperboard easily delaminates, and the ease of delamination is controlled by the shear strength profile of the paper, i.e. the local shear strength through the thickness of the material. Despite the importance of shear strength for paperboard performance, only two previous investigations, to the knowledge of the author, have been devoted to examining how the shear strength of paper develops with moisture removal (Alince et al. 2006; de Oliveira et al. 2008). These studies investigated how shear strength developed with moisture removal in specimens made of previously dried blotting paper sheets rewetted and couched together. The fact that the blotting papers had previously been dried might limit the applicability of the results when it comes to explaining shear strength development in paperboards in which various layers of never-dried fibres are couched together. Furthermore, the shear strength was only tested at a relatively low solids content (<60%), and not all the way to complete dryness. The objective of the work presented in paper VI was accordingly to study the shear strength development during drying to complete dryness in couched sheets made of both never-dried and previously dried fibres.

4

An understanding of the development of shear strength during drying is valuable for optimising the shear strength profile in paperboards, and will give vital clues to the fibre interactions during drying, the nature of these interactions, and how fibres should be treated to optimise these interactions. Residual stresses are the stresses that remain in a piece of material when all external forces are removed. In paper, the residual stress state can influence the likelihood of crack formation during the folding of paperboard, and may thus be important for the convertibility of paperboard. Beating the pulp is known to induce additional shrinkage gradients in multi-ply paperboard and to increase the magnitude of the residual stresses (Östlund et al. 2004a). However, there are no published studies on the effect of chemical additives on the residual stresses in paper. Accordingly, studies were undertaken to clarify whether fibre modification with polyelectrolytes can influence and control the residual stress state in paper (Paper V).

5

Background Paper is made from a suspension of fibres in water and when the water is removed the fibres bond to each other without additional additives. Paper is thus not a composite but a self-bonding fibrous material. The fibres are usually obtained from wood, but sometimes also from other plants, by either a mechanical or a chemical process. During paper production, the fibre suspension is dewatered by different machine elements such as foils, blades, suction boxes, suction rolls and press nips, and finally dried on steam-heated cylinders so that only a little water is left in the final paper. At a low fibre concentration, the fibre network is held together mainly by mechanical entanglement. When water is removed during the production process, the fibres are brought into close contact by capillary forces arising from the water meniscus formed between the fibres (Lyne and Gallay 1954). At a high solids content the fibres are brought into such close contact that strong forces acting at the molecular level can develop, giving the final paper its high strength. Forces that may possibly act to hold the fibres together in the dry paper include: mechanical entanglement of fibres and fibre surface fibrils, inter-diffusion of polymers across the fibre/fibre interface, covalent bonds, ionic bonds, hydrogen bonds, polar interactions and van der Waals interactions (Lindström et al. 2005). The complex structure of paper and its constituent fibres, in combination with the wide variety of possible forces acting to hold the fibres together, make paper one of the most complex engineering materials available (Haslach 2000; Alava and Niskanen 2006). It is not therefore clear how all these different factors influence the mechanical properties of paper, even though a lot of research has been dedicated to the subject. The mechanical behaviour of paper is also difficult to fully characterise, since it shows a complex time-dependence, so that the mechanical response depends on the rate of straining or stressing (Haslach 2000; Coffin 2009). For example, if the stress is applied rapidly, paper shows a higher strength and stiffness but a lower extensibility than if the stress is applied at a lower rate. Paper also shows creep and stress-relaxation. Creep is the slow continuous deformation of a material subjected to a constant stress, whereas stress-relaxation is the slow decay of stress in a material subjected to a constant strain.

6

Both the fibres and the forces holding them together are strongly influenced by moisture, and water can easily disrupt the fibre/fibre joints, unless they are protected by chemical means. Therefore moisture has a large effect on the mechanical properties of paper. For example, the strength and stiffness of paper are reduced under high humidity conditions. Similarly, creep in paper specimens occurs at a faster rate under high moisture conditions (i.e. with larger amounts of adsorbed water).

Paper strength and the formation of fibre/fibre joints The exact nature of the different forces acting to hold the fibres together in the paper sheet, and the contribution of each kind of force, is not known. However, it is known that the fibre/fibre joints are very important for how large a load a paper can withstand before failure (Van den Akker 1950; Davison 1972). Davison (1972) found that the potential strength of paper calculated from the strength of single fibres is less than the strength found in practice. He also found that intact fibres are often pulled out in the rupture-zone, even in strong papers, and he thus concluded that the fibre/fibre joint is the weak link in a dry paper. Of course the fibre strength ultimately limits the strength that theoretically can be reached in a paper. Page (1969) developed a theory that describes how paper strength increases asymptotically with increasing shear bond strength and increasing relative bonded area to approach a maximum limited by the fibre strength. Since paper strength to a large extent is determined by the strength of the fibre/fibre joints, it is necessary to understand how the fibre/fibre joints are formed during the paper manufacturing process and what determines the number of created fibre/fibre joints and their strength. Van der Waals forces, i.e. dispersion forces and dipole interactions, between two atoms or molecules have the range of a few nanometres or less. But for two colloidal particles suspended in a medium, the atoms in one particle are to some extent able to interact with all of the atoms in the other particle and these interactions are to some degree additive. The very important consequence of this partial additivity is that the van der Waals force between colloidal particles has a much longer range than the van der Waals force between individual atoms or molecules. Since long-range van der Waals forces originate from fluctuations in the electron clouds of the

7

atoms (i.e. dispersion forces) they exist between all materials. Cellulosic fibres are to some extent coated with hemicelluloses and in water the fibres will have an anionic charge due to carboxylic groups on the hemicelluloses. An electrostatic double-layer force will hence also always be present between fibres in water. If two fibres are brought sufficiently close together, the entropy loss of confining the dangling hemicellulose chains will result in a repulsive entropy force, referred to as steric repulsion. When water is removed from the fibres during the papermaking process, so that air enters the sheet, water menisci will form between fibres and fibrils and pull them closer due to capillary action (Lyne and Gallay 1954; Clark 1985; van de Ven 2008). The balance between the attractive forces, (capillary and van der Waals forces) and the repulsive forces (electrostatic double layer forces and steric repulsion), determines how close together the fibres come at a given solids content (Wågberg and Annergren 1997). Once the fibres come into sufficiently close contact different interactions such as mechanical entanglement of fibre surface fibrils, inter-diffusion of polymers across the fibre/fibre interface, covalent bonds, ionic bonds, hydrogen bonds, polar interactions, and van der Waals interactions can develop between the fibres (Lindström et al. 2005). The relative importance and contribution of each of all the possible interactions between cellulosic fibres have not yet been quantified. Much more research in this area is probably necessary to accomplish this demanding task. One feasible way to study fibre/fibre interactions is to use model cellulose surfaces. The AFM colloidal probe technique has for example been used to study the interaction between two cellulose surfaces (Notley et al. 2004; Stiernstedt et al. 2006). The JKR method has been applied to study the adhesion between poly(dimethylsiloxane) (PDMS) caps and Langmuir–Blodgett cellulose surfaces (Rundlöf et al. 2000). However, even though the model experiments are well defined, they suffer from the drawback of not using actual fibres and it is also difficult to dry the surfaces together in a manner similar to that occurring in paper making. Shear testing of papers that have been couched together can be another feasible way to study fibre/fibre interaction (Paper VI). Shear testing, compared to in-plane tensile testing, to a large extent avoids the effect of mechanical fibre entanglement, as pointed out by Alince et al. (2006). Shear testing should therefore be a relatively pure measurement of the interactions between the fibres. The author of the present thesis is of the

8

opinion that the measurements on model systems and shear strength measurements on wet and dry paper complement each other, and that together they are a powerful tool for understanding how the fibre/fibre interactions develop during paper manufacture.

Mechanical behaviour of paper prior to failure The deformation behaviour of paper prior to failure has been the subject of considerable research over the years. A key issue has been whether the time-dependent deformation prior to failure depends on the properties of the fibre/fibre joints, on the fibres themselves or on the network structure, and also under which conditions these different factors are of importance. The literature contains conflicting data and different opinions (Haslach 2000). One view is that plastic and viscous deformation is due to inter-fibre effects, such as bond breakage and inter-fibre frictional effects. For example, Rance (1948) hypothesised that permanent deformation is due to bond breaking and that additional time-dependent effects are due to inter-fibre frictional effects. Sanborn (1962) also found that permanent deformation includes bond-breakage. The second view is that both plastic and viscous deformation is due to an irreversible deformation within the fibre wall, and that bond breakage occurs only close to failure. Page and Tydeman (1961) found that total bond breakage is infrequent and argued that it should not be considered as a primary contributor to the shape of the stress-strain curve of paper, which they attribute more to partial bond breakage at microcompressed bonding sites. Furthermore, they concluded that delayed elastic deformation (primary creep) is not associated with frictional effects but is purely an intra-fibre phenomenon. Another study supporting the view that the fibres themselves play an important role was that of Parker (1962), who studied the influence of ethylamine decrystallisation on the viscoelastic properties of paper. He found that wet-pressing had a large influence on creep up to a certain point, but as the sheet strength increased, differences in intra-fibre structure became more important. Later, Hill (1967) studied the creep of single fibres loaded in tension and found that the basic creep results for single fibres are similar to those for paper.

9

Seth and Page (1981) examined the influence of a bonder (locust bean gum) and a de-bonder (surfactant) on the stress-strain curve of paper. In the case of strong efficiently loaded paper sheets, they found that the additives did not influence the shape of the stress-strain curve, only the ultimate failure strength and strain. In contrast, in the case of weaker not fully efficiently loaded sheets, the additives influenced the shape of the curve. To account for the effect of the fibre/fibre joint on the stress-strain curve of paper, they introduced the concept of an efficiency factor to describe how efficiently the stress is transferred between the fibres in a sheet. Seth and Page suggest the efficiency factor to depend on the properties of the fibres (length, width and shear modulus) and the relative bonded area in the sheet, as determined by light scattering. It must be stressed, however, that scattering of visible light only is sensitive down to a fibre separation of approximately 200 nm (half the wavelength of light), and can hence not be used to obtain information regarding the true molecular bonded area. A better measure of the relative bonded area can be obtained by the use of nitrogen adsorption (Haselton 1955; Kallmes and Eckert 1964; Eriksson et al. 2006). A value of unity of the efficiency factor corresponds to a fully efficiently loaded sheet structure. As the strength and number of active fibre/fibre joints are increased by beating or wet-pressing, the efficiency factor asymptotically approaches unity. Simultaneously, the modulus of the paper approaches a limiting value determined by the properties of the fibres (Page et al. 1979; Page and Seth 1980a; Page and Seth 1980b). Since the scaling with efficiency factors transposes the stress-strain curves for incompletely efficiently loaded sheets to a single curve, Page and his colleagues concluded that the viscoelasticity of paper originates from within the fibre wall. DeMaio and Patterson (2005) studied the effect of similar additives as used by Seth and Page (1981) on the creep properties of sheets made from a previously dried, bleached softwood kraft pulp. Creep compliance data showed that the creep curves for treated and untreated sheets were the same in the case of sheets wet pressed under a high press load but different for sheets wet pressed under a low press load. In agreement with Seth and Page, they concluded that for sheets wet pressed at a high load, treatment-induced changes in specific bond strength do not influence the creep deformation because fibre-fibre bonding is at a level where the sheets are efficiently loaded structures. They also concluded that the concept of an

10

efficiency factor can be used to account for the degree of bonding in incompletely efficiently loaded sheets. Clearly the general view today is that the inherent creep response of paper is primarily a result of intra-fibre creep, so that, in strong fully efficiently loaded papers, the creep is determined by the creep of the component fibres. However, in weaker paper with an insufficient relative bonded area, the load is not optimally distributed within the sheet and the stress in the individual fibres is greater. In such sheets, changes in inter-fibre bonding influence the creep response in a manner equivalent to that of a shift in load or time (Coffin 2005). If the creep of a strong efficiently loaded paper depends only on the creep of the component fibres, what then are the mechanisms responsible for creep in fibres composed of cellulose fibrils? Since cellulosic fibres are made up of different biopolymers (cellulose, hemicelluloses and lignin), it is natural that they show an inherent creep as do all polymeric materials (Gedde 1995). There are however few direct studies on the micromechanics and molecular mechanisms involved in the viscoelastic behaviour of pulp fibres, and the field is not yet fully explored. The first study of the creep of individual pulp fibres found that the creep behaviour of a single fibre resembles the creep behaviour of a paper sheet (Hill 1967). The crystallinity and crystallite orientation in the fibres before and after creep was also measured in this study, and it was found that the crystallinity did not change but that the crystallite orientation increased. Thus, it was concluded that there had been a movement of crystalline regions within the fibrils or, more probably, of the fibrils within the fibres during the creep. Byrd (1972a) also found that the fibril angle decreased (increased fibril orientation) during creep under constant humidity conditions. Molecular mechanisms involved in the creep deformation of paper have also been studied using infrared spectroscopy (Olsson and Salmén 2001). Changes in the mid-IR spectra were observed which indicated an orientation of the cellulose molecules and a sliding between cellulose chains as a result of the creep deformation. Raman spectroscopy has also been used to probe molecular deformation mechanisms in natural cellulose fibres (Eichhorn et al. 2001). During the tensile deformation of fibres, the 1095 cm-1 Raman band shifts toward lower wave numbers, and this is believed to be due to a deformation of the molecular backbone of the cellulose. The Raman spectra were recorded during a constant strain rate

11

tensile test and not during a creep test, but it is possible that similar molecular deformations take place also during tensile creep deformation. The fact that considerable effort has been spent on understanding the mechanisms behind creep in paper materials is not surprising, since the occurrence of creep deformation has important practical implications. For example, when corrugated boxes are piled on top of each other and stored for extended times, they creep under the compressive load from the boxes above and eventually fail. It is therefore vital to minimise the creep rate to achieve the maximum stacking life (Koning and Stern 1977; Leake and Wojcik 1993; Henriksson et al. 2007). However, when it comes to the performance of boxes, it is important to keep in mind that the rate of creep of paper is accelerated by changes in humidity, so that the creep during cycling between low and high humidity exceeds the creep observed at a constant high humidity (Byrd 1972a; Byrd 1972b). This phenomenon is usually referred to as mechano-sorptive creep or accelerated creep and will be discussed separately.

Moisture sorption in fibres Dry fibres contain no pores (Stone and Scallan 1967), but in contact with water (whether as liquid or vapour) the fibres spontaneously sorb water due to the favourable interaction between water and hydroxyl groups in the cellulose. So thermodynamically, the driving force for moisture sorption is the gain in enthalpy when water molecules are adsorbed to the cellulose. Breakage of hydrogen bonds and adsorption of water between fibril lamellae will separate them, causing an expansion of the fibre wall (Scallan 1977). Real fibres also contain amorphous and hydrophilic hemicelluloses that will readily adsorb water and further increase the volume of the fibre wall. These processes are usually referred to as hygroexpansion. The water up-take is much faster if fibres are subjected to liquid water than in moist air, since it can take several hours for paper in moist air to reach moisture equilibrium (Jarrell 1927). In paper, water can be absorbed not only in the fibres but also in the pores between the fibres and in the lumen inside the fibres.

12

The interaction between water and paper is possibly the most important factor affecting its end-use performance. Water disrupts the forces holding fibres together so that the strength of a wet paper, assuming that it contains no wet strength agents, is only a small fraction of the strength of the dry paper. Many of the mechanical properties described in this thesis are directly related to the interaction between fibres and water. Creep deformation in paper specimens occurs at a faster rate under high moisture conditions (i.e. with larger amounts of adsorbed water), which may be explained by the fact that sorbed water increase the mobility of the cellulose chains (Froix and Nelson 1975). Mechano-sorptive creep is a consequence of the sorption-induced swelling of fibres and paper (Habeger and Coffin 2000). Residual stresses arise because paper dries and shrinks inhomogeneously (Östlund et al. 2004b).

Mechano-sorptive creep As mentioned earlier, the creep rate of paper is accelerated if the paper is subjected to variations in humidity (Byrd 1972a; Byrd 1972b). Since paper packaging is often exposed to compressive loads and variations in humidity during use, storage and transportation, it is primarily the creep rate during varying humidity conditions, i.e. the mechano-sorptive creep rate, that determines the stacking life-time (Leake and Wojcik 1993; Henriksson et al. 2007). If the moisture content in the air surrounding a paper specimen is altered, this results in moisture gradients in the paper. However, Back et al. (1983) showed that a constant moisture gradient did not lead to mechano-sorptive effects in stress relaxation tests on fibre building board. Hence, it is suggested that mechano-sorptive creep is caused by changing moisture gradients. Coffin and Boese (1997) measured the creep in tension of single fibres and hand-sheets, and found that the single fibres did not exhibit mechano-sorptive creep while the hand-sheets did. Later, both ramie fibres (Habeger et al. 2001) and wood fibres (Olsson et al. 2007) have been shown to exhibit mechano-sorptive creep. It is possible, as pointed out by Habeger et al. (2001) that the rate of change in relative humidity in the surrounding air has to be related to the rate of sorption for the material investigated in

13

order to obtain significant moisture gradients and mechano-sorptive creep. This can explain the lack of single-fibre mechano-sorptive creep in the study by Coffin and Boese (1997). The fact that wood fibres have been shown to exhibit mechano-sorptive creep suggests that the mechano-sorptive creep of the individual fibres is one of the mechanisms behind the mechano-sorptive creep observed in paper. Since the discovery of mechano-sorptive creep in the 1970s, it has attracted considerable attention and two dominant models to describe mechano-sorptive creep in paper have evolved. Habeger and Coffin (2000) suggest that humidity variations give rise to sorption-induced swelling and that this, either by moisture gradients or material heterogeneity, results in localised load cycling that in combination with the non-linear creep of paper gives rise to an accelerated creep. Alfthan et al. (2002) suggest that the anisotropic hygroexpansion of the fibres on exposure to moisture leads to a mismatch of hygroexpansive strains at the fibre/fibre bonds, causing large stresses at the bond sites, and that these, together with the non-linear creep, give rise to an accelerated creep. Since the model of Alfthan is an example of sorption-induced swelling in combination with material heterogeneity, it can be considered as a special case of the more general model presented of Habeger and Coffin.

Residual stresses in paper Residual stresses are the stresses that remain in a material when all external forces are removed. Stress is typically the response of a material to an external load, but it can exist irrespective of load provided that the average force on every cross-section is zero. In paper, residual stresses originate from the fact that paper dries inhomogeneously (Östlund et al. 2004b). For a paper dried so that vapour is allowed to escape from both sides of the paper (i.e. two-sided drying), the outer layers of the paper dry first (Bernada et al. 1998). As the outer layers dry and shrink, the interior of the paper will be compliant, due to its high moisture content, and no significant stress build-up will take place. However, towards the end of the drying, when the interior of the paper dries and shrinks, the shrinkage will be opposed by the already dry and thus stiff surface layers. This will cause tensile stresses in the middle layers of the

14

paper that are not able to shrink and compressive stresses in the outer layers that oppose the shrinkage. Residual stresses have several important implications for paper materials. An uneven residual stress distribution leads to shape distortion, i.e. curl and twist. Curl is an important quality problem within the paper industry, especially for papers subjected to high speed printing (Uesaka 1991). It is thus desirable to have good control of the residual stress development during manufacture. Compressive stress at the surface, as usually found in commercial papers, is a disadvantage in the sense that upon exposure to water the stress-state is modified leading to dimensional changes that may be difficult to predict Tensile stresses on the surface may on the contrary be harmful by facilitating fracture and thus reducing the strength of the material (Lindström et al. 2005). Surface stresses could for example influence the risk for crack formation during the folding of paperboard.

Fibre modification and strength additives Beating can be seen as a physical way to modify fibres and is commonly used to improve the strength and mechanical performance of paper. Beating has several effects on the fibres. The fibres become more flexible with beating (Samuelsson 1964) and thus conform better to each other during the consolidation of the paper sheet, increasing the number of fibre/fibre joints and the molecularly bonded area in each joint. Beating produces fibrillar fines. Fibrillar fines are definitely important for the strength of mechanical pulps (Mohlin 1977), but the influence of fibrillar fines on the strength of chemical pulps is not as great (Sandgreen and Wahren 1960). Beating also results in surface fibrillation, which probably increases sheet strength by mechanical entanglement of the surface fibrils. Page (1985) showed that beating straightens out fibres and suggested that straighter fibres give a better stress distribution in the sheet and thus a higher tensile strength (Page et al. 1979). Although beating is effective for increasing the paper strength, it has several drawbacks. The density of the manufactured paper increases with beating, due to the fact that beating makes the fibres more flexible so that they can pack closer together during the consolidation of the paper (Samuelsson 1964). Density is a key property for paper since it has a great

15

influence on bending stiffness, which in turn, has a considerable impact on the performance of many paper grades. For example, high bending stiffness reduces the tendency for boxes to buckle and fail under a compressive load, prevents printing grade papers from folding under their own weight when being read, and increases the runnability of sack paper during converting. So, in order to maximise the bending stiffness, it is desirable to have as low a density as possible while having a high tensile stiffness. Secondly, beating increases the swelling and water-retaining ability of the fibres thus rendering the dewatering of the paper web more difficult, influencing the rate at which paper can be produced on dryer-limited paper machines. Thirdly, the energy consumption required for beating is costly. The literature contains numerous examples of substances that have been used to increase both the dry strength and wet strength of paper (Lindström et al. 2005). The additives used to increase the dry strength are generally water-soluble polymeric substances. The cellulosic fibre has an anionic charge, and cationic polyelectrolytes are therefore often used, also in combination with anionic polyelectrolytes. By using charged polymers, it is possible to increase the adsorption to the fibres and hence achieve a greater increase in the dry strength. One example of combining anionic and cationic polyelectrolytes is the polyelectrolyte multilayer technique (Decher 1997; Wågberg et al. 2002), in which fibres are consecutively treated with oppositely charged polyelectrolytes to form a multilayer. This makes it possible to achieve very large adsorbed amounts and to dramatically increase the paper strength. In practice, starch is the most common dry strength additive, and it is added both at the wet-end and in the size press. For wet-end application, a water-soluble cationic or amphoteric starch is usually used. Chemical additives can be one way to improve certain properties without affecting other properties in a negative manner. For example, the build-up of polyelectrolyte multilayers on pulp fibres can be used to increase the tensile strength with less densification than if chemical fibres are PFI-beaten (Wågberg et al. 2002) or if the energy input in the refining of mechanical pulp fibres is increased (Lundström 2009). Increasing the surface charge of chemical pulp fibres by irreversible adsorption of carboxymethyl cellulose onto the surface of the fibres has also been shown to increase the in-plane tensile strength with less densification than PFI-beating (Laine et al. 2003).

16

Since chemical additives and beating have been shown to have different impact on the properties of paper, the influence of the two was studied in more detail in the work presented in Paper V. As described previously, the types of interaction active in the fibre/fibre joints and the relative contributions of each type have long been an active area of research, but there is still considerable debate as to the true nature of the fibre/fibre joint. Similarly, the exact molecular mechanisms by which different additives increase paper strength are not fully known, and different polyelectrolytes probably function by different mechanisms. Even if the exact mechanisms are not fully understood, there is considerable knowledge in the published literature about factors that are important for the performance of strength additives. A recent review (Pelton 2004) emphasised the importance of polyelectrolyte structure for paper strength; the more hydrophilic the polyelectrolyte the greater being its effect. The influence of polyelectrolyte multilayers on fibre wettability and wet adhesion (studied by the AFM colloidal probe technique) has also been reported (Lingström et al. 2007). In contrast to the conclusions drawn by Pelton, these authors showed that polyelectrolyte multilayers which present a high advancing contact angle for water (i.e. poor water wettability) give higher wet-adhesion and also higher paper strength. The structure and viscoelasticity of polyelectrolyte multilayers have been studied using a quartz crystal microbalance (Notley et al. 2005). The more viscous and water-rich the polyelectrolyte layer, the higher was the adhesion between multilayer-covered silica surfaces, as determined by the AFM colloidal probe technique. When the effect of the polyelectrolyte multilayers on the adhesion was compared with the effect on the paper strength (Eriksson et al. 2005), it was evident that the water-rich conformable multilayers gave rise to a higher paper strength. New methods to determine fibre/fibre joint strength and contact area have recently been developed (Stratton and Colson 1993; Torgnysdotter and Wågberg 2003). The technique has also been applied to study the effect of polyelectrolyte multilayers on paper strength (Eriksson et al. 2006), where it was shown that the multilayers increased the paper strength via an increased number of fibre/fibre contacts per sheet volume, an increased degree of molecular contact in each fibre/fibre joint and the introduction of covalent bonds.

17

These different results show the complexity of the influence of additives on the fibre-joint and paper strength. It also stresses the need for combinations of model experiments and paper testing to increase the fundamental understanding of the topic. This is important both from a scientific and product optimisation point of view. As described by Lindström et al. (2005), there is plenty of literature on the effect of polymeric additives on paper strength. The influence of polymeric additives on the time-dependent mechanical behaviour prior to failure has however been less well examined. Two studies on how a bonder (locust bean gum) and a de-bonder (surfactant) influence creep and the shape of the stress-strain of paper (Seth and Page 1981; DeMaio and Patterson 2005) suggest that polymeric additives do not influence the mechanical behaviour prior to failure (i.e. viscoelasticity) for strong efficiently-loaded paper sheets. There is however a lack of systematic data on how polymeric additives of different types and different molecular weights influence the viscoelasticity of paper. Since the viscoelasticity is considered to be determined by the mechanical properties of the fibres themselves, it should be possible to influence the mechanical behaviour prior to failure if the added polyelectrolytes alter the properties of the fibres themselves and not just the fibre/fibre-joints. Low molecular weight polymeric additives can access the fibre wall, as shown by Wågberg et al. (1987). The use of low molecular weight polymeric additives should thus make possible the modification of the entire fibres and thereby possibly affect the viscoelasticity of the fibres and the resulting paper sheets. It is possible that the introduction of covalent cross-links into the fibre wall might hinder molecular motions and relaxations, and thus influence the viscoelasticity of the fibre. The wet-strengthening mechanisms of polymers containing primary amines have recently been investigated (Laleg and Pikulik 1991; DiFlavio et al. 2005). The conclusion drawn from these studies is that primary amines can react with aldehydes present in lignocellulosic fibres to form imine and aminal linkages. Primary amines can also form amide linkages with carboxylic groups at elevated temperatures, although this reaction is believed to be less important (DiFlavio et al. 2005). Since it is possible for amines to form crosslinks, the possibility of using polyallylamine (a polyelectrolyte containing primary amines) to reduce the creep of paper was explored in Paper I and II.

18

In order for a polyelectrolyte to influence the mechanical properties of the fibres, it must access the fibre wall. The access to the fibre wall can be controlled by the molecular mass of the polyelectrolyte and the electrolyte concentration during adsorption. The adsorption of polyelectrolytes generally increases with increasing electrolyte concentration to pass through a maximum at some intermediate salt concentration and then decrease at high salt concentrations (Lindström and Wågberg 1983). The initial increase in electrolyte concentration acts to coil the polyelectrolytes. This, in combination with the porous nature of the cellulose fibres, means that a greater surface area is available to the polyelectrolytes, and this result in an increased adsorption. At very high electrolyte concentrations, however, the interaction between the polyelectrolytes and the charged fibre diminishes, resulting in a decrease in the amount adsorbed. The adsorption of a polyelectrolyte on cellulosic fibres generally increases with decreasing molecular mass. A lower molecular mass is associated with a lower radius of gyration, and this increases the number of charges accessible to the polyelectrolyte. This has been observed with several different types of polyelectrolytes: C-PAM (Tanaka et al. 1990), polyethyleneimine (Alince 1990), and polyDADMAC (Wågberg and Hägglund 2001). To summarise, a sufficiently low polyelectrolyte molecular mass and a sufficiently high electrolyte concentration should lead to such a small effective hydrodynamic radius that the polyelectrolyte would probably access pores throughout the entire fibre wall, although too strong an electrolyte concentration could make the driving force for adsorption too small and result in no adsorption at all (van de Steeg et al. 1992; van de Steeg et al. 1993). A recently developed technique, involving labelling the polyelectrolyte with a fluorescent dye and examining fibres with adsorbed polyelectrolyte in a confocal laser scanning microscope, makes it possible to obtain a visual record of where the adsorbed polyelectrolyte is located (Horvath et al. 2008a). With this technique, the localisation of adsorbed polyallylamine was visualised in Paper I and II. Fibre modification can be achieved not only by the addition of chemical additives but also by performing chemical reactions directly on the wood polymers. As described previously one of the main focuses of the work

19

described in this thesis was to identify ways to reduce the mechano-sorptive creep of paper. It has previously been shown that reacting paper with low molecular weight multifunctional carboxylic acids significantly decreases the mechano-sorptive creep (Caulfield 1994). The non-polymeric nature of these substances allows them to diffuse throughout the fibre walls. In combination with a catalyst, they are also very reactive, and can thus cross-link the fibres. It was suggested that the mechanism for the creep reduction was that the covalent cross-links stabilises arrays of moisture-sensitive hydrogen bonds and reduce their tendency to creep into a stress-relaxed configuration. No direct experimental support for the validity of this molecular mechanism was however presented. Cross-linking with carboxylic acids renders paper very brittle and this may thus not be a practical way to reduce mechano-sorptive creep, since the toughness of the paper is also very important for the performance of paper packaging (Henriksson et al. 2007). However, since cross-linking is the only commonly known way to significantly decrease mechano-sorptive creep, the effects of cross-linking were explored in Papers III and IV. The intention was to increase the understanding of the mechanism by which cross-linking reduces mechano-sorptive creep. This might in turn generate ideas how to achieve mechano-sorptive creep reduction without causing embrittlement. Cross-linking by multifunctional acids requires an extra immersion step after the paper is produced, which might be impractical in an industrial perspective. Accordingly, reactive groups were introduced in the fibres prior to sheet-making in the present work. The introduced groups then reacted during drying and cross-links were formed. The reactive groups were introduced by oxidising the fibres with periodate ions. Periodate is known to selectively oxidise the C2-C3 bond of 1,4-glucans forming two reactive aldehyde groups. These aldehydes can subsequently react with other hydroxyl groups in the fibre during drying to form hemiacetal linkages (Zeronian et al. 1964; Back 1967; Ghosh and Dalal 1988). The oxidation of fibres with periodate has been shown to cause an unevenly distributed oxidation and also to reduce the crystallinity of the cellulose (Kim et al. 2000). Zeronian et al. (1964) studied the influence of periodate oxidation of fibres on the mechanical properties of paper sheets. They found that both the dry and wet strength increased with increasing degree of oxidation up to a maximum and then decreased on further oxidation.

20

Experimental

Materials

Fibres Never-dried softwood kraft pulp fibres (supplied by StoraEnso Skoghall Mill, Sweden) that had been oxygen-delignified to kappa 18 and refined to a dewatering resistance corresponding to 22 SR were used in Paper I. Never-dried unbleached softwood kraft fibres (supplied by Kappa Kraftliner, Piteå, Sweden) that had been cooked to kappa 76 and Escher-Wyss beaten to 30 M°SR (corresponding to about 16 SR) were used in Papers II, III and VI. Never-dried unbeaten softwood kraft pulp fibres (supplied by SCA, Östrand Mill, Sweden) bleached according to a (OO)Q(OP)(ZQ)(PO) sequence were used in Papers IV och VI. Never-dried laboratory-pulped spruce wood kraft fibres with a yield of 49.7% and kappa number of 34 were used in paper V. In order to prepare a pulp that is suitable for evaluating the influence of fibre properties on sheet properties, it is necessary to remove most of the fines material from the pulp. In Papers I, II, III, and IV, the fines were removed from the pulp by successive spraying through a spray disk filter fitted with a plastic wire with 75 µm openings. In Paper V, the fines were removed using a Britt Dynamic Drainage Jar according to the Tappi T 261 cm-94 standard. The long fibre fraction was washed at both high and low pH in order to remove most of the remaining adsorbed metal ions and dissolved and colloidal material. In Paper VI, fines material was not removed from the pulp, nor was the pulp washed at low and high pH.

21

Chemicals Polyallylamine hydrochloride (PAH) with molecular masses of 15 kDa and 70 kDa and polyacrylic acid (PAA) with a molecular mass of 8 kDa were purchased from Sigma-Aldrich, Sweden. Polyallylamine with a molecular mass of 150 kDa was kindly provided by Nittobo Boseki, Japan. Cationic potato starch was supplied by Lyckeby Stärkelsen, Kristianstad, Sweden and had a degree of substitution of cationic groups of 0.065. The potassium polyvinyl sulphate (KPVS) used for polyelectrolyte titration was purchased from Waco Pure Chemical Industries, Japan. Ortho-toluidine blue (VWR, Sweden) was used as an indicator during the titration. Fluorescein isothiocyanate (FITC), used for labelling PAH, sodium metaperidoate used for fibre cross-linking and hydroxylamine hydrochloride used for carbonyl content determination were all purchased from Sigma-Aldrich, Sweden. The hydrochloric acid, sodium hydroxide, sodium bicarbonate and sodium chloride were all of analytical grade.

Methods

Polyelectrolyte adsorption It is extremely important to control the pH during the adsorption of weak polyelectrolytes such as PAH and PAA, because changes in pH and thus changes in polyelectrolyte charge density influence both the amount adsorbed and the polyelectrolyte conformation (Wågberg 2000). Consequently the pH was carefully controlled to pH 8 during all the adsorption experiments in Papers I, II and V. All the adsorptions were conducted at a fibre concentration of 5 g/L. To determine how much of the added PAH was adsorbed, the nitrogen content of the fibres was determined using an elemental analyser (ANTEK 7000, Model 737). By testing small amounts of fibres or sheets, it was possible to determine the amount of nitrogen and hence the adsorbed amount PAH from prepared calibration curves.

22

Visualising the location of adsorbed polyelectrolyte In order to qualitatively determine the distribution of the adsorbed polyallylamine, confocal scanning laser microscopy was used to obtain images of thin optical sections of fibres saturated with fluorescently labelled polyallylamine. PAH was labelled with fluorescein isothiocyanate (FITC) using a general protocol developed for biological macromolecules (Hermanson 1996). The charge density of the labelled polymers was checked by polyelectrolyte titration, and no differences compared to the unlabelled polymers could be detected. After adsorption of labelled PAH, the fibres were washed in order to remove weakly bound polyelectrolyte. The fibres were then immediately frozen using liquid nitrogen. The frozen fibre sample was then freeze-dried in order to dry the fibres without collapsing the lumen. A Bio-Rad Radiance 2000 confocal system mounted on a Nikon Eclipse 800 microscope was used to obtain images of thin optical sections of the fibres. A Krypton Argon laser was used for excitation at 488 and 568 nm. Images of the fibres were taken using a 100x N.A. 1.4 oil-immersion lens.

Fibre cross-linking Fibres suspended in de-ionised water was oxidised by adding different amounts of sodium metaperiodate for different reaction times (see Paper III and IV for the exact dosages and reaction times). The periodate oxidises the C2-C3 bond of cellulose forming two aldehyde groups that can cross-link with adjacent hydroxyl groups during drying as outlined in Figure 1 (Back 1967; Ghosh and Dalal 1988). The oxidation reaction was stopped by dewatering the fibres in a Büchner funnel fitted with a filter paper and repeatedly washing with de-ionised water until the conductivity of the filtrate was below 5 μS/cm. The content of carbonyl groups in the fibres was determined by the hydroxylamine hydrochloride method (Zhao and Heindel 1991; Vicini et al. 2004). Hydroxylamine hydrochloride reacts quantitatively with carbonyls in the fibres to form the corresponding oximes, releasing an equivalent amount of hydrochloric acid. The amount of hydrochloric acid released and hence the carbonyl content can easily be determined by a simple potentiometric neutralisation titration.

23

Figure 1. Schematic representation of how the periodate ion oxidises the C2-C3 bond of the cellulose into dialdehyde cellulose followed by a possible mechanism for the subsequent cross-linking reaction. (Papers III and IV)

Sheet preparation The majority of the sheets prepared in this work were isotropic sheets made in a Rapid-Köthen sheet preparation apparatus. These sheets were dried under restrained conditions at 93°C. In Paper I, some sheets were further heat-treated at 160°C for 15 minutes. In Paper III, thin anisotropic sheets with a grammage of 20 g/m2 were prepared on a dynamic sheet former (Formette Dynamique). These sheets were roll pressed, restrained dried at ambient conditions and then post-dried at 93°C for 15 minutes.

Paper testing Dry tensile testing was carried out according to the SCAN P:67 standard for the tensile testing of laboratory-made sheets. The thickness of the prepared sheets was measured as structural thickness (Schultz-Eklund et al. 1992) and was used to calculate the apparent density.

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

24

Creep testing In Papers I and II, creep was measured under a tensile load at constant climates of 50% RH and 90% RH using an apparatus developed at STFI (now Innventia AB), Stockholm, Sweden. A detailed description of the apparatus can be found elsewhere (Panek et al. 2004). The applied load and the resulting strain were monitored as a function of time for 100 or 300 seconds. The creep behaviour was evaluated by means of isochronous stress-strain curves (Kolseth and de Ruvo 1983; Haraldsson et al. 1994; Panek et al. 2004). An isochronous stress-strain curve indicates how much stress needs to be applied to achieve a certain creep strain at a specific time. The mechano-sorptive creep of the sheets with a grammage of 140 g/m2

prepared in Paper III was tested in compression using the same apparatus. The apparatus enables the creep to be measured in compression since the paper specimen is prevented from buckling by supporting columns. The test employed three 50 to 90% RH cycles with each cycle being seven hours long and having a ramp time of approximately 20 minutes. The result was subsequently analysed using isocyclic stress-strain curves as proposed by Panek et al. (2004). An isocyclic stress-strain curve is constructed from the measured strain versus time data as shown in Figure 2, and gives the relation between stress, total strain and number of humidity cycles. When studying mechano-sorptive creep, it is important to consider that paper can permanently shrink when first exposed to cyclic humidity. The magnitude of the shrinkage decays with increasing number of cycles until no further permanent shrinkage can be detected. Different terms for this phenomenon have been used in the literature, but the term “release of dried-in strains” seems to describe the situation best. Without proper preconditioning, the strain measured in a mechano-sorptive creep test will be a combination of creep strain and release of dried-in strains. Accordingly, all the samples used for mechano-sorptive creep testing were preconditioned by exposing them to six 50 to 90% RH cycles prior to testing.

25

Figure 2. Construction of an isocyclic stress-strain curve for three humidity cycles, from strain versus time data. The applied load and sheet grammage is used to calculate the specific stress and combined with the strain measured at the end of the third humidity cycle.

-0.2

-0.1

0

0.1

0.2

0.3

0 200 400 600 800 1000 1200 1400

Stra

in (%

)

Time (min)-4

-3

-2

-1

0-0.2-0.15-0.1-0.050

Spec

ific

Stre

ss (k

Nm

/kg)

Strain (%)

Specific stress is calculatedfrom grammage and applied load

Low load

High load

RH

26

The mechano-sorptive creep behaviour of single fibres and thins sheets (20 g/m2) was examined using the method developed by Olsson et al. (2007). Prior to testing, all the individual fibres were heated in an oven at 93°C for 15 minutes, so that they had received the same heat treatment as the fibres in the sheets. Dried single fibres and sheet strips were loaded in tension in a Perkin Elmer dynamic mechanical analyser (DMA). The samples were first loaded under constant humidity (80% RH) for 3 hours to establish the creep rate at constant humidity. Thereafter the humidity was cycled 10 times between 80 and 30% RH, each cycle being one hour long, to establish the creep rate at varying humidity. By normalising the creep rate at cyclic humidity with respect to the creep rate at constant humidity, it was possible to eliminate the experimental scatter due to the large difference between individual fibres. This normalised value gives a good measure of how much the creep is accelerated by varying humidity compared with the creep under constant humidity conditions.

Moisture sorption measurements In Paper III, the moisture up-take during the humidity cycling used for the mechano-soprtive creep testing was continuously recorded by a balance (Sartorius BP 110 S) connected to a PC. The samples used to study the moisture up-take were, like the samples used for mechano-sorptive creep testing, preconditioned by exposure to six 50 to 90% RH cycles. In Paper IV, a dynamic vapour sorption equipment (DVS) from Surface Measurement Systems Ltd. was used to obtain near equilibrium sorption isotherms and to study the sorption kinetics at a temperature of 33±2°C. To achieve the desired relative humidity, dry and water vapour saturated air currents were mixed in appropriate proportions.

27

Determination of residual stresses The through-thickness distribution of residual stress in the plane of the paper was determined by a layer removal method published previously (Östlund et al. 2005). Thin layers of the paper were removed by surface grinding, which changes both the stress distribution and the bending stiffness of the substrate, resulting in a change of curvature of the substrate. By grinding paper specimens to the middle, from alternate sides, and measuring the curvature versus the grinding depth, the stress distribution in the original specimen could be calculated. Since the sheet preparation was symmetrical, the stress state was assumed to be equibiaxial.

Preparation of specimens for shear testing Both sheets made of never-dried fibres and rewetting of dry sheets were used for specimen preparation. In the case of specimens made from rewetted sheets, papers with a grammage of 300 g/m2 were first prepared and dried in the Rapid-Köthen equipment. The paper sheets were then cut into 15 mm wide strips, which were subsequently soaked in water for 2 hours. The wet paper strips were then arranged together as shown in Figure 3, with an overlap of 15 mm. By configuring the samples as shown in the side-view of the specimens in Figure 3 a relatively pure shear stress field was created along the contact zone between the two strips when the samples were loaded in tension, as shown by the calculations presented by Nygårds et al. (2009a). The specimens were subsequently dried for various times in the Rapid-Köthen dryers. The specimens prepared were tested as described below. In the case of sheets of never-dried fibres, sheets with a grammage of 300 g/m2 or 500 g/m2 were prepared in the Rapid-Köthen sheet former. The wet paper sheets were cut in half and arranged together according to Figure 3, with the sheets overlapping 5 or 15 mm. The sheets were then dried for various times in the Rapid-Köthen dryers. After being dried, the moist sheets were cut into 15 mm wide strips and tested as described below. The specimens were tensile tested on a horizontal tester. The measured peak load was combined with the overlap area to calculate the shear strength. Three to ten specimens were tested for each drying time. Immediately after testing, the solids contents of the samples were determined.

28

Figure 3. Schematic of how the shear testing specimens were prepared. The overlap between the strips was 15 mm except for the high-dry-content samples made of never-dried fibres, in which the overlap was reduced to 5 mm. (Paper VI)

29

Results and discussion

Fibre modification with polyelectrolytes Fibres were modified with polyelectrolytes in Papers I, II and V. In Papers I and II, the effect of adsorbing polyallylamine (PAH) on the creep of resulting hand-sheets was studied. As described in the Background, the creep of paper is primarily considered to be determined by the properties of the fibres themselves and not by the fibre/fibre joints. Hence, the effects of modifying fibres surfaces alone or together with fibre walls were studied in Paper II. In Paper V, the influence of polyelectrolyte multilayers of polyallylamine and polyacrylic acid on the residual stresses in hand-sheets was studied. In order to determine whether the adsorbed polyelectrolyte could penetrate throughout the fibre walls, a technique involving labelling of PAH with a fluorescent dye and examination of single fibres in a confocal laser scanning microscope was used (Papers I and II). Charge density measurements showed no difference between labelled and unlabelled PAH which was also to be expected considering the low degree of substitution of fluorophores (< 0.01). Thus the presence of the fluorophores should have no significant effect on the polyelectrolyte conformation or the adsorption behaviour (Tanaka et al. 1990). Hence, the images obtained reflect how the unlabelled polyallylamine was adsorbed to the fibres. Two different adsorption conditions were studied: a short adsorption time (30 minutes) at low ionic strength (5·10-3 M NaHCO3) and a long adsorption time (24 hours) at high ionic strength (5·10-3 M NaHCO3 + 10-1 M NaCl). The results for the two cases are presented and discussed separately below.

Low ionic strength The idea behind using a low ionic strength and a short adsorption time (5·10-3 M NaHCO3 and 30 minutes) was that a low ionic strength gives the polyelectrolyte molecules an extended conformation. It is thus improbable that they have access to the interior of the porous fibre wall. Figure 4 summarises the results of the confocal laser scanning microscopy of fibres with a kappa number of 76 (Paper II), where, it is seen that the PAH

30

molecules were adsorbed only on the surfaces of all the fibres examined. In no case were the molecules able to diffuse into the lumen. For thin-walled earlywood fibres, however, the thickness of the adsorbed layer was close to the thickness of the fibre wall. Thus, in this case, the adsorption was not purely limited to the fibre surfaces. Since the adsorbed layer had a considerable thickness, it is more appropriate to call it an adsorption to the external parts of the fibre wall. Only the results for the adsorption to fibres with a kappa number of 76 are shown here. The results for fibres with a kappa number of 18 were however very similar (see Paper I).

Figure 4. CLSM micrographs of radial cross-sections of fibres (Kappa 76) with fluorescently labelled PAH adsorbed at 5·10-3 M NaHCO3 for 30 minutes. The red and green pictures are obtained by excitation at 488 nm and 568 nm respectively. At 568 nm there is auto-fluorescence from the lignin in the fibre wall, so that the thickness of the fibre cell wall is shown. The PAH molecules could reach only the external parts of the fibres, however, a larger fraction of the fibre wall was reached by the PAH molecules in the early wood fibres than in the latewood fibres. (Paper II)

31

High ionic strength The idea of using a high ionic strength (5·10-3 M NaHCO3 + 10-1 M NaCl) and a long adsorption time (24 hours) was that a high ionic strength gives the polyelectrolyte molecules a coiled conformation and reduced size, so that they were thus more likely to be able to access the entire porous fibre wall. A high ionic strength screens the electrostatic interaction between the charged moieties along the polymer backbone. The polyelectrolyte molecules thus coil up and the effective size is reduced compared to that at low ionic strength. High ionic strength was combined with a long adsorption time and, as seen in Figure 5, this did in fact allow the PAH molecules to reach throughout the fibre wall. To summarise, it has been clearly demonstrated that adsorption at low ionic strength for a short time resulted in an adsorption to the external parts of the fibres, whereas adsorption at high ionic strength for a long time resulted in an adsorption throughout the cell wall. Thus any impact on the mechanical properties of the fibres should be considerably greater in the case of adsorption at high ionic strength.

Figure 5. CLSM micrographs of radial cross-sections of fibres (Kappa 76) with F-PAH adsorbed at 5·10-3 M NaHCO3 + 10 -1 M NaCl for 24 hours. The red and green pictures are obtained by excitation at 488 nm and 568 nm respectively. The adsorbed PAH molecules reached throughout all the examined fibres. (Paper II)

32

Kinetics of fibre wall penetration The above results established the conditions suitable for the adsorption of PAH to the fibre exterior or throughout the fibre wall. However, the results did not reveal whether it was the long adsorption time or the high ionic strength that gave fibre wall penetration. Nor did the experiment yield any information about the migration kinetics at high ionic strength. In order to achieve a better understanding of the adsorption process, the ionic strength and the adsorption time were varied in smaller steps. The migration of polyelectrolyte molecules into the fibre wall was easier to study in thick-walled latewood fibres and consequently only such fibres were examined in this case. Figure 6 shows that a combination of both high ionic strength and long adsorption time was required to achieve full fibre wall penetration. Dynamic laser light scattering analysis of the polyelectrolyte showed that the 15 kDa PAH had a hydrodynamic radius of approximately 6 nm at 1 M NaCl. The pore radius in unbleached kraft pulp fibres ranges from 13 nm to 17 nm according to NMR relaxation measurements (Andreasson et al. 2005). These two facts together indicate that the PAH molecules should be able to enter the porous fibre wall, from a pure geometrical point of view. Calculations suggested that the PAH molecules were also small enough at lower ionic strengths to be able to enter the fibre pores. However, Figure 6 shows that, despite the high ionic strength, penetration into the fibre wall did not occur at a fast rate, but was a rather slow process since an adsorption time between 3 and 24 hours was needed. The kinetics of the migration process seemed to be controlled not by polyelectrolyte size but rather by chain flexibility and interaction with the anionic fibre, which is in accordance with recently published results for similar polyelectrolytes (Horvath et al. 2008b; Horvath et al. 2008c). Chain flexibility increases with increasing ionic strength and the interaction with the anionic fibre decreases. An increase in chain flexibility and a decrease in the electrostatic interaction with the fibre wall material would speed up the migration. Since this type of behaviour was found in Figure 6, this hypothesis seems probable. In Figure 6, it is also evident that, for certain adsorption conditions, adsorption to the fibre exterior is associated with a high average adsorbed amount as measured by nitrogen analysis of the fibre material. Since a high adsorbed amount suggests that the polyelectrolyte molecules have access to

33

the fibre wall, this result seems to be somewhat contradictory. This can, however, probably be explained by the fact that under these conditions the polyelectrolyte did have access to the fibre wall of thin and medium-thick walled fibres but not to the fibre wall of thick latewood fibres (see Paper II). Access to the fibre wall of thin and medium-thick fibres may account for the high average adsorbed amount.

Figure 6 Micrographs of axial cross-sections of latewood fibres saturated with F-PAH at different ionic strengths and for different times. The images were recorded at an excitation wavelength of 568 nm so that the thickness of the fibre cell wall is shown by the auto-fluorescence from the lignin. The figure includes the adsorbed amounts measured by nitrogen analysis. The 95% confidence intervals for these values are approximately ± 8 mg/g. (Paper II)

34

Influence of polyelectrolytes on creep and mechano-sorptive creep

Adsorption to the exterior parts of the fibre wall At a low ionic strength and a short adsorption time, polyallylamine molecules were preferentially adsorbed onto the exterior parts of the fibre wall. This fact was utilised to study the effect which polyallylamine present only on the exterior parts of the fibre wall had on the mechanical properties of the sheets. Sheet strength is primarily determined by the strength of and the number of fibre/fibre joints per sheet volume (Davison 1972). As mentioned in the introduction, polyelectrolytes usually increase the sheet strength by strengthening the fibre/fibre joints and the number of efficient joints, at least in low density sheets. Sheet strength is thus a good estimate of the joint-strengthening ability of an additive.

Figure 7. Tensile strength of the sheets produced from the two pulps with different kappa number, with and without additives. Initially, the two pulps had slightly different strengths. After the addition of polyelectrolytes there was, however, no significant difference in strength between the sheets. (Paper I and II)

50

70

90

110

-5 0 5 10 15 20 25 30

Adsorbed Amount (mg/g fibre)

Tens

ile S

tren

gth

Inde

x (k

Nm

/kg)

Cationic Strach - Kappa 18PAH 150k - Kappa 18PAH 15k - Kappa 18Reference - Kappa 18PAH 150k - Kappa 76PAH 15k - Kappa 76Reference - Kappa 76

35

The tensile strengths of the different samples are presented in Figure 7. The paper sheets produced from the oxygen-delignified kappa 18 pulp were slightly weaker than the paper sheets produced from the high-yield unbleached kappa 76 pulp. After the adsorption of polyelectrolytes, however, they reached the same strength, and there was no significant difference between the effects of starch and polyallylamine, at least for the adsorbed amounts studied here. This indicates that the numbers and strengths of fibre/fibre joints in all the sheets with added polyelectrolyte were similar. The viscoelasticity of the sheets was evaluated by creep testing. The creep strain was recorded as a function of time for the different samples at different load levels using the specially designed creep-testing equipment. The deformation response in a creep test can be divided into instantaneous, delayed elastic and permanent creep. The instantaneous response is an idealisation, since it always takes some time for deformation to develop. In this study, the instantaneous response was evaluated as the deformation detected after one second. In Figures 8 to 11, the delayed creep strain, calculated as the difference between the strain at 100 seconds and the strain at 1 second, is shown for different load levels at 50% and 90% RH. Earlier studies have shown that the creep of strong, i.e. efficiently loaded, paper sheets is not influenced by the properties of the fibre/fibre joint (Parker 1962; DeMaio and Patterson 2005). The sheets produced in this study have a high strength and a high density, and should thus be efficiently loaded structures. Nevertheless, it is seen in Figures 8 and 9 that the creep under different loads at both 50% and 90% RH was significantly reduced by the addition of polyallylamine. The filled and unfilled symbols, which indicate whether or not the samples were heat-treated at 160ºC for 15 minutes, in Figures 8 and 9 intermix randomly, and it can hence be concluded that this extra heat-treatment had no significant effect on the creep behaviour of the samples.

36

Figure 8. Creep deformation during 100 seconds at 50% RH for sheets made from kappa 18 fibres with different polyelectrolytes adsorbed to the exterior parts of the fibre wall. Filled symbols indicate heat-treated and unfilled non-heat-treated samples. The adsorbed PAH reduced the creep deformation significantly. (Paper I) Starch had a small effect on the creep at 50% RH, but at 90% RH no effect at all was seen. The sheets with adsorbed starch or polyallylamine were definitely efficiently loaded structures, and had a similar tensile strength and thus a similar state of fibre/fibre bonding, but they nevertheless showed differences in creep behaviour. This difference in viscoelastic behaviour cannot therefore be explained by the concept of efficiency factors. The results also show that, even if different polyelectrolytes have the same effect on the maximum strength, they can still have different effects on the deformational behaviour prior to failure. The exact molecular mechanisms behind these differences are still not known.

0

10

20

30

40

0 0.1 0.2 0.3 0.4

ε(100 s) - ε(1 s) (%)

Spec

ific

Stre

ss (k

Nm

/kg)

PAH 150 kDa 6.7 mg/g PAH 150 kDa 3.3 mg/g PAH 70 kDa 10.0 mg/g PAH 70 kDa 6.6 mg/g PAH 15 kDa 15.2 mg/g PAH 15 kDa 9.1 mg/g Starch 16.3 mg/g Reference

37

Figure 9. Creep deformation during 100 seconds at 90% RH for sheets made from kappa 18 fibres with different polyelectrolytes adsorbed to the exterior parts of the fibre wall. Filled symbols indicate heat-treated and unfilled non-heat-treated samples. The adsorbed PAH reduced the creep deformation significantly. (Paper I)

In Figures 10 and 11, the delayed deformation during 300 seconds long creep tests is shown for sheets made from fibres with a kappa number of 76. Here, it is seen that the added PAH had no significant effect on the creep at 50% RH. At 90% RH however, the addition of 15 kDa PAH reduced the delayed deformation significantly. The 150 kDa PAH had no significant effect, however, in any of the examined climates. There are thus significant differences in effect between the kappa 76 pulp and the kappa 18 pulp, since for the kappa 18 pulp, significant reductions in creep were found at both 50% and 90% RH and for all the different molecular weights of PAH.

0

3

6

9

12

15

0 0.1 0.2 0.3 0.4

ε(100 s) - ε(1 s) (%)

Spec

ific

Stre

ss (k

Nm

/kg)

PAH 150 kDa 6.7 mg/g PAH 150 kDa 3.3 mg/g PAH 70 kDa 10.0 mg/g PAH 70 kDa 6.6 mg/g PAH 15 kDa 15.2 mg/g PAH 15 kDa 9.1 mg/g Starch 16.3 mg/g Reference

38

The only reasyield of the twlignin contentway in whichfibre/fibre joinof reducing crthe potential fadded in the ri

Figure 10. Creepfibres with PAH wall. The solid linsets. The broken loverlap, there wastreated samples. (P

sonable explanawo pulps. Due ts and differenth the adsorbent properties. Teep under cons

for significant imight way.

p deformation durinH of two different m

nes are the model plines are 95% confis no significant diffePaper II)

ation of these dto the differen

t porosities, anded polyallylaminThese results shstant humidity wmprovements p

ng 300 seconds at 5molecular weights adproposed by Panek idence limits for the ference in delayed def

differences is tnce in yield, thed this could pone affects the

how the complewith chemical adprovided the rig

50% RH for sheets mdsorbed to the exteret al. (2004) fittedpredicted fits. Since

eformation between

the difference iey have differenossibly affect the fibre and thexity of the topidditives, but alsght chemicals ar

made from kappa 7rior parts of the fibd to the different dae the confidence limithe reference and t

in nt he he ic so re

76 bre ata its the

Figure 11. Creepfibres with 15 kDmodel proposed byconfidence limits fodeformation for stsignificant effect (d

To summariseto the exterioThe effect howused, and thehowever, diffibehind the red The fact that tthe fibres sugaffected by thcreep behavioPrior literaturePatterson 200

p deformation durinDa PAH adsorbed y Panek et al. (200for the predicted fitstresses > 8 kNm/data not shown). (Pa

e, the creep testor of fibres inflwever dependee type and moicult to draw anduction in the cr

the polyelectrolggests that it whe adsorbed polour was due toe on the topic (P5) has, howeve

ng 300 seconds at 9to the exterior parts04) fitted to the tw. The addition of 1/kg. The addition aper II)

ts clearly showeluenced the vised on several faolecular weightny decisive conreep.

lyte molecules wwas primarily thlyelectrolyte. Tho changes in thParker 1962; Se

er, shown the im

0% RH for sheets ms of the fibre wall. T

wo data sets. The b15 kDa significantly

of 150 kDa PA

ed that polyallyscoelasticy of thactors such as tt of the polyeclusions about

were adsorbed the fibre/fibre his indicates thhe fibre/fibre eth and Page 19mportance of t

3

made from kappa 7The solid lines are tbroken lines are 95ly reduced the delay

AH, however, had n

ylamine adsorbehe paper sheetthe type of fibrelectrolyte. It ithe mechanism

to the exterior ojoints that wer

hat the change ijoint propertie

981; DeMaio anhe fibre wall fo

39

76 the %

yed no

ed ts. re is,

ms

of re in

es. nd or

40

the viscoelastic behaviour of strong and fully efficiently loaded paper sheets. As was shown by the fluorescent labelling technique, the polyallylamine molecules reached a considerable part of the fibre wall in the case of thin earlywood fibres. Taken together, this suggests that a change in the fibre wall properties in thin walled fibres could be the cause of the difference in creep behaviour, but before any absolute conclusions can be drawn, it is necessary to measure the mechanical properties of individual fibres, with and without additive.

Adsorption throughout the fibre wall The microscopy studies of the fibres with polyallylamine adsorbed at high ionic strength for a long time showed that polyelectrolyte molecules were present throughout the entire fibre wall of the fibres. This means that the adsorbed polyallylamine had a significantly greater potential to affect the properties of the fibre cell wall than when the polyallylamine was adsorbed only to the exterior parts of the fibre wall. This was utilised to study the effect which polyallylamine had on the mechanical properties of the sheets when adsorbed throughout the fibre wall. The result of the tensile testing of the sheets is presented in Table 1, where it is evident that the strength of the polymer-treated sheets was at the same level as that of the reference. This is somewhat unexpected, since the results for the samples treated at low ionic strength clearly showed that the presence of polyallylamine on the exterior part of the fibre wall resulted in an increase in sheet strength. The absence of any strength gain can however be explained by the fact that cationic polyelectrolytes are known to deswell cellulosic fibres so that both the fibre/fibre joint strength and the molecular contact area in each joint decreases (Torgnysdotter et al. 2007). Fibre swelling and surface flexibility are very important for sheet consolidation. Swollen and flexible fibres give an efficient packing and good fibre/fibre contact, which is a prerequisite for the formation of strong fibre/fibre joints. Swerin et al. (1990), found that a low molecular mass polyelectrolyte that has access to all fibre charges gives greater deswelling than a high molecular mass polyelectrolyte that has access only to fibre surface charges. This means that the fibre deswelling was probably less when polyallylamine was adsorbed to the fibre exterior than when it was adsorbed throughout the fibre wall. It can hence be suggested that when adsorption occurred only to the exterior parts of the

41

fibre wall, the loss in fibre surface flexibility was compensated for by the increase in fibre/fibre joint strength given by the adsorbed polyallylamine, and the sheet strength consequently increased. When the polyallylamine was adsorbed into the fibre wall, however, the loss in fibre swelling was so severe that it could not be fully compensated for by the strengthening effect of PAH on the fibre/fibre joints. Accordingly, the sheet strength did not increase when PAH was adsorbed throughout the cell wall of the cellulosic fibres.

Table 1. The result of tensile testing of sheets made from fibres with PAH adsorbed at high ionic strength. The sheets were not significantly weaker or stronger than the reference sheets. (Paper II)

Reference PAH 15 kDa (High ionic strength)

Tensile Strength Index (kNm/kg) 74.5 ± 1.3 73.4 ± 1.8

Strain-at-Break (%) 2.87 ± 0.13 2.69 ± 0.09

Tensile Stiffness Index (MNm/kg) 7.54 ± 0.11 7.50 ± 0.10

The effect of the adsorbed PAH on the delayed deformation during creep tests performed at 90% RH is shown in Figure 12. The creep actually increased, which refutes the initial hypothesis that the creep would decrease if the fibre walls were treated with polyallylamine. The creep at 50% RH was not however significantly affected by the adsorbed PAH. It is suggested that the increased creep at 90% RH can also be explained by the deswelling and thus less effective consolidation induced by the adsorbed PAH. The decreased consolidation results in fewer fibre/fibre joints per sheet volume and also a lower molecular contact area in each joint which means a less efficient distribution of stresses in the fibre-network. The less efficient stress distribution implies that some fibres will be more stressed than the fibres in the reference sheets, and this will in turn explain the observed increase in creep. The initial hypothesis was that the creep of individual fibres would be decreased if the fibre cell wall were treated with polyallylamine. The results presented here, however, provide no conclusive indication of whether the creep of the individual fibres was increased or decreased by the adsorbed polyallylamine. Even if it is suggested that the increased creep was the

42

result of fewerthe individualmentioned eafibres would regarding this.

Figure 12. Creepfibres with 15 kDproposed by Panekconfidence limits fdeformation for sp

r fibre/fibre conl fibres was inrlier, measurembe beneficial t

p deformation durinDa PAH adsorbed k et al. (2004) fittefor the predicted fitsecific stress > 5 kN

ntact points, it increased by thements of mechto enable defin

ng 300 seconds at 9throughout the fibre

ted to the two differes. The addition of P

Nm/kg. (Paper II)

is still possible e adsorbed po

hanical propertinitive conclusio

90% RH for sheets me cell walls. The solent data sets. The bPAH significantly

that the creep oolyelectrolyte. Aies of individu

ons to be draw

made from kappa 7lid lines are the modbroken lines are 95increased the delay

of As al

wn

76 del %

yed

43

Mechano-sorptive creep The initial hypothesis was that the creep at 90% RH must be reduced in order to reduce mechano-sorptive creep, and for this reason the creep at constant relative humidities of 50 and 90% was initially investigated. As shown above, the adsorption of polyallylamine to the fibre surfaces reduced the creep at constant relative humidities of 50 and 90% (especially for the kappa 18 pulp used in Paper I). Therefore sheets with adsorbed PAH were tested for mechano-sorptive creep. Figure 13 shows that there was no significant difference in mechano-sorptive creep during three humidity cycles between the reference sheets and the sheets with adsorbed PAH. This means that, even though PAH can be used to reduce creep at constant humidity, the use of PAH does not seem to be a feasible way to reduce mechano-sorptive creep under a tension load. It is not likely that other cationic polyelectrolytes commonly used to increase paper strength could reduce mechano-sorptive creep. However, it is possible that specifically tailored polyelectrolytes, with for example cross-linking capabilities, might prevent moisture sorption and fibre swelling, and could in the future be used to reduce mechano-sorptive creep.

Figure 13. Isocyclic stress-strain data recorded after three 7 hour long 50 – 90% RH cycles. The sheets with adsorbed PAH were heat-treated at 160°C. As seen there was no significant difference in mechano-sorptive creep during three humidity cycles between the reference sheets and the sheets with adsorbed PAH when loaded in tension.

0

2

4

6

8

0 0.2 0.4 0.6 0.8 1 1.2

Spec

ific

Stre

ss (k

Nm/k

g)

Strain (%)

Reference

PAH 70k 14mg/g

PAH 15k 30mg/g

44

That no reduction of mechano-sorptive creep could be seen could be explained by the results of Panek et al. (2005). In their study, pulps were beaten in different ways to obtain different fibre shapes and structural densities. Figure 14 is reproduced from their work and shows the relationship between tensile stiffness index and creep stiffness index for their pulps. When the tensile stiffness index was increased by beating, there was a substantial increase in isochronous creep stiffness index at 50% RH, but the corresponding increase in isocyclic creep stiffness index was significantly less (as shown by the slope of the top and bottom curves in Figure 14). In fact, the increase in isocyclic creep stiffness index was approximately 22% of the increase in isochronous creep stiffness index. The isochronous creep stiffness index at 50% RH for the kappa 18 pulp increased by approximately 23% when PAH was added to the fibre surfaces. If it is assumed that the same relationship between improvement in isochronous and isocyclic creep stiffness index as found by Panek et al. (2005) holds for fibre modification with PAH, the equivalent increase in isocyclic creep stiffness index would be 5%. Considering the experimental scatter of the mechano-sorptive creep data, it is unlikely that such a small improvement as 5% would have been seen in Figure 13. It is not certain that the relationship found by Panek et al. (2005) holds for fibre modification with PAH, but these results indicate that reductions in creep at constant 50 and 90% RH must be large to have a significant effect on the mechano-sorptive creep.

45

Figure 14. Data reproduced with permission from (Panek et al. 2005). In this study, pulps were beaten in different ways to obtain different fibre shapes and structural densities. The figure shows the relationship between tensile stiffness index and creep stiffness index for these pulps. When the tensile stiffness index was increased by beating, there was a substantial increase in isochronous creep stiffness index at 50% RH, but the corresponding increase in isocyclic creep stiffness index was significantly less.

46

Influence of polyelectrolytes on residual stresses Polyelectrolyte adsorption was used to study not only the effect on creep but also the effect on residual stresses. The effect of adding a single layer of polyallylamine (PAH) or a multilayer of polyallylamine (PAH) and polyacrylic acid (PAA) compared to beating of the fibres was studied in Paper V. The polyelectrolytes were adsorbed in a 10-2 M NaHCO3 buffer for 15 minutes. As shown in Figure 6, the treatment with PAH in 10-2 M NaCl resulted in adsorption to the exterior of the fibres so the multilayers prepared in Paper V should only have been formed on the exterior parts of the fibres. The effect on the tensile strength index of the prepared laboratory sheets is presented in Figure 15. The addition of polyelectrolytes increased the strength of the sheets considerably, in fact more than could be achieved by beating, without any significant densification. The fact that the tensile strength index increased so significantly without densification demonstrates the possibility of producing paper possessing both high strength and bulk. The fact that the strength increased without densification also indicates that the addition of the current additives caused no significant flexibilising or collapse of the fibres during drying, since this would have resulted in densification (Samuelsson 1964; Page 1985). Previous BET measurements and fibre contact zone analysis of sheets treated with PAH/PAA multilayers (Eriksson et al. 2006) suggest that the main effect of the additives was to increase the number of fibre/fibre joints, the molecular contact area in each joint and the fibre/fibre joint strength. The results also indicate that fibre flexibilisation is not necessary to obtain high paper strength, but that the surface properties of the fibres largely determine the strength of the fibre/fibre joint and the resulting paper. Such results and ideas have been previously described by, for example, Wågberg et al. (2002), Laine et al. (2003), and Torgnysdotter and Wågberg (2004).

47

Figure 15. Tensile strength index for sheets made of beaten fibres or fibres with adsorbed polyelectrolytes plotted as a function of the apparent density. (Paper V)

The effect of beating and polyelectrolyte addition on the residual stress state was determined according to the above-described procedure and the results are shown in Figure 16. The residual stresses were small in the reference paper, which agrees with previous results for paperboard made from unbeaten pulp (Östlund et al. 2004a). Nevertheless, the stress was definitely compressive at the surface of the reference paper, as in the stronger papers. This is typical of paper after drying, as indeed of many plastics after cooling. The magnitude of the residual stresses increased with the number of both beating revolutions and adsorbed polyelectrolyte layers. Figure 17 shows the residual stress amplitude, i.e., the absolute value of the surface stress at the side where it was the highest, as a function of the sheet modulus. In the method of residual stress determination, the stiffness is simply a scaling factor applied to the measured deformation and geometry data. It is thus natural for the influence on the stiffness of beating and chemical additives to be manifested also in the residual stress amplitude. However, as seen in Figure 17, the residual stress amplitude did not scale linearly with the stiffness in the prepared sheets, as the data points would then have fallen on the broken line in Figure 17.

No beating, no added polyelectrolyte

PAH

Multilayer PAH/PAA/PAH

500 revs

5000 revs

7000 revs

40

50

60

70

80

90

100

110

120

600 650 700 750 800 850 900

Tensile

 strength inde

x (kNm/kg)

Density (kg/m3)

No beating, no added polyelectrolyte

Beating

Chemistry

48

Figure 16. Distribution of in-plane residual stress in the thickness direction for a) papers with added polyelectrolyte and b) papers made from pulps beaten to different degrees. (Paper V)

‐12

‐10

‐8

‐6

‐4

‐2

0

2

4

‐150 ‐100 ‐50 0 50 100 150

In‐plane

 stress (M

Pa)

Position in thickness direction (µm)

Chemistry

No chemistry

a)

PAH

Multilayer PAH/PAA/PAH

‐12

‐10

‐8

‐6

‐4

‐2

0

2

4

‐150 ‐100 ‐50 0 50 100 150

In‐plane

 stress (M

Pa)

Position in thickness direction (µm)

500 revs

Beating5000 revs

No beating

7000 revs

b)

49

Figure 17. Residual stress amplitude plotted as function of sheet stiffness. The broken line shows were the residual state amplitude would have ended-up if the residual stresses scaled linearly with the stiffness of the sheets. (Paper V)

The fact that the residual stress amplitude did not scale linearly with the stiffness means that there must be some additional mechanism apart from increased stiffness underlying the increase in residual stresses with beating and polyelectrolyte addition. Increased beating is known to increase shrinkage during drying (Brecht et al. 1956), whereas the build-up of PAH/PAA multilayers on fibres has been demonstrated not to influence shrinkage during free drying (Larsson and Wågberg 2008). The fact that beating increases dewatering resistance and shrinkage during drying implies that there will be greater moisture and shrinkage gradients and thus greater residual stresses in sheets made of beaten fibres. The finding of Ivarsson (1954), that sheets start to shrink at higher moisture contents the more beaten their constituent fibres, also suggests that beating increases the moisture content at which load starts to be transferred between fibre layers. Earlier load transfer between fibre layers further increases the residual stresses.

No beating, no chemistry

PAH

Multilayer PAH/PAA/PAH

500 revs

5000 revs

7000 revs

0

2

4

6

8

10

12

0 2 4 6 8 10

Residu

al stress amplitud

e (M

Pa)

Elastic modulus (GPa)

No beating, no chemistry

Beating

Chemistry

50

Since the presence of polyelectrolyte multilayers did not influence de-watering resistance or drying shrinkage (Larsson and Wågberg 2008), the increase in residual stresses with the addition of multilayers cannot have been because the multilayers increased the moisture and shrinkage gradients through the sheet. This suggests that the added polyelectrolytes mainly influenced the dry content at which fibres started to interact and transfer load and thereby the residual stresses. The fact that adding polyelectrolytes produced higher residual stresses than did beating for a given sheet modulus in combination with the fact that polyelectrolyte addition did not increase de-watering resistance and thus moisture gradients, means that the load transfer between fibre layers in sheets containing adsorbed polyelectrolyte must have started at a significantly higher moisture content than in sheets made of beaten fibres. The presence of adsorbed polyelectrolytes probably facilitated capillary forces, the development of van der Waals forces, and the interdiffusion of polymer chains between adjacent fibres, so that the fibres with adsorbed polyelectrolyte started to interact and transfer load at a significantly lower solids content.

51

Influence of fibre cross-linking on mechano-sorptive creep of hand-sheets Since the use of polyelectrolytes did not seem to be a feasible way to reduce the mechano-sorptive creep, the use of fibre cross-linking through periodate oxidation was explored in Paper III. Fibres were oxidised using two different periodate dosages and rections times and their carbonyl content was determined as described earlier. During drying, the carbonyl groups introduced can react with hydroxyl groups in the fibres, as suggested in Figure 1, thus introducing cross-links. The sheets prepared from oxidised fibres were tested for thickness, tensile properties (dry and wet) and mechano-sorptive creep. The results of the tensile tests are shown in Table 2. The fact that the wet tensile strength for the most oxidised sample was 40% of the dry tensile strength is a clear indication of the existence of covalent links, since covalent linkages are assumed to be a prerequisite for achieving a relative wet strength greater than 30% (Espy 1995). The change in solid-state 13C-NMR spectra upon oxidation of the fibres as shown in Figure 18 also clearly indicates the formation of cross-links, especially since treatment under basic condition (pH 11.8), which is known to hydrolyse hemiacetals, changed the spectra back to what it was before cross-linking. As shown in Table 2, cross-linking had no major influence on the apparent density, tensile strength index or tensile stiffness index of the sheets. This indicates that the oxidation did not influence the interaction between fibres in the dry state or during the consolidation process, since this would have manifested itself as a difference in tensile strength and/or density. The strain at break was, however, severally impaired. These results are similar to those found by Caulfield who used multifunctional carboxylic acids to crosslink the fibre wall (Caulfield 1994).

52

Table 2. Results of tensile testing of sheets made from periodate-oxidised unbleached pulp, mean values and 95%-confidence limits (Paper III).

Carbonyl content

(mmol/g fibre)

Desnisty (kg/m3)

Tensile index

(kNm/kg)

Tensile stiffness index

(MNm/kg)

Strain-at-Break

(%)

Wet Tensile index

(kNm/kg)

0 682 75 ± 1 7.5 ± 0.1 2.9 ± 0.1 3.0 ± 0.1

0.47 ± 0.02 666 83 ± 3 7.8 ± 0.2 2.2 ± 0.1 29 ± 1

1.12 ± 0.02 695 76 ± 2 7.6 ± 0.1 1.6 ± 0.1 30 ± 1

Figure 18. Solid state 13C NMR-spectra of untreated bleached kraft fibres and periodate-oxidised fibres (1.2 mmol carbonyls per gram fibre). After oxidation, sheet formation and drying, one part of the sheet was soaked in water for 24 hour and the other part soaked under alkaline conditions (pH 11.8) until the fibres had disintegrated from each other. The signal intensities of the different samples have been normalised with respect to the C4-carbon signal of the starting material. (Larsson 2010)

405060708090100110120Shift (ppm)

Starting materialOxidisedOxidised, soaked in water 24 hOxidised, dissolved in NaOH (pH 11.8)

C4

C1

C2, C3, C5

C6

C3 C 2

C 1

OC5

C4

C6OH

O

OHOH

O

53

Mechano-sorptive creep was evaluated as described earlier. Combining the strain measured after three humidity cycles with the applied load, the results shown in Figure 19 were obtained. It is seen that, for a given stress, the creep strain after three cycles was considerably lower for the oxidised samples, i.e. the sheets made from oxidised fibres had a greater resistance to mechano-sorptive creep. If the nomenclature proposed by Panek et al. (2004) is adopted, the slope of the linear part of a isocyclic stress-strain curve is called the isocyclic creep stiffness index. The calculated isocyclic creep stiffness indices and corresponding carbonyl contents can be found in Table 3. For the most oxidised sample, more than a three-fold increase compared to the reference was detected, which must be considered to be a major improvement. Hence, it is shown that cross-linking can significantly decrease the mechano-sorptive creep of paper. Because of the embrittlement, cross-linking may not however be a feasible way to achieve a resistance to mechano-sorptive creep.

Figure 19. Isocyclic stress-strain data recorded after three humidity cycles (50-90% RH, each cycle being 7 hours long) for paper sheets (140 g/m2) with two different levels of cross-linking. Cross-linking through periodate oxidation significantly reduced the mechano-sorptive creep strain after three humidity cycles. (Paper III)

-4.5

-4

-3.5

-3

-2.5

-2

-1.5

-1

-0.5

0-0.25-0.2-0.15-0.1-0.050

Strain (%)

Spec

ific

Stre

ss (k

Nm

/kg)

Reference, 0.0 mmol carbonyls / g

Oxidised, 0.47 mmol carbonyls / g

Oxidised, 1.12 mmol carbonyls / g

54

The two dominant theories for mechano-sorptive creep (Habeger and Coffin 2000; Alfthan et al. 2002) have sorption-induced swelling as a common factor. The strain response measured during a mechano-sorptive creep test is a combination of hygroexpansion and creep. The amplitude of the hygroexpansion during the humidity cycling used for mechano-sorptive creep testing of the 140 g/m2 sheets was estimated as shown in Figure 20. The result is presented in Table 3 together with the corresponding isocyclic creep stiffness indices. The hygroexpansion amplitude decreased with increased degree of oxidation. A similar behaviour was found for sheets made of cross-linked bleached fibres in Paper IV, where the hygroexpansion between 20 and 85% RH decreased with increasing degree of cross-linking, as shown in Table 4. The lower hygroexpansion was primarily due to a smaller change in moisture content in the oxidised sheets. It is thus clear that the periodate-oxidation reduced the sorption-induced swelling of the sheets during the humidity cycles used for mechano-sorptive creep testing. The suggestions of Habeger and Coffin (2000) that sorption-induced swelling induces stress concentrations either by moisture gradients or material heterogeneity, mean that the reduction in sorption-induced swelling with cross-linking probably reduces the amplitude of stress concentrations within the sheets that arise due to the humidity cycling. However, the reason for the reduced hygroexpansion is not clear, nor whether stress concentrations are reduced both at the fibre network level and at the fibre level (so that the fibres themselves showed reduced mechano-sorptive creep). Accordingly, the effects of cross-linking on moisture sorption and mechano-sorptive creep of single fibres were specifically studied and the results are presented in the following sections.

Table 3. Carbonyl content, hygroexpansion amplitude (during 7 hour long 50-90% RH cycles) and isocyclic creep stiffness index at three humidity cycles for sheets made from unbleached fibres with different degrees of oxidation. (Paper III)

Degree of oxidation

Carbonyl Content (mmol/g fibre)

Isocyclic Creep Stiffness Index (MNm/kg)

Hygroexpansion Amplitude (%)

Reference 0 0.71 ± 0.08 0.34

Low 0.47 ± 0.02 1.54 ± 0.06 0.27

High 1.12 ± 0.02 2.23 ± 0.13 0.22

55

Figure 20. Strain recorded in a typical mechano-sorptive creep test. By least-squares-fitting lines to the maximum and minimum points (indicated by circles), the hygroexpansion during the creep test could be estimated.

Table 4. Carbonyl content, hygroexpansion, and change in moisture content when changing the relative humidity from 85 to 20% RH for sheets made from bleached unbeaten fibres with different degrees of oxidation. The hygroexpansion and moisture content change were measured after the release of dried-in strains. (Paper IV)

Degree of oxidation

Carbonyl Content (mmol/g fibre)

Hygroexpansion 85-20% RH (%) ΔMC (%)

Reference 0 -0.69 ± 0.03 7.83

Low 0.26 -0.62 ± 0.01 7.23

Medium 0.57 -0.56 ± 0.01 6.24

High 1.21 -0.50 ± 0.01 4.66

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

0 200 400 600 800 1000 1200

Time (min)

Stra

in (%

)

Hygroexpansionamplitude

56

Influence of fibre cross-linking on mechano-sorptive creep of single fibres As indicated above, the reduction in hygroexpansion with increasing cross-linking is the basic mechanism behind the reduction in mechano-sorptive creep. In order to understand whether the reduction in fibre swelling reduces the stress concentrations, and hence the mechano-sorptive creep, at the fibre level or at the fibre network level, the creep behaviour of single fibres was also studied. The result is shown in Figure 21 where the creep rate in cyclic humidity is plotted as function of the creep rate at a constant humidity, which means that each fibre is normalised with respect to itself and the scatter in data due to large differences between different fibres is avoided. The ratio between creep at cyclic humidity and creep at constant humidity gives a good measure of how much the creep is accelerated by varying the humidity. As seen in Figure 21, the increase in creep rate due to varying humidity is the same for the reference and cross-linked fibres, suggesting that the fibres themselves were not improved from an accelerated creep perspective, even though sheets made from the fibres showed significant improvements. This also suggests that any stress concentrations in the fibres due to humidity cycling were not affected by the cross-linking. The acceleration ratio in this case was 2.4, which is in fairly good agreement with previous results where the ratio for single wood fibres isolated by maceration with hydrogen peroxide and acetic acid, was found to be about three (Olsson and Salmén 2008). The fact that the mechano-sorptive creep of the fibres themselves was not reduced by the chemical cross-linking shows that the improvement found at the sheet level does not originate from an improvement in the creep resistance of the individual fibres but instead from mechanisms at the fibre network level. The fact that fibre/fibre joints are necessary in order for cross-linking to reduce mechano-sorptive creep suggests that the reduction in hygoexpansion with cross-linking reduced the mismatch of hygroexpansive strains and stress concentrations in the fibre/fibre joints, and thus reduced the mechano-sorptive creep according to the model proposed by Alfthan et al. (2002).

57

Figure 21. Creep strain rate at cyclic humidity versus creep strain rate at constant humidity (logarithmic time) for reference and periodate-oxidised single fibres. (Paper III)

Since the mechano-sorptive creep of the fibres themselves was not improved by the chemical treatment it can be concluded that the hypothesis proposed by Caulfield (1994) that the cross-links introduced stabilise arrays of moisture-sensitive hydrogen bonds and hinder them from creeping into a stress-relaxed configuration cannot hold true at the fibre level.

y = 2.3909x

0

0.2

0.4

0.6

0.8

1

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

Creep rate in constant RH%/log (time, min.)

Cre

ep ra

te in

cyc

lic R

H%

/log

(tim

e, m

in.)

Reference, 0.0 mmol carbonyls / gOxidised, 1.12 mmol carbonyls / g

58

Influence of cross-linking on moisture sorption In order to understand the mechanisms behind the reduced hygroexpansion of sheets made from cross-linked fibres, the moisture sorption of sheets made from cross-linked fibres was studied in Papers III and IV. In Paper IV, the moisture sorption of sheets made from cross-linked bleached fibres was studied in a dynamic vapour sorption equipment. The moisture content change in the different samples when changing from 30% RH to 90% RH at a speed of 0.5% RH/min is shown in Figure 22. The rate of moisture sorption was much slower for the most oxidised sheets that had not previously been exposed to a high relative humidity. However, after the sheets had been subjected to 98% RH, there was no longer any difference in the moisture sorption kinetics. This suggests that something happens with the cross-linked fibres when they are exposed to a high relative humidity and this will be discussed in detail later. The fact that the sheets from oxidised fibres sorb moisture at a lower rate suggests that the reason why the cross-linked sheets showed less hygroexpansion during the mechano-sorptive creep testing was that they sorbed less moisture. However, the equilibrium moisture sorption isotherms in Figure 23, show no significant difference in adsorption behaviour between the bleached cross-linked fibres and the reference fibres. Hence, it depends upon whether or not the unbleached cross-linked sheets reached moisture equilibrium during the mechano-sorptive creep testing.

59

Figure 22. Moisture content as a function of time for sheets made from bleached fibres cross-linked to different degrees, when subjected to a change in relative humidity from 30 to 90% RH (a) before and (b) after a first exposure to 98% RH. (Paper IV)

0

10

20

30

40

50

60

70

80

90

100

2

4

6

8

10

12

14

-20 0 20 40 60 80 100 120 140 160 180

Targ

et R

H (%

)

Moi

stur

e co

nten

t (%

)

Time (min)

Reference0.3 mmol/g0.6 mmol/g1.2 mmol/gRH

(a)

0

10

20

30

40

50

60

70

80

90

100

2

4

6

8

10

12

14

-20 0 20 40 60 80 100 120 140 160 180

Targ

et R

H (%

)

Moi

stur

e co

nten

t (%

)

Time (min)

Reference0.3 mmol/g0.6 mmol/g1.2 mmol/gRH

(b)

60

Figure 23. Equilibrium moisture sorption isotherms for sheets made from bleached fibres cross-linked to different degrees. The cross-linked fibres sorb approximately the same amount of water as the reference fibres, but retain more water during the desorption. (Paper IV)

To ascertain whether or not the cross-linked samples reached moisture equilibrium during the mechano-sorptive creep testing, the moisture contents of the unbleached reference and cross-linked sheets during the humidity cycling used for mechano-sorptive creep testing were determined. The result is shown in Figure 24, where it is evident they did not reach moisture equilibrium and that the moisture content variation amplitude during the 7 hour long 50-90% RH cycles decreased with increasing oxidation. It can therefore be concluded that the reason for the reduced hygroexpansion of the cross-linked sheets during mechano-sorptive creep testing was that they were subjected to smaller variations in moisture content. The basic mechanism behind the reduction in mechano-sorptive creep with cross-linking was hence a reduction in the moisture content changes in the oxidised samples, and not that the cross-links introduced stabilised arrays of moisture-sensitive hydrogen bonds and hindered them from creeping into a stress-relaxed configuration as suggested by Caulfield (1994).

0

2

4

6

8

10

12

14

16

18

0 10 20 30 40 50 60 70 80 90 100

Moi

stur

e co

nten

t (%

)

Relative humidity (%)

Reference

0.3 mmol/g

0.6 mmol/g

1.2 mmol/gDesorption

Adsorption

61

The correlation between the degree of mechano-sorptive creep and the moisture content variation during the humidity cycling has been found previously. Alfthan (2004) found that the creep strain after a certain number of moisture cycles increased with increasing moisture content variation amplitude. Fellers and Panek (2007) found that the creep acceleration due to varying humidity, measured as a mechano-sorptive creep factor (creep stiffness at varying humidity divided by creep stiffness at constant humidity), increased with increasing difference in moisture ratio at low and high humidity.

Figure 24. Moisture-sorption for sheets made from unbleached fibre cross-linked to different degrees, during the three 50-90% RH cycles used for the mechano-sorptive creep testing. Prior to testing, the samples, like the samples used for mechano-sorptive creep testing, had been preconditioned by exposing them to six 50 to 90% RH cycles. As seen, the moisture variation amplitude decreased with increasing oxidation. (Paper III) The fact that the cross-linked sheets in Figure 22 that had been subjected to 98% RH no longer showed a slower moisture sorption rate might suggest that the cross-links are hydrolysed at high relative humidity. However, the solid-state 13C-NMR spectra in Figure 18 show a change in spectra for cross-linked sheets of bleached fibres that is not reversed by soaking in water for 24 hours, but is reversed by treatment under alkaline conditions (pH 11.8), which are known to hydrolyse hemiacetals. This clearly shows that water soaking or high relative humidity does not hydrolyse the cross-links.

30

40

50

60

70

80

90

100

5

6

7

8

9

10

11

12

13

14

15

0 200 400 600 800 1000 1200

RH

(%)

Moi

stur

e co

nten

t (%

)

Time (min)

Reference, 0.0 mmol carbonyls / gOxidised, 0.47 mmol carbonyls / gOxidised, 1.12 mmol carbonyls / gRH

62

What is then the reason why the cross-linked sheets no longer sorbed moisture at the slower rate after being subjected to 98% RH? A clue to this can be found in Figure 24, where for the most cross-linked samples it is seen that, at the beginning of the second cycle, the moisture content quickly reached the moisture content that had been reached during the first cycle, whereupon the moisture sorption was again slow. The same phenomenon is also seen in the third cycle, where the moisture content quickly reached the moisture content that had been reached during the second cycle. This suggests that regions in cross-linked fibres that had previously been reached by water take up moisture at a relatively fast rate after re-drying. The adsorption isotherms in Figure 23 also suggest that at high relative humidities the cross-linked fibres take up as much water as the reference fibres, if they are allowed to reach moisture equilibrium. Taking these facts into consideration, the following scenario (which is schematically outlined in Figure 25) is suggested: In ordinary fibres, fibrillar lamellae that have collapsed during drying re-expand rather easily and quickly upon re-moistening. This means that the moisture sorption in the reference samples is a rather fast process. In cross-linked fibres, the periodate ions are able to introduce aldehyde groups into all the water-swollen parts of the fibre wall. When the samples are dried, the fibrillar lamellae collapse and the aldehyde groups introduced react with hydroxyl groups in neighbouring lamellae, linking the lamellae irreversibly together. When the cross-linked samples are re-moistened, the lamellae that have been linked together are unable to expand, but new lamellae (not previously reached by water) are separated. The process of separating lamellae previously inaccessible to water is however proposed to be slow, so that the moisture sorption in cross-linked fibres becomes a slow process. If, however, the cross-linked fibres are re-dried after new lamellae have been separated, these will quickly separate on re-moistening. This explains why the fibre regions reached by water in the first humidity cycle in Figure 24 can quickly expand during the second cycle so that the moisture content reached during the first cycle is quickly reached in the second cycle. Thereafter, new lamellae must be separated so that the moisture sorption rate in the second cycle becomes slow after the moisture content reached during the first cycle again had been reached. This proposed mechanism also explains why the moisture sorption in cross-linked samples in Figure 22 again becomes rapid after the samples have been exposed to a high relative humidity (98% RH) for a long time.

63

Figure 25. Schematic drawing of how the moisture sorption in untreated and cross-linked samples is suggested to take place. In the reference fibres, fibrillar lamellae that have collapsed during drying re-expand rather easily and quickly upon re-moistening. This make the moisture sorption in reference samples a rather fast process. In periodate-treated fibres, the periodate ions are able to introduce aldehyde groups into all the water-swollen parts of the fibre walls. When the samples are dried, the fibrillar lamellae collapse and the aldehyde groups react with hydroxyl groups in neighbouring lamellae, linking the lamellae together (|). When these cross-linked samples are re-moistened, the lamellae that have been linked together are unable to expand. However, the cross-linked material still takes up moisture at a slow rate. It is therefore suggested that new lamellae (not previously reached by water) are separated, but that the process of separating lamellae previously inaccessible to water is slow, so that the moisture sorption in cross-linked fibres becomes a slow process. If the cross-linked fibres are re-dried after a new pore system has been opened, the re-moistening will again be rapid, since the lamellae have now previously been separated by water. This explains why the moisture sorption in cross-linked samples again becomes rapid after the samples have been exposed to a high relative humidity for a long time.

From these data, it is clear that cross-linking does not influence the equilibrium moisture uptake (as shown in Figure 23) but only the kinetics of the sorption process. Hence, if the moisture cycles shown Figure 24 had been continued for a large number of times, the cross-linked fibres would eventually have taken up as much water as the reference fibres. Here it is important to remember that the samples in Figure 24 had been preconditioned with six 50 to 90% RH cycles, which means that the moisture changes in the cross-linked samples were even less during the preconditioning cycles than in the cycles shown in Figure 24.

64

Since the cross-linked samples would eventually have taken up as much water as the reference samples, the hygroexpansion, and thus stress concentrations in fibre/fibre joints, would eventually have been the same. At this point, there would probably be no difference in the mechano-sorptive creep rate between the cross-linked and the reference papers. The positive effect of cross-linking on the mechano-sorptive creep rate is therefore probably only temporary, but it is possible that cross-linking reduces the mechano-sorptive creep rate over such long time periods that it would significantly improve the stacking life-time of corrugated boxes. For cross-linking to be a really feasible way to reduce mechano-sorptive creep, the problem of embrittlement must however be solved. The single fibres tested for mechano-sorptive creep were subjected to constant 80% RH for 3 hours before being exposed to a cyclic humidity. This might suggest that new pores were opened in these fibres during the 3-hour period so that they did not show any reduced moisture sorption rate during the humidity cycling as did the sheets in Figure 24. This alone could be the reason why the cross-linked single fibres exhibited the same degree of creep acceleration in Figure 21 as the reference fibres. However, the creep properties of sheets with a thickness of only 45 μm (20 g/m2), i.e. only a few fibres thick, were also determined in Paper III. Since these sheets were very thin, they should absorb moisture almost at the same rate as single fibres. These sheets were also subjected to a 3-hour period at constant 80% RH prior to exposure to a cyclic humidity. As seen in Figure 26, the creep rate was still accelerated significantly more for the reference sheets than for the sheets made from oxidised fibres. This shows that the 3-hour long 80% RH period was not the reason why the cross-linked fibres did not show any reduction in mechano-sorptive creep when tested individually. The testing of the thin sheets also shows that fibre/fibre joints are necessary for cross-linking to affect the mechano-sorptive creep.

65

Figure 26. Creep strain rate in a cyclic humidity (80-30% RH, 1 hour long cycles) versus creep strain rate in a constant climate (80% RH) for thin (20 g/m2) reference and cross-linked sheets. (Paper III)

Development of shear strength during drying Shear strength is very important for the how well a paperboard performs during creasing and subsequent folding (Nygårds et al. 2009a; Nygårds et al. 2009b). Understanding the development of shear strength during drying is valuable not only in order to be able to optimise the shear strength profile in paperboards but also because it may give vital clues to how the fibres interact during drying and the nature of these interactions. Shear testing of wet paper instead of in-plane tensile testing can give more information regarding the interaction between fibres, since mechanical entanglement of the fibres can play a vital role besides the adhesion between fibres for the results obtained by in-plane tensile testing, as pointed out by Alince et al (2006). Accordingly, shear strength development during drying was studied in Paper VI.

y = 5.4011x

y = 2.2767x

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.05 0.1 0.15 0.2 0.25

Creep rate in constant RH%/log (time, min.)

Cre

ep ra

te in

cyc

lic R

H%

/log

(tim

e, m

in.)

Reference, 0.0 mmol carbonyls / g

Oxidised, 1.12 mmol carbonyls / g

66

The result of the shear testing of never-dried and rewetted sheets, made of beaten and of unbeaten fibres, that had been couched together as described in the Experimental section, is shown in Figure 27. In the figure, it is obvious that the shear strength of all combinations of the two types of fibres (i.e., beaten and unbeaten) and the two specimen preparation methods (i.e., never-dried and rewetted sheets) was low (<120 kPa) up to a solids content of approximately 60–70%, after which the shear strength increased rapidly with increasing solids content. The more water that was removed from the sheets, the closer the fibres in the opposite contacting sheets could be drawn together by capillary forces, leading to intimate contact between the fibres. The kink in the shear strength versus solids content plot at 60–70% suggests that interactions important for the strength at complete dryness started to develop at this particular solids content. The exact nature of the fibre/fibre interactions at different solids content cannot however be deduced from the data in this study.

Figure 27. Shear strength of the various sheets couched together, plotted as a function of the solids content. The specimens with a 15-mm overlap between the strips are indicated by symbols with a thin outline, whereas the specimens with a 5-mm overlap are indicated by symbols with a bold outline. The regression lines included in the figure merely serve to guide the eye.

0

500

1000

1500

2000

2500

3000

3500

30 40 50 60 70 80 90 100

Shear stren

gth (kPa

)

Solids content (%)

Beaten unbleached  fibres  ‐ Never‐dried sheets

Beaten unbleached  fibres  ‐ Rewetted sheets

Unbeaten bleached  fibres  ‐ Never‐dried sheets

Unbeaten bleached  fibres  ‐ Rewetted sheets

67

As indicated in Figure 27, the shear strength of the paper sheets couched together increased with increasing solids content at all solids contents in the investigated interval. The results of Alince et al. (2006), who examined blotter papers made of bleached unbeaten fibres, indicate that maximum shear strength was reached at a solids content of approximately 39%, after which the shear strength dropped, reaching zero at a solids content of approximately 47%. The rewetted sheets made of unbeaten fibres were not found to behave in this way in the present study. The reason for this discrepancy is unknown, but may relate to differences between the paper sheets used. As also seen in Figure 27, the highest shear strength, throughout the solids content range examined, was achieved with never-dried sheets made of beaten fibres, which is as expected since beating is known to increase the strength of paper and never-dried fibres are known to give stronger paper than previously dried fibres. The shear strength of the specimens made from never-dried sheets of unbeaten fibres was lower than that of the sheets of beaten fibres. This may partly be because the unbeaten fibres have a fibre wall with a higher wet modulus (Chhabra et al. 2005) and swell less than do the beaten fibres, and thus produce weaker fibre/fibre joints. However, it is also likely that the more fibrillated surface of the beaten fibres facilitates mechanical interlocking and interdiffusion of surface polymers, thus further increasing the shear strength of the sheets of beaten fibres (McKenzie 1984; Clark 1985). The specimens made from rewetted sheets had significantly lower shear strength than did those made from never-dried sheets. The fact that the previously dried fibres had a fibre wall with a higher wet modulus (Scallan and Tigerström 1992) and swelled less (Jayme 1944; Forsström et al. 2005) than did the never-dried fibres could explain why the specimens made from rewetted sheets were significantly weaker than those made from never-dried sheets. However, it is likely that the reduced swelling of the rewetted sheets also affected their conformability thus reducing their ability to come in close contact and thus achieve greater strength.

68

The author of this thesis is of the firm opinion that shear strength measurements on modified fibres could increase the understanding of how the modification influences the fibre/fibre interactions and thus give vital clues regarding the mechanisms by which the modification improves the mechanical performance of the resulting paper. By increasing the understanding of the mechanisms by which different fibre modifications improve paper performance, the time for the development of new and improved fibre modifications can hopefully be reduced.

69

Concluding remarks Labelling of polyallylamine (PAH) with fluorescent marker and examination of fibres with the aid of a confocal scanning microscope (Horvath et al. 2008a) generated a visual record of where the adsorbed PAH was located. This technique was used to locate the adsorbed PAH and to establish the adsorption conditions necessary to ensure that the PAH molecules were adsorbed either to the exterior parts of the fibres alone or to both the external parts and throughout the fibre walls. By preparing and testing the mechanical behaviour of sheets made from fibres with polyelectrolyte adsorbed either to the exterior parts alone or into the fibre cell walls as well, a link was established between the localisation of the adsorbed polyelectrolyte and the mechanical properties. Creep testing clearly showed that the creep at constant humidity was significantly reduced by the adsorption of PAH to the exterior parts of the fibre wall. The effect obtained depended, however, on the type of fibres used. PAH resulted in a significantly greater reduction in creep than cationic starch, even though both treatments induced the same increase in tensile strength. To the knowledge of the author this is the first time that the use of polyelectrolyte treatment has been shown to be able to reduce the creep of paper with high density and strength. This also shows that different polyelectrolytes can have different effect on the viscoelastic behaviour of paper materials. The adsorption of polyallylamine into the entire fibre cell wall, however, resulted in an increase in creep of the resulting sheets at high humidity conditions (90% RH). It is suggested that this may be due to a deswelling of the fibres by the adsorbed polyelectrolyte, which leads to fewer fibre/fibre contact points and hence a less efficient distribution of stresses in the sheets. These results show the complexity of the topic of reducing creep under constant humidity with chemical additives, but also that there is a potential for significant improvements provided that the right chemicals are added in the right way.

70

It is, however, primarily the mechano-sorptive creep rate and not the creep rate at constant humidity that determines the stacking lifetime of paper-based boxes. No significant effect of adsorbed polyallylamine could be detected on the creep during cyclic humidity conditions. It is therefore unlikely that any other polyelectrolyte commonly used to improve the strength of paper would have any significant effect on the mechano-sorptive creep. New specifically tailored polyelectrolytes, with for example cross-linking capabilities, might however prevent moisture sorption and fibre swelling, and could perhaps in the future open new possible routes for reducing mechano-sorptive creep. Polyelectrolytes can however be efficient in improving other mechanical properties important for the performance of packaging grade papers, since the use of multilayers consisting of polyallylamine (PAH) and polyacrylic acid (PAA) significantly increased the strength of paper with much less densification and build-up of residual stress than is obtained by beating. The use of polyelectrolytes should thus enable the production of paper with high strength but low residual stress and density. To the knowledge of the author, this is also the first published study on how chemical additives influence the residual stress state in paper. The results indicate that the added polyelectrolytes increased the residual stress by decreasing the solids content at which load transfer between fibre layers starts. In contrast to the use of polyelectrolytes, cross-linking by the oxidation of the fibres with periodate radically decreased the mechano-sorptive creep of the resulting sheets. By testing sheets and individual fibres, it was clearly shown that the basic mechanism behind the reduction in mechano-sorptive in this case was that the cross-linking slows down the moisture sorption. A slow sorption rate led to smaller moisture content variations during the mechano-sorptive creep testing. Due to the anisotropic swelling of fibres, large stress concentrations in the fibre/fibre-joints will arise upon changes in moisture content (Alfthan et al. 2002). Smaller moisture content variations in the cross-linked sheets during humidity cycling mean less fibre swelling and thus less stress concentrations in the fibre/fibre joints. Less stress concentrations in the fibre/fibre joints in combination with the non-linear creep of paper mean less creep acceleration. These results also indicate that the mechanism proposed by Caulfield (1994), i.e., that the cross-links introduced stabilise arrays of moisture-sensitive hydrogen bonds and hinder them from creeping into a stress-relaxed configuration, is most probably not the mechanism behind mechano-sorptive creep reduction with cross-linking.

71

A major problem with cross-linking is that it causes severe embrittlement. However, an understanding that the reduction in mechano-sorptive creep with cross-linking is due to a slower moisture sorption might generate new ideas as to how to achieve mechano-sorptive creep reduction without causing embrittlement. One way to circumvent the problem of embrittlement might be to use polyelectrolytes with cross-linking capabilities, which might be able to retard the moisture sorption without impairing the toughness of the fibres. The effect of cross-linking on the moisture sorption kinetics and thus also mechano-sorptive creep is however only temporary, although it might nevertheless be sufficient to improve the stacking life of boxes. Considering that it has only a temporary effect, cross-linking might not in general be the best way to reduce moisture content variations under varying humidity and thus mechano-sorptive creep. Barrier approaches could be an alternative way to slow down the moisture sorption kinetics and thus the mechano-sorptive creep. It is therefore suggested that such approaches should be further explored in the future. Lamination with plastic foils might be costly, but the development of cheap barrier coatings that can be applied in a standard on-line coater or size-press could be economically feasible. Another option might be to use the multilayer technique to apply a moisture barrier coating to each individual fibre. This would require an extra process step in the furnish preparation, but no extra operation on or after the paper-machine. The effect of cross-linking on mechano-sorptive creep is relatively well understood, but the effects of other fibre modifications and additives on the mechanical properties of paper are not however always so well understood. The method of determining shear strength at different dry solids content presented in this thesis could in the future be combined with modified fibres and hopefully increase the understanding of how different modifications influence the fibre/fibre interactions. By understanding the influence on fibre/fibre interactions, clues regarding the mechanisms by which the modifications improve the mechanical performance could be gained. A greater understanding of the mechanisms by which different fibre modifications improve paper performance could hopefully speed up the development of new and improved fibre modifications.

72

The data presented here also appear to be the first presentation of how shear strength develops between couched wet sheets during drying all the way to completely dry sheets. The results indicate that the shear strength increased with increasing solids content at all solids contents investigated. The results of Alince et al. (2006), who examined blotter papers made of bleached unbeaten fibres, indicate that the maximum shear strength was reached at a solids content of approximately 39%, after which the shear strength dropped, reaching zero at a solids content of approximately 47%. The rewetted sheets made of unbeaten fibres were not found to behave in this way in the present study. The reason for this discrepancy is unknown, but it may be due to differences between the paper sheets used. The presented results also showed that the shear strength was low up to a solids content of approximately 60–70%, after which the shear strength increased rapidly with increasing solids content. The kink in the shear strength versus solids content plot at 60–70% suggests that interactions important for the strength of dry paper start to develop at this particular solids content.

73

Acknowledgements First, I would like to express my gratitude to my supervisor Lars Wågberg for giving me the opportunity to do my PhD-studies within the fibre technology group, and for all the guidance, encouragement and optimism. You make the fibre technology group a great place to work. I am very grateful to SustainPack (a project within EU's Sixth Framework Research Programme) and BiMaC Innovation for financial support. Tom Lindström is also acknowledged for a great job with both BiMaC Innovation and sub-project no. 2 within SustainPack. All co-authors are thanked for good collaboration. I thank all present and former colleagues at the Department of Fibre and Polymer Technology for creating a stimulating and friendly work environment. Special thanks goes to Brita, Inga and Mona for the assistance with adimistrative and practical tasks; Per, Oskar, Caroline, Julia, Andrew, Lars-Erik, Stefan and Lars Ö for good company, fruitful discussion and many happy moments over the years. My assistant supervisor Sören Östlund and my other colleagues at KTH Solid Mechanics, Magnus Östlund and Mikael Nygårds are acknowledged for always being supportive and helpful. Innventia is gratefully acknowledged for granting me access to their facilities. Special thanks to all the kind people there who have helped me over the years, especially Petri Mäkelä, Anne-Mari Olsson, Lennart Salmén, Anni Hagberg, Tomas Larsson, Åsa Blademo, Gunnborg Glad-Nordmark, Christer Fellers, Ann Fält, Lars-Erik Åkerlund and Anette Lindé. Felix Lindström is acknowledged for his skilful assistance with the laboratory work in Paper V and VI. Anthony Bristow is acknowledged for excellent linguistic revision over the years. Finally, I would like to express my deepest gratitude to my family for always being there and being the best family one could have. And Madeleine, thank you for everything, I love you.

74

References Alava, M. and Niskanen, K. (2006): The physics of paper, Reports on Progress in

Physics, 69(3): 669-723.

Alfthan, J. (2004): The Effect of Humidity Cycle Amplitude on Accelerated Tensile Creep of Paper, Mechanics of Time-Dependent Materials, 8(4): 586-597.

Alfthan, J., Gudmundson, P. and Östlund, S. (2002): A micromechanical model for mechanosorptive creep in paper, Journal of Pulp and Paper Science, 28(3): 98-104.

Alince, B. (1990): The role of porosity in polyethylenimine adsorption onto cellulosic fibers, Journal of Applied Polymer Science, 39(2): 355-62.

Alince, B., Vanerek, A., de Oliveira, M. H. and van de Ven, T. G. M. (2006): The effect of polyelectrolytes on the wet-web strength of paper, Nordic Pulp & Paper Research Journal, 21(5): 653-658.

Andreasson, B., Forsström, J. and Wågberg, L. (2005): Determination of fibre pore structure: influence of salt, pH and conventional wet strength resins, Cellulose, 12(3): 253-265.

Back, E. L. (1967): Thermal auto-crosslinking in cellulose material, Pulp & Paper Magazine of Canada, 68(4): 165-171.

Back, E. L., Salmen, L. and Richardson, G. (1983): Transient Effects of Moisture Sorption on the Strength Properties of Paper and Wood-Based Materials, Svensk Papperstidning-Nordisk Cellulosa, 86(6): R61-R71.

Bernada, P., Stenström, S. and Månsson, S. (1998): Experimental study of the moisture distribution inside a pulp sheet using MRI. Part II: drying experiments, Journal of Pulp and Paper Science, 24(12): 380-387.

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

Byrd, V. L. (1972a): Effect of Relative Humidity Changes During Creep on Handsheet Paper Properties, Tappi, 55(2): 247-252.

75

Byrd, V. L. (1972b): Effect of Relative Humidity Changes on Compressive Creep Response of Paper, Tappi, 55(11): 1612-1613.

Caulfield, D. F. (1994): Ester crosslinking to improve wet performance of paper using multifunctional carboxylic acids, butanetetracarboxylic and citric acid, Tappi Journal, 77(3): 205-12.

Chhabra, N., Spelt, J., Yip, C. M. and Kortschot, M. T. (2005): An investigation of pulp fibre surfaces by atomic force microscopy, Journal of Pulp and Paper Science, 31(1): 52-56.

Clark, J. d. A. (1985): Fibrillation and fibre bonding, In: Pulp technology and treatment for paper, 2 ed., Miller Freeman, San Francisco.

Coffin, D. W. (2005): The creep response of paper, In Advances in Paper Science and Technology, 13th Fundamental Research Symposium, Cambridge, UK, September, 2005, 651-747.

Coffin, D. W. (2009): Developing a deeper understanding of the constitutive behavior of paper, In Advances in Pulp and Paper Research, 14th Fundamental Research Symposium, Oxford, UK, Sept. 2009, 841-875.

Coffin, D. W. and Boese, S. B. (1997): Tensile creep behaviour of single fibers and paper in a cyclic humidity environment., In Moisture and creep effects on paper, board and containers: 3rd international symposium, Rotorua, New Zealand, February, 1997, 39-52.

Davison, R. W. (1972): Weak link in paper dry strength, Tappi, 55(4): 567-73.

de Oliveira, M. H., Maric, M. and van de Ven, T. G. M. (2008): The role of fiber entanglement in the strength of wet papers, Nordic Pulp & Paper Research Journal, 23(4): 426-431.

Decher, G. (1997): Fuzzy nanoassemblies: Toward layered polymeric multicomposites, Science, 277(5330): 1232-1237.

DeMaio, A. and Patterson, T. (2005): Influence of fiber-fiber bonding on the tensile creep compliance of paper, In Advances in Paper Science and Technology, 13th Fundamental Research Symposium, Cambridge, UK, September, 2005, 749-775.

76

DiFlavio, J.-L., Bertoia, R., Pelton, R. and Leduc, M. (2005): The mechanism of polyvinylamine wet-strengthening, In Advances in Paper Science and Technology, 13th Fundamental Research Symposium, Cambridge, UK, September, 2005, 1293-1316.

Eichhorn, S. J., Sirichaisit, J. and Young, R. J. (2001): Deformation mechanisms in cellulose fibres, paper and wood, Journal of Materials Science, 36(13): 3129-3135.

Eriksson, M., Notley, S. M. and Wågberg, L. (2005): The influence on paper strength properties when building multilayers of weak polyelectrolytes onto wood fibers, Journal of Colloid and Interface Science, 292(1): 38-45.

Eriksson, M., Torgnysdotter, A. and Wågberg, L. (2006): Surface Modification of Wood Fibers Using the Polyelectrolyte Multilayer Technique: Effects on Fiber Joint and Paper Strength Properties, Industrial & Engineering Chemistry Research, 45(15): 5279-5286.

Espy, H. H. (1995): The mechanism of wet-strength development in paper: a review, Tappi Journal, 78(4): 90-9.

Fellers, C. and Panek, J. (2007): Effect of relative humidity cycle start point and amplitude on the mechano-sorptive creep of containerboard, In International Paper Physics Conference, Gold Coast, Austrailia, 6-9 May 2007, 291-296.

Forsström, J., Andreasson, B. and Wågberg, L. (2005): Influence of pore structure and water retaining ability of fibres on the strength of papers from unbleached kraft fibres, Nordic Pulp & Paper Research Journal, 20(2): 176-185.

Froix, M. F. and Nelson, R. (1975): Interaction of Water with Cellulose from Nuclear Magnetic-Resonance Relaxation-Times, Macromolecules, 8(6): 726-730.

Gedde, U. W. (1995): Polymer physics, Chapman & Hall, London.

Ghosh, P. and Dalal, J. C. (1988): Studies on crosslinking of dialdehyde cellulose and dialdehyde starch using poly(vinyl alcohol), Journal of Polymer Materials, 5(4): 241-7.

77

Habeger, C. C. and Coffin, D. W. (2000): The role of stress concentrations in accelerated creep and sorption-induced physical aging, Journal of Pulp and Paper Science, 26(4): 145-157.

Habeger, C. C., Coffin, D. W. and Hojjatie, B. (2001): Influence of humidity cycling parameters on the moisture-accelerated creep of polymeric fibers, Journal of Polymer Science, Part B: Polymer Physics, 39(17): 2048-2062.

Haraldsson, T., Fellers, C. and Kolseth, P. (1994): A Method for Measuring the Creep and Stress-Strain Properties of Paperboard in Compression, Journal of Pulp and Paper Science, 20(1): J14-20.

Haselton, W. R. (1955): Gas adsorption by wood, pulp, and paper. 2. The application of techniques to the study of the area and structure of pulps and the unbonded and bonded area of paper, Tappi, 38: 716-23.

Haslach, H. W., Jr. (2000): The moisture and rate-dependent mechanical properties of paper, Mechanics of Time-Dependent Materials, 4(3): 169-210.

Henriksson, L., Bredemo, R., Fellers, C. and Stubbfält, G. (2007): Mechanical Properties of Paper and Their Relation to Packaging Performance, In 61st Appita Annual Conference and Exhibition, Gold Coast, Austrailia, 6-9 May 2007, 305-312.

Hermanson, G. T. (1996): Bioconjugate techniques, Academic Press, San Diego.

Hill, R. L. (1967): The creep behavior of individual pulp fibers under tensile stress, Tappi, 50(8): 432-40.

Horvath, A. T., Horvath, A. E., Lindström, T. and Wågberg, L. (2008a): Adsorption of low charge density polyelectrolytes to an oppositely charged porous substrate, Langmuir, 24(13): 6585-6594.

Horvath, A. T., Horvath, A. E., Lindström, T. and Wågberg, L. (2008b): Adsorption of highly charged polyelectrolytes onto an oppositely charged porous substrate, Langmuir, 24(15): 7857-7866.

Horvath, A. T., Horvath, A. E., Lindström, T. and Wågberg, L. (2008c): Diffusion of cationic polyelectrolytes into cellulosic fibers, Langmuir, 24(19): 10797-10806.

78

Ivarsson, B. W. (1954): Introduction of stress into a paper sheet during drying, Tappi, 37: 634-9.

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

Jayme, G. (1944): Mikro-Quellungsmessungen an Zellstoffen., Der Papier-Fabrikant, Wochenblatt für Papierfabrikation, 6: 187-194.

Kallmes, O. and Eckert, C. (1964): The structure of paper: VII. The application of the relative bonded area concept to paper evaluation, Tappi, 47: 540-548.

Kim, U.-J., Kuga, S., Wada, M., Okano, T. and Kondo, T. (2000): Periodate oxidation of crystalline cellulose, Biomacromolecules, 1(3): 488-492.

Kolseth, P. and de Ruvo, A. (1983): The measurement of viscoelastic behavior for the characterization of time-, temperature-, and humidity-dependent properties, In: Handbook of physical and mechanical testing of paper and paperboard. Vol. 1, Marcel Dekker, New York, 255-322.

Koning, J. W., Jr. and Stern, R. K. (1977): Long-term creep in corrugated fiberboard containers, Tappi, 60(12): 128-131.

Laine, J., Lindström, T., Bremberg, C. and Glad-nordmark, G. (2003): Studies on topochemical modification of cellulosic fibers part 5. Comparison of the effects of surface and bulk chemical modification and beating of pulp on paper properties, Nordic Pulp & Paper Research Journal, 18(3): 325-332.

Laleg, M. and Pikulik, I. I. (1991): Wet-web strength increase by chitosan, Nordic Pulp & Paper Research Journal, 6(3): 99-103, 109.

Larsson, P. (2010): Hygro- and hydroexpansion of paper - Influence of fibre-joint formation and fibre sorptivity, Doctoral Thesis, KTH, Dept. of Fibre and Polymer Technology.

Larsson, P. A. and Wågberg, L. (2008): Influence of fibre-fibre joint properties on the dimensional stability of paper, Cellulose, 15(4): 515-525.

Leake, C. H. and Wojcik, R. (1993): Humidity cycling rates: how they influence container life spans, Tappi Journal, 76(10): 26-30.

79

Lindström, T. and Wågberg, L. (1983): Effects of pH and electrolyte concentration on the adsorption of cationic polyacrylamides on cellulose, Tappi Journal, 66(6): 83-5.

Lindström, T., Wågberg, L. and Larsson, T. (2005): On the Nature of Joint Strength in Paper - A Review of Dry and Wet Strength Resins used in Paper Manufacturing, In Advances in Paper Science and Technology, 13th Fundamental Research Symposium, Cambridge, UK, September, 2005, 457-562.

Lingström, R., Notley, S. M. and Wågberg, L. (2007): Wettability changes in the formation of polymeric multilayers on cellulose fibres and their influence on wet adhesion, Journal of Colloid and Interface Science, 314(1): 1-9.

Lundström, L. (2009): Polyelectrolyte multilayers of cationic and anionic starch and their use for improving the strength of papers made from mechanical pulps, Licentiate thesis, KTH, Dept. of Fibre and Polymer Technology.

Lyne, L. M. and Gallay, W. (1954): Studies in the fundamentals of wet-web strength, Pulp and Paper Magazine of Canada: 128-138.

McKenzie, A. W. (1984): The Structure and Properties of Paper .21. the Diffusion-Theory of Adhesion Applied to Interfiber Bonding, Appita, 37(7): 580-583.

Mohlin, U. B. (1977): Mechanical Pulp Properties - Importance of Fines Retention, Svensk Papperstidning-Nordisk Cellulosa, 80(3): 84-88.

Notley, S. M., Eriksson, M. and Wågberg, L. (2005): Visco-elastic and adhesive properties of adsorbed polyelectrolyte multilayers determined in situ with QCM-D and AFM measurements, Journal of Colloid and Interface Science, 292(1): 29-37.

Notley, S. M., Pettersson, B. and Wågberg, L. (2004): Direct measurement of attractive van der waals' forces between regenerated cellulose surfaces in an aqueous environment, Journal of the American Chemical Society, 126(43): 13930-13931.

Nygårds, M., Fellers, C. and Östlund, S. (2009a): Development of the notched shear test, In Advances in Pulp and Paper Research, 14th Fundamental Research Symposium, Oxford, UK, Sept. 2009, 877-897.

80

Nygårds, M., Just, M. and Tryding, J. (2009b): Experimental and numerical studies of creasing of paperboard, International Journal of Solids and Structures, 46(11-12): 2493-2505.

Olsson, A.-M. and Salmén, L. (2001): Molecular mechanisms involved in creep phenomena of paper, Journal of Applied Polymer Science, 79(9): 1590-1595.

Olsson, A.-M. and Salmén, L. (2008): Mechanosorptive creep a fibre property, comparison between the behaviour of fibres and paper, In Progress in Paper Physics Seminar, Espoo, Finland, June 2-5, 2008, 49-51.

Olsson, A.-M., Salmén, L., Eder, M. and Burgert, I. (2007): Mechano-sorptive creep in wood fibres, Wood Science and Technology, 41(1): 59-67.

Page, D. H. (1969): Theory for the tensile strength of paper, Tappi, 52(4): 674-81.

Page, D. H. (1985): The mechanism of strength development of dried pulps by beating, Svensk Papperstidning, 88(3): R30-R35.

Page, D. H. and Seth, R. S. (1980a): The elastic modulus of paper. II. The importance of fiber modulus, bonding, and fiber length, Tappi, 63(6): 113-116.

Page, D. H. and Seth, R. S. (1980b): The elastic modulus of paper. III. The effects of dislocations, microcompressions, curl, crimps, and kinks, Tappi, 63(10): 99-102.

Page, D. H., Seth, R. S. and De Grace, J. H. (1979): The elastic modulus of paper. I. The controlling mechanisms, Tappi, 62(9): 99-102.

Page, D. H. and Tydeman, P. A. (1961): A new theory of the shrinkage, structure and properties of paper, In Formation and Structure of Paper, 2nd Fundamental Research Symposium, Oxford, UK, Sept. 1961, 397-413.

Panek, J., Fellers, C. and Haraldsson, T. (2004): Principles of evaluation for the creep of paperboard in constant and cyclic humidity, Nordic Pulp & Paper Research Journal, 19(2): 155-163.

81

Panek, J., Fellers, C., Haraldsson, T. and Mohlin, U.-B. (2005): Effect of fibre shape and fibre distortions on creep properties of kraft paper in constant and cyclic humidity, In Advances in Paper Science and Technology, 13th Fundamental Research Symposium, Cambridge, UK, Sept. 2005, 777-796.

Parker, J. L. (1962): The effects of ethylamine decrystallization of cellulose fibers on the viscoelastic properties of paper, Tappi, 45: 936-43.

Pelton, R. (2004): On the design of polymers for increased paper dry strength - a review, Appita Journal, 57(3): 181-190.

Rance, H. F. (1948): Some new studies in the strength properties of paper, Proc. Tech. Sect., Paper Makers' Assoc. Gt. Brit. & Ireland, 29: 449-69, discussion 470-6.

Rundlöf, M., Karlsson, M., Wågberg, L., Poptoshev, E., Rutland, M. and Claesson, P. (2000): Application of the JKR method to the measurement of adhesion to Langmuir-Blodgett cellulose surfaces, Journal of Colloid and Interface Science, 230(2): 441-447.

Samuelsson, L. G. (1964): Stiffness of Pulp Fibres, Part 2. The Effect of Mechanical Treatment, Svensk Papperstidning, 67(23): 943.

Sanborn, B. I. (1962): A Study of Irreversible, Stress-Induced Changes in the Macrostructure of Paper, Tappi, 45(6): 465-474.

Sandgreen, B. and Wahren, D. (1960): Studies on pulp crill. III. Influence of crill on some properties of pulp and paper, Svensk Papperstidning, 63: 879-83.

Scallan, A. M. (1977): Accommodation of water within pulp fibres, In Fibre-water interactions in papermaking, 6th Fundamental Research Symposium, Oxford, UK, Sept. 1977, 397-413.

Scallan, A. M. and Tigerström, A. C. (1992): Swelling and Elasticity of the Cell-Walls of Pulp Fibers, Journal of Pulp and Paper Science, 18(5): J188-J193.

Schultz-Eklund, O., Fellers, C. and Johansson, P. Å. (1992): Method for the local determination of the thickness and density of paper, Nordic Pulp & Paper Research Journal, 7(3): 133-9, 154.

82

Seth, R. S. and Page, D. H. (1981): The stress strain curve of paper, In The Role of Fundamental Research in Papermaking, 7th Fundamental Research Symposium, Cambridge, UK, Sept. 1981, 421-452.

Stiernstedt, J., Nordgren, N., Wågberg, L., Brumer, H., Gray, D. G. and Rutland, M. W. (2006): Friction and forces between cellulose model surfaces: A comparison, Journal of Colloid and Interface Science, 303(1): 117-123.

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

Stratton, R. A. and Colson, N. L. (1993): Fiber wall damage during bond failure, Nordic Pulp & Paper Research Journal, 8(2): 245-9,257.

Swerin, A., Ödberg, L. and Lindström, T. (1990): Deswelling of hardwood kraft pulp fibers by cationic polymers: the effect on wet pressing and sheet properties, Nordic Pulp & Paper Research Journal, 5(4): 188-96.

Tanaka, H., Ödberg, L., Wågberg, L. and Lindström, T. (1990): Adsorption of cationic polyacrylamides onto monodisperse polystyrene lattices and cellulose fiber: effect of molecular weight and charge density of cationic polyacrylamides, Journal of Colloid and Interface Science, 134(1): 219-28.

Torgnysdotter, A., Kulachenko, A., Gradin, P. and Wågberg, L. (2007): The link between the fiber contact zone and the physical properties of paper: a way to control paper properties, Journal of Composite Materials, 41(13): 1619-1633.

Torgnysdotter, A. and Wågberg, L. (2003): Study of the joint strength between regenerated cellulose fibers and its influence on the sheet strength, Nordic Pulp & Paper Research Journal, 18(4): 455-459.

Torgnysdotter, A. and Wågberg, L. (2004): Influence of electrostatic interactions on fibre/fibre joint and paper strength, Nordic Pulp & Paper Research Journal, 19(4): 440-447.

Uesaka, T. (1991): Dimensional stability of paper. Upgrading paper performance in end use, Journal of Pulp and Paper Science, 17(2): J39-J46.

83

van de Steeg, H. G. M., Dekeizer, A., Stuart, M. A. C. and Bijsterbosch, B. H. (1993): Adsorption of Cationic Amylopectin on Microcrystalline Cellulose, Colloids and Surfaces a-Physicochemical and Engineering Aspects, 70(1): 77-89.

van de Steeg, H. G. M., Stuart, M. A. C., Dekeizer, A. and Bijsterbosch, B. H. (1992): Polyelectrolyte Adsorption - a Subtle Balance of Forces, Langmuir, 8(10): 2538-2546.

van de Ven, T. G. M. (2008): Capillary forces in wet paper, Industrial & Engineering Chemistry Research, 47(19): 7250-7256.

Van den Akker, J. A. (1950): The elastic and rheological properties of papermaking fibers. Some of the problems to be met in developing an analytical approach to the subject, Tappi, 33: 398-402.

Vicini, S., Princi, E., Luciano, G., Franceschi, E., Pedemonte, E., Oldak, D., Kaczmarek, H. and Sionkowska, A. (2004): Thermal analysis and characterisation of cellulose oxidised with sodium methaperiodate, Thermochimica Acta, 418(1-2): 123-130.

Wågberg, L. (2000): Polyelectrolyte adsorption onto cellulose fibers -- a review, Nordic Pulp & Paper Research Journal, 15(5): 586-597.

Wågberg, L. and Annergren, G. (1997): Physicochemical characterization of papermaking fibers, In Fundamentals of Papermaking Materials, 11th Fundamental Research Symposium, Cambridge, UK, Sept. 1997, 1-82.

Wågberg, L., Forsberg, S., Johansson, A. and Juntti, P. (2002): Engineering of fibre surface properties by application of the polyelectrolyte multilayer concept. Part I. Modification of paper strength, Journal of Pulp and Paper Science, 28(7): 222-228.

Wågberg, L. and Hägglund, R. (2001): Kinetics of polyelectrolyte adsorption on cellulosic fibers, Langmuir, 17(4): 1096-1103.

Wågberg, L., Winter, L., Ödberg, L. and Lindström, T. (1987): On the charge stoichiometry upon adsorption of a cationic polyelectrolyte on cellulosic materials, Colloids and Surfaces, 27(1-3): 163-73.

Zeronian, S. H., Hudson, F. L. and Peters, R. H. (1964): The mechanical properties of paper made from periodate oxycellulose pulp and from the same pulp after reduction with borohydride, Tappi, 47(9): 557-64.

84

Zhao, H. and Heindel, N. D. (1991): Determination of degree of substitution of formyl groups in polyaldehyde dextran by the hydroxylamine hydrochloride method, Pharmaceutical Research, 8(3): 400-2.

Östlund, M., Östlund, S., Carlsson, L. A. and Fellers, C. (2004a): The influence of drying restraints and beating degree on residual stress build-up in paperboard, Journal of Pulp and Paper Science, 30(11): 289-293.

Östlund, M., Östlund, S., Carlsson, L. A. and Fellers, C. (2004b): The influence of drying conditions on residual stress build-up in paperboard, Journal of Pulp and Paper Science, 30(11): 312-316.

Östlund, M., Östlund, S., Carlsson, L. A. and Fellers, C. (2005): Experimental determination of residual stresses in paperboard, Experimental Mechanics, 45(6): 493-497.