Improvement of the retention-formation relationship using three-component retention aid systems

Anna Svedberg Doctoral Dissertation Stockholm 2012 Royal Institute of Technology Department of Fibre and Polymer Technology SE-100 44 Stockholm, Sweden

Typeset in Garamond. TRITA-CHE REPORT 2012:19 ISSN 1654-1081 ISBN 978-91-7501-314-5 Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan i Stockholm framlägges till offentlig granskning för avläggande av teknologie doktorsexamen, fredagen den 8 juni 2012 klockan 13.00 i hörsal K1, Teknikringen 56, Stockholm. Avhandlingen försvaras på engelska. Copyright © Svedberg Anna 2012. All rights reserved. Except for the purpose of academic education or research, no part of this publication may be reproduced by any means without the permission of the author. Stockholm 2012, Universitetsservice US AB.

To Oskar and Selma

Abstract Paper formation and filler retention are working in opposite directions during papermaking, where high filler retention correlates with poor formation, and vice versa. A change in this interdependency would gain increased competitiveness of paper materials in terms of better paper properties at profitable process conditions (e.g. less filler content in the white water, reduced wear of forming fabrics and shorter response times to changes in chemical additives). As raw material prices increase, there is a greater interest in new applications in papermaking, such as the production of highly filled papers, and in these contexts the retention-formation relationship also becomes essential. Over the years, the use of and demand for retention aids have increased significantly as a consequence of an enhanced white-water system closure, increased production rates and increased filler content. Knowledge about whether some retention aid systems are more detrimental to paper formation than others has, however, been very limited. At the same time, knowledge regarding other chemical and mechanical factors influencing the retention-formation relationship has also been insufficient. In order to investigate retention-formation relationships, a pilot web former (R-F machine) has been optimized during this thesis work. The R-F machine provides a short circulation of the white-water and a short residence time to chemical equilibrium and is suitable for systematic studies predicting a ranking order that corresponds to what could be expected in a full-scale machine. In this respect this equipment fills an essential gap in earlier available experimental techniques for studying retention-formation relationships. After establishing the routines for this pilot web former, a number of different retention aid systems were evaluated with respect to retention and formation under a given set of experimental conditions. These experiments were performed using a fine paper stock consisting of 90% bleached hardwood and 10% softwood and with an addition of 25% CaCO3 filler (based on solids content). It was shown that single-component retention aids (three cationic polyacrylamides of varied molecular weights) and polyacrylamide-based microparticulate retention aid systems (two systems with different microparticles) resulted in similar retention-formation relationships. The results from all these systems suggested a linear relationship, where an increased filler retention was accompanied by a deterioration in paper formation. The investigation of different retention aid systems also included three-component retention aid systems. These systems combined a cationic polyacrylamide-based microparticulate retention aid system with a formation aid. Three high molecular weight anionic polymers of different structure were used as formation aids (linear, partly structured and structured). It was shown that with such three-component systems, the retention-formation relationship could be improved. At a given level of retention, the formation was significantly improved. The key for this improvement was to add the high molecular weight polymers at such high amounts that the anionic polymer was free in solution and able to disperse the papermaking fibres. It was also found that the structure of the polymer was decisive for the favourable formation effect. The drainage rate, however, was also affected by high added amounts of anionic polymer. In contradiction to the formation improvements, a high amount of added polymer correlated to reduced drainage properties of the fibrous dispersions.

The continued thesis work was focused on finding molecular mechanisms that could explain the findings with the R-F machine, showing a deviation from the linear relationship between retention and formation. In these studies the same three high molecular weight anionic polymers were investigated as single-component additives in a flow loop system which simulated forming conditions (fibre consistency of 5 g/l). In these studies it was shown that the high molecular weight anionic polymers reduced the mean size of the flocs when present in high amounts in the suspension, meaning that these adjuvants can act as deflocculation aids. Again, the polymer structure was found to be important for the dispersion effect. The main hypothesis was that the formation aids, when added at such high amounts that they were free in solution, could suppress the turbulence and thereby gain favourable dispersion effects. From that point of view, both phase flow encoded magnetic resonance imaging (MRI) experiments and adsorption studies on model surfaces were performed. In addition, R-F machine trials were performed to elucidate whether the changed retention-formation relationship was an effect of changed drainage performance. However, these R-F machine trials showed that the detected effect was related to changed chemistry upon the addition of formation aids in high amounts, and not to changed drainage rate of the fibrous furnish. Adsorption studies using QCM and DPI techniques were performed for two three-component retention aid systems, including the linear and the structured anionic polymer. The objective of these studies was to investigate whether surface effects, due to adsorption of the anionic polymer, contributed to changed retention-formation relationships. Since both systems demonstrated similar results and only slight adsorptions of the anionic polymers were observed, it was concluded that interaction effects between the fibre surfaces were not a contributing factor to the changed relationships. Hence, the mechanism was attributed to solution effects where the free anionic polyelectrolyte caused a dispersion of the papermaking fibres. The effect of formation aids on flow properties of cellulosic fibrous suspension in turbulent flows was studied in MRI experiments. The objective was to correlate changes in rheological properties of the suspending medium to changes in formation and fibre flocculation. Flow imaging measurements and pressure drop analysis were performed for water and fibre suspensions (0.5% consistency) with and without additions of formation aids. The results indicated that high molecular weight anionic polymers as formation aids effectively act as drag reducing additives in turbulent flows. Drag reduction effects were shown in terms of reduced turbulence fluctuations (as determined by signal intensities) and reduced pressure drops. Synergy effects were observed when fibres and formation aids were used in combination. Again, the importance of the polymer structure was revealed in these MRI studies in terms of drag reduction effects. The overall results in this thesis work, given from independent R-F machine trials, flow loop experiments and MRI experiments, are all consistent regarding the effect of the formation aids and the influence of the polymer structure. Consistently, the deflocculation effects were shown only for the linear and the partly structured polymer. No effects were shown for the structured polymer. Hence, the extension of the polymer in solution and the polymer flexibility are important characteristics for drag reduction phenomena and the beneficial effects that can be generated therefrom.

To sum up, this work documents for the first time that the established retention-formation relationship can be changed; that is, the formation can be improved at a given level of retention. This was shown in pilot web former trials for three-component systems, combining polyacrylamide-based microparticulate retention aid systems with formation aids. These trends were later also supported by flocculation investigations, model experiment studies and MRI measurements. The mechanism of changed retention-formation relationships and fibre dispersion could hence be related to turbulence damping effects and not to adsorption effects of the formation aids. Turbulence is generated within the headbox nozzle passage to ensure a uniform dispersion of the fibres in the jet stream delivered to the forming fabric, so that an even cross-directional mass distribution of the fibres is obtained in the paper sheet. However, the fibres tend to reflocculate in the decaying turbulence, after passing the turbulence generator in the headbox, and when distributed on the wire. A high degree of turbulence introduced in the headbox, and hence an increased inherent energy in the fibre networks, correlates to a more pronounced degree of reflocculation of the fibres. By using formation aids as drag reducing additives, the turbulence was damped and the degree of reflocculation was thereby reduced, which contributed to improved paper formation. Owing to turbulence damping effects, there is a reduction in the flow hydraulic resistance that is expected to reduce the internal stress of the coherent fibre network. When the internal stress is reduced, the stability of the flocs is reduced and hence a favourable fibre dispersion is achieved. KEYWORDS: Dewatering, drag reduction, fibre flocculation, filler retention, fine paper, flow imaging, flow loop, formation aid, kraft pulp, magnetic resonance imaging, MRI, paper formation, pilot web former, pressure drop, retention aid, turbulence damping

List of publications This thesis is based on the following papers and additional unpublished data. The papers are appended at the end of the thesis.

I. A pilot web former designed to study retention-formation relationships Svedberg, A. and Lindström, T. (2010) Nord. Pulp Paper Res. J., 25(2), pp. 185-194.

II. The effect of various retention aids on retention and formation Svedberg, A. and Lindström, T. (2010) Nord. Pulp Paper Res. J., 25(2), pp. 195-203.

III. Improvement of the retention-formation relationship using three-component retention aid systems Svedberg, A. and Lindström, T. (2012) Nord. Pulp Paper Res. J., 27(1), pp. 86-92.

IV. The effects of formation aids on flow properties of cellulosic fibrous suspensions Svedberg, A., Tozzi, E. J., Lavenson, D. M., Lindström, T., McCarthy, M. J. and Powell, R. L. (2012) Manuscript.

Other relevant disseminations not included in the thesis: 1. Storage of chemical pulp - Inactivation of retention aids

Nordin, L. (2006). Supervised by: Westlund, P-O., Wågberg, L. and Svedberg, A. Degree project in Engineering Chemistry, Umeå University.

2. Valuation of the retention/formation (R/F) relationship using a laboratory pilot-paper machine

Svedberg, A. (2007) Intertech Pira Conference Filler and Pigments for Papermakers, Berlin, Germany.

3. Valuation of retention/formation relationships using a laboratory pilot-paper machine

Svedberg, A. (2007) Licentiate thesis, Department of Fibre and Polymer Technology, Royal Institute of Technology, Stockholm, Sweden.

4. On-line applikation för signifikanta studier av avvattning på en pappersmaskin i pilotskala

Nilsson, P. (2009). Supervised by: Hägglund, J-E. and Svedberg, A. Degree project in Technical Physics, Umeå University.

5. Förbättring av formations/retentionssambandet

Svedberg, A. (2010) Ekmandagarna, Stockholm, Sverige.

6. A strategy of improving the retention-formation relationship by using drag reducing additives

Svedberg, A. and Lindström, T. (2012) 8th International Paper and Coating Chemistry Symposium, Stockholm, Sweden.

7. A strategy of improving the retention-formation relationship by using drag reducing additives Svedberg, A., Tozzi, E. J., Lavenson, D. M., Lindström, T., McCarthy, M. J. and Powell, R. L. (2012) The XVIth International Congress on Rheology - August 5-10, 2012.

Patent applications: Filed patent application “Composition and Process for Improving Paper and Paper Board” on December 15, 2011. Application no. in the US: 61576250. Application no. in Sweden: 1151205-0. Manuscripts in preparation: The effect of formation aids on fiber suspension flocculation Svedberg, A., Ankerfors, M. and Lindström, T. (2012) Adsorption study – The interactions of cationic polyacrylamide, high molecular weight anionic polymer and anionic montmorillonite clay on model surfaces Svedberg, A., Utsel, S., Lindström, T. and Wågberg, L. (2012)

Contents 1 Introduction ........................................................................................................................ 1

1.1 Setting the scene ................................................................................................................ 1

1.2 Objective ........................................................................................................................... 2

1.3 Outline of the thesis ........................................................................................................... 2 2 Background ........................................................................................................................ 3

2.1 Retention and formation .................................................................................................... 3

2.2 The retention-formation relationship ................................................................................. 3

2.3 Paper formation, fibre flocculation and network strength .................................................. 4

2.4 Machine design and operational variables ........................................................................... 5

2.5 Retention aid systems – aspects on retention mechanisms and reversibility of flocculation 7

2.6 Other chemical adjuvants and dosage strategies - aspects on dispersion mechanisms ....... 13

2.7 Drag reduction of turbulent flows by additives ................................................................ 15 3 Materials ........................................................................................................................... 17

3.1 Pulp ................................................................................................................................. 17

3.2 Filler ................................................................................................................................ 18

3.3 Retention aid systems ....................................................................................................... 18

3.4 Formation aids ................................................................................................................. 19 4 Methods ............................................................................................................................ 21

4.1 The retention and formation machine .............................................................................. 21

4.2 Retention and formation .................................................................................................. 24

4.3 Dewatering analysis .......................................................................................................... 24

4.4 Flocculation analysis – The flow loop system ................................................................... 25

4.5 Magnetic resonance flow imaging .................................................................................... 26

4.6 Adsorption analysis (QCM and DPI) ............................................................................... 28 5 Results and Discussion .................................................................................................... 29

5.1 The effect of the jet-to-wire speed ratio on paper formation ............................................ 29

5.2 The effect of different retention aids on retention and formation .................................... 30

5.2.1 Single-component polyacrylamide retention aids ....................................................... 30

5.2.2 Polyacrylamide-based microparticulate systems ......................................................... 32

5.2.3 Three-component retention aid systems .................................................................... 35

5.3 The effect of formation aids on fibre suspension flocculation .......................................... 41

5.4 The effects of formation aids on flow properties of cellulosic fibrous suspensions .......... 43 5.5 Adsorption studies – The interactions of cationic polyacrylamide, high molecular weight

anionic polymer and anionic montmorillonite clay on model surfaces.................................... 46 6 Conclusions ...................................................................................................................... 49 7 Future work ...................................................................................................................... 51 Acknowledgements ............................................................................................................. 52 References ........................................................................................................................... 54

1

1 Introduction

1.1 Setting the scene

Papermakers strive for uniform networks of fibres when producing paper materials in order to optimize paper strength and optical performance. Paper formation describes how the fibres are distributed, arranged and mixed in the sheet structure. The paper formation is important for obtaining good strength and printability properties. Forming of paper materials starts from a dilute aqueous suspension of papermaking pulp (0.2-1% mass consistency depending on the product), which is transformed to a coherent sheet structure during forming on the wire. Most of the water is drained on the wire (96-99%), yielding a wet paper web which is pressed and dried in subsequent steps to form the final product. The resulting paper formation depends on the shearing conditions in the forming zone and the degree of fibre flocculation in the papermaking furnish. Filler is a finely divided, usually white pigment (e.g. calcium carbonate, kaolin and talc). It is added to the stock in order to raise the opacity and smoothness of the paper and also for economic reasons related to the increased costs of wood as a raw material. A fine paper consists of approximately 20% filler and the rest is a mixture of fully bleached chemical hardwood and softwood fibres. Retention aids are stock additives known to be effective for improving the retention of fine particles during the forming of paper. However, retention aids are not selective; they help in attaching fine materials such as fillers to fibres but also cause additional fibre flocculation, which is detrimental to paper formation. Hence, there is a delicate balance between retention and formation. The composition of retention aid systems varies among the supplier portfolios, and it has often been questioned whether some systems are more detrimental to paper formation than others. Although it has been communicated in conference articles that certain developments in retention strategies are superior to formation, there are very few convincing academic investigations found in the open literature. The general trend reported in the literature describes difficulties in breaking the interdependency between retention and formation (Albinsson et al. 1995; Brouillette et al. 2004; Huber et al. 2004; Jokinen, Palonen 1986; Krogerus 1994; Rooks 2004). Although other strategies to improve the retention-formation relationship have been suggested, such as preflocculation of filler (Mabee 2001; Silenius 2003) and the usage of formation aids (Jokinen, Palonen 1986; Lindström et al. 1986), there are few publications available in the literature. Exploration of the literature revealed the need for more focused developments in this area, including systematic studies on the influence and interplay of chemical and mechanical aspects on the interdependency between retention and formation. A simple and cost-effective method for evaluating retention and formation under chemical equilibrium situations has also been identified. The low number of systematic studies found may be explained by the limited number of suitable evaluation methods. Industrial trials are rarely reported, since they are costly and do not include controlled experimental conditions because of variations in the process parameters (e.g. grammage, reuse of broke from different paper machines, etc.). Moreover, the residence time to chemical equilibrium is long on full-scale paper machines, due to large white-water systems. Further, most of the laboratory experimental set-ups

2

for flocculation analysis do not simulate recirculation of the white water, which results in lower concentration of fine materials in the investigated furnishes.

1.2 Objective

The objective of this thesis was threefold. The first objective was a reliable R-F machine, being a significant tool for systematic reproducible studies of retention, formation and dewatering under realistic conditions. The second objective was a strategy to improve the retention-formation relationship. Finally, the third objective was a fundamental understanding about flocculation mechanisms that contribute to an improved retention-formation relationship in papermaking.

1.3 Outline of the thesis

The thesis is based on the four papers appended at the end and additional unpublished data. The thesis is divided into seven main chapters, which are followed by acknowledgements and references. The first two chapters introduce and give background information on the investigations. The next two chapters are self-explanatory, being entitled Materials and Methods. Chapter 5 is the central part of the thesis, presenting the results and discussions. Chapter 5 is divided into five main subchapters. Chapter 5.1 presents the effect of the jet-to-wire speed ratio on paper formation, results which were presented in Paper 1. Chapter 5.2 gives a survey of the impact of various retention aids on retention and formation. The impacts of single-component polyacrylamide retention aids (5.2.1), polyacrylamide-based microparticulate retention aid systems (5.2.2) and three-component retention aid systems (5.2.3) are presented and compared. The main results presented in Paper 2 are discussed in chapters 5.2.1 and 5.2.2, and the main results presented in Paper 3 are discussed in chapter 5.2.3. Chapter 5.3 presents the effect of formation aids on fibre suspension flocculation, which was investigated in a flow loop system. These data are not yet published. The effects of formation aids on flow properties of cellulosic fibrous suspensions, investigated by using magnetic resonance imaging (MRI), are presented in chapter 5.4. These results refer to Paper 4. The last paragraph in chapter 5 presents the results obtained from adsorption studies on model surfaces, in which the interaction between cationic polyacrylamide, high molecular weight anionic polymers and montmorillonite clay was studied. Finally, the thesis work is summarized and the most important findings are given in the Conclusion chapter. The last chapter also presents suggestions for future work.

3

2 Background

2.1 Retention and formation

Two parameters that are almost always critical to good papermaking are paper formation and filler retention.1 Paper formation is one of the most important quality characteristics of paper materials, while filler retention is an important issue with regard to productivity and system stability of the wet end of the paper machine. A sheet with good formation has positive impacts on paper strength (Norman, Wahren 1973; Hallgren, Lindström 1989; Nazhad et al. 2003), printability (Huang, LePoutre 1994; Kajanto 1989; Bernié et al. 2006) and optical properties such as opacity (Jordan 1985; Wahren 1987). On the other hand, there are several advantages to be derived from improved filler retention, such as higher paper machine and additive efficiency, faster responses to changes in process conditions, cost savings and reduced effluent loading. Higher retention gives smaller amounts of circulating materials, which gives smaller changes in the mix composition and hence a more constant product quality but also less material carry-over between paper machines with connected white-water systems. Poor filler retention can also cause sheet quality problems. Uneven filler distribution, as one example, might result in poor optical properties and increased two-sidedness of the produced paper. 1 Retention is defined as the proportion of a component present in the original mixture which remains in the mixture at some stage of the process or in the final product (Fellers, Norman 1998).

2.2 The retention-formation relationship

It is generally recognized that retention and paper formation are opposing parameters in papermaking, where high filler retention correlates to poor paper formation and vice versa. The papermaker has several economic and practical reasons for wanting to attain both high filler retention and good paper formation, that is, to break the interdependency between retention and formation. General trends in modern papermaking include a higher degree of white-water system closure, faster paper machines, more highly filled papers and twin-wire forming. These trends, together with the demands of producing higher-quality paper, shed new light on the retention-formation relationship. The retention-formation relationship is influenced by several chemical and mechanical factors. Retention aid choices, machine designs, operational variables, furnish characteristics and dosage strategies are all factors that influence retention and formation (e.g. Solberg, Wågberg 2003). Other important chemical aspects are the chemistry of the suspending medium and the addition of other chemical adjuvants, such as formation aids. The influences of mechanical factors on paper formation have been reviewed by Norman and Söderberg (2001), and factors influencing the retention-formation relationship have been reviewed by Swerin and Ödberg (1997). To balance retention and formation, the papermaker must acknowledge the interplay between these chemical and mechanical factors. Chemical and mechanical factors influencing the retention-formation relationship are discussed in more detail in the following chapters.

4

2.3 Paper formation, fibre flocculation and network strength

Formation is a concept which describes the grammage distribution of paper on a scale up to 50 mm (Norman 1989; Norman, Söderberg 2001). The final distribution of fibres in paper materials is principally influenced by two interacting phenomena: fibre flocculation and shearing conditions in the forming section (Kiviranta, Dodson 1995; Kiviranta 1996). Fibre flocs, defined as local fibre concentration variations, have a negative impact on paper formation. The correlation between fibre flocculation and paper formation has been addressed by several authors (Wahren 1972; Jokinen, Ebeling 1985; Jokinen, Palonen 1986; Swerin, Ödberg 1996a; Yan et al. 2006). Fibre flocculation depends on both chemical and mechanical factors. Before Mason (1954), the research claimed that flocculation was mainly governed by chemical factors, based on classical colloidal phenomena. Mason first showed that collisions and subsequent mechanical entanglement of fibres were primary factors affecting flocculation of fibres, a concept described as mechanical flocculation. Mason’s mechanistic concept laid the foundation for the subsequent research on fibre flocculation. Mechanical flocculation gives rise to local fibre networks (flocs), where the fibres are mechanically interlocked in the network if the fibre is in contact with at least three other fibres. Regarding fibre flocculation, the number of contacts per fibre is critical. To quantify the number of contacts, Kerekes and Schell (1992) introduced the crowding factor concept (N), where values for N can be calculated from: N=2Cv(L/d)2/3 [1] where Cv is the volumetric concentration of fibres, L is the fibre length and d is the fibre diameter. The crowding factor concept combines fibre geometry with fibre consistency and can be used to characterize the flocculation tendency of a given furnish. When the N-value increases (increased crowding), the probability that fibres will collide increases. Longer fibres, stiffer fibres and an increased consistency will increase the N-value, that is, the flocculation of fibres. Kerekes and Schell (1992) classified fibre suspensions in three regimes (a-c), depending on the crowding factor (a: N<1, dilute; b: 1<N<60, semi-concentrated; c: N>60, concentrated). In commercial papermaking, the range of 1<N<60 is of key importance. From laboratory sedimentation experiments, Martinez et al. (2001) concluded that within the range 1<N<60, there were two different sub-regimes present regarding fibre flocculation. These sub-regimes were delineated at the critical level N=16, where formation was found to be slightly dependent on N in the region N<16 and then to worsen significantly with N>16. The machine design and the operational variables should hence be carefully considered when critical levels for fibre flocculation are discussed. Fibre flocculation affects the behaviour and the rheology of the pulp suspension, influencing the fibre network strength. The network strength of fibre suspensions arises from the cohesive forces that act between fibres at the fibre-fibre contact points. The fibre network strength is one important parameter affecting paper formation (Swerin, Ödberg 1996a; Wahren 1979).

5

The fibre network strength is affected by several factors, such as the addition of retention aids or other chemical adjuvants, fibre consistency, fibre dimensions, fibre properties and several other factors. One simple method to reduce the network strength and thereby to improve the formation is to reduce the fibre consistency to such a low level that the network strength is negligible. However, this method has a negative influence on the retention, which decreases with decreased fibre consistency and naturally on the volumes that have to be passed through the headbox during papermaking operations. The shearing conditions, including turbulence, are essential for fibre flocculation and the subsequent paper formation. Due to shearing of the stock during the dewatering process, fibres and flocs are more evenly distributed, which results in better paper formation in the sheets being produced. Reflocculation of the stock in decaying turbulence is one very important aspect of paper formation. The pulp being deflocculated in the headbox turbulence generator reflocculates quickly in the decaying turbulence when distributed on the wire, and the resulting paper formation will suffer. Meyer and Wahren (1964) have elucidated how and why fibre flocs form after dispersion, offering the following explanation: “…When a fiber suspension is agitated, the fibers are exposed to viscous and dynamic forces that tend to bend them. When agitation ceases, the fibres tend to regain their original unstrained shape. However, if there are many fibers per unit volume, they can not straighten out freely but will come in contact with other fibers. A fraction of the fibers will come in contact with so many others that they will come to rest in a strained position, and forces will be transmitted from fiber to fiber. These fibers become interlocked by normal and frictional forces, constituting a fiber network, where forces can be transmitted through the fibers and from fiber to fiber. Thus, fiber networks are coherent because of internal stresses…”

2.4 Machine design and operational variables

The design of the wet end of the paper machine (e.g. the former type, headbox design, and dewatering and shearing conditions) and the operational variables (e.g. the stock consistency, jet-to-wire speed ratio, machine speed, wire tension and slice opening) have significant impacts on paper formation. By optimizing these variables, it is possible to improve the formation without any loss in retention. The effects of machine design and operational variables on paper formation have been reviewed by Norman (1989) and Norman and Söderberg (2001). Paper machine performance has developed over time to embrace new web-forming techniques and headbox designs, both including increased shearing conditions. The main reasons for these developments were the calls for increased running speeds with higher production, better grammage profiles, better paper formation and less two-sidedness. In medium or high-speed paper machines the turbulence generator is a key part of the headbox. A turbulence generator is basically a tube bank that assists in deflocculating the pulp, since the stock which enters the tube bank is accelerated and fibre flocs are broken. The basic function of the turbulence generator is to even out the cross-directional (CD) mass distribution of fibres, so an even CD mass distribution of fibres in the paper sheet is obtained. When the deflocculated stock leaves the tube bank in the headbox, it starts to reflocculate in the decaying turbulence. The reflocculation tends to be more pronounced with a higher degree of turbulence. Hence, there is a need to dampen the turbulence produced in the headbox. The design of the headbox is important for the resulting paper formation. Recent developments in headbox designs can be

6

seen on some paper machine supplier home pages. For instance, a company offering equipment with optimized turbulence generation and nozzle geometry reports improved paper formation as one of its main benefits. The new turbulence generator configuration and nozzle geometry reduce the residence time and accelerate the flow, which minimize the fibre reflocculation. Both retention and formation are influenced by the stock consistency (Albinsson et al. 1995; Norman 1989). A higher stock consistency gives a higher filler retention but, on the other hand, it also results in poorer paper formation. It has been suggested that paper formation is improved, to a certain extent, by dewatering, adjusting elongational shear and supplying oriented shear and turbulence to the stock during forming (Norman, Söderberg 2001). A higher machine speed correlates to a higher dewatering rate; both properties have consequences for the web formation. Dewatering has a self-healing effect on the grammage distribution (Norman, Söderberg 2001; Norman 1974). A rapid initial drainage, however, tends to set the sheet too early in the forming zone, which affects the formation negatively. A higher stock temperature correlates to a more rapid initial drainage. When accelerating the stock in the headbox nozzle, an elongational flow is developed. This elongational flow tends to orient fibres and cause flocs to stretch or rupture. By changing the headbox contraction ratio, the overall acceleration of the flow can be modified. Norman and Söderberg (2001) have suggested that a high contraction ratio2 should be set in order to achieve improved paper formation. By adjusting the jet-to-wire speed ratio to such an extent that a small difference in velocity of the jet and wire occurs, a certain oriented shear is achieved, which is favourable for paper formation (Swerin, Mähler 1996; Svensson, Österberg 1965). The jet-to-wire speed ratio has not, however, shown any considerable effect on the retention level (Swerin, Mähler 1996; Lindström et al. 2006). Another way to introduce shear during forming is by varying the wire tension in a blade former, influencing the dewatering pulses. For filled papers, Albinsson et al. (1995) investigated, the effect of varied wire tensions on paper formation, using the experimental paper machine, EUROFEX, in a twin-wire blade former configuration. This study reported that an increased wire tension gave an improved formation without affecting the retention level. Another factor that has been shown to be important for paper formation is the jet impact, including the free jet length and the jet angle (Norman 1989). 2 Here contraction ratio is defined as the ratio between headbox height at the exit from the tube bank into the contraction and the headbox jet thickness (slice opening).

7

2.5 Retention aid systems – aspects on retention mechanisms and reversibility of flocculation

Retention aids are used to retain filler and fines in the wet paper web during the forming process, by aggregating these stock components to larger units. Hence, retention aids also cause fibres to aggregate, which causes mass formation to deteriorate. This chapter briefly reviews the development of retention aids with some common materials used, and it discusses retention mechanisms and reversibility of flocculation. The mechanism of action and the development of retention aids has been reviewed by several authors, including Lindström (1989), Eklund and Lindström (1991), Horn and Linhart (1996) and Swerin and Ödberg (1997).

Development of retention aid systems Early retention aids were single-component products, most often based on acrylamide chemistry, alum, starch, polyamines or polyethyleneimines (PEI). A further development of these retention aids was the dual-component systems, which are usually based on interactions between a long-chain charged polyelectrolyte and a second polymer with the opposite charge. The microparticulate systems were introduced some 25 years ago, and these systems still dominate the retention aid market. Microparticulate retention aids are normally based on combinations of cationic polymers and anionic inorganic microparticles (Eklund, Lindström 1991). Some work using inorganic cationic microparticles has also been reported in the literature (e.g. Ovenden, Xiao 2002). The first two commercial microparticle retention aid systems used anionic colloidal silica in combination with cationic starch (Andersson 1984) and anionic montmorillonite clay in combination with cationic polyacrylamide (Langley, Litchfield 1986). Later, Smith (1991) reported on systems based on cationic polyacrylamide together with anionic silica sol. Since these precursors, there have been several further developments in the use of microparticulate systems, based on, for example, aluminium hydroxide sols (Lindström et al. 1989), polymeric aluminium species (Carré 1993) and nano-sized inorganic oxides (Gerli et al. 1999). Moreover, modifications of silica sols have also been developed (Jaycock, Swales 1994; Swerin et al. 1995) together with on-site production of the microparticles (Moffet 1994). Today, there are still ongoing developments in the area of retention and dewatering aid systems. For certain paper grades, a development of new systems is the multi-component systems. In 1986, Lindström et al. reported the use of long-chain polymers in conjunction with dual-component systems. It has recently been demonstrated, both in laboratory evaluations and in machine trials, that it is possible to change the retention-formation relationship by using a new multi-component organic/inorganic system (Ledda et al. 2005). The retention-formation relationship was obtained by decoupling the effects of retention, drainage/dewatering and formation. There has also been considerable progress in structured microparticles, such as micro-aggregated silica sols (Andersson, Lindgren 1996), and also systems based on organic polyacrylate gel particles (Honig et al. 1993, 2000, 2005).

8

Aspects on retention mechanisms and reversibility of flocculation There are several functional mechanisms describing the action of different retention aid systems. Table 1 summarizes some different flocculation mechanisms and the influence of shear and soluble substances on these mechanisms (Eklund, Lindström 1991). Table 1. Survey of retention mechanisms and the impact of shear and soluble substances (Eklund, Lindström 1991). Mechanism Shear resistance of the

floc Effect of increased content of dissolved anionic polymers

Effect of simple electrolytes

Charge neutralization - -- +

Patch flocculation + -- -

Bridging -adsorption flocculation

++ -- --

Bridging -sensitization flocculation

++ -/+ ++

Complex flocculation -classical dual systems

+++ -- --

Complex flocculation -microparticulate systems

+++ -- --

Network flocculation +++ -/+ -/+

-, -- means retention decrease +, ++ means retention increase The flocculation mechanism for the particles in a suspension is determined by the physical and chemical surface characteristics of the particles and by the physical and chemical nature of the retention aid system used (Eklund, Lindström 1991). Particles (fibres, fillers and fines) assume an electric charge when dispersed in water. This charge originates from dissociation of ionic groups on the particle, adsorption of ions and isomorphous substitution3 (van Olphen 1977). The charged particles will affect the distribution of ions in the surrounding interfacial region, resulting in an electrostatic double layer around each particle (inner Stern layer and outer diffuse layer) (Evans, Wennerström 1999).

The charged surface might be described in terms of surface charge density and surface potential. The surface potential decays exponentially with distance, with a decay constant given by the Debye length (1/Κ), which is an estimate of the thickness of the electrostatic double layer (Evans, Wennerström 1999). The Debye length is strongly influenced by the ionic strength and the valency of the counter-ions. Two forces are dominating the interaction between particles in solution: electrostatic repulsion and attraction by dispersion forces (London forces). The DLVO theory (Derjaguin-Landau-Verwey-Overbeek) describes the colloidal stability of a dispersed system from the total potential energy between particles, which is the sum of the attractive and repulsive energies (Evans, Wennerström 1999). According to the total energy balance from the DLVO theory, illustrated in Fig 1, coagulation between colliding particles occurs in primary and secondary minima of the potential energy, where the attractive energy (VA) overcomes the repulsive energy (VR). 3 Isomorphous substitution: The replacement of one atom by another of similar size in a crystal lattice without disrupting or changing the crystal structure of a mineral.

9

In comparison with coagulation in the primary minimum, coagulation in the secondary minimum gives rise to flocs which are easier to redisperse. The order of the repulsive barrier (Vmax) is decisive for primary coagulation to occur and correlates to the Debye length (Vmax decreases with smaller Debye length). By reducing the surface potential, the electrostatic repulsive barrier (Vmax) is reduced and coagulation is promoted.

Fig 1. Total potential energy between two colloidal particles according to the DLVO theory (Kruyt 1952). According to the charge neutralization mechanism, the charges on the particle surfaces are neutralized by adsorbed polyelectrolytes. Attractive van der Waals forces will then, as described by the DLVO theory (Evans, Wennerström 1999), destabilize the system and flocculation occurs. However, if the amount of polyelectrolyte is increased further, above the amount required for charge neutralization, the particle charge is reversed and the dispersion is again stable. Even if the charge neutralization mechanism can be demonstrated in the laboratory at low shear conditions, this mechanism alone is not widely used in practical papermaking today because of the high shear forces on fast-running paper machines. For single-component retention aids based on polymers, bridging (LaMer, Healy 1963) and patch flocculation (Gregory 1973; Kasper 1971) are the two principal flocculation mechanisms. The bridging and patch flocculation mechanisms are illustrated in Fig 2.

10

a) b) c) Fig 2. Schematic illustrations of (a) patch flocculation (Horn, Linhart 1991), (b) bridging (Horn, Linhart 1991) and (c) bridging, effect of shear (Hubbe 2005). Cationic condensation polymers of high charge density and low or moderate molecular weight are normally considered to be patch flocculants (Fig 2a). The concept of patch flocculation is to form cationic “patches” on the negatively charged particles, which will attract oppositely charged parts on another particle. The patches are formed by adsorbing cationic polymers of high charge density in a flat conformation. Patch flocculation is promoted by the long-range attractive van der Waals forces and attractive electrostatic forces. Hence, flocs formed by patch flocculation are more shear resistant than flocs formed by charge neutralization, where only attractive dispersion forces act. Patch flocculation is favoured by sensitization, that is, reducing the thickness of the double layer with electrolytes. Too high an electrolyte concentration, however, has a negative effect on patch flocculation, which distinguishes the patch flocculation mechanism from charge neutralization. The degree of patch flocculation is dependent on the charge density of the polymer and the degree of surface coverage, where maximum flocculation is often found around the isoelectric point (i.e.p.). Fibre flocculation induced by polycations of high molar weight and low or moderate charge density is usually described by the bridging adsorption flocculation model (Fig 2b). Bridging is a common feature for retention aid systems used today (single-component products and dual and microparticulate systems). In bridging adsorption theory, polymers adsorb to particle surfaces, forming loops and tails, which act as bridges between particles. The anchoring of the polymer on the surface and the conformation of the formed loops and tails are two parameters essential for bridging. These two parameters are determined by the charge balance between polymer and particle, the contact time affecting the conformation of the adsorbed polymer, non-ionic interactions between the polymer and the surface and other properties of the solution (electrolyte content, etc.). In general, the protruding loops and tails must exceed twice the double layer thickness (2/Κ) for bridging to occur. As emphasized, the charge density of the polymer and the particles are decisive for bridging. Too low a charge density of the polymer or the surface gives poor anchoring and, with that, the formation of weak flocs. On the other hand, too high a charge density (polymer or surface) will result in too strong interactions, giving polymers adsorbed in a flat conformation, which are unfavourable for bridging. In such situations, the mechanism is transferred to patch flocculation or charge neutralization. Moderate charge interaction between polymer and surface is therefore

11

desirable. A moderate charge density also favours the polymer to adopt a more extended conformation, due to charge repulsion. According to La Mer’s bridging theory, Ufloc~ α(1- α), the rate of flocculation (Ufloc) is proportional to both the fraction of surface covered by polymer (α) and the fraction of uncovered surface (1- α). Hence, optimum flocculation occurs when α = 0.5, but the flocculation starts before charge reversal. Increased electrolyte content negatively affects bridging flocculation. The adsorption is reduced because the polymers coil up and will have a lower extension out from the surface once adsorbed on the particle surface. The thickness of the double layer may, however, be reduced by a smaller electrolyte content, which promotes bridging (sensitization). Flocculation of negatively charged colloids by anionic polyelectrolytes or non-ionic polymers (e.g. polyacrylamide) is described as sensitization flocculation (Ives 1978). Sensitization flocculation is only possible if an appropriate electrolyte concentration is present in the solution, which screens the surface charges and then reduces the thickness of the double layer. In these conditions, the flocculation results from adsorption of polymers/polyelectrolytes on the colloid surface and from the bridging of the polymer chains between the colloids. Flocs created in bridging flocculation are generally more shear resistant than flocs created by patch flocculation. Patch formation, however, in comparison with bridging, is considered to be more reversible due to the reversible cohesive electrostatic attractions. Patch flocculants are adsorbed in a flat conformation, where chain breakage is less likely to occur during shearing. For bridging flocculants, however, both polymer chain cleavage (Tanaka et al. 1993; Sikora, Stratton 1981) and kinetic factors such as reconformation towards a flatter conformation (Wågberg et al. 1988) can account for the more or less irreversible nature of the flocculation. The floc structure, important for retention, formation and dewatering, may vary depending on the flocculation mechanism. Patch-charge flocs are generally compact and their dimensions are considerably smaller than those formed by bridging flocculation, which usually have a loose, voluminous structure (Horn, Linhart 1991). Charge neutralization, patch formation and bridging as separate retention mechanisms are not sufficient on modern high-speed paper machines including high shear, because the formed flocs are too weak and more or less irreversible. Patch flocculation aids, however, are normally used on low-speed board machines as dewatering agents, where their more reversible nature of flocculation is favourable for dewatering. They are also used in wood containing furnishes for production of publication papers where their main action is to retain dissolved and colloidal material from the mechanical pulps and also to improve filler retention. As such they are still successful since they can avoid large scale fibre flocculation while maintaining their retention increasing action. By using combinations of flocculants, synergistic effects can be achieved resulting in stronger flocs and a higher degree of reflocculation after dispersion. Hence, the dual complex flocculation systems were developed from the single-component retention aids. Depending on the interacting components, the dual-complex flocculation systems are divided into classical dual systems and microparticulate systems (Eklund, Lindström 1991; Swerin, Ödberg 1997). The idealized mechanisms of classical dual systems and microparticulate systems are illustrated in Fig 3.

12

a) b) Fig 3. Schematic illustrations of (a) classical dual systems (Eklund, Lindström 1991) and (b) microparticulate systems (Hubbe 2005). The classical dual systems are often based on a short-chain cationic polymer of high charge, followed by an anionic polymer of high molecular weight. First, the cationic polymers create a primary flocculation according to the patch model, and second, the anionic polymers link together the primary flocs according to interparticle bridging (Aksberg, Ödberg 1990). The combination of both patch and bridging flocculation and the strong electrostatic interaction between the oppositely charged polymers explains the higher floc strength and higher resistance to shear. Like patch and bridging flocculants, the classical dual systems are also favoured by sensitization. However, at too high salt concentrations the efficiency is again decreased due to a decrease in interaction between the polymer and the solids and eventually a desorption of polymer will occur. None of the classical dual systems have the ability to reflocculate sufficiently after floc disruption, resulting in poorer retention levels at high shear rates. The incapability to reflocculate can partly be explained by polymer chain cleavage during dispersion, which hinders polymers to reflocculate. Microparticulate systems are most often based on combinations of cationic polymers and anionic microparticles. The general idea behind these systems is to add the cationic polymer at an early stage to the stock, inducing a primary large-scale flocculation (bridging) which is broken down by shear. Next, the microparticles are added close to the headbox, creating a secondary flocculation (bridging), mainly governed by charge interactions with the adsorbed cationic polymers. Due to the electrostatic interactions and the higher amount of added polyelectrolyte the flocculation induced by microparticle addition is less sensitive to reconformation and chain cleavage of the adsorbed cationic polymers (Swerin et al. 1996). Most of today’s commercial microparticulate retention aids are able to achieve acceptable levels of filler retention, even in high-speed twin-wire formers. This is partly explained by their ability to produce shear-resistant flocs which can also reflocculate after dispersion. Lindström (1989) first described the microparticulate retention aid systems as reversible, giving a more rapid dewatering in the wire and press section and also a higher porosity of the dried sheet. Swerin (1995) and Swerin et al. (1997) compared cationic polyacrylamide (C-PAM), as a single-component product, with a microparticulate system in terms of the flocculation reversibility. They found that C-PAM alone gave a decaying trend in floc size with time, whereas the addition of microparticles resulted in a high degree of reflocculation; that is, the decaying trend was reduced. That microparticulate systems have a more reversible flocculation pattern has also been shown by Alfano et al. (1999), Hedborg and Lindström (1996) and also Clémençon and Gerli

13

(1999). Swerin et al. (1996) also developed a semiquantitative model for description of the mechanism behind the microparticle systems where the surface coverage of the polymer and the microparticle, as well as the conformation of the polymer on the particle surfaces were taken into account. The real reason why reflocculation after dispersion is possible with microparticulate systems is probably due to the lack of failure at the interface between the microparticles and the cationic polymer adsorbed on the particle surface (Eklund, Lindström 1991). Moreover, a number of studies also claim that the stock reflocculates to smaller and denser flocs when microparticulate retention aids are used (e.g. Hedborg, Lindström 1996; Wall et al. 1992 and Swerin et al. 1993). This also explains the increased drainage performance of microparticulate retention aid systems. In contrast to the earlier-mentioned flocculation mechanisms, network flocculation is envisioned to occur in solution rather than on particle surfaces (Lindström 1989). The mechanism of network flocculation is basically to form diluted transient three-dimensional networks, in which dispersed material can be occluded. The transient network may be formed by cross-linking a cationic polyelectrolyte with an anionic component, by electrostatic interactions or hydrogen bonds, for example, the polyethylene oxide (PEO) –phenolic resin systems. The transient networks are unstable and a collector is needed to collect the dispersed material. Without the collector, which is often made of fibres, the transient network breaks apart to a noneffective colloidal dispersion before flocculation occurs. The irreversible nature of the network, however, is essential for the dewatering, since a gelatinous network structure would diminish the dewatering capacity. The network flocculants are also characteristically independent of the dispersed particle surface characteristics, where the particle separation instead takes place with respect to particle size. These type of retention aids have been used a lot for paper production, specially in Canada, and have attracted a considerable amount of research (e.g. van de Ven and Alince (1996) and Xiao et al. (1996). In comparison with, for example, microparticulate systems, many non-ionic network flocculants are generally less sensitive to high contents of dissolved organic material and colloids in the system if the interaction between the network components is not based on electrostatic interactions. Considering the shear resistance of the flocs created by network flocculants, the stability to shear is relatively high. This may be explained by the increased number of junction points between the constituent parts, generated when the networks concentrate after flocculation (Lindström 1989). It should be stressed, however, that the knowledge regarding the mechanisms of microparticulate retention aid systems is by no means extensive, which indeed holds for the other flocculation mechanisms as well. Since there is a complexity in defining real systems (fibres, chemicals, chemical equilibrium situations and hydrodynamic situations), further research on retention mechanisms is needed.

2.6 Other chemical adjuvants and dosage strategies - aspects on dispersion mechanisms

Apart from retention aid systems, the state of flocculation is also affected by other chemical adjuvants (e.g. dispersion aids), dosage strategies (e.g. preflocculation of filler) and properties of the suspending medium (e.g. medium temperature).

14

Dispersion aids are chemical additives known to decrease fibre flocculation, thereby also enhancing paper formation when added to the stock prior to the headbox. They have been known for a long time and rather significant studies were performed already in the 1930´s by Woolwage (1939) and Erspamer (1940). Dispersion aids include chemical adjuvants affecting the properties of the suspending medium and/or surface characteristics (fibre-fibre friction, surface charge of fibres and steric hindrance between fibres). Dispersion aids may be grouped as follows:

1. Additives increasing the dispersion medium viscosity (Soszynski, Kerekes 1988; Zhao, Kerekes 1993).

2. Formation aids (Class I): Additives (e.g. gums, mucilages, xyloglucan and

carboxymethylated cellulose) which are believed to decrease the coefficient of friction between fibres (Woolwage 1939; Erspamer 1940; De Roos 1958; Beghello, Lindström 1998; Yan et al. 2006; Stiernstedt et al. 2006; Horvath, Lindström 2007).

3. Formation aids (Class II): High molecular weight additives affecting the rheological

properties of the suspending medium (Wasser 1978; Jokinen, Palonen 1986; Lee, Lindström 1989; Yan et al. 2006).

Besides the additives, the temperature of the suspending medium also influences the state of fibre flocculation by influencing the medium viscosity (Yan et al. 2006). Reducing the medium temperature decreases the fibre suspension flocculation due to a decreased inertia/drag ratio of the fibres, resulting in decreased fibre-fibre collision efficiency factor. Formation aids include both natural (e.g. root extracts, gums, mucilages) and synthetic polymers (high molecular weight anionic or non-ionic polymers). Class I and Class II formation aids are two classes separated with respect to their anticipated functional mechanism. Class I formation aids are known to decrease the friction between fibres by changing the surface chemistry of the fibres. Beghello and Lindström (1998) reported that carboxymethylation of cellulosic fibres had a profound dispersing effect by decreasing the fibre-fibre frictional forces provided by the electrostatic repulsion between fibres. This report was supported by the work of Yan et al. (2006) and Horvath and Lindström (2007), demonstrating that fibre surface modifications (carboxymethyl cellulose adsorption and xyloglucan adsorption) can reduce the fibre flocculation in suspension by decreasing the fibre-fibre friction and hence weakening the network strength. The probable mechanism of dispersing fibre suspensions by friction-reduced aids has been discussed by Beghello and Lindström (1998), who claim that decreased fibre friction decreases the internal stress in the flocs. When the internal stress, which keeps the flocs together, is released, the stability of the flocs is reduced and dispersion is favoured. There has also been some extensive work on xyloglucan in cellulose modification (Stiernstedt et al. 2000; Stiernstedt et al. 2006; Zhou et al. 2007). These studies reveal that xyloglucan adsorbs strongly to cellulose, which gives a reduced coefficient of friction between cellulosic surfaces. Class II formation aids, including high molecular weight polymers such as anionic polyacrylamide, are known to decrease flocculation of fibres in suspension as single-component additives, a mechanism favourable to paper formation (Wasser 1978; Jokinen, Palonen 1986; Lee, Lindström 1989; Yan et al. 2006). In this specific case, adsorption is not required for dispersion.

15

There are also a few publications claiming that high molecular weight anionic polymers can be used in conjunction with retention aid systems to improve the retention-formation relationship (Jokinen, Palonen 1986; Lindström et al. 1986). The most beneficial effect on formation was reported when the addition of the high molecular weight anionic polymer was raised above the polymer dosage for maximum retention. It has been hypothesized that Class II formation aids suppress the intensity of turbulence in turbulent flows (drag reduction), which, in turn, favours paper formation (Lee, Lindström 1989). It has been reported (e.g. by Lee, Lindström 1989) that the use of high molecular weight polymers as dispersion aids has a strong negative impact on the drainage of the wet web, and that this type of system therefore has limited practical utility. However, in recent years there has been renewed interest in applications including dispersion aids. In the mid 80’s, Bown (1985) and Hayes (1985) reviewed the field of preflocculation of fillers, a field often referred to when discussing the production of highly filled papers. Preflocculation of fillers as a strategy to improve the retention-formation relationship has been widely discussed among papermakers, although the literature reveals only a few publications on this issue. However, there are investigations showing that the retention-formation relationship could be improved by using preflocculation of fillers as an alternative dosage strategy in papermaking (Mabee 2001; Silenius 2003).

2.7 Drag reduction of turbulent flows by additives

Drag reduction is manifested by decreased friction in turbulent flows. Drag reduction of turbulent flows can be obtained by using additives or by passive means (e.g. by adding special roughness elements to the wall), where the efficiency for additives is higher than that produced by passive means. The effect of drag reduction is used in many applications for energy savings and increasing the flow rate (e.g. the use of polymers in crude-oil pipelines). The three main classes of drag reducing additives are polymers (e.g. polyacrylamide and polyethylene oxide), surfactants (e.g. soaps) and fibres (e.g. cellulosic and asbestos). Equivalently, these drag reducing additives reduce the pressure drop (ΔP) over a considered pipe length by reducing the wall shear stress. The phenomenon of drag reduction by additives has been described in earlier reviews (e.g. Gyr, Bewersdorff 1995; Radin et al. 1975; Ptasinski et al. 2001). One way to report on drag reduction is by giving the percent drag reduction (DR %), calculated from: DR (%)=(1-(ΔPsolution/ΔPNewtonian solvent))*100 [2] where ΔPsolution and ΔPNewtonian solvent are the recorded pressure drops in the solution and Newtonian solvent, respectively. As described in a previous chapter, reflocculation of fibres in decaying turbulence impairs the paper formation. The reflocculation tends to be more pronounced with a higher degree of turbulence. In this work, drag reduction is referred to in the discussion of turbulence damping by formation aids. Drag reduction by fibres was discovered 75 years ago, and by polymers and surfactants, 65 years ago. Toms (1948) was the first to discover that additions of minute amounts of high molecular

16

weight polymers could reduce the pressure drops in a turbulent pipe flow. At that time it was already known that cellulosic fibre suspensions showed the same effect. Further, the use of fibrous suspensions in combination with polymers can generate synergistic effects when compared to their individual effects on drag reduction under turbulent flow conditions (Lee et al. 1974; McComb, Chan 1981). The mechanism of drag reduction by additives is very complex and cannot be reduced to one effect; hence a broad description is needed. One hypothesis explains drag reduction through stretching and alignment of the additives in solution, which is mainly manifested in changed energy balances (energy dissipation) and changed structures of turbulence (large-scale versus small-scale turbulence). Turbulent flows are characterized by velocity fluctuations. It is thought that the drag reducing additives reduce the internal stress of the fibre flocs and also enhance the anisotropy of the dispersed fibres. The velocity fluctuations in the flow direction are normally enhanced, whereas the ones perpendicular to this direction are reduced. Relevant parameters for drag reduction by additives are the geometry and flexibility of the additive, the concentration of the additive, the degree of turbulence (present Reynolds number) and the wall condition (roughness). Drag reducing additives are characterized by high molecular weight (usually of the order of 106 to 107g/mol), an extreme aspect ratio and rather high extensibility. The additive needs to become oriented and elongated in a shear flow for drag reduction to occur. The drag reduction effect is concentration dependent, meaning that for any type of drag reducing additive, the effect is largest at a specific concentration (the particle-to-flow interaction is dominant). In concentrations higher than the specific concentration the effect is decreased, and, finally, at even higher concentrations, disappears since the interactions will be dominated by the particle-to-particle interactions. The drag reduction effect starts at a certain Reynolds number of the flow, larger than some threshold which depends on the concentration and the nature of the additive. The strain field of the flow needs to be strong enough to stretch and align the additives. After a maximum effect at a certain Reynolds number, the effect is decreased and finally terminated when the Reynolds number is increased even more. Two possible reasons for a terminated effect at a certain Reynolds number are degradation of the additive due to high shear and misalignment of the additives.

17

3 Materials

3.1 Pulp

All investigations in this work were performed with wet refined bleached kraft pulp, consisting of 90% hardwood (mainly birch 90-96%) and 10% softwood (about 50% spruce, 50% pine). The bleaching sequence was O2/ClO2/H2O2/ClO2/H2O2 for both the hardwood and softwood pulp. Unless stated elsewhere, the pulp was collected from M-real Husum Paper Mill in Sweden, at a consistency of 4.5-5.0% (by weight). All concentrations given in the text are defined by weight. The hardwood and softwood pulps were collected from separate lines in the mill, at positions before the blending chest (long circulation). The pulps were subsequently mixed to the ratio 90:10 (hardwood : softwood). The average size of the hardwood fibres was approximately 0.85-0.95 mm in length and 26 µm in width, and for the softwood fibres the average size was approximately 1.9-2.0 mm and 31 µm, respectively (PulpEye measurements, Eurocon Analyzer, Sweden). The drainability of the pulp was about 18.0-19.2 SR (determined according to ISO 5267-1). Henceforth, the reference characteristics of the pulp described in this paragraph are referred to when using the denomination pulp* later in the text. In all trials with the R-F-machine (chapter 5.1, 5.2), the pulp* was diluted with tap water in the pulp chest to a target consistency of 3.0% (down from 4.5-5.0%) and to a volume of 18 m3. The pH and the conductivity of the 3.0% furnish were adjusted to specific intervals by adding salts. NaHCO3 was used to adjust the pH to 8.0-8.4 and CaCl2 was used to adjust the conductivity to 600-800 µS/cm. There were also routines for other furnish characteristics, such as drainability (water retention value), temperature, cationic demand and storage time of pulp. In all cases, the pulp was stored for five days in the pulp chest before experiments, and no experiments were performed with pulp that had been stored for more than seven days. By following this routine, the dissolution of dissolved substances from the pulps and the temperature was controlled. The standard furnish characteristics in R-F-machine experiments are summarized in Table 2. Table 2. Standard furnish characteristics in R-F-machine experiments. Characteristics Reference interval

Consistency 2.5-3.0%

Hardwood:softwood ratio 90:10

Temperature 23-30 ºC

Drainability 18.0-20.0 SR

Conductivity 600-800 µS/cm

pH 8.0-8.4

Cationic demand 25-46 µeq/l (PCD-03 Mütek™ measurements)

For the flocculation studies performed with the aid of the flow loop system, fibre suspensions were prepared from pulp* and tap water, to a consistency of 0.5%. The pulp (unbeaten) was obtained at the last filter in the bleach plant, at a concentration of 17% (exception from pulp*). For all flow loop studies, the stock temperature was 20 °C, the pH was adjusted to 7.0 (with NaHCO3) and the conductivity was adjusted to 700 µS/cm (with CaCl2). Cellulose fibre suspensions were prepared for the MRI experiments (chapter 5.4) from pulp* and deionized water. For all MRI experiments, the fibre concentration was 0.5%, the pH was 8.2±0.2, the conductivity was 540±10 µS/cm and the temperature varied between 21 °C and 24

18

°C. The pulp (unbeaten) used in the MRI experiments was obtained at the last filter in the bleach plant, at a concentration of 17% (exception from pulp*).

3.2 Filler

Two different fillers were used in this work: a ground calcium carbonate (GCC) and a precipitated calcium carbonate (PCC). Both fillers were delivered from Imerys Minerals Ltd., Sweden. The GCC was used for the experiments reported in chapter 5.1 (The effect of the jet-to-wire speed ratio on paper formation), 5.2.1 (Single-component polyacrylamide retention aids) and 5.2.2 (Polyacrylamide-based microparticulate systems). The PCC was used in the experiments reported in chapter 5.2.3 (Three-component retention aid systems). The characteristics of the fillers were given by the supplier (note the differences in characterization of the two fillers). The GCC slurry was delivered with an anionic dispersant at a dry solids content of 67.0±1%. The cationic demand of the GCC was 8.7 µeqv/g and the Z-potential was -56.2 mV (MütekTM System Zeta Potential 06). The density of the GCC was 1.73±0.02 g/cm3, and 60±3% of all filler particles had a particle size less than 2 micrometers in diameter. The PCC slurry was delivered undispersed at a dry solids content of 19.0%. The charge of the PCC was slightly anionic; the cationic demand was approximately 0.3-0.4 µeqv/g. The charge was measured by titrating 5 ml of slurry in deionized water (total weight 10 g) with Polydiallyl-dimethylammonium chloride (Mütek™ Particle Charge Detector). The density of the filler was 1.14±0.01 g/cm3 and the D50 was 2.50-2.60 µm.

3.3 Retention aid systems

The effect of different retention aids on retention and formation has been evaluated in this work, including single-component, microparticulate and multi-component retention aid systems. The effects of single-component retention aids are presented in chapter 5.2.1, where three linear cationic polyacrylamides (C-PAM) of different molecular weight were investigated (Table 3). These cationic polyacrylamides, all supplied by Eka Chemicals, were co-polymers of acrylamide and N,N,N-trimethylamino-ethylacrylate. In chapter 5.2.2 the investigation of two different microparticulate retention aids systems is presented. These systems were based on a linear cationic polyacrylamide in combination with one of two different microparticles: anionic montmorillonite clay and anionic colloidal silica sol (Table 4). The montmorillonite clay (Hydrocol SH) was supplied by BASF and the silica sol (Silica NP) was supplied by Eka Chemicals. Silica NP is an alumina modified, to maintain charge at lower pH, silica sol designed for interactions with synthetic polymers. Three-component retention aid systems have also been evaluated with respect to retention, formation and drainage (chapter 5.2.3). These three-component systems combine formation aids with a polyacrylamide-based microparticulate retention aid system (Table 5). As formation aids, three high molecular weight anionic polymers were used, of different degree of cross-linking, here defined as linear (Percol 156), partly structured (M305) and structured (M200). The microparticulate retention aid system was based on cationic polyacrylamide (Percol 3035) and montmorillonite clay (Hydrocol SH). The addition order was cationic polyacrylamide, anionic high molecular weight polymer and, last, the microparticles. All components in these three-

19

component retention aid systems were supplied by BASF. The cationic polyacrylamide (Percol 3035) was a co-polymer of acrylamide and N,N,N-trimethylamino-ethylacrylate. Table 3. Characteristics of the single-component retention aids, all given by the supplier. Component Molecular weight

(million Dalton) Charge density (meq/g)

C-PAM (A) 2.9-4.4 0.82A

C-PAM (B) 6.0-7.9 1.02A

C-PAM (C) >10.8 1.06A

Table 4. Characteristics of the polyacrylamide-based microparticulate retention aid systems, all given by the suppliers. Components 1st comp./2nd comp.

Charge density (meq/g)

Surface area (when hydrated), (m2/g)

Particle size (nm)

C-PAM (B)/Montmorillonite clay

1.02A/-0.12B -/800 -/1*600*300 (flakes)

C-PAM (B)/Silica sol 1.02A/-0.071B -/850 -/3 (colloids)

Table 5. Characteristics of the three-component retention aid systems, all given by the supplier. Components 1st comp./2nd comp./3rd comp.

Charge density (meq/g)

Intrinsic viscosity

C-PAM (D)/Linear polymer/Montmorillonite clay 1.15A/-1.76C/-0.34B 11D/14D/30E

C-PAM (D)/Partly structured polymer/Montmorillonite clay

1.15A/-2.16 C/-0.34B 11D/10D/30E

C-PAM (D)/Structured polymer/Montmorillonite clay 1.15A/-2.50 C-0.34B 11D/2E/30E

A The charge density of the cationic polyacrylamides was measured by polyelectrolyte titration, with polyvinylsulfate of potassium (0.001N) as titrating reagent. B The charge density of the microparticles was measured with MütekTM Particle Charge Detector (PCD), according to the PAP-SOP 01-19 method. C The charge density of the anionic polymers was measured by polyelectrolyte titration, with polydiallyldimethylammonium chloride (0.001N) as titrating reagent. D Intrinsic viscosity values given in dl/g: A suspended-level viscometer was used to determine the specific viscosity of the test component at various concentrations in buffered 1M sodium chloride solution. A plot was made of reduced specific viscosity against concentration, and the intrinsic viscosity was obtained by extrapolation to infinite dilution. The longer the polymer chains, the higher the intrinsic viscosity (dl/g). The test method is referred to as ACSMOT No: 7. E Standard viscosity values given in mPas. The value given for the montmorillonite clay is the direct bulk viscosity of a 5% solution. A Brookfield LVT viscometer was used to characterize the standard viscosity of the anionic polymer (0.1% solution). This method is referred to as L.A. Test Method 20.

3.4 Formation aids

High molecular weight anionic polymers of different structures (cross-linking levels) were used in this work as formation aids (Class II formation aids). The cross-linking level was altered to different degrees, defined as linear (Percol 156), partly structured (M305) and structured M200. The formation aids used are listed in Table 6. All formation aids were supplied by BASF. The formation aids were studied with the R-F machine as one part in the three-component retention aids systems (Table 5). Besides these studies, the formation aids were investigated as single additives in flocculation studies (flow loop system) and in MRI investigations studying the effect on flow properties.

20

During the MRI experiments, the formation aids were used as aqueous solutions with a concentration of 0.2% that were prepared the same day as the experiment (stirred for two hours with a bar magnet in a 1000 ml beaker). During the flow loop experiments, aqueous solutions (0.02-0.06%) of the formation aids were prepared on the same day as the experiments (stirring time: two hours). Table 6. Characteristics of the formation aids, all given by the supplier. Component Structure Charge density

(meq/g) Intrinsic viscosity (dl/g)

Gyration radius (nm)

Linear anionic polymer

-1.76A 14B n/a

Partly structured anionic polymer

-2.16A 10B >500C

Structured anionic polymer

-2.50A n/a Approx. 300C

A Measurements were made with Mütek™ Particle Charge Detector (PCD) with Polydiallyldimethylammonium chloride (0.001N) as titrating reagent. B A suspended-level viscometer was used to determine the specific viscosity of the test component at various concentrations in buffered 1M sodium chloride solution. A plot was made of reduced specific viscosity against concentration, and the intrinsic viscosity was obtained by extrapolation to infinite dilution. The longer the polymer chains, the higher the intrinsic viscosity (dl/g). The test method is referred to as ACSMOT No: 7. C The determination of the gyration radius was based on fractionation of polydisperse samples followed by elution-time-resolved mass measurements of each fraction by light scattering.

21

4 Methods

4.1 The retention and formation machine

The R-F (retention and formation) machine used in this thesis work was a pilot-scale fourdrinier former which was designed to study retention, paper formation and drainage rates on the wire part. This pilot former mimics the short circulation of full-scale paper machines and produces, without press and drying parts, a wet paper web at speeds of up to 400 m/min (grammages between 40-80 g/m2). The R-F machine is photographically and schematically illustrated in Fig 4 and Fig 5 (including surrounding equipment and dosage system), respectively. The R-F machine is owned by MoRe Research AB, Örnsköldsvik, Sweden. A general description of the R-F machine is given here, whereas details are thoroughly described in Paper 1 (Svedberg, Lindström 2010a). With a construction of static mixers, the shear forces applied in the R-F machine are similar to those in earlier full-scale paper machines, albeit not today’s state-of-the-art paper machines. The R-F machine provides a short circulation of the white-water system, and the volume of the white-water system was designed to give a short residence time to chemical equilibrium. Sampling (retention, formation, etc.) was performed five minutes after a change in machine settings, which corresponds to the residence time to chemical equilibrium in the white-water system (Svedberg, Lindström 2010a). The dosage system of the R-F machine includes eight possible chemical injections; five of the injection points are found in the stock flow. All injection points are characterized by a certain degree of shear and residence time for the additive (Fig 5). The residence times of the additives in the pulp and stock flow, given in seconds, are calculated from the addition point to the headbox when running at standard machine operating conditions (Table 7). As well as evaluations of machine settings, the R-F machine allows for studies of different retention aid systems, fillers, furnishes and dosage strategies (e.g. different degrees of shear and dwell times) and other chemical adjuvants (starch, dispersion aids, etc.), with high reproducibility. The R-F machine was shown to be useful for predicting a ranking order that corresponds to full-scale machine observations (Svedberg, Lindström 2010a). Unless stated elsewhere, the R-F machine trials were performed with the standard operating conditions given in Table 7. These standard operating conditions were found to give the best performance with respect to paper formation and minimized streak formation (Svedberg, Lindström 2010a). The produced paper contained, unless stated elsewhere, approximately 20% filler (equivalent to an addition of filler of 25% based on total solids content). Table 7. R-F machine standard operating conditions. Variables Standard operating conditions

Machine speed 260 m/min

Slice opening 16 mm

Jet-to-wire speed ratio 1.2

Stock consistency 5 g/l

Grammage 60 g/m2

Volumetric headbox flow rate 910 l/min

22

Fig 4. Photograph showing the R-F machine.

23

Fig 5. Schematic diagram of the R-F machine with its surrounding equipment and dosage system (not to scale). Details of the R-F machine have been described by Svedberg and Lindström (2010a). The residence times of the additives in the pulp and stock flow (in parenthesis), given in seconds, are calculated from the addition point to the headbox when running at standard operating conditions (Table 7).

24

4.2 Retention and formation

The retention value evaluated in the R-F machine experiments was the first pass retention with respect to filler (Rf), in percent (%), which is defined by:

100*11

2

C

CRf [3]

where C1 is the concentration of filler in the headbox and C2 is the concentration of filler in the wire pit. The mass formation was determined using a FUJI method at MoRe Research, Sweden. The FUJI method measures local variations in grammages according to a beta radiographic method (Norman, Wahren 1974; Norman 2009). The formation results are presented as formation numbers, calculated for different wavelength intervals (Johansson, Norman 1996). The main wavelength intervals correspond to the total formation number (0.4-30.0 mm), the small-scale formation number (0.4-3.0 mm) and the large-scale formation number (3.0-30.0 mm). Moreover, these formation numbers can be subdivided into narrower intervals with respect to the following wavelength intervals: 0.4-0.5 mm, 0.5-1.0 mm, 1.0-2.0 mm, 2.0-4.0 mm, 4.0-8.0 mm, 8.0-16.0 mm and 16.0-32.0 mm. It should be pointed out that the wavelength is twice the floc size and that the formation numbers are normalized to a grammage of 60 g/m2. Unless stated elsewhere, the formation results are presented as the total formation numbers (0.4-30.0 mm) in the machine direction (MD). A lower formation number means better paper formation (lower coefficient of grammage variation). The presented formation number is a mean value from measurements of two separate paper sheets. Based on reproducibility tests, the retention and formation data are shown with standard deviations of 2.0 percentage units and 0.36 units, respectively (Svedberg, Lindström 2010a). How the standard deviations were determined is described in more detail in Paper 1.

4.3 Dewatering analysis

An on-line application has been developed at MoRe Research, Sweden, for the purpose of characterizing differences in dewatering in terms of vertical displacements of the dry line4 on the wire section. This application is thoroughly described in a degree project report written by Peter Nilsson, Umeå University. The applied method was based on light scattering and used a DSLR camera to image dry line displacements. The dry line was identified as the boundary between the reflecting and non-reflecting areas. A series of image-processing steps quantified the change in dewatering (displacement of dry line) as the area of the adjoining wet surface (Fig 6). The results generated from the dewatering analysis are given as areas (10^3 pixel) with standard errors, where a high number correlates to poor dewatering. 4 The dry line is the boundary between the reflecting and non-reflecting regions of the upper surface of the fibre mat on a fourdrinier wire.

25

a)

b) Fig 6. Example images: (a) Photo image of the dry line on the R-F machine fourdrinier wire, (b) The corresponding contour of the area of the adjoining wet surface, which was used to quantify the dewatering performance.

4.4 Flocculation analysis – The flow loop system

A flow loop system, developed at KTH and Innventia AB, Stockholm, Sweden, has been used in this work for flocculation analysis. The details of the flow loop system and image analysis have been described by Yan et al. (2006), thus only a short description will be given in this thesis. The flow loop system (Fig 7) was designed to simulate realistic headbox flows (headbox contraction) and to mimic flow accelerations during dewatering (second contraction). Flocculation of fibres was studied in the system by using an optical detection method. In this work, images were taken through the extended nozzle, with the second contraction part not included. The flow loop system includes an 810-liter pulp storage tank (1) equipped with appropriate agitation, a frequency controlled centrifugal pump (2), a radial flow distributor consisting of eight output tubes (5), a transparent plastic headbox with a step diffuser pipe package (6) and an extended nozzle (7) followed by a secondary contraction (8). Moreover, the flow loop system includes a dosage system (3) with an inline static mixer (4), where an additive can be injected online. The maximum flow rate of the flow loop system was 300 l/min. After imaging, the stock exited to waste (general issue when adding chemicals) or follows the loop. The headbox nozzle contraction ratio was 16.7, and at the exit of the headbox nozzle contraction the turbulence had decayed (Yan et al. 2006). The extended nozzle was adopted as a normal slice opening on full-scale machines. Images were taken through the extended nozzle by a high-resolution camera with transmitted light pulse illumination (stroboscope). Hence, images of fibre flocculation were evaluated by power spectrum analysis, and the mean floc size was calculated (Yan, Norman

26

2006). The mean floc size (mm) is interpreted as half the value of the mean wavelength. A higher mean floc size number correlates to a higher degree of fibre suspension flocculation.

Fig 7. Schematic diagram of the flow loop system (Yan et al. 2006). See text for description of the components. (Roughly scaled).

4.5 Magnetic resonance flow imaging

The effects of formation aids on flow properties of cellulosic fibrous suspensions were investigated using phase flow encoded nuclear magnetic resonance imaging (MRI). These investigations were performed at the Department of Food Science and Technology at the University of California in Davis, US. The key element of the investigations was to measure velocity profiles and turbulence intensity distributions in fibre suspensions. Besides flow imaging, it was possible to characterize the rheology by knowing the pressure drop along the pipe. Velocity profiles and turbulence intensity profiles were obtained using a velocity encoded spin-echo pulse sequence in an Aspect Imaging 1 Tesla permanent magnet based imaging spectrometer (Aspect Imaging Shoham, Israel), with 30 G/cm peak gradient strength. The spin-echo pulse sequence has been described by Li and McCarthy (1995). The measured signal derives from the hydrogen nucleus (1H) in the water phase. Due to the short relaxation time and broad line width, the 1H contribution of cellulose fibres has a negligible effect on the signal, even at concentrations exceeding 10%. To obtain velocity profiles, magnetic resonance images were processed using the methods described by Arola et al. (1997; 1998). In order to study turbulence intensity at high mean bulk flow rates, it was necessary to implement a signal averaging technique due to signal loss (Li et al. 1994a; Li et al. 1994b). Unless stated elsewhere, the acquisition time for a typical profile was 1.05 minutes corresponding to 8 averages. Figure 8 shows a schematic diagram of the experimental set-up including the imaging assembly, the fibre flow and the means of introducing polymer. Fibres and water were loaded and agitated (4) in a mixing tank (3) (~80 dm3) connected to a Moyno pump (6) of variable speed drive. The mean bulk flow rate was determined during the flow imaging experiments by timed collection of the suspension. Subsequent polymer additions (1) were made through an injection quill (5). A gear pump (2) with variable speed drive was used to control the flow rate of the polymer addition. Downstream from the injection point, a static mixer (7) was used to disperse the polymer into the flow. After leaving the static mixer, the suspension entered a section of tubing

1.

3.

4.

5.

6. 7. 8.

2.

27

(8) whose length was chosen to provide a dwell time of 4.6 seconds for the polymer. The suspension then entered a 2.08 m long acrylic tube (9) with an internal diameter of 18.6 mm that passed through the imaging section of a magnet (10). Pressure drops (13) and temperatures (14) were measured at the ends of the acrylic tube. The radio frequency coil (11) was a solenoid with 3 turns, encasing a cylindrical volume 38 mm in diameter and 36 mm long. After imaging, the suspension exited into a waste tank (12). The pressure drop, ∆P, over a tube length, L (2.04 m), was measured to study drag reduction effects. The ∆P value ((P2-P1)/L) was determined by measuring the inlet pressure (P1) and the outlet pressure (P2) of the acrylic tube, by using a pressure measuring instrument (manometer). By knowing the pressure drop along the tube, the local shear stress (σ) can be determined using: (Arola et al. 1999)

σ(r) rL

P*

2 [4]

where r is the radial coordinate in the pipe.

Fig 8. Schematic diagram of the flow imaging system. See text for description of the components. (Not to scale).

28

4.6 Adsorption analysis (QCM and DPI)

A quartz crystal microbalance (QCM) E4 from Q-Sense AB (Västra Frölunda, Sweden) was used to study adsorption with a continuous flow of 100 µl/min (Marx 2003). All studied samples had a concentration of 100 mg/l. The polymer solutions and the rinsing solution were all pH adjusted to pH 8, and 10 mM NaCl was added to the solutions. For each sample, adsorption for five minutes was followed by rinsing for another five minutes. The resonating QCM crystals used were AT-cut quartz crystals (5 MHz resonant frequency) with an active surface of sputtered silica, which were plasma treated using an air plasma cleaner (Model PDC 002, Harrick Scientific Corporation, NY, USA) under reduced air pressure for 120 s at high effect (30 W). The quartz crystal oscillates due to the piezoelectric effect, at a constant resonance frequency. Adsorption onto the crystal is sensed as a decrease in the resonance frequency, indicating mass uptake. To convert a change in frequency into an adsorbed mass per area unit, including immobilized solvent, the Sauerbrey model was used (Sauerbrey 1959):

n

fCm * [5]

where m=adsorbed mass per area [mg/m2] C=sensitivity constant, -0.177 [mg/(m2*Hz)] Δ=change in resonance frequency of the AT-cut quartz crystal [Hz] n=overtone number The change in energy dissipation during adsorption can also be examined, yielding information about the visco-elastic properties of the adsorbed substances. To determine these properties, the electric current to the crystal is turned off and the decay of the oscillating amplitude is registered. By studying changes in dissipation we can get hints about changes in the stiffness of the adsorbed layers and how water-rich the layers are. A rigidly attached substrate is characterized by no change in dissipation during adsorption. A more loosely attached and water-rich (visco-elastic) layer is characterized by an increase in dissipation during adsorption. The dissipation factor, D, is defined by:

stored

dissipated

E

ED

2 [6]

where Estored is the energy stored in the oscillating system and Edissipated is the energy dissipated during one oscillating period. A dual polarization interferometer (DPI) Analight Bio200 from Farfield Sensors Ltd (Manchester, UK) was used to study adsorption with a continuous flow of 50 µl/min (Swann et al. 2004). The same samples were used as in the QCM experiments and Nitrogen-doped silica chips were used as deposition substrates. DPI analyses do not include associated water in the mass uptake determination, whereas the QCM analysis does.

29

5 Results and Discussion

5.1 The effect of the jet-to-wire speed ratio on paper formation

The standard machine operating conditions given in Table 7 were determined during the trimming of the R-F machine. During the trimming, a number of mechanical properties were evaluated, such as the jet-to-wire speed ratio, and the best configurations were chosen as standards. The effect of the jet-to-wire speed ratio (i.e. the speed of the jet compared to the speed of the wire) on paper formation was investigated on the R-F machine in one series of trials. These trials were performed without chemical additions (filler, retention aids, etc.), with standard furnish characteristics (Table 2) and with standard operating conditions according to Table 7 (except the slice opening). The jet-to-wire speed ratio was varied by varying the slice opening. In this way, a constant machine speed, volumetric headbox flow rate (910 l/min), headbox consistency (0.5%) and hence grammage were maintained. The flow rate was controlled by the frequency steering electric motor connected to the mixing pump. The jet speed was not measured in this work. By maintaining the machine speed and flow rate at constant levels, it was possible to change the jet-to-wire speed ratio by knowing its proportion to the slice opening. The jet-to-wire speed ratio was varied in four steps (1.0, 1.1, 1.2 and 1.3) which corresponded to slice openings between 19.2 mm and 14.8 mm. A reduced slice opening correlates to a higher jet speed, and hence to a higher jet-to-wire speed ratio. Studying ratios lower than 1.0 was not practically possible at the standard R-F machine speed due to the low discharge rate. The strong impact of the jet-to-wire speed ratio on paper formation is shown in Fig 9, showing the total formation number in the machine direction as a function of the jet-to-wire speed ratio. A higher jet-to-wire speed ratio results in significantly better formation. Poorest formation was observed for the lowest jet-to-wire speed ratio investigated (1.0), which included no speed difference between the jet and the wire. With jet-to-wire speed ratios of 1.1, 1.2 and 1.3, a small speed difference arises by decelerating the jet from a higher velocity to the speed of the wire and this results in a certain shear of the flocs, which is favourable for paper formation. At zero speed difference, the paper formation was poorer because no extra shear forces were applied in the forming zone and the flocs froze in the formed paper web. The paper formation was improved with increased jet-to-wire speed ratios up to a certain level of speed difference. When the speed difference was too large (jet-to-wire speed ratios >1.4), the shear of the flocs was so intense that the fibre network broke up, resulting in a pronounced deterioration of the formation. Since the jet-to-wire speed ratio was varied by varying the slice opening (19.2 mm to 14.8 mm), the headbox contraction ratio was also varied. The headbox contraction ratio was varied between 7.3 to 9.5 (calculated), with the lowest contraction ratio referring to the highest slice opening (lowest jet-to-wire speed ratio). Interpreting the results in Fig 9 in terms of the contraction ratio, the results indicate that a higher contraction ratio should be set in order to achieve good formation. Contrary to this interpretation, Nordström (2003) reported that for chemical pulps

30

without any additives, including 100% hardwood, the headbox nozzle contraction had no significant effect on the resulting paper formation. The standard operating jet-to-wire speed ratio was determined to 1.2, based on both the paper formation and the streakiness in the produced sheet. A tendency of increased streakiness was observed with jet-to-wire speed ratios higher than 1.2.

10

14

18

22

26

0,9 1 1,1 1,2 1,3 1,4

Jet-to-wire speed ratio

To

tal

form

ati

on

nu

mb

er

(MD

)

Fig 9. The total formation number in the machine direction as a function of the jet-to-wire speed ratio. The jet-to-wire speed ratio was varied by changing the slice opening. The study was performed on the R-F machine with standard pulp characteristics (Table 2) and standard operating conditions (Table 7) and without chemical additions.

5.2 The effect of different retention aids on retention and formation

Several different retention aids have been investigated with the R-F machine with respect to the retention-formation relationship. These retention aids include both single-component (Paper 2) and dual-component microparticulate retention aid systems (Paper 2) and, finally, three-component retention aid systems (Paper 3). The influence of the different retention aid combinations on retention and formation are shown in the following three chapters. All R-F machine investigations accounted for here were performed with standard furnish characteristics (Table 2) and standard operating conditions (Table 7).

5.2.1 Single-component polyacrylamide retention aids

Three linear cationic polyacrylamides (C-PAM A-C) with different molecular weights were investigated on the R-F machine, as single-component retention aids (Svedberg, Lindström 2010b). Characteristics of the cationic polyacrylamides are given in Table 3. In order to reduce the shear stress on the polymers, the polyacrylamides were added at a position after the static mixers,

31

at the last dosage point (no. 8), which corresponds to a residence time of 2.0 s. Different retention levels were obtained by varying the polyacrylamide addition level between 500 g/t to 1500 g/t. The filler (GCC) was added in the R-F machine pulp flow, at dosage point no. 3 (residence time 8.8 s) with an addition of 25% based on the total stock consistency. Figure 10 shows the filler retention (%) as a function of the added amount of cationic polyacrylamide (g/t). C-PAM (A) corresponds to the polyacrylamide with the lowest molecular weight, which gave, at a certain dosage level, a lower level of filler retention in comparison with the other two polyacrylamides. This is explained in terms of the extension of the protruding polymer bridges, which is less for a polymer of lower molecular weight (Wågberg et al. 1988; Solberg, Wågberg 2003). The retention level was changed from approximately 30% to 80% with the present polyacrylamide addition intervals (500 g/t to 1500 g/t). With respect to the retention performance, the same polyacrylamides were also investigated in Britt dynamic drainage jar (BDDJ) experiments. These results are presented in Paper 2 (Svedberg, Lindström 2010b).

0

20

40

60

80

100

200 600 1000 1400 1800

Added amount of cationic polyacrylamide (C-PAM), g/t

Fil

ler

rete

nti

on

, %

C-PAM (A)

C-PAM (B)

C-PAM (C)

Fig 10. The filler retention (%) as a function of the added amount of cationic polyacrylamide (g/t), for three cationic polyacrylamides of different molecular weight (C-PAM A-C). Symbols: ● C-PAM (A) 2.9-4.4 million Daltons; ○ C-PAM (B) 6.0-7.9 million Daltons; ▲ C-PAM (C) >10.8 million Daltons (according to the supplier). The study was performed on the R-F machine for a fine paper stock with addition of filler (25% GCC based on the total solids content). The total formation number in the machine direction as a function of the retention level (%) is shown in Fig 11 for C-PAM (A-C). The same linear relationship was shown for all three cationic polyacrylamides. Independent of the molecular weight of the C-PAM, an increased addition level, and hence an increased retention level, resulted in increased formation numbers, which means poorer formation. These results hence indicate that the flocculation and retention mechanism is the same for these types of polymers.

32

10

12

14

16

18

20

0 20 40 60 80 100

Filler retention, %

To

tal

form

ati

on

nu

mb

er

(MD

)

C-PAM (A) C-PAM (B) C-PAM (C) C-PAM (C) repeated trial

Fig 11. The total formation number in the machine direction as a function of the filler retention (%), for three cationic polyacrylamides of different molecular weights (C-PAM A-C). Symbols: ● C-PAM (A) 2.9-4.4 million Daltons; ○ C-PAM (B) 6.0-7.9 million Daltons; ▲ C-PAM (C) >10.8 million Daltons (according to the supplier). The study was performed on the R-F machine for a fine paper stock with addition of filler (25% GCC based on the total solids content). The cationic polyacrylamides (C-PAM A-C) have also been evaluated with respect to other formation wavelength intervals (Svedberg, Lindström 2010b). No differences between the polyacrylamides were observed in those evaluations, neither in the small-scale (0.4-3.0 mm) and large-scale (3.0-30.0 mm) formation numbers nor in any other wavelength interval. For all three polyacrylamides, the large-scale formation number was the principal contribution to the increase in the total formation number. All formation data are given in Paper 2 (Svedberg, Lindström 2010b).

5.2.2 Polyacrylamide-based microparticulate systems

In this chapter, two polyacrylamide-based microparticulate retention aid systems based on different microparticles (montmorillonite clay and silica sol) are compared. These two systems were investigated with the R-F machine with respect to retention and formation (Svedberg, Lindström 2010b). The microparticles were investigated in conjunction with a cationic polyacrylamide (denoted here as C-PAM B). Characteristics of the polyacrylamide-based microparticulate systems are given in Table 4. The cationic polyacrylamide (C-PAM B) was added before the microparticles, at dosage point no. 5, which corresponds to a residence time of 5.6 s (position before the mixers). The microparticles were added at the last dosage point (no. 8), which corresponds to a residence time of 2.0 s (position after the mixers). The addition of C-PAM was changed in order to vary the retention level, whereas the additions of microparticles were constant (2 kg/t of montmorillonite

33

clay and 3 kg/t of silica sol). As in the previous chapter (5.2.1), the filler (GCC) was added in the R-F machine pulp flow, at dosage point no. 3 (residence time 8.8 s) with an addition of 25% based on the total stock consistency. The two polyacrylamide-based microparticulate retention aid systems were compared with respect to the retention performance (Fig 12) and the retention-formation relationship (Fig 13). The filler retention (%) vs. the added amount of cationic polyacrylamide (g/t) is shown in Fig 12. It is shown in this figure that the two microparticulate systems diverge in their retention performance. The system including montmorillonite clay gave a higher retention level, at the same C-PAM addition, in comparison with the silica-based system. The retention-formation relationships are demonstrated in Fig 13, showing the total formation number in the machine direction as a function of the filler retention (%). Both systems, based on different microparticles, showed the same relationship between retention and formation. Both systems suggested a linear relationship, where an increased filler retention level corresponds to a deteriorated formation. Both microparticulate systems were investigated twice, indicating a good reproducibility in the experiments (Fig 13). Figure 14 combines the previous results (single-component polyacrylamide retention aids) with the present study of microparticulate retention aid systems. Also, when comparing all five systems together, no difference in the retention-formation relationship was obtained (Fig 14). For all systems, a high retention level was accompanied by an impaired formation, indicating an overall difficulty in changing the interdependency between retention and formation.

0

20

40

60

80

0 300 600 900 1200 1500 1800

Added amount of cationic polyacrylamide (C-PAM), g/t

Fil

ler

rete

nti

on

, %

C-PAM (B)+Silica sol

C-PAM (B)+Montmorillonite clay

Fig 12. The filler retention (%) as a function of the added amount of cationic polyacrylamide (g/t), for two polyacrylamide-based microparticulate systems with different microparticles (montmorillonite clay and silica sol). Symbols: □ C-PAM (B)+montmorillonite clay; ■ C-PAM (B)+silica sol. The study was performed with the R-F machine for a fine paper stock with addition of filler (25% GCC based on the total solids content).

34

10

12

14

16

18

20

0 20 40 60 80 100

Filler retention, %

To

tal

form

ati

on

nu

mb

er

(MD

)

C-PAM (B)+Silica sol

C-PAM (B)+Silica sol repeated trial

C-PAM (B)+Montmorillonite clay

C-PAM (B)+Montmorillonite clay repeated trial

Fig 13. The total formation number in the machine direction as a function of the filler retention (%), for two polyacrylamide-based microparticulate systems with different microparticles (montmorillonite clay and silica sol). Symbols: □◊ C-PAM (B)+montmorillonite clay; ■♦ C-PAM (B)+silica sol. The study was performed with the R-F machine for a fine paper stock with addition of filler (25% GCC based on the total solids content).

35

10

12

14

16

18

20

0 20 40 60 80 100

Filler retention, %

To

tal

form

ati

on

nu

mb

er

(MD

)

C-PAM (B)+Silica sol

C-PAM (B)+Montmorillonite clay

C-PAM (A)

C-PAM (B)

C-PAM (C)

Fig 14. The total formation number in the machine direction as a function of the filler retention (%), for five different retention aids. Three cationic polyacrylamides of different molecular weights (C-PAM A-C) and two polyacrylamide-based microparticulate systems with different microparticles (montmorillonite clay or silica sol in conjunction with C-PAM (B)) were investigated. Symbols: ■ C-PAM (B)+silica sol; □ C-PAM (B)+montmorillonite clay; ● C-PAM (A) 2.9-4.4 million Daltons; ○ C-PAM (B) 6.0-7.9 million Daltons; ▲ C-PAM (C) >10.8 million Daltons. The study was performed with the R-F machine for a fine paper stock with addition of filler (25% GCC based on the total solids content).

5.2.3 Three-component retention aid systems

Finally, three-component retention aid systems were investigated with the R-F machine with respect to retention, formation and drainage (Svedberg, Lindström 2012). The three-component systems were based on cationic polyacrylamide (C-PAM), high molecular weight anionic polymer and montmorillonite clay, added in sequence according to the earlier described procedure. That is to say, a microparticulate retention aid system was added in combination with a dispersion aid (Class II formation aids). As formation aids, three high molecular weight polymers of different structure (degree of cross-linking) were investigated, denoted as: linear, partly structured and structured polymers. Characteristics of the three-component retention aid systems are given in Table 5. The high molecular weight polymers were added from a small amount up to an addition level where papermaking fibres in suspension become dispersed (200 g/t-1200 g/t). The cationic polyacrylamide was added at dosage point no. 5 (before the mixers), the anionic polymer at dosage point no. 7 (between the mixers) and the montmorillonite clay at dosage point no. 8 (after the mixers). The additions of cationic polyacrylamide and montmorillonite clay were constant, 400 g/t and 2.0 kg/t, respectively. PCC was added as filler in these experiments, added in the R-F machine pulp flow, at dosage point no. 3 (residence time 8.8 s) with an addition of 25% based on the total stock consistency.

36

The combination of the overall results illustrated in Fig 14 with the performance of a three-component retention aid system based on the partly structured polymer is shown in Fig 15. The results in Fig 15 demonstrate a way to change the retention-formation relationship, that is, by improving the formation at a given level of retention, or at even higher retention levels. Figure 16 shows the total formation number in the machine direction as a function of the added amount of the high molecular weight anionic polymer (g/t) for all three polymers of different structure. The results in Fig 16 demonstrate different trends depending on the polymer structure. The formation was significantly improved when the linear or the partly structured polymer was used, and as the added amount was increased. The lowest formation number, which means best formation, was attained at the highest polymer addition level. However, the formation was not influenced by an addition of the structured polymer, at any dosage level. The improvement in formation by using the linear polymer in excess is illustrated in Fig 17, showing two comparative radiograph images which correspond to the zero addition and the highest addition of polymer (1200 g/t).

8

10

12

14

16

18

20

0 20 40 60 80 100

Filler retention, %

To

tal

form

ati

on

nu

mb

er

(MD

)

C-PAM (B)+Silicasol

C-PAM(B)+Montmorilloniteclay

C-PAM (A)

C-PAM (B)

C-PAM (C)

C-PAM (B)+Partlystructuredpolymer+Montmorillonite clay

Fig 15. The total formation number in the machine direction as a function of the filler retention (%), for six different retention aids. Symbols: ■ C-PAM (B)+silica sol; □ C-PAM (B)+montmorillonite clay; ● C-PAM (A) 2.9-4.4 million Daltons; ○ C-PAM (B) 6.0-7.9 million Daltons; ▲ C-PAM (C) >10.8 million Daltons; Blue circle: C-PAM (B)+partly structured polymer+montmorillonite clay. The study was performed with the R-F machine for a fine paper stock with addition of filler (25% based on the total solids content).

37

6

9

12

15

18

0 200 400 600 800 1000 1200 1400

Added amount of anionic polymer, g/t

To

tal

form

ati

on

nu

mb

er

(MD

)

Structured polymer

Partly structured polymer

Partly structured polymer repeated trial

Linear polymer

Fig 16. The total formation number in the machine direction as a function of the added of anionic polymer (g/t). Data are shown for three anionic polymers of different structure (linear, partly structured and structured), which were investigated in conjunction with C-PAM and montmorillonite clay. The study was performed with the R-F machine for a fine paper stock with addition of filler (25% PCC based on the total solids content). The partly structured polymer was also investigated in a repeated trial.

Fig 17. Beta radiogram images (real size 10*10 cm). Left: 0 g/t of anionic polymer (total formation number was 13.48 in MD). Right: 1200 g/t addition of the linear polymer (total formation number was 10.78 in MD).

38

The added amount of anionic polymer had a slight or insignificant effect on the retention level, which remained around 50% irrespective of the addition level. These retention results, together with the trends reported in Fig 16, give rise to the relationships shown in Fig 18. In Fig 18, the retention-formation relationships for all three-component retention aid systems are shown and compared to the corresponding microparticulate system (C-PAM (B)+mont-morillonite clay). As discussed earlier, an addition of the microparticulate retention aid system resulted a linear relationship between retention and formation, where the formation deteriorated at increased retention levels. The results in Fig 18 demonstrate that the interdependency between retention and formation can be broken. A key for changing the retention-formation relationship is, hence, to add a high molecular weight polymer in excess, in conjunction with a microparticulate system. These pioneering results were valid for the systems including the linear and the partly structured anionic polymer, that is, polymers that have high extension in solution and that have the ability to elongate. When the structured polymer was used, no effects were observed. The trend lines in Fig 18 indicate shifts in addition levels, from low to high addition levels. The higher the added amount of the linear and the partly structured polymers, the better was the formation. The trends shown in Fig 16 and Fig 18 were repeated in a separate series of trials, not shown in detail in this summary. However, the high degree of reproducibility is also indicated in Fig 16 by comparing the two trials performed with the partly structured polymer.

1200 g/t

0 g/t

6

9

12

15

18

0 20 40 60 80 100Filler retention, %

To

tal

form

ati

on

nu

mb

er

(MD

)

Structured polymer Partly structured polymer Reference system Linear polymer

Fig 18. The total formation number in the machine direction as a function of the filler retention (%). Data are shown for a dual reference microparticle system (C-PAM (400 g/t) and montmorillonite clay (2 kg/t)) and three-component systems (reference system plus an anionic polymer of varied structure). The anionic polymer was changed in three structures, defined as linear, partly structured and structured. The study was performed with the R-F machine for a fine paper stock with addition of filler (25% based on the total solids content).

39

The addition of high molecular weight anionic polymers (dispersion aids) at high addition levels had an effect not only on the retention-formation relationship, but also on the dewatering performance on the wire (Fig 19). Fig 19 shows the dewatering, in terms of dewatering area (see figure 6 for definition) (10^3 pixel) versus the added amount of anionic polymer (g/t), for all three polymers of different structure. It can be seen in Fig 19 that the dewatering number was significantly increased as the added amount of the linear and the partly structured polymer was increased, which means reduced drainage rates. No effect on dewatering was observed when the structured polymer was used, which is in line with the trends reported in Fig 16. Hence, these investigations, combining a formation aid with a microparticulate retention aid system indicate an interdependency between paper formation and drainage rate of the fibre suspension.

0

300

600

900

1200

0 200 400 600 800 1000 1200 1400

Added amount of anionic polymer, g/t

Dew

ate

rin

g a

rea,

10^

3

pix

el

Structured polymer Partly structured polymer Linear polymer

Fig 19. The dewatering in terms of dewatering area (10^3 pixel) (defined in figure 6) as a function of the added amount of anionic polymer (g/t). Data are shown for three anionic polymers of different structure (linear, partly structured and structured), which were investigated in conjunction with C-PAM and montmorillonite clay. The study was performed with the R-F machine for a fine paper stock with addition of filler (25% PCC based on the total solids content). Since the dewatering performance was also affected by high additions of anionic polymer, it was questioned whether the formation improvements were caused by changed chemistry or by the effect of changed dewatering. An answer to this question can be found in the results in Fig 20. The underlying trials to the results presented in Fig 20 were designed to vary the dry line position on the wire, both mechanically and chemically, from a reference position. The reference position was obtained for a dual reference system (C-PAM (B) (400 g/t) and montmorillonite clay (2 kg/t)) and with standard operating conditions. The dry line position was changed to the same upper register, both mechanically by reducing the number of foils and vacuum, and chemically by adding anionic polymer in excess (1200 g/t). The partly structured anionic polymer was used in these trials, added together with C-PAM (B) (400 g/t) and montmorillonite clay (2 kg/t). The C-PAM was added at dosage point no. 5, the anionic polymer at dosage point no. 7 and the

40

montmorillonite clay at dosage point no. 8. The dry line was also moved down by increasing the vacuum. The dewatering in terms of area (10^3 pixel) and the total formation number in the machine direction are shown as functions of the dry line position, in Fig 20. A higher dewatering number corresponds to a higher position of the dry line (reduced drainage rate). From Fig 20 it can be concluded that the formation improvements were caused by a chemical mechanism which was introduced when adding high amounts of anionic polymer, and not by the effect of changed dewatering. The formation was not affected when the dry line position was mechanically changed up and down in relation to the reference position.

0

300

600

900

1200

1500

Dry line moved

down by vacuum

Reference Dry line moved up

by anionic polymer

Dry line moved up

by foils and

vacuum

Dew

ate

rin

g a

rea,

10^

3 p

ixel

6

9

12

15

18

To

tal

form

ati

on

nu

mb

er

(MD

)

Fig 20. The dewatering in terms of area (10^3 pixel) and the total formation number in the machine direction as functions of the dry line position. The dry line was moved from a reference position in three ways: down by increased vacuum, up by overdosage of anionic polymer, and up by reduced vacuum and number of foils. The study was performed with the R-F machine for a fine paper stock with addition of filler (25% PCC based on the total solids content).

41

5.3 The effect of formation aids on fibre suspension flocculation

The effect of formation aids on fibre suspension flocculation under simulated forming conditions was investigated with the aid of a flow loop system. Three high molecular weight anionic polymers of different structure were investigated as single-component additives, defined as linear, partly structured and structured polymers. Characteristics of the formation aids used are given in Table 6. The objective of this study was a correlation between the deflocculation of a furnish as determined with the flow loop system and the change in formation as determined with the R-F machine for the three-component retention aid systems. One batch of fibre suspension (400 l) was used for each trial point. Details of the fibre suspensions are given in chapter 3.1. The standard fibre concentration in the suspension was 5 g/l. The flow rate was 252 l/min in the trials and the dwell time of the anionic polymers was 3.4 s (corresponds to the time from polymer injection to imaging through the extended nozzle). First, a reference curve was established with respect to the fibre concentration, with no polymers added. The fibre concentration was varied from almost 6 g/l down to 1 g/l. The mean floc size in the machine direction (mm) versus the fibre concentration (g/l) is shown in Fig 21. Second, the three polymers were added in separate trials. All the three polymers were investigated at a dosage of 2.5 mg/g. The partly structured polymer was also investigated at a lower dosage amount (0.8 mg/g) in order to study the effect of the dosage amount. Each trial point that included a polymer occurred after the corresponding reference trial point (only the fibre suspension). Hence, one and the same batch of fibre suspension was used for the polymer trial point and the corresponding reference trial point. The effect of high molecular weight anionic polymers as formation aids on fibre suspension flocculation is shown in Fig 21. By comparing the mean floc size number for a certain polymer trial point with the corresponding reference trial point, the dispersion effect of the respective polymer can be deduced. All data in Fig 21 are shown with standard errors of each trial point. Figure 21 reveals that the mean floc size decreased significantly when the linear and the partly structured polymer were added to the fibre suspension, at a dosage level of 2.5 mg/g. At the lower addition of the partly structured polymer (0.8 mg/g), no dispersion effect was observed. Hence, the results indicate that the addition of polymer needs to be high in order to gain a dispersing effect. If the standard deviations for all trial points are taken into consideration, the effect of the structured polymer is viewed as low or even negligible. The results indicate that the structure of the polymer is highly important for the dispersion effect. These results are consistent with those found for the R-F machine trials when evaluating the same three polymers in conjunction with a microparticulate retention aid system (see chapter 5.2.3). The flow loop experiments indicate that fibre dispersion is possible by adding formation aids to a fibrous suspension. The highest obtained reduction of fibre suspension flocculation was observed for the linear polymer at a dosage level of 2.5 mg/g. A translation of this maximum reduction by additives into a reduction in fibre consistency (synonymous with a consistency of approximately 2.2 g/l) gives a hint of the deflocculation capability of formation aids. In one trial, the linear polymer was added directly to the stock (one injection of 2.5 mg/g). The stock was then recirculated for ten minutes and images were recorded during that time. The objective was to study the influence of shear/residence time on fibre dispersion by formation aids. It was shown that a direct addition of polymer to the stock had no effect in terms of

42

reduced mean floc size. The shear introduced to the polymer by pumping the fibre suspension probably results in polymer deformations (see e.g. Tanaka et al. 1993), and the dispersion effect fails. These results indicate that the dispersion effect of formation aids cannot be explained by electrostatic mechanisms only. The extension of the polymer in solution seems to be highly relevant for the deflocculation mechanism. In another series of experiments in the flow loop system, several linear anionic polymers of different molecular weights and different charge densities were investigated with respect to fibre suspension flocculation. The objective was to find an optimal anionic polymer for fibre dispersion. These experiments are not described in detail in the thesis. However, a manuscript is planned that will present the overall results. Briefly, it was shown that a high molecular weight of the polymer was preferable considering the reduction of fibre suspension flocculation. The charge density of the polymer, on the other hand, had little effect on the flocculation performance. The polyelectrolyte effect on fibre dispersion was also investigated by varying the suspension conductivity from 800 µS/cm to 4000 µS/cm by adding salt (CaCl2). It was shown in these studies, using a linear anionic polymer of high molecular weight, that high expansion of the anionic polymer structure was a prerequisite for fibre dispersion. In the presence of high ionic strength, the polymer structure became more coiled due to electrostatic interactions, and the dispersion effect disappeared.

1

1,5

2

2,5

3

0 1 2 3 4 5 6

Fiber concentration in suspension, g/l

Me

an

flo

c s

ize

in

MD

, m

m

Reference curve

Partly structured polymer 0 mg/g (ref. to 2.5 mg/g)

Partly structured polymer 2.5 mg/g

Partly structured polymer 0 mg/g (ref. to 0.8 mg/g)

Partly structured polymer 0.8 mg/g

Structured polymer 0 mg/g (ref. to 2.5 mg/g)

Structured polymer 2.5 mg/g

Linear polymer 0 mg/g (ref. to 2.5 mg/g)

Linear polymer 2,5 mg/g

Fig 21. The mean floc size in the machine direction (mm) as a function of the fibre concentration in suspension (g/l). Data are shown for a reference curve (no additives) and for dosages of high molecular weight anionic polymers to the fibre suspension (5 g/l). The structure of the anionic polymers was varied in three levels, defined as linear, partly structured and structured.

43

5.4 The effects of formation aids on flow properties of cellulosic fibrous suspensions

In chapter 5.2.3 and chapter 5.3 it was shown that formation aids can be used to improve the retention-formation relationship and reduce fibre suspension flocculation, respectively. High molecular weight anionic polymers were used as formation aids (Class II) in these studies. It was hypothesized that these high molecular weight anionic polymers were basically presented free in solution and hence, affected the rheological properties of the suspending medium by suppressing the intensity of turbulence. The present investigation aimed to study the effect of formation aids on flow properties of cellulosic fibrous suspensions and to examine whether or not turbulence damping is a mechanism correlated to fibre dispersion and, hence, to improved paper formation. The effects were studied using phase flow encoded magnetic resonance imaging (MRI). The intensity of the MRI signal is influenced by the presence of velocity fluctuations (Li, McCarthy 1995). Velocity profiles and turbulence intensity distributions derived from the flow imaging measurements were evaluated together with pressure drop analysis (∆P). Experiments were performed for water and fibre suspensions (0.5%), with and without formation aids added, at different mean bulk flow rates. The same three high molecular weight anionic polymers as used in the previous studies were investigated in the MRI experiments as single-component additives (characteristics are given in Table 6). The dwell time of the anionic polymers in the system was 4.6 seconds, and the dosage order was varied up to 1200 g/t. The overall results in this study demonstrated that high molecular weight polymers act as drag reducing additives in turbulent flows. The results show how both fibres and formation aids dampen the turbulence, in terms of reducing turbulence fluctuations (as determined by signal intensities) and reduced pressure drops. Good synergy effects were observed when fibres and polymers were used together. Regarding the polymer structure, the same trends were observed as in the previous R-F machine trials and flow loop experiments. Turbulence damping effects were observed for the linear polymer and the partly structured polymer, but not for the structured polymer. It was revealed that turbulence damping by additives correlates positively to improved formation. Hence, drag reduction by additives can be described as a mechanism for changing retention-formation relationships. Figure 22 shows velocity images (upper row) and signal intensity profiles (lower row) at a velocity of 1.2 m/s, for water, fibrous suspension and fibrous suspension with polymer. The polymer used in Fig 22 was the linear polymer, added at an amount of 1000 g/t (equivalent to an addition of polymer of 5 ppm based on the aqueous phase). In the velocity images, the signal intensity is a function of the radial position (mm) in the tube (horizontal coordinate) and velocity in arbitrary units (vertical coordinate). The direction of the flow in the velocity images is from bottom to top (the pipe was horizontal in the physical apparatus). In the lower row, the horizontal axis represents the radial coordinate (mm) in the tube and the vertical axis represents I(r); the signal intensity (maximum) at each radial coordinate. Imax is the maximum value of I(r) in a given profile, usually occurring in the middle of the profiles. A comparison between two signal intensity profiles measured at the same flow rates provides an indication of the differences in velocity fluctuations between the two profiles.

44

The results in Fig 22 indicate that the signal intensity increases if fibres are added, and more so if fibres and polymer are present, as shown both in the velocity profiles and the corresponding signal intensity profiles. The addition of polymer to the fibre suspension increased the value of Imax approximately three times and generated a less blunt velocity profile. The increase in signal intensity is explained by damping of the turbulent fluctuations. The trends in Fig 22 were also observed for mean bulk flow rates corresponding to 1.0 m/s and 1.4 m/s. High turbulence intensity correlates to a high pressure drop along the tube. Hence, effects of drag reduction are shown in pressure drop analyses as reduced pressure drops, due to reduced flow resistance. The corresponding pressure drop analyses were consistent with the flow imaging results shown in Fig 22. The pressure drop was reduced if fibres were added, and even more so if fibres and polymer were present.

Water Fibrous suspension Fibrous suspension with polymer

Fig 22. Velocity images (upper row) and signal intensity profiles (lower row) for water, fibrous suspension and fibrous suspension with polymer added, at a mean bulk flow rate of 1.2 m/s. The fibre concentration was 0.5% and the polymer was linear, added at 1000 g/t. The benefit of the linear polymer as a drag reducing additive was also observed in water. Table 8 summarizes the maximum signal intensities, and pressure drop values for all experiments ran at the flow rate 1.4 m/s. Good synergy effects were observed when fibres and polymers were used together, as shown in Table 8. This synergy effect was greater than the sum of the two independent effects of fibre and polymer in water. The results in Table 8 also indicate that the polymer addition is mainly responsible for the loss in pressure drop. Table 8. Maximum signal intensities (Imax) and pressure drop values ∆P for different experiments, all

performed at a mean bulk flow rate of 1.4 m/s. Sample Imax (Arb. units) ∆P (Pa/m)

Water 4.4 10.6

Water+Linear polymer (1000 g/t) 11.2 5.4

Fibrous suspension 10.1 9.2

Fibrous suspension+Linear polymer (1000 g/t) 35.4 5.9

45

A comparison of the results in Table 8 shows more pronounced effects for the Imax values, compared to the ∆P values. This could be explained by a limitation in the pressure measurement

technique. Even though the present effect of drag reduction was higher, the pressure measurements could

not serve this information. A ∆P value of about 5 seemed to be the limit for these measurements. The

evaluation window was larger in the imaging measurements and hence larger drag reduction effects could be seen in these results. Figure 23 shows the maximum signal intensity (Imax) as a function of the added amount of each polymer in fibrous suspensions at 0.5% consistency. The results in Fig 23 demonstrate that the rheological properties of the suspending medium are influenced by the structure of the added polymer. The corresponding pressure drop analysis was consistent with the results given in Fig 23. No effect on the flow behaviour was observed for the structured polymer, presumably due to the reduced capability to stretch and align in shear flows. Neither the signal intensity nor the pressure drop was influenced by an addition of the structured polymer at any dosage level studied. However, turbulence damping effects in terms of increased signal intensity and reduced pressure drop were observed for the linear polymer and the partly structured polymer. These polymers possess high extension in solution and can elongate in shear flows. An increased addition level of the two drag reducing polymers indicated an increased turbulence damping effect (valid until a state of equilibrium was reached).

0

20

40

60

80

0 300 600 900 1200 1500

Added amount of polymer in fibrous suspension, g/t

Ma

xim

um

sig

na

l in

ten

sit

y,

Arb

.un

its

Structured Partly structured Linear

Fig 23. The maximum signal intensity (Imax) as a function of the added amount of polymer (g/t) in fibrous suspension (0.5% consistency). Data are shown for three high molecular weight anionic polymers of different structure (linear, partly structured and structured). The correlation between turbulence damping and formation improvement by formation aids is illustrated in Fig 24. Figure 24 shows the total formation number in the machine direction and the pressure drop as functions of the added amount of polymer in fibrous suspension, for the linear and the structured polymer. The formation data were obtained in R-F machine trials and as can be seen the data for the partly structured polymer followed the same trends as for the linear

46

polymer regarding both the formation number and the pressure drop. The flow imaging results (maximum signal intensity, etc.) were also consistent with the pressure drop results. The results in Fig 24 show that the pressure drop measurements support the findings of the R-F machine experiments. The linear and partly structured polymer can dampen the turbulence and hence improve paper formation, whereas the structured polymer cannot.

5

8

11

14

17

0 200 400 600 800 1000 1200 1400 1600

Added amount of polymer in fibrous suspension, g/t

To

tal

form

ati

on

nu

mb

er

(MD

)

3

6

9

12

15

Pre

ssu

re d

rop

, P

a/m

Linear polymer (Formation) Structured polymer (Formation)

Linear polymer (Pressure drop) Structured polymer (Pressure drop)

Fig 24. The total formation number in the machine direction and the pressure drop (Pa/m) as a function of the added amount of polymer (g/t) in a fibrous suspension at 0.5% consistency. Data are shown for a linear and a structured high molecular weight polymer. Filled symbols refer to formation number and unfilled symbols to pressure drop. The study on formation was performed with the R-F-machine for a fine paper stock with addition of 25% filler (PCC) (based on solids content), with the anionic polymers added in conjunction with a microparticulate retention aid system (Svedberg, Lindström 2012).

5.5 Adsorption studies – The interactions of cationic polyacrylamide, high molecular weight anionic polymer and anionic montmorillonite clay on model surfaces

The mechanism of improved retention-formation relationship has previously been discussed in terms of drag reduction, that is, turbulence damping in turbulent flows by additives, in this case by high molecular weight anionic polymers. It has been assumed that the adsorption of these anionic polymers to the fibre surfaces was negligible and that all polymers were present free in solution. In order to clarify if adsorbed layers could influence the hydrodynamic properties of the fibres, adsorption studies have been performed using the DPI and QCM techniques. The objective of these adsorption studies was to investigate whether the mechanism is explained only by a solution effect or also by an effect of adsorption of the anionic polymers to the fibre surface. If there was a significant adsorption of the anionic polymer to the fibre surface there could be an interaction between the high molecular weight anionic polymer and the montmorillonite clay that

47

in turn could contribute to the improved retention-formation relationship, by, for example, reducing the fibre-fibre friction. These experiments were planned as a comparison between the two three-component retention aid systems, including the linear and the structured anionic polymer. The linear anionic polymer has shown effects throughout this thesis work, which the structured anionic polymer, on the other hand, has not. If these two three-component retention aid systems indicate the same results regarding adsorption, it would be possible to conclude that the mechanism is linked only to solution effects and not to surface effects. The three-component retention aid systems were also compared with a polyacrylamide-based microparticulate retention aid based on the same cationic polyacrylamide and montmorillonite clay. All characteristics of the components used are summerized in Table 5. The results from the adsorption study are shown in Fig 25. The left column shows the DPI results (no associated water) in terms of mass (g/m2) and thickness (nm). The right column shows the QCM results (with associated water) in terms of mass (g/m2) and dissipation. The first row in Fig 25 corresponds to the dual system, the second row to the three-component system with the linear anionic polymer and the third row to the three-component system with the structured polymer. For the dual system, both the C-PAM and the montmorillonite clay are adsorbed to the model surface, indicated by the mass uptakes (g/m2). The thickness of the layer was reduced by the adsorption of the montmorillonite clay, indicating that the layer was compacted by this second adsorbent. The dissipation normalized with frequency indicates a rigid attachment of the C-PAM, whereas the same number for the montmorillonite clay indicates a change in the visco-elastic properties, that is, a more water-rich layer. The results for both three-component systems with the DPI and QCM show the same trends for both systems regarding the mass uptake for all three constituents. Only a slight increase in mass was observed when the anionic polymer was added, whereas the subsequent adsorption of montmorillonite clay was more significant. As for the dual component system, the thickness of the layers was reduced when the montmorillonite clay was adsorbed, indicating denser layer structures after this addition. If we, also for the three-component systems, consider the dissipation normalized with frequency, we obtain different appearances of the two systems. The dissipation normalized with frequency was higher after montmorillonite clay adsorption for the system based on the linear polymer. This indicates a higher degree of swelling, that is, more associated water in a gel-like layer structure. For the system including the structured polymer, the results indicate a more rigid adsorption of the montmorillonite clay. Since the same trends for both three-component retention aid systems was obtained, these results support the suggested mechanism that the dispersion effect of the added anionic polyelectrolyte is linked to solution effects (turbulence damping) and not to surface effects. There is a small difference of the properties of the adsorbed layers between the linear and the structured polymers but this is not considered to be large enough to explain the detected dispersion effects. These results are planned to be a part of a future manuscript together with adsorption studies on fibre surfaces (polyelectrolyte titration). These adsorption results are planned to be a part in a future manuscript together with similar adsorption studies with fibres.

48

Left column: DPI results Right column: QCM results

0

2

4

6

8

10

C-PAM Montmorillonite clay

Added component

Th

ickn

ess

(n

m)

/ M

ass

(g

/m2)

Thickness, nm Mass, g/m2

0

5

10

15

20

25

30

C-PAM Montmorillonite clay

Added component

Mass (

g/m

2)

/ D

issip

ati

on

Dissipation Mass, g/m2

DPI: Dual system QCM: Dual system

0

2

4

6

8

10

C-PAM Linear anionic polymer Montmorillonite clay

Added component

Th

ickn

ess

(n

m)

/ M

ass

(g

/m2)

Thickness, nm Mass, g/m2

0

5

10

15

20

25

30

C-PAM Linear anionic polymer Montmorillonite clay

Added component

Mass (

g/m

2)

/ D

issip

ati

on

Dissipation Mass, g/m2

DPI: Three-component system

(linear anionic polymer)

QCM: Three-component system

(linear anionic polymer)

0

2

4

6

8

10

C-PAM Structured anionic

polymer

Montmorillonite clay

Added component

Th

ickn

ess

(n

m)

/ M

ass

(g

/m2)

Thickness, nm Mass, g/m2

0

5

10

15

20

25

30

C-PAM Structured anionic

polymer

Montmorillonite clay

Added component

Mass (

g/m

2)

/ D

issip

ati

on

Dissipation Mass, g/m2

DPI: Three-component system

(structured anionic polymer)

QCM: Three-component system

(structured anionic polymer)

Fig 25. DPI results (left column) and QCM results (right column) for three systems which are described in the text. The DPI results are not with associated water, presenting the adsorbed mass per area unit (●, g/m2) and the thickness of the adsorbed layer (○, nm). The QCM results include associated water and present the adsorbed mass per area unit (●, g/m2) and the dissipation (○, unitless).

49

6 Conclusions The following conclusions were drawn from this thesis work:

I. The R-F machine was found to be a significant and important tool for predicting a ranking order of different retention aid systems, and the overall results demonstrated high reproducibility. Changed settings in R-F machine trials (e.g. machine settings, addition levels, furnish characteristics, etc.) gave changes in paper properties and process parameters which correspond well to that observed in full-scale mill evaluations.

II. In one study evaluating the jet-to-wire speed ratio at four levels (1.0, 1.1, 1.2 and 1.3), it was found that a higher jet-to-wire speed ratio resulted in better paper formation. Poorest formation was obtained at the lowest ratio.

III. No change in retention-formation relationship was observed when comparing the single-component retention aids and the two polyacrylamide-based microparticulate retention aid systems. All systems suggested a linear relationship, where an increased retention level corresponded to impaired paper formation.

IV. Considering the single-component retention aids, it was found that a lower molecular weight of the cationic polyacrylamide gave a lower filler retention level. The polyacrylamide-based microparticulate system including montmorillonite clay gave a higher retention level, at the same added amount of cationic flocculant, in comparison with the silica-based system.

V. By using three-component systems, the retention-formation relationship was changed. It was found that the formation could be significantly improved at a given level, or even at an increased level of filler retention. These three-component systems combine a polyacrylamide-based microparticulate retention aid system with a formation aid. The key for these pioneering results was to add the formation aid of a certain structure in excess. The higher the addition of formation aid, the better the formation was.

VI. Contrary to the formation improvements, the drainage rate was reduced when using the three-component systems that included high dosages of formation aids. The improved retention-formation relationship was, however, correlated to a changed chemical composition of the furnish and not by the effect of changed dewatering performance on the wire.

VII. The structure of the high molecular weight anionic polymers as formation aids was decisive for the obtained effects. Three structuring levels of the polymers were investigated, defined in this work as linear, partly structured and structured. The linear and the partly structured polymers showed effects both with respect to formation and dewatering performance. No effects were demonstrated for the same responses when the structured polymer was used.

VIII. The results obtained in the flow loop experiments were consistent with those results found in the R-F machine trials, also regarding the influence of the polymer structure. As single-component additives, both the linear and the partly structured high molecular weight anionic polymer showed effects on the fibre suspension flocculation in terms of reduced mean floc size. No effects were observed for the structured polymer. Also, the results indicated that the addition of polymer needed to be high in order to gain dispersion effects.

IX. The fundamental mechanism of changed retention-formation relationships is related to an effect of having high molecular weight anionic polymers present in the solution and not to adsorption effects involving interactions between adsorbed layers on the fibre surfaces. This was concluded from the adsorption studies and the flow imaging experiments.

X. The flow imaging experiments showed that formation aid acts as drag reducing additive in turbulent flows. This was shown in terms of reduced signal intensities indicating damped turbulent fluctuations and reduced pressure drops. Consistent with all other results revealed

50

in this work, the polymer structure was found to be crucial for drag reduction effects. The overall results indicate that turbulence damping is a favourable mechanism for improving paper formation, and hence for changing the retention-formation relationship.

51

7 Future work As mentioned in the thesis text, two additional papers are planned. One will focus on the effect of formation aids on fibre suspension flocculation, investigated in the flow loop system. The basis of this paper is described in chapter 4.4. The other paper will focus on adsorption studies of cationic polyacrylamide, anionic high molecular weight polymers and montmorillonite clay on model surfaces (DPI and QCM analyses) and on fibre surfaces (polyelectrolyte titrations). The adsorption studies on fibre surfaces will be performed at the lab during 2012. The aim is to publish both papers in 2012. I would find it very interesting to further study the present three-component retention aid systems in larger-scale formers, such as EUROFEX at Innventia and also industrial state-of-the-art paper machines. From EUROFEX trials we could acquire additional information about what levels in formation improvements we can obtain with such systems and how the dewatering performance is influenced, in more authentically industrial settings. This additional information can promote full-scale industrial trials as a next step. For these kinds of three-component systems, where the key is to add a drag reducing additive in surplus together with a polyacrylamide-based microparticulate system, optimization is needed for each system when we discuss larger-scale trials. Optimizations in terms of added amounts, type of drag reducing additive and machine configurations (dewatering capacity etc.) are needed. Nanocellulose as a drag reducing additive could be one idea to include in a future large-scale investigation aiming to improve the retention-formation relationship. I can see two areas of interest regarding product development coming out from this thesis work. The first is the development of new kinds of retention aid systems or better specification of existing systems, which can improve paper formation at the same level, or at a higher level, of retention. The second is the development of better headbox designs regarding generation of turbulence and reflocculation.

52

Acknowledgements I have so many to thank in so many ways. I want to thank everybody who took the initiative in starting this industrial PhD project in 2004 and all the industrial partners that have contributed time, materials, knowledge and financing during the project. That is to say, thanks to Processum, KTH, Innventia, MoRe Research, M-real, Eka Chemicals, BASF and Imerys. A great thank-you particularly to Kempe Stiftelserna for financing this project over all five years. Continuing the acknowledgements in some chronological order, I want to say thanks to Peter Blomqvist (CEO at that time) and Jennie Söderström for the warm welcome when I was employed by Processum. Thanks, Peter, for inspiring me to do a good job and for all the encouraging conversations in which you always found a new way to see problems - things never felt impossible. Thanks, Jennie, for the hearty support you have given me, both at work and after hours. Thanks for your encouragement when I was applying for scholarships before the trip to the US; you believed in me at a time when I needed it. Next, I want to thank Pär Odén and Torbjörn Sjölund for all their efforts related to the R-F machine. Thanks, Pär, for two fantastic frustrating years when we were trimming the machine together. I will, despite all the breakdowns, remember great patience, fighting spirit and joy. Thanks, Torbjörn, for running the machine in the best of ways. I have really appreciated your angelic patience and your flexibility during the trials. Thanks to all of you who have been involved in the project steering committee, at specific times or over the course of the entire project, including Sune Wännström, Åsa Möller, Marie-Louise Wallberg, Tage Konradsson, Gudrun Pauler-Johansson, Per-Ola Eriksson, Sture Noreus, Stefan Svensson, Jan-Erik Hägglund, Francois Lambert, Petra Kreij and Kjell Andersson. Thank you all for your engagement, fruitful discussions and for sharing experience, knowledge and new thinking related to the research. I especially want to thank Per-Ola Eriksson for valuable and interesting discussions and for so many times showing that you really care. Many thanks to Lisa Nordin and Peter Nilsson for your collaboration and good degree projects. Both of the degree projects have contributed a lot to my research. I also want to thank the present CEO of Processum, Clas Engström, and my dear colleagues Ing-Mari DeWall and Emma Johansson. Clas, you have just been damn good in your role as a leader. A lot of what I learned during the time we worked together I will carry with me for the rest of my life. Ing-Mari, thanks a lot for helping with all paperwork. Emma, thanks for being such a nice room-mate, and for your support and our “five-minute talks”. I also want to acknowledge the personnel at M-real in Husum and MoRe Research who have been involved in pulp collection, analysis, chemical preparation, R-F machine trials and other tasks. Next, I want to express my thanks to Mikael Ankerfors for your collaboration in the flow loop experiments at Innventia. It has been a real pleasure working with you. Thanks also to Åsa Blademo and Gunborg Glad-Nordmark for your assistance and analysis associated with these flow loop experiments. That the Gunnar Sundblad Foundation gave attention to my research and awarded my project the 2010 competence development prize made me very glad and proud. This scholarship of

53

500 000 SEK was used for a stay at the University of California in Davis, US, to develop knowledge about flocculation mechanisms in papermaking. The research at UC Davis became successful and contributed a lot towards arriving at a good fundamental understanding of the other results obtained within this thesis work. So, a great thanks to the Gunnar Sundblad Foundation for making all of this possible. I also want to express my gratitude to Emilio Tozzi, David Lavenson, Mike McCarthy and Robert Powell, all at UC Davis. Thanks for the warm welcome, good research and enjoyable collaboration. My stay in Davis will be a memory for life thanks to you. Simon Utsel, thanks a lot for your contribution to this thesis work, in the form of adsorption studies at our lab at KTH. Lars Ödberg, thanks a lot for the preliminary examination of my thesis work. Two people who have followed me from the beginning of this PhD work and have meant a lot to me are my supervisors Tom Lindström and Lars Wågberg at KTH. As my main supervisor, I want to dedicate my largest thanks to you, Tom. All of your experience in research as well as in life and all your knowledge have meant so much, both for this thesis work and also for me as a person. We have had an eventful and interesting journey together during these years and I am very glad to have had the opportunity to get to know you so well. Thanks also for introducing me to the world of Russian River Pinot Noirs. Lars, thanks for your helpful supervision, fruitful discussions and splendid comments on results and design of experiments. Last but certainly not least, I want to thank my family: my mother Eivor, my father Tomas, my sister Karin, my children Oskar and Selma and my husband-to-be Rolf. Thanks, Mum, for everything; for being so kind to me in all situations, for always believing in me, always encouraging me and for all your love. Thanks, Dad, for being the best dad ever. Karin, thanks for all your support and encouragement during these years. I am looking forward to celebrating with you in NY. Rolf, this thesis would not have been possible without you. I cannot thank you enough. Rolf; I love you for being so great during this time, for taking such good care of our children, for your patience and willingness to assist as best you can in all situations. Oskar and Selma, this thesis is dedicated to you.

54

References Aksberg, R. and Ödberg, L. (1990): Sequential adsorption of cationic and anionic polyelectrolytes on bleached cellulosic fibres. Nord. Pulp Pap. Res. J. 5, p.168. Albinsson, C. J., Swerin, A. and Ödberg, L. (1995): Formation and retention during twin-wire blade forming of a fine paper stock, Tappi J., 78(4), pp. 121-128. Alfano, J. C., Carter, P. W. and Whitten, J. E. (1999): Use of scanning laser microscopy to investigate microparticle flocculation performance, J. Pulp Paper Sci., 25(6), pp. 189-195. Andersson, K. (1984): Compozil – a multifunctional system of paper chemicals, Proceedings 1984 SPCI Conference, Stockholm, Sweden, SPCI, pp. 241-243. Andersson, K. and Lindgren, E. (1996): Important properties of colloidal silica in microparticulate systems, Nord. Pulp Pap. Res. J., 11(1), pp. 15-21. Arola, D. F., Barrall, G. A., Powell, R. L., McCarthy, K. L. and McCarthy, M. J. (1997): Use of nuclear magnetic resonance imaging as a viscometer for process monitoring, Chemical Engineering Science, 52(13), pp. 2049-2057. Arola, D. F., Powell, R. L., McCarthy, M. J., Li, T. Q. and Ödberg, L. (1998): NMR imaging of pulp suspension flowing through an abrupt pipe expansion, AIChE Journal, 44(12), pp. 2597–2606. Arola, D. F., Barrall, G. A., Powell, R. L. and McCarthy, M. J. (1999): Pointwise observations for rheological characterizations, Journal of Rheology, 43(1), pp. 9-30. Beghello, L. and Lindström, T. (1998): The influence of carboxymethylation on the fibre flocculation process, Nord. Pulp Pap. Res. J., 13(4), pp. 269-273. Bernié, J-P., Pande, H. and Gratton, R. (2006): An instrumental determination of the effect of sheet formation on the printability of uncoated fine paper, Tappi Journal, 5(10), pp. 28-32. Bown, R. (1985): Review of methods for increasing filler loadings, Paper Tech. and Ind., 26(6), pp. 289-292. Brouillette, F., Morneau, D., Chabot, B. and Daneault, C. (2004): Paper formation improvement through the use of new structured polymers and microparticle technology, Pulp Paper Can., 105(5), pp. 108-112. Carré, B. (1993): Starch and alumina/silica based compounds as a microparticle retention aid system – an explanation of the synergisms, Nord. Pulp Pap. Res. J., 8(1), pp. 21-26. Clémençon, I. and Gerli, A. (1999): The effect of flocculant/microparticles retention programs on floc properties, Nord. Pulp Pap. Res. J., 14(1), pp. 23-29. De Roos, A. (1958): Stabilization of fibre suspensions, Tappi, 41(7), p. 343-358.

55

Eklund, D. and Lindström, T. (1991): Paper chemistry - an introduction, DT Paper Science Publications, Grankulla, Finland. Erspamer, A. (1940): The flocculation and dispersion of papermaking fibers, Annual meeting of the Technical Association of the Pulp and Paper Industry, New York, pp. 132-137. Evans, F. and Wennerström, H. (1999): The colloidal domain: where physics, chemistry, biology and technology meet (2nd edition), John Wiley & Sons, New York. Fellers, C. and Norman, B. (1998): Pappersteknik, Institutionen för Pappersteknik, Royal Institute of Technology, Stockholm, ISBN: 91-7170-741-7, p. 100, (In Swedish). Gerli, A., Oosterhof, F. and Keiser, B. A. (1999): An inorganic nanosize particle – part of a new retention/dewatering system, Paper Technology, 40(10), pp. 41-45. Gregory, J. (1973): Rates of flocculation of latex particles by cationic polymers, J. Coll. Interface Sci., 42, pp. 448-456. Gyr, A. and Bewersdorff, H.W. (1995): Drag reduction of turbulent flows by additives, Fluid mechanics and its applications, Volume 32, Kluwer Academic Publishers, Netherlands, ISBN 0-7923-3485-X. Hallgren, H. and Lindström, T. (1989): The influence of stock preparation on paper forming efficiency on a paper machine, Paper Techn. Ind., 20(2), pp. 35-39. Hayes, A. J. (1985): 40 % filler loaded paper…dream or reality?, Paper Techn. Ind., April/May, pp. 129-132. Hedborg, F. and Lindström, T. (1996): Some aspects on the reversibility of flocculation of paper stocks, Nord. Pulp Pap. Res. J., 11(4), pp. 254-259. Honig, D. S., Harris, E. W., Pawlowska, L. M., O`Toole, M. P. and Jackson, L. A. (1993): Formation improvements with polymeric microparticle systems, Tappi Papermakers’ Conference, Atlanta, 1979, Tappi Press, Atlanta, pp. 153-162. Honig, D. S., Farinato, R. S. and Jackson, L. A. (2000): Design and development of the micro-polymer system – an organic microparticle retention/drainage system, Nord. Pulp Pap. Res. J., 15(5), pp. 536-544. Honig, D. S. (2005): Organic microparticles, Micro and Nanoparticles in Papermaking, Ed. By Rodriguez, J.M., Tappi Press, Norcross, GA, USA. Horn, D. and Linhart, F. (1991): Retention aids, Paper Chemistry, J. C. Roberts (Ed.), Blackie, Glasgow and London, Chapman and Hall, New York, ISBN: 0-216-92909-1, Chap. 4. Horvath, E. and Lindström, T. (2007): The influence of colloidal interactions on fiber network strength, J. Coll. Interface Sci., 309, pp. 511-517. Huang, T. and LePoutre, P. (1994): Effect of basestock roughness, absorbency and

56

formation on uncalendered coated paper properties, 1994 Coating Conference, San Diego, California, USA, May 1-5, Tappi Press, p. 263-268. Hubbe, M. A. (2005): Mechanistic aspects of microparticle systems, Tappi J., 4(11), pp. 23-28. Huber, P., Pierre, C., Bermond, C. and Carré, B. (2004): Comparing the fiber flocculation behaviour of several wet-end retention systems, Tappi J., 3(4), pp. 19-24. Ives, K. J. (1978): The scientific basis of flocculation, Ed. Sijthoff & Nordhoff, Series E:Applied science, ISBN: 9028607587. Jaycock, M. J. and Swales, D. K. (1994): The theory of retention, Paper Technology, 35(8), pp. 26-33. Johansson, P-A. and Norman, B. (1996): Methods for evaluating formation, print unevenness and gloss variations developed at STFI, Process and Product Quality Conference, Cincinnati, OH, USA, October 14-17, 1996, Tappi Press, pp. 139-145. Jokinen, O. and Ebeling, K. (1985): Flocculation tendency of papermaking fibres, Paperi ja Puu (Papper och Trä), 67(5), pp. 317-325. Jokinen, O. and Palonen, H. (1986): Interdependence of retention and formation in the manufacture of SC paper, Paperi ja Puu (Papper och Trä), 68(11), pp. 801-808. Jordan, B. D. (1985): Predicting the effect of formation on opacity and scattering coefficient, J. Pulp Pap. Sci., 11(2), pp. 56-59. Kajanto, I. M. (1989): Effect of formation on print quality with woodfree offset papers, Nord. Pulp Paper Res. J., 4(1), pp. 8-15. Kasper, D. R. (1971): Theoretical and experimental investigations of the flocculation of charged particles in aqueous solutions by polyelectrolytes of opposite charge, Ph.D. Thesis, California Institute of Technology, Pasadena. Kerekes, R. J. and Schell, C. J. (1992): Characterization of fibre flocculation regimes by a crowding factor, J. Pulp Pap. Sci., 18(1), pp. 32-38. Kiviranta, A. J. and Dodson, C. T. J. (1995): Evaluating fourdrinier formation performance, J. Pulp Pap. Sci., 21(11), pp. 379-382. Kiviranta, A. J. (1996): Fiber and forming-related mechanisms affecting formation, TAPPI 1996 Papermakers Conference Proceedings, Tappi Press, Atlanta, pp. 239-245. Krogerus, B. (1994): Impact of retention polymers on flocculation, retention, drainage and sheet formation – a laboratory study, 1994 TAPPI Papermakers Conf., April 24-27, San Francisco, California, USA, TAPPI Press, Atlanta, GE, USA, pp. 445-452. Kruyt, H. R. (1952): Colloidal science, Elsevier, New York. LaMer, V. K. and Healy, T. W. (1963): Adsorption-flocculation reactions of macromolecules at the solid-liquid interface, Rev. Pure Appl. Chem., 13, pp. 112-133.

57

Langley, J. G. and Litchfield, E. (1986): Dewatering aids for paper applications, Tappi Papermakers Conference, New Orleans, USA, 1986, Tappi Press, pp. 89-92. Lee, P. and Lindström, T. (1989): Effects of high molecular mass anionic polymers on paper sheet formation, Nord. Pulp Pap. Res. J., 4(2), pp. 61-70. Lee, W. K., Vaseleski, R. C. and Metzner, A. B. (1974): Turbulent drag reductions in polymeric solutions containing suspended fibres, AIChE Journal, 20, pp. 128-132. Ledda, J., Ford, P. A., Miyazaki, R. and Hasegawa, M. (2005): Ciba Telioform system – A new multi component organic/inorganic system, Japan Tappi Journal, 59(8), pp. 1181-1187. Li, T. Q. and McCarthy, K. L (1995): Pipe flow of aqueous polyacrylamide solutions studied by means of nuclear magnetic resonance imaging, J. Non-Newtonian Fluid Mech., 57, pp. 155-175. Li, T. Q., Seymour, J. D., Powell, R. L., McCarthy, K. L., Ödberg, L. and McCarthy, M. J. (1994a): Turbulent pipe flow studied by time-averaged NMR imaging: Measurements of flow velacity profile and turbulent intensity, Magnetic Resonance Imaging, 12(6), pp. 923-934. Li, T. Q., Powell, R. L., McCarthy, M. J., McCarthy, K. L. and Ödberg, L. (1994b): Velocity measurements of fiber suspensions in pipe flow by the nuclear magnetic resonance imaging method, TAPPI (Technical Association of the Pulp and Paper Industry) Journal 77(3), pp. 145-149. Lindström, T. (1989): Some fundamental chemical aspects on paper forming, Fundamentals of Papermaking, Baker C. F., Punton V. W. (Eds.), Trans. Ninth Fundam. Res. Symp., Cambridge, Mech. Publ. Eng. Publ. Ltd., London, 1989, pp. 311-412. Lindström, T., Hallgren, H. and Hedborg, F. (1989): Aluminum based microparticle retention aid systems, Nord. Pulp Pap. Res. J., 4(2), pp. 99-103. Lindström, T., Mähler, A. and Norman, B. (2006): Formation/retention relationships from STFI-Packforsk EUROFEX trials, 6th International Paper and Coating Chemistry Symp. 2006, June. Lindström, T., Wågberg, L. and Hallgren, H. (1986): Chemical additives strategies for the production of highly filled fine papers, EUCEPA/ATICELCA, 22nd Conf., Florence, Devt. & Trends in Sci. & Technol. of Pulp & Papermaking, Vol. 1, Paper no 13, pp. 13:1- 13:17. Mabee, S. W. (2001): Controlled filler preflocculation – improved formation, strength and machine performance, 2001 Tappi Papermakers Conferences Proceedings, March 11- 14, Cincinnati, Ohio, USA, Tappi Press. Martinez, D. M., Buckley, K., Jivan, S., Lindström, A., Thiruvengadaswamy, R., Olson, J. A., Ruth, T. J. and Kerekes, R. J. (2001): Characterizing the mobility of papermaking fibres during sedimentation, 12th Fundamental Research Symposium, Oxford, September, pp. 225-254.

58

Marx, K. A. (2003): Quartz crystal microbalance: a useful tool for studying thin polymer films and complex biomolecular systems at the solution-surface interaction, Biomacromolecules, 4(5), pp. 1099-1120. Mason, S. G. (1954): Fibre motions and flocculation, Pulp Pap. Mag. Can., 55(13), pp. 96- 102. McComb, W. D. and Chan, K. T. J. (1981): Drag reduction in fibre suspension, Nature, 292, pp. 520-523. Meyer, R. and Wahren, D. (1964): On the elastic properties of three-dimensional fiber networks, Svensk Papperstidn., 67(10), pp. 432. Moffet, R. H. (1994): On-site production of a silica based microparticulate retention and drainage aid, 1994 Tappi Papermakers’ Conference, San Francisco, USA, Tappi Press, Atlanta, pp. 243-255. Nazhad, M. M., Karnchanapoot, W. and Palokangas, A. (2003): Some effects on fibre properties on formation and strength of paper, Appita Journal, 56(1), pp. 61-65. Nordström, B. (2003): Effects of headbox tube design and flow rate on formation and other sheet properties in twin-wire roll forming, Nord. Pulp Pap. Res. J., 18(3), pp. 296-302. Norman, B. (1974): An investigation of the mass distribution in paper sheets, Ph.D. Thesis, Royal Institute of Technology, Stockholm, Sweden. Norman, B. and Wahren, D. (1974): The measurement of mass distribution in paper sheets using a beta radiographic method, Svensk Papperstidn., 77(11), pp. 397-406. Norman, B. and Wahren, D. (1973): Mass distribution and sheet properties of paper, Transactions of the 5th International Fundamental Research Symposium, Cambridge. pp. 7-73. Norman, B. (1989): Overview of the physics of forming, Fundamentals of Papermaking, C. F. Baker (Ed.), Transactions of the Ninth Fundamental Research Symposium held at Cambridge, Mechanical Engineering Publications Ltd., London, Volume 3, pp. 73-149. Norman, B. (2009): Beta-radiation based on grammage formation measurement - Radiogram methods applicable to paper and light weight board, Nordic Standardization Programme Report No. 5. Norman, B. and Söderberg, D. (2001): Overview of forming literature, 1990-2000, Science of Papermaking, C. F. Baker (Ed.), Transactions of the 12th Fundamental Research Symposium held in Oxford, September 2001, The Pulp & Paper Fundamental Research Society, Vol. I, pp. 431-558. Ovenden, C. and Xiao, H. (2002): Flocculation behaviour and mechanisms of cationic inorganic microparticle/polymer systems, Colloids and Surfaces, A: Physicochemical and Engineering Aspects, 197 (2002), pp. 225-234.

59

Ptasinski, P. K., Nieuwstadt, F. T. M., Van Den Brule and Hulsen, M. A. (2001): Experiments in turbulent pipe flow with polymer additives at maximum drag reduction, Flow, Turbulence and Combustion, 66, pp. 159-182. Radin, I., Zakin, J. L. and Patterson, G. K. (1975): Drag reduction in solid-fluid systems, AIChE Journal, 21, pp. 358–371. Rooks, A. (2004): Managing retention, drainage and formation, Solutions, 87(6), pp. 40-42. Sauerbrey, G. (1959): The use of quartz oscillators for weighing thin layers and for microweighing, Zeitschrift fuer Physik, Volume: 155, pp. 206-222. Sikora, M. D. and Stratton, R. A. (1981): The shear stability of flocculated colloids, Tappi, 64(11), pp. 97-101. Silenius, P. (2003): Improving the combinations of critical properties and process parameters of printing and writing papers and paperboards by new paper-filling methods, Dr of Science in Technology Thesis, Dept. of Forest Products Technology, HUT, Finland. Smith, J. H. (1991): Laboratory comparisons of various multicomponent retention systems, Tappi Papermakers’ Conference, Seattle, USA, 1991, Tappi Press, pp. 481-499. Solberg, D. and Wågberg, L. (2003): On the mechanisms of cationic-polyacrylamide-induced flocculation and re-dispersion of a pulp fiber dispersion, Nord. Pulp Pap. Res. J., 18(1), pp. 51-55. Soszynski, R. and Kerekes, R. J. (1988): Elastic interlocking of nylon fibres suspended in liquid, Part 1, Nature of cohesion between fibers, Nord. Pulp Pap. Res. J., 3(4), pp. 172- 179. Stiernstedt, J., Brumer, H., Zhou, Q., Teeri, T. T. and Rutland, M. W. (2000): Friction between cellulose surfaces and effect of xyloglucan adsorption, Biomacromolecules, 7(7), pp. 2147-2153. 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, J. Colloid Interface Sci., 303(1), pp. 117-123. Svedberg, A. and Lindström, T. (2010a): A pilot web former designed to study retention-formation relationships, Nord. Pulp Paper Res. J., 25(2), pp. 185-194. Svedberg, A. and Lindström, T. (2010b): The effect of various retention aids on retention and formation, Nord. Pulp Paper Res. J., 25(2), pp. 195-203. Svedberg, A. and Lindström, T. (2012): Improvement of the retention-formation relationship using three-component retention aid systems, Nord. Pulp Paper Res. J., 27(1), pp. 86-92. Svedberg, A., Tozzi, E. J., Lavenson, D. M., Lindström, T., McCarthy, M. J. and Powell, R. L. (2012): The effects of formation aids on flow properties of cellulosic fibrous suspensions, Manuscript.

60

Svensson, O. and Österberg, L. (1965): Våta suglådors funktion, Del 3 - Inverkan på arkets egenskaper, Svensk Papperstidn., 68(11), pp. 403-418 (In Swedish).

Swann, M. J., Peel, L. L., Carrington, S. and Freeman, N. J. (2004): Dual-polarization interferometry: an analytical technique to measure changes in protein structure in real time, to determine the stoichiometry of binding events, and to differentiate between specific and nonspecific interactions, Anal. Biochem., 329(2), pp. 190-198. Swerin, A. (1995): Flocculation and fibre network strength in papermaking suspensions flocculated by retention aid systems, Ph.D. Thesis, Royal Institute of Technology, Stockholm, Sweden. Swerin, A., Glad-Nordmark, G. and Sjödin, U. (1995): Silica based microparticulate retention aid systems, Paperi Puu (Papper och Trä), 77(4), pp. 215-221. Swerin, A., Ödberg, L. and Wågberg, L. (1996): An extended model for the estimation of flocculation efficiency factors in multicomponent flocculant systems, Colloids and Surfaces A: Physicochemical and Engineering Aspects 113, pp. 25-38. Swerin, A. and Mähler, A. (1996): Formation, retention and drainage of a fine paper stock during twin-wire roll-blade forming - implications of fibre network strength, Nord. Pulp Pap. Res. J., 11(1), pp. 36-42. Swerin, A., Risinger, G. and Ödberg, L. (1997): Flocculation in suspensions of microcrystalline cellulose by microparticle retention aid systems, J. Pulp Paper Sci., 23(8), p. 374-381. Swerin, A., Sjödin, U. and Ödberg, L. (1993): Flocculation of cellulosic fibre suspensions by model microparticulate retention aid systems – effect of polymer charge density and type of microparticle, Nord. Pulp Pap. Res. J., 8(4), pp. 389-398. Swerin, A. and Ödberg, L. (1996a): Flocculation and floc strength – from the laboratory to the FEX paper machine, Das Papier, 50(10a), pp. 45-47. Swerin, A. and Ödberg, L. (1996b): Flocculation of cellulosic fibre suspensions by a microparticulate retention aid system consisting of cationic polyacrylamide and anionic montmorillonite – effect of contact time, shear level and electrolyte concentration, Nord. Pulp Pap. Res. J., 11(1), pp. 22-29. Swerin, A. and Ödberg, L. (1997): Some aspects of retention aids, Fundamentals of Papermaking Materials, Vol. 1, C. F. Baker (Ed.), Transactions of the 11th Fundamental Research Symposium held at Cambridge, September 1997, Pira International, Surrey, UK, pp. 265-350. Tanaka, H., Swerin, A. and Ödberg, L. (1993): Transfer of cationic retention aid from fibers to fine particles and cleavage of polymer chains under wet-end papermaking conditions, Tappi J., 76(5), pp. 157-163.

61

Toms, B. A. (1948): Some observations on the flow of linear polymer solutions through straight tubes at large Reynolds numbers, North Holland, Amsterdam, Proc. 1st Intern. Congr. on Rheology 2, pp. 135-141. Van de Ven, T. G. M. and Alince, B. (1996): Association-Induced Polymer Bridging: New insights into the retention of fillers with PEO, J. Pulp Paper Sci., 22(7), pp. 257–J263. Van Olphen, H. (1977): An introduction to clay colloid chemistry, Interscience Publishers, Div. of John Wiley & Sons, 605 Third Ave., New York 16, N. Y. Wahren, D. (1972): Flocculation phenomena in wet processing, Proceedings Symp. Man-made Polymers in Papermaking, Helsinki, Finland, 1972, The Finnish Paper Engineers’ Association, The Finnish Pulp and Paper Institute, Helsinki, Finland, pp. 241- 255. Wahren, D. (1979): Fiber network structures in papermaking operations, Proceedings, Conference in Paper Science and Technology – The Cutting Edge, Institute of Paper Chemistry, Appleton, USA, pp. 112-129. Wahren, D. (1987): Numerical analysis of the influence of formation on the optical properties of paper, Int. Paper Physics Conference, Mont-Rolland, Montreal, Canada, 1987, CPPA, TAPPI, PITA, p. 1. Wall, S., Samuelsson, P., Degerman, G., Skoglund, P. and Samuelsson, A. (1992): The kinetics of heteroflocculation in the system cationic starch and colloidal anionic silicic acid, J. Coll. Interface Sci. 151(1), pp. 178-188. Wasser, R. (1978): Formation aids for paper - an evaluation of chemical additives for dispersing long-fibered pulps, Tappi, 61(11), pp. 115-118. Woolwage, J. C. (1939): The flocculation of papermaking fibres, TAPPI Papers Series 22, pp. 578. Wågberg, L., Ödberg, L., Lindström, T. and Aksberg, R. (1988): Kinetics of adsorption and ion-exchange reactions during adsorption of cationic polyelectrolytes onto cellulosic fibers, J. Coll. Interface Sci, 123(1), pp. 287-295. Xiao, H., Pelton, R. and Hamielec, A. (1996): Retention mechanisms for two-component systems based on phenolic resins and PEO or new PEO-copolymer retention aids. Journal of Pulp and Paper Science, 22 (12), pp. 47-485. Yan, H., and Norman, B. (2006): Fibre floc behaviour of softwood kraft pulp in flowing suspensions, Nord. Pulp Pap. Res. J., 21(19), pp. 24-29. Yan, H., Lindström, T. and Christiernin, M. (2006): Some ways to decrease fibre suspension flocculation and improve sheet formation, Nord. Pulp Pap. Res. J., 21(1), pp. 36-43. Yan, H., Norman, B. and Lindström, T. (2006): A flow loop system for study of fibre suspension flocculation, Nord. Pulp Pap. Res. J., 21(1), pp. 19-23.

62

Zhao, R. and Kerekes, R. (1993): The effect of suspending liquid viscosity on fiber flocculation, Tappi J., 76(2), pp. 183-188. Zhou, O., Rutland, M.W., Teeri, T.T. and Brumer, H. (2007): Xyloglucan in cellulose modification, Cellulose, 14(6), pp. 625-641.