UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Flexographic ink behaviour during newspaper repulping Nesbit, Susan Elizabeth 1999

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1999-389499.pdf [ 10.13MB ]
Metadata
JSON: 831-1.0058614.json
JSON-LD: 831-1.0058614-ld.json
RDF/XML (Pretty): 831-1.0058614-rdf.xml
RDF/JSON: 831-1.0058614-rdf.json
Turtle: 831-1.0058614-turtle.txt
N-Triples: 831-1.0058614-rdf-ntriples.txt
Original Record: 831-1.0058614-source.json
Full Text
831-1.0058614-fulltext.txt
Citation
831-1.0058614.ris

Full Text

FLEXOGRAPHIC INK BEHAVIOUR DURING NEWSPAPER REPULPING by Susan Elizabeth Nesbit B.A, The University of British Columbia, 1981 B.A.Sc, The University of British Columbia, 1988 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Chemical Engineering) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA March 1999 © Susan Elizabeth Nesbit, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of C ^ H t Q ^ (r . f r i . U >^<JlLM^fTlLi ll)Ox The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract Newspapers printed with water-based flexographic ink create great difficulties in deinking plants by causing unacceptably low levels of product pulp brightness. Therefore, notwithstanding this ink's relatively benign environmental impact and low safety hazard during printing, deinking mills shun flexographically printed newspaper stock. The paper is avoided because mitigating strategies based on a fundamental understanding of flexographic ink behaviour during deinking mill unit operations are scarce. One reason that little scientific work in this area has been published relates to difficulties in quantifying the response of the ink to different deinking conditions. Flexographic ink behaviour has been studied here by performing experiments on a model repulping system. A novel technique has been developed to measure the degree of residual ink in the repulped paper fibre from experimental trials. This technique, along with pulp brightness, has been used to investigate the effect of different operating conditions during repulping on the net detachment of ink. It is hypothesised that the mechanisms of flexographic ink detachment during repulping are pH dependent and relate primarily to the solubility of the ink's resin component. Once detached, the interaction of ink particles with pulp fibres follows the competing capture and escape phenomena predicted by Langmuir. Thus, if the ink particles are sufficiently small and the ratio of colloidal to hydrodynamic forces is appropriate, then deposition onto both the outside and inside surfaces of fibres can occur. ii Measuring the effects of pH, repulping time, and repulping power during the model experiments has tested the hypothesis. The effect of pH on both ink particle size and ^-potential has been measured. Also, the solubility of the ink's acrylic polymer binder has been measured indirectly as a function of pH. Results from these trials, as well as analysis of published data, support the hypothesis. It has been found that the residual ink on fibres repulped in model experiments is low when the repulping pH is high and, conversely, high when fibres are repulped under acidic conditions. Ink binder solubility appears to correlate closely with these results. The relatively high degree of ink detachment achieved during basic repulping likely results from binder dissolution and subsequent ink particle dispersion. An estimate of colloidal interaction energies between ink particles and fibres under high pH conditions indicates that the counterion concentration in these repulping experiments is not sufficiently large to extinguish the repulsive electrostatic interaction energy between ink particles and pulp fibres. However, calculations have shown that the dampening of this repulsive energy is significant during conditions typically found during commercial repulping. Thus, although ink particles may be stabilised by both electrostatic and steric effects during the model experiments, during commercial repulping they are likely stabilised only by a steric barrier that results from ink binder adsorption onto the surface of ink pigment particles. Interestingly, the high levels of ink detachment achieved during high pH repulping may not be attained by binder dissolution alone. Experimental results suggest that some mechanical action (i.e., repulping energy) improves ink detachment even under alkaline conditions. It is suggested here that the repulping energy^ whose iii purpose is to defibre the paper, also liberates at least some ink particles that would otherwise remain trapped within the tortuous confines of the newsprint. Too much repulping energy decreases the final pulp brightness of repulped fibres. Thus mechanical energy should be optimised in the repulping unit if the pulp brightness is to be maximised. Because ink binder does not dissolve in acidic solutions, it is proposed that ink detachment during low pH repulping occurs as a result of shear induced by mechanical energy at the fibre-ink interface. Experimental results support this explanation of ink detachment. However, it has also been found that high levels of repulping power (i.e. approximately 800 W/kg pulp) cause an increase in the residual ink in pulp. This increase may be due to comminution and subsequent deposition of detached ink particles under high power conditions. It has been shown here that such deposition likely follows the Langmuir model mentioned previously. As in the case of alkaline repulping, the mechanical energy imparted to the acidic repulping system should be optimised if the maximum net ink detachment is to be achieved. Photomicrographs of repulped fibres indicate that, while ink particle deposition onto the inside surfaces of pulp fibres can occur under basic repulping conditions, lumen loading of ink does not occur during acidic repulping. A Langmuir model, that neglects lumen loading has successfully been fitted to data from acidic repulping experiments during which previously dried and ground ink was added to defibred newsprint that had not been printed. This supports the qualitative experimental results and suggests that lumen loading of flexographic ink during neutral or acidic industrial iv repulping would be insignificant. Data collected under acidic conditions during the model experiments during which printed paper was repulped do not follow the trend predicted by the Langmuir equation. This implies that phenomena other than deposition and detachment, such as particle comminution, occur during low pH repulping. Table of Contents Abstract ii Table of Contents vi List of Tables x List of Figures xii Acknowledgements xvi Chapter 1. Introduction 1 Chapter 2. Basic Concepts 4 2.1 Newspaper Printing 4 2.1.1 Offset Printing 4 2.1.2 Water-based Flexography 4 2.2 Water-Based Flexographic Ink 7 2.2.1 Carbon Black 8 2.2.2 Flexographic Binders 9 2.3 Newspaper 10 2.3.1 Wood Fibres 10 2.3.2 Mechanical and Chemical Pulps 12 2.3.3 Newsprint 16 2.4 The Printing Process: Ink-Paper Interactions 17 2.5 Newspaper Deinking 20 2.6 Flexographically Printed Newspaper Deinking 23 Chapter 3. Flexographic Newsprint Deinking Literature Review 24 vi 3.1 Applied Studies 24 3.2 Model Systems 30 3.3 Discussion and Summary 38 Literature Review of Relevant Subject Areas 44 4.1 Introduction 44 4.2 Laundering Systems 45 4.2.1 Particulate Soiling 46 4.2.2 Solid Soil Removal 47 4.2.3 Solid Soil Deposition 51 4.3 Filler Retention Systems 53 4.3.1 Thermodynamics 55 4.3.2 Kinetics 56 4.3.3 Lumen Loading 60 4.4 Discussion 66 4.4.1 Laundering 66 4.4.2 Filler Retention 70 Thesis Objectives 74 Dependent Variable Study 76 6.1 Introduction 76 6.2 The Search for a New Method 78 6.3 Sample Preparation for Thermogravimetric Analysis of Residual Ink 81 6.3.1 Detached Ink Separation 81 vii 6.3.2 Thermogravimetric Analysis Samples 84 6.4 Thermogravimetric Analysis 86 6.5 Thermogravimetric Analysis Sensitivity 92 Chapter 7. Experiments 96 7.1 Experimental Approach 96 7.2 Apparatus 96 7.2.1 Repulping 96 7.2.2> Photographic Equipment 97 7.2.3 Brightness Measurement Equipment 98 7.2.4 Light Absorbance Equipment 98 7.2.5 Electrophoretic Mobilities 98 7.2.6 Particle Size Analysis Equipment 99 7.3 Experimental Materials 100 7.3.1 Chemicals and Photographic Film 100 7.3.2 Model Flexographic Ink 101 7.3.3 Newsprint 103 7.3.4 Printed Paper 103 7.4 Experimental Procedures 105 7.4.1 Pulping Experiments 105 7.4.2 Model Ink Particle Size Experiments 110 7.4.3 Zeta Potential Experiments 111 7.4.4 Light Absorbance of Ink Binder 112 viii 7.4.5 Pulping pH Change Experiments 113 Chapter 8. Results from Factorial Experiments 115 Chapter 9. Detailed Experimental Results 123 Chapter 10. Discussion 146 Chapter 11. Conclusions, Contribution, and Recommendations 168 10.1 Conclusions 168 10.2 Contribution 168 10.3 Recommendations 170 References 173 Appendix 1. Details of the Ink Separation System 184 Appendix 2. Detachment Efficiency Measurements 188 Appendix 3. Experimental Data 199 Appendix 4. Print Characteristics of Model and Commercial Inks 208 Appendix 5. The Light Absorbance of Ink Binder Solutions 210 Appendix 6. Factorial Effects Sample Calculations 213 Appendix 7. The Generation of Total Interaction Energy Curves 216 Appendix 8. The Fit of Equation [10.1] to Published Data 220 ix List of Tables Table 2.1. Repulping chemicals and conditions Table 3.1 The ccc and iep of model ink particles Table 3.2 Model ink particle size and polyelectrolyte adsorption at different NaCl concentrations Table 7.1 Recipe of laboratory manufactured ink Table 7.2 Independent repulping variables Table 7.3 Repulping constants Table 8.1 The effects of sodium hydroxide and brightening chemicals on the carbon black content of rinsed pulp Table 8.2 The effects of repulping time and pH on the carbon black content of rinsed pulp Table 8.3 The effects of repulping power and pH on the carbon black content of rinsed pulp Table 8.4 The effects of initial pH and binder presence on the change in slurry pH during repulping Table 9.1 Analysis of results from repulping pH runs Table 9.2 Analysis of carbon black content data from high pH runs Table 9.3 Analysis of brightness data from high pH runs Table 9.4 Analysis of carbon black content data from low pH runs Table 9.5 Analysis of brightness data from low pH runs Table 9.6 A[ION] with and without binder at different initial pH levels Table A3.1 TGA sensitivity results Table A3.2 Carbon black content vs repulping pH Table A3.3 Ink absorbance/transmittance data Table A3.4 Model ink particle size at different pH levels Table A3.5 Ink particle ^-potentials at different pH levels 21 38 38 102 107 108 118 119 120 121 125 136 137 139 140 144 199 200 201 202 202 Table A3.6 Carbon black pigment ^-potentials at different pH levels 203 Table A3.7 Carbon black content and brightness of rinsed pulp from high pH runs at different repulping times 203 Table A3.8 Carbon black content and brightness of rinsed pulp from low pH runs at different repulping times 204 Table A3.9 Carbon black content and brightness of rinsed pulp from high pH runs at different repulping intensities 205 Table A3.10 Carbon black content and brightness of rinsed pulp from low pH runs at different repulping intensities 205 Table A3.11 pH change during repulping in the absence of ink binder 206 Table A3.12 pH change during repulping in the presence of ink binder 206 Table A3.13 Carbon black content of fibres repulped with and without sodium hydroxide and bleaching chemicals 207 Table A4.1 Comparison of a commercial flexographic ink with the laboratory manufactured ink 208 Table A8.1 Data from flexographic ink deposition experiments in acidic conditions 222 List of Figures Figure 2.1 An illustration of the flexographic printing process 5 Figure 2.2 A plan view illustration of the fibre wall 11 Figure 2.3 The distribution of chemical compounds in the fibre wall 13 Figure 2.4 A plan-view sketch illustrating the difference between chemical and mechanical pulping 13 Figure 2.5 Scanning electron micrographs of chemical and mechanical pulp fibres 15 Figure 2.6 A sketch of the papermaking process 16 Figure 2.7 An illustration of Voith flotation cells 22 Figure 2.8 A drawing of the Black-Clawson Double-Nip Thichener 22 Figure 3.1 The interaction energy between two like spheres 28 Figure 3.2 The effect of pH on z-potential of pigment particles and ink particles 31 Figure 3.3 Brightness data from fibres repulped under acidic conditions 33 Figure 3.4 A plot illustrating pulp brightness as a function of pH, with and without both calcium and ink 34 Figure 4.1 An illustration of soil detachment mechanisms 52 Figure 4.2 The interaction energy plot between titanium a dioxide particle and a fibre surface 62 Figure 4.3 The change with agitation time of external and internal adsorption of titanium dioxide particle on pulp 65 Figure 6.1 A sketch of the detached ink separation apparatus 83 Figure 6.2 An illustration of the vacuum filtration system used to prepare samples for thermogravimetric (TG) analysis 85 Figure 6.3 A plot of TG profiles of carbon black, model ink, and model ink previously washed with solvents 87 Figure 6.4 A photograph of the TG analysis apparatus 88 xii Figure 6.5 TG profiles of carbon black samples heated at different rates Figure 6.6 TG profiles of blank newsprint, repulped paper, and carbon black Figure 6.7 An illustration of TG analysis sensitivity Figure 6.8 A dot plot of the differences between TG data and the microbalance data Figure 6.9 A plot of percent error of the TG data Figure 7.1 A diagram of the experimental repulping system Figure 7.2 An electron micrograph of model ink particles Figure 7.3 Cross-sections of newspapers Figure 8.1 The cube plot showing the effects of brightening chemicals and sodium hydroxide on carbon black content Figure 8.2 The cube plot showing the effects of pH and repulping time on carbon black content Figure 8.3 The cube plot showing the effects of pH and repulping power on carbon black content Figure 8.4 The cube plot showing the effects of initial pH and binder presence on the change in pH during repulping Figure 9.1 The plot of carbon black content vs. repulping pH Figure 9.2 An electron micrograph of fibres repulped under high pH conditions Figure 9.3 An electron micrograph of fibres repulped under low pH conditions Figure 9.4 The plot of light absorbance of binder vs solution pH Figure 9.5 The plot of light absorbance of binder and carbon black content vs repulping pH Figure 9.6 The plot of ink size vs pH Figure 9.7 The plot of model ink and pigment ^-potentials vs pH Figure 9.8 The plot of the ^-potentials of different inks vs pH 90 91 93 94 94 97 102 105 117 119 120 121 124 126 127 128 129 130 131 132 xm Figure 9.9 The plot of the ^-potentials of different pigments vs pH 133 Figure 9.10 The plot of carbon black content of fibres repulped under alkaline conditions vs repulping energy 135 Figure 9.11 The plot of the brightness of fibres repulped under alkaline conditions vs repulping energy 136 Figure 9.12 An micrograph of a fibre repulped at high pH, high shear conditions 137 Figure 9.13 The plot of carbon black content of fibres repulped under acidic conditions vs repulping energy 138 Figure 9.14 The plot of the brightness of fibres repulped under acidic conditions vs repulping energy 139 Figure 9.15 A micrograph of fibres that illustrates lumen loading during alkaline repulping 141 Figure 9.16 The plot of pH change during repulping vs carbon black content 142 Figure 9.17 The plot of pH during repulping vs repulping time 143 Figure 9.18 An illustration of the effect of initial pH on the pH change during repulping 144 Figure 9.19 A micrograph of rinsed fibres that had been repulped under acidic conditions with pigment particles 145 Figure 10.1 The plot of the interaction energy between a sphere and a plate at pH 10 150 Figure 10.2 The plot of the brightness of fibres repulped under basic conditions vs repulping time 155 Figure 10.3 The plot of the ink deposition during acidic repulping vs repulping time 160 Figure A1.1 A diagram showing the method used to seal the mddj 185 Figure A1.2 A sketch of the mddj impeller 186 Figure A2.1 An illustration of the procedure followed to prepare a rinsed pulp sample for TG analysis 191 Figure A2.2 An illustration of the procedure followed to prepare an mddj filtrate sample for TG analysis 194 xiv Figure A5.1 The plot of the absorbance of light at different wavelengths of ink binder vs solution pH 210 Figure A5.2 The plot of the binder absorbance of 470 nm light at 20 °C and30°CvspH 211 Figure A5.3 The plot of the absorbance of binder solutions of differing concentrations vs pH 212 xv Acknowledgements With warmth and respect I gratefully acknowledge the tremendous support given to this project and to me by my advisors Ken Pinder and Bruce Bowen. I also thank my other committee members, Janus Laskowski, Arbeit Horng, and Dick Kerekes, for their advise at critical times. I am especially grateful to Dr. Kerekes for his financial support during the first five years and for his enthusiastic encouragement throughout the project. Several people from Paprican's Vanlab and Pulp and Paper Centre provided valuable assistance to the thesis. They include Phil Allen, Raj Seth, George Williams, and especially Ben Chan and James Drummond. I am particularly grateful to Tim Pattison and Peter Taylor, and also to Rita Penco, who constantly amazes me with her resourcefulness. I would also like to thank Lisa Bradley, Georgina White, and Brenda Dutka for their assistance. Gilles Dorris from Paprican at Pt. Claire, and Bruno Chabot helped a great deal in getting me started in the project and I thank them for this. Also during the early stages, assistance was provided by several people from MB Research. These include Arbeit Horng and also Greg Mickleborough and Les Zanyi. Sally Finora, from the Mining and Mineral Processing Department at UBC, was very helpful and I am grateful to her. I also thank Ed Lim and Tom Troczynski from the Metals and Material Science Department. I acknowledge the assistance of all the staff in the Chemical Engineering Department at UBC, particularly Helsa Leong, Lori Tanaka, Shelagh Plenty, Cecilia Jankowski, Horace Lam, Yeena Feng, Qi Chen, Alex Thng and John Baronowski. I xvi thank these people, and my other friends and collegues in the department, including Colin Oloman (who took the time to proof read the first draft), Peter Roberts, Yuan-Shing Perng, and Cathrine Gaarder, for their encouragement and help at various times. Due to somewhat unusual circumstances, this work has only been finished because of my interaction with friends outside of UBC. Thanks is therefore due to Monica Bittle, Susan D'Aloisio, Mark Force, Brian Gavin, Jennifer Nener, Albert Souza, Darren Quist, Irma Jamont, Alison Fraser, Bob Lewis, Heather Sutherland, Liz Ryan, Cathy Maurer, and especially Martha Whitehead. I would also like to express my deep sense of gratitude to my family. These people include Elmer and Ed Rud, Lynn, Wayne, Lauren, and Kyle Axford, Dick and Joan Dexter, my grandfather John Nesbit, Kevin Doyle, my brother John Nesbit, Kazuko Okuda, and most of all, Maya and Erina Okuda Nesbit. There are five people to whom I feel profoundly grateful: I would like to thank Layne Dexter for encouraging me to enter engineering in 1982 and whose continued encouragement and support allowed me to enter the post-graduate programme. I acknowledge with deep love the emotional and financial support of my parents, Don and Helen Nesbit, to whom I dedicate this thesis. I thank Liam Kain for happily (at least most of the time) putting up with a mother who spent too much playtime working on a thesis. Finally and most importantly, I want to thank Paul Hannig-Kain for so may things, but especially for taking on a thank-less job, and, in so doing, allowing me to finish that which I started. Susan Nesbit, March 1999. xvii Chapter 1. Introduction Global agreements such as those signed in the 1992 Earth Summit Conference (Rio de Janeiro) (e.g., the Agenda 21 action plan) indicate that closed-loop and sustainable processes are recognised as being fundamental to meeting the needs of present and future generations. The renewable and recyclable nature of paper makes the pulp and paper industry an ideal candidate to attain the industrial goals of the 21st century. Indeed, significant work has already resulted in completely contained (i.e., closed-cycle) wood pulping technology. While paper recycling is not yet a closed-loop process and therefore produces waste sludge that is typically deposited as land-fill, it can consume less energy than virgin pulp production (iied, 1996, p. 184). It is thus a natural partner of sustainable paper product manufacturing. Approximately 13% of all paper and board consumed on an annual, global, basis is newsprint (International Fact and Price Book, 1995). The technology of newspaper recycling, which involves repulping and deinking old newspaper to produce a market pulp, is utilised primarily in Japan, Western Europe, and North America. In Canada, 15 repulping and deinking facilities (CPPA, 1997), including two deinking pulp mills, produce up to 1.5 million tonnes of pulp annually. This Canadian recycled fibre is incorporated into newsprint products destined for the American market where several states have enacted legislation demanding up to 50% recovered fibre in published newspapers. Offset and letterpress technology dominate global newspaper printing facilities. However water-based flexography for newsprint is expected to increase its current 15% 1 share of the printing market (Rangamannar et al., 1992) because water-based flexographic ink is relatively inexpensive and produces clean, crisp lettering that does not easily transfer to the hands of the reader. Also, flexography is associated with reduced volatile organic chlorine compound emissions and a safer work environment due to the elimination of fire hazard during press clean up. Unfortunately, even 10% of flexographic newsprint in the deinking mill feed stock produces an unacceptable level of final pulp brightness. Mill experience and bench scale tests have indicated that the difficulties linked to the recycling of flexographic newspaper are due to the relatively hydrophilic, finely dispersed, pigment particles that are released when such newsprint is treated in conventional deinking systems. But fundamental research into the mechanisms responsible for this low final pulp brightness is difficult because the deinking process involves simultaneously repulping paper and detaching ink from a fibrous substrate, then removing ink and other contaminants from the resulting pulp slurry. Paper repulping is particularly complex because the system contained in a deinking mill's repulping unit consists of, not only colloidal particles with a range of surface characteristics, but also relatively large, flexible, somewhat porous, and swelling wood fibres whose surfaces consist of tangles of lignin-rich fibrils. These solid components are suspended in an alkaline soup comprised of pulp brightening chemicals as well as polyelectrolytes and surface active agents contributed by the wood pulp and sometimes by the newsprint ink, but also added to the system to aid in ink detachment. 2 The objective of this thesis is to elucidate and quantify the behaviour of flexographic ink during newspaper repulping. It is hypothesised here that, if the ratio of the attractive colloidal forces to the hydrodynamic forces is sufficiently great, then the ink first detaches from the fibrous substrate and subsequently deposits onto, and re-detaches from, pulp fibres according to a Langmuir kinetic model. The specific ink detachment mechanism and subsequent ink particle kinetic behaviour is primarily dependent on repulping pH. Background information pertinent to flexographic newspaper deinking is reviewed in Chapter 2. A literature review of previous scientific studies of flexographically printed newspaper deinking is presented in Chapter 3 and reviews of two systems similar to that of flexographic newsprint deinking, namely soiled textile laundering and paper filler retention, are presented in Chapter 4. Chapter 5 outlines the thesis objectives. The development and testing of a novel method that allows the direct measurement of residual flexographic ink in deinked pulp produced in a model experimental system is contained in Chapter 6. Chapter 7 gives experimental details of this model system which was designed to address the objectives stated in Chapter 5. Experimental results are presented in Chapters 8 and 9 without discussion except for mention of statistics and experimental difficulties of particular interest. Chapter 10 discusses the thesis hypothesis, recalling data presented in Chapters 8 and 9 where appropriate. Finally, Chapter 11 summarises the important conclusions that can be drawn from the discussion, clearly states the novel contributions of this thesis and makes recommendations for future studies. 3 Chapter 2. Basic Concepts 2.1 Newspaper Printing Letterpress is the traditional, although somewhat outdated, newspaper printing process. It employs an inked, raised, plate, made of photopolymers on a rigid plastic or metal backing (Ferguson, 1992), which is pressed onto newsprint. This technology is largely being replaced by either offset printing or water-based flexography (Ferguson, 1992; Aspler, 1987), both of which produce a higher quality, less expensive, print. 2.1.1 Offset Printing Offset printing uses a photopolymer printing plate that possesses a hydrophobic image area and a hydrophilic non-image area. During newspaper printing, a water-based solution which adheres to the non-image area, is applied to the plate. Next, oil-based ink is applied, thus covering the plate's image area. A compressible rubber blanket, onto which the image is transferred, is then rolled with the plate. Finally, the image is transferred to the paper by rolling the compressible blanket with the paper web (Ferguson, 1992). While offset was the first technology to replace the more cumbersome and poorer quality letterpress, flexography is now competing with offset for the newsprinting market. 2.1.2 Water-based Flexography The flexographic process was originally designed to print solvent-based ink on flexible, non-porous surfaces such as plastic wrap and foil labels and packaging (Benemelis, 1991; Aspler, 1987; Ferguson, 1992; Nunez and Deatherage, 1996). It has subsequently evolved to print water-based ink on permeable substrates such as 4 newsprint (Nunez and Deatherage, 1996). Flexographic printing is similar to letterpress except that a flexible, raised, photopolymeric sheet is inked rather than a raised rigid plate. The printing pressure used during flexography (believed to be between 0.8 and 1.3 MPa (LaDuce, 1984)) is less than that used during either letterpress printing or offset (approximately 3 to 4 MPa (Aspler, 1993,; Aspler and Taylor, 1991)) and the ink, at approximately 0.0.1 Pas, is much less viscous (Aspler, 1987). A schematic drawing of the basic flexographic printing system is shown in Figure 2.1. The fountain roller, made of smooth rubber, transfers ink from an ink bath to an anilox roller. The anilox cylinder is engraved with small wells that enable even distribution of the ink onto the printing plate. The printing plate, once inked, transfers the image to be printed onto the paper web (Nunez and Deatherage, 1996). Water-based flexography has several advantages compared with offset printing (Aspler, 1987; Schipke, 1991; Hruzewiczet al., 1991; Werther, 1991). The press design and operation is simple compared to the offset press (Aspler, 1987). Because the printing start-up is much quicker, it wastes relatively little newsprint. Also, fewer rollers Fountain Roll Impression Roll Figure 2.1. An illustration of the flexographic printing process. 5 mean that a lighter weight of paper can be employed. Because the ink used is water-, rather than oil-based, emission reduction equipment is minimal (Benemelis, 1991). In addition, the ink is less expensive to store, transport, and produce (Werther, 1991). In fact, one estimate suggests that the 1991 annual operating cost of a typical flexographic press is $2,600,000; $400,000 less than that associated with a corresponding offset press (Schipke, 1991). The print quality of flexographic newspaper is at least as high as that of newspaper produced by the offset process (Benemelis, 1991; Aspler, 1987). In addition, flexographically printed newspaper possesses reduced rub-off, reduced show-through, and an improved contrast compared to newspaper printed via the offset process. The pollution and safety hazards of newsprint flexography are significantly less than those associated with offset printing because the ink used is water- rather than oil-based (Benemelis, 1991; Aspler, 1987; Werther, 1991). This means that no volatile organic compounds are emitted during printing (Benemelis, 1991) and that press cleaning does not require the use of organic solvents. All the advantages listed above have resulted in several newspaper printers across North America switching to flexographic technology (over sixteen North American newspapers are currently printed with flexography (Benemelis, 1991)). In Europe, presses in Britain and Italy have switched to flexography. However, there are disadvantages associated with the water-based process. These include ink holdout problems, linting, and printing plate fill-in (Aspler, 1987; Werther, 1991), all of which are 6 caused by relatively fast ink drying and ink flocculation problems. The overwhelming problems associated with water-based newspaper flexography though, are the difficulties that arise when paper printed by flexography is deinked during the recycling process. Before these problems are mentioned in section 2.6 below, a description of the ink and newsprint components and their known interactions, as well as an overview of the deinking process, are provided. 2.2 Water-Based Flexographic Ink The primary constituents of water-based flexographic newspaper ink are its pigment, solvent, and binder or resin. Typically the ink may also contain bactericides, fungicides, and anti-foaming agents (Werther, 1991). The pigment in black newsprint ink is carbon black, which has been ground to a fine powder (with a particle size range of 0.1 to 1 um). Water, the ink's solvent, allows the ink to flow and thus be applied with ease to the printing sheet. Resin disperses the pigment throughout the solvent ensuring even distribution of pigment on the printed page, binds the pigment to the paper substrate and also helps to "set" the ink during ink drying. Different water-based flexographic inks possess different binders, but ink intended to be printed on newspaper usually contains a resin that is a water-soluble polymer of styrene, acrylic acid, or acrylate, associated with ammonia or amines which dissociate in an alkaline environment (Werther, 1991). Ink drying mechanisms may include absorption of the solvent into the paper substrate (Werther, 1991), but recent work has shown that the primary drying mechanism is evaporation of water. 7 Detailed information about basic ink components, which is provided below, may be useful in understanding the difficulties encountered during deinking of black, water-based, flexographic newspaper. 2.2.1 Carbon Black Furnace black, the carbon black pigment found in flexographic newsprint ink, is finely divided carbon formed in the gas phase by oxidation of heavy aromatic hydrocarbon feedstocks (Dannenberg, 1978; Lahaye and Prado, 1991). In general, the chemical composition of this type of carbon black is 97.9% C, 0.7% O, 0.6% S, and 0.4% H. The precise composition, however, is dependent on its formation history (Puri, 1970). The character of carbon black is also dependent on the manufacturing process. During formation, very fine spherical particles develop which cluster into primary aggregates of various shapes of approximately 0.2 to 1.0 ixm diameter. These amorphous aggregates have a microcrystalline structure characterized by disordered platelets of graphite-like layers (Dannenberg, 1978). The surfaces of carbon black particles contain edges of single, unstacked, graphitic layers as well as many defects, discontinuities, and dislocations in the layer planes. These strongly developed surfaces have many active sites which encourage chemisorption of oxygen and hydrogen complexes (Puri, 1970). The most common oxygen functional groups are carboxylic and phenolic, while hydrogen is present in chemisorbed water, as well as is phenolic, hydroquinonic and carboxylic groups. 8 The oxygen groups present on the surface of carbon black particles influence its surface behaviour. For example, while pure carbon is hydrophobic, the wettability of carbon black increases dramatically as the oxygen content increases (Kampe and Sarangapani, 1983; Puri, 1970). Oxygen complexes also have a strong influence on the adsorption of acids and bases. It is known that the aqueous phase of furnace black suspensions have pH values in the alkaline range. Sulfur, also present on the surfaces of carbon black particles, arises from sulfur in the feedstock hydrocarbon from which the carbon black is made. Evidently it is chemically bonded to the carbon. Thus it is a stable feature of the carbon black surface (Dannenberg, 1978; Puri, 1970). 2.2.2 Flexographic Binders Binder (also known as resin) addition to ink ensures pigment dispersion in the solvent prior to printing and subsequent pigment adhesion to the substrate. It also gives the ink desirable printing characteristics such as appropriate viscosity, resolubility (this is important during press cleaning), and print quality (Vash, 1991). The binder type used depends on the printed substrate and the print qualities desired. Resins added to water-based flexographic ink are of three types: 1. water-soluble polymers; 2. emulsion polymers; and 3. alkali-soluble acidic polymers (Sen, 1987; Burke, 1980; Blom and Feig, 1988; Brown, 1980). Flexographic newspaper ink usually contains alkali-soluble binders (Vash, 1991) which allow excellent dispersion of the ink pigment (this is important because pigment is the most costly component in ink; the smaller the pigment particles, the less pigment is 9 needed to obtain a specified colour intensity). They provide wet ink with a good shelf-life, as well as reasonable resolubility and drying rates (Sen, 1987). Also, these resins imbue the dried ink with good water resistance and print gloss (Vash, 1991; Sen, 1987). Alkali-soluble resins are usually acrylic polymers made from a variety of so-called backbone, modifying, and functional monomers (Clarke, 1990). The backbone monomers give the ink its film forming characteristics such as toughness and chemical resistance. Methyl methacrylate and styrene are typical backbone monomers. Ethyl acrylate is often used as a modifying monomer, which, in general, provides flexibility to the dry ink. Functional monomers, such as acrylic acid, give ink specific qualities like alkali solubility. The molecular weight of these acrylic polymers can range from 300 (Sen, 1987) to 50,000 (Brown, 1980), but is typically around 20,000 g/mol. Inks containing alkali-soluble binders usually rely on a volatile alkali to increase the ink pH (as well as increase the ink drying speed). Common bases are ammonia and lower molecular weight alkanolamines such as monoethanolamine (Clarke, 1990). The resulting ink pH is typically around 7.5 to 8.5. 2.3 Newspaper 2.3.1 Wood Fibres Newspaper usually consists of coniferous wood fibres that are bound together by hydrogen bonds. Each wood fibre is approximately 15 to 40 urn in diameter and 1 to 4 mm long (Smook, 1992). Each is composed of cellulose, hemicellulose, and lignin polymers. 10 Cellulose regions in wood fibres are made of aggregates of long cellulose molecules known as fibrils that give the fibre its strength. These fibrils are oriented in a variety of patterns thus making several distinct layers within the fibre wall. Lignin is a complicated amorphous polymer whose primary role is to cement cellulose fibrils in specific orientations. The role played by the hemicellulose polymer is not clearly understood; however, it may facilitate bonding between cellulose and lignin molecules Wall Figure 2.2. A plan view illustration of the fibre wall (not to scale). The schematic diagram of fibre structure given in Figure 2.2 illustrates the layers of typical wood fibres which are named, from the outside inward, the middle 11 lamella, the primary wall, the secondary wall, and the tertiary wall. Also seen in the figure is the fibre's central void that is known as the fibre lumen. Access to the lumen is possible from either end of the fibre but is most likely attained through radially oriented pit pores which occasionally form in the fibre wall (Treimanis, 1996). The frequency and characteristics of pit pores are typically related to wood species. Figure 2.3 shows the distribution of wood fibre components in the fibre wall. Lignin is the dominant molecule in the middle lamella and the primary wall where its role as the matrix holding fibres together is fulfilled. Conversely, close to the fibre's inner (i.e., tertiary) wall, lignin plays only a minor role while cellulose and hemicellulose, both of which add strength to wood, are the primary polymers. 2.3.2 Mechanical and Chemical Pulps Fibres are liberated from wood by either chemical or mechanical pulping, or a combination of the two processes. During chemical pulping, lignin and hemicellulose are dissolved, leaving a pulp of over 90% cellulose. Mechanical pulping, on the other hand, involves shredding and grinding wood chips to produce a high yield pulp containing almost native proportions of lignin and cellulose. Figure 2.4 illustrates the two different mechanisms by which papermaking fibres are obtained from wood. Mechanical pulp fibres are stiff and of varying length. Their surfaces are rough due to shearing forces imparted to fibres during pulping which cause tearing of the middle lamella and unraveling of primary wall layer fibrils (Mohlin, 1977). In addition, mechanical pulp contains a high fines fraction which generally consists of delaminated, disconnected fibril ribbons and flake-like pieces of the middle lamella (see Figure 2.5). 12 In contrast, chemical pulp contains long flexible fibres whose surfaces are relatively smooth and undisturbed (see Figure 2.5). The fines fraction of chemical pulps is relatively low. 100iT • cellulose • lignin S hemicellulose Figure 2.3. The distribution (mass percent) of chemical components in the fibre wall after Scott (1995). ml = middle lamella, p = primary wall, s = secondary wall, t = tertiary wall. —<A Chemical Pulping Mechanical Pulping Figure 2.4. A plan-view sketch illustrating the difference between chemical and mechanical pulping mechanisms. The surfaces of both chemical and mechanical pulp fibres are often characterized by a negative zeta-potential (Pelton, 1993; Lindstrom, 1989). The surface charge arises from ionizable groups of which the most common is the carboxylic type (Lindstrom, 1989; Lindstrom, 1992). On chemical pulp fibres, carboxylic groups result 13 from stopping reactions of cellulose chain degradation during pulping and also from oxidation reactions during bleaching (Scott, 1995). On both mechanical and chemical pulp fibres, the carboxylic groups result from the presence of non-cellulosic wood components, i.e., lignin and hemicellulose. Particularly in the case of mechanical pulp fibres, the charged groups are not evenly distributed on the fibre surface. In addition, the fines fraction of mechanical pulps usually contain a higher portion of charged groups than the fibres because of the relatively high concentration of lignin in the fines (Mohlin, 1977). Much work in determining the lignin distribution throughout the chemical pulp fibre wall has been reported (Roberts, 1996). Like mechanical pulp fibres, the lignin concentration in chemical pulp is greatest on the outer wall of the fibres. (For some wood species, at least, this may be due to precipitation of lignin during the final stages of the cooking process.) The presence of lignin polymers in wood fibres imbues the fibres with polyelectrolytic characteristics (Lindstrom, 1989). Furthermore, it has been suggested that wood fibres have gel-like surfaces (Pelton, 1993; Lindstrom, 1989; Lindstrom, 1992) due to the presence of hemicellulose, which are highly hydrated and possibly charged (depending on the suspending solution) (Pelton, 1993). This seems likely, especially in the case of chemical pulp fibres where so little lignin is present. The above discussion is limited to the outside surface of wood fibres. The inside surface, which is probably similar in both mechanical and chemical pulp fibres, possesses little lignin even before pulping takes place (see Figure 2.3). Because the 14 inner surface is not exposed to the ravages of the pulping process to the same degree as the fibre's outside surface, it seems reasonable to assume that the inner surface is relatively smooth. However, no direct investigation of the characteristics of the lumen surface has been reported. Figure 2.5. Scanning electron micrographs of chemical and mechanical pulp fibres after the Papricourse '91, March 1991. In summary, the outside surfaces of both mechanical and chemical pulp fibres are likely covered non-uniformly by a steric barrier of hydrated, possibly charged, polymers that originate from the hemicellulose present in the fibres' primary walls. In addition to possessing more lignin and therefore being more electrosterically stabilized 15 than chemical pulp fibres, mechanical pulp fibres are considerably rougher due to unraveled fibrils present on the fibre surface. In contrast to the outside, the inside surfaces of both chemical and mechanical pulp fibres, while not completely devoid of charge and a steric barrier, are likely relatively smooth and free of electrolytic polymer. 2.3.3 Newsprint Newsprint is made by directing a uniform jet of a low consistency pulp (approximately 0.5% by weight pulp fibres in water) onto a moving fabric mesh on which a fibre mat is formed via filtration. Water drained from the pulp slurry passes through the fibre mat and is collected for recycle. Once de-watered, the fibre mat is pressed, dried, and calendered. Hydrogen bonds, formed between the cellulose portions of the fibres during drying, give the paper its strength. A conceptual sketch of the papermaking process is shown in Figure 2.6. Paper spool purge Figure 2.6. A sketch of the papermaking process. Paper made with mechanical pulp fibres displays different qualities from those possessed by paper made with chemical pulp fibres. For example, mechanical pulp 16 paper has good printing properties but has low tensile and tear strength compared to paper made with chemically pulped fibres. Newsprint manufacturers take advantage of these differences by combining the pulps to obtain the desired qualities in the paper they produce. It is for this reason that newspaper typically contains 80 to 85% mechanical pulp fibres and 15 to 20% chemical pulp fibres. 2.4 The Printing Process: Ink-Paper Interactions Printing involves the transfer and setting of ink. During the printing of conventional oil-based inks such as letterpress and offset on newsprint, ink transfer is typically described as a three-stage process involving: a) contact between plate and paper; b) immobilization of part of the ink by the paper; and c) splitting of the remaining ink film (Zang, 1993; Aspler, 1988). It is known that these conventional inks "set" largely by penetration of the oil vehicle into the paper pores. In comparison to oil-based newspaper inks, flexographic ink is up to 1000 times less viscous and dries rapidly (Aspler et al., 1993). Also, it is known that the nip pressure applied to the newspaper sheet in a flexographic press is low compared to that of a letterpress or offset system (Aspler, 1988). (This is due to the flexographic anilox cylinder system which allows uniform transfer of the low viscosity ink without resorting to high printing pressure.) Thus, one might expect the mechanisms of ink transfer and ink setting in flexography to be unlike those encountered in oil-base ink printing. However, no models of ink transfer and setting exist for water-based flexographic ink and the mechanisms by which these processes take place are not well understood. 17 In the absence of proven mechanisms, it has been assumed that flexo ink is transferred to paper in a manner similar to the three-stage model describing oil-based newspaper ink transfer, and that, while the amines in the ink vehicle may evaporate, the primary ink setting mechanism is absorption of the ink's water component into the pores and individual fibre walls of the paper substrate (Lyne, 1993; Wasilewski, 1987; Werther, 1991). It has further been assumed that the acidic nature of the paper enhances ink setting by lowering the pH of the ink film thus encouraging ink binder precipitation. The carbon black pigment particles to which binder molecules are adsorbed, then bond to the paper by entanglement of precipitated binder polymers in the paper matrix. Initial studies of flexographic ink transfer and setting have uncovered some interesting facts. Aspler (1988) has found that, while between 2 and 3 g/m2 of letterpress ink are typically printed onto the news sheet, only 1 g/m2 of flexographic ink is printed to obtain the same print density. It is not surprising then, that cross-sections of flexo-printed samples show that, unlike oil-based inks which penetrate into the paper, flexo ink remains on the paper surface. However, Gregersen et al. (1995) have contradicted this by showing that flexographic ink can penetrate beyond the surface layer of newsprint fibres. A second study (Aspler et al., 1993) showed that flexo ink transfer increases with increased water absorbency of the news sheet. Likewise, the more acidic the paper, the more ink is transferred. However, both these effects are small within the pH and absorbency ranges found in commercial newsprint. Although no formal mechanism has been offered, Aspler (1988) suggests that flexographic ink 18 transfer may be controlled by plate-to-paper contact in a similar fashion to oil-based ink transfer. The studies mentioned above also investigated flexographic ink drying. It was shown that, under normal conditions, the water absorbency of the paper does not affect ink drying (Aspler, 1988). The second study found that only when at least ten times the usual amount of ink was applied to the paper surface did water absorbency affect ink setting (Aspler et al., 1993). Interestingly, the second study also showed that ink setting was not affected by paper pH. This is in direct contradiction to the assumed mechanism which suggests that paper pH aids in ink binder precipitation. Aspler (1988) suggests that water evaporation in a thin ink film is the primary mechanism of flexo ink setting. A study by Borhan et al. (1996) of the drying mechanism of water-based ink jet printing on bond paper corroborates this idea. (In the case of ink jet printing, the final film thickness is about 12 um - six times greater than the final film thickness of flexographic news ink. For such a thick ink film, the authors point out that, first evaporation, then ink penetration are the drying mechanisms.) Presumably, the binder polymer concentration increases as the ink vehicle evaporates. The polymer sets the ink by precipitation once its concentration is suitably high. In summary, the mechanisms of flexographic ink transfer and setting are not well known. It has been assumed that flexo ink transfer is similar to the transfer of oil-based news inks and initial studies support this. However, these studies do not confirm the assumptions that the ink dries by absorption of the water into the paper substrate and that ink setting is enhanced by the acidic pH of the paper. Instead, it is suggested the 19 primary mechanism of flexographic ink drying and setting is related to evaporation of the ink vehicle and precipitation of ink binder molecules. 2.5 Newspaper Deinking During the deinking process, paper is defibred and ink particles are detached from the fibre substrate, dispersed in an aqueous medium, and then separated from the pulp fibre suspension. Several unit operations are employed in a deinking mill to accomplish these tasks. Ink detachment and dispersion take place in the repulping stage where old newspaper and lightweight coated stock (in an approximate ratio of 3 to 1) are repulped in an alkaline environment for roughly 12 minutes. Modern batch repulpers are generally run at 11 to 15 % consistency (where consistency refers to the mass percent of fibres in the pulp slurry) with a helical impeller. The repulper volume can range from roughly 10 to 70 m 3. Typical repulping chemicals and conditions are listed in Table 2.1. Sodium hydroxide is added to the repulper to increase fibre flexibility and also to encourage dissolution of the ink collector. Calcium ions, contributed to the repulping system by hard process water, react with the fatty acid collector to produce a calcium soap that precipitates on offset or letterpress ink particles. The particles then acquire sufficient hydrophobicity to agglomerate and adhere to air bubbles during the flotation step (Larsson, 1987). Hydrogen peroxide, sodium silicate, and a chelating agent are often added to brighten the otherwise alkali darkened mechanical pulp that, unlike chemical pulp, contains chromophores associated with the pulp's lignin component. While the peroxide reacts with pulp chromophores under high pH conditions, sodium silicate is 20 added as a buffer, and, along with the chelating agent, scavenges metal ions contributed by the process water that would otherwise degrade the peroxide (Ali et al., 1991). Once repulped, the repulper slurry is diluted to approximately 2.5 wt% with recycled plant water and transferred to a dump chest from which it is pumped, in a continuous fashion, through the plant's separation units. Large contaminants, such as dirt, sand, and plastic materials, are removed during screening and centrifuging stages; ink particles are separated by froth flotation and/or washing stages; and, finally, the pulp is brightened during a bleaching stage. Table 2.1. Repulping Chemicals and Conditions, after Ferguson (1991) CHEMICAL DOSE (based on the weight of oven dried pulp in the system) Chelant (typically DTPA) 0.15%-0.4% sodium silicate 1.0%-3.0% sodium hydroxide 0.8%-1.5% hydrogen peroxide 0.5% - 2.0% COllector(typically a fatty acid soap such as sodium stearate) 0.25%-1.0% CONDITIONS RANGE temperature 45 -55°C PH 9.5-10.2 consistency 5%-15% time 4 - 60 min. Froth flotation, a technology adapted from mineral processing, separates hydro-phobic particles of roughly 10 to 100 um in diameter from a diluted suspension of pulp and ink. Once the pulp stock has been diluted to approximately 1% consistency (i.e., 1 21 introduced under turbulent conditions in order to encourage collision between bubbles and hydrophobic ink particles. The ink-laden froth is removed and the cleansed stock is pumped to another cell where the flotation process is repeated. One flotation configuration is illustrated in Figure 2.7. % wMrre WATER OISCHARGE Figure 2.8. A drawing of the Black-Clawson double-nip thickener after supplier specifications (1990). 22 Pulp washing is a relatively simple step that involves stock thickening via vacuum filtration. A washer is shown in Figure 2.8. The stock enters the thickener through the headbox and is dewatered on the breast and couch rolls to approximately 15% consistency. Other industrial units involve series of disk vacuum filters that dewater and reslush stock several times. All unit operations in a mill are, of course, significant. However, repulping is the heart of the deinking process and the success of all subsequent ink separation stages depends on the degree of ink detachment accomplished during repulping. 2.6 Flexographically Printed Newspaper Deinking Several mills throughout Europe and North America have found that addition of newspaper printed with water-based flexographic ink makes deinking more difficult (Bast, 1990; Mah et al., 1993; Kubler, 1988; Clewley et al., 1990). Flexographic ink particles, detached in typical repulping conditions, are difficult to remove by a flotation process due to their small size and their fundamentally hydrophilic nature. While pulp washing is more successful at removing detached flexographic ink particles, such washing overloads the water clarification system, making it impossible to close the water loop. As deinking mills move towards closed-loop systems, finding solutions to the problems associated with flexographically printed newspaper stock is crucial to the survival of this otherwise benign printing technique. Indeed currently, some European mills will not accept any flexographically printed paper in their paper supply (Gottsching, 1994). 23 Chapter 3. Flexographic Newsprint Deinking Literature Review 3.1 Applied Studies Several laboratory investigations have tried to improve the deinking of flexographically printed newspapers (Jarrehult et al., 1989; Putz et al., 1989; Galland and Vernac, 1993a; Weide, 1987; Clewley et al., 1990; Wallstrom, 1988; Putz et al., 1989; Putz and Gottsching, 1989; Liphard et al., 1990; Heimburger, 1992; Harrison, 1995; Borchardt et al., 1994; Ellis et al., 1993). The experiments performed in the course of these studies combined the processes of ink detachment, deposition and ink separation by measuring the final pulp handsheet brightness of stock that was repulped and then floated and/or washed under different conditions. By showing that standard laboratory repulping and flotation procedures do not successfully deink pulp containing flexographic newspaper, these studies echo industrial experience (Mah et al., 1993). They also indicate that the brightness of both post-repulper and deinked pulp drop dramatically as the percentage of flexographic newspaper in the feed stock is increased. Surfactants have been developed that allow flotation of post-repulper stock containing from 10% to 30% flexographic newsprint (Jarrehult et al., 1989; Galland and Vernac, 1993a; Clewley et al., 1990; Heimburger, 1992; Harrison, 1995; Borchardt et al., 1994; Ellis et al., 1993; Horng, 1993). However, significant fibre and fines losses during flotation are associated with these systems (Horng, 1993). Both repulping and flotation performed under neutral rather than alkaline conditions produce a positive effect on handsheet brightness of both post-repulper stock 24 (Putz et al., 1989) and deinked pulp (Jarreult, 1989; Putz et al., 1989; Mah et al., 1993). Brightness remains constant below roughly 6.5 pH. It has also been found that deinking solutions containing C a + + similarly improve the brightness of post-flotation pulp (Putz et al., 1991). Galland and Vernac (1993a) have also found that acidic or neutral repulping followed by non-alkaline flotation increases pulp brightness. However, running under low pH conditions in commercial units makes paper defibreing difficult and also causes excessive fibre loss (Mah et al., 1993; Borchardt et al., 1994). Addition of a post-flotation wash step, particularly one in which no fibre mat is formed (Rangamannar et al., 1992), improves the brightness of laboratory repulped samples where no wash-water recycle loop is employed (Jarrehult et al., 1989; Galland and Vernac, 1993b; Clewley et al., 1990; Heimburger, 1992). However, because industrial recycle water clarification techniques are incapable of adequately removing small ink particles, flexographic ink present in the recycle loop may be able to deposit onto pulp fibres during repulping. Recently, ultrafiltration has been found to clarify flexographic-contaminated recycle water (Chabot et al., 1997; Upton et al., 1996, 1997a, 1997b). But addition of this unit operation undoubtedly adds cost to the deinking process and remains untried in the industrial setting. Galland and Vernac (1993a) reported that the brightness of infinitely washed pulp, made from commercial newsprint and repulped in laboratory equipment under typical deinking conditions, decreases as repulping time is increased. Ciampa (1995) reported similar results when repulping at pH 5. Ackermann et al. (1994) found that 25 repulping consistency has a negative effect on pulp brightness under unstated, but presumably alkaline, conditions. Several authors have illustrated that pulp brightness loss is due, at least in part, to the small size of the detached flexographic ink particles (less than 1 um) entering the flotation cell (Jarrehult, 1989; Putz et al., 1989; Putz and Gottsching, 1989; Clewley et al. 1990; Galland and Vernac, 1993b; Borchardt et al., 1994). And it is the small size of the ink particles that has led researchers to invoke the principles of classic colloid science to explain flexographic ink behaviour observed in these applied studies. The classic theory of Derjaguin, Landau, Verwey, and Overbeek (DLVO) models the interaction energy between two particulate surfaces suspended in an aqueous medium as a function of the distance between them (see Figure 3.1). It accounts for the contributions of attractive dispersion forces (van der Waal's forces) and repulsive Coulombic forces, to the total interaction energy. This theory, which assumes that the suspending medium is continuous, does not adequately explain the interaction between surfaces at short distances of separation (i.e., < 5 nm). At very small distances of separation, solvation forces arise due to the behaviour of the solvent (i.e., water) molecules in the confined space between the two surfaces. Such forces are known as either the hydration (repulsive) force, if the interacting surfaces are hydrophilic, and the hydrophobic (attractive) force, if the surfaces are hydrophobic (Hunter, p. 420, 1986; Israelachvili, pp. 276-286, 1991). Thus the DLVO theory can be modified to account for the solvation forces such that: V T = V d + V e + V s [3.1] 26 where Vj is the total interaction energy between the two spheres, is the attractive interaction energy due to dispersion forces, V e is the electrostatic force (either attractive or repulsive) and V s is the interaction energy due to the solvation forces. A colloidal dispersion is stable if the net repulsive force acting between the particles is large because this creates an energy barrier (see Figure 3.1) that cannot easily be overcome by Brownian or shear-engendered motion. Recently, Yoon and Ravishankar (1996a, 1996b) have shown that the attractive hydrophobic force can act over a relatively large interparticulate distance (i.e., approximately 24 nm). This long-range force begins to appear when the contact angles (measured through water) of the interacting surfaces are greater than 90°; that is, when the surfaces exhibit hydrophobic behaviour. The origin of this force has been the subject of considerable debate. These authors suggest that it may be attributed to the capillary force associated with the spontaneous nucleation of a vapour phase (known as cavitation) on or near the hydrophobic surfaces. As in the case of the solvation forces, the distance between surfaces over which dispersion and electrostatic forces act varies with substance and suspending medium. Van der Waal's forces are known to act over an interparticulate distance as large as 1 um (Hunter, 1993, p. 273). The electrostatic force acts over an intersurface distance that is dependent on the electric double layer of each surface, the thicknesses of which are negatively influenced by the electrolyte concentration in the suspending medium. Thus, in a 1:1 electrolyte at 25°C, the double layer thicknesses of a plane carrying a 27 uniform charge are: 30.4 nm in a 10"4 M solution, 9.6 nm in a 10"3 M solution, 3.0 nm in a^  10~2 M solution, and 0.96 nm in a 10"1 M solution (Everett, p. 44, 1988). D) i— <D C UJ c g o «_ d) 0 i i I i r Intersurface Distance Figure 3.1. The interaction energy due to the attractive van der Waal's (Vd) and repulsive electric double layer (Ve) interactions between two similar spheres as a function of the distance between them. Colloidal stability can occur by either electrostatic or polymeric stabilization, or by a combination of the electrostatic and the polymeric effects. Furthermore, Israelachvili (p. 280, 1991) provides evidence that the hydration force between colloidal particles can also influence their stability. Electrostatic and/or hydration stabilization occurs if the energy barrier predicted by the modified DLVO theory is sufficiently large (i.e., approximately 15 kT). Polymer stabilization is possible because polymers that are adsorbed onto the surface of either one or both of the interacting surfaces can create a 28 physical barrier that is of similar dimension to the range of the van der Waal's attraction force. This phenomenon is known as steric stabilization (Napper, p. 14, 1983; Hunter, p. 454, 1986; Israelachvili, pp. 293-298, 1991). The magnitude of the repulsive force resulting in the steric stabilization of colloidal particles is dependent on the quantity or coverage of polymer on each surface, on whether or not the polymer is physically adsorbed onto the surfaces or irreversibly grafted onto the surfaces, and on the ability of the solvent to dissolve and expand the polymer (i.e., on the quality of the solvent). Electrosteric stabilization occurs when a polyelectrolyte is adsorbed onto the surfaces of the particles. While colloidal forces exist for all surfaces, it is only when the masses of the interacting particles are sufficiently small that these forces may dominate over inertial effects. In addition to affecting interparticulate behaviour, colloidal forces also influence the interactions between dispersed particles and large surfaces. These interactions, which are frequently modeled as forces between small spheres and an infinite plane, describe, in an ideal way, the interactions between ink particles and fibre surfaces. For example, the positive effect of low pH deinking on post-flotation pulp brightness has been attributed to a decrease in surface potential of flexographic ink particles and consequent particle aggregation at low pH, thus allowing more successful ink/fibre separation (Hodgson, 1996). It has been assumed that addition of divalent calcium dampens the ink particle electric double layer thickness, which reduces the electrostatic repulsion and similarly allows particle aggregation. Furthermore, it has been suggested (Galland and Vernac, 1993a, and Ackermann et al., 1994, respectively) that the loss in 29 pulp brightness as repulping time and pulp consistency are increased is due to ink deposition. However, because the studies mentioned above have been aimed at solving pressing industrial difficulties, no clear distinction between possible ink detachment difficulties, problems with ink separation, and the possibility of detached flexographic ink deposition, have been established. 3.2 Model Systems Investigations of well-defined systems have been reported which aim at more concise identification of the problems and mechanisms associated with flexographic newspaper deinking. Liphard et al. (1990) report large negative zeta potentials of both ink (-55 mV) and carbon black (-40 mV) particles in an alkaline solution (pH 10). The authors suggest that the difference between the carbon black and ink particle zeta potentials is due to the anionic polyelectrolytic ink binder molecules (in this case, a co-polymer of styrene and acrylic acid) which adsorb onto the surfaces of carbon black particles, thus giving the ink particles a large, negative, surface charge in alkaline solutions. This surface charge, as well as a possible steric effect associated with adsorbed binder molecules, allows dispersion of flexographic ink in deinking solutions. Liphard et al. also show that the zeta potentials of both ink and carbon black particles decrease as either the pH of the suspending solution is decreased or the C a + + concentration is increased. The lower the pH or the higher the C a + + concentration, the less negative are the potentials of both the carbon black and ink particles, with the absolute value of the zeta potential being consistently greater for ink particles than for carbon black. Figure 3.2 illustrates the results relating to pH. The authors show that 30 decreasing ^-potentials correspond to decreases in pulp brightness. It is implied that flexographic ink behaviour reflects classic DLVO theory and the assumptions of the more applied studies are supported. That is, as the potential-determining hydroxyl ion concentration decreases, the ^-potential of the particles approaches zero (it is -10 mV at pH 5). Similarly, addition of multivalent cations at constant pH reduces the magnitude of the ^-potential. In this case, the potential's magnitude is reduced to approximately -20 mV, which is low enough to allow some particle aggregation. 6 10 11 PH Figure 3.2. The effect of pH on the zeta potential of carbon black particles and commercial flexographic ink after Liphard et al. (1990). Concentrations of ink and carbon black are 20 and 10 mg/L respectively. The carbon black used here is identical to that used in the commercial ink. A styrene acrylate copolymer is used as binder in the ink. Both the ink and carbon black suspensions were ultrasonified for 30 minutes before electrophoretic mobility measurements were taken. Dorris and Nguyen (1995) performed flotation experiments on well-defined flexographic ink suspensions in the absence of fibres at various pH levels, with and without calcium and oleate. They found that flexographic ink floats poorly in an alkaline solution where the particle size is below 0.8 um, but 80% is separated from a 31 suspension at pH 3 where the mean particle size is approximately 20 urn. In high pH solutions where calcium is present, 50% of the ink is separated by flotation. Almost complete flotation of ink is possible if both C a + + and sodium oleate are added to the suspending alkaline solution. It is interesting to note that these conditions are typical of those found in commercial flotation cells where flexographic ink separation has been notoriously difficult. Ciampa (1995) repulped blank newsprint at pH 5, 0.1% commercial surfactant (Neodol 45-7), 25°C, and 6% consistency in a bench-scale repulper at 2900 rpm for 20 minutes, added previously dried and ground commercial flexographic ink to the pulp (0.015 g/g paper), then continued repulping for another 30 minutes, taking samples at 5, 10, 15, and 30 minutes. These samples were diluted to 0.5% consistency then rinsed 15 times in a dynamic drainage jar (rpm = 2000) with a solution containing sodium carbonate and a dispersing agent. The brightness of the pulp samples before and after washing was measured. These tests were then repeated with a different flexographic newspaper ink, at a lower ink concentration (i.e., 0.005 g/g paper). The results, shown in Figure 3.3, indicate that brightness loss is greatest during the first 5 minutes of repulping but may continue up to 15 minutes after the dried ink particles are added to the system. Ciampa (1995) similarly repulped newspaper printed with flexographic ink via a method designed to simulate industrial printing. During these runs, the repulping time, the repulping temperature, and the surfactant concentration were varied, then the brightness of the washed pulp was measured. The results showed that, at low pH, 32 regardless of the repulping time and surfactant level, increasing the repulping temperature from 25 to 55°C decreased pulp pad brightness. O w CO c/> c SI g> DO Q. 3 70 60 50 40 30 20 10 10 15 20 25 Time (minutes) 30 • Ink 1 After Washing • Ink 2 After Washing A Ink 1 After Slushing • Ink 2 After Slushing Figure 3.3. Brightness of newsprint pulps that have been repulped then mixed (slushed) with previously dried ink (either ink 1 or ink 2) at 6% consistency in a solution of pH 5 for extended times then rinsed (washed) several times (Figure 27 of Ciampa's thesis). Chabot et al. (1995) added diluted commercial flexo ink to 1% consistency TMP suspensions (40 mg/L of carbon black pigment in the final suspension), stirred the mixtures at a rate of 1200 rpm for 5 minutes at 45°C with pH levels ranging from 7.5 to 10.5, a sodium oleate concentration of 100 mg/L and calcium ion concentrations of from 0 to 200 mg/L. The slurries were then washed in a dynamic drainage jar (DDJ) (rpm 1200) with deionized water whose pH was adjusted to that of the ink/fibre suspension. Finally, the rinsed pulp brightness was measured. The brightness results of these pulps were compared with those resulting from other, similar, tests which were performed over a wider pH range, without DDJ washing. The results from some of these tests, 33 illustrated in Figure 3.4, show that pH and calcium concentration affect pulp brightness via alkali darkening and the presence of calcium ions. (Note that the mechanism by cn « c 4-* s: u> m O (0 60 58 56 54 52 50 48 Calcium concentration 100 mg/L sodium olaata (mg/L) 0 X . " 100 X X X . . . ^ S X**" S X v "—With Ink 2 ^ S " > ^ ^ X ; - No Ink R = 0.92 "—Wtth Ink I . I . 9 pH 10 11 Figure 3.4. This figure illustrates results from a central composite experimental design which tested the effects of pH, ink addition, and calcium concentration on pulp brightness after Chabot et al. (1995). The functions shown here result from a regression model which fits the experimental data with an R 2 value or 0.92. which calcium affects pulp brightness is not well understood.) Flexo ink particles were almost completely removed from virgin pulp fibres during washing in the DDJ. Thus, at an rpm level of 1200 during 5 minutes of conditioning and subsequent washing, ink particles, not previously dried during a printing process, do not bond to pulp fibres — or do not bond with sufficient strength to withstand high shear forces. Ben et al. (1996) have conducted experiments that test commercial flexographic ink deposition during displacement washing of the long fibre fraction of a virgin Kraft (chemical) pulp. Degassed, 0.5% consistency pulp was first conditioned for 10 minutes with diluted ink and various chemicals at a stirring speed of 500 rpm. The mixture was then placed in a specially designed displacement washing cell and gradually thickened to a final consistency of 10% after which displacement washing, at a flow of 20 mL/min, 34 was implemented. The washed pulp pad was removed and brightness measurements were made. All runs were performed at a pH level of 8.5 and with varying concentrations of sodium oleate and calcium chloride. The results suggest that the presence of both excess positively charged calcium ions and calcium soaps adversely affect ink washing by thickening then solution displacement. Under these conditions, small ink aggregates (0.5 - 3 urn) become adsorbed onto the fibre surface — presumably by a colloidal deposition mechanism. These aggregates could not be detached from the pulp mat by application of a calcium-chelating chemical in the displacement fluid or by increasing the displacement fluid flow rate. The effect on the detachment of ink aggregates, of increased shear during pad formation, is not known. However, chelation of calcium with EDTA before mat formation was found to eliminate ink deposition. Fernandez and Hodgson (1996) characterised two model flexographic inks, composed of the same ink binder but different carbon black pigments, by measuring the ^-potential and particle size of both the carbon black pigments and the pigment/binder combination. The surface acidity of the carbon blacks, the pigment surface area available for binder adsorption, as well as an adsorption isotherm of binder onto the carbon blacks, were also determined. The Flexo "F" ink binder, obtained from Johnson Wax Inc., is a solution of acrylic polymer whose molecular weight is approximately 50,000 g/mol. Before combining the binder with the two different pigment types, the ammonium and monoethanolamine salts in the binder solution were removed. The Regal 99 and Black Pearls 420 (BP 420) carbon black pigments, both obtained from Cabot Corp., are furnace blacks which are typical of those found in many commercial 35 flexographic newsprint inks. The inks were made by first dispersing the pigments in water via ultrasonic bombardment, then mixing the dispersed particles in the polymer solution. The ^-potentials of the inks and their particle size distributions were compared with the same variables measured for particles in the filtrate from laboratory repulped and washed newspaper printed flexographically. By comparing the particle sizes of the pigments with and without binder (approximately 0.19 urn average diameter for both model inks compared to 0.47 \xm average diameter of BP 420 and 0.62 pm average diameter of Regal 99), Fernandez and Hodgson showed that the ink binders disperse otherwise agglomerated pigment particles. Furthermore, the authors confirm that flexographic ink dispersions are stabilized by an electrostatic effect that appears strongest around pH 8. These results, along with those of Liphard et al. (1990), prove that this effect is due to adsorption of the acrylic ink binder onto the carbon black pigment particles. Also, it appears that the initial pigment surface charge (in the absence of ink binder) does not affect the final ink particle charge over the entire pH range. Interestingly, the adsorption isotherm suggests that a monolayer of binder molecules does not exist on the pigment surfaces at the polymer concentration range typically found in water-based inks. Fernandez and Hodgson estimate that the smallest area occupied by a completely adsorbed monomer molecule of the acrylic binder is 0.3 nm 2. Since this estimate is larger than most of the estimates of the surface area available for adsorption per acrylic binder monomer (depending on initial polymer concentration, these values range from 0.12 to 0.29 nm2 for Regal 99 ink and from 0.13 to 0.47 nm2 for BP 420), it is likely that the binder 36 polymer extends away from the particle surface, creating a steric barrier around the pigment particles. Comparison of the size distribution and ^-potential data with the corresponding information for a laboratory wash-deinked filtrate suggests that these model inks can accurately simulate the commercial product. An investigation of the steric effect of binders on ink stabilization has been reported by Fernandez and Hodgson (1997). The authors continued their study of model inks by measuring ink turbidity and ink particle ^-potentials at different pH levels and electrolyte concentrations. A summary of their results is provided in Table 3.1. Here the critical coagulation concentration (ccc) is defined as the electrolyte concentration in a model ink sample where, after 20 seconds of shaking and 30 minutes of quiescence, the ink turbidity is 200 NTU. Turbidity measurements lower than 200 NTU indicate that the salt concentration has exceeded the ccc. The isoelectric point (iep) is simply the electrolyte concentration at which the ink particle zeta potential is zero. Although not stated explicitly, it is presumed that all measurements were taken at room temperature. Fernandez and Hodgson (1997) found that, at a pH level of 4, the ink particles follow the Schultz-Hardy rule for low potentials. However, at pH 10, a steric effect is evident. Complete coagulation of ink particles is observed at pH values of 3.2 and lower, at which point the zeta-potential is -20 mV. The authors conclude that, at this level, the reduction in a steric barrier, which is likely caused by polyelectrolyte precipitation, is sufficient to create instability. Results from tests performed at pH 10, summarized in Table 3.2, indicate that an increase in polyelectrolyte adsorption onto 37 pigment particles can occur as the salt concentration increases. This adsorption, which is likely due to charge screening by the monovalent electrolyte, may have a significantly positive effect on ink stability. Finally, Fernandez and Hodgson report that, at pH 10, with and without 0.4 M NaCI, ink stability is not affected by temperature changes over the range of 20 to 40°C. Table 3.1: The ccc and iep of Model Ink Particles at Low and High pH Levels (Fernandez and Hodgson, 1997) Salt [Sal t ] C C C p H 4 [Salt] c c c p H 1 0 [Salt] j e p p H 4 NaCI 2.910 mol/L — 0.2 mol/L CaCI2 0.816 mol/L 1.090 mol/L 0.02 mol/L FeCI3 0.148 mol/L 0.166 mol/L 0.0037 mol/L Table 3.2: Model Ink Particle Size and Polyelectrolyte Adsorption at Different NaCI Concentrations (Fernandez and Hodgson, 1997) [NaCI](mol/L) Average Particle Size (nm) Gram Polyelectrolyte /Gram Pigment 0.0 262 0.0651 0.2 282 0.0882 0.4 293 0.1170 3.3 Discussion and Summary No direct investigations of the flexographic ink detachment step appear in the literature and it is acknowledged that, in general, ink detachment mechanisms have not been studied in detail (Borchardt, 1993). Liphard et al. (1990) and Fernandez and 38 Hodgson (1996 and 1997) have shown that flexographic ink binder molecules adsorb onto and disperse minute (approximately 0.2 um dia.) carbon black pigment particles in alkaline solutions and it has been assumed that ink detachment in alkaline conditions occurs via conversion of the acrylate groups in the acrylic binder resins to carboxylic acid groups, making the otherwise solid binder resins dispersible and/or soluble (Borchardt, 1997). The resulting electrostatic charge of, and steric barrier around, each individual pigment particle (Fernandez and Hodgson, 1996 and 1997) enable dispersal of ink particles into the aqueous phase of a repulping slurry. Fernandez and Hodgson have shown that ink particles detached during alkaline repulping are in the size range of 0.2 p.m while results from a test by Borchardt et al. (1994) indicate a detached ink size of 0.02 urn. This order of magnitude discrepancy may be due to the different methods by which the size measurements were obtained. (Fernandez and Hodgson measured sedimentation rates with a centrifuge sedimentation device that converts the sedimentation rates of suspended particles to particle diameters via the Stokes equation. Borchardt et al. employed an image analysis technique using a scanning probe microscope to measure the size of dispersed flexographic ink particles after repulping.) In both cases, the results suggest that ink particles do indeed detach into a fine dispersion during alkaline repulping. Questions regarding the flexographic ink detachment step remain. While it is known that flexographic ink particles in the repulper suspension are well dispersed, it is not clear if these particles include both those previously attached to the fibres and those previously set in the ink layer above the ink/fibre interface, or just the particles of the 39 latter group. Also, during repulping, shear forces induced by imposed mixing disintegrate the newsprint sheet and cause individual pulp fibres to flex and bend. The effects of this pulp behavior, as well as the effect of shear on ink itself, are not understood. Furthermore, the mechanism by which flexographic ink detaches during non-alkaline repulping has not been studied. Likewise, the effect of the counterion concentration on ink detachment, as well as those of binder type and printing conditions, are not fully understood. The studies by Fernandez and Hodgson (1996, 1997) also raise questions about the nature of binder adsorption onto pigment particles. Is this adsorption reversible (i.e., is it chemi- or physi-sorption), and can binder molecules adsorb onto the pulp fibres? Ciampa's study (1995) indicates that repulping temperature over the range of 25 to 55°C only affects the washed pulp brightness under acidic repulping conditions. Ciampa attributed the negative temperature effect under low pH conditions to the surfactant cloud point which, in this case, was 44 °C. That no temperature effect under alkaline conditions was evident may be due to the stabilizing effect of the ink binder. Fernandez and Hodgson (1997) have found that model ink stability at pH 10 is not affected by temperatures ranging from 20 to 40 °C. Further study of temperature's effect on the flexographic ink detachment mechanism might prove fruitful. Ciampa (1995) has conclusively shown that flexographic ink deposition occurs during low pH repulping. Because the author's results from tests of different inks and different ink concentrations are not similar, Ciampa's study also indicates that ink concentration, composition and/or pigment size distribution may influence the extent of 40 deposition during repulping. The author suggests that the drop in pulp brightness with repulping time also is due to ink deposition. However, the effect of low pH repulping time on ink particle size is not known. Although it is not clear whether the drop in brightness with repulping time is due to ink particle comminution or to increased ink deposition, the latter explanation is more likely because the dried ink was ground to a fine powder with mortar and pestle before it was added to the pulp. Galland and Vernac (1993a) suggest that deposition occurs during alkaline repulping, however the results of Chabot et al. (1995) suggest that deposition during a relatively short contact period - even in the presence of calcium ions — may be insignificant. Thus experimental results are contradictory and the conditions under which ink deposition occurs during high pH repulping are not well understood. The influence of shear on ink deposition during repulping is not known, and the role of indifferent electrolytes during high pH repulping is not well understood. In the absence of salt from the ink binder, Fernandez and Hodgson (1997) have measured the critical coagulation concentration (ccc) of CaCI 2 at different pH values. That measured at pH 10 (1.1 M CaC^) is considerably higher than the calcium ion concentration used by Chabot et al. (0.005 M Ca + + ) in a pH 8.5 solution. However, Chabot et al. saw evidence of pigment coagulation at this low calcium ion concentration, as did Ben et al. (1996) during displacement washing experiments where the C a + + concentration was 0.013 M. (Tests by Chabot et al. showed that the brightness of pads made from pulp mixed in an alkaline solution of 0.005 M (200 mg/L) calcium ion with 40 mg/L of never-dried flexographic ink is approximately 30 points lower than similarly made pads of pulp 41 mixed in the absence of calcium.) These observations may be explained by the salt concentration contributed by the never-dried ink. However, a calculation based on data provided in Chabot et al.'s article indicates that the monovalent counterion concentration contributed by the binder in their experiments was much less than that necessary for coagulation of the carbon black particles (i.e., only about 0.16 mM — much smaller than what might be considered typical values provided by Fernandez and Hodgson (1996) and shown in Table 3.1). Ben et al. (1996) used an ink which was similar to that of Chabot et al. but at a much lower concentration. Further study of both the effects of shear and electrolyte concentration on ink deposition would be useful. The mechanism of ink deposition during repulping is not known. Penetration of ink particles into fibre pores is alluded to by some authors (Ciampa, 1995; Gottsching and Putz, 1994; Oye et al., 1991), however no evidence supporting this hypothesis is in the flexographic deinking literature. Ben et al. (1996) suggest that the mechanism by which particles adsorb onto fibres during displacement washing with an alkaline solution containing calcium soaps and calcium ions may be colloidal because disiodgment of the particles from fibres does not occur when fluid flow around the fibres is increased. This hypothesis supports the suggestion by Nguyen and Dorris (1995) that flexographic ink may not be separated during flotation because of ink deposition onto fibres in an alkaline environment containing calcium ions and sodium oleate. However, Chabot et al. (1995), who have shown that flexographic ink can be washed from fibres suspended in an identical medium, suggest that deposition is due to ink entrapment in the fibre web. 42 Knowledge of flexographic ink deposition kinetics is limited. Ciampa's study (1995) — the only published information on the topic — indicates that most ink deposition occurs during the initial stages of repulping and reaches a plateau after approximately 15 minutes under the experimental conditions tested. It is not clear whether this is a dynamic equilibrium following the Langmuir model where the number of ink particles depositing onto fibres equals the number being detached, or a saturation point when all possible adhering surfaces are covered with ink. 43 Chapter 4. Literature Review of Other Relevant Subject Areas 4.1 Introduction Although some studies of flexographic newsprint deinking systems have been published in the last ten years, many questions regarding the deinking of flexographic ink remain unanswered. Additional insight about the flexographic deinking process may be gained by reviewing studies of similar systems. Thus the following chapter provides summaries of the laundering (i.e., soil removal from textile fabric) literature and investigations into filler pigment retention during papermaking. Both laundering and filler retention systems are analogous to the flexographic deinking system. Each consists of fine particles and fibrous material suspended in an aqueous environment on which agitation is imposed. Laundering is particularly similar to deinking because, like the deinking system, its objective is the detachment and separation of fine particles from a fibrous substrate via agitation in an aqueous environment. Although the objective of filler retention is to deposit (rather than remove) fine particles on wood pulp just prior to papermaking, this system is nevertheless similar to that of deinking because both involve interactions between similar materials (i.e., colloidal particles and wood pulp fibres) in a sheared suspension. Of particular interest in the filler retention literature is the subject of lumen loading of fibres with filler particles. A review of this subject may shed light on the mechanism of flexographic ink deposition proposed by some authors (i.e., Ciampa, 1995; Gottsching and Putz, 1994). Thus the following two sections review relevant aspects found in both the laundering and the filler retention literature. 44 4.2 Laundering Systems (Carroll, 1993; Kissa, 1987; Cahn and Lynn, 1983) Like deinking, laundering is complex and difficult to study. Theoretical investigations are therefore limited to simplified systems that attempt to model specific concepts of the laundering process. These studies are divided into investigations of liquid soil systems, and solid or particulate soil systems. Liquid soils are necessarily non-volatile and typically non-aqueous. Their shape can change dramatically on deposition - the final form being dictated by capillary action and textile shape. Particulate soil exists in many shapes and can range in size from approximately 0.1 um to 5 um in equivalent diameter. The shape and size of these soils remain constant throughout the laundering process (unless the process temperature is sufficiently high to melt the solid). Thus, although exhibiting similar colloidal properties, liquid and solid soils also possess some markedly different properties. Borchardt (1994) has performed an environmental scanning electron microscopy (ESEM) investigation on newspaper inks. (The ESEM is a scanning electron microscope (SEM) that can operate at the low pressure of a typical SEM through to the pressure required to observe liquid distilled water (609 Pa at 0 °C).) He suggests that both oil- and water-based inks are analogous to liquid soils because the study indicates that these liquid-based inks exist as films on the paper surface. But, unlike oil-based newsprint inks, whose oil component tends to absorb into the paper substrate (hence the experience of oil-based ink "rub-off1 while reading one's daily paper), dry flexographic ink is a flexible solid polymer in which the pigment particles are embedded. Thus, dried water-based flexographic newsprint ink is more properly analogous to solid 45 soils. An understanding of the mechanisms involved in solid soil laundering may help illuminate some of those involved in deinking flexographically printed newspaper. What attractive forces are broken during the detachment of soil particles from a fabric substrate? This fundamental laundering question requires an understanding of soil deposition mechanisms as well as those of soil removal; hence studies of both soil deposition and detachment are found in the literature. Whether or not it is possible to deposit detached soil is a separate but related issue that is also addressed in the laundering literature. The following sections summarise conclusions from investigations of all three phenomena. 4.2.1 Particulate Soiling Soiling occurs by either direct deposition of particulates onto the textile, transfer of particulates from a soiled surface to a clean surface, or deposition by electrostatic attraction (Kissa, 1987b). Once deposited, it is believed that solid soil attaches to the textile via van der Waals and/or Coulombic interaction (Kissa, 1987b; Cahn and Lynn, 1983; Schwartz, 1972), although occasionally short-range forces such as hydrogen bonding may play a role (Schwartz, 1972). Earlier researchers believed microocclusion of fine particles in the irregularities of natural fibres was the primary mechanism of particulate soiling. Although microphotographs of particulates clinging to smooth (synthetic) fibre surfaces, as well as other data, have discounted this idea (Cahn and Lynn, 1983), surface roughness does apparently encourage attachment and will be discussed presently. 46 4.2.2 Solid Soil Removal The removal of solid soil from the surface of a textile fibre occurs in three distinct stages (Carroll, 1993; Kissa, 1987b; Schwartz, 1972): 1. displacement of air by the detersive bath; 2. detachment of the soil from the substrate, which normally involves breaking the adhesive bond between the particulate and the fibre surface; and 3. transport of the soil particle into the bulk bath solution. All three stages in the soil removal process can be described in terms of classic colloid science theory where a soil particle attached to a textile fibre can be appropriately treated as a sphere attached to an infinite plane. Because the interacting surfaces are made of different materials and possess different geometries, the intersurface colloidal forces may be either attractive or negative. During the first stage of soil removal, in which air is displaced by a surfactant solution, the interaction energy between the soil particle and the textile fibre is altered. The change is due to both the increased dielectric constant of the dispersing medium and the creation of the electric double layer. Soil detachment, the second and most critical soil removal stage, happens only if the interaction energy well of either the primary or secondary minimum is overcome (see Figure 3.1). This can occur by several mechanisms that will be discussed presently. The third stage of the soil removal process involves transport of the soil into the bulk bath solution, that is, transport beyond the distance over which the interaction energy acts. 47 Factors that affect soil removal have been identified as solid soil size and the shape and surface roughness of both soil and substrate, mechanical energy put into the textile/particulate/bath system, which can also be described in terms of the shear rate at the textile surface, bath temperature, and the presence of an adsorbed layer on the surfaces in question (Carroll, 1993; Kissa, 1987b; Cahn and Lynn, 1983; Schwartz, 1972). Mechanisms by which these variables affect soil removal have been speculated on in the literature. The attachment energy between a soil particle and a fabric surface depends on the contact area between soil and fibre (Kissa, 1987b, p. 244). The contact area, in turn, is dependent on particle size and shape, as well as the surface roughness of both the textile fibre and the solid soil. During detachment, the force required to overcome attachment depends on either the ratio of contact area to particle area not in contact with the textile, or the ratio of contact area to the particle mass (Schwartz, 1972, p. 209). The latter condition is true because the magnitude of the inertial forces needed to detach a particle is proportional to the particle's mass. Similarly, non-inertial forces related to hydrodynamic drag can break bonds by exerting traction over the particle area not in contact with the textile, and are, therefore, proportional to that area. In practice, it is typically found that particles greater than 5 u.m in equivalent diameter are easily removed from fabric. Particles less than 0.01 um are virtually impossible to remove and cause gradual, irreversible, graying of white fabric with repeated laundering (Cahn and Lynn, 1983, p.393). 48 Surface roughness of both textile and particle will influence the contact area between the two surfaces but it will also affect wetting of the soiled fabric during immersion in the bath as well as wetting of the contact area surfaces subsequent to soil detachment. These effects are not well understood. In practice, soaking the soiled fabric before washing usually aids in soil removal. This may be due to increased wetting of a fabric that would otherwise remain at least partially dry during laundering. Related to surface roughness is the geometry of the soiled fabric; i.e., whether a fabric is woven or knitted. Like surface roughness, fabric geometry also affects soil removal. This is illustrated in one study that compared removal of an ideal soil (carbon black) from cellulose and polyester in film and fabric form (Schwartz, 1972, p. 237). The carbon was more difficult to remove from cotton fabric than from polyester fabric, but easier to remove from cellophane than from polyester film. Mechanical energy increases the rate and extent of solid soil removal (Schwartz, 1972; Kissa, 1987a) by affecting each stage of the soil removal process. During the first stage, agitation helps to liberate trapped air pockets from the fabric. During the second soil removal stage, shear force applied to the soil particle is often a key step in overcoming the forces of attachment. Furthermore, the form in which the energy is applied is crucial to efficient soil detachment. For example, ultrasonic action, boiling, and even vigorous flexing of the material are more effective methods of cleaning a soiled fabric than hydraulic agitation (Cahn and Lynn, 1983, p. 404; Schwartz, 1972, p. 233). This is due to increased shear on the soil particles in the first three cases compared to that in the fourth. And finally, during the third stage of soil removal, 49 mechanical energy enhances mixing which ensures transport of the detached soil particle from the adhesion site. Laundering temperatures typically range from approximately 10 °C to 50 °C. Increasing bath temperature increases the kinetic energy of components in the detersive system. This weakens the bonds between soil and substrate thus enhancing the probability of soil detachment. Indeed it has been suggested that Brownian movement, which increases in proportion to increases in temperature, may be as important as hydrodynamic action in detaching fine particulate soil from the substrate (Schwartz, 1972, p. 231). An increase in kinetic energy also increases the rate of diffusion mass transport of detached particles from the vicinity of the textile surface. In addition, temperature influences the behaviour of bath components. For example, temperature affects surfactant solubility and adsorption rate (Kissa, 1987b, p. 320). Finally, temperature affects the swelling rate of textile fibres. Fibre swelling is beneficial because it improves soil detachment by hydrating and expanding the fibre surface and also allowing diffusion of bath components into the fibre interior where the components can easily diffuse to the fibre-soil interface. Laundering baths invariably contain surface active agents which migrate to the bath/soil, and/or bath/fibre interface and adsorb onto the solid surface thus lowering the surface energy of that surface. This decreases the force needed to remove the soil particle (Kissa, 1987b, p. 244; Schwartz, 1972, p. 239). But, although surfactant spreading pressure and ionic repulsion between surfactant molecules adsorbed onto 50 the soil and textile fibre aid in soil detachment, an external (i.e., mechanical) force is usually required to overcome the particle/fibre attachment. 4.2.3 Solid Soil Deposition It is well known that soil removed from a fabric substrate can deposit onto the fabric. And, while an equilibrium can be obtained between soil attached to the textile and soil dispersed in the bath if laundering continues for several hours (Kissa, 1987b, p. 283), the equilibrium is not what it might first appear and is not directly related to the kinetics of the soil removal step. When a soiled fabric is immersed and agitated in a detersive bath, the fabric and desorbed soil undergo irreversible changes involving metastable states. This is largely due to a change in the continuous medium (from air to aqueous solution) and the adsorption of surfactant molecules onto the surfaces in the system (for example, the adsorption of surfactant onto the textile surface vacated by a detached soil particle). Thus, soil detachment is not reversible. How is it that removed soil can be deposited onto the textile if soil detachment is irreversible? This incongruity can be explained by investigating the nature of solid soils. A soil particle is typically an agglomerate of either dissimilar or similar small particles. Only a fraction of the particles in the agglomerate contact and attach to the textile substrate. During laundering the bond between these particles and the substrate may break, releasing the entire agglomerate. Alternatively, the bonds between the agglomerate's constituent particles may break, dispersing the small particles into the bath. Figure 4.1 illustrates these two mechanisms. 51 Dispersed particles and agglomerates are free to collide with adhesive sites on the textile fibre and soiling, as previously discussed, may occur. Whether or not such a particle becomes attached to the textile during laundering depends on the interfacial energies that prevail, the energy and/or steric barriers against close approach and the kinetic energy and directional velocity with which the particle and textile fibre collide (Schwartz, 1972, p. 226). a) Fbre b) Fbre Fbre Fbre Figure 4.1. a) An agglomerated particle is dispersed into constituent parts. b) An agglomerated particle is detached from the textile substrate. The deposition behaviour of detached soil is considered to be a kinetic process because the laundering time (typically about 20 minutes) precludes attainment of equilibrium. It is dependent on several factors, including the laundering conditions (e.g., temperature, agitation time and intensity), soil and fabric characteristics, and bath components. A bath component of particular interest is the anti-deposition agent. It is 52 known that adsorption of polyelectrolyte molecules onto the textile surface can inhibit deposition of detached soil particles. For example, sodium carboxymethylcellulose (NaCMC), adsorbed by hydrogen bonding onto cotton fibres, decreases the deposition of carbon black particles by electrostatic repulsion (Greminger et al., 1978; Kissa, 1987b, p. 296). In fact, it has generally been found that hydrophilic fabric finishes with carboxylic groups reduce the deposition of negatively-charged carbon in the same way that NaCMC reduces deposition. 4.3 Filler Retention Systems Although the goal of filler addition to a paper slurry prior to paper formation is the opposite of the deinking objective (i.e., fine particle deposition onto, rather than detachment from, fibres), knowledge of filler-fibre interactions may help to illuminate ink detachment and deposition mechanisms occurring in a deinking mill's repulping unit. Therefore, a review of the thermodynamic and kinetic aspects of filler-fibre colloidal interactions, both without and within the fibre, is given. Fillers are pigments added to pulp stock during production of writing grade papers in order to lower production costs or, more likely, to improve paper properties such as opacity and brightness. While clay and talc are examples of fillers used in the former application, specialty fillers such as titanium dioxide, alumino-silicates and calcined clay attain the goals of the latter (Brown, 1996; Scott, Chpt. 10, 1996). The most important properties of fillers are their size, which typically range from 0.1 to 5 um, their shape, specific gravity, and surface chemistry. All filler particles disperse in water and most are negatively-charged. 53 Generally, the manner of addition and point at which fillers are added to paper stock varies from papermachine to papermachine. However, introduction of a mechanically dispersed pigment slurry of approximately 50 wt% concentration typically takes place just before the pulp enters the headbox (refer to Chapter 2, section 3.3 for a brief description of the papermaking operation) where pulp consistencies can range from 0.5 to 1.5 % (Biermann, 1996) and the shear rate to which fibres are exposed is 3 4 - 1 around 10 to 10 s (van de Ven and Mason, 1981). The conventional concentration of filler in the headbox is approximately 0.1 to 0.6 g per gram dry pulp (i.e., approximately 1 to 6 grams filler/L) and the final mass percent filler in the sheet is between 5 and 30 % (Brown, 1996). The residence time of fibres in the papermachine wet-end is in the order of several seconds; the sheet formation time on the wire is typically less than 10 seconds. Fillers are retained during sheet formation by either entrapment during drainage on the papermachine wire or via colloidal deposition (i.e, hetero-flocculation) onto individual fibres in the papermachine wet-end (Scott, Chpt. 10, 1996; Jaycock and Swales, 1994). Well-dispersed filler particles are generally too small to become trapped in the sheet during formation (Alince et al., 1991), hence successful retention by entrapment occurs when pigment particles are able to coagulate in the papermachine headbox. Thus, both entrapment and deposition mechanisms are dependent on colloidal interactions in the headbox. 54 4.3.1 Thermodynamics (Lindstrom, 1989; Jaycock and Swales, 1994) Thermodynamic interactions between filler particles and pulp fibres in the papermachine headbox have traditionally been explained by classic DLVO and steric stabilisation theory. In the absence of electrostatic charge, interactions between closely approaching filler particles and fibres are due to attractive van der Waals forces. In a polar medium, particle surface charge affects interactions by either enhancing attraction if the interacting particles are oppositely charged, or by encouraging repulsion if the particles are of like charge. Since both fillers and fibres almost always have negative surface charges (Brown, 1996; Alince et al., 1991), filler deposition can only occur via dampening the double layer thickness or by charge reversal. The number and strength of filler-fibre bonds on both the inside and outside fibre surfaces are affected by pulp type, degree of pulp refining (beating), pulp bleaching, solution pH, electrolyte concentration and polyelectrolyte concentration, all of which affect the surface chemistry of the fillers and/or fibres. Adsorption of cationic polyelectrolytes on filler surfaces, such as polyacrylamine or polethylenimine, results in either charge reversal and subsequent attachment via Coulombic attraction, or polymer bridging when loops and tails of the adsorbed polymer extend into solution and adsorb onto fibres. Increasing the electrolyte concentration above that of critical deposition dampens the electrostatic double layer thickness surrounding both fillers and fibres, suppressing repulsive electrostatic forces and allowing filler deposition if close approach of filler particles to fibre surfaces can be 55 achieved. Such dampening of particulate double layer thickness occurs when alum (aluminum sulfate) is introduced into the papermachine wet-end. Solution pH alters the surface charge on fibres because hydrogen ions adsorb onto carboxyl surface groups. Pulp bleaching and pulp type affect the lignin content in pulp fibres (see Chapter 2, section 2 for a description of fibre characteristics). The more lignin present in the fibre, the greater is the fibre's surface potential, and the more likely oppositely charged filler particles will be retained on the fibre surface. Similarly, by encouraging fibre fibrillation, refining pulp fibres increases fibre specific surface area, thus improving filler deposition efficiency by increasing the available surface area. 4.3.2 Kinetics A short residence time in the papermachine wet-end dictates that kinetic considerations, and more specifically, perikinetic (due to Brownian motion) and orthokinetic (caused by flow) collision rates, as well as the shear rate to which component surfaces are exposed, have a large influence on filler and fibre behaviour in the system. Model systems of latex particles and nylon fibres have been shown to generally follow modified Langmuir kinetics (Boughey et al., 1978), and Middleton and Scallan (1991) have shown that the kinetics of submicrometric filler particle deposition in a 0.5% pulp fibre suspension subject to shear can also be modeled by competing capture and escape mechanisms. Kamiti and van de Ven (1994) have employed the following similarly modified Langmuir equation to describe deposition and detachment of a dilute suspension of calcium carbonate particles onto pulp fibres: 56 de/dt = ki(c0-e)(i-e)-k29 [4.1] where 8 = (amount of particles actually deposited on fibres (g filler/g pulp))/(maximum amount of fillers than can deposit (g filler/g pulp)), C 0 = (initial amount of particles in suspension (g filler/g pulp))/(maximum amount of particles that can deposit (g filler/g pulp)), and k-i and k2 are the rate constants for deposition and detachment respectively. Here the initial filler concentration (expressed as grams of filler per grams of pulp fibre) is assumed to be not much greater than the maximum amount of fillers that can deposit on the pulp fibres (also expressed as g filler per g fibre). Both k-i and k2 are strong functions of the collision rates between fibres and fillers. These collisions are primarily influenced by the turbulent hydrodynamic behaviour of the headbox slurry where approximately 10% of all pulp fibres experience maximum shear stresses on their walls of approximately 102 to 104 Pa in a machine running at 1000 m/min (Tarn Doo et al., 1984). The rate of collisions between fillers and fibres due to hydrodynamic forces can be estimated using Smoluchowski's equation (van de Ven, 1989) which models collisions between spheres as a function of shear rate, particle concentration, the size ratio of the approaching particle to the reference particle (which is approximated as zero in the case of fillers and fibres), and the radius of the reference particle. The fraction of collisions that result in filler deposition depends on the ratio of colloidal to hydrodynamic forces in the system. It has been shown theoretically that the efficiency with which fillers deposit onto fibres goes through a maximum as flow intensity is increased if no repulsive electrostatic force is present (Petlicki and van de Ven, 1992). When the flow 57 intensity is low, the number of particles depositing on fibres is determined by Brownian diffusion. As flow intensity increases, collisions are orthokinetic as well as perikinetic. But when the flow intensity is very high, filler particles simply orbit around the fibres, close approach takes place solely via Brownian movement, and once again, deposition is infrequent. Indeed, Kamiti and van de Ven (1994) have shown that, for 0.8 um diameter CaC0 3 particles (which possess a positive surface charge at the pH investigated (i.e., pH = 9.5)), 0 decreases as mixing increases from 80 rpm (estimated to be a shear rate at the fibre surfaces of 100 s"1) to 265 rpm. If particles are sufficiently small (i.e. <0.1 um) then, regardless of hydrodynamics, close approach to the fibre surface occurs via Brownian motion. In the absence of polyelectrolytes, the theory predicts that fillers cannot deposit onto fibres because of electrostatic repulsion between the two surfaces. This has been recorded experimentally by Middleton and Scallan (1991) in the case of 0.3 um diameter, negatively-charged, Ti0 2. When electrolytes are present in the aqueous phase of the headbox slurry - and, because of salts in wood pulp, they are invariably present - the electric double layer between approaching particles may be dampened, in which case deposition becomes possible via perikinetic and/or orthokinetic collisions as discussed above. This occurs above the electrolyte's critical deposition concentration where the energy barrier predicted by the DLVO theory disappears or appears within the negative range of interaction energies. (An electrolyte's critical deposition concentration is usually smaller than its critical coagulation concentration, i.e., the salt concentration where like particles will coagulate. This phenomenon leads to the possibility that filler 58 deposition onto fibres can occur in the absence of filler coagulation if the electrolyte concentration is greater than the concentration of critical deposition but less than that of critical coagulation.) The critical deposition concentration is proportional to the inverse of the counterion valency to the sixth power (this is the Schultz-Hardy rule) and Boluk and van de Ven (1990) have witnessed this dependency during deposition tests of titanium dioxide on a glass surface. The rate of detachment of deposited fillers is a function of the filler-fibre bond strength, the deposited particle's thermal motion, the magnitude and direction of hydrodynamic forces (Kamiti and van de Ven, 1994), and the concentration of fillers and fibres in the headbox (Middleton and Scallan, 1991). Bond strength is, of course, directly related to the magnitude of the attractive forces existing between the filler and fibre, which in turn, depends of the presence or absence of surface charge modifiers such as electrolytes and polyelectrolytes. In the absence of polymers, a deposited submicrometre particle is usually caught in the deep primary energy well predicted by the DLVO theory. Also, its small size dictates that dislodgment caused by hydrodynamic forces is negligible (the hydrodynamic force is proportional to the product of the squared diameter of the particle and the shear rate (van de Ven, 1989)). Thus, it is well known that such a particle is unlikely to detach readily. Indeed, van de Ven was unable to detach colloidal Ti0 2 particles from a cellophane sheet subjected to a wall shear rate of 103 s"1. The detachment rate of such particles depends strongly on the shape and depth of this primary well. Interestingly, both Kamiti and van de Ven (1994) and Middleton and 59 Scallan (1991) were able to detach particles ranging in size from 0.2 to 0.8 um, from pulp fibres by increasing the stirring rate of pulp suspensions. 4.3.3 Lumen Loading The disadvantage of adding fillers to paper is that particles tend to obstruct fibre-fibre bonding and thereby substantially reduce paper strength (Miller and Paliwal, 1985). This can be overcome by deposition of fillers inside the hollow core of pulp fibres. That is, by depositing pigment on the lumen surface rather than on the fibres' external surfaces, optical properties can be improved without compromising the sheet's physical attributes. For lumen loading to occur, two criteria must be met; filler particles must be small enough to travel through the fibre's pit pores and the slurry chemistry must favour filler adherence onto the lumen surface (Scallan and Middleton, 1985). Although several factors, which will be discussed below, can affect the extent of lumen loading, it is thought that, under ideal conditions, complete monolayer coverage of the lumen surface is possible. Under favourable thermodynamic conditions, it is known that lumen loading of up to approximately 15 wt% (i.e., the ratio of filler mass to the mass of paper product times 100) of certain types of pulp (Middleton and Scallan, 1985; Miller and Paliwal, 1985) can occur. (Of course the weight percent of filler in the final paper product depends, not only on filler concentration in the paper, but also on the filler density. The number quoted here was obtained using titanium dioxide filler with a specific gravity of 4.2. For comparison, the specific gravity of clay is only 2.5; thus, for the same filler volume in the final product, the weight of clay in the product would be only 8.9 %.) 60 Many factors affect the extent of lumen loading of fillers in pulps. Several researchers (Green et al., 1982; Middleton and Scallan, 1985; Petlicki and van de Ven, 1994; Middleton and Scallan, 1993; Okayama et al., 1989) have found that the degree of turbulence, length of agitation time, filler concentration and pulp consistency as well as pigment particle size can affect the kinetics of lumen loading. The number and type of filler-fibre bonds are affected by pulp type (Scallan and Middleton, 1985; Okayama et al., 1989), whether or not the pulp to be loaded has been previously dried (Middleton and Scallan, 1993), beaten (Miller and Paliwal, 1985; Middleton and Scallan, 1993), or bleached (Scallan and Middleton, 1985), the electrolyte concentrations (eg. NaCl or aluminum sulfate (alum)) (Middleton and Scallan, 1985), the pH of solution, and whether or not ionic polymers have been adsorbed onto the fillers and fibres prior to lumen loading. As in the case of fillers depositing on the external surfaces of fibres, classic DLVO theory is invoked to explain bonding of filler particles to fibre lumens (Middleton and Scallan, 1985). By employing equations from the literature that estimate the contribution of these forces to the interaction energy, Middleton and Scallan were able to estimate the total interaction energy between a Ti02 particle and the lumen surface at different pH values. Their plot of these approximations is shown in Figure 4.2. Middleton and Scallan (1985) note that, once a particle has found sanctuary in the fibre lumen, it is no longer subject to the shear experienced by particles on the outsides of the fibre walls. Thus the magnitude of attraction required for adsorption of fillers on the internal surface of a fibre is less than that needed for permanent 61 attachment of fillers on the fibre's external surface. Indeed, the authors (1993) have found that, not only is the detachment rate of lumen loaded fillers undisturbed by increased turbulence, but filler deposition in the lumen increases as the shear rate is increased. It is speculated that an increase in bulk concentration of filler, which occurs as adsorbed particles are detached from the external surfaces of fibres when the shear is increased, causes an increased concentration gradient across the fibre wall and subsequently greater filler movement into the fibre's interior. Figure 4.2. A plot of the total interaction energy between a titanium dioxide and fibre surface as a function of the inter-surface distance at different pH levels (Middleton and Scallan, 1985). Researchers have noted that never-dried pulps can be lumen loaded to a greater degree than pulps that have been previously dried. This is explained by the fact that pulps that have been dried and rewetted contain some collapsed fibres whose lumens 62 are inaccessible to the filler particles (Middleton and Scallan, 1993). Also, that beaten pulps can be lumen loaded to a greater degree than never-beaten pulps is explained by the suggestion that beating will remove pit membranes which otherwise obstruct filler entry into the fibre lumen (Miller and Paliwal, 1985). The observations noted above have resulted in the development of hypotheses which explain lumen loading behaviour. Petlicki and van de Ven (1994) suggest that lumen loading kinetics is exclusively dependent on particle diffusion that proceeds in three distinct stages. The first is characterised by what is called an advancing diffusion front that describes the increasing concentration of pigment inside the fibre lumen occurring within the first minute of mixing. During this period, particle deposition onto the lumen surface is negligible. Most deposition takes place in the second stage when the number of pigment particles entering the lumen equals the number being deposited on the internal wall. During the second stage, the concentration of particles inside the fibre remains almost constant, at about one half the level found in the bulk phase outside the fibre. The rate determining step of the third stage is no longer the rate at which particles can enter the lumen but becomes that at which particle deposition occurs on the remaining empty sites on the lumen wall. Like Middleton and Scallan (1993), Petlicki and van de Ven explain the dependence of lumen loading on the degree of turbulence by suggesting that increasing the shear during agitation increases the rate at which particles escape from the external fibre surface thus increasing the bulk pigment concentration which, in turn, results in a larger particle concentration gradient between the outside and inside of the fibre. 63 Of the three hypothesised stages, clearly the most influential is the second, which Petlicki and van de Ven model as equation [4.1] except that ki, the deposition rate constant, is replaced by a rate constant that is a function of both the external and internal surface deposition time constants. It has been shown that the external surface deposition time constant can be estimated as an inverse function of the deposition efficiency, the number of fibres per unit volume and Smoluchowski's collision rate constant (van de Ven, 1993). Petlicki and van de Ven approximate the lumen loading time constant as a function of the flux of particles entering the lumens which, in turn, is estimated by considering both diffusion and convection mechanisms. Significantly, the authors show that pumping action caused by fibre flexing does not contribute to the particle flux into the fibre since the penetration depth of liquid into the pit apertures due to pressure differences over the openings is always less than the fibre wall thickness. Finally, the authors estimate the particle detachment time constant, which is used to calculate k2, as the product of a constant that is dependent on the hydrodynamic conditions to which the fibres and particles are exposed and an exponential function of the ratio of filler-fibre bond energy to particle thermal energy. Calculations based on the theory outlined here suggest that the typical time for complete lumen loading to occur under ideal conditions is approximately one hour, which explains the apparent dependence of the degree of lumen loading on agitation time. The positive effect of particle concentration and the negative effect of pulp consistency seen experimentally are reflected theoretically by the influence these variables have on the concentration gradient across the fibre wall. Also of interest, 64 Petlicki and van de Ven (1994) modelled the two components of lumen loaded particles; those adhering to the interior fibre surface and those suspended in the liquid phase that occupies the lumen void. Figure 4.3 compares the curves that describe adsorption of filler onto the fibre surface according to the equations of Petlicki and van de Ven with the experimental data collected by Middleton and Scallan (1991). Petlicki and van de Ven remarked that, although the amount of adsorbed filler particles is almost constant for most of the experiment, the number of particles deposited on the external surface is decreasing with time and the number within the lumen is increasing with time. Thus, if the initial filler concentration (c0) is less than the maximum amount that can cover the lumen surface, then all the particles will eventually be inside the lumen and none will be adsorbed on the outside fibre surfaces. 0.00 50 100 150 200 TIME, m i n Figure 4.3. The change with agitation time of external (T) and internal (rL) adsorption of TiC>2 filler particles on bleached softwood Kraft pulp at 1 % consistency. c L is the filler concentration (g filler/g pulp) within the fibre lumens. Mixing occurred in a British Disintegrator whose stirrer rpm was 1000 with an initial filler concentration (c0) of 0.097 g filler per g fibre. The data were collected by Middleton and Scallan (1991). The solid lines are generated with equations developed by Petlicki and van de Ven (1994). 65 As in the case of filler deposition on the exterior fibre surface, particle size affects lumen loading kinetics. In this case, size is crucial because the dominant rate-determining step is the passage of particles through the fibre wall pit apertures. The smaller the particles, the greater is the flux of particles through these passageways. 4.4 Discussion 4.4.1 Laundering Several key ideas in the laundering literature can be applied to flexographic newspaper deinking. For example, the emphasis on the effect of the soiling process on subsequent soil removal suggests that the ink printing procedure may affect ink detachment during paper repulping. Mah et al. (1993) have shown that the degree of flexographic ink removal during the deinking process differs from newspaper to newspaper, but no direct study of the effect of printing on subsequent ink detachment has been made. And, unlike the bond between a substrate and an adhered soil layer, the bond between the paper and an ink layer is not well understood (see Chapter 2, section 2.4). Therefore, while it is known that the contact area between the soil particle and the textile plays a crucial role in soil detachment, the role played by the contact area between an ink layer and the paper substrate is not known. Just as soil removal is subdivided into three stages, so it may be useful to similarly describe ink detachment. By analogy then, the first stage of ink detachment is displacement of air by the deinking solution. This changes the thermodynamic conditions and alters the interaction energies between ink and pulp fibres and between individual ink particles. 66 During the second stage of ink detachment, the bonds between ink and paper are broken. In the flexographic deinking literature, it is suggested that the thermodynamic change experienced by flexo ink binders when exposed to an alkaline deinking solution causes ink pigment release (i.e., the binder dissolves). Under such circumstances the thermodynamic change itself breaks the bonds that hold the ink to the paper, and stages one and two are combined. During acidic repulping, where the flexographic ink presumably remains set on the paper after displacement of air by the deinking solution, it is likely that bonds must be mechanically broken if ink is to detach from the paper substrate. Finally, stage three of ink detachment can be described as the removal of the detached ink particle from the fibre surface and into the bulk solution of the repulping vessel. Caution in comparing this stage with that described in laundering is necessary since the ratio of aqueous phase to solid phase differs dramatically between the two systems. Because pulp fibres absorb a significant volume of solution, repulping even at 10 % consistency makes it difficult to distinguish a continuous aqueous phase from the pulp mass. Soil size and shape, surface roughness of both soil and substrate, mechanical energy imposed on the laundering system, bath temperature, and an adsorbed layer on the soil and substrate surfaces have been identified as variables that affect soil detachment. At least some of these variables also affect flexographic ink detachment. For example, Fernandez and Hodgson (1996, 1997) have indicated that an adsorbed layer of binder on the ink pigment surface creates a steric barrier in an alkaline solution. 67 And, under conditions where binder dissolution affects ink detachment, it is unlikely that particle size, shape, and surface roughness, fibre roughness, mechanical energy imparted to the deinking system, and repulping temperature affect ink removal. Indeed, Ciampa (1995) has shown that the final pulp brightness of deinked paper is not affected by repulping temperature ranging from 25 to 55 °C under basic repulping conditions. During acidic repulping however, some of the variables mentioned above may affect ink detachment. But pigment size, shape and surface roughness are likely not relevant because the ink pigment, which remains embedded in the binder matrix, does not exist as discrete particles. The contact area between ink and paper is likely affected by paper roughness and porosity; thus these variables may affect bond breakage during stage two of ink detachment under acidic conditions. The effect of mechanical energy on ink detachment during low pH repulping is not known. However, by analogy to particulate laundering, it is likely crucial to overcoming the forces of attachment between the ink layer and the paper substrate. In contrast, increasing repulping temperature may decrease ink detachment during low pH repulping if the temperature changes from below to above the surfactant cloud point (Ciampa, 1995). The effect of increasing the temperature above 55°C is not known but is probably not relevant since such high temperatures are likely uneconomical in a commercial deinking operation. Laundering studies have shown that soil detachment is not reversible. But, unlike soiling, which usually takes place in an air environment, ink attachment to paper during printing may take place in the alkaline aqueous environment of the ink solution. 68 More likely however, ink attachment probably does not occur until the ink has dried; thus, similarly to soil detachment, ink detachment is likely irreversible. (Once again, better knowledge of the flexographic ink printing mechanism might help illuminate ink behaviour during deinking.) Detached soil has been modelled as agglomerates of small particles and as individual soil particles. Detached flexographic ink particles can be similarly modelled, depending on repulping pH. During alkaline repulping, the dissolved binder molecules disperse individual pigment particles. Flexo ink particles detached during acidic repulping are likely agglomerates of pigment particles held together by the solid binder matrix. Just as in the case of laundering, deposition of these detached particles and agglomerates onto pulp fibres is likely dependent on the electrostatic energy and steric barriers prohibiting close approach, and on the energy of collision between the agglomerate or particle and the pulp fibre. Detached soil deposition is dependent on laundering temperature, time and intensity, soil and textile characteristics, and bath components. As previously mentioned, Ciampa (1995) has measured the effect of mixing time on ink deposition in an acidic solution on rinsed pulp brightness. Ink deposition under such conditions is likely time dependent (see Figure 3.3). However, the effects of other deinking conditions, ink and pulp characteristics and deinking solution components on detached flexographic ink deposition during repulping are not well understood. Interestingly, the polyacrylic binder molecules may act in a similar fashion to anti-deposition agents such as NaCMC which adsorb onto cotton fibres via hydrogen bonds. Similar adsorption by 69 binder molecules onto pulp fibres and ink pigment during high pH repulping would decrease the likelihood of pigment deposition by creating a large electrostatic energy barrier and steric barrier between the two surfaces. 4.4.2 Filler Retention Although the objective of filler addition during papermaking is opposite to that of ink detachment, the systems of filler retention and flexographic newspaper repulping are remarkably similar. The fibre components are virtually identical. Flexographic ink pigment and filler properties are also similar. The ink binder molecules in the deinking system are analogous to the polyelectrolytes added during filler retention. However, differences between the two systems exist. The pulp consistency during papermaking is as much as an order of magnitude lower than that found in repulping vessels, and the filler concentration is an order of magnitude higher than the pigment concentration during repulping, Also, the shear to which filler particles are exposed is likely greater than that experienced by ink particles. Nevertheless, the fundamental forces dictating the behaviour of fillers during papermaking and detached ink particles during repulping are surely similar. Just as with papermaking, ink particles and fibres have a negative surface charge during high pH repulping; thus detached ink deposition can only occur either via dampening of the particulate double layer thickness or by charge reversal. Unlike papermaking where charge reversal occurs by adsorption of cationic polyelectrolytes, charge reversal of ink particles during repulping via polyelectrolyte adsorption is unlikely because the ink binder molecules are anionic. However, the presence of calcium ions 70 contributed by the domestic water supply in some regions may affect particle and/or fibre charge. Charge in both cases is also affected by pH. The lower the pH, the less negative are the ^-potentials of all components. In addition, the more lignin in the fibres, the greater is the surface charge and the less likely will be the occurrence of particle deposition. Finally, during both filler retention and flexographic newspaper repulping, electric double layer thickness dampening will occur if the electrolyte concentration is above that of critical deposition. Kinetic considerations are extremely important for filler deposition during papermaking because the residence time in the papermachine wet-end is only several seconds. Although the typical batch repulping time is approximately 10 to 20 minutes, Galland and Vernac (1993a) and Ciampa (1995) have shown that final pulp brightness drops as repulping time increases. While several tests have proven that filler deposition follows Langmuir kinetic behaviour, it is not known if detached flexographic ink deposition during repulping follows a similar model. van de Ven (1989) argued that colloidal particles deposited on fibres are caught in the DLVO primary energy well (see Figure 3.1) because fine Ti02 particles deposited onto cellophane do not detach when subjected to a large shear rate (1000 s"1). This suggests that, if similarly small ink particles are caught in this deep well, then 100% ink detachment cannot occur. Detached ink deposition of this type is unlikely during high pH repulping due, if nothing else, to the steric barrier created by binder adsorption onto ink pigment particles. However, such deposition may be possible during low pH repulping. 71 If a particle is sufficiently small, then entrapment within the pulp fibre lumen is possible. Extensive study of filler deposition within fibre lumens has shown that those factors governing filler deposition on the external surfaces of fibres also affect filler deposition on internal fibre surfaces. Monolayer deposition of TiG*2 filler on fibre lumen surfaces (15 wt% of the pulp mass according to Middleton and Scallan, 1993), corresponds to approximately 6.5 % by mass of flexographic ink pigment. However, Middleton and Scallan (1993) have shown that previously dried fibres do not lumen load as well as fibres that have never been dried. Thus the nature of the fibre to be repulped (it has been dried) probably precludes the levels of lumen loading possible during papermaking with virgin pulp. Nevertheless, it seems possible that detached flexographic ink particles may enter fibre lumens during repulping and may subsequently deposit onto the internal surfaces if colloid forces permit. Detachment of such ink particles is unlikely because they do not experience the shear that exists outside the fibre wall. No study of flexographic ink particle lumen loading during paper repulping has been performed. Perhaps lumen loading of ink particles may also take place during printing. If this happens, then subsequent detachment from the lumen surface should occur during high pH repulping where the ink binder dissolves and disperses pigment particles. Such pigment particles will leave the lumen sanctuary only if the pigment concentration difference between the outside solution and that within the lumen is negative and only then if an exit passageway of appropriate size presents itself. Detachment of lumen-loaded ink during acidic repulping is unlikely because the solid pigment-binder 72 composite may not be exposed to sufficiently large detachment forces. However, the bending and flexing of fibres that occur during repulping may indeed crack the ink composite within the lumen, breaking the fibre-ink bond. Once again, if the cracked pieces of the composite are sufficiently small, then a concentration difference similar to that mentioned above may induce such an ink/binder matrix to leave the fibre lumen. 73 Chapter 5. Thesis Objectives This thesis aims to continue the work performed in the course of earlier studies of flexographic newsprint deinking, making use of relevant information from research into the analogous systems reviewed in Chapter 4. Although significant and interesting investigations have contributed to the understanding of flexographic ink behaviour during deinking, striking gaps in knowledge remain. Specifically, the direct measurement of ink mass retained on (or in) the pulp fibres upon completion of the deinking process has not been reported. Secondly, detailed knowledge of flexographic ink behaviour during the repulping process is missing. Therefore, the first objective of this thesis has been to develop a reliable method of measuring the mass of ink remaining attached to paper fibres after completion of flexographic newsprint repulping. The search for such a method, a description of the experimental procedure developed, and sensitivity tests of the method are provided in Chapter 6. The second objective of the thesis is to develop and test hypotheses that describe flexographic ink behaviour during repulping. To this end, the experiments described in Chapter 7 employ a model system designed to confirm the influence of potential determining ions (i.e., repulping pH) on the ink content in repulped newsprint fibres, as discussed by several authors. Because mechanical energy imparted to both laundering and filler retention systems plays such a pivotal role in both particulate detachment and deposition, the same model system is then used to investigate the 74 influences of both repulping time and repulping intensity on ink behaviour. Also, lumen loading of ink during both printing and repulping is investigated in a qualitative manner. 75 Chapter 6. Dependent Variable Study 6.1 Introduction The degree of ink pigment removal during deinking is an obvious dependent variable to use during repulping experiments and previous deinking studies have relied on the optical qualities of pulp sheets to estimate this variable. Pulp brightness, image analysis techniques (McCool and Taylor, 1983; Nguyen et al., 1992), and the Effective Residual Ink Concentration (ERIC) (Jordan and Popson, 1994), have all been employed. Pulp brightness, or whiteness, is an optical quality that is obtained by standard procedures in which the reflectance of 457 nm light from a pulp handsheet (Canadian Pulp and Paper Association (CPPA) Standard C.5) is compared to that of a 100% white standard (see Chapter 7, section 5 for a description of the measuring procedure). Image analysis techniques measure the total ink area or ink particle size distribution in a sample via digitised microscopic images of either a pulp pad or a monolayer of deinked pulp deposited on a membrane filter of small pore diameter. ERIC analysis, which measures the pad reflectance of infrared light (at a wavelength of approximately 1000 nm), eliminates the effect of pH on pulp because the influence of lignin's chromophores subsides at long wavelengths (Jordan and Popson, 1994). Furthermore, ERIC transforms the reflectance information to Kubelka-Monk coefficients (Robinson, 1975) and allows calculation of an effective ink concentration in the pulp sheet. The equation used to make such a calculation is: c i n k = (Sp/ki„k)((1- RP)2/(2 Rp) - (1- Rmix)2/(2 Rmix)) [6.1] 76 where q nk is the concentration of ink in the pulp (ppm), s p is the scattering coefficient of paper (m2/kg), kink is the light absorption coefficient of paper (m2/kg), Rp is the brightness of blank paper, and Rm i x is the reflectance of the deinked pulp. (Appendix 8 provides details of the development of equation [6.1] and an example of its use.) Such a transformation assumes that the light absorption coefficient of ink is much greater than that of paper, that the ink concentration in the sheet is a fraction of a percent, and that the ink particles within the sheet are uniformly distributed. Pulp brightness is an apparently reasonable dependent variable not only because it is sensitive to trace concentrations of ink in the pulp handsheet (Jordan and Popson, 1994), but also because high pulp brightness is the criterion used by paper manufacturers to market their product. However, as a dependent variable aimed at measuring successful ink detachment and removal during deinking experiments, pulp brightness possesses inherent difficulties because it depends on many factors other than ink mass in the pulp sheet. For example, pulp brightness is affected by the fines content in the sheet, pH of the pulp slurry from which the handsheet is made, filler content and size, as well as the size and shape of the ink particles. Also, it is affected by non-uniformities in ink distribution throughout the sheet. Thus, when comparing brightness results to determine the effects of repulping and/or separation conditions on ink detachment, and detached ink deposition and/or removal, care must be taken to account for possible effects from factors other than ink particle concentration in the sheet. 77 Both image analysis and ERIC are very useful under specific conditions with some ink types (Carre et al., 1994). But both are of limited value as methods of measuring the ink content in flexographic newsprint pulp from deinking experiments for at least two reasons. The minimum particle size for microscopic ink analysis is approximately 1 urn, which is larger than the typical size of flexographic ink particles in an alkaline environment. ERIC analysis is not limited by particle size per se, however ERIC results are dependent on both the ink concentration in the sample, and the ink particle size distribution in that sample. Since flexographic ink particle size is dependent on pH and possibly other repulping conditions (see Chapters 8 and 9), ERIC analysis of samples has the same limitations as brightness measurements. That is, comparison of ERIC results is reasonable only when the specific conditions that affect the ink particle size distribution are controlled. Beyond the difficulties associated with pulp brightness, ERIC and image analysis measurements discussed above, the sample preparation method employed during each of these analysis techniques, including preparation chemistry, can have a dramatic effect on the experimental results (Dorris, 1997; Scott, 1993). Hence it is crucial that sample preparation procedures be identical if results are to be successfully compared. 6.2 The Search for a New Method of Ink Pigment Content Measurement Difficulties inherent in pulp brightness, ERIC, and image analysis measurements precipitated a search for a more reliable method of indicating the degree of flexographic ink detachment during paper repulping performed over a range of repulping conditions. 78 First, a feature of ink particles that might be used to distinguish these particles from pulp fibres was identified. Then an effective analytical procedure was developed. The most distinguishing characteristic of ink particles is their remarkable ability to absorb light. However, light absorption data cannot always be easily transformed to mass data, as previously discussed. Separation via sedimentation is not possible because the specific gravities of carbon black and pulp fibres are similar (approximately 1.9 for the former and 1.3 for the latter (Perry and Green, pp. 3-95 & 3-11, 1984)). Separation via filtration then gravimetric analysis of separated portions is not possible because pulp fines can be of the size range of ink particles (i.e., less than 1 nm). Surface groups on carbon black cannot be used to identify the detached pigment in a slurry because pulp fibres possess identical surface groups (see Chapter 2). Methods of altering ink to allow its concentration to be measured in the presence of pulp fibres were investigated. Mixing radioactive carbon black particles with the ink pigment was considered because Grindstaff et al. (1970) had successfully used radioactive carbon black particles to trace soil behaviour during particulate soiling experiments. However, radioactive carbon black is no longer available from suppliers and production of such a species is prohibitively expensive. (A preliminary quote of between $5000 and $10000 for a total activity of 125 millicurie (approximately 10 grams of carbon black) was obtained from the only supplier (New England Nuclear Corp., Boston, U.S.A.) willing to discuss the doping of carbon black. An immunochemical technique of carbon black analysis in the pulp sheet was also considered. Such a method is often the analytical method of choice in clinical 79 chemistry and endocrinology because of its sensitivity, specificity, speed of analysis, ease of automation, cost effectiveness, and general applicability (Hall et al., 1990). Immunoassays are based on the idea that antibodies produced in animals can recognize and attach with extreme accuracy to certain chemical groups displayed on the surface of molecules. By combining specific antibodies with an appropriate indicator system, immunochemical assays can identify the presence of a particular chemical at levels approaching 1 ppb or less. Such a technique might work well in identifying carbon black in the pulp sheet (or suspended in the aqueous phase of the repulper) if a surface group specific to carbon black could be identified or such a group could be permanently adsorbed to the pigment surface. However, similar surface groups on both pulp fibres and ink pigments, and the uncertainty of permanent adsorption of an identifying group to the pigment, brought into question the effectiveness of the immunoassay technique. Milanova and Dorris (1993) used thermogravimetry to determine the solid components in small samples of froth from a laboratory flotation unit in which model offset ink was floated. During the analysis, previously dried froth samples were placed in open crucibles and heated in an oxygen atmosphere at a rate of 5°C/min. The analysis provided mass ratios of sample components to the initial mass of the sample. The thermogravimetric behaviour of flexographic ink and repulped newsprint fibres was investigated. It was found that the analysis results were sufficiently different to identify the ink pigment component in a dry repulped slurry. The analysis procedures 80 used during the model repulping experiments described in Chapter 7 are presented in this chapter. 6.3 Sample Preparation for Thermogravimetric Analysis of Residual Ink The ideal technique would measure the mass of detached ink, un-detached ink, and ink that has detached then deposited onto fibres, in situ, while repulping takes place. Thermogravimetric analysis measures the change in a sample's mass as the sample is heated at a specified rate, to a specific temperature, under controlled conditions. Although this technique can identify ink in the repulping slurry, it cannot distinguish between detached ink and ink attached to pulp fibres. Therefore, ink that has been detached from fibres during repulping must be removed from a pulp slurry sample before the thermogravimetric analysis can proceed. This situation is undesirable since, by performing a separation step before analysis, the final ink content of pulp may be affected, leaving one with data similar to those of previous researchers who combined the unit operations of repulping and ink separation. The effect of introducing an ink separation step can be mitigated by designing the step such that only a systematic error is introduced into the final results. The analysis equipment and procedures described in the next sections have attempted to achieve this. 6.3.1 Detached Ink Separation Separation Equipment Ink pigment, detached from newsprint during model repulping, was separated from the repulped paper in a 3.5 L modified dynamic drainage jar (mddj) whose components were made of hydrophobic materials to discourage deposition of the 81 hydrophilic ink particles. This device, sketched in Figure 6.1, was a filtration unit that allowed thorough rinsing of repulped paper samples. Ink pigment detached from the fibres during repulping, as well as small pulp fines to which ink may or may not have been attached, passed through the jar's 30 um nylon mesh screen and out through the exit port. Rinse solution was pumped into the mddj with a peristaltic pump. The flow of the filtrate was controlled by a needle valve to ensure that the slurry volume in the upper chamber of the mddj remained constant. Tygon tubing connected the exit port to a polyethylene tank in which the filtrate was collected. The slurry in the upper chamber of the mddj was stirred slowly (approximately 70 rpm, 2 W power) to avoid the formation of a filter cake on the screen. The impeller was placed as close to the screen as possible but did not contact it. Detailed sketches of the mddj apparatus including dimensional information are provided in Appendix 1. Ink Separation Procedure Ink detached during repulping was separated from the pulp fibres by rinsing in the mddj. A 1 L sample of diluted repulper slurry was added to the upper chamber of the mddj where it was diluted with 1 L of a solution having an identical pH. The jar's stirring motor was then started and at least 15 L of rinse solution of identical pH to that of the diluted repulper slurry sample was pumped into the mddj while, simultaneously, at least 15 L of mddj filtrate was collected. Such rinsing, which continued until the filtrate leaving the mddj appeared to be clear of ink, typically required 1.5 to 2 hours, and theoretically (i.e., using a continuous stirred tank analysis) removed 99.9% of all free ink 82 particles. The pulp consistency in the mddj was approximately 0.2%, a value at which pulp fibres are rarely in continuous contact (Kerekes et al., 1985). The mixing energy and intensity imparted to the rinsing system were no more than 2.7 kJ/L and 1 W/L respectively. Due to the relatively low pulp consistency in the mddj (0.2 %), most of this energy went into mixing the aqueous solution. Thus the energy imparted to the fibres was small compared to that imparted during repulping where the minimum mixing energy and intensity were 9 kJ/L and 5 W/L respectively, at 6% pulp consistency. stirrer rinse solution (15 L) -MI pump mddj 3 L capacity stirrer impeller mddj screen needle valve mddj filtrate >-••*•----•••• • i ' - • s Figure 6.1. A sketch of the detached ink separation apparatus. Once rinsed, the remaining slurry was siphoned from the jar and the jar was dismantled. Samples of the diluted repulper slurry, the diluted repulper slurry after rinsing in the mddj, and the mddj filtrate were stored in either glass or polyethylene 83 containers at 3°C. The mddj was cleaned with a standard laboratory glass cleaning detergent, then rinsed with several litres of distilled water, in preparation for the next experiment. 6.3.2 Thermogravimetric Analysis Samples Rinsed Pulp Carbon Black Content Three dry samples of rinsed pulp were typically prepared from each experiment's rinsed slurry. The slurry, which had been siphoned into a glass beaker, was gently stirred, then a portion (~ 200 mL) was poured into a 500 mL glass beaker. This portion was stirred and then used to fill a 100 mL glass graduated cylinder. The entire sampling process was then repeated twice to obtain the second and third samples. Three polycarbonate filters of 47 mm diameter, 10 um thickness, and 1.2 um pore diameter (Isopore track-etched filters from Millipore) were placed in the vacuum filtration system illustrated in Figure 6.2. Each 100 mL sample of rinsed slurry was acidified (to a pH level of ~ 1) with dilute (~ 0.1 M) sulphuric acid and filtered. The pulp pad formed on each filter's surface was then rinsed with 100 mL of acetone and 100 mL of methanol. Acidification of samples ensured that pulp fines and small precipitated binder particles that might be present agglomerated and did not plug the polycarbonate filters. Solvent rinsing, which removed ink binder from the pulp pad (Dorris and Nguyen, 1995), was needed to remove binder that would otherwise confound the thermogravimetric results. Also, initial experiments suggested that the presence of ink binders in samples increased the oxidation rate of carbon black. 84 Experiments were performed to determine the effect of acidification and solvent washing on newsprint. Six samples of blank newsprint identical to that used in all other experiments, were weighed at constant temperature and humidity, rinsed with a dilute solution of sulphuric acid (pH level of ~ 1) then washed with acetone and methanol. Once dry, the samples were re-weighed, again at constant temperature and humidity. Results show an average weight loss of 1.3%. Since ~85% of the newsprint is mechanical pulp, it may be that a small amount of wood resin was dissolved by the solvents. Because of this weight loss, the dry mass of acetone and methanol washed solids in the rinsed slurry was chosen as the basis weight of the carbon black content data. Figure 6.2. An illustration of the vacuum filtration system used to prepare samples for thermogravimetric analysis. A test of the solvent rinsing procedure was performed by comparing thermogravimetric data from three samples each of carbon black, laboratory manufactured flexographic ink, and the same ink washed with acetone and methanol. Figure 6.3 illustrates the data from all nine tests. The thermogravimetric (TG) analysis procedure is explained in section 6.4. The average dry weight percent of the solvent washed ink residue was 96.0% (st. dev. = 0.08). When compared to the average dry 85 weight percent of the carbon black residue (98.0%, st. dev. = 0.004%), it seems likely that not all binder was removed from the ink by using the solvent rinsing technique. (Note that the binder weight fraction of the dried laboratory ink was approximately 30% if no evaporation of binder occurred during printing (see Chapter 7, section 7.3.2).) But, because the binder fraction of rinsed pulp samples from model repulping experiments was small (i.e., less than 0.5% of the total sample weight), the binder content was neglected in subsequent calculations of the carbon black content. Detachment Efficiency Measurements During early experiments, mass balances around the modified dynamic drainage jar were performed to calculate ink detachment efficiencies (i.e. the mass ratio of detached ink to ink initially printed on the pulped paper). This required several added procedures during detached ink separation and TG analysis which included measuring the total volumes of mddj filtrate, rinsed slurry, and initial pulper slurry added to the mddj, accurate consistency measurements for these volumes, and accurate mass measurements of the polycarbonate filters and TG samples. (The complete procedure followed for all these measurements, as well as the MathCad programmes which calculated the mass percent carbon black in each sample and the ink detachment efficiency for each run, are provided in Appendix 2.) These added procedures, some of which resulted in very inaccurate data (e.g., consistency measurements typically have standard deviations of up to 5%), led to large errors in the detachment efficiency estimates. 86 6.4 Thermogravimetric Analysis Apparatus Thermogravimetric analysis of pulp samples was performed on a Perkin-Elmer TGS-2 system thermogravimetric analyzer (TGA). Figure 6.4 shows the TGA equipment which consists of an analyzer unit including a furnace and microbalance, a balance control unit, a heater control unit, and a Perkin-Elmer system 4 microprocessor controller which was used to program temperature profiles. Both the balance control and the heater control units are connected to a 286 microcomputer which records and stores time, sample weight, weight percent, and sample temperature data in an ASCII file. 100 90 r -80 f— Wt% 70 60 50 Dotted lines solid lines dashed lines \ \ \ v V \ w \ carbon black lab ink washed with acetone and methanol lab ink 100 150 200 250 300 350 400 450 500 Temperature (°C) Figure 6.3. A plot of thermogravimetric profiles of samples of carbon black, laboratory manufactured ink, and laboratory ink that have been washed with 100 mL of acetone and 100 mL of methanol. The heart of the TGA system is its balance and furnace. The furnace consists of a machined, thin-walled alumina mandril wound with a platinum filament which acts 87 alternately as a platinum resistance thermometer and resistance heater. The platinum weighing pan, which is approximately 0.75 cm in diameter, is supported by a platinum stirrup suspended from the microbalance hang-down wires. During TGA operation the furnace was continually flushed with approximately 100 mL/min of purge air. Oxygen was not used as a purge gas because it encourages rapid oxidation and since oxidation is an exothermic reaction, use of 0 2 as a purge gas elevated the sample temperature out of the desired range, thus altering the results of the analysis. Figure 6.4. A photograph of the thermogravimetric analysis apparatus. Procedure During TG analysis, a small, central strip (weighing approximately 10 to 25 mg) was cut from a randomly selected rinsed pulp sample, rolled into a spiral shape and carefully placed into the TG analyser crucible that had been previously tared. The crucible was then returned to the micro-balance. Care was taken to ensure that the 88 strip to be analysed was a vertical cross-section of the entire sample so that misrepresentation due to component segregation during filter-cake formation could be avoided. Also, the sample strip was placed entirely within the walls of the crucible so that no segment of the strip could fall out during analysis. Next, the loaded crucible was placed inside the furnace assembly. With air as the furnace purge gas, the TG furnace temperature was brought to approximately 100°C while the ASCII file, in which data was collected, was named. Both the data collection system, which recorded sample mass, mass percent and temperature every 20 seconds throughout the run, and the TG heating protocol, were then initiated. During the heating protocol used to analyse each sample, the furnace was held for 15 minutes at 105°C so that moisture in the sample evaporated before the sample temperature was increased. The furnace temperature was then brought to 500°C at a constant rate. This standard ramping rate was employed for reasons that are discussed presently. Once the furnace temperature reached 500°C, it was held at that value and the first 10 data recorded at this temperature were used in subsequent calculations. It was found that the thermogravimetric results were dependent on the temperature ramping rate as demonstrated in Figure 6.5, which shows TG data sets for carbon black samples collected at different temperature ramping rates. The heating protocol used in these runs was different from the standard run described above because the runs did not end at 500°C. Instead, each sample was allowed to oxidise completely or, in the case of the run whose ramping rate was 40°C/min, heated to 900°C. The effect of temperature ramping rate on TG data was likely due to the position 89 of the equipment's thermocouple, which was situated just outside the curcible that contained the pulp sample. As the ramping rate was increased, the lag between the time at which the thermocouple reached a specific temperature and the time when the sample reached that same temperature, also increased. The standard ramping rate of 20°C/min. was the quickest rate that also enabled a clear distinction between the primary sample components (i.e., carbon black and newsprint). 100 1 1 1 1 I I 1 1 ^ \ '*.'•. \ \ \ \ 80 v \ \ \ \ 1 l I 1 \ \ — 60 - 1 \ 1 I \ Wt% \ \ •:, i V > 1 •: i 1 40 Dashed Line = Solid Line = 5°C/min 10°C/min \ \ i \ \ \ \ \ \ 20 Dotted Lines = 20°C/min \ \ i \ \ -Dot-Dash Line = 40°C/min \ \ v \ 0 i i i i i V i \ _ 1 100 200 300 400 500 600 700 800 900 Temperature (°C) Figure 6.5. Thermogravimetric profiles of carbon black samples heated at different rates. Figure 6.6 shows typical TG profiles of repulped printed paper samples as well as profiles of blank newsprint and carbon black. Note that the TG heating protocol used to obtain the data from the newsprint and carbon black samples deviated from the standard protocol. These samples were oxidised with a temperature ramping rate of 20 °C/min (the standard) until the sample's weight percent was close to zero and did not change. In each of these cases the samples did not oxidise completely. This is due to inorganic compounds in the paper (ash) and in the carbon black (see Chapter 2, section 90 2.2.1). The ash content of the newsprint used in the experiments was found to be 0.7% (st. dev. = 0.2%), based on TG analysis of 7 newsprint samples. The drop in weight of the repulped paper samples that starts at around 285°C and ends just before 500°C is clearly due to combustion of the samples' newsprint component. By approximately 500°C, the samples' residue should be carbon black and ash. Not all TG run data ended at precisely 500°C because the thermocouple within the TG furnace did not always record the final temperature as that to which the furnace was set. This behaviour was due to occasional fouling of the thermocouple during analysis runs performed by other users of the TG equipment. The foulant was removed by running the TG to temperatures of approximately 1000°C. But, because such a cleaning routine caused degradation of the thermocouple, it was followed only when the fouling was significant. 100 200 300 400 500 600 700 800 900 Temperature (°C) Figure 6.6 Replicate thermogravimetric profiles of blank newsprint, pulped paper, and carbon black samples. Each profile was generated under standard conditions (i.e. with a temperature ramping rate of 20°C/min.). 91 On completion of the TG analysis, the mass fraction of residuals in the bone dry rinsed pulp sample was calculated by dividing the average of the first 10 mass data collected at 500°C by the average of the last ten sample mass data collected at 105°C. It was assumed that the mass loss during analysis was due to the fibre component in the newsprint. Thus the mass fraction of ash in the rinsed pulp sample was calculated and subtracted from the residual fraction to obtain the mass fraction of carbon black pigment in the sample. A sample calculation is provided in Appendix 3. 6.5 Thermogravimetric Analysis Sensitivity The sensitivity of thermogravimetric analysis to low levels of carbon black in a pulp sample was measured using the following procedure. 1. Laboratory manufactured ink was dried on a plastic sheet, removed from the plastic sheet, ground with mortar and pestle, and then rinsed with acetone and methanol on a 1.0 urn pore diameter filter. 2. A small (approx. 20 mg) piece of blank newsprint was dried and weighed in the TGA apparatus by setting the load temperature to 105°C and waiting until the weight of the sample did not change (approx. 15 minutes). This weight was recorded. 3. A small sample of rinsed ink from step 1 was dried and weighed in the TGA as per step two, this weight was recorded then the previously removed paper sample was added to the ink sample and the standard TGA heating protocol was initiated. 4. The final mass % of carbon black of the treated ink and paper samples which was determined by the TGA apparatus was recorded and compared to the mass % calculated from the initial balance measurements. 92 5. Steps 2 through 5 were repeated several times using different carbon black to paper ratios in the samples. The results from these tests are shown in Figures 6.7 which is a plot of mass percent carbon black in samples measured by the microbalance compared to the measurements from the TGA. Complete agreement between measurements would result in a Y = X function on the plot. Deviations from this line indicate errors. All data collected and calculations that generated the data shown in these figures are found in Appendix 3. % Carbon Black From Microbalance Figure 6.7. This plot shows all data collected during the experiment. The solid line is a X = Y curve. This comparison indicates that the TG analysis and microbalance data are in good agreement. Figure 6.8 is a dot plot of the differences between the data collected via TG analysis and those collected by the microbalance. The dot plot indicates that the differences are normally distributed around 0.0% with a standard deviation of 0.2%. 93 The variation in the differences is largely due to the variance in ash content in the newsprint portion of the samples. Not surprisingly, as the mass of pigment in the pulp samples decreases, the error associated with the TG analysis increases. This increase is shown clearly in Figure 6.9 which is a plot of the percent error in each measurement against the microbalance data. • • -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 Figure 6.8. A dot plot of the differences between the TG data and the microbalance data. 175 • 150 125 100 (0 (0 Q ^ c CD 1 o t € 75 «- O § 3? 2 w © Q. 5025 0 2 4 6 8 10 % Carbon Black From Microbalance Figure 6.9. This plot of percent error (defined as the 100 times the absolute difference between microbalance and TG data divided by the microbalance data) against microbalance data shows that the error in data increases as the amount of carbon black in pulp samples decreases. 94 Results from these tests show that the TG method loses sensitivity at low values of carbon black content. The loss reflects the increasing ratio of ash content to carbon black content. Below values of approximately 0.5%, changes in carbon black content cannot be detected reliably via the thermogravimetric method of analysis presented in this chapter. 95 Chapter 7. Experiments 7.1 Experimental Approach Model experiments have been designed to identify some of the repulping variables affecting ink detachment during repulping and to illuminate the mechanism(s) by which ink detachment takes place. Several two level factorial trials were performed during the first experimental stage. The effects of selected repulping chemicals, repulping time and repulping intensity were found by bench-scale repulping of paper printed in the laboratory with a model flexographic ink. All other repulping variables were held constant. Subsequent experiments explored results from the initial factorial trials by repulping over a range of pH levels, repulping times, and repulping power levels. Ink particle size, ink surface (zeta) potential, and ink solubility at different pH levels were studied as well. The possibility of ink entering fibre lumens during repulping was also investigated. Finally, because it was noted that repulping pH could change dramatically during the course of an experiment, the effect of ink binder on repulping pH, during both high pH and low pH repulping conditions, was investigated. 7.2 Apparatus 7.2.1 Repulping Newsprint, printed with flexographic ink, was repulped in either a 600 mL glass beaker (8 cm diameter, 9 cm long), or a similarly sized polycarbonate vessel (10 cm inside diameter, 0.5 cm thick, polycarbonate tube, closed at one end with a 11 cm diameter, 0.5 cm thick polycarbonate disc) which contained four equi-spaced, 96 rectangular baffles extending 0.5 cm into the vessel's interior. A 1/8 hp motor, to which a stirring shaft and impeller were attached, enabled paper "deflaking". A 6-vaned, stainless steel turbine style impeller "defibred" paper in the glass vessel. A polycarbonate impeller with six vanes of approximately 2 cm length and depth was used in the plastic unit. During experimental runs, the repulping vessel was suspended in a water bath. A wattmeter was used to measure the specific repulping power employed during each experiment. Figure 7.1 is a sketch of this repulping apparatus. watt meter »^Repulper (600 mL glass beaker) / I J 200 mL water bath Figure 7.1. Diagram of the repulping system. 7.2.2 Photographic Equipment Microscopic observation of rinsed pulp fibres was performed with a Nikon Optiphot microscope equipped with a Microflex HFX-II photomicrographic attachment. This microscope allows top-viewing of specimens via transmitted light. Light filters were not employed. 5x, 10x, 20x, and occasionally 40x objectives were used. 97 7.2.3 Brightness Measurement Equipment The Elrephro 2000 Spectrometer manufactured by Datacolor Inc., which was used to measure pulp brightness, is an industry standard. This device measures the light reflectance of pulp at various wavelengths. A CPPA standardized white tile and a box lined with black cloth are used to calibrate the machine to 100% and 0% reflectance respectively. The 12 mm diameter aperture was used during all measurements because pulp pad diameters were only 47 mm. All measurements were performed at 457 nm, which is the standard light wavelength at which brightness measurements are performed. 7.2.4 Light Absorbance Equipment Absorbance is defined as logio(Po/P) where P 0 is the initial power of a light beam and P is the power of the attenuated beam due to its travel through an absorbing suspension. Ink solubility was measured indirectly via absorbance of 470 nm light by solutions of ink binder possessing different pH levels. Zero absorbance was calibrated with a 0.1 wt% binder solution whose pH was 9.5 or greater. The measurements were performed with a Hach DR/2000 Direct Reading Spectrophotometer, which is a microprocessor-controlled, single-beam instrument. Standard 13 mm diameter, glass test tubes that measure approximately 10 cm in length were used to hold each sample of the binder solution. 7.2.5 Eiectrophoretic Mobilities Electrophoresis is the migration of particles through a solution under the influence of an electric field. The mobilities of ink and pigment particles at different pH 98 levels were obtained with a Malvern Zetasizer III. This instrument defines a volume in the sample suspension, located in the stationary layer of the suspension, by crossing two laser beams. The beams produce interference fringes in their intersection volume. Particles inside this volume interact with the fringes to produce scattered light that oscillates in time, depending on the speed of the particles. During analysis, the sample volume and an equivalent electrolyte volume are suctioned from syringe ports into the electrophoresis cell and electrode chambers. The cell employed during all tests reported here was the AZ4, which is a general purpose cell. The general relationship between electrophoretic mobility and the (^ -potential of a particle is given by Henry's law: U E = C(e/r,o)f(Ka) [7.1] where U E is the electrophoretic mobility, £ is the zeta-potential, s is the dielectric constant of the suspending liquid, r\0 is the viscosity of the suspending liquid, and f(ica) is a function of the particle radius (a) multiplied by the reciprocal of the Debye length (1/K, where K characterises the rate at which the potential decreases with distance from the particle surface). The Zetasizer software generates f(Ka) using a program, developed by O'Brien and White (1978), that accounts for displacement of charge effects and electro-osmotic flow around the spherical particle during electrophoresis. 7.2.6 Particle Size Analysis Equipment Ink particle size distribution measurements were performed using an Horiba model CAPA-500 centrifuge particle size analyzer. This equipment employs the relationship 99 S = (1/©2r)(dr/dt) = Me/f [7.2] where S is the sedimentation coefficient (i.e., the steady state velocity of a particle), dr/dt is the particle velocity, co is the angular velocity (radians/sec) and r is the radial distance of the particle from the axis of rotation, M e is the effective mass of the particle, and f is the frictional coefficient. Data from a centrifugal sedimentation experiment was expressed as a distribution of the sedimentation coefficient, S. This coefficient is a function of particle diameter according to where p and p 0 are the densities of the particle and the suspending medium respectively, and V is the particle volume. The frictional coefficient is given by Stokes' law as f = 37ir|od; so, if it is assumed that the particle is spherical, Equation [7.3] becomes where d is the particle diameter and r|0 is the viscosity of the suspending medium. Because this relationship assumes that the suspended particles are spherical, all mean particle size data are presented as Stokes equivalent diameters. 7.3 Experimental Materials 7.3.1 Chemicals and Photographic Film Aqueous solutions used in the experiments reported here were made with distilled water. Repulping pH was adjusted with solutions made with 98.5% pure S = V(p- P o)/f [7.3] S = (p-p0)d2/(18r|0) [7.4] 100 sodium hydroxide pellets (Fisher Scientific) and 98% pure sulphuric acid (also from Fisher Scientific). In preparation for thermogravimetric analysis, pulp samples containing ink were rinsed with Fisher Scientific's analytical grade acetone and methanol. Sodium chloride, used during some zeta-potential measurements, was also analytical grade. Black and white photographs of fibres were taken with Kodak TMax 400 film. All colour photographs were obtained with Kodachrome T64 (tungsten) film. 7.3.2. Model Flexographic Ink Newspaper printed with a model flexographic ink was used during all deinking experiments. The ink was manufactured in the laboratory with carbon black pigment (a furnace black) supplied by Cabot Inc. (Boston, U.S.A.) and a flexographic newspaper ink binder solution from S.C. Johnson & Son Inc. (Racine, U.S.A.). Both components are available to commercial ink manufacturers. Table 7.1 provides the composition of the ink binder solution and the ink recipe used during manufacturing. The carbon black supplied by Cabot Inc. consisted of aggregates of agglomerates that, in turn, were made of primary particles. Before adding pigment to the ink mixture, the carbon black was ground with a mortar and pestle to a fine powder and, although the size distribution of the powder is unknown, scanning electron micrographs of diluted and dried ink, one of which is shown in Figure 7.2, indicate an approximate primary particle diameter of 0.13 um. 101 Ink was manufactured by mixing the pigment powder with water and binder then grinding the mixture in a ball mill for 24 hours to aid in the comminution of the carbon black particles. The ink was subsequently kept at 3°C in a sealed, polypropylene tank. Before being used to print paper, the ink was stirred thoroughly so that the mass distribution of the ink components was homogeneous. Table 7.1: Recipe of Laboratory Manufactured Ink Ink Component Mass % Distilled Water 63.5 Cabot's Black Pearl 120 Carbon Black 14.5 S.C. Johnson & Son's Flexo "F" Ink Binder Solution 66-78% water 25-30% acrylic polymer with a monoethanolamine salt 2-4% acrylic polymer with an ammonium salt 22.0 Figure 7.2. A micrograph of ink particles. 102 The model ink used in the repulping experiments possessed neither the waxes nor fungicides found in proprietary industrial inks. And, although the print characteristics of the model ink are inferior to those of commercial inks (see Appendix 4 for a comparison), the binder used in this ink is one employed commercially (and is identical to the binder in the commercial ink whose print characteristics are compared with those of the model ink in Appendix 4). By using a simplified ink, the repulping experiments balanced simulation of industrial newsprint repulping with the scientific necessity of being able to identify each component in the repulping system. 7.3.3. Newsprint A commercial newsprint consisting of 85% TMP and 15% Kraft fibres, which was manufactured in McMillan Bloedel's Powell River facility, was used in each repulping trial. Although it was stored in constant humidity conditions at 3°C in darkness, the product experienced a 0.5 ISO brightness decrease over the course of the experiments. However, this brightness loss is within the experimental error of the brightness results (see Chapter 9). 7.3.4. Printed Paper All deinking experiments were performed on paper printed in the laboratory. Before printing, each newsprint sheet was conditioned at 20°C and 50% humidity for several hours then weighed with an analytical balance. The sheet was then immersed to its mid-line in laboratory manufactured ink and immediately removed from the ink. The dripping sheet was then hung to dry at constant temperature and humidity. The printing and drying times for each sheet were approximately 5 seconds and 4 hours 103 respectively. Unlike commercial drying that occurs by evaporation, the drying of the laboratory ink may have occurred by both evaporation and absorption of the water vehicle into the paper sheet (see Chapter 2 for a discussion of ink drying mechanisms). Once, dried, the printed paper was weighed and the ink to paper mass calculated. Like the model ink, the printing technique used during these experiments was simpler than that used commercially. Its purpose was to increase the mass fraction of ink on printed paper from the 1 or 2% commonly found in flexographically printed newspaper, to approximately 26 to 32%, and, in so-doing, flood the repulper with ink. These large amounts of ink in the system were needed to minimize the subsequent measurement errors and likely magnified the ink detachment mechanisms. Although the printing method was simpler than that used during commercial printing, the mechanism of ink adhesion to the paper occurring during laboratory printing was likely similar to that associated with the commercial process. Flexography was originally designed to print ink onto non-porous substrates (see Chapter 2, section 2.1.2). Therefore the objective of flexography is to lay the ink on the substrate, using minimal printing pressure. Figure 7.3 shows cross-section micrographs of the Vancouver Province newspaper (which is printed flexographically), paper printed in this laboratory with model ink, and a blank sheet of the newsprint used in the model experiments. Sample fibres are stained using basic green (according to CPPA standard B.3P) such that the yellow and blue-green colours indicate chemically and mechanically produced fibres respectively. These micrographs show that the laboratory printing technique obtained 104 its objective by increasing the ink/fibre contact area over that found in commercial newsprint. Ink layer depth is also increased in the laboratory printed paper over that of the commercial sample. Most significantly, these photographs show that, while occasional lumen loading of fibres with ink occurs during commercial printing, the phenomenon is common during laboratory printing. Ink particle movement into and out of fibre lumens during repulping is discussed in Chapter 10. i) ii) iii) Figure 7.3. Cross-sections of paper, i) Vancouver Province newspaper (200x magnification); ii) paper dipped in the model ink (100x mag.); iii) blank newsprint used in model experiments (100x mag.). 7.4 Experimental Procedures 7.4.1 Repulping Experiments Pre-Repulping Soak Approximately 17 grams of paper printed with model ink (i.e., three sheets of the paper, each having been previously immersed in ink and dried) were torn into 2 cm square pieces and soaked with 200 mL of repulping solution whose pH had previously 105 been adjusted, in a repulping vessel at room temperature for approximately 24 hours. This pre-repulping soak was necessary to ensure complete paper "deflaking" during repulping. Repulping After the soak, the repulping vessel was lowered into a water bath whose temperature was held at 30°C. The temperature of the vessel's contents was then allowed to equilibrate. Next, the stirring impeller was lowered into the repulping vessel to a preset distance from the bottom of the vessel and, after setting the impeller's stirring speed, repulping at approximately 8% solids (or 6% pulp consistency) was initiated. Repulping chemical concentrations, repulping time, and repulping power were varied from run to run. During initial factorial tests, independent factors were set at one of two levels as indicated in Table 7.2. Note that bleaching chemicals were not tested independently. This design feature undoubtedly confounded the individual effects of each chemical. However, the strategy is reasonable given that these chemicals are never added independently of one another in the industrial setting. Repulping power is defined as the difference between the total power input while repulping, measured via a wattmeter during each trial, and the no-load power input, which is that measured by a wattmeter while running the stirring motor in air. While the low repulping power level was set by repulping in the glass vessel with the stainless steel impeller at a stirring speed of 190 rpm, the high power runs were performed in the plastic, baffled, vessel with the paddle impeller turning at approximately 460 rpm. The 106 change in repulping vessel and impeller configuration were necessary to achieve a significantly high power level. Concern about how such a change might affect the energy dissipation pattern during repulping was mitigated by laboratory experience which shows that changing the vessel during toner ink repulping at 20 °C from a conical type to a kneader type does not change the mean size of detached ink particles (Carriere and Pinder, 1997). Additional repulping runs were performed after the factorial trials, during which the initial repulping pH was randomly varied from 2 to 12. Also, more runs were performed at both high and low repulping pH levels, during which repulping time and repulping power were varied. When not under study, the independent variables of all repulping experiments were held constant at the levels indicated in Table 7.3. Table 7.2: Independent Repulping Variables Factor Low Level High Level (~ pHof6) (~ pHof 11) NaOH (g/L) 0.0 1.0 Bleaching Chemicals(g/L): Na2Si03 0.0 2.0 DTPA 0.0 0.2 H 2 0 2 0.0 1.5 Repulping Time (min) 30 90 Repulping Power (W) ~2 ~8 After the specified time (30 minutes for most runs), the repulping vessel was removed from the water bath and the pH of the repulped slurry was measured. The slurry was then mixed with approximately 3800 mL of a solution whose pH had been 107 adjusted with H 2 S0 4 and/or NaOH to that of the repulped slurry, then the total volume of this diluted repulper slurry was brought to 4 L and the pH adjusted to that of the repulper slurry again. Samples from this diluted repulper slurry were then rinsed in the mddj as described in Chapter 6. Table 7.3: Repulping Constants Factor Level Pre-Repulping Soak Time (hr) -24 Pre-Repulping Soak Temperature (°C) -22 Repulping Temperature (°C) 30 Repulping Time (min) 30 Repulping Power (W) 2 In the case of the zero repulping power experiments, the laboratory printed paper was torn into pieces of approximately 4 cm square, soaked in the standard fashion, then prepared for mddj rinsing as described above. When rinsed slurry brightness measurements were required, the diluted repulper pulp pH was adjusted to either 3, in the case of low pH repulping conditions, or 9, in the case of high pH repulping conditions to ensure that the effects of alkali darkening and ink particle size distribution (both of which are dependent on pH - see Chapter 6 for a discussion) did not confound the brightness results. Dependent Variables Rinsed pulp carbon black content and ink detachment efficiency, both of which are described in Chapter 6, were used as the primary dependent variables in most of the repulping experiments. However, while varying the repulping time and repulping 108 intensity during high pH trials, it was noticed that, although the response of the carbon black content was insensitive to changes in these independent variables, the rinsed pulps of some runs were clearly brighter than those of others. Therefore, during the investigations of both of these variables, the ISO brightness (reflectance of 457 nm light) of the rinsed pulp pads, as well as the carbon black content, were measured. Each rinsed pulp brightness pad was manufactured by filtering 100 mL of rinsed pulp slurry (pH 9 or 3) through a 1.2 um pore diameter, 47 mm diameter, polycarbonate filter which had previously been placed in a Millipore glass filtering funnel. This procedure was identical to that used in preparation of the TG samples. As with the TG samples, the pulp pad was then rinsed with first 100 mL of acetone, then 100 mL of methanol. Next, the pulp pad and filter were removed from the filtration apparatus, and the pad was carefully peeled from the filter and placed on blotting paper. (Blotting paper is approximately 1 mm thick, highly absorbent, bleached paper. It is used during production of CPPA Standard C.5 handsheets and is common to all pulp testing laboratories.) The procedure was then repeated such that five replicate rinsed pulp pads were produced. The five pads were placed on the same blotting paper, covered with first another sheet of blotting paper, then the CPPA handsheet couch plate, and finally rolled ten times with the CPPA handsheet roller. Next the pads were placed on a single handsheet plate, covered with more blotting paper, pressed according to CPPA Standard C.5 for handsheet manufacture, and placed in the dark, at 22°C and 50% humidity, overnight. Next, the Elrephro Spectrophotometer was calibrated. Then, the 109 average ISO brightness of the five rinsed pulp pads was measured. This was done by stacking the pads, taking five brightness measurements on the top side of the stack, placing the top pad on the bottom then taking another five measurements of the top of the reconfigured stack, placing the top pad of the reconfigured stack on the bottom of the stack and taking another five brightness readings, and so on until five measurements of each of the top sides of each pad in the stack had been obtained. The stack was then turned over and the procedure repeated so that five brightness measurements of the bottom sides of each of the pads in the stack were recorded. Thus the average brightness of each run was the average of fifty measurements. The procedure for producing the pulp pad and obtaining brightness measurements is similar to the CPPA standard for brightness measurements (E.1), except for the method of pad formation. In the case of the zero repulping power runs, the rinsed 4 cm square pieces of paper were dried and stacked together. The brightness of each piece was then measured by following the procedure just described. 7.4.2 Model Ink Particle Size Experiments The pH of a diluted sample of model ink (approximately 50 mL of ink was mixed with 200 mL of distilled H 20) was adjusted to 13 by adding NaOH. A sample cell from the Horiba particle size analyser was then charged with a portion of the diluted ink sample and the size distribution of the ink particles was measured. The pH of the diluted ink was then decreased by adding H2SO4 and the analysis repeated. The pH was adjusted in consecutively descending order to a final value of 3 because the ink 110 suspension became unstable at pH levels below approximately 6. The counterion concentration in the ink suspension is affected by this method of pH adjustment. Since this concentration may affect the size distribution of particles (see Chapter 3), the particle size distributions of new diluted ink samples were measured at a high (10) and a low (6) pH level. This allowed an estimate of the error associated with the particle size measurements. 7.4.3 Zeta Potential Experiments The zeta-potentials of model ink particles were measured in the presence and absence of 0.01 M NaCI, in solutions with pH levels ranging from 2 to 11. Six samples containing 2 mL of a diluted model ink suspension (1 mL never-dried ink per 250 mL of distilled water) added to 40 mL of filtered, distilled, water in a 150 mL glass beaker were prepared. The pH level of each sample was adjusted to either 2, 3, 5, 7, 9, or 11 with sulphuric acid or sodium hydroxide. Six samples of filtered, distilled, water, to which either sulphuric acid or sodium hydroxide was added, in a similar fashion to the diluted ink samples, were then made. These six samples provided a blank electrolyte solution during each ^-potential measurement to which the Malvern Zetasizer automatically compared its data. A third set of 6 samples consisting of ground Black Pearl 120 carbon black pigment suspended in 0.01 M NaCI was also prepared. In this case, the pH of all samples was first increased to 10 before final pH adjustment was made because the initially alkaline conditions encouraged dispersion of the pigment particles which would otherwise remain unmixed and settle to the bottoms of the glass beakers. 111 The (^ -potential of particles in each sample was obtained by injecting approximately 10 mL of sample into the Malvern Zetasizer III along with 10 mL of the corresponding electrolyte solution. Each ^-potential and size distribution measurement was repeated three times. 7.4.4 Light Absorbance of Ink Binder Direct measurement of ink binder solubility is problematic because the binder precipitate is difficult to filter from a solution. (The precipitate sol is apparently less that 0.1 L i m in size because it passed easily through the filter with the smallest pore size, which is 0.1 um. Several passes through the filter were not possible because the of filter clogging problems.) Therefore, an indirect measurement of binder solubility was performed by observing changes in the light absorbance of a binder solution at different pH levels. The pH level of 200 mL of binder solution, whose concentration was 0.1 wt%, was raised to approximately 10. A sample of this solution was placed in a clean, 50 mL test-tube. Similar samples were obtained as the pH of the initial solution was lowered to various pH levels by adding H 2 S0 4 . Replicate data were obtained by repeating the entire sample preparation procedure with a fresh batch of 200 mL of binder solution. Random adjustment of binder solution pH was not performed because the binder polymer was sensitive to pH shock. Thus, unless the pH of the initial solution was slowly and carefully decreased, large polymer strings of various shapes and sizes would form and settle. 112 The absorbance samples thus prepared were immersed in a constant temperature water bath whose temperature was set to either 20 or 30 °C. Each sample's absorbance of various light wavelengths was measured by simply inserting the individual test tube into the spectrophotometer, adjusting the meter's wavelength indicator to the desired level and recording the absorbance of the sample. The spectrophotometer, which is designed to determine trace concentrations in waste water samples, required calibration by insertion of a zero absorbance sample. Therefore, because the binder solution is completely soluble at alkaline levels, the samples with pH values above 9.5 were used in this application. Absorbance sensitivity to binder precipitate was found to be the greatest at the light wavelength of 470 nm, hence this was chosen as the level at which all absorbance measurements were taken. Appendix 5 presents the experimental data and discussion on which these absorbance procedures were based. 7.4.5 Repulping pH Change Experiments During the course of early experimentation, it was noted that suspension pH could change dramatically during repulping. Of the three components in the repulping system, namely pulp fibres, ink binder, and carbon black pigment, the former two seemed more likely to affect repulping pH because carbon black is largely inert. Therefore, a set of experiments was initiated to investigate the repulping pH change in the absence and presence of ink binder at both high and low initial pH conditions. First, an "ink" made of dilute Flexo "F" binder and distilled water (i.e., without carbon black) was made in the same proportions as that found in the model ink (see 113 Chapter 7, section 3.2). Three sheets of blank newsprint (approximately 12.7 g) were "printed" with this binder solution in identical fashion to the paper used in the repulping experiments (see Chapter 7, section 3.4). The sheets were then shredded, soaked, and repulped according to the procedures outlined in Chapter 7, section 4.1. In all cases, the repulping vessel and impeller were glass and stainless steel respectively. Except for the repulping pH, which was set by adjusting the repulping solution in which shredded paper was soaked to either 2, 10, or 12 with either H 2 S0 4 or NaOH, the repulping conditions were not varied from run to run. The standard levels of pre-repulping soak time and temperature, as well as repulping temperature, time, and power input, are shown in Table 7.3. The repulping pH was measured after the soaking period and during repulping. Additional runs were performed which were identical to those described above except that blank paper was repulped rather than paper previously printed with dilute binder solution. All runs were performed in random order. 114 Chapter 8. Results from Factorial Experiments The objective of these initial experiments was to identify repulping variables having a significant effect on the carbon black content of rinsed pulp. Factorial experiments are ideally suited to such an objective because main effects can be estimated with minimal laboratory time, no presumption of underlying mechanisms is required, and, unlike the classic "one-factor-at-a-time" experimental design, interactions between effects can be estimated (Montgomery, 1997) from factorial experiment results. Therefore, sodium hydroxide concentration, brightening chemicals concentration (i.e., hydrogen peroxide, sodium silicate, and D.T.P.A.), repulping time and repulping power were studied via two level factorial experiments. The effects of ink binder and initial repulping pH on pH change during repulping were also investigated by the factorial technique. Each set of factorial experiments tested two variables at one time. Although this experimental strategy meant that the effects of interactions between three and four variables were ignored, it allowed experimental designs to evolve in an efficient manner. Also, clarity in presenting the results was facilitated by limiting each factorial experiment to a test of only two variables. Two level factorial experiments are performed by selecting a high and low value of each variable or factor under investigation. Four experimental runs are then performed so that all possible combinations of factor levels are tested. With only four test results, two estimates of each main effect can be calculated, as well as the interaction effect between the two variables. A main factor effect is defined as the 115 average change in experimental response produced by a change in the level of that factor. An interaction effect is the average difference in experimental response at different levels. Thus, an interaction between factors is seen when the difference in the dependent variable between the levels of one factor is not the same at all levels of the other factor. Interpretation of factorial results is typically performed by calculating an effects table. Such a table presents the average main and interaction effects, the standard error of each effect (which is the square root of the estimated variance of that effect), the standard deviation in experimental response, and the degrees of freedom inherent in the standard deviation. The reader is referred to statistical experimental design texts such as Box et al. (1978) and Montgomery (1997) for further details on basic factorial experimental design techniques. The following is a presentation and short discussion of all factorial experiments performed. Results from each set of two factor, two level experiments are presented in box-plots and effects tables. Each circle in a box-plot represents one experimental condition. The number within the circle is the average response of all runs performed at that condition. A discussion of the results accompanies the presentation of each set of results. Sample calculations are provided in Appendix 6. The first factorial design investigated effects of repulping chemicals on ink detachment. The results, shown in the box-plot in Figure 8.1, are presented as the residual pigment content in repulped and rinsed fibres. For example, the average carbon black content in rinsed pulp when paper was repulped with high levels of 116 brightening chemicals and a low level of sodium hydroxide, is 0.5 wt%. To determine significant effects, each effect is compared with its standard error. Although the main effects in Table 8.1 are relatively large, the significant nature of the interaction effect precludes a discussion of main effects without mention of the independent variables' influences on each other. This interaction indicates that the sodium hydroxide concentration has a positive effect on ink detachment only at low levels of brightening chemicals. Similarly, the effect of brightening chemicals on ink detachment is only seen if sodium hydroxide concentration is set at the low factorial level. These results suggest an effect of repulping pH on ink detachment since both hydrogen peroxide, sodium silicate, and sodium hydroxide increase solution pH. No evidence of an effect of the brightening chemicals beyond pH is seen in Table 8.1, where the magnitude of the main sodium hydroxide effect is within the standard error of the effect of the brightening chemicals. Carre et al. (1995), who tested the effect of H2O2 on deinking of a mixture of Brightening Chemicals Rinsed Pulp C.B. Content Figure 8.1. The box-plot of results from a 22 experiment designed to test the effects of pulp brightening chemicals and sodium hydroxide concentrations during pulping on the pigment content of rinsed pulp. Low and high factor levels are given in Table 7.2. 117 50 wt% old newspaper and 50 wt% old magazines, where the type of old newsprint was unstated but presumably offset, found similar results. Table 8.1. The Effects of Sodium Hydroxide and Brightening Chemicals on the Carbon Black Content of Rinsed Pulp FACTOR EFFECT (wt% C.B.) STANDARD ERROR OF EFFECT Average 4.4 1.1 sodium hydroxide -6.8 2.3 brightening chemicals -7.6 2.3 interaction effect 6.9 2.3 Pooled standard deviation of data = 2.7 with 3 degrees of freedom The influence of repulping pH is seen in the results and calculated effects of the second factorial experiment shown in Figure 8.2 and Table 8.2. Increasing the initial repulping solution pH from 2 to 12 decreases the carbon black content in rinsed pulp by about 14 wt%. The effect of repulping time is only slightly higher than the standard error but may indicate a small but significant increase in the pigment content of the rinsed pulp as repulping time is increased. Figure 8.3 and Table 8.3 show the results and effects from two level runs during which repulping power and pH were varied. The interaction effect is significant. This indicates that, under acidic repulping conditions, repulping power has a positive effect on the carbon black content but, when repulping occurs in an alkaline environment, no effect of repulping power is evident. 118 Repulping Time Rinsed Pulp C.B. Content Figure 8.2. The box-plot of results from a 2 2 experiment designed to test the effects of pulping energy (time) and pH on the pigment content of rinsed pulp. Pulping time levels were set at 30 and 90 minutes. The initial pH of the pulping solutions was set at either 2 or 12. Table 8.2. The Effects of Repulping Time and pH on the Carbon Black Content of Rinsed Pulp. FACTOR EFFECT (wt% C.B.) STANDARD ERROR OF EFFECT Average 8.8 0.7 Pulping time 2.1 1.4 PH -13.8 1.4 Interaction effect 0.4 1.4 Pooled standard deviation of data = 2.5 wit h 8 degrees of freedom 119 Repulping Power Rinsed Pulp C.B. Content Figure 8.3. The box-plot of results from a 2 2 experiment designed to test the effects of pulping power and pH on the pigment content of rinsed pulp. Factor levels are given in Table 7.2. Table 8.3 The Effects of Repulping Power and pH on the Carbon Black Content of Rinsed Pulp FACTOR EFFECT (wt% C.B.) STANDARD ERROR OF EFFECT Average 11.3 0.7 Pulping power 4.5 1.4 pH -19.4 1.4 Interaction effect -4.7 1.4 Pooled standard deviation of data = 2.2 wit fi 9 degrees of freedom The influence of ink binder on pH change during repulping was studied via a two level factorial experiment in which the presence of binder and initial repulping solution pH were the varied factors. As described in Chapter 7, the binder was either present or absent and the initial repulping solution pH was set at either 2 or 12. The results and effects from these experiments are shown in Figure 8.4 and Table 8.4. To enable an 120 unencumbered factorial analysis of data, the results have been coded by addition of absolute value bars around each measured pH change. Such coding is suggested when the dependent variable values can be either negative or positive (Hicks, 1993). Figure 8.4. The box-plot of results from a 2 2 experiment designed to test the effects of binder and initial pulping solution pH on the change in pulping solution pH that occurs during repulping. Table 8.4. The Effects of Initial pH and Binder Presence on the Change in Slurry pH During Repulping FACTOR EFFECT (A pH) STANDARD ERROR OF EFFECT Average 1.44 0.02 Binder presence 1.03 0.04 Initial pulping solution pH 1.57 0.04 Interaction effect -0.08 0.04 Pooled standard deviation of data = 0.05 with 2 degrees of freedom 121 Table 8.4 shows that binder affects slurry pH change during repulping regardless of the initial pH. The change at high pH conditions is due to the carboxylic acid groups in the binder polyelectrolyte. The large initial pH effect indicates that initial pH influences pH change during repulping thus reflecting the carboxylic acid component of the pulp fibres (Scott, 1996). 122 Chapter 9. Detailed Experimental Results The results presented in this chapter complement those presented in Chapter 8. The factorial experiments showed a large pH effect on the carbon black content in pulp from model repulping experiments. Therefore more repulping runs performed over the pH range between 2 to 10 were completed. The results from these runs are presented here. Also, the results from tests of the light absorbance of an ink binder solution, ink particle size measurements, and ink and carbon black ^-potential measurements are reported. These help to explain the change in ink pigment content observed over the pH range examined. The results from experiments that varied repulping time within and beyond the range tested in the factorial experiments are reported. Similarly, more results from experiments that studied the effect of repulping power are presented in this chapter. Both rinsed pulp brightness and carbon black content were the dependent variables measured during these repulping energy experiments. The factorial experiments that explored the change in pH during the model repulping trials were expanded. All data pertaining to this phenomenon are presented here. Microscopic observations of rinsed pulp fibres from the model repulping experiments reveal ink situated inside the central void of some fibres. Therefore experiments were also performed to establish whether lumen-loading of ink might occur during repulping. Photomicrographs of rinsed pulp fibres from these tests as well as those from the model repulping experiments are presented in this chapter. 123 c © c o o o « CD c o .a (0 O As with the preceding chapter, all quantitative data presented in this chapter are listed in Appendix 3. A comprehensive discussion of the results is reserved for Chapter 10. 20 _ 18 I 16 ~ 14 12 10 8 6 4 2 0 i i • i • • • • 1 6 7 PH 8 10 11 Figure 9.1. This plot illustrates rinsed pulp carbon black content as a function of repulping pH. The Effect ofpH Figure 9.1 illustrates the dependence on pH, of the ink pigment content in pulp that was printed with laboratory manufactured ink, soaked and repulped, then rinsed in the mddj. The repulping pH changed during the soaking/repulping process (results from an investigation into this phenomenon are presented towards the end of this chapter). Therefore, the repulper slurry pH values plotted in Figure 9.1 are those obtained upon completion of the 30 minute repulping time. Carbon black content is defined as the 124 mass percent of pigment in pulp or 100 x (carbon black mass)/(pulp and carbon black mass) (see Chapter 6). Precise replication of the data presented in Figure 9.1 was not possible because of the pH changes that took place during the repulping process (see Figure 9.16). However, statistical scrutiny of the data grouped into low pH repulping runs (pH<5.5) and high pH repulping runs (pH>7), supports the trend shown in Figure 9.1. The results from this analysis are shown in Table 9.1. The probability statistic, p, which is the probability that the two means describe data from the same population, was calculated by using a two-sided student-t test. Because it is much smaller than 0.01, a strong negative effect of repulping pH is evident. Table 9.1: Analysis of Results from Repulping pH Runs Repulping pH Average Carbon Black Content of Rinsed Pulp (wt%) Standard Deviation (%CB) High (>7) 1.5 ±0.8 Low (<5.5) 15.6 ±3.5 student-t test: the means are srgnil at the 0.01 (p = 3.12 X icantly different level 10"11) The visual difference between fibres repulped at high and low pH is shown in Figures 9.2 and 9.3. Of particular note is the number of low pH repulped fibres that contain ink within their lumens. Micrographs of fibres repulped and rinsed at low pH also show relatively large ink flakes outside the fibres, some possibly attached to the fibres, others obviously detached. Scant lumen-loading of individual ink particles can be 125 seen in fibres repulped and rinsed in high pH solutions. These fibres are relatively clean compared to those repulped at low pH. The solubility of Flexo "F" ink binder has been measured indirectly over a wide pH range as the light absorbance of a 0.1 wt% binder solution. The absorbance data were obtained at 20°C. The data, along with light transmittance data measured by Aspler et al. (1988), are shown in Figure 9.4 (see Appendix 3 for a data list and Appendix 5 for results from experiments that tested this method of measuring solubility). Figure 9.2. A micrograph of fibres pulped and rinsed under high pH conditions (pH = 9.8). Magnification is 200x. Aspler et al. (1988) studied the effects of newsprint properties on the quality of newspapers printed with water-based flexographic ink. They found that both low pH and the presence of aluminum ions caused ink instability. Their transmittance data, 126 presumably obtained at room temperature, are presented to support the results from the absorbance tests reported here. Figure 9.3. A micrograph of fibres pulped and rinsed under low pH conditions (pH = 5.2). Magnification is 400x. Light absorbance is a function of the mass of solids suspended in the absorbing solution, but it is also influenced by the size of the suspended particles. Therefore the results shown in Figure 9.4 must be analysed cautiously. The zero absorbance level was set at pH 9.5 because the solution appeared to be completely transparent at this pH level. The absorbance of the solution remained at zero for pH values exceeding 8.0. As the pH is lowered, the absorbance starts to increase significantly at approximately pH 7, corresponding to the high light transmittance values obtained by Aspler et al.. As 127 the solution becomes more acidic, the light absorbance increases (and the transmittance de-creases). It may be that the size of the previously precipitated binder particles increases as pH is decreased. However, increasing the size of precipitated binder particles by agglomeration should decrease the solid phase surface area and consequently decrease the suspension's light absorbance. Thus the increased absorbance can only be explained by a greater mass of binder precipitate in the solution. Both the light absorb-ance and transmittance behaviours of the binder suspensions reflect what one would expect if binder solubility increases with increases in pH. 100 80 o c 60 re O (ft § 40 20 0 i • • • ---• • ----• • a • + Absorb. (0.1 wt% sol'n) • Transmit. (0.5 wt% sol'n) 1 • * 1 • i h • • i 1 — • 1 • i 100 80 60 0) o c re C E w c re 40 jo 20 4 6 pH 8 10 12 Figure 9.4. This plot illustrates the relative solubility of low concentration binder solutions at 20°C as a function of solution pH. Absorbance data standard deviation is 5.8%. The transmittance data are from Aspler et al. (1988). A comparison of binder solution absorbance and rinsed pulp carbon black content is provided in Figure 9.5. This plot illustrates that both the carbon black content 128 and light absorbance increase as pH is lowered - the increase in both cases beginning at similar pH levels, i.e., between 7.5 and 7. The dependence on pH of the final ink detachment levels after repulping may be due to ink binder behaviour at different pH levels. •D CD </> c cE c c 2 c o o c a> E O) 20 15 £-10 a 3 Q. 0 o <^o o o Carbon Black Content of Rinsed Pulp # 0.1 wt% Binder Sol'n Absorbance 100 75 0) u JQ C o v> 25 < 6 8 PH 10 12 Figure 9.5. Ink binder solubility as reflected by light absorbance at different pH levels Figure 9.4) is superimposed on the plot of carbon black content vs. repulping pH shown in Figure 9.1. The average ink particle sizes of the model flexographic ink used to print the paper employed in the repulping experiments were measured at different pH levels. The data are shown in Figure 9.6 (see Appendix 3 for a data list). Never dried ink samples were diluted with distilled water before pH adjustment occurred. The results, given in terms of the equivalent Stokes diameter, become less accurate at low pH levels (note the replicates of model ink measurements taken at pH 6 and pH 10) because the ink dispersions become unstable under low pH conditions. Instability may be caused by 129 the protonation of carboxylic groups which, in turn, lowers the surface potential of the particles and precipitates the ink binder. Such effects may explain the large difference in replicate data at pH 6. Instability may also be caused by the increasing levels of counterions in solution, particularly at the lowest pH levels, that dampen the electric double layer thickness around each particle and prompt further coagulation. 100 E O N W o o r CO 0. c (0 'S 0 10 0.1 ET • 1 2 3 4 5 6 7 8 9 10 11 12 13 14 PH Figure 9.6. A plot of never-dried ink particle size vs. pH of the model flexographic newsprint ink. The star points are replicate data. The ^-potentials of the model ink particles and the carbon black pigment used in the model ink were measured at various pH levels. These results are presented in Figure 9.7 and are compared to values for similar systems found in the literature in Figures 9.8 and 9.9. The standard deviations of measurements collected in this labor-atory ranged from approximately ±0.2 to ±1.0 mV. (See Appendix 3 for a complete data list.) The potentials of ink particles and pigment particles, both suspended in 0.01 M 130 NaCl, are presented in Figure 9.7. Adsorption of binder molecules onto pigment surfaces in an electrolyte solution is illustrated by the greater negativity of the ink potentials relative to the pigment particles in the absence of binder. 10 > g « c a o Q. i h-jn 0 -10 -20 -30 -40 -50 I • Model Ink (0.01 M NaCl) • Carbon Black (0.01 M NaCl) • • mm • • • • • • 6 8 PH 10 12 14 Figure 9.7 ^-potentials, at different pH levels, of ink pigment particles suspended in electrolyte with (i.e., the model ink) and without (i.e., the carbon black) ink binder. In Figure 9.8, the ^-potentials of the never-dried, model ink particles with and without salt (0.01 M NaCl) are compared to the potentials of two model inks made with Flexo "F" binder and different pigment particles (i.e., Cabot's Black Pearl 420 (BP 420) and Regal 99), reported by Fernandez and Hodgson (1996). The ^-potentials of the commercial ink particles were reported by Liphard et al. (1990). Figure 9.9 compares the ^-potentials of the model ink pigment (Cabot's Black Pearl 120) with those of BP 420 and Regal 99 reported by Fernandez and Hodgson (1996), and with those of an unknown carbon black pigment reported by Liphard et al. (1990). Unlike the published potentials, those of BP 120 have been measured in a solution of 0.01 M NaCl. 131 > E 75 c I Q. i O N 10 0 •10 -20 •30 -40 -50 -60 -70 -80 • L i + A r~l I 1 t L r "1 / s X r i L J V + A * < > + + j A D , t L P . _ a c J 0 4 6 8 PH 10 12 14 Figure 9.8. (^ -potentials, at different pH levels, of the model flexographic ink are compared with values found in the literature. The stars and crosses represent data from the model ink used in this thesis obtained with and without 0.01 M NaCI respectively. The diamonds represent data from a commercial ink after Liphard et al. (1990); the triangles represent data from model ink #1 after Fernandez and Hodgson (1996), and the squares are data from model ink #2 after Fernandez and Hodgson (1996). Like the potentials of the other flexographic ink particles, those of the model ink used in experiments presented in this thesis, plotted in Figure 9.8, are close to neutral at very low pH values, but obtain a sufficiently negative value (i.e., -20 mV or more) to inhibit coagulation above pH 4. When compared to the potentials of other flexographic ink particles, the values obtained here are similar to those measured by Liphard et al. (1990), but are significantly less negative than those tested by Fernandez and Hodgson (1996) - particularly at pH values greater than 6. Fernandez and Hodgson removed the salt from the inks they tested (by precipitating the binder, washing it, and then dissolving 1 3 2 it in 0.1 M NaOH before adding it to the ink pigment). Consequently, the double layer thickness of the pigment particles in their studies was not significantly dampened. In contrast, a substantial concentration of counterions was present in the model ink samples used in the present study. The exact system studied by Liphard and coworkers is not known, however it is likely that no pre-treatment of the commercial ink was performed. It is therefore likely that the double layer thickness of these ink particles was also dampened, thus lowering the C-potentials of these particles. 20 > 15 10 0 -10 | " 2 0 I -30 i I -40 N -50 -60 i \ L I A U n D C i 0 X o V / k I i r IT Jf\ L c \ ] L a i ] $ c / ] • \ r i  A i i t A \ 6 8 PH 10 12 14 Figure 9.9. ^-potentials, at different pH levels, of carbon black pigment particles are compared. The stars represent model ink pigment (PB 120) data. The diamonds represent data for an unidentified carbon black after Liphard et al. (1990); the triangles and squares represent PB 420 and Regal 99 data respectively after Fernandez and Hodgson (1996). The phenomenon of double layer thickness dampening is apparent when the C,-potentials of the model ink particles obtained with and without added salt are compared (Figure 9.8). Except at very low and very high pH levels, when the values are similar, 133 the ^-potentials of ink particles suspended in the salt solution are less negative than those that have been diluted in distilled water. It is interesting to note that the effect of double layer compression by NaCl does not dominate the plot of the ^-potentials of the carbon black pigments used in each of the model inks, shown in Figure 9.9. It appears that other factors, such as differences in the carbon blacks, also influence the zeta-potential of pigment particles. The Effect of Repulping Energy The effect of repulping time on rinsed pulp pigment content and ISO brightness during both low and high pH repulping, investigated in factorial experiments, was revisited. The results from these tests, manifested as repulping energy, are shown in Figures 9.10 through 9.13. The variance in data seen in these plots is primarily due to inconsistent repulping power and pulp consistency levels that varied slightly from run to run. Statistical analysis was performed on data grouped into low and high repulping pH values; zero, low and high repulping times; and zero, low and high repulping power levels. High pH Conditions (pH > 7) The carbon black content results from the high pH trials are presented in Figure 9.10. Carbon black content data at zero repulping energy is not shown because, although the initial carbon black content of the printed sheet is approximately 29% (st. dev. = 3%), the paper was soaked for 24 hours subsequent to the printing process. Thus the carbon black content data at zero repulping time (and after soaking) is not known. Statistical analyses of the data in Figure 9.10, shown in Table 9.2, indicate that 134 data variance is sufficiently great to overwhelm any effects repulping time and power may have. That is, a student-t test comparing the carbon black content in rinsed fibres from the 90 minute repulping runs with that in rinsed fibres from the 30 minute runs indicates that the mean carbon black content measurements only become significantly different when the calculated probability is bordering on a high level (i.e., the sample means from the two populations are different only 80% of the time). Such a result suggests that, while the variation in data is great, a trend towards increasing carbon black content with repulping time might exist. The result from the statistical analysis that compares results from high power runs (>400 W/kg pulp) with low power runs (<400 W/kg pulp) is more conclusive. At a value of 0.6, the calculated probability statistic is high, indicating that there is no significant difference between the two populations. 5 c ^ 4 c ? O a ° "5 « t 2 5 ? o .E 1 re O 0 o O Time • Power < > EL o B o • O 0.5 1 1.5 2 2.5 Repulping Energy (MJ/kg Pulp) Figure 9.10. This plot shows the carbon black remaining in fibres that have been repulped and rinsed in high pH solutions at different repulping energy levels. Repulping energy was varied by changing either the repulping time (the diamonds) or the repulping power (the squares). The two runs represented by overlaping square and diamond symbols are replicates. 135 Table 9.2: Analysis of Carbon Black Data from High pH Runs that Varied Repulping Energy by Increasing Repulping Time or Repulpinc Power. Repulping Energy Variable Low Level Results (%CB) St. Dev. (%CB) High Level Results (%CB) St. Dev. (%CB) Student-t Test Time (30 or 90 minute) 1.5 1.4 3.1 1.5 The means are not significantly different (p = 0.2 = 0.2) Power (<400W/kg pulp or >400W/kg pulp 1.1 0.9 1.5 0.2 The means are not significantly different (p = 0.62 > 0.2) The effects of repulping time and power on rinsed pulp brightness under high pH conditions are illustrated in Figure 9.11. The standard deviations of these brightness measurements ranged from 0.4 to 3.4. A statistical analysis of the results, provided in Table 9.3, indicates that, under high pH repulping conditions, rinsed pulp brightness improves when some mechanical energy is added to the system but too much repulping has a negative effect on the final brightness of the pulp fibres. The trend of increasing pigment in the rinsed pulp with increasing repulping time, seen in Figure 9.10, is magnified here, thus illustrating the sensitivity of ISO brightness to the presence of carbon black (Jordan and Popson, 1994). However, the conclusion that the decrease in brightness is entirely due to increased carbon black in the pulp cannot be made since the decrease may also be due to comminution of the already submicrometric pigment particles. The fact that increasing repulping power also decreased final pulp brightness even though no corresponding increase in the carbon black content in the pulp occurred, suggests that particle comminution may indeed take place. 136 0) <0 <D C JC -O) • MM m o Q_ •o (0 c 45 40 35 30 25 20 15 10 • i < > • A a o • o O Time • Power El u 0.5 1 1.5 2 2.5 Repulping Energy (MJ/kg Pulp) Figure 9.11. Rinsed pulp brightness from high pH runs is plotted against repulping energy, varied by increasing either repulping time (the diamonds) or repulping power (the squares). Table 9.3: Analysis of Brightness Data from High pH Runs that Varied Repulping Energy by Increasing Repulping Time or Repulping Power Repulping Energy Variable Zero (ISO Br.) (St. Dev.) Low Level (ISO Br.) (St. Dev.) High Level (ISO Br.) (St. Dev.) Student-t Test Between Zero and Low Level Br. Responses Student-t Test Between Low and High Level Br. Responses Time (30 or 90 minute) 15.8 (3.6) 40.3 (0.9) 31.5 (3.5) p = 1.7 x 10-5 < 0.01 p = 5.9 x 10-4 < 0.01 Power (<400W/kg pulp or >400W/kg pulp 15.8 (3.6) 39.9(1.1) 34.5(1.3) p = 0.003 < 0.01 p = 0.02 < 0.05 Microscopic observation of fibres repulped in a high pH, high power environment revealed ink particles adhering to those sections of fibres exhibiting nodes, crimps and kinks. A sample micrograph of fibres repulped in high pH solutions with high mixing intensity conditions is shown in Figure 9.12. 137 Figure 9.12. Microphotograph of a fibre repulped under high pH, high shear conditions. (Magnification = 1600x) Low pH Conditions (pH < 5.5) Similar runs to those performed under high pH repulping conditions were also performed in acidic solutions. The effect of repulping energy on the rinsed pulp carbon black content and ISO brightness, affected by changing both the repulping intensity (power) and the repulping time, is illustrated in Figures 9.13 and 9.14. In Figure 9.13 the carbon black content at zero repulping time can be assumed to be about 29% because the ink binder does not dissolve during the 24 hours soak and no ink pigment is released into the soaking solution. Statistical analyses of results were performed in a 138 similar manner to those from alkaline runs and are presented in Tables 9.4 and 9.5. The statistical analysis of the carbon black content results from runs that varied 30 •«-> c o o U ra CO c o ja L . ra O X X * > < A A * L A Time X Power 0 0 5 1 1 5 : > 2 5 : 25 20 15 10 5 0 Repulping Energy (MJ/kg pulp) Figure 9.13. The carbon black remaining in fibres that have been repulped and rinsed in low pH solutions at different repulping energy levels. Repulping energy was varied by changing either the repulping time (the triangles) or repulping power (the x's). Table 9.4: Analysis of Carbon Black Content Data from Low pH Runs that Varied Repulpini 3 Energy by lncreasin< 3 Repulping n fime or Repulping Power Repulping Energy Variable Low Level Results (%CB) St. Dev. (%CB) High Level Results (%CB) St. Dev. (%CB) Student-t Test Time (30 or 90 minute) 14.9 3.3 16.5 3.1 The means are not significantly different (p = 0.6 > 0.2) Power (<400W/kg pulp or >400W/kg pulp 16.4 2.4 25.6 1.8 The means are significantly different (p = 0.0058 < 0.01) power indicates a positive effect at low repulping pH levels. This echoes the interaction effect seen in the previous chapter. In contrast, repulping time does not affect the 139 pigment content in the rinsed pulp. 30 25 20 c sz m o 15 Q. CO 3 w n T3 O c 10 5 0 A i A 7T > X X A Time X Pov /er £ — 0 0 5 I 1 .5 ; > 2 .5 : Repulping Energy (MJ/kg Pulp) Figure 9.14. Rinsed pulp brightness from low pH runs is plotted against repulping energy, varied by increasing either repulping time (triangles) or repulping power (x's). The standard deviation of these brightness measurements ranged from 0.2 to 7.7. As in the case of high pH repulping, a positive effect on rinsed pulp brightness of repulping power up to 400 W/kg pulp is seen during both high and low pH repulping. In contrast, an increase of repulping power from 400 to 800 W/kg pulp has no significant influence on the rinsed pulp brightness from acidic runs, even though this level of power has a reasonably significant negative influence on the brightness from alkaline runs. Increasing repulping time from 30 to 90 minutes has the opposite effect at low and high pH, increasing the brightness at low pH while reducing it at high pH. Repulping Energy Variable Zero (ISO Br.) (St. Dev.) Low Level (ISO Br.) (St. Dev.) High Level (ISO Br.) (St. Dev.) Student-t Test Between Zero and Low Level Br. Responses Student-t Test Between Low and High Level Br. Responses Time (30 or 90 minute) 4.5 (0.0) 21.1 (2.3) 26.4(1.2) p = 1.7 x10"5< 0.01 p • 5.9 x 10"4< 0.05 Power (<400W/kg pulp or >400W/kg pulp) 4.5 (0.0) 21.1 (2.3) 18.4 (0.9) p = 6.6 x 10"4 < 0.01 p = 0.2 > 0.05 140 Lumen Loading Handsheets made from bleached Kraft pulp were printed with model ink via the dipping technique, then repulped at either high or low pH with blank newsprint containing approximately 85% TMP fibre in a manner similar to ail previous low mixing intensity experiments. Samples of the rinsed pulp were then treated with stain capable of colouring chemical pulp yellow and mechanical pulp blue-green and photographs were obtained. Figure 9.15 shows a sample microphotograph of fibre samples pulped in an alkali solution. Observation of five samples from each test indicates that blank TMP fibres be-come lumen loaded with ink during high pH repulping. (Such lumen loading can be distinguished by an ink boundary that corresponds to the inside diameter of a fibre.) Figure 9.15. Kraft and TMP fibres pulped and rinsed in a high pH solution (mag. = 200x). The yellow fibres are Kraft, most of which are from printed handsheets. The blue-green fibres are thermomechanical, from blank newsprint pulped with the Kraft fibre handsheets. 141 pH Change During Repulping A phenomenon noted during the repulping tests performed at different pH levels was the change in repulping solution pH during the soaking/repulping process. The change in pH was dependent on the initial pH of the solution in which the shredded newsprint was soaked and repulped. This is illustrated in both Figures 9.16 and 9.17. The plot of data from model repulping experiments is presented Figure 9.16. The initial pH values were measured prior to the addition of shredded paper to the repulping solution. Clearly, repulping pH decreases during the soaking/repulping process of alkaline runs and increases when the initial pH is low. 14 12 10 8 6 4 2 0 X X initial pH • final pH / V ) tOL C < 6 ii c ] u m Q i I i 9, t I B 0 2 4 6 8 10 12 14 16 18 20 Rinsed Pulp Carbon Black Content (wt%) Figure 9.16. The change in pH occurring during the soaking and repulping process. Figure 9.17 shows the results from a set of runs designed to characterise the phenomenon shown in the previous figure. During these runs, shredded newsprint that 142 was either blank or had been previously dipped in a dilute solution of Flexo "F" binder (similar concentration to the model ink formulation) then dried, was soaked and repulped using the identical procedure to that employed in earlier tests that measured carbon black content as a function of pH. Three initial pH levels were tested. The pH of the system was measured throughout the repulping step of each run. I j'SThr".'soak "} - f — — j 30 min. pulping [• Time Figure 9.17. Repulping pH as a function of time is shown in this plot. Three levels of initial repulping solution pH were tested. Figure 9.17 illustrates that, during these bench-scale tests, the pH change previously noted primarily occurs, not surprisingly, in the 24 hour soak period prior to the repulping step. Also, when the initial pH is 12, the repulping pH drops more when binder is present than in its absence. If the initial pH is 10, then repulping pH drops slightly more in the absence of binder. When the initial pH is acidic, the final pH increases more in the presence of binder than in its absence. This is seen more clearly in Figure 9.18 where the change in pH (i.e. final pH - initial pH) at the three initial pH 143 levels, with and without binder, is illustrated. Error associated with thes data is small; the pooled stand-ard deviation, which is based on two replicates of three experimental conditions, being 0.07. The low and high pH data presented in Figure 9.18 were converted to changes in hydroxyl and hydrogen ion concentrations (i.e., final concentration - initial concentration) and these results are shown in Table 9.6. 1.5 |-1 0.5 0 -z a | -0.5 'c I "1 a 1 -1.5 ii -2 -2.5 -3 -2 10 12 Initial pH Figure 9.18 The effect of Flexo "F" binder on pH change during the soaking/repulping process at different initial repulping solution pH levels. Table 9.6: A[ION] with and without Binder at Different Initial pH Levels Initial pH of the Pulping Solution A[ION] in the Presence of Flexo "F" Binder A[ION] in the Absence of Flexo "F" Binder 2 1.48 x 10-1 1 mol/L OH" 2.6 x 10-13mol/L OH" 12 3.97x10-10mol/L H + 4.42 x 10-1 1 mol/L H + Insight into the influence of binder on ink deposition and re-detachment behaviour during acidic repulping was obtained by repulping blank newsprint with ground carbon black pigment. Almost 100% of the carbon black added to the repulper 144 remained in the pulp after mddj rinsing (initial carbon black content in the repulper was 31% (dry basis), and the rinsed carbon black content was 30%). Although the initial carbon black concentration in this run was similar to those in which laboratory printed paper was repulped under low pH conditions, the residual pigment concentration is significantly more than that measured during the more standard low pH repulping runs. A photograph of the rinsed pulp fibres from this single run is shown in Figure 9.19. Figure 9.19. A photograph of rinsed fibres repulped in an acidic solution (pH=2) under standard conditions except that, rather than repulping laboratory printed paper, ground carbon black pigment (Black Pearl 120) was added to blank newsprint. (Magnification = 200x) 145 Chapter 10. Discussion The Hypothesis It is hypothesised in this thesis that flexographic ink behaviour during newspaper repulping can be described as initial ink detachment and subsequent ink deposition/re-detachment kinetic equilibrium. Ink detachment takes place through one of two mechanisms, depending on the repulping pH. During acidic repulping, ink binder does not dissolve and the binder-pigment mass is ripped from the fibrous substrate by shear introduced in the repulping system by mixing energy. Large ink particles so detached may subsequently comminute and are thus susceptible to deposition. In the case of alkaline repulping, ink detachment occurs via an irreversible change in thermodynamic conditions resulting in ink binder dissolution and consequent steric stabilisation of pigment particles in the aqueous phase of the repulper slurry. Electrostatic stabilisation may also occur if the counterion concentration in the repulping solution is sufficiently low. Regardless of repulping pH, once ink detachment has occurred, it is proposed that detached ink deposits onto fibres if thermodynamic conditions and particle size are appropriate. This deposition behaviour is balanced by ink particle re-detachment, due to shear induced by pulp mixing, according to a Langmuir-type model. The discussion presented in this chapter argues the hypothesis outlined above by reviewing the experimental results presented in the preceding chapters as well as recalling other data in the literature. 146 Ink Detachment: Alkaline Repulping Several authors have shown that the final pulp brightness of alkaline repulped fibres improves when the pulp is thoroughly washed and have concluded that the low brightness of high pH repulped fibres is due, not to poor ink detachment, but to the subsequently inefficient separation of the detached ink from the pulp slurry (Jarrehult et al., 1989; Clewley et al., 1990; Galland et al., 1995; Heimburger, 1992; Borchardt, 1997). The experimental results shown in Figure 9.1 verify that ink detachment under basic repulping conditions is high. It has also been proposed in the literature that ink detachment during alkaline repulping occurs via binder dissolution which results in an electrostatic energy barrier surrounding each pigment particle as predicted by the DLVO theory (e.g., Liphard et al., 1990). This barrier stabilises the ink particles once the binder is dissolved (Borchardt, 1997). Certainly, ^-potential measurements of never-dried ink particles show that potentials are of sufficient magnitude to stabilise a suspension by electrostatic repulsion in solutions of pH greater than about 4 (see Figure 9.7). Thus, the assumption that binder dissolution, electrostatic stabilisation, and ink detachment occur at the same pH is reasonable. However, Figure 9.1 illustrates that ink detachment does not occur until relatively high pH levels. Figure 9.6, which shows that the size of model ink particles increases once the suspending pH is below approximately 6, also brings into question the assumption that binder dissolution, ink detachment, and electrostatic stabilisation occur at the same pH value. 147 Investigations of ink binder solubility, the findings of which are presented in Figure 9.4, provide the foundations for an alternate explanation of the ink behaviour. While ink binder begins to dissolve at a pH level of approximately 2.5, complete dissolution does not occur until approximately pH 7. Once binder precipitation is initiated by decreasing pH from an alkali level to around 7, the diameter of single primary aggregate pigment particles should increase and more than one pigment particle may be caught in a matrix of solid polyelectrolyte molecules. These relatively large particles should (and do) remain somewhat charged until complete precipitation of binder occurs, at a pH level of approximately 2.5. Thus, the ^-potential of ink particles may not be crucial to ink stability during high pH deinking. The effect of binder solubility is further revealed in Figure 9.5. Here, the residual ink in the rinsed pulp correlates closely with binder solubility. Data presented by Fernandez and Hodgson (1997) also erode the accuracy of characterising ink detachment in alkaline solutions as discrete particles detached via DLVO and hydrodynamic forces during repulping by showing that flexographic ink particles are stabilised, at least in part, by a steric barrier resulting from ink binder adsorption on the pigment. The above discussion suggests that an accurate model of ink detachment in alkaline solutions may be one where binder dissolution causes liberation of ink pigment particles. Such particles are prevented from re-depositing by a steric effect that results from adsorbed binder molecules on the surface of the pigment particles. Although plausible, the model can be challenged because, under such circumstances, the significance of a steric barrier relative to electrostatic repulsion is not known. 148 To estimate the electrostatic effect during high pH repulping, plots of interaction energies between an ink particle and one fibre, modeled respectively as a submicrometric sphere and an infinite plate, against intersurface distance have been generated. These are shown in Figure 10.1, where each curve represents interactions under different counterion concentrations. In each case the solution pH is assumed to be 10. The attractive van der Waal's interaction energy, V^, was found by using the relationship developed by Hamaker (1937). The Hamaker constant was estimated as the geometric mean of the values for wood fibres in water (Jaycock and Pearson, 1976) and acid treated carbon black particles in water (Horwatt et al., 1989). The repulsive double layer interaction energy, V e , was calculated via the method outlined by Hogg et al. (1966), assuming that the characteristic radius of the fibre was much larger than that of the ink particle. The ink particle radius was obtained from Figure 9.6. Figure 9.7 provided the ^-potential of the ink particle at pH 10 and this value was substituted for the surface potential. The ^-potential of the fibre surface was obtained from Jaycock and Pearson (1976). The total interaction energy, Vj , shown in the plot is simply the sum of V e and Vd. The short-range hydration force that likely acts between the ink particle and fibre surfaces, both of which are hydrophilic, is not calculated because, as yet, no quantitative theory for these forces is available (van de Ven, 1989). Appendix 7 provides a sample calculation as well as the data used to produce the curves. A brief discussion of the assumptions made for these calculations is also provided. The electrolyte concentration in the model repulper was estimated as that arising from contributions made by the initial repulping solution (pH = 12). (The 149 monoethanolamine ions associated with the ink binder do not dissociate at pH 10 (refer to the discussion on p. 164).) The curve generated by this estimate, which corresponds to a counterion concentration of 0.01 M, shows that large electrostatic effects are not significantly dampened. The second curve in Figure 10.1, with an electrolyte concentration of 0.04 M, illustrates the counterion concentration at which deposition of particles onto fibres can occur (i.e., the critical deposition concentration (Everett, 1988, p. 132)). The third curve (counterion concentration = 0.66 M) estimates the interaction under commercial repulping conditions. 70 r— 60 d 50 40 30 LU 20 I «H CO Q3 0 -10 -20 -30 i i I i i i i I i i i i I i i i i I i i i i I i i i i I i i i i I i i i counterion concentration = 0.01 M T - r i i i i counterion concentration = 0.04 M counterion concentration = 0.06 M i i i I i i i i I i i i i l i i i i l i i i i l i i i i I i i i i I i i i i l i i i i l i i i i 0 5 10 15 20 25 30 35 40 45 50 Intersurface Distance (nm) Figure 10.1. The interaction plots between a submicrometric sphere (ink particle) and a plate (pulp fibre). V T is the total interaction energy which is the sum of V e and Vcj. The pH level in these systems is 10 and the three curves are obtained at different counterion concentrations. 150 It is generally recognised that an interaction energy barrier of about 15 kT is sufficient to prevent colloidal particles in Brownian motion from depositing on a surface (Middleton and Scallan, 1985; van de Ven, 1989b, p. 40). In the model repulping system, this should occur at approximately 0.05 M. Therefore, the curves in Figure 10.1 suggest that, at pH 10, with a particle diameter of 0.2 u,m and an ink particle ^-potential of -50 mV, the counterion concentration in the model repulper does not attenuate electrostatic repulsive interactions. It seems likely then, that ink particle stabilisation during high pH repulping in the model system is due, at least in part, to electrostatic stabilisation of the pigment particles. Although not sufficiently high in the model experiments, the counterion concentration may significantly subdue electrostatic interactions in commercial repulping units where sodium hydroxide, sodium silicate, and sodium oleate are typically added at the levels of 1.2 wt%, 2.0 wt%, and 0.63 wt% of the mass of the dried pulp respectively (see Table 2.1). If repulping takes place at 10% consistency, then the concentration of counterions in solution contributed by these additives is approximately 0.06 M. Figure 10.1 suggests that such a concentration renders ineffective the energy barrier caused by electrostatic interactions. Increases in the counterion concentration, contributed by the process water, and/or pulp, enhance the likelihood that the electric double layers of the interacting colloidal surfaces will compress such that the electrostatic forces between the negatively-charged surfaces during industrial repulping become insignificant. 151 Visual examination of rinsed pulp samples from alkaline repulping runs (Figure 9.2) shows that most of the ink deposited inside the fibre lumens during printing is able to escape into the bulk solution during subsequent soaking and/or repulping. High magnification observation of the fibres from these samples reveals that those few ink pigment particles remaining within the fibres (which are presumably primary aggregates) exhibit Brownian motion. (Although particulate motion is obvious while observing the samples, the motion cannot be captured in still photographs.) These particles may reflect the ink particle concentration outside the pulp fibres because it has been shown that particle transport into and out of fibre lumens is governed by diffusion (Petlicki and van de Ven, 1994). Alternatively, they may have become trapped inside the lumens during repulping by fibre contortions that caused blockage of the particle escape route. Because ink inside the fibre lumen can detach from the fibre's inside surface and diffuse through pit apertures into the alkaline bulk solution, it is clear that ink binder dissolution promotes ink detachment regardless of shear. However, a comparison between brightness results obtained after no repulping and those from tests where repulping took place (shown in Figures 9.11 and 9.14) indicates that some mechanical energy is essential for high levels of ink detachment to occur. The need for this energy may be related to significant penetration of flexographic ink into the interfibre voids of the paper's surface layer that occurs during commercial flexographic newsprint printing (Gregersen et al., 1995) and most certainly occurred during model printing in this laboratory. Just as some ink particles remain in the fibre lumens even under high pH conditions, it may be that, unless the printed paper is pulled apart during the repulping 152 process (something that obviously does not occur under zero repulping energy conditions), ink trapped in these voids cannot physically escape. Detached Ink Deposition and Re-detachment: Alkaline Repulping When flexographic ink detaches from newsprint via binder dissolution, the detached pigment particles are likely prevented from depositing onto pulp fibres by the electrosteric effect created by adsorbed binder (polyelectrolyte) molecules. Results published by Chabot et al. (1995) support this assumption by showing that, under the experimental conditions summarised in Chapter 3, section 3.2, flexographic ink deposition does not occur when ink (never-dried) is mixed with a mechanical pulp slurry whose pH is 8.5, for 5 minutes. However, the brightness results provided in Figure 9.11 suggest that, under somewhat different repulping conditions, deposition of flexographic ink during alkaline repulping occurs when the repulping time is extended from 30 to 90 minutes. The data shown in this figure echo those published by Galland and Vernac (1993a). Although experimental results showing the effect of repulping time on the carbon black content in rinsed pulp (Figures 8.2 and 9.10) are vague (due to the insensitivity of thermogravimetric analysis at low levels of ink content), a trace of deposition can be seen. Also, rinsed pulp brightness data from experiments that change repulping intensity, reported in Figure 9.11, indicate increased ink deposition as repulping energy is increased. It can be concluded that, while ink particle deposition may be negligible under low repulping energy conditions, particle deposition increasingly occurs as repulping energy levels are increased. 153 For particle deposition to occur, a situation that degrades the electrosteric stability of the particles must develop. Such a change may result from increasing the repulping energy because the increased exposure to shear experienced by each particle may cause the breakage of bonds between the binder molecules and the pigment surfaces. Curiously, photomicrographs of rinsed fibres repulped under high energy (see Figure 9.12 for example) indicate that ink pigment particles deposit on fibre kinks and micro-compressions. Such fibre deformations only occur at relatively high energy levels (Bennington and Seth (1989) showed that Kraft fibres at 6% consistency developed microcompressions, kinks, and curl when treated at energy levels above approximately 2 MJ/kg.). It may be that, given sufficient time or intensity, enough binder is ripped from particles at any one time that the electrosteric barrier surrounding some particles is degraded, allowing particles in suspension to deposit onto vulnerable sites on fibre surfaces. A calculation assuming a particle diameter of 0.2 um, area per binder-ink bond site of 0.2 nm2 (which is an average of data presented by Fernandez and Hodgson, 1996), 2.5 g pigment/g pulp (i.e., approximately that found in the repulper), pigment specific density of 2, and a bond strength between each adsorption site of 360 kj/mole (which is the covalent bond energy between two carbon atoms in C 2 H 6 (Israelachvili, 1992, p. 32)), predicts that approximately 112 kJ/kg pulp is needed to break all bonds between binders and pigment particles. This is an order of magnitude less than the mixing energies put into the repulping systems. Of course, most repulping energy is expended in paper defibreing. But the calculation suggests that rupture of the binder-pigment bond is possible in the paper repulping system. 154 Interestingly, Chabot et al. (1995) reported no ink deposition (see Chapter 3, section 3.2). The mixing energy employed during their experiments is unknown but the impeller stirring speed was moderate (1200 rpm). That these researchers did not witness ink deposition may relate to a different energy dissipation pattern during their experiments compared with those found during the model repulping reported in Chapters 8 and 9 and reported by Galland and Vernac (1993a). Such a difference would be due to differing pulp consistencies (1% in the case of Chabot et al., approximately 10% in the case of Galland and Vernac, and 6% in the experiments reported here), and impeller and vessel configurations. The conflicting results may also have resulted from the different mixing times used (Chabot et al. only mixed the ink with the pulp for five minutes). 52.0 1 0 50 100 150 Time (minutes) Figure 10.2. A plot of pulp reflectance against time after Galland and Vernac (1993a). 155 The deposition/re-detachment behaviour of detached ink particles during alkaline repulping may follow a Langmuir model similar to equation [4.1]. Although no data is available with which to test directly this hypothesis, the brightness data of Galland and Vernac (1993a), shown in Figure 10.2, suggests that ink deposition may indeed follow this capture and escape behaviour. (Conversion of this data to ink content in the pulp via calculations similar to those performed by ERIC would be useful. However, sufficient information regarding the data in Figure 10.2 is not available.) Even if Langmuirian behaviour is exhibited by the detached ink, the data in Figure 10.2 (and Figures 9.10 and 9.11) are surely also manifestations of both the changes in pulp fibre morphology that accompany prolonged or intense repulping, and the link between newsprint defibring and complete ink detachment mentioned previously. Ink Detachment: Acidic Pulping It is hypothesised here that ink detachment under acidic conditions occurs if the binder/pigment matrix is broken and ink particles consisting of pigment embedded in the solid binder polymer are mechanically removed from the fibre substrate. Since the repulping step is designed to dismantle paper, rather than shear ink from the paper surface, one might expect ink detachment occurring via this mechanism to be inefficient. Indeed, Figure 9.1 shows that the pigment content in rinsed pulp produced by acidic repulping of flexographic newsprint is high relative to that obtained in alkaline repulping. A representative micrograph of rinsed fibres repulped in acidic conditions, shown in Figure 9.3, also supports the hypothesis stated above. Several ink flakes, some of which possess dimensions exceeding the diameter of the fibres, are shown detached 156 from the fibre surface. Interestingly, this sample also shows that ink trapped in the fibre lumen during laboratory printing does not escape during acidic repulping. This is consistent with the stated hypothesis since the ink inside the fibres is protected from the hydrodynamics associated with repulper mixing. If ink detachment under low pH conditions occurs as the ink film cracks and shears from the substrate due to paper disintegration and fibre flexing, then the carbon black content of rinsed fibres repulped in an acidic solution should decrease with increasing repulping intensity or power. Not surprisingly then, Figure 9.14 and Table 9.4 show that the brightness of rinsed paper produced by 24 hours of soaking in a low pH solution (i.e., no mixing) is lower than that of rinsed pulp produced by repulping paper at a power level of approximately 200 W/kg pulp (i.e., 0.36 MJ/kg pulp), for 30 minutes after a similar soaking time. The increase in brightness resulting from initiation of mixing during repulping suggests that shear is crucial to ink detachment under low pH repulping conditions. The carbon black content levels (i.e., approximately 18 wt%) corresponding to a repulping power of around 200 W/kg pulp, may indicate that substantial amounts of ink remain attached to the fibres. Higher levels in rinsed fibres repulped at 800 W/kg pulp (1.75 MJ/kg pulp) (see Figures 8.3 and 9.13) may be caused by comminution of detached pigment/binder agglomerates under the higher repulping power condition. Detached Ink Deposition and Re-detachment: Acidic Repulping Sufficiently small binder/pigment agglomerates (i.e., small enough that inertial effects are insignificant), detached from fibres during low pH repulping, will deposit onto 157 pulp fibres during repulping via attractive colloidal forces as predicted by DLVO theory. It is hypothesised here that these particles also follow modified Langmuir kinetics (see equation [4.1]). Although the effect of low pH repulping time on rinsed pulp brightness has been measured, application of a modified Langmuir model to the data presented in Figure 9.10 is not possible for two reasons. The first arises from the possibility that ink continues to detach from the fibre surface and comminute concurrently with the deposition/re-detachment phenomena. While the brightness data reflects all such behaviours, the Langmuir model does not take into account either of these factors. The second, more compelling, difficulty is that the measured dependent variable is brightness. While Popson and Jordan (1994) have shown that brightness can be converted to ppm pigment in the pulp pad, the theory for doing so is only valid if, among other criteria, the pigment concentration in the sheet is low. Since the ink concentration in the low pH pulped pads is high (approximately 18% of the total pad mass), the accuracy of an ink mass measurement obtained by conversion of brightness data is questionable. Fortunately, excellent data are available in the literature to which the modified Langmuir kinetic equation can be reasonably applied. Figure 10.3 shows brightness data collected by Ciampa (1995) that have been converted to grams ink per gram fibre and plotted against mixing time. In Ciampa's experiments, dried ink was ground, then mixed using a Morden Laboratory Slush-Maker operating at 2900 rpm, with previously de-fibred, 6% consistency newsprint, in a solution of pH 5, at a temperature of 25°C. 158 The mass ratio of ink to paper was 0.015. The following equation has been used to model Ciampa's data: de/dt = (i/T)(n0-e)(i-e)-(e/xesc) [10.1] where: e X esc time (mass of particles deposited on fibres at time t)/ (max. mass of particles than can deposit on fibres) (mass of particles initially present in suspension)/ ( max. mass of particles that can deposit onto fibres) particle deposition time constant particle lumen loading time constant particle re-detachment time constant This equation was developed by Petlicki and van de Ven (1994) for filler particle deposition and detachment from fibres in the papermachine headbox. It is identical to equation [4.1] except that the constant k-j has been replaced with a function accounting for the time scales of both particle deposition on the external surface of fibres and lumen loading of the fibres by the particles. x e s c is simply the inverse of k2. To fit equation [10.1] to the reported data, it was assumed that the ink particle diameter was 0.2 urn, that the maximum possible coverage of ink particles on fibres was 0.44 g ink/g fibre (i.e., the same as that of 0.2 urn diameter titanium dioxide particles on black spruce fibres (Middleton and Scallan, 1991), and that lumen loading of ink particle was negligible (lumen loading of titanium dioxide particles is small up to times of approximately 50 minutes). The last assumption allowed an analytic solution of equation [10.1] to be applied to Ciampa's data. Appendix 8 provides the analytic 159 solution as well as sample calculations for the conversion from sheet brightness to the ink content of the pulp sheet, and the programme used to generate the curve seen in Figure 10.3. 0 f i — i — i — i — | — i — i — i — i — | — • — i — i — i — | — i — i — i — i — | — i — i — i — i — | — i — i — i — r 0 5 10 15 20 25 30 Time (minutes) Figure 10.3. Deposition of ink particles on newsprint fibres in a solution of pH 5. The least squares fit of equation [10.1] to Ciampa's data yielded a TdeP of 18 minutes and a xesc of 8.8 minutes. Inset are Ciampa's data prior to conversion to mass ratios. The fit of the modified Langmuir kinetic equation to Ciampa's deposition data is impressive. Although the time constants are likely only estimates because lumen loading has been neglected, the trend of the data clearly shows that small flexographic 160 ink particles behave according to the competing phenomena of particle capture and escape predicted by the Langmuir equation. A comparison of Ciampa's brightness data to the brightness results shown in Figure 9.14 reveals opposing trends. That the brightness data in Figure 9.14 increase with time, even up to 90 minutes of repulping, suggests that ink detachment from newsprint is taking place throughout the repulping period and that the detached ink comminution and deposition phenomena are overwhelmed by the detachment behaviour. Whether or not initial ink detachment also dominates over deposition and re-detachment of binder/pigment agglomerates during repulping of commercial newsprint, whose ink/fibre attachment area is considerably less than that of the laboratory printed paper, is not known. Finally, a comparison of both responses to the low pH runs, shown in Figures 9.13 and 9.14, with results from a similar run performed by repulping blank newsprint with crushed Black Pearl 120 carbon black pigment is useful. Microphotographs of fibres from the compared runs are presented in Figures 9.2 and 9.19. (Interestingly, the fibres in Figure 9.19 are covered by pigment particles to such an extent that observation of lumen-loading is not possible.) In both cases the mass of pigment in the repulping system was similar, i.e., approximately 31% of the total solids mass. The carbon black content in the rinsed fibres obtained from runs performed by repulping printed paper was approximately 13% (ranging from 9 to 19%) giving an average ink retention of 42%. This is high compared to results from alkaline repulping runs but markedly less than the 161 retention of carbon black during the acidic repulping of blank paper which resulted in approximately 95% pigment retention. A difference in particle size may explain the contrasting behaviours described above. Neither electrostatic nor steric barriers inhibit attachment of either the submicrometric pigment particles or the detached binder/pigment agglomerates to pulp fibres. Both types of particles may be captured in the deep primary energy well due to van der Waals attraction and possibly the hydrophobic force. But, since the hydrodynamic force needed for subsequent detachment is proportional to the product of the squared radius of the particle and the shear rate (van de Ven, 1989), the small diameter of the pigment particles probably results in a negligible detachment rate once adhesion has occurred. In contrast, the detachment rate of the pigment/binder agglomerates, whose dimensions are several micrometres, is surely large. pH Change During Repulping Figures 8.4, 9.16 and 9.17 illustrate that solution pH decreases during alkaline soaking/repulping, and increases during acidic soaking/repulping. The figures also show that the presence of binder in the repulping vessel has a significant effect on the observed pH changes. Although not directly related to the ink detachment phenomenon, these changes are largely a result of ink behaviour during repulping. Explanations lie in the acidic nature of the pulp fibre surface groups, and the buffering effect of the binder. The dried binder consists primarily of precipitated polyacrylic acid and monoethanolamine (MEA), although some molecules of the acrylic polymer-MEA salt are probably also present. (The ammonium salt likely evaporates as 162 ammonia with the water during printing.) When immersed in an acidic solution, the acidic polymer remains insoluble, but the MEA dissociates according to the following reaction: NH 2CH 2CH 2OH + H 20 > (NH 3CH 2CH 2OH) ++ OH' MEA, which is a strong base, is therefore able to partially neutralise the strong acid (H2SO4) in the repulping solution. In addition to this reaction, the anion of the polyacrylic-MEA salt will form polyacrylic acid. This reaction will also increase the pH of the solution. The small increase in pH that occurs in the absence of binder is probably due to an unidentified salt in the pulp. In a highly basic solution, all the carboxylic acid groups in both the pulp fibres and the binder molecules dissociate. This contribution of hydrogen ions partially neutralises the base and the pH decreases. More acid is available to dissociate when the binder is present therefore the pH drops more in this case compared to the drop noted when binder is absent. Also, the reaction between the hydroxyl ion and the cation of the polyacrylic-MEA salt will lower the solution pH. Curiously, when the initial pH is 10 rather than 12, repulping pH changes more in the absence of binder than when it is present. This behaviour is likely related to the buffering nature of the MEA in the binder. At pH levels above the buffer pH, the acid will neutralise hydroxyl ions and the pH of the solution will decrease. The pH is prevented from decreasing below the buffer pH because the base (i.e., MEA) will start to contribute hydroxyl ions under these conditions. The exact pH of this basic buffer (which is approximately 7.5 in this case) is dependent on the ratio of base to cation (the pH would 163 be the pKg of the base (which is 9.4 at 25 °C) if the ratio is one). That the buffer pH (i.e., 7.5) is not reached when the initial pH is 12 is likely because the concentration of NaOH is larger than the carboxylic acid concentration. The final pH of the slurry during runs where the initial pH was 10 and binder was absent is dependent only on acid dissociation. The Implications of Model Experiments For Industrial Deinking The model experiments discussed here differed in several ways from commercial flexographic newsprint repulping. The mass ratio of ink to fibre in these experiments was an order of magnitude larger than that of commercially printed paper; the model ink composition was relatively simple; and the technique used to print the paper in preparation for the repulping experiments was different from commercial flexographic newspaper printing. Moreover, pulp consistency in commercial repulping units can be up to three times that used in the model experiments (i.e., 18% rather than 6% consistency) and paper is not usually soaked before repulping as it was in the model experiments. Also, repulping energy in commercial units, ranging from approximately 0.05 to 0.16 MJ/kg pulp (McKinney, 1995, p82), is less than that used in model repulping. Soap, hydrogen peroxide, sodium silicate and EDTA or DTPA as well as sodium hydroxide are usually added to the pulp slurry in an industrial repulper whereas only sodium hydroxide or sulphuric acid was added to the repulping vessel during most of the model experiments. Finally, industrial repulping of flexographically printed newspaper is not performed in isolation as it was in the model experiments. As mentioned previously, the fraction of flexographic newspaper in the repulper stock can 164 be as high as 10 to 15%, with the remainder of the paper stock consisting of letterpress, or more commonly, offset printed newsprint. Although it is unlikely that the differences between the model repulping experiments and the industrial experience might change ink detachment mechanisms and subsequent ink deposition and re-detachment behaviour during repulping, the success of detachment and repression of detached ink deposition may well be affected by the different independent variable levels listed above. For example, while all flexographic newsprint ink binders are acidic, the effect of pH on binder solubility will change depending on the specific binder used to disperse the ink pigment. Thus the pH ranges over which ink detachment occurs via binder dissolution may differ from ink to ink. Also, binder solubility may be affected by changes in solvent characteristics resulting from different repulping additives. The printing process and consequent mass ratio of ink to paper may affect the extent of ink detachment, particularly during acidic repulping because, as shown in Figure 7.3, the printing system can influence the degree of ink lumen loading during printing. Also, because ink is sheared from the fibrous substrate during low pH repulping, the mass ratio of printed ink to paper fibre may affect ink detachment if the increased mass increases the fibre surface coverage of printed ink. The effect of soaking newsprint prior to repulping on ink detachment efficiency is not known. However it may alter the ink-substrate bond strength that must be overcome by shear during acidic repulping. Also, it may speed binder dissolution and transport of dispersed ink pigment particles into the bulk aqueous phase during alkaline repulping. 165 Repulping energy also affects ink detachment as well as detached ink deposition. Levels lower than those used during model experiments would discourage detached ink deposition during both low and high pH repulping but would also detract from ink detachment during repulping under acidic conditions. (The relatively high brightness values of flexographic newsprint fibres repulped and floated in commercial units under acid conditions is accompanied by low pulp yield; therefore, it is likely that the removed ink remains attached to fibres which are also removed, along with the ink, during the flotation step.) A soap such as sodium stearate, added to the repulping vessel, prepares oil-based inks for flotation separation by adsorption onto the ink particle surfaces. During alkaline repulping, it is unlikely that these molecules will adsorb onto the surface of the flexographic ink particles; however the counterion concentration in the repulping unit is increased by the addition of soaps. During acidic repulping, it is possible that the soap molecules may adhere to the surface of removed flexographic ink agglomerates since the surface charge of the pigment/binder particles is relatively small. Such adsorption would inhibit ink deposition by increasing the surface charge of the particles. Hydrogen peroxide, sodium silicate, and a chelating agent such as DTPA are sometimes added to the basic repulping slurry to brighten mechanical pulp fibres whose chromophores otherwise lower pulp brightness under high pH conditions. Experimental results (Figure 8.1) show that, except for affecting solution pH, these chemicals do not have an effect on the carbon black content in the rinsed pulp. A similar situation can be expected during large-scale industrial repulping. 166 Although several experimental studies have combined flexographically printed paper with that printed by the offset technique in repulping stock, no unequivocal information on possible interactions between different ink types is available. For example, it is not known how flexographic ink binder interacts with offset ink particles during high pH repulping. Perhaps these binder molecules adsorb onto and disperse the otherwise hydrophobic offset ink particles in a manner similar to interactions with flexographic pigment particles. Fundamental studies in this area may yield interesting results. 167 Chapter 11. Conclusions, Contributions, and Recommendations 11.1 Conclusions Model experiments have shown that ink detachment during the repulping of flexographic printed paper is affected positively by repulping pH and further, that the dependence is related to the solubility behaviour of the flexographic ink binder molecules. Also, the results from investigations of repulping energy and lumen loading during repulping indicate that ink detachment mechanisms and subsequent detached ink particle behaviour are strongly dependent on the repulping solution pH. Ink Behaviour During Alkaline Repulping In basic solutions, the flexographic ink binder dissolves, thus releasing pigment particles into solution. In model repulping experiments, these particles, whose surfaces are covered with adsorbed binder molecules, possess potentials that are sufficiently high to stabilise the detached particles. During commercial repulping, however, electrostatic stabilisation of detached particles is probably thwarted by the relatively high electrolyte concentration that results from contributions made by deinking chemicals such as sodium hydroxide, sodium oleate, and sodium silicate. Under such circumstances, detached particles, which are known to be submicrometric in size, may be dispersed in the repulping aqueous phase via the steric barrier surrounding each pigment particle. Although it has been shown that ink, set inside fibres during printing, can be detached during high pH repulping, brightness measurements indicate that, even in basic solutions, mechanical energy is necessary to obtain high levels of ink detachment. 168 However, too much energy will lower pulp brightness. This decrease may be due to deposition of particles on the microcompressions, kinks, and crimps that fibres develop when exposed to high energy levels. During alkaline repulping, the ink binder components act as a buffer of approximate pH 7.5. Therefore repulping pH decreases are due to contributions of hydrogen ions from the carboxylic acid groups in the ink binder (and on the pulp fibres). Ink Behaviour During Acidic Repulping At low pH conditions, ink binder does not dissolve; thus it is hypothesised that flexographic newsprint ink detachment during repulping is achieved via mechanical shear induced at the substrate surface by mechanical energy. Some experimental results support this hypothesis but, surprisingly, the brightness of rinsed pulp declines and its pigment content increases as the repulping power in the system increases. This may be due to comminution of detached ink particles that results in deposition during repulping. Such deposition would likely follow the Langmuir model that describes the competing phenomena of capture onto and escape from fibre surfaces. Finally, the positive pH change that occurs during acidic repulping is likely due to the hydrolysis of monoethanolamine, which is contributed to the pulp slurry by the ink binder. Protonation of carboxylic groups, contributed by the binder, may further increase the solution pH under this condition. 11.2 Contributions 1. A novel method of measuring ink pigment mass in pulp has illuminated ink behaviour during paper repulping. This method involves 169 thermogravimetric analysis of pulp from which detached ink has been removed. 2. By using this novel method, verification of the effect of pH on ink detachment, independent of optical measurements such as brightness, has been presented. 3. It has been established that lumen loading of ink particles during alkaline repulping is possible. 4. The influence of binder solubility on ink detachment during alkaline repulping has been established and the assumed mechanism by which detached ink particles stabilise has been questioned by highlighting the role of counterions. 5. It has been determined that some repulping energy is required to achieve high levels of ink detachment in the model system during both alkaline and acidic repulping. However, too much energy causes a decrease in rinsed pulp brightness. 6. Hypotheses of an ink detachment mechanism and detached ink deposition/re-detachment behaviour during acidic repulping have been proposed, and experimental evidence supporting these hypotheses has been presented. 11.3 Recommendations Although this thesis has enhanced the understanding of flexographic ink behaviour during newsprint repulping, the knowledge gained by reading it is perhaps 170 overwhelmed by the questions it has raised. The proposed investigations outlined below are aimed at answering those questions that are most relevant. 1. The effect of soiling conditions on particulate soil laundering analogously suggests that printing conditions may affect the extent to which ink can be detached during subsequent newsprint deinking. Therefore it is suggested that an analysis of printing conditions on flexographic ink detachment be performed. 2. This thesis indicates that increased ink pigment deposition onto fibres during alkaline repulping occurs as repulping energy is increased. However the mechanism by which such deposition takes place is not clear. The results from an investigation of ink deposition under alkaline conditions might be directly applicable to commercial deinking systems. 3. A study of acid repulping in an ideal system would clarify the roles played by ink detachment, detached ink comminution, and detached ink deposition onto and re-detachment from fibres. 4. A fundamental study revealing the nature of the bonds between binder molecules and pigment particles would be useful. Included in such a study should be an investigation into the possibility of binder adsorption onto pulp fibres. Answers to these questions would help elucidate the role played by ink binders during flexographic newspaper repulping. 5. It would be useful to develop a quantitative method of measuring ink lumen loading. A sensitive technique could be used to study the behaviour of ink types other than flexographic newspaper ink. 171 6. Development of a model capable of simulating flexographic newsprint repulping would be extremely useful. Such a model would necessarily include such features as paper defibring and the comminution of ink particles, as well as ink detachment behaviour. 7. Mechanical energy has been shown to affect ink detachment during model repulping experiments run at low pH conditions. An investigation of the effect of energy dissipation patterns in the repulper on ink detachment would be useful. Part of such a study should include an investigation into the adhesive bond strength between printed ink and its paper substrate. 8. Finally, the reader is left with the daunting challenge to develop a direct, in situ, measuring technique capable of determining exact amounts of ink that remain attached, that detach, and that redeposit, during the repulping process. 172 References Ali, T., McLellan, F., Adiwinata, J., May, M. and Evans, T., "Functional and Performance Characteristics of Soluble Silicates in Deinking. Part I: Alkaline Deinking of Newspaper/Magazine", Pre-prints, CPPA Recyc. Res. Forum, Toronto, 21-29, (1991) Alince, B., Petlicki, J. , van de Ven, T.G.M., "Kinetics of Colloidal Particle Deposition on Pulp Fibres: 1. Deposition of Clay on Fibres of Opposite Charge", Colloids and Surfaces, 59: 265-277, (1991) Aspler, J.S., "Newsprint Requirements from Water-Based Flexography", Tappi, 70(9): 69-74, (1987). Aspler, J.S., "Tack Development in Water-Based Inks", Nordic Pulp Paper Res. J., (1): 4-7, 39, (1991). Aspler, J.S., "Newsprint Requirements for Water-Based Flexography, Part III: Influence of Newsprint Properties on Flexo Ink Holdout", J. Pulp Paper Sci., 14(3): J66-J72, (1988). Aspler, J.S., Perreault, Z.F., "Newsprint Requirements for Water-Based Flexography, Part II: Newsprint Chemistry and Water-Based Inks", J. Pulp Paper Sci., 14(1): J12-J16, (1988). Aspler, J.S., DeGrace, J.H., Beland, M.-C, Maine, C , Piquard, L, "Transfer and Setting of Water-Based Ink Part II: pH, Water Absorbency and Uncoated Paper Structure", J. Pulp Paper Sci., 19(5): J203-J208, (1993) Ackermann, C , Putz, H.-J., and Gottsching, L, "Deinkability of Waterborne Flexo Inks by Flotation", P. & P. Can., 95(8): 28-33 (1994). Bast, I., "Recovery of Computer Paper and Flexographic Paper by the Use of Special Deinking Additives", Preprints, EUCEPA Conf., Barcelona, 95-104, (1990). Ben, Y., Pelton, R. and Dorris, G.M., "Retention and Dislodgement of Flexographic Ink in a Fibrous Mat During Thickening and Displacement Washing", J. Pulp Paper Sci., 22 (11): J411-J419, (1996). Benemielis, R., "Water Flexography in the United States in the 1990's", Ink & Print, 8(4): 14-15,(1991). Bennington, C.P.J., and Seth, R.S., "Response of Pulp Fibres to Mechanical Treatment During MC Fluidization", Fundamentals of Papermakinq: Transactions of the Ninth Fundamental Research Symposium Held at Cambridge: September 1989.. Vol. 1. (ed.s, C.F. Baker and V.W. Punton), Mechanical Engineering Publications Ltd., London, 87-104,(1989). 173 Biermann, C.J., "Stock Preparation and Additives for Papermaking", Chpt. 8, Handbook of Pulping and Paper Making. 2 n d Ed.. Academic Press Inc., Sand Diego, (1996). Blom, B. E., and Feig, G., "Aqueous Printing Ink Technology For Newspaper Printing", Proc. TAPPI Intern. Printing & Graphic Arts Conf., Washington, 29-33, (1988). Boluk, M.Y., van de Ven, T.G.M., "Effects of Polyelectrolytes on Flow-Induced Deposition of Titanium Dioxide Particles onto a Cellophane Surface", Colloids and Surfaces, 46: 157-176, (1990). Borchardt, J.K., "Mechanistic Insights into De-Inking", Colloids and Surfaces A: Physicochemical and Engineering Aspects, 88: 13-25, (1994). Borchardt, J.K., Raney, K.H., Shpakkoff, P.G., Matalamaki, D.W., Denley, D.R., "Insights into Flexographic Deinking: Laboratory and Polot Scale Deinking Results", Proc, TAPPI Pulping Conf., Atlanta, 1067-1103, (1994). Borchardt, J.K., "Deinking Paper Printed With Water-Based Inks: An Overview", Chpt. 10, Paper Recycling Challenge, Vol. II: Deinking & Bleaching, (ed.s, M.R. Doshi and J.M. Dyer), Doshi & Associates Inc., Appleton, 122-138, (1997). Borhan, A., Pence, S.B., Sporer, A.H., "Mechanisms of Drying of Water-Based Inks on Bond Papers", J. Imaging Techn., 16(2): 65-69, (1990). Boughey, M.T., Duckworth, R.M., Lips, A., Smith, A.L., "Observation of Weak Primary Minima in the Interaction of Polystyrene Particles with Nylon Fibres", J. Chem. Soc. Faraday Trans. I, 74: 2200-2209, (1978). Box, G. P. E., Hunter, W. G., Hunter, J. S., Statistics for Experimenters, John Wiley and Sons Inc., New York, (1978). Brown, G.R., "Resins for Water Dilutable and Gravure Printing Inks", British Ink Maker, 23(3): 79-81,(1980). Brown, R., "Physical and Chemical Aspects of the Use of Fillers in Paper", Chpt. 11, Paper Chemistry. 2 n d Ed. (ed., J.C. Roberts), Chapman & Hall, London, 194-230, (1996). Burke, R.E., "Resins and Vehicles for Aqueous Inks", American Ink Maker, 52(2): 44, 46, 48-50, (1980). Busk, H., "How Different are Newsinks in Practice", Proc, Topical Themes in Newsprint-Printing Research, Lidingo, 143-153, (1987). Cahn, A., Lynn, J.L., "Surfactants and Detersive Systems", Encyclopedia of Chemical Technology. 3 r d Ed.. Vol. 22. (ed., Kirk-Othmer), Wiley-lnterscience Publications, New York, 332-432, (1983). 174 Carre, B., Galland, G., Vernac, Y., Suty, H., "The Effect of Hydrogen Peroxide Bleaching on Ink Detachment During Pulping and Kneading", Proceedings, TAPPI Recylc. Symp., 189-198, (1995). Carre, B., Galland, G. and Julien Saint Amand, F., "Estimation of Ink Detachment and Removal", Progress in Paper Recycling, 4(1): 80-87, (Nov.,1994). Carriere, S., Pinder, K.L., "Kneader-Type Pulpers: Lab Test", Pulp and Paper Canada, 98 (1): 76-79,(1997). Carroll, B.J., "Physical Aspects of Detergency", Colloids and Surfaces A: Physciochemical and Engineering Aspects, 74:131-167, (1993). Chabot, B., Krishnagopalan, B.A., Abubakr, S., "Flexographic Newspaper Deinking: Treatment of Wash Filtrate Effluent by Membrane Technology", Pre-prints, CPPA 4th Recyc. Res. Forum, Quebec, 233-242, (1997). Chabot, B., Daneault, C. and Dorris, G.M., "Entrapment of Water-Based Inks in Fibre Mats", J. Pulp Paper Sci, 21(9): J296-J301, (1995). Ciampa, S., "The Effect of Repulping Variables on Deinking of Flexographic Ink", M.Sc.Thesis, Western Michigan University, Kalamazoo, (1995). Clarke, B. J., "The Development of Alkali Soluble Acrylic Resins for Use in Aqueous Printing Inks", JOCCA, 73(9): 3800-384, (1990). Clewley, J.A., Bever, M. and Vautier, F.E., "Deinking Newspapers Printed by the Flexographic Method" (Translation), Wochenbl.Papierfabr 11/12: 512-515, (1990). CPPA, "Newsprint Data", Canadian Pulp and Paper Association, (1997). Danilatos, G.D., "Foundations of Environmental Scanning Electron Microscopy", Advances in Electronics and Electron Physics, 71: 109-250, (1988). Dannenberg, E.M., "Carbon Black", Encyclopedia of Chemical Technology. 3 r d Ed.. Vol. 4, (ed., Kirk-Othmer), Wiley-lnterscience Publications, New York, 632-643, (1978). Daoust, P., Undergraduate Thesis, University of British Columbia, Vancouver, (1992). Dorris, G.M., "Determination of Ink Detachment and Flotation Efficiencies From the Sidedness of Handsheets",Pre-prints, CPPA 4th Recyc. Res. Forum, Quebec, (1997). Dorris, G.M. and Nguyen, N., "Flotation of Model Inks Part II: Flexo Ink Dispersions Without Fibres", J. Pulp Paper Sci, 21(2): J55-J62, (1995). 175 Ellis, M., Hou, M.J. and Seenivasan, N., "Deinking Study on Furnish Contaminated by Water-Based Ink", Notes, TAPPI Recycling Symposium, New Orleans, 157-162, (1993). Everett, D.H., Basic Principles of Colloid Science, Royal Society of Chem., London, (1988). Ferguson, L.D., "The Role of Pulper Chemistry in Deinking", Pre-Prints, TAPPI Pulping Conf., Orlando, 793-799, (1991). Ferguson, L.D., "Deinking Chemistry: Part 1", TappiJ., 75(7): 75-83, (July, 1992). Fernandez, E.O. and Hodgson, K.T., "Steric Stability of Flexographic Newsprint Deinking Dispersions", Pre-Prints, CPPA 4th Recyc. Res. Forum, Quebec, 129-132, (1997). Fernandez, E.O. and Hodgson, K.T., "Stabilization Mechanisms of Water-Based Newsprint Inks" J. Pulp Paper Sci., 22(11): J452-J456, (1996). Galland, G. and Vernac, Y., "De-inking of Wastepaper Containing Water-Based Flexo-Printed Newsprint", P. & P. Can., 94(6): 71-75, (1993a). Galland, G. And Vernac, Y., "A New Two-Stage Deinking Process Adapted to Mixtures Containing Flexoprinted Newsprint", Revue A.T.I.P., 47(5): 154-162, (1993b). Galland, G., Vernac, Y., Carre, B., "The Advantages of Combining Neutral and Alkaline Deinking", Pre-Prints, CPPA 3rd Recyc. Res. Forum, Vancouver, 235-245, (1995). Gottsching, L., "De-Inking in Europe: Tough Competition", European Papermaker, 37-42, (Aug., 1994). Gottsching, L., and Putz, H.-J., "Current Research Activities on Waste Paper Recycling at the IFP", Proceedings, TAPPI Recycling Symposium, Boston, 207-224 (1994). Green, H.V., Fox, T.J., Scallan, A.M., "Lumen-Loaded Paper Pulp", P. & P. Can., 83(7): 39-42, (1982). Greenwald, H.L., and Luskin, L.S., "Poly(Acrylic Acid) and its Homologs", Chpt. 17 in Handbook of Water-Soluble Gums and Resins, (ed., R.C. Davidson), McGraw-Hill Book Co., New York, (1980). Gregersen, O.W., Johnsen, P.O., Helle, T., "Small-Scale Topographical Variations of Newsprint Surfaces and their Effects on Printing Ink Transfer Distribution", J. Pulp Paper Sci., 21(10): J331-J336, (1995). Greminger, G.K., Teot, A.S., Sarkare, N., "Anti-redeposition Additives Face New Opportunities", J. Am. Oil Chemists' Soc, 55:122-127, (1978). 176 Grindstaff, T.H., Patterson, H.T. and Billica, H.R., "Studies of Soiling and Detergency Part III: Detergency Experiments with Particulate Carbon Soils", Tex. Res. J., 564-573, (July, 1967). Hall, J.C., Deschamps, R.J.A., McDermott, M.R., "Immunoassays to Detect and Quantitate Herbicides in the Environment", Weed Technology, 4:226-234, (1990). Hamaker, H.C., "The London-van der Waals Attraction Between Spherical Particles", Physica, 4(10): 1058-1072, (1937). Harrison, A., "Customized Deinking Chemistry Solves Mill-Specific Problems" Pulp & Paper, 69(2): 75-77, (Feb., 1995). Heimburger, S., "Upgrading the Quality of Deinked Newsprint/Magazine Containing Flexographic Ink", Proceedings - Pan-Pacific P. and P. Tech. Conf., Tokyo, 95-100, (1992). Hicks, C.R., Fundamental Concepts in the Design of Experiments, 4 t h Ed., Saunders College Publishing, New York, (1993). Hodgson, K.T., "Deinking Considerations with Flexographic Wastepaper", Progress in Paper Recycling, 5(2): 71-74, (May, 1996). Hogg, R., Healy, T.W., Fuerstenau, D.W., "Mutual Coagulation of Colloidal Dispersions", Trans. Faraday Soc, 62: 1638-1652, (1966). Horng, A.J., "Experiences in De-Inking Waste Papers Printed with Water-Based Inks", Pre-Prints, CPPA Pacific Paper Expo., Vancouver, 139-142, (1993). Hruzewicz, J.N., Reiter, R., Newsome, R., "The Mileage of Newspaper Flexo Ink Vs Offset Ink: A Study", Flexo, 13-21, (Sept., 1991). Horwatt, S.W., Rwei, S.-P., Manas-Zloczower, I., Feke, D., "The Influence of Interstitial Liquids on the Cohesive Strength of Carbon -Black Agglomerates", Rubber and Technology, 62(5): 928-938, (1989). Hunter, R.J., Introduction to Modern Colloid Science. Oxford University Press, Oxford, (1993). Hunter, R.J., Foundations of Colloid Science Vol. I.. Clarendon Press, Oxford, (1986). iied (International Institute for Environment and Development), Towards a Sustainable Paper Cycle: Report to the World Business Council for Sustainable Development, (ed., M. Grieg-Gran), London, (1996). Imamura, T., and Tokiwa, F., "Consideration of Soil Deposition on Fabrics From the Viewpoint of Heterocoagulation", Nippon Kagaku Kaishi, (11): 2177- 2184, (1972). 177 Israelachvili, J., Intermolecular and Surface Forces. 2 Edition, Academic Press, London, (1991). Jarrehult, B., Lindquist, M.-L., Horacek, R.G., "Deinking of Wastepaper Containing Flexographic Inks", Proceedings, TAPPI Pulping Conf., Seattle, 391-405, (1989). Jaycock, M.J., and Pearson, J.L., "A Study of the Retention of Pigment during Paper Formation", J. of Colloid and Interface Sci., 55(1): 181-190, (April, 1976). Jaycock, M.J. and Swales, D.K., "The Theory of Retention", Paper Tech., 35(8): 26-33, (Oct., 1994). Jordan, B.D., and Popson, S.J., "Measuring the Concentration of Residual Ink in Recycled Newsprint", J. Pulp Paper Sci., 20(6): J161-J167, (1994). Kamiti, M. and van de Ven, T.G.M., "Kinetics of Deposition of Calcium Carbonate Particles onto Pulp Fibres", J. Pulp Paper Sci., 20(7): J199-J205, (1994). Kampe, D.J. and Sarangapani, S., "Wettability of Carbon Blacks", Proceedings of the Workshop on The Electrochemistry of Carbon, (ed.s, S. Sarangapani, J.R. Akridge, and B. Schumm) The Electrochemical Society Inc., Pennington, 179-185, (1983). Katz, S., Beatson, R.P., Scallan, A.M., "The Determination of Strong and Weak Acidic Groups in Sulfite Pulps", Svensk Papperstidn., 87(6): R48-R53, (1984). Kerekes, R.J., Soszynski, R.M., Tarn Doo, P.A., "The Flocculation of Pulp Fibres", Papermaking Raw Materials: Their Interactions with the Production Process and Their Effect on Paper Properties, Vol. 2. (ed., V.W. Punton), Mechanical Engineering Publications Ltd., London, 265-310, (Sept., 1985). Kissa, E., "Evaluation of Detergency", Chpt. 1 in Detergencv: Theory and Technology, (ed.s, W.G. Cutler and E. Kissa), Surfactant Sci. Series, Vol. 20, Marcel Dekker Inc., New York, (1987a). Kissa, E., "Kinetics and Mechanisms of Soiling Detergency", Chpt. 4 in Detergencv: Theory and Technology, (ed.s, W.G. Cutler and E. Kissa), Surfactant Sci. Series, Vol. 20, Marcel Dekker Inc., New York, (1987b). Kubler.R., "Deinking According to the Type of Printing Ink", (trans.), Wochenblatt fur Papierfabrikation, 116(10): 386-391, (1988). LaDuce, A.J., "Ink Transfer Studies in Flexography", M.Sc. Thesis, Lehigh University, (June, 1984). Lahaye, J., and Prado, G., "Mechanisms of Carbon Black Formation", Chpt. 3 in Chemistry and Physics of Carbon Vol. 14. (eds. P.L. Walker and P.A. Thrower), Marcel Dekker Inc., New York, (1978). 178 Larsson, A., "Surface Chemistry in Flotation Deinking", PTI, 28(1): 388-390, (Feb., 1987). Lindstrom, T., "Some Fundamental Chemical Aspects on Paper Forming", Fundamentals of Papermaking: Transactions of the Ninth Fundamental Research Symposium Held at Cambridge: September 1989.. Vol. 1. (ed.s, C.F. Baker and V.W. Punton), Mechanical Engineering Publications Ltd., London, 311-412, (1989). Linstrom, T., "Chemical Factors Affecting the Behavior of Fibres During Papermaking", Nordic Pulp Paper Res. J., (4): 181-192, (1992). Liphard, M., Schreck, B. and Hornfleck, K., "Interfacial Studies and Application Tests on the Flotation of Printing Inks and Fillers", Proceedings, TAPPI Pulping Conf., Toronto, 965-973, (1990). McCool, M.A., and Taylor, C.J., "Image Analysis Techniques in Recycled Fibre", Tappi J., 66(8): 69-71, (1983). McKinney, R.W.J., "Wastepaper Preparation and Contaminant Removal", Technology of Paper Recycling, (ed., R.W.J.McKinney), Blackie Academic and Professional, London, (1995). Mah, T., Reid, J. and Yau, A., "Contaminants Encountered in Washing and Flotation Deinking Processes", Proceedings, CPPA Technical. Sec. 79th Annual. Meeting, A97-A102, Montreal, (Jan., 1993). Middleton, S.R., and Scallan, A.M., "Partial Lumen Loading", Nordic Pulp Paper Res. J., (1): 204-207,231, (1993). Middleton, S.R., and Scallan, A.M., "A Kinetic Model for the Adsorption of Fillers by Pulp Fibres", J. Pulp Paper Sci., 17(4): J127-J133, (1991). Middleton, S.R., and Scallan, A.M., "Lumen-Loaded Paper Pulp: Mechanism of Filler-to-Fibre Bonding", Colloids anad Surfaces, 16: 309-322, (1985). Milanova, E., and Dorris, G.M., "Flotation of Model Inks. Part I: Experimental Methods", J. Pulp Paper Sci., 19 (5): J194-J202, (1993). Miller, M.L., and Paliwal, D.C., "The Effects of Lumen-Loading on Strength and Optical Properties of Paper", J. Pulp Paper Sci., 11(3): J84-J88, (May, 1985). Mohlin, U.-B., "Distinguishing Character of TMP", P. & P. Can., 78(12): 83-88, (Dec, 1977). Montgomery, D.C., Design and Analysis of Experiments. 4 t h Ed.. John Wiley and Sons, New York, (1997). 179 Nguyen, N.G., O'Neill, M.A., Jordan, B.D. and Dorris, G.M., "Measurement of Ink Particle Size in Deinked Stocks by Image Analysis", J.Pulp Paper Sci., 18 (5): J193-J196, (1992). Nunez, C M . , and Deatherage, G.W., "Evaluation of Innovative Ink Feed Systems for the Flexographic and Gravure Printing Industries", J. Air Waste Management Assoc., 46: 267-272, (March, 1996). O'Brien, R.W., and White, L.R., "Electrophoretic Mobility of a Spherical Colloidal Particle", J. Chem. Soc. Faraday II, 1607-1626, (1979). Okayama, T., Kumo, K., Hanada, A., Oye, R., "Behavior of Particles in Lumen-Loading", TappiJ., 43(5): 495-506, (May, 1989). Oye, R., Okayama, T., Yamazaki, Y., Yoshinaga, N., "Changes of Pulp Fibre Cell Wall by Recycling", Pre-Prints, CPPA 1st Recyc. Res. Forum, Toronto, 191-194, (1991). Papricourse '91, Pulp and Paper Research Institute of Canada's Bi-Annual Seminars for Employees and Students, (1991). Pelton, R.H., "A Model of the External Surface of Wood Pulp Fibers", Nordic Pulp Paper Res. J., (1): 113-119, (1993). Perry, R.H., Green, D.H., (ed.s), Perry's Chemical Engineers' Handbook. 6 t h Edition, McGraw-Hill Book Co., New York, (1984). Petlicki, J., and van de Ven, T.G.M., "Kinetics of Lumen Loading", J. Pulp Paper Sci., 20(12): J375-J382, (Dec, 1994). Pulp and Paper International, International Fact and Price Book 1995, Miller Freeman, Brussels, (1995). Puri, B.R., "Surface Complexes on Carbons", in Chemistry and Physics of Carbon Vol., 6, (ed., P.L. Walker), Marcel Dekker Inc., New York, 191-266, (1970). Putz, H.-J., Schaffrath, H.J. and Gottsching, L, "Deinking of Oil- and Water-Borne Printing Inks: A New Flotation De-Inking Model", P. & P. Can., 94(7): 16-21, (1993). Putz, H.-J., Torok, I., Gottsching, L., "The Deinkability of Various Printing Inks", (trans.), Das Papier, 43(1 OA): 48-57, (1989). Putz, H.-J., Gottsching, L., "Deinking by Flotation and Washing: Laboratory Comparison of Possibilities and Limitations", Proceedings, EUCEPA Conf., Ljubljana, 379-399, (Oct., 1989). 180 Rangamannar, G., Grube, G. and Karneth, A.M., "Behavior of Water-based Flexographic Inks in Newsprint Deinking", Proceedings, TAPPI Pulping Conf. Vol. 3, Boston, 933-939, (1992). Roberts, J.C.The Chemistry of Paper. Royal Society of Chemistry, Cambridge, (1996). Scallan, A.M., Katz, S., Argyropoulos, D.S., "Conductometric Titration of Cellulosic Fibres", in Cellulose and Wood - Chemistry and Technology, (ed., O. Schuerch), John Wiley, New York, 1457-1471, (1989). Schipke, S.A., "Flexo Newspaper Costs are Favorable Compared with Offset Costs", Flexo, 26-32, (June, 1991). Schwartz, A.M., "The Physical Chemistry of Detergency", in Surface and Colloid Science. Vol. 5. (ed., E. Matijevic), Wiley-lnterscience, 195-243, (1972). Scott, D.L., "What is the Appropriate Procedure for Handsheet Preparation When Evaluating Optical Properties of Recycled Pulp?", Progress in Paper Recycling, 106-109, (Aug., 1993). Scott, W.E., Abbott, J.C., Trosset, S., Properties of Paper: An Introduction. TAPPI Press, Atlanta, (1995). Scott, W.E., Principles of Wet End Chemistry, TAPPI Press, Atlanta, (1996). Sen, G., "Resins for Water-Based Inks: Color Development/Grinding", American Ink Maker, 37-40, 180, (Dec, 1987). Shaw, D.J., Introduction to Colloid and Surface Chemistry. 3 r d Ed., Butterworths, London, (1985). Smook, G.A., Handbook for Pulp and Paper Technologists. 2 n d Ed., Angus Wilde Pub., Vancouver, 1992. Steele, F.A., "The Optical Characteristics of Paper", Paper Trade J.t TAPPI Section Pages 151-156, (Mar. 21,1935). Tarn Doo, P.A., Kerekes, R.J. and Pelton, R.H., "Estimates of Maximum Hydrodynamic Shear Stresses on Fibre Surfaces in Papermaking", J. Pulp Paper Sci., 10(4): J80-J88, (July, 1984). Tessier, H.G., Carmichael, D.L., "Deinking of Newsprint with Hydrogen Peroxide", Pre-prints, CPPA Tech. Sect. Annual Meet, Montreal, (1991). Treimanis, A., "Wood Pulp Fibre Structure and Chemical Composition, their Influence on Technological Processes", Nordic Pulp Paper Res. J., (3): 146-151, (1996). Unbehend, J.E., "The 'Dynamic Retention/Drainage Jar': Increasing the Credibility of Retention Measurements", TappiJ., 60(7): 110-112, (1977). 181 Upton, B.H., Krishnagopalan, G.A., Abubakr, S., "Deinking Flexographic Newsprint: Using Ultrafiltration to Close the Water Loop", Tappi J., 80(2): 155-164, (Feb., 1997). Upton, B.H., Cullinan, H.T., Abubakr, S., Krishnagopalan, G.A., "Effects of Operating Variables on Ultrafiltration of Flexographic Pigments From Wash Deinking Filtrate", AppitaJ., 50(2): 113-120, (Mar., 1997). Upton, B.H., Krishnagopalan, G.A., Abubakr, S., "Flexographic Newsprint Deinking: Wash Filtrate Clarification by Membrane Filtration", Progress in Paper Recycling, 5(1): 36-48, (Feb., 1996). Vash, R., "Pigment Wetting and Dispersing Additives for Water-Based Coatings and Inks", in: Surface Phenomena and Additives in Water-Based Coatings and Printing Technology, (ed., M.K. Sharma), Plenum Press, New York, 139-149, (1991). van de Ven, T.G.M., "Particle Deposition on Pulp Fibres", Nordic Pulp Paper Res. J., (1): 130-134,147, (1993). van de Ven, T.G.M., "Physiocochemical and Hydrodynamic Aspects of Fines and Fillers Retention", Fundamentals of Papermaking: Transactions of the Ninth Fundamental Research Symposium Held at Cambridge: September 1989.. Vol. 1. (ed.s, C.F. Baker and V.W. Punton), Mechanical Engineering Publications Ltd., London, 471-494, (1989). van de Ven, T.G.M., Colloidal Hydrodynamics, Academic Press, London, (1989b). van de Ven, T.G.M., and Mason, "Comparison of Hydrodynamic and Colloidal Forces in Paper Machine Headboxes", Tappi J., 64:171-175 (1981). Wallstrom, E., "Water-Borne Flexo for Newspaper Printing", Proceedings, IARIGAI Research Conf., Eisenstadt, 199-210, (1988). Wasilewski, O., "Composition and Chemistry of Novel Inks Used in the Newspaper Industry and Deinking", Proceedings, TAPPI Pulping Conf., Atlanta, 25-28, (1987). Weide, W., "Deinkability of Flexoprinted Newspapers", Ifra Special Report 3.1, Darmstadt, (Oct., 1987). Werther, H.-V., "Water-Based Printing Inks: Limitations and Opportunities", Flexo, 36-39, (Sept., 1991). Yoon, R.-H., Ravishankar, S.A., "Long-Range Hydrophobic Forces between Mica Surfaces in Dodecylammonium Chloride Solutions in the Presence of Dodecanol", J. Coll. Interf. Sci., 179: 391-402, (1996a). 182 Yoon, R.-H., Ravishankar, S.A., "Long-Range Hydrophobic Forces between Mica Surfaces in Alkaline Dodecylammonium Chloride Solutions", J. Coll. Interf. Sci., 179: 403-411, (1996b). Zang, Y.-H., "A New Approach for Modelling Ink Transfer", TappiJ, 76(7): 97-103, (July, 1993). 183 Appendix 1. Details of the Ink Separation System The upper chamber of the mddj consisted of a 19.0 cm inside diameter polycarbonate cylinder of 0.5 cm thickness, a polycarbonate 1.0 cm thick top in which a 2.5 cm diameter inlet port was bored 1.5 cm from one edge and through which the stirring motor's shaft was fitted. The primary purpose of the top was to steady the stirring shaft. The lower chamber was made of a 25.0 cm diameter solid polycarbonate disk into which a funnel shape of 19.4 cm diameter and 4 cm depth was bored. A 25.2 cm diameter, 64 u.m thick piece of 30 u,m mesh size nylon screen manufactured by Spectrum Laboratory Products (Spectra/Mesh nylon filters, 21 % open area), which had been glued with epoxy resin to a 0.5 cm thick polycarbonate annulus having a width of 2.4 cm and the outside diameter as the screen, separated the upper chamber of the mddj from the lower chamber. This mesh was supported by a 0.5 cm thick, 20.2 cm diameter polycarbonate plate into which several 1.2 cm diameter holes had been bored. Figure A1.1 shows the profile of the mddj and indicates how the seal between the two chambers was accomplished. Ink pigment, detached from fibres during repulping, as well as small pulp fines to which ink may or may not have been attached, passed through the jar's screen and through the 0.64 cm diameter, threaded, exit port. Meanwhile, the rinsed pulp and its attached ink remained in the mddj. Rinse solution, contained in a 20 L polyethylene tank, was peristaltically pumped into the mddj via several feet of 0.32 cm i.d. Tygon tubing. A needle valve was attached to the mddj's threaded exit port which was located at the apex of the jar's funnel-shaped bottom. (Control of the outlet flow was necessary to ensure that the slurry 184 in the mddj remained at a constant level.) A 0.64 cm polypropylene elbow joint connected the valve to 0.64 cm i.d. Tygon tubing. The mddj filtrate flowed through the exit port, the valve, elbow joint, and tubing, and was collected in a second 20 L polyethylene tank. Locking Mechanism Detail Figure A1.1. A diagram showing the method used to prevent leakage of the mddj. 185 a) 2.0 c m — 5.0 b) 2.2 cm f — 9.3 1 Figure A1.2. mddj impeller diagrams: a) plan view; b) elevation view. A polycarbonate impeller consisting of four tilted blades, shown in Figure A1.2, slowly stirred (approximately 70 rpm and 2 W power) the pulp suspension in the mddj, 186 thus inhibiting filter-cake formation on the nylon screen (cake build-up on the nylon screen would effectively decrease the mesh size of the screen and might inhibit the removal of detached ink particles). The impeller and its 1/8 hp motor were placed such that the impeller blades did not touch the mddj screen but were as close to the screen as possible. 187 Appendix 2. Detachment Efficiency Measurements & MathCad Programs To measure the ink detachment efficiencies during each repulping experiment, samples are prepared, consistencies are measured, and the weight percent of carbon black in each sample is determined. Sample Preparation Procedures The following paragraphs describe the preparation techniques for the initial pulp samples, the rinsed mddj pulp samples and the mddj filtrate solids samples. Initial Pulp Three polycarbonate filters of 47 mm diameter, 10 jam thickness, and 1.2 u.m pore diameter (Isopore track-etched filters from Millipore) are weighed and placed in the vacuum filtration system using a similar procedure to that followed for carbon black content sample preparation (see Chapter 6). 100 mL of the diluted repulper slurry are then acidified with diluted H2SO4 to a pH of approximately 1.5. Next, this acidified slurry is poured into one of the three filtration funnels and filtered through the vacuum filtration system. This procedure is repeated twice so that three samples (a total of 300 mL) are filtered at one time. Once filtered, each sample is rinsed first with 100 mL of acetone and then 100 mL of methanol. The wet filters with filter-cakes are removed from the filtration system, and three more repulper slurry samples are produced in an identical fashion to the first three. The six samples are dried in an oven for several hours, then cooled in a dessicator and weighed with an analytical balance. Once weighed, the samples are ready for thermogravimetric analysis. 188 Rinsed Pulp Three pre-weighed, 1.2 nm pore diameter, 47 mm diameter, polycarbonate filters are placed in the vacuum filtration system. Three MDDJ rinsed pulp filter cakes are prepared with the filtration system sketched in Figure 6.3 and described in Chapter 6, section 5.1. The samples are then removed from the filtration system. Three more samples are prepared. The six wet filters with their cakes are dried for several hours in an oven, cooled in a dessicator then weighed with an analytical balance. mddj Filtrate Solids These samples are prepared in an identical manner to mddj rinsed pulp samples except that 300 mL of the filtrate (rather than 100 mL) are used to prepare each sample. Consistency Measurements Each sample of initial pulp, rinsed pulp, and mddj filtrate solids is prepared using pre-weighed polycarbonate filters. Known volumes of the diluted repulper slurry, rinsed pulp slurry, and mddj filtrate are filtered through the filters and the filters and filter-cakes are subsequently oven-dried and weighed. These gravimetric measurements enable calculation of the consistency of each of the diluted repulper slurry initially poured into the mddj, the rinsed pulp slurry, and the final mddj filtrate. Carbon Black Measurements The mddj mass balance necessitates obtaining the weight percent of carbon black in samples of the dried repulper slurry, the dried pulp that has been rinsed in the mddj, and the mddj filtrate solids. These are found by thermogravimetric analysis (TGA) of two samples of each sample type. 189 Thermogravimetric Analysis Within two weeks of preparation the samples are analysed with the Perkin-Elmer TGS-2 system thermogravimetric analyser (TGA). Thermogravimetric Analysis of Initial Pulp Once one of the six initial pulp samples has been randomly selected, the dried pulp is peeled from the 1.2 u,m pore diameter polycarbonate filter. This procedure is illustrated in Figure A2.1. Once a portion of the dried repulper slurry sample has been carefully placed in the weigh pan, the pan is returned to the microbalance. The furnace is then placed around the weigh pan and the TGA heating protocol is started. Using tweezers, a small portion of the sample (approximately 10 to 25 mg) is removed and placed into the platinum weigh pan. Care is taken to ensure that the portion to be analysed is a vertical cross-section of the entire sample so as to avoid sample misrepresentation due to possible component segregation during the filter-cake formation. In addition, it is important that the sample portion to be analysed be placed entirely within the walls of the weigh pan so that no segment of the portion breaks off during analysis and falls to the bottom of the furnace. The TGA heating protocol and data collection system is then initiated. Once the analysis is complete, the sample residue is removed from the weigh pan, the pan is cleaned with ethanol and returned to the stirrup in preparation for the next analysis. At least one more initial pulp sample is analysed in an identical fashion. 190 a) c ) b) d) Figure A2 .1 . a) The dried sample on the polycarbonate filter; b) The dried sample that has been peeled from the polycarbonate filter; c) The portion of the dried sample that will be analysed; d) The portion of sample to be analysed placed in the platinum weigh pan. 191 Thermogravimetric Analysis of Rinsed Pulp Each of two samples of dried rinsed pulp is treated in an identical fashion to the dried initial pulp samples. Like the initial pulp samples, the initial mass of each dried rinsed pulp sample is approximately 10 to 25 mg. Thermogravimetric Analysis of mddj Filtrate Solids Unlike both the repulper slurry and rinsed mddj slurry samples, the entire sample of dried mddj filtrate solids is analysed with the TGA. Removing the filtrate solids from the polycarbonate filter is difficult. Each filter is held, upside-down, over a clean sheet of white paper. The filter is then crinkled rapidly such that flakes of the dried mddj filtrate solids fall off the filter and onto the white paper. Once most of the sample has been removed from the polycarbonate filter, the white paper is folded then tapped from underneath such that the sample collects at the paper crease. The sample is then carefully "funnelled" along the paper fold and into the weigh pan. This procedure is illustrated in Figure A2.2. Approximately 2 to 15 mg of sample are collected in this fashion. When the sample has been successfully placed in the weigh pan, the pan is returned to the analyser and the heating protocol is initiated. Carbon Black Weight Percent Calculation The computer associated with the TGA stores temperature versus percent of original mass of each analysed sample in a user specified ASCII file. On completion of each thermogravimetric analysis, the data file is accessed by a MathCad programme which calculates the percent of carbon black in each sample type. A second MathCad pro-192 gramme calculates ink detachment efficiency from the data. Both programmes and sample calculations are attached to this appendix. 193 a) Figure A2.2. a) The mddj filtrate solids on the polycarbonate filter and the white paper sheet; b) Crinkling the filter over the white paper; c) Filling the crucible with the mddj filtrate solids flakes. 1 9 4 An Example of the MathCad Program Designed to Analyse the Thermogravimetric Results: Experiment U 96: Analysis of Pulper Slurry TGA Results December 30, 1995 Read the data files containing subsample TGA info and find the average carbon black weight % in the dry samples at the end of 30 minutes at 500oC: U : - READPRN( e96p2) WT1 is the matrix containing the last 10 wt% data taken at 105 oC: W T 1 U :=M 3 S 3 W T V = M 3 6 3 W T l ^ M ^ , W T V = M 3 S 3 WT1 4 :=M3 y 3 W T l ^ - M ^ , W T 1 < : - M 4 1 J W T 1 7 : - M 4 2 J W T l g : - M 4 3 3 WT1, : - M < 4 3 wl is the matrix containing the last 10 weight data taken at 105 oC: W 1 0 : = M 3 J 1 W 1 1 : = M 3 < ! 1 W l ^ M ^ , W 1 , : = M , 8 - | W l ^ M , , , W l :=M Wl :=M Wl :=M Wl :=M Wl :=M 3 40.1 V 41.1 7 ll42.l 8 43.1 9 44.1 WT2 is the matrix containing the last 10 data recorded at 500 oC: w l V = M , « * . . , W T V = M , , o . , w i V ^ . . ^ ^ , ! W T V = M „ , . 3 AV105 :-mean(WTl) AV105 = 99.335 AV500 :-mean(WT2) AV500 = 14.357 AVW105 := mcan(Wl) AVW105 = 15.076 And. the weight percent of carbon black remaining in the bone (Note expt#70 showed that an dry sample at 500oC for this data set is: average of 93.17% of the acetone/methanol treated ink CB o = 15.513 sample remained after the TGA runs.) CBo:=if1 A V 1 0 5 r 0 . ( ^ J " L | . 0 \AV105 .9317 Also, all the data from the file are read into vectors for plotting purposes: temol l : = M < 2 > wtl 1 :=— 100 timel 1 :=— AV105 ~ 3 M.= READPRN(e96p4) WT1 is the matrix containing the last 10 wt% data taken at 105 oC: WT1 0 :=M 3 5 3 WT1, :=M36 3 WT1 2 :=M J 7 >, WT13 :=M,8>3 WT14 ^ V = M 4 0 . 3 ^ V = M 4 , . 3 ^ V = M 4 2 . 3 W V = M 4 3 . 3 ^ V ^ O 195 w l is the matr ix con t a in ing the last 1 0 weight da t a t a k e n at 1 0 5 o C : W 1« : = M,,., W V = M V , , W V = M , 7 . , WV=M,«., W 1 4 : = M , , . , w V = M 4 < u w l« : s M«.i W 1 7 : = M « . » W 1 « : = M 4 3 . . m r \ , W T 2 is t he matr ix con t a in ing the last 1 0 da ta r ecorded at 5 0 0 o C : W ^ o = M m . 3 ^ 1 = ^ 5 . 3 ^ 2 = M U , . 3 ^ 3 = M » 8 7 . 3 ^ 4 = M ,88.3 W T 2 5 : = M 1 » . S W T 2 « : = M 1 9 0 . 1 ^ ^ m ^ r V i ^ ^ 1 9 3 . 3 A V 1 0 5 :=mean(WTl) A V 1 0 5 - 9 7 . 2 2 7 A V 5 0 0 :=roean(wT2) A V 5 0 0 - 12.43 A V W 1 0 5 :=mean(Wl) A V W 1 0 5 = 15.183 A n d . t he we igh t pe rcen t of c a r b o n b l a c k r ema in ing in t he b o n e (Note expt#70 s h o w e d that a n d r y s a m p l e at 5 0 0 o C for this d a t a se t i s : a v e r a g e of 9 3 . 1 7 % of t he a c e t o n e / m e t h a n o l t rea ted i nk CD, :=if| A V 1 0 5 , 0 . ^ ™ - ! ~ 1 \ A V 1 0 5 .9317/ CDj - 1 3 . 7 2 2 s a m p l e r e m a i n e d after t he T G A runs . ) A l s o , a l l the da t a f rom the file a re r ead into vectors for plotting p u r p o s e s : M<3> M<o> tempi 2 : = M < 2 > wtl 2 : = — 100 t imel_2:=—— v - ~ A V 1 0 5 3 A n d , t he ave rage c a r b o n b l a c k w t% in the bone dry s a m p l e s is : mcan(CB) -=14.617 stdcv(CB) - 0 . 8 9 5 i := 1.. 196 j := 1.. 195 loo K— W t % 100 150 200 250 300 350 400 450 500 Tempera tu re (oC) 196 An Example of the MathCad Program Designed to Calculate Ink Detachment Efficiencies Experiment # 96 Calculations and Results D e c e m b e r 3 0 , 1 9 9 5 Overall Mass Balance Calculations: R e a d i n the d i lu ted P u l p e r S lu r ry c o n s i s t e n c y m e a s u r e m e n t s ( m g / 1 0 0 m l ) : (note that th i s i s t h e first d i lu t ion of t h e pu lpe r s lurry (i.e.. before c l ea r so lu t i on is a d d e d t o t he M D D J . ) ) -351.7 P 3 :-352.1 p , = 344.5 P 4 := 336.3 = 337.0 P 5 = 346.6 m«an(P) =344.7 stdev(D =6.288 R e a d i n the M D D J S l u r r y c o n s i s t e n c y m e a s u r e m e n t s ( m g / 1 0 0 m l ) : M V = 1 2 5 - 7 M S M V - 1 2 1 - 6 M s " M V = 122.8 = 125.5 = 130.8 = 131.1 mean(MS) = 126.25 stdev(MS) = 3.62 R e a d in the M D D J Filtrate c o n s i s t e n c y m e a s u r e m e n t s ( m g / 3 0 0 m l ) : F o :=20.3 F 3 :=20.3 F ! :-20.3 F 4 -20.2 F2 = 20.2 F 5 = 20.2 mean(F) -20.25 stdev(F) =0.05 R e a d i n the v o l u m e s (in ml) of the pu lpe r s lurry (1st dilution), the M D D J s lur ry a n d the M D D J filtrate: Pvol := 1000 MSvol := 1950 Fvol := 14744 Tota l M a s s Into the M D D J (in mg . s ) : Irtmass :=Pvol-m e a n^ P^ 100 Tota l M a s s O u t of the M D D J (in mg . s ) : M a s s in the R i n s e d M D D J Slur ry : iss :-MSvol- m C 100 Inmass - 3.447-103 MSoutma  - S v o l - m C a n ( M S ) MSoutmass - 2.462-103 M a s s in the M D D J Filtrate: Foulraass : - F v o l - m e 3 n ^ F ^ 300 Foutmass = 995.22 Oulmass :-MSoutraass +• Fouunass Outmass =3.457-103 O v e r a l l M a s s B a l a n c e Er ro r (%): _ „ Inmass - Outmass , . . Overall_error :=—:—: ——100 Inmass Overull_error —0.293 Carbon Black Mass Balance Calculations: R e a d in the T G S results for % c a r b o n b lack in the pu lpe r so l ids , the rinsed M D D J s lur ry so l id s a n d the M D D J filtrate so l ids : Tcb := 14.62 MScb :-0.33 Feb := 51.92 C a l c u l a t e t he m a s s of c a r b o n b l a c k enter ing the M D D J (in mg . s ) : Pcb mCB:=Inmass._ TnCR-503.951 C a l c u a l t e the m a s s of c a r b o n b l a c k in the rinsed M D D J s lurry (in m g . s ) : MSoutCB :=MSoutmass MSoutCB - 8.124 100 C a l c u l a t e the m a s s of c a r b o n b lack in the M D D J filtrate (in mg . s ) : FoutCB :-Foulmass— FoutCB -516.718 100 C a l c u l a t e the total m a s s of c a r b o n b l a c k exit ing the M D D J (in m g . s ) : OutCB := MSoutCB + FoutCB OutCB = 524.842 C a r b o n B l a c k M a s s B a l a n c e Er ro r (%): ™ . InCB - OutCB CBerror:- -100 TnOB CBeiror—4.145 C a l c u l a t e the c a r b o n b l a c k D e t a c h m e n t Eff iciency (%): D e t F . f f : > C B - M S o u t C B l o o InCB DetF.ff = 98.388 Appendix 3. Experimental Data Table 3.1: TGA Sensitivity Results Mass % Carbon Black Determined by TGA Actual Mass % Carbon Black Determined by Microbalance (TGA Mass % Carbon Black) - (Actual Carbon Black) Absolute Percent Error (on the basis of actual carbon black content) 9.14 9.27 -0.13 |-1.4| 5.04 5.16 -0.12 |-2.3| 1.71 1.67 +0.04 |2.4| 0.82 0.81 +0.01 |1.2| 0.21 0.0 +0.21 — 3.09 2.87 +0.22 |7.7| -0.10 0.14 -0.24 |-171.4| 0.61 0.71 -0.10 |-14.1| 0.3 0.0 +0.3 — 0.26 0.28 -0.02 |7.1| 0.8 0.20 +0.08 |40.0| Sample Calculation of C.B. Content: CB Content = 100((wt5oooc/wt1o5oc)avg - (Ash Fraction in the Sample)) where: n V W t j j s^(n-9) «, wtios oc = . wti is the i sample weight recorded by the TGA at 105 C 10 and n is the last sample weight recorded at 105 °C. Similarly, 199 m+9 wt5oo oc = > ' i s t n e ^ sample weight recorded by the TGA at 500 °C, and m is the first weight recorded at 500 °C. The ash fraction in the sample is Ash Fraction = 0.007(1 - wt 5oooc/wti 05oc)/(1 - 0.007). Thus, the carbon black content in the rinsed pulp from experiment 120.4 is CB Content = 100(0.0397 - (0.007(1 - 0.0397))/0.993) = 3.29 Table A3.2 Carbon Black Content vs Repulping pH RUN NO. POST-REPULPING pH C.B. IN RINSED PULP (wt%) 94 6.7 2.9 95 9.4 2.3 96 9.8 1.2 97 6.1 1.7 98 5.2 9.3 100 7.5 2.1 101 2.0 12.8 102 5.7 2.4 105 3.5 13.1 106 2.5 17.5 108 5.0 12.6 110 2.9 18.6 118.1 2.2 19.6 118.5 2.0 16.8 118.3 7.5 1.9 118.4 10.0 1.7 120.1 2.4 18.8 120.2 2.6 19.0 120.6 2.4 13.0 121.1 9.0 1.7 121.4 9.0 1.1 200 Table A3.3 Ink Absorbance/Transmittance Data (Expt. 115) PH 1% INK BINDER SOLUTION (by weight) ABSORBANCE OF 470 nm LIGHT 0.5% INK BINDER SOLUTION LIGHT TRANSMITTANCE after Aspler et al (1988) 2.0 0.438 — 2.3 0.246 — 2.5 — 15 2.6 0.259 — 2.9 0.122 — 3.5 — 17 3.6 0.080 — 3.6 0.121 — 4.5 — 25 4.9 0.078 — 5.5 — 40 6.2 0.052 — 6.5 — 65 6.8 0.062 — 7.5 — 90 7.8 0.002 — 8.5 — 90 10.5 0.000 — 201 Table A3.4: Model Ink (never-dried) Median Particle Size at Different pH Levels (Expt. 42) PH MEDIAN INK PARTICLE SIZE (um) 3 11.0 4 5.0 6 2.0 6 0.28 8 0.19 10 0.19 10 0.19 13 0.19 Table A3.5: Ink Particle C-Potentials at Different pH Levels. (Expt. 112) PH MODEL INK (Black Pearl 120 + Flexo "F") [st. dev.] (mV) MODEL INK in 0.01 M NaCI (Black Pearl 120 + Flexo "F") [st. dev.] (mV) UNIDENTIFIED COMMERCIAL INK (after Liphard et al (1990)) (mV)* MODEL INK (Black Peart 420 + Flexo "F") (after Fernandez and Hodgson (1996)) (mV) MODEL INK (Regal 99 + Flexo "F") (after Fernandez and Hodgson (1996)) (mV) 1 — -7 2 +3.7 [1.2] +3.4 [0.9] — -5 -22 3 -5.4 [0.2] -11.0 [0.2] -15 -20 4 — -30 -40 5 -45.9 [0.4] -29.0 [1.3] -28 -50 -55 6 -40 -58 -60 7 -47.9 [3.1] -41.4 [1.1] -42 -58 -55 8 -45 -67 -75 9 -49.2 [1.1] -38.9 [0.1] -55 -70 -67 10 -60 -57 -72 11 -42.1 [0.5] -38.7 [0.3] — — — 12 — — -35 -40 *Zeta potentials measured by Liphard et al (1990) were obtained from electrophoretic mobility measurements and transposed to shear surface potentials via Smoluchowski's equation. Potentials reported by Fernandez and Hodgson (1996) and those pertaining to the model ink used in experiments presented in this thesis were transformed from mobility measurements via the method outlined by O'Brien and White (1978). 202 Table A3.6 Carbon Black Pigment ^-Potentials at Different pH Levels (Expt. 112) PH BLACK PEARL 120 in 0.01 M NaCl [st. dev.] (mV) UNIDENTIFIED CARBON BLACK (after Liphard et al (1990)) (mV)* BLACK PEARL 420 (after Fernandez and Hodgson (1996)) (mV) REGAL 99 (after Fernandez and Hodgson (1996)) (mV) 2 — — +12 +10 3 — — 0 -10 4 — — --28 -13 5 -18.2 [0.2] -10 -52 -45 6 — — -55 -30 7 -22.1 [0.8] -12 -52 -36 8 — — -55 -34 9 -23.8 [0.4] -35 -60 -34 10 — -40 -54 -34 11 -20.0 [--] — — — 12 — — -20 -28 *Zeta potentials measured by Liphard et al (1990) were obtained from electrophoretic mobility measurements and transposed to shear surface potentials via Smoluchowski's equation. Potentials reported by Fernandez and Hodgson (1996) and those pertaining to the model ink used in experiments presented in this thesis were transformed from mobility measurements via a method outlined by O'Brien and White (1978). Table A3.7 Carbon Black Content and Pulp Brightness of Rinsed Pulp from High pH Runs at Different Repulping Times. RUN NO. REPULPING TIME (min.) REPULPING ENERGY (MJ/kg pulp) C.B. CONTENT (wt%) BRIGHTNESS (iso) [st. dev.] 120.4 30 1.0 3.3 42.7 [1.8] 120.5 90 2.6 4.8 30.4 [3.4] 120.3 90 2.2 2.4 27.3 [0.9] 127.1 90 2.1 2.0 35.4 [0.9] 100 30 0.14 2.1 39.2 [3.2] 122.8 60 0.86 0.0 39.3 [1.0] 122.7 60 0.86 0.8 40.2 [1.9] 122.1 30 0.71 0.0 41.2 [1.5] 203 RUN NO. REPULPING TIME (min.) REPULPING ENERGY (MJ/kg pulp) C.B. CONTENT (wt%) BRIGHTNESS (iso) [st. dev.] 122.3 30 0.71 0.0 39.4 [0.4] 127.4 90 2.1 — 32.9 [0.5] 128.6 0 0 — 19.4 [1.7] 128.9 0 0 — — 128.14 0 0 — 12.2 [1.7] 118.3 30 0.18 1.9 — Table A3.8 Carbon Black Content and Brightness if Rinsed Pulp from Low pH Runs at Different Repulping" fimes RUN NO. REPULPING TIME (min.) REPULPING ENERGY (MJ/kg pulp) C.B. CONTENT (wt%) BRIGHTNESS (iso) [st. dev.] 120.2 90 2.2 19.0 24.6 [5.0] 120.6 90 3.0 13.0 27.1 [7.7] 127.2 90 2.6 17.5 27.1 [3.5] 105 30 0.14 13.1 19.4 [5.1] 101 30 0.14 12.8 24.5 [2.3] 110 30 .20 18.6 20.7 [5.6] 127.3 90 2.6 — 26.6 [2.5] 128.7 0 0 — 4.5 [0.4] 128.10 0 0 — 4.4 [0.2] 204 Table A3.9 Carbon Black Content and Brightness of Rinsed Pulp From High pH Runs at Different Repulping Intensities RUN NO. REPULPING POWER (W/kg pulp) REPULPING ENERGY (MJ/kg pulp) RINSED PULP CARBON BLACK CONTENT (wt%) RINSED PULP BRIGHTNESS [st. dev.] (iso) 125.3 787.4 2.8 1.4 33.2 [0.61] 125.5 781.1 2.8 1.6 35.8 [1.3] 100 76.0 0.14 2.1 39.2 [3.2] 118.3 99.0 0.18 1.9 — 121.1 250.0 0.45 1.7 — 121.4 83.0 0.15 1.1 — 122.1 392.6 0.71 0.0 41.2 [1.5] 122.3 392.6 0.71 0.0 39.4 [0.5] 128.6 0.0 0.0 — 19.4 [1.7] 128.14 0.0 0.0 — 12.2 [1.7] Table A3.10 Carbon Black Content and Brightness of Rinsed Pulp from Low pH Runs at RUN NO. REPULPING POWER (W/kg pulp) REPULPING ENERGY (MJ/kg pulp) RINSED PULP CARBON BLACK CONTENT (wt%) RINSED PULP BRIGHTNESS [st. dev.] (iso) 125.1 787.4 1.4 26.9 17.4 [2.6] 125.4 793.7 2.1 24.3 19.3 [5.5] 118.5 300.0 0.54 16.8 — 118.1 297.0 0.53 19.6 — 101 83.0 0.15 12.8 24.5 [2.3] 105 87.6 0.16 13.1 19.4 [5.1] 106 269.6 0.49 17.5 19.9 [5.4] 110 109.7 0.20 18.6 20.7 [5.6] 128.7 0.0 0.0 — 4.5 [0.4] 128.10 0.0 0.0 — 4.4 [0.2] 205 Table A3.1 pH Change During Repulping in the Absence of Ink Binder (Expt. 117) PULPING pH PULPING TIME (min.) TRIAL # 1 TRIAL #2 TRIAL #3 TRIAL # 4 TRIAL # 5 24 x (-60) 2.0 2.0 10.0 12.1 12.1 0 2.0 2.0 7.1 10.4 10.4 0.5 2.1 2.1 7.0 10.4 — 1 2.1 2.2 7.0 10.4 10.4 2 — 2.2 — — — 3 2.1 2.2 7.1 10.3 10.4 4 2.1 — — — — 5 — 2.2 7.1 10.4 10.4 10 2.1 2.2 7.1 — — 15 — — — 10.3 10.4 20 2.1 2.1 7.0 — — 30 2.1 2.1 7.0 10.3 10.4 Table A3.12 pH Change During Repulping in the Presence of Ink Binder (Expt. 117) PULPING pH PULPING TIME (min.) TRIAL # 1 TRIAL #2 TRIAL # 3 TRIAL # 4 24 x (-60) 2.0 10.0 10.0 12.1 0 2.9 7.4 7.5 9.6 0.5 3.1 7.4 7.4 9.5 1 3.1 7.4 7.5 9.4 2 7.4 3 3.1 7.4 7.5 9.3 4 5 3.2 7.4 7.5 9.4 10 3.2 7.4 7.5 15 — — 9.3 20 3.2 7.5 7.7 — 30 3.2 7.5 7.6 9.4 Table A3.13 Carbon Black Content of Fibres Repulped with and without Sodium Hydroxide and Bleaching Chemicals RUN NO. NaOH (g/L) H 2 0 2 (g/L) Na2SiO 3 (g/L) D.T.P.A. (g/L) POST-PULPING pH RINSED PULP C.B. CONTENT (wt%) 74 0.0 0.0 0.0 0.0 5.1 18.8 77 1.0 1.5 2.0 0.2 11.8 0.6 78 0.5 0.75 1.0 0.1 11.0 1.2 79 0.0 1.5 2.0 0.2 10.3 0.5 81 0.0 0.0 0.0 0.0 5.6 13.5 82 0.0 0.0 0.0 0.0 5.9 12.8 85 -1.5 0.0 0.0 0.0 9.8 1.2 91 -1.5 0.0 0.0 0.0 10.7 1.5 Appendix 4. Comparison of Model Ink to Commercial Ink The print characteristics of the laboratory manufactured ink have been compared to those of a commercial flexographic newsprint ink composed of similar binders and pigment. The results of this comparison are shown in Table A4.1. Table A4.1. Comparison of a Commercial Flexographic Ink with the Laboratory Manufactured Ink Commercial Proprietary Ink Laboratory Ink Primary Particle Size of Carbon Black Pigment (um)* 0.084 0.13 Weight Percent of Carbon Black in the Dried Ink ? 675 Print Density (measured with shell cup #2, vise. = 20 sec.)* 0.7 0.55 Example of Print (using a Speedflex 13 inch Commercial Web Press) H El •After Daoust (1992) Clearly, the laboratory ink has inferior print qualities compared with those of the proprietary flexo ink. Its relatively low print density is likely the result of its course pigment particle size and low carbon black concentration (Busk, 1987). Mottling, evident on the print sample of the laboratory ink, is likely due to ink drying on the printing 'plate' (Aspler and Perreault, 1988; Busk, 1987). The relationship between print characteristics and deinkability is unknown because the role played by all ink components during deinking is not understood. During the course of the investigation presented in this thesis it has been assumed that 208 the laboratory ink is sufficiently similar to commercial inks to emulate their deinking characteristics. 209 Appendix 5. The Light Absorbance of Ink Binder Solutions Light absorbance of a 20 °C ink binder solution (1.8 wt %) under different pH conditions was tested at various wavelengths with the absorbance at pH 9.6 set to zero. The results are shown in Figure A5.1. The wavelength showing the greatest contrast was 470 nm; therefore all subsequent absorbance data was obtained at this light wavelength. 2.5 2 <D S 1.5 O (0 < 1 0.5 0 A ft A * • o ft - 4 * — • + 370 nm • 400 nm A 470 nm X 601 nm X 801 nm O 997 nm 0 4 6 PH 8 10 Figure A5.1. Light absorbance at different pH levels of a solution of a 1.8 wt % solution of ink binder. Next, the absorbance of 470 nm light in a 30 °C solution of 1.8 wt% binder was tested at different pH levels and compared to similar data collected at 20 °C. The results, shown in Figure A5.2, indicate that the absorbance response is not affected within this 10 degree temperature range; therefore, all subsequent measurements were made at 20 °C. 210 2.5 5 2 | g 1.5 o 1 c CO •e ° 0.5 < I -1 ---• • T = 30°C • T = 20°C • 6 PH 8 10 12 Figure A5.2. Absorbance of 470 nm light a binder solution (1.8 wt %) at 20 and 30 °C. Thermogravimetric analysis shows that the binder content in the dried model ink used in the repulping experiments is about 25 %. Therefore, since approximately 4 grams of ink is printed on blank newspaper in preparation of each repulping experiment, the concentration of binder in the pulping slurry is approximately 0.5 wt %. This is considerably smaller than the binder concentration used in light absorbance tests presented above. Therefore absorbance tests were performed using dilute solutions of the ink binder. The results, presented in Figure A5.3, show that, regardless of the binder concentration, the solution absorbance increases as the pH of the solution is lowered, and that the initial increase occurs around a pH level of between 6.5 and 7. The Figure also shows that absorbance behaviour is not a linear function of binder concentration. This is reasonable since a high concentration of colloidal binder particles will obscure subsequently precipitated particles. Therefore a dilute concentration responds better than a concentrated solution to the precipitation behaviour of the binder. 211 It is for this reason that the most dilute solution (0.1 wt%) was compared with carbon black content data in Figure 9.4. 2.5 o r~-Z 1.5 as <D O c 1 JO 1 0.5 X • X AA X • 1.8 wt% Sorn X 0.9 wt% Sol'n A 0.1 wt% Sol'n x 6 pH 10 Uoo % i-300 % 200 % L100% 12 Figure A5.3. A plot of the light absorbance of different concentrations of ink binder against solution pH at 20 °C. The scale on the right abscissa indicates that used to present the data seen in Figure 9.4. While light absorbance is an indirect method of measuring binder solubility, the tests presented here indicate that ink binder in the model repulping system is in solution during repulping experiments whose pH level is above approximately 7 and precipitates during tests where the pH level is below about 6.5. 212 Appendix 6. Effects Table Sample Calculations The following calculations show the method by which the effects tables are generated. All tables in this appendix were calculated by JASS software from Joiner Assoc. Inc., Madison, Wl., U.S.A.. Factorial results from experiments that varied pH and time, where the dependent variable is rinsed pulp carbon black content, are used in all sample calculations and are shown in Figure 8.2. Main Effects Effectdme = ((16.5 -14.8) + (3.1 - 0.7))/2 = 2.1 EffectpH = ((3.1 -16.5) + (0.7 -14.8))/2 = -13.8 Interaction Effect EffecWraction = ((3.1 -16.5) - (0.7 -14.8))/2 = 0.4 Standard Error The standard error of effects is calculated from the estimate of the response variance. In the case of the experimental results in Figure 8.2, the variance values of all replicates are pooled and the resulting data standard deviation is 2.5. The following calculations show how this value was obtained: Number of replicates at the ith experimental condition = n, Degrees of Freedom at the ith experimental condition = v, Average C.B. content response at the ith experimental condition = CBavgj Estimated variance of results obtained at the ith experimental condition = sf 213 /?! = 3 v^ = 2 CBavgi = 16.5 s2i = (16.5 -17.5)2 + (16.5 -13.0)2 + (16.5 -19.0)2 )/2 = 9.8 i 2 3 4 3 3 3 2 2 2 CBavg. 14.8 3.1 0.7 S2; 10.7 2.3 1.5 Therefore the pooled variance of the data = s2p = (2(9.8) + 2(10.7) + 2(2.3) + 2(1.5))/(4x2) = 6.04 Box et al. (1978) show that, if the experimental design is balanced, an effect variance {Veffect) can be estimated as: V - - i s 2 w effect ~ 0 , where N = the total number of two level experiments. Thus, Veffect = 4(6.04)/12 = 2.01, and the standard error of each effect is (2.01 ) 0 5 = 1.4. The standard error of the average effect is s/(N)° 5 = 0.7. 214 If a design contains unbalanced data (i.e., n differs with experimental conditions), then the variance of each effect is Venect = (1/n+ + 1/n")s2 where n + and n" are the total number of runs performed at the high and the low levels respectively. 215 Appendix 7. Total Interaction Energy Calculations The total interaction energy (Vj) between a single ink particle and one pulp fibre, as a function of the distance between the surfaces of the two solids, has been estimated by modelling each solid as a sphere and plate respectively. A sample calculation at pH 10 is provided below as well as the data used for all other calculations. The total interaction energy is given by: V T = V d + V e where is the attractive London - van der Waals interaction energy between the two surfaces and V e is the repulsive electrical double layer interaction energy between the two surfaces. Estimating Ve Hogg et al. (1966) developed the following general equation to model the repulsive energy between two similar or dissimilar spherical particles: V e = (aia2£/4kT(ai +a2)) x {(29ie2/(e12 + 02 ))(ln[(1 + exp(-KH))/(1 - exp(-KH))] + ln[1 - exp(-2KH)]} [A7.1 ] where ai and a 2 = the radii of the spheres 8 = the permittivity of the suspending medium 0! = the electric potential at the surface of sphere 1 02 = the electric potential at the surface of sphere 2 H = the distance between the two surfaces K = the Debye reciprocal length k = Boltzmann's constant (1.381 x 10~23 J K"1) T = Temperature (298 K) 216 Equation [A7.1] is exact only for surface potentials less than 25 mV and for solution conditions such that the double layer thickness is small compared to the particle size. However, the equation gives a good approximation of the interaction for surface potentials less than 60 mV. The approximation is mediocre when the product of K and a is greater than 5 and it is quite good for Ka > 10. Equation [A7.1] was used to model the repulsive energy between an ink particle and a pulp fibre. In this case, the limit 3 2 - ^ ° ° was employed in a similar fashion to Middleton and Scallan (1985) who modelled the interaction between a filler particle and a pulp fibre. The following values were used for the current case: a = 0.1 x 10"6 m (from Figure 9.6) e = 6.95 x 10" 1 0CV 1m" 1 (the permittivity of water) 9i = -0.05 V at pH 10 (from Figure 9.7) 02 = -0.035 V (Jaycock and Pearson, 1976) K = 0.329 x 1010(cz2)0'5 m"1 for an aqueous solution of a symmetrical electrolyte at 25 °C (Shaw, 1985, p. 154). z is the counterion charge number, c is the counterion concentration. Three cases of counterion concentrations in the repulper were tested. The first case assumed that the counterions were contributed by the sodium hydroxide added to increase the repulping pH to 10 (i.e., 0.01 M, which corresponded to an initial pH of 12). In this case, the binder contributes no counterions because monoethanolamine does not react with water to contribute an hydroxyl ion (see page 162). The second case illustrates the counterion critical deposition concentration. This was found to be 0.04 M. 217 The final counterion concentration tested here was estimated to be that present in a commercial repulping unit. In this case, 0.018 M Na+ is contributed by the silicate, 0.032 M by the sodium hydroxide, and 0.005 M by the collector. Estimating The estimates of the attractive interaction energy between the sphere and plate were found using Hamaker's equation: Vrj = -(A123/6kT)((2a(H + a))/(H(H + 2a)) - ln((H + 2a)/H)) [A7.2] where A123 is the Hamaker's constant for the sphere/plate/aqueous solution system. A 1 2 3 was estimated using a similar method to that used by Middleton and Scallan (1985) as: A i 2 3 = (Afibre/water X Apjgment/water) = 4.5x10 J where Aflbre/water = 6.2 x 10" J (after Jaycock and Pearson, 1976), and A p i g m e n t f w a t e r = 3.3 x 10"20 J (after Horwatt et al., 1989). The following Mathcad program was used to generate the curves seen in Figure 10.1. 218 // MicroMath Scientist Model File // A model that estimates the free interaction energy of dispersion and coulombic forces between // a sphere and a plate. The Hamaker model is used to estimate van der Waal's force. V electrical // is estimated using Hogg et al's (1966) equation and assuming that Rf»Rp (Middleton & Scallan ( //(1985)) IndVars: H DepVars: Va.Vr.VtC, L, N, M.KT Params:Rp, R, Ap, At, A, Zp, Zf, K, Er, E, B, J Rp=1*10A(-7) R=Rp Ap=3.3*10A(-20) Af=6.2*10 -^20) A=4.5*10A(-20) J=2*R L=H+J Va=-(A/6)*((J*(H+R)/(H*L))-LN(L/H)) Zp=-0.050 Zf=-0.035 K=6.87*10A(8) Er=78.5 E=Er*8.85*10A(-12) B=ZpA2+ZfA2 C=EXP(-K*H) M=2*Zp*Zf N=EXP(-2*K*H) Vr=(Er*Rp*((1.054*10A(-5))A2)*B/4)*((M/B)*LN((1 +C)/(1 -C))+LN(1 -N)) Vt=Va+Vr KT=Vt/(4.1*10A(-21)) *** Appendix 8. Calculations to Fit Kinetic Data to the Modified Langmuir Equation (Equation [10.1]) Equation [10.1] has been fitted to data found in the literature. The data set consists of washed pulp brightness measurements taken after repulping for various time lengths. The first step in the fitting process involved translating the brightness data to mass ratios of ink to pulp. Next, an analytic solution to Equation [10.1], which assumed negligible lumen loading of ink during repulping, was used to obtain estimates of the time constants. A computer program that employed least squares analysis of the data at set time intervals was used to obtain these estimates. Once the time constants had been generated, then the curve described by the analytic solution was plotted along with the transposed experimental data. Transposing Brightness Data Data from Ciampa (1995), shown in Table A8.1, have been converted to mass ratio data using calculations outlined by Jordan and Popson (1994) who explained that the adsorption coefficient of paper is a function of the scattering coefficient of the paper and the paper's reflectance. Thus, for a paper containing trace amounts of ink: k m i x = S(1-Rinf)2/(2 Rinf) where km* = absorption coefficient of paper and containing ink (m2/kg) s = scattering coefficient of the paper (m2/kg) R = the brightness or reflectance of the paper containing ink Let Ak = kp - kmix, where kp is the absorption coefficient of paper (m2/kg) = " (kp Qnkkjnk) 220 where Q n k is the ink concentration in the paper and is approximately less than 1% of the total paper/ink mass, and kj„k is the absorption coefficient of the ink (m2/kg) also, Ak = sp((1- Rp) 2/(2 Rp) - (1- Rm«) 2/(2 Rmb{)) therefore: Q n k = (Sp/k ink)((1- Rp)2/(2 Rp) - (1- Rmix)2/(2 Rmix)) [6.1] Here, ink concentration is in parts per million. A conversion factor (B) is necessary to convert this to g ink per g fibre: Q n k = (B)(Sp/k ink)((1- Rp)2/(2 Rp) - (1- Rmix)2/(2 Rmbc)) [A8.1] Ciampa's first data point is used in the sample calculation presented here. R p = 67.1; Rmix = 37.1 which corresponds to an ink content of 0.015 g ink /g fibre; and Rmix,w, the brightness of the washed pulp, is 66.5. If it is assumed that the ink particle size does not change as repulping proceeds, then 0.015 = Aac((1-67.1)2/(2x67.1)-(1-37.1)2/(2x37.1)) = 15.0 AaC where Aac is (B)(Sp/kj nk) = 0.001 g ink/g fibre. Thus, the concentration of ink in the pulp pad after washing is: Q n k = 0.001(32.6 - (1-66.5)2/(2 x 66.5)) = 3.4 x 10"4 g/g fibre Of note in Ciampa's data set is the drop in post-pulping, pre-washing, pulp brightness. The reason for the decrease in not clear because the ink concentration within the slurry remains constant. It may be that the ink particle size distribution changes with pulping 221 time. To mitigate the effect of brightness change with time, AaC has been recalculated for each time period using the appropriate Rm i x value. Table A8.1: Data from Flexographic Ink Deposition Experiments in Acidic Conditions (Ciampa, 1995) Time (min) Pulp Brightness Before Washing (ISO) Pulp Brightness After Washing (ISO) Estimated Ink Concentration (g/g fibre) (0.5) 37.1 66.5 3.4 X10"4 5 28.3 59.7 2.9 x 10"3 10 26.7 56.6 3.9 x 10"3 15 25.0 54.6 4.5 x 10'3 30 20.3 51.9 4.9 x 10"3 the initial pulp brightness was 67.1 ISO Ciampa reported that the first brightness datum corresponded to a time of zero minutes. However, this point must have been measured slightly after t = 0 because the ink was added to blank paper at t = 0. Therefore this first datum was shifted to t = 0.5 minutes. (Whether the datum is recorded at t = 0 or t = 0.5 makes little difference to the fit of the curve. For t = 0, the equation ([10.1]) actually fits slightly better to the data (i.e., r2 = 0.9999 instead of r2 = 0.9996.)) Analytic Solution to Equation [10.1]: By assuming that lumen loading is negligible, x reverts to Tdep and subsequent integration of Equation [10.1] where 9 = G0 at t = 0, yields: 6 = 2n0(?i -1) + 60[A(1 - X) + B{X + W[A(X - 1) + B(X + 1) + 0O(1 - X)] [A8.2] 222 where: X = exp(Bt/xdep) A = K + n 0 + 1 B = [(n0-1)2 + 2K(n0+1) + K 2 ] 0 5 K = Tdep/^esc A least squares fit of Equation [A8.2] to Ciampa's (1995) ink content data yielded estimates of T d e p and Tesc • These values were then used to generate the curve seen in Figures 10.3. The program used to generate the curve is: //Langmuir model of ink particle deposition.fibres. //TO is the initial amount of ink deposited on the fibres divided by the max than can deposit. //NO is the initial concentration of ink in suspension divided by the max that can deposit. IndVars: MINUTES DepVars: THETA, NUM, DEN, MAXDEP, INK, B, A, L, K, TO, NO Params: K1, K2 MAXDEP=0.44 TO = 0.00034/MAXDEP NO=0.01466/MAXDEP |<=K2/K1 B=((NO-1 )A2 + 2*K*(NO+1) + ^ 2)^.5 A=K + NO + 1 L=EXP(B*K1*MINUTES) NUM=(2*NO*(L-1 )+TO*(A*(1 -L)+B*(L+1))) DEN=(A*(L-1 )+B*(L+1 )+TO*(1 -L)) THETA=NUM/DEN INK=THETA*MAXDEP 223 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058614/manifest

Comment

Related Items