UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development of an on-line technique to assess mixing quality in pulp suspensions Kamal, Noreen Rashda 1997

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

Item Metadata

Download

Media
831-ubc_1997-0353.pdf [ 4.82MB ]
Metadata
JSON: 831-1.0058521.json
JSON-LD: 831-1.0058521-ld.json
RDF/XML (Pretty): 831-1.0058521-rdf.xml
RDF/JSON: 831-1.0058521-rdf.json
Turtle: 831-1.0058521-turtle.txt
N-Triples: 831-1.0058521-rdf-ntriples.txt
Original Record: 831-1.0058521-source.json
Full Text
831-1.0058521-fulltext.txt
Citation
831-1.0058521.ris

Full Text

I  DEVELOPMENT O F AN ON-LINE TECHNIQUE TO ASSESS MIXING QUALITY IN PULP SUSPENSIONS By Noreen Rashda Kamal B.Sc, The University of Calgary, 1994 A THESIS SUBMITTED IN PATRIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE 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 1997 © Noreen Rashda Kamal, 1997  In  presenting  degree  this  at the  thesis  in  partial  fulfilment  of  University  of  British Columbia, I agree  freely available for reference and study. I further copying  of  department  this or  thesis for by  his  or  the  representatives.  that the  for  an advanced  Library shall make  agree that permission for  scholarly purposes may be her  requirements  It  is  granted  by the  understood  that  it  extensive  head of copying  my or  publication of this thesis for financial gain shall not be allowed without my written permission.  The University of British Columbia Vancouver, Canada  DE-6 (2/88)  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 copying  agree that permission for extensive  of this thesis for scholarly purposes may be granted  department  or  by  his or  her  representatives.  It  is  by the head of my  understood  that  copying or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department  of  C/k&<M{taJl  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  Pio -  tfeSdu/'CA  £tg  >nC.e.<sttLj  11  ABSTRACT  Assessing the mixing quality in pulp suspensions is very important to pulp and paper processing because inhomogeneities lead to a low quality product. In the past, mixing has been assessed using a number of techniques including: The uniformity of inert tracers added to the suspension; the distribution of residual bleaching chemical following mixing and reaction; the temperature uniformity around the periphery of process piping; and by benchmarking the bleaching stage using various pulp quality parameters.  Many of these techniques require  removing and testing samples, which is a tedious procedure and it gives no indication of the variability of the mixing process over time. Furthermore, the major disadvantage with most of the techniques is that the response or sampling time is too large. Therefore, the objective of this work is to develop an on-line sensor to assess mixing quality. An on-line technique is described using an in-situ fibre optic probe.  The technique  measures the concentration of an inert ultraviolet fluorescent tracer added to the pulp suspension at the ppm level. The method has been tested in the laboratory and has been shown to give mixing quality indices that agree with other methods. The applicability of this technique in a mill is discussed; however, there are limitations in using this technique in an aggressive bleaching environments, especially in the presence of C10 . 2  This technique can easily be  applicable in assessing mixing in stock chests, in the headbox of the paper machine, and in the extraction stage of the bleach plant.  Ill  TABLE OF CONTENTS ABSTRACT  ii  TABLE OF CONTENTS  iii  LIST OF TABLES  vi  LIST OF FIGURES  vii  ACKNOWLEDGEMENTS  x  1. INTRODUCTION  1  2. LITERATURE REVIEW  4  2.1 2.2 2.3 2.4 2.5 2.6 2.7  Mixing With Fibre Suspensions Criteria that Need to Be Met for Assessing Mixing Quality Assessing Pulp Mixing on the Laboratory Scale Assessing Mixing in Pulp and Paper Mills Assessing Mixing in Other Industries Use of a Fluorescent Tracer to Assess Mixing Quality Using Fibre Optics to Measure Fluorescence  4 7 7 11 12 14 15  3. EXPERIMENTAL WORK 19 3.1 Description of Equipment Used 19 3.2 Finding Appropriate Fluorescent Dyes . 20 3.3 The Relationship Between Fluorescent Intensity and Concentration With Fibre Present 22 3.4 Fluorescent Dye Adsorption on the Pulp Fibres 23 3.5 Reaction Between Fluorescent Dyes and Typical Bleaching Chemicals 24 3.6 Construction of Fibre Optic Probes 25 3.7 Hand Mixing Experiments 29 3.8 On-line Mixing Tests 30 3.9 Comparison with Lithium Tracer Tests 31 4. RESULTS AND DISCUSSION 4.1 Finding Appropriate Dyes 4.2 Using Fluorescent Dyes With Pulp Fibres 4.3 Adsorption Equilibria 4.3.1 Background on Adsorption Equilibria 4.3.2 Adsorption of Mephenesin and 2-Naphthalene Sulphonic Acid 4.4 Effect of Bleaching Chemicals on Fluorescent Dyes  33 33 39 44 44 45 49  iv 4.4.1 Background on Kinetics of Fluorescent Dyes and Bleaching Chemicals 49 4.4.2 Reaction Kinetics Between Fluorescent Dyes and Bleaching Chemicals 51 4.5 Assessing Mixing With Fluorescent Dyes as a Tracer 58 4.5.1 Relationship of Mixing Quality and Mixing Energy in Hand Mixing . . . 58 4.5.2 Continuous On-line Mixing Tests of a Batch Mixer 60 4.5.3 Comparison Between a Fluorescent Tracer and Lithium Tracer 65 5. CONCLUSIONS  67  6. RECOMMENDATIONS FOR FURTHER WORK  70  7. NOMENCLATURE  72  8. REFERENCES  74  APPENDIX I: Details of Techniques That Have Attempted to Assess Mixing Quality 1. Mixing Assessment in Pulp Suspensions on a Laboratory Scale 1.1 Kouppamaki (1985) . . . 1.2 Kouppamaki et al. (1992) 1.3 Paterson and Kerekes (1985) 1.4 Breed (1985) 1.5 Francis and Kerekes (1990) 1.6 Bennington and Thangavel (1993) 2. Mixing Assessment in Pulp Suspensions on a Mill Scale 2.1 Elliott and Farr (1973) 2.2 Backlund, Bergnor, Sandstrom, and Teder (1987) 2.3 Torregrossa (1983) 2.4 Torregrossa (1983), Sinn (1984), and Pattyson (1985) 2.5 Paterson and Kerekes (1986) 3. Mixing Assessment in Other Industries 3.1 Nagase and Yasui (1983) 3.2 Shenoy and Toor (1990) 3.3 Mann, Knysh, Rasekoala, and Didara (1988) 3.4 Lee and Brodkey (1964) ; 3.5 Haam ad Brodkey (1992) 3.6 Holmes, Voncken, and Dekker (1964) 3.7 Patterson, Bockelman, and Quigley (1982) 3.8 Gaskey, Vacus, David, J.C. Andre, and Villermaux (1988) 3.9 C. Andre, David, J.C. Andre, and Villermaux (1992)  . . 78 78 78 79 80 81 82 83 86 86 86 87 88 .88 91 91 91 92 93 93 94 94 95 96  V  APPENDIX II: Fluorescence 1. 2. 3. 4. 5.  Introduction The Excitation Process The Deexcitation Process Quenching of Fluorescence Fluorometers  99 99 99 101 102 103  APPENDIX III: Decay of Fluorescent Dyes with Bleaching Chemicals  106  APPENDIX IV: Study of the Noise in the Fluorescent Signal  117  APPENDIX V: Results of Continuous Mixing Tests  121  vi LIST O F TABLES  1. Summary of dyes that were tested  21  2. Typical conditions and chemical concentration in the bleach plant  25  3. Dimensions of Perkin Elmer probe, probe 1, and probe 2  27  4. How well all the fluorescent dyes met each criteria  36  5. Summary for rate constants and activation energies for mephenesin and 2-NSA  53  6. Summary of activation energies and frequency factors for mephenesin and 2-NSA . . . 53 7. Typical half life of mephenesin and 2-NSA with each bleaching chemical  54  8. Summary of mixing times  63  9. Comparing the mixing index for the fluorescence method and the lithium method . . . 65 1-1. Summary of methods attempted in the pulp and paper industry on a laboratory scale  85  1-2. Summary of methods attempted in the pulp and paper industry in the mill  90  1-3. Summary of methods attempted in other industries  98  IV-1. The slope shown in equation IV-1 for each probe  118  IV-1. Summary of variances for each probe  119  Vll  LIST OF FIGURES  1. Schematic of proposed technique  17  2. Schematic of Perkin Elmer LS50B with fibre optic probe attachment  20  3. Cross-sectional view of Perkin Elmer probe, probe 1, and probe 2 with a scale of approximately 5x  28  4. The cross-sectional area of the excitation optical fibres is directly related to emitted light intensity  28  5. Schematic of mixing apparatus used for on-line mixing test, where M l and M2 show the probe locations  31  6. Relationship between intensity and concentration for (A) EEDQ, (B) Azaindol, (C) Anthracene, and (D) Piperoxan  . 37  7. Relationship between intensity and concentration for (A) Mephenesin and (B) 2-NSA . 38 8. Chemical structure of mephenesin, 2-NSA, and the lignin monomer  39  9. Effect of consistency on the fluorescence intensity for mephenesin  42  10. Effect of consistency on the fluorescence intensity for 2-NSA  42  11. Fluorescence of lignin at 304 nm with mephenesin in 10% consistency pulp suspension (excitation at 280 nm)  43  12. Fluorescence of lignin at 334 nm with 2-NSA in 10% consistency pulp suspension (excitation at 277 nm)  43  13. (A) Adsorption isotherm of mephenesin with the error bar depicting sigma. (B) Adsorption isotherm of 2-NSA with the error bar depicting sigma  47  14. Adsorption at 3%, 7%, 10%, and 15% consistency without normalizing the mass of fibres present for (A) mephenesin and (B) 2-NSA  48  15. Reaction between 2-NSA and H 0 shown as (A) raw data and (B) determination of rate constants  55  16. Determination of activation energy for 2-NSA and H 0  56  2  2  2  2  Vlll  17. Relationship between mixing index and energy dissipated found by hand mixing mephenesin, and the error is calculated using the standard deviation  59  18. Observed movement of the fluid in the stirred tank  61  19. On-line mixing test result using Perkin Elmer probe at 1 % consistency and the rotor stirring at 150 rpm  61  20. On-line mixing test result using Probe 1 at 1 % consistency and the rotor stirring at 150 rpm  62  21. On-line mixing test result using Probe 2 at 1% consistency and the rotor stirring at 150 rpm  62  22. Conceptual probe designed to facilitate pulp floes to move over the probe tip easily  64  23. Proposed schematic for mill installation of a fluorescent tracer method utilizing a fibre optic probe  71  II-1. Jabolonski energy-level diagram depicting absorption and emission processes II- 2. Six major components of a basic fluorometer  . . .  100 105  III- 1. Reaction between 2-NSA and C10 (A) raw data (B) determination of 2  rate constants  107  III-2. Determination of activation energy for 2-NSA and C10  108  2  III-3. Reaction between 2-NSA and NaOH (A) raw data (B) determination of rate constants  109  III-4. Determination of activation energy for 2-NSA and NaOH  110  III-5. Reaction between mephenesin and C10 (A) raw data (B) determination of 2  rate constants  Ill  III-6. Determination of activation energy for mephenesin and C10  112  III-7. Reaction between mephenesin and H 0 (A) raw data (B) determination of rate constants III-8. Determination of activation energy for mephenesin and H 0  113 114  2  2  2  2  2  ix III-9. Reaction between mephenesin and NaOH (A) raw data (B) determination of rate constants  '.  115  111-10. Determination of activation energy for mephenesin and NaOH  116  V - l . On-line mixing test with no fibres at 75 rpm using Perkin Elmer probe  122  V-2. On-line mixing test with no fibres at 75 rpm using Probe 1  123  V-3. On-line mixing test with no fibres at 75 rpm using Probe 2  124  V-4. On-line mixing test at 1 % consistency at 150 rpm using Perkin Elmer probe . . . . 125 V-5. On-line mixing test at 1% consistency at 150 rpm using Probe 1 V-6. On-line mixing test at 1 % consistency at 150 rpm using Probe 2  126 127  V-7. On-line mixing test at 2% consistency at 300 rpm using Perkin Elmer probe. . . . 128 V-8. On-line mixing test at 2% consistency at 300 rpm using Probe 1  129  V-9. On-line mixing test at 2% consistency at 300 rpm using Probe 2  130  ACKNOWLEDGMENTS  I would sincerely like to acknowledge the following people for there help in allowing this work to be completed. I would like to thank my supervisor Chad Bennington for his advice and guidance. I thank Denys Leclerc and Thanh Trung at PAPRICAN's Vancouver Laboratory for allowing me to work in their laboratory and helping me to learn the instrumentation. Also, I acknowledge everyone at the Pulp and Paper Centre for assisting the completion of my thesis. Especially, I thank Peter Taylor for all his help in constructing the fibre optic probes, and Rita Penco for her thorough literature searches.  1.INTRODUCTION Mixing is very important in pulp and paper processing. It is used throughout the pulp mill for blending pulp in stock chests, for producing uniform slurries in the headbox of paper machines, and for mixing chemicals with pulp in the bleach plant. The purpose of mixing in the stock chests is to maintain uniformity of composition and consistency. Furthermore, the purpose of mixing at the headbox of paper machines is to achieve mass uniformity in order to control flocculation and formation. Poor mixing on the paper making side can cause poor sheet formation.  Finally, the purpose of mixing in the bleach plant is to ensure that each fibre  contacts the correct amount of bleaching chemical to ensure effective pulp bleaching. Poor mixing in the bleach plant can have detrimental effects on the product; for example, it can lead to a product that is not evenly bleached or a product that has lower brightness. This is often offset by adding more bleaching chemicals, which has an adverse effect on the strength of the product. Excessive use of bleaching chemicals are costly and it can make it difficult to meet environmental regulations. As can be seen from the above examples, mixing is used to achieve a homogenous slurry through out the pulp and paper mill. Homogeneity is required on many scales: on the macroscale (> 10 mm), on the fibre scale (0.05 mm - 10 mm), and on the microscale (<0.05 mm). Mixing is rarely measured in the mill; however, it can be assessed by indirect methods such as measuring pulp brightness or strength. For example, if the product has low brightness or low strength, it can be assumed that there is poor mixing occurring in the bleach plant. However, this is not an unequivocal indicator of mixing quality since low brightness may be due  -2-  to other factors; for example, poor mixing can be due to channelling in a bleaching tower or to poor washing of the pulp in the washers (due to residual chemicals) rather than from the mixers. Where mixing has been measured, direct sampling techniques have been used which are tedious. Further, such techniques only give an indication of mixing at the sampling point, and there is so much work involved in handling of the samples that only a small number of samples can be taken. The limited number of samples leads to a very narrow picture or "snap shot" of how mixing is occurring leaving little scope for trouble-shooting. A more desirable way to assess mixing in a pulp and paper mill is to use an on-line technique; however, none have been developed or used. The goal of my research is to assess mixing quality in pulp suspensions in situ and in real time. By measuring mixing on-line we can determine two things. First, it will give us an indication of how well the mixer is working. Secondly, an on-line technique can also be used, as a diagnostic tool. That is, the on-line technique can be used to trouble shoot what is affecting mixing up-stream of the mixer. For example, a positive displacement pump maybe causing poor mixing at every pulse, which can be determined by measuring the change in mixing quality with respect to time. In other words, one can compare the points in time where mixing is poor to changes in process conditions. A number of criteria must be satisfied when developing a method to assess mixing. First, the technique must measure something directly related to mixing quality. Second, it must be safe to handle (that is toxic and radioactive additives cannot be used). Third, the technique must be applicable with the presence of fibres. Fourth, it cannot affect process conditions (such as temperature), and it cannot affect the end product in any way (such as coloration of the pulp).  -3Fifth, it must be applicable as an on-line method in a mill (in-situ and in real time), and it must measure fibre-scale mixing. Sixth, the method should be applicable to as many bleaching stages as possible. Seventh, its use should be cost effective.  _  4  _  2. LITERATURE REVIEW  2.1 Mixing With Fibre Suspensions Measuring the quality of mixing has been attempted by chemical engineers for the last 45 years. Dankwerks (1953 a) presented fundamental thought on the theories of mixing, and specifically he talks of the "scale of scrutiny", which is applied to the minimum size of the regions of segregation in the mixture that would cause it to be regarded as imperfectly mixed for a specific purpose. The scale of scrutiny can be expressed as an order of magnitude such as length, volume, or area. Dankwerks (1953 b) also described quantitative ways of measuring mixing using concepts common in turbulence theory. Two measures are used. Thefirsthe calls the scale of segregation, which is the size of the regions of segregation, or "clumps", of unmixed components in an imperfect mixture. Mixing cuts or breaks the "clumps" of unmixed components into smaller "clumps", which cause the quality of mixing to increase and the scale of segregation to decrease. The second measure is called the intensity of segregation, which is a measure of the departure of the composition from the mean value, averaged over all points in the mixture.  However, the problem of assessing mixing is further complicated with fibre  suspensions, since fibres tend to flocculate, which imposes non-uniformities. Bennington (1996 a) has thoroughly described mixing in pulp bleaching. He states that the purpose of mixing is to eliminate non-uniformities on a number of scales, or simply mixing scales. The large scale variations is called the macroscale.  This represents non-uniformities  around the size of pipe diameters or more, and this is given the approximate size of 10 mm or larger in chemical engineering literature. Mixing is achieved on this scale by substantial back  -5-  flow. The next mixing scale is the fibre-scale, which represents the non-uniformities having a size from fibres to floes, and the approximate size of this scale is from 0.05 mm (fibre diameter) to 10 mm (floe size). Mixing on this scale is achieved primarily by laminar and turbulent shear and through diffusion. Finally, the last mixing scale described is the microscale, which has dimensions of less than 0.05 mm, and represents non-uniformities approaching the molecular level. Mixing is achieved here by diffusion, which is aided by small-scale fluid motion. When considering these mixing scales for my on-line mixing technique, it is obviously advantageous to have a technique that will measure mixing on all three of these mixing scales. However, it is difficult to conceive of an on-line measurements at scales below 0.05 mm, thus a good technique would measure on both the fibre-scale and the macroscale. Furthermore, micro-scale mixing is achieved by diffusion, and therefore it is not directly affected by mechanical mixing, which shows that only fibre-scale and macroscale mixing will be affected by improvements in the mixing equipment. The next area to consider is what sort of response time will be needed to obtain the quality of mixing on the fibre-scale. This can be calculated from how fast the pulp is moving through the pipe. Typical pipe velocities for medium consistency ( C =8%-20%) suspension m  are between 1.8 m/s to 2.3 m/s after the mixer in a bleach plant. This means that a response of 230 Hz is required to be just inside the fibre-scale (10 mm) as given by the following calculation: (2.3m/5)  (0.01/w)  = 2 3 0 5  _  1 = 2 3 ( ) g z  (  1  )  -6However, the Nyquist sampling criterion must be taken into consideration, which states that in order to obtain data every 10 mm with 100% accuracy, then the response time needs to be multiplied by 2.2 (Kersey, 1991); therefore, the response required for medium consistency suspensions is 506 Hz. Furthermore, for low consistency suspensions (C =0%-8%), pipe m  velocities are typically between 5.2 m/s to 6.5 m/s after the mixer in a bleach plant, which corresponds to a response of 1430 Hz after taking the Nyquist frequency into account. Medium consistency mixing is becoming very widespread in the pulp and paper industry. Consequently, a response of 506 hz will be aimed for, but if a response of 1430 hz can be achieved, then it would be preferred. There are several ways to quantify mixing; however, a common and easy way is to use a statistical parameter called the mixing index, M , which is the coefficient of variation of the parameters of interest. It is readily calculated from:  n  M  (2)  XA r N n-l M=sjx=1  where s is the standard deviation of the measured quantity and x is its mean value. A mixing index of zero corresponds to perfect mixing; however, for pulp bleaching good mixing is referred to any mixture with a mixing index less than or equal to 0.1.  This is because the  amount of chemical savings obtained by improving mixing quality from 0.1 to 0.05 is typically very small (Bennington, 1996 a).  -72.2 Criteria that Need to be Met for Assessing Mixing Quality There are seven criteria that an ideal method to assesses mixing should fulfil. While performing a thorough literature review, each method found was critiqued according to the following seven criteria:  1.  It must measure something directly related to mixing quality.  2.  It must be non-hazardous and non-toxic.  3.  It must be applicable with the presence of fibres.  4.  It cannot affect process conditions, such as temperature, and it cannot affect the end product in any way, such as coloration of the pulp.  5.  It must be applicable as an on-line method in a mill (in-situ and in real time), and it must measure fibre-scale mixing (response time of at least 506 Hz).  6.  It must be applicable to as many bleaching stages as possible.  7.  Its use should be cost effective.  2.3 Assessing Pulp Mixing on the Laboratory Scale Since mixing is such an important operation in a pulp and paper mill, there have been several attempts made to assess mixing on laboratory scale equipment. Many of the methods attempted were on-line, or they can be applied to an on-line system. An analysis was done to see how many of my criteria shown in section 2.2 were met. Each method is described in detail in Appendix I. Three of the attempts made to assess pulp mixing on the laboratory scale were done using a tracer technique. A tracer is a material that is added at a location upstream of the mixer, and  -8it then goes through the mixer. The spatial distribution of the tracer as a function of time is determined after mixer. Good mixing would correspond to a concentration that remains the same, and poor mixing would correspond to high fluctuations in the concentration. There were three tracers used: 137m-Ba was the radioactive tracer used (Kouppamaki, 1985 & 1992), NaCl was the ionic tracer used (Breed, 1985), and Astra Red P Liquid was the tracer dye used (Francis and Kerekes, 1990). Another attempt made to assess pulp mixing was by sampling the residual chlorine concentration in a suspension using syringes (Paterson and Kerekes, 1985), and finally a mixing sensitive chemical reaction to study microscale mixing and turbulence intensity with in the liquid phase of the pulp fibre suspension (Bennington and Thangavel, 1993). Kouppamaki (1985; 1992) attempted to assess mixing by using a radioactive tracer. This method can be on-line; however, it is not readily applicable in a mill environment, since it is very difficult to use radioactive material in a mill. Radiation detectors can be mounted on the pipe wall which means that there would be no need for sampling ports; however, the cost of radioactive generators and their detector are very expensive. This method has a response of 50 Hz, which is still considerably slower than 506 Hz.  The advantages to this method is that it  does not change process condition or the end product, and it is applicable with fibres and to all bleaching stages. Breed (1985) assessed mixing in the laboratory by measuring the distribution of NaCl concentration at the discharge of the mixer. This method can be applied to a mill by measuring another ion such as lithium, since both sodium and chloride are already present in the process in high concentrations.  Furthermore, Breed uses a sampling method to determine the  concentration of the salt; however, an ion selective electrode can be used to apply this method  -9to an in-situ technique. The problem with this method is that ion selective electrodes have a very slow response time (2 Hz) (Nagase et al.,1983). This response time would measure on a length scale of 70 m in medium consistency mixing, which is well into the macroscale range. Francis and Kerekes (1990) assessed mixing in high consistency pulp suspensions (C =20%-40%) by use of a tracer dye. The sampling method they used was quite involved and m  tedious because from one sample 50 pulp floes were picked, and the dye concentration was determined by chemical treatment and spectrometry; furthermore, the floes were oven dried and weighed.  This work's goal (to assess mixing in high consistency pulp suspensions) is not  directly related to my work, since most bleach plants in mills do not use high consistency pulp suspensions; however, it does bring about a useful idea, that is to assess mixing by using a tracer dye. This idea can be in-situ by applying a fibre optic probe; an optical method would have a very fast response time. The main problem with this method is that it colours the pulp, and tracer dyes that absorbs light in the ultra-violet range can not be used, since the absorbance of lignin cover the entire ultra-violet range (Francis and Kerekes, 1990). Furthermore, dyes might react readily with the bleaching chemicals, but more research needs to be done to confirm this. Paterson and Kerekes (1985) measured mixing quality in the fibrescale range by measuring the residual chorine concentration. For this method samples are removed, which means that it is not an on-line method in real time, and it only gives a snap shot of the mixing at particular point in time; however, later it was applied to mills (discussed in section 2.4). This method is only applicable to the chlorine stage, although it could by adapted to other chemistries. Furthermore, extracting samples from a sampling line can contribute to mixing or segregation.  -10Bennington and Thangavel (1993) used a mixing sensitive chemical reaction to study microscale mixing and turbulence intensity within the liquid phase of pulp fibre suspension. Although this method does not study mixing on thefibre-scaleand macroscale (which is the goal of my research), it is worth mentioning since it could provide a useful avenue to assess mixing for my research.  This technique assessed micromixing in pulp fibre suspensions by using  competitive, consecutive chemical reactions between 1-naphthol and diazotized suphanilic acid. Mixing quality was determined from the distribution between mono and bis substituted reaction products once a correlation was made for the adsorption of the product dyes onto the suspended fibres. This method could be applied to an on-line technique by using fibre optics to measure the products dyes; however, this would colour the pulp, which shows that criteria #4 is not met. Finally, the reactants were extremely expensive. Many of the above mentioned techniques of assessing mixing are not applicable in the mill since they often colour the pulp. Paterson and Kerekes' work (1985) and Breed's work (1985) was later applied to a mill, and Kouppamaki's work (1985 and 1992) shows the most promise despite the problem of bringing radioactive materials into a mill. These methods are further summarized in Table 1-1 which is shown in Appendix I.  -112 . 4 Assessing Mixing in Pulp and Paper Mills The literature has cited several attempts to assess mixing in pulp and paper mills. In all cases sampling of the suspension at some point in the process was done, with the samples analyzed later in the laboratory. The methods are examined to determine how many of the criteria that are laid out in section 2 . 2 are met. Each method is described in detail in Appendix I. One of the first attempts made to assess mixing in the chlorination stage of a mill was in 1973 by Elliott and Farr.  They assessed mixing quality in a mill by measuring the free  available chlorine, and by plotting the change in oxidation reduction potential (ORP) with time. This method can be applied today by using a specially designed electrode that measures the residual chlorine, such as the one developed by Paterson and Kerekes (1984). This would allow this technique to be on-line, and measure small volume samples. The response of the electrode would by relatively slow (0.0167 Hz) (Paterson and Kerekes, 1984). Furthermore, it would only be applicable to the chlorine stage. One method performed excessively in mills and accepted widely is the addition of an ionic tracer (Backlund et al., 1987; and Torregrossa, 1983).  This method can be made  continuous by utilizing an ion selective probe or an optode; however, even the fastest of these probes are too slow for our purposes, since they have a response of 2 Hz or smaller because they rely on diffusion. Another method was to make a temperature profile around the discharge pipe of a C10  2  mixer, which was done by Torregrossa in 1983. It works in the C10 stage because the C10 2  2  is chilled, and when it is mixed with the pulp, the C10 acts as a temperature tracer. The C10 2  2  -12enters at about 10 °C, and the pulp temperature is around 70 °C. This method is only applicable to the C10 stage. However, a cold stream could be added as a temperature tracer for the other 2  stages, although this would change the process conditions; consequently, criteria #4 would not be met. The fastest thermocouples (10 Hz) (Haam et al, 1992) would still be too slow for our purpose. In 1986, Paterson and Kerekes extended their laboratory technique (discussed in section 2.3) into the mill.  This method is not an on-line method in real time because samples are  withdrawn from the line mixer, and it only gives a snap shot of mixing quality at particular point in time. Furthermore, an entire sampling line needs to be added to the sampling port which can contribute to mixing or segregation in the sampling line. Mill data illustrated that this technique does give results that are indicative of mixing quality. Many of these techniques can assess mixing on-line by applying relevant changes (Elliott and Farr, 1973; Backlund et al., 1987; Torregrossa, 1983); however, the main problem is that the response time is too slow. Ion selective electrodes are very slow (0.033 Hz), and even very fast thermocouples tend to be too slow for our purposes (10 hz). These methods are summarized in Table 1-2 which is shown in Appendix I.  2.5 Assessing Mixing in Other Industries A literature review was conducted of mixing assessment techniques used outside the pulp and paper industry to see if they could be adapted for pulp suspensions. Again, the methods are described in detail in Appendix I. Many of the techniques that are applicable to the pulp and  -13paper industry have already been attempted in pulp suspensions and mentioned above. Some of these methods are the addition of an ionic tracers (Nagase and Yasui, 1983), using a dye to assess mixing quality (Shenoy et al, 1990; Mann et al, 1988; and Lee et al., 1964), and measurement of temperature uniformity to assess mixing (Haam et al, 1992). There are two methods that are different from methods that have been attempted in the pulp and paper industry, but may be applicable to it. The first method is measuring conductivity of an ionic tracer (Holmes et al, 1964). Holmes obtained a response of about 50 Hz, the measurement were continuous, and it was very easily done. However, it would be difficult to apply this to a mill because of the interference from the other ions that are present in the mill environment; furthermore, the concentration of the ions in the process would not remain constant, and it would be difficult to separate whether the change in conductivity is occurring due to changes in the background ions or due to poor mixing. The second method is using a fluorescent tracer to assess mixing (Patterson et al, 1982; Gaskey et al, 1988; and C. Andre, 1992). This method has the potential for a fast response time (response times of up to 5000 Hz have been obtained by Gaskey et al. in 1988).  Furthermore, it is hazard free and is very  sensitive and selective to a particular dye (detectable in the ppm range). Furthermore, it has already been brought to a mill as a tracer to test the residence time in a high density tower ahead of the bleach plant (Mitro, 1995). However, there are certain unknowns such as how fibres will effect the signal, if lignin will fluoresce, and if the bleaching chemicals will breakdown the fluorescent dye, since compounds that have a colour or fluoresce have aromatic structures that are similar to that of lignin. From this literature review it seems that the fluorescent tracer method might prove  -14to be applicable to our situation; however, more work needs to be done. These methods are further summarized in Table 1-3 which is shown in Appendix I.  2.6 Use of a Fluorescent Tracer to Assess Mixing Quality By evaluating all the advantages and disadvantages from the previous section and Appendix I, the fluorescent tracer methods appears to be the most promising. The use of a fluorescent tracer is the one method (aside from the radioactive tracer methods) that meets all of the criteria outlined in section 2.2:  1.  This technique measures mixing quality as demonstrated by Gaskey et al. (1988).  2.  Many fluorescent dyes are non-hazardous and non-toxic.  3.  Laboratory work needs to be done to determine if a fluorescent tracer is still applicable with the presence of fibres.  4.  By using a colourless tracer at the ppm level, the pulp will not be affected.  5.  Fluorescence has a very fast response time of 5000 Hz, and can be made into an on-line process by applying fibre optics to measure the tracer concentration in-situ (see section 2.7).  6.  It might be applicable anywhere in the mill; however, we need to determined if the bleaching chemicals react significantly with the fluorescent dyes chosen.  7.  Fluorescent dyes range widely in price. A suitable one would be inexpensive since significant dye would be used in an industrial scale plant. The provision of a suitable sampling location may also be required which is expensive.  Background on how fluorescence occurs and how it is measured is discussed in Appendix II. The radioactive tracer method actually has more advantages than the fluorescent tracer  -15method because no sampling ports are required. However, it is difficult to bring radioactive material into a mill due to the hazards involved, and a fluorescent tracer can be non-toxic and hazard free. Although the average cost of a fluorometer is quite high ($15 000), it is still much cheaper than radiation detectors. The largest disadvantages with the fluorescence method is that there are many unknowns. This is because an on-line technique of measuring fluorescence in-situ in pulp suspensions has never been attempted before. This raises questions about how the fibres will affect the signal, how lignin's fluorescence will effect the signal, and if the bleaching chemicals will attack the dye. All this needs to be addressed.  2.7 Using Fibre Optics to Measure Fluorescence The usual method to measure fluorescence is to sample by separating the aqueous phase from the dispersed phase and measuring the intensity of fluorescence in the aqueous phase. Sampling methods are quite tedious, since samples need to be withdrawn and analyzed. For this reason only a small number of samples can be taken, which only give a "snap shot" of mixing at particular point in time. An on-line technique is advantageous because it would show the change in mixing quality in real time. It would measure the concentration of the fluorescent dye every fraction of second (the time interval could be controlled), and therefore a large set of data points can be collected. In order to measure fluorescence on-line with the presence of pulp, one needs to either continuously remove filtered samples after the mixer, or to insert a fibre optic probe and measure the fluorescence in-situ. The fibre optic probe would give the most accurate description  -16of how the tracer is being mixed.  However, this presents questions about the nature of  fluorescence and whether this can be done because fluorescence is usually measured at right angles (see Appendix II section 4). The reason for measuring fluorescence at right angles is because it minimized interferences between transmitted and scattered exciting light, which means that a 90° angle gives the greatest emission intensity. Therefore, while fluorescence can be measured at any other angle, the emission intensity would be less. Afibreoptic probe, with half thefibretransmitting the excitation light, and the other half taking the emitted fluorescence back to the fluorometer, utilizes a 360° angle or reflectance.  The use of a fibre optic probe to  measure fluorescence has already been applied by Perkin Elmer in their fluorometer model LS50B (Perkin Elmer, Vancouver, BC), which is what I will use for my experiments. Now that a methods exists for measuring fluorescence on-line, a proposed scheme can be devised for in-situ measurements as is illustrated in figure 1. The fluorescent tracer will be added directly before the mixer, and the fibre optic probe will be inserted directly after the mixer. A light source will shine the excitation wavelength through a bundle of optical fibres. The light source will be a xenon discharge lamp because it is capable of UV-visible excitation. Then the emitted fluorescence will go to the fluorometer for intensity detection, and the signal will then be analyzed with a computer through a program that will calculate the concentration of the fluorescent tracer as a function of time, and then various mixing indexes at different time intervals is determined. This will measure the spatial distribution of the fluorescent tracer, and give an indication of how "good" the mixing is.  -17-  point of continuous insertion of fluorescent tracer Pulp flow  fibre optic probe inserted in a sample port Xenon discharge lamp—^  Mixing Indices  CPU  A/D  fluorometer  Figure 1: Schematic of proposed technique.  There are several criteria that need to be addressed because fluorescence has never been used in a pulp mill to assess mixing quality. First, a colourless dye that fluoresces in the UV range needs to be found so that the paper or pulp is not effected. The second criteria that needs to be taken into consideration is that the fluorescence intensity of the dye must have a monotonically increasing relationship with concentration. Thirdly, the dye must be sensitive enough to be detected in the ppm range and by the reflectance method. The dye should be nontoxic and water soluble. Lastly, the cost of the fluorescent tracer must be assessed. Once a dye is found that meets all these criteria, it needs to be determined if the concentration of the fluorescent dye can be measured accurately with the presence of pulp, or  -18if lignin interferes with the signal. Since this method will be in-situ, how much of the dye will adsorb on to the fibres will be studied.  Finally, the stability of the fluorescent dyes in the  presence of typical bleaching chemicals under typical bleaching condition will be determined.  -19-  3. EXPERIMENTAL WORK 3.1 Description of Equipment Used A Perkin Elmer (Vancouver, B.C.) fluorometer model LS50B was used. This unit is capable of measuring fluorescence, phosphorescence, chemiluminescence, and bioluminescence. It has a xenon discharge lamp as a light source, and the detector is a gated photomultiplier. The LS50B is capable of excitation wavelengths from 200-800 nm, and emission wavelengths from 200-900 nm. Its general set up is similar to that shown in figure  II-1 (in Appendix II).  However, in the Perkin Elmer LS50B fluorometer the user has the option of using a sample cell or a fibre optic probe. The probe cannot be immersed in water; therefore, it can only scan the top of a sample, and it is placed in the fluorometer as shown in Figure 2. The probe can be easily detached and the sampling cell installed in its place.  The instrument is capable of  providing a spectrum of the fluorescent intensity over a range of emmision wavelengths at a specific excitation wavelength. The LS50B can also read the intensity of a known emission wavelength by setting the excitation wavelength. Finally, the fluorometer is capable of dynamic measurements as well (by using what Perkin Elmer calls the "time drive mode"), which records a number of emission intensity values collected at a specified time interval for a specified amount of time. This allows a chart to be made with shows the change in fluoresce intensity with time. This is useful for conducting on-line mixing tests, and finding how fast the bleaching chemicals attack the dye.  -20-  Detector  Figure 2: Schematic of Perkin Elmer LS50B with fibre optic probe attachment.  3.2 Finding Appropriate Fluorescent Dyes Of the numerous fluorescent dyes that were screened, 14 were purchased and tested. Eleven of the dyes are readily available from Anachemia (Richmond, BC); however, three dyes were found through Molecular Probes Inc. (Eugene, OR). The fluorescent dyes that were tested are summarized in Table 1 with their excitation and emission wavelengths, supplier, lot number, percent purity, and unit cost.  21  I  CN  O o  o o o  os os  OS  os  o o  mm  c  s ft  o CN  CN  I  00 00 CN  o  o o  o cn  oo ON  00  Os  <  o o  Os OS  6£ o OS  i  so  o  SO^  SO  o  00  cn  oo"  SO OS  Z  CN  CO  o  S  SI  a <0  •3  CO  a •s < =5 o  o  0  a  J3  o CO  8  J5  o  < a C  Ti-  oo cn  CM  CN  cn  o  o  o  O  o  CN  00 CN  CN  CN  CN  OS  OS  en cn  C~ CN  CN  •o  J3  mm  PQ  as 3  £>  a  fSI  « S ¥ » S  u  V-i >s  a,  5 C O O  w CO  >- e x "JS o  o  •o5'£ 3 '  c  ~  S  5? *  V ^  CN  s  'o  6 o  a O '  0  1 s C O <  (D  e  tl  15  a lo  c  u.  CM  CO  0  o  CO  •a o  C O X C O  •S  a l« c  •S tl  -22The potential applicability of each dye was found by testing whether they were colourless and water soluble, which was easily done by visual investigation. The sensitivity of each dye was also determined.  This was done by making samples of varying fluorescent dye  concentrations (0.1 ppm, 1 ppm, 10 ppm, and 100 ppm), and then the fluorescent intensity is read from the fluorometer. The intensity value given by the LS50B is an arbitrary number between 0 and 1000; however, this can be adjusted by changing the multiplication factor. I kept the multiplication factor at 1 in order to keep the background noise constant. The sample that gives the highest fluorescent intensity indicates the greater sensitivity of the dye, but the intensity needs to be well below the value of 1000. This determines the concentration range needed for that particular dye. Once the sensitivity range is found, the relationship between fluorescence intensity and dye concentration (without fibres) can be determined. This was done by creating samples in the sensitivity range with known dye concentration (for example, if the sensitivity range was 1 ppm then samples of 0 ppm, 0.5 ppm, 1 ppm, 1.5 ppm, 2 ppm, 2.5 ppm, and 3 ppm were made), and then each sample was put under the fibre optic probe, and its intensity was measured.  3.3 The Relation Between Fluorescent Intensity and Concentration With Fibre Present In section 3.2, the relationship between fluorescent intensity and dye concentration was determined, now this relationship can be found with the presence of fibres by utilizing the fibre optic probe to determine how fibres effect the fluorescent signal. This is done in a similar manner to that without fibres. Samples were created with known dye concentrations in distilled water. The sensitivity range with fibres was slightly higher than the sensitivity range without  -23fibres due tofluorescentquenching (further studied in section 4.3). Oven dried fully bleached kraft (FBK) fibres were then added to the samples to prepare suspensions of the desired consistency. Oven dried fibres were used to eliminate the dilution due to the water in the fibre suspensions.  The desired consistencies were 3% and 9%. The fibres were left to soak for 4  hours, and then mixed with for 15 minutes in a Beach House milkshake mixer (San Diego, CA). The samples were then placed under the fibre optic probe, and thefluorescentintensity was read from the fluorometer.  3.4 Fluorescent Dve Adsorption on the Pulp Fibres The amount offluorescentdye adsorbed on the fibres was determined. Two sets of six samples having increasing amounts of a fluorescent dye to make solutions having dye concentrations in the sensitivity range of the particular dye being tested (e.g. 0 ppm, 1 ppm, 2 ppm, 3 ppm, 4 ppm, and 5 ppm for a dye with a sensitivity range in the 1 ppm range). One set was put aside (this was labelled the first set). In the other set (this was labelled the second set), oven dried FBK fibres were added to each sample. The amount of fibres that were added depended on the desired consistency. The consistencies that were utilized were 3%, 7%, 10%, and 15%. The fibres were allowed to soak for 4 hours, and then mixed for 15 minutes with the Beach House milkshake mixer to create a pulp suspension, and allowed to stand for 30 minutes to permit thefluorescentdye and the fibres to reach equilibrium (Bennington et al, 1996). The fibres were then filtered with a metal mesh to be certain that none of thefluorescentdye would adsorb on the filter (which could occur with a paper filter). Thefluorescentintensity was then determined for both sets of samples by using the fibre optic probe on thefluorometer.The  -24intensity can then be related to the concentration, and the concentration of the dye that was adsorbed on the fibres can be calculated by comparing the first set to the second set. Section 4.3.1 describes how the data was analyzed.  3.5 Reaction Between the Fluorescent Dyes and Typical Bleaching Chemicals The reaction between the fluorescent dyes and C10 , H 0 , and NaOH was determined 2  2  2  by using the "time drive mode" on the LS50B fluorometer. As discussed in section 3.1, this mode records the concentration at every time interval (every second for 10 minutes was utilized), which shows how the concentration of the fluorescent dye decreases in real time to determine the reaction rate constant.  This was done by first preparing a small aqueous sample of  fluorescent dye with a concentration of 5 ppm.  The time drive was activated, and a  concentration of bleaching chemical was added (at t=0) and quickly mixed with a stirring rod. It was allowed to run for 10 minutes. The amount of bleaching chemical added is shown in Table 2 in the initial concentration column. The distilled water before the addition of the bleaching chemical and the fluorescent dye was placed on a hot plate and heated to a desired temperature.  This was done for a few  temperatures so that the activation energy can be determined. The temperatures that were used for the H 0 and NaOH systems were 20 °C, 30 °C, 40 °C, and 55 °C, and for the C10 system 2  2  2  the temperatures were 10 °C, 20 °C, and 40°C (since the reaction between C10 and the 2  fluorescent dyes was very fast at higher temperatures as will be discussed in the Results and Discussion section). The 10 °C was obtained by placing the sample in an ice water bath. Once the activation energy is determined the reaction rate constant at bleach plant temperatures (60-80  -25°C)  can be determined. A full discussion of how the data was treated is given in section 4.4.1.  Table 2: Typical conditions and chemical concentration in the bleach plant. Bleaching Chemical  Bleaching Stage and Reaction Conditions Stage Designation  C (ft)  Temperalur  •lIBlllilll <°C)  C10  70 70 70 70  4 4 3.5 4  2.0 1.5 0.8 0.3  0.0092 0.0303 0.0161 0.0060  (E+O+P) P  12 12  80 75  10 10.5  0.25 1.0  0.0100 0.0401  E  12. 12  70 60  10.8 10  3.5 3  0.119 0.102  0  0  2  2  2  Tnitial Cone. (mol/L)  3 12 12 12  D D Di D  2  H 0  Chemical Charge (% wt/wt)  NaOH (E+O)  3.6 Construction of Fibre Optic Probes Two  fibre optic probes were constructed. The fibre optic probe that comes with the  Perkin Elmer LS50B fluorometer is not an immersion probe, and can only scan on top of the samples. The probes that I constructed had the capability of being inserted into harsh chemical environments. The Perkin Elmer fibre optic probe arranges the fibres such that the fibres are bundled into a cylinder tubing. The probe is divided in half from the middle, and one half of the optical fibres carry the excitation light, and to the fluorometer. My  the other half take the emitted fluorescence back  probes slightly deviated from this method. I arranged the optical fibres  in a similar manner to that done by Ouellet et al. (1996). In this technique, the optical fibres  -26that carry the excitation light are placed around the circumference of a large diameter optical fibre. This large diameter fibre takes the emitted fluorescence back to the fluorometer. The small fibres that carry the excitation light have a diameter of 0.02 cm, and have nylon jackets around the optical fibres, which brings the total diameter up to 0.039 cm. The large optical fibre has a diameter of 0.1 cm, and it also has a nylon jacket around it, which brings the diameter up to 0.13 cm. Both the small and large diameter optical fibres were supplied by Fiberguide Industries (Stirling, NJ). The optical fibres that were purchased were the Superguide G UV-visible fibres with nylon jacket to protect the fibres from excessive heat (up to 120°C). The first probe that was constructed had one large optical fibre in the middle, and 25 small fibres around its circumference. 25 optical fibres were used to make the surface area of the excitation fibres equal to that of the emission's optical fibre. That is, the excitation light should be kept constant in relation to the emission surface area. A second probe was constructed to measure the effects of cross-sectional area on the emitted fluorescence intensity. The second probe had three times the cross-sectional area of thefirstprobe; therefore, the second probe had three large fibres in the centre (to carry the emitted fluorescence), and 75 fibres around the three large fibres. The fibres were glued together with epoxy, which took up some space between the fibres. The fibres were encased in a 15 cm long stainless steel tube at the end of which was a sapphire window, which was purchased through Edmund Scientific (Barrington, N.J.). This window transmits ultra-violet light, and it can withstand high temperatures and harsh chemicals. A schematic of the end on view of all three fibres is shown in figure 3, and the dimensions are given in Table 3. Once the probes were constructed, the relationship between the cross sectional area of  -27the fibres fJiat carry tbe emitted light back to the fluorometer and the emitted light intensity can be found. This is basically the relationship between sample volume and emitted light intensity. It was found that the emitted light intensity has a linear relationship with the cross sectional area of the receiver or exciter optical fibre as shown in Figure 4.  Table 3: Dimensions of Perkin Elmer probe, probe 1, and probe 2. Type of Probe  Perkin Elmer Probe  Probe 1  Probe 2  Total Diameter (cm)  1.0  0.386  0.824  0.196  0.00785  0.0236  Dense fibre bundle  0.1  0.1  0.13  0.13  0.02  0.02  0.039  0.039  Excitation Surface Area = Detection Surface Area (cm ) 2  Large Fibres Small Fibres  Inside Diameter (cm) Outside Diameter (cm) Inside Diameter (cm) Outside Diameter (cm)  Dense fibre bundle  Perkin Elmer Probe  Small fibres (excitation)  -28Figure 3: Cross-sectional view of Perkin Elmer probe, probe 1, and probe 2 with a scale of approximately 5x.  Surface Area of Excitation Optical Fibrefs) (m ) 2  Figure 4: The cross-sectional area of the excitation optical fibres is directly related to emitted light intensity. 3.7 Hand Mixing Experiments Hand mixing is done by putting a fibre suspension of a given consistency in a 1 Litre beaker, and then kneading the pulp with your hands at a constant rate in order to maintain a constant energy input. The energy exerted by hand mixing per unit time estimated by measuring the increase in pulp temperature after mixing the pulp for a given amount of time. This was done by applying the following equation E=AQ=mC AT p  (3)  where m is the mass of the sample, Cp is the specific heat capacity in J/kgK (the heat capacity  -29of water was used), and T is the temperature (Felder et al., 1986). Heat transfer from body temperature of my hands to the pulp suspension was minimal. This was determined by placing my hands (no mechanical action) in the pulp suspension for 300 seconds, and there was no change in pulp temperature. The hand mixing trails were carried out by first making four pulp samples all at the same consistency. A certain amount of concentrated aqueous fluorescent dye is added at one spot to a sample. The sample was hand mixed for 300 seconds. The sample was then put under the Perkin Elmer probe and the surface was scanned. Fifteen locations on the surface were read for fluorescence intensity and repeated for three more samples, but each sample was mixed for a different amount of time (90 seconds, 30 seconds, and 5 seconds). There were only 15 points that were scanned because 15 points could be scanned with 100% confidence that none of the points were scanned more than once. This was done at 3%, 5%, and 10% consistency.  3.8 On-line Mixing Tests On-line mixing tests with all three probes were done using a stirred tank mixer. A schematic of the apparatus is shown in Figure 5. The stirred tank used was 35 cm long and 20 cm in diameter. The pulp suspension was 30 cm high in the tank. The mixing device has three blades that are shaped similar to airplane impellers, and each blade has an r/R=0.5. The mixing rotor was z/H=0.333 from the bottom of the tank. The Perkin Elmer probe was placed just above the surface and at a radial location of r/R=0.25 (shown as point M l on Figure 5). My first and second probes were immersed to a depth of z/H=0.10 into the pulp suspension and at the same radial location as the Perkin Elmer probe (shown as point M2 on Figure 5). The dye  -30was added 10 cm below the surface of the fibre suspension and a radial location of r/R=0.5. The on-line mixing trials were carried out with water and FBK suspensions at C = 1 m  and 2%. Fluorescence was measured at 1 second intervals, which was fairly rapid compared with the pulp flow, which was moving at 0.4 cm/s. This gives a mixing scale of 0.9 cm (talcing the Nyquist sampling criterion into account), which is in the fibre scale range. The mixer was then turned on. A concentrated aqueous solution of fluorescent dye was added as a pulse at the dye insertion point, and the time drive in the fluorometer is turned on simultaneously. The run lasted for 900 seconds.  -31-  10  cm  i 20  cm  Figure 5: Schematic of mixing apparatus used for on-line mixing test, where M l and M2 show the probe locations.  3.9 Comparison with Lithium Tracer Tests Test were carried out with the same stirred tank as the one described in section 3.8. First a pulp suspension was made at 2% consistency. The fibre suspension was treated with 1 M HC1 to ensure that none of the fluorescent dye and lithium ions will not adsorb on the fibres (Bennington et al., 1996 b). The fibre suspension was then placed into the mixer, and the mixer started. A solution containing concentrated aqueous fluorescent dye and concentrated aqueous lithium ions were inserted to the mixer at the tracer insertion point, and the timer was started.  -32The solution was allowed to mix for 100 seconds. 10 samples were withdrawn from the mixer each containing about 25 mL of pulp suspension. Each sample was then filtered with a metal mesh to be sure that no dye or lithium ions were removed by adsorption on to the filter. Once each sample was filtered, it was centrifuged for 15 minutes to settle any of the fibres that remained in suspension. This was done to ensure that the atomic adsorption unit, which was used to measure lithium concentration, would not plug up with fibre. Each sample was first placed underneath Perkin Elmer's fibre optic probe, and its fluorescent dye concentration is determined. The lithium concentration is then determined by atomic adsorption using Perkin Elmer model 2380 spectrophotometer (Vancouver, B.C.).  This was also done for a mixing time of 300  seconds, and at 1 % consistency. The lithium ion were from lithium chloride crystals that were ordered from Anachemia. These results were also compared with the results from the on-line mixing test as will be further described in the Results and Discussion Section.  -33-  4. RESULTS AND DISCUSSION 4.1 Finding Appropriate Dyes Similar to the seven criteria that needed to be met when choosing the ideal method to assess mixing (section 2.2), there are six criteria that need to be met when choosing the ideal fluorescent dyes:  1.  In order to apply this method in a mill, the end product must be unaffected. This means that a colourless tracer needs to be found because the tracer cannot colour the pulp in any way. Furthermore, the fluorescence must occur in the ultra-violet range. This ensures that if any tracer is left on the fibres after the paper is made, it cannot be detected with the human eye.  2.  The dye's fluorescent intensity should have a monotonically increasing relationship to the dye concentration. This will ensure that a fluorescent intensity is only representative of one concentration.  3.  The fluorescent tracer must be very sensitive. This will minimize the amount of tracer needed, and maximize the signal; furthermore, a highly sensitive dye is needed in order to be detected by the fibre optic probe.  4.  The dye should be non-toxic because the pulp maybe used in food packaging (Roy, 1995).  5.  The dye should have a high water solubility to allow for a concentrated dye solution to enter the pulp suspension before the mixer. This will minimize the amount of water that will enter the process.  6.  There should be consideration given to the cost of the dye in a mill run. It is desired to run the tracer test for a significant amount of time, that is in the order of 1 hour. This means that the dye should be relatively inexpensive and highly sensitive. This will reduce the overall tracer cost. The dyes with the lowest cost/hr would be considered.  Table 4 summarizes how all these factors effect the dyes that were tried. The first column on Table 4 states whether the dye would change the end product, which shows if the dye  -34coloured the pulp in any way, and if excites or emits in the visible range (400 - 700 nm). The actual excitation wavelengths and emission wavelengths are given in table 1 on page 21. The next column states whether the relationship between the fluorescence intensity and dye concentration was monotonically increasing. The third column shows the sensitivity in ppm, which is the ppm range that is needed to obtain a fluorescence intensity of 50 with the Perkin Elmer fibre optic probe attachment. Next, the table states whether the dye is toxic or not. The solubility is shown in the next column, and the last column states the approximate dye cost to perform a 30 minute mill trail in the bleach plant. The first dyes purchased were rhodamine 6G, anthracene, acridine orange, rhodamine B, and fluorescein. All these dyes coloured the pulp, and all fluoresced in the visible range except for anthracene. However, anthracene was not soluble in water. The next set of dyes tested were purchased from a company that specialized in fluorescent dyes (Molecular Probes; Eugene,  OR).  These  dyes  were  1-pryrenebutanol, 2-ethoxy-l-ethoxycarbony 1-1,2-  dihydroquinoline (EEDQ), and 7-acetoxy-4-(bromomethyl)coumarin (Br-Mac). All these dyes were very expensive, and they had low sensitivity.  EEDQ did not have a monotonically  increasing relationship between fluorescent intensity and concentration. This is shown on Figure 6 (A).  The next set of dyes purchased were azaindol mephenesin, piperoxan,  naphthaleneacetamide, naphthaleneacetic acid, and 2-naphthalene sulphonic acid (2-NSA). All these dyes work very well, since they met all the criteria. Figure 6 also shows the relationship between intensity and concentration for azaindol, anthrcene and piperoxan. It can be seen on this figure that azaindol is the most sensitive dye and EEDQ is the least sensitive.  Figure 7  shows the how thefluorescentintensity changes with concentration for Mephenesin and 2-NSA;  -35-  the precise linear fit should be noted along with how 2-NSA is far more sensitive than Mephenesin. By taking all six of these criteria into account, two fluorescent tracers were selected for further study: mephenesin and 2-NSA. The chemical structure of these dyes is given in Figure 8. As can be seen from this figure the chemical structure of both dyes are very similar to that of the lignin monomer. This brings to mind the question of how these dyes will react with the bleaching chemicals. investigated.  This, along with how pulp fibres affect the fluorescence, needs to be  iffs  o o o  o o o o  00  SSSS  XI  3  o  3 O  o o o o > / >  o o  o o o  o  <N  _tD  <u  3  3  "3  3  "5b  00  "So  o  o o  o  CO Ui  o  •a  X  o  0  o  1  c  X! 00  o  z  z  3  •a  X o  o  c o Z  o  •a 0  1  c  z  ill  AS  ,  «*  CD  o X tu  i i i  •  <$} <0  X CU  CU  « a*  III « O  tu X  CD  2. x  '3  'o  tu  a>  . & u ?  03  O . Ui 3 _tu CV,  tU  fX  3  "3 o  "3 o  "3 o  CO  CO  i>  tu  >-  tu  60  a '3 o  3  o  Z  z  tu  c  8  o  o  u,  CU  co  3  o  tu  *_ S>  urs  _g. U l  If « V  CO *—>  tU  <u  <u  tu  VlSl  CU  E g c«  z  g  2 3  X>  cu  Ul  >,  a q  o  a Q  w w  8 CD  X!  P. CU  s _eu  •s c  ea  a  CU  tu  _« tu  "2 "3  c«  CO  -a  xi R. O CO  Z §•  Z  <"  <N "  j-  §1  ^ 00  1  1  1 O  •  1 v->  1  I • O  1  I ' I • I—'—I—'—I— —I—I—I—I—I—I" « O O v > O v - v O v - » o 1  <D — I<  3$  GO  Zi  CN  -a C/5  CD  c . cx CD  g *+-» is c CD O  e o o  -a c  1 c  CU CD CD  J5 C  o  <D )-. 3 bp  -39-  ^-SOjNa  /  \  OH OH  | |  OH—v'  (/\ — - /V-0—CH-CH-CHj  o  N  Mephenesin  2-NSA  /  CH3O'  /  =  y-CH—CH-CH OH  V  2  =  =  /  lignin monomer  Figure 8: Chemical structure of Mephenesin, 2-NSA, and the softwood lignin monomer.  4.2 Using Fluorescent Dyes With Pulp Fibres Both mephenesin and 2-naphthalene sulphonic acid are highly sensitive, which enables them to be detected with relative ease with a fibre optic probe; however, the effects of pulp fibre suspensions are still unknown.  It needs to be determined how pulp fibres will effect the  fluorescence, and if concentration still has a linear relationship with the fluoresced light intensity. Experiments were done at low and medium consistency with mephenesin.  The  relationship with concentration and fluorescent intensity was still linear with a R =0.903. 2  However, the overall intensity did go down substantially. At 3% consistency, the signal was about 12.5% of the signal with no pulp fibres, and at 9% consistency, the signal dropped to 4% of the signal with no pulp fibres.  This effect is due to the net effect of adsorption and  quenching. The adsorption of fluorescent dye on to the fibres will be fully investigated in the  -40next section. 2-naphthalene sulphonic acid (2-NSA) was one of the most sensitive fluorescent dyes I worked with. Also investigated was how 2-NSA fluoresced in the presence of pulp fibres. Again the relationship between concentration and fluorescence intensity was very good (R = 2  0.90); however, the intensity did go down. This time the signal drop was not as dramatic as it was with mephenesin. 90 % of the signal was maintained at 3% consistency. The signal at 9% consistency was about 60% of the intensity with no fibres. Again a complete study on how 2naphthalene sulphonic acid adsorbs on the fibres will be done in section 4.3. Figures 9 and 10 show how the intensity drops as consistency increases for both mephenesin and 2-naphthalene sulphonic acid. It is evident from these figures that 2-naphthalene sulphonic acid is superior in signal strength, and maintains most of its signal strength in the presence of pulp fibres. The effect of lignin on the fibres was also investigated since lignin is known to fluoresce (Olmstead et al, 1995). 10% consistency pulp samples were used with 4.2% lignin (28 Kappa) in them, which is the amount of lignin that would be in the pulp stock entering the 1st stage of a conventional bleach plant. Mephenesin excites at 280 nm and fluoresces at 304 nm, and 2naphthalene sulphonic acid excites at 277 nm and fluoresces at 334 nm. Figures 11 and 12 show 28 Kappa pulp have negligible change in the fluorescence intensity from FBK fibres. Also shown on these figures is two concentrations of fluorescent dyes mixed with 28 Kappa pulp. The results indicate that the same fluorescence intensity value was obtained from the dye in 28 Kappa pulp and FBK pulp. One final area that was investigated was how much the black liquor in the aqueous phase fluoresces compared to the background fluorescence of water. It was found that the black liquor did not fluoresce at the excitation and emission wavelengths for both  -41mephenesin and 2-naphthalene sulphonic acid. In summary, it is still possible to detect fluorescence with high accuracy in the presence of pulp fibres. The fluorescence intensity still has a monotonically increasing relationship with concentration; however, the overall intensity does drop in pulp fibre suspensions. Lignin was found to have no effect at the wavelengths that mephenesin and 2-naphthalene sulphonic acid fluoresces. The extent of adsorption will be studied in detail in the next section.  -42-  Figure 10: Effect of consistency on the fluorescence intensity for 2-NSA.  -43-  1  -r-  1  1  1  /  &9  •  1  '  i  1  •  ^ \ 2 8 Kappa pulp (20ppm)  -//  N28 Kappa pulp (lOppm)  j 1•  ;  •—  === 2§Kapj» pulp (Oppm)  -  =  FBK. pulp (Oppm)  1  i  I  300  305  310  .  I  .  •  I  315  325  320  Wavelength (nm) Figure 11: Fluorescence of lignin at 304 nm with Mephenesin in 10% consistency pulp suspension (excitation at 280 nm). 60 j  1  1  1  1  1  •  1  >  1  •  1  >  1  Wavelength (nm) Figure 12: Fluorescence of lignin at 334 nm with 2-NSA in 10% consistency pulp suspension (excitation at 277 nm).  -44-  4.3 Adsorption Equilibria 4.3.1 Background on Adsorption Equilibria The adsorption of a substance from one phase to the surface of another in a specific system leads to a thermodynamically defined distribution of that substance between the phases when the system reaches equilibrium; that is, when no further net adsorption occurs. The common manner in which to depict this distribution is to express the amount of substance absorbed per unit weight of adsorbent, q , as a function of the residual equilibrium e  concentration, C , of substance remaining in the "solution" phase. In our situation q is the moles e  e  of fluorescent dye absorbed per weight of pulp fibre (oven dried), and C is the moles of e  fluorescent dye at equilibrium in the water phase after the fibres are removed. A plot of q  e  versus C is termed an adsorption isotherm, and it defines the functional equilibrium distribution e  of adsorption with concentration of adsorbate in solution at constant temperature (Weber, 1985). There are several models that describe different adsorption isotherms, and they are described in detail by Weber (1985). However, the linear model best describes the isotherm at the concentrations of concern here (ppm range) in addition to being the simplest isotherm model. The linear isotherm is given by:  where K is the partition coefficient. This model has the advantage of describing a given set of p  adsorption data in terms of a single parameter, K p , for simplicity of modelling purposes. The  -45linear partitioning model is not solely a mathematical convenience without basis in theory. All the other models discussed by Weber in 1985 can be reduced to linear relationships under special conditions. However, it should be noted that the linear relationship is seldom valid over large ranges of concentration, and it should not be used for extrapolation beyond the limits of a particular data set.  It was experimentally determined that for the concentrations that are of  concern in this situation, the linear model fits the data point. This is shown in section 4.3.2.  4.3.2 Adsorption of Mephenesin and 2-Naphthalene Sulphonic Acid The adsorption equilibria for mephenesin and 2-naphthalene sulphonic acid are shown on Figure 13 (A) and (B). The partitioning coefficient for mephenesin is 0.006 L/g NSA it is slightly lower at 0.0045 L/g  fibre  fibre  , and for 2-  . The effect of adsorption on fluorescent intensity is  further decreased for 2-NSA since it is ten times more sensitive, and less dye is adsorbed at lower concentrations. For example, at 3% consistency only 17% of the dye was adsorbed at 4 ppm, and at 10% consistency only 38% of the dye was adsorbed at 4 ppm. This contrasts to mephenesin where 27.5% of the dye is absorbed at 40 ppm with 3% consistency, and at 10% consistency, 42.5% of the dye is absorbed at 40 ppm. The confidence interval was also determined for both Figure 13 (A) and (B) as is shown by the error bars on the figures (cr ). For mephenesin the error in each point was 5%, and for 2  2-NSA the error was 1%. Also in Figure 13, the mass of pulp fibres were normalized, which was achieved by dividing the moles dye adsorbed by the mass of fibres in the sample. Figure 14 shows that as consistency increases the amount of dye adsorbed will increase, which is done by not normalizing the mass of pulp fibres.  -46The question is how does the adsorption effect the assessment of mixing quality. It is evident from the previous discussion that both fluorescent dyes adsorb on the fibre surface, and from section 4.2 it is also evident the presence of fibres quenches fluorescence. However, when mixing is being assessed, we really want to know the distribution of tracer concentration due to mixing by hydrodynamic or diffusional effects. We can either correct for fibre distribution by knowing q as a function of consistency, and then calculating the mixing index from equation e  (2), or by sampling on a sufficiently large scale to avoid the effect of non-uniformities due to fibre distribution (mixing scale of 10 mm).  -47-  ure 13: (A) Adsorption isotherm of mephenesin with the error bar depicting sigma. (B) Adsorption isotherm of 2-NSA with the error bar depicting sigma.  -48-  -10  10  0  20  30  40  50  60  70  80  C (mol/L) e  O  S  10  15  20  C (mol/L) e  Figure 14: Adsorption at 3%, 7%, 10%, and 15% consistency without normalizing the mass of fibres present for (A) mephenesin and (B) 2-NSA.  -494.4 Effects of Bleaching Chemicals on Fluorescent Dyes 4.4.1 Background on Kenetics of Fluorescent Dyes and Bleaching Chemicals As shown in Figure 8 (section 4.1), the chemical structure of Mephenesin is very similar to the monomer of lignin. Furthermore, 2-NSA is also an aromatic hydrocarbon. This makes both mephenesin and 2-NSA susceptible to attack by bleaching chemicals. The rate at which the fluorescent dye will react with each bleaching chemical needs to be determined. The reaction between the bleaching chemical (B) and fluorescent dye (D) can be shown by equation (5) , and the reaction between the bleaching chemical and lignin (L) can be shown by equation (6) . k  (5)  D+B—DB  k L+B—LB (6) The consumption of a fluorescent dye by a bleaching chemical can be characterized by equation (7) , and the consumption of lignin by a bleaching chemical can be characterized by equation (8).  4Z>1  =* [5][0]  dt  C7)  D  dm k [L][B] dt  (8)  L  Where [D] is the concentration of the fluorescent dye, [B] is the concentration of the bleaching chemical, [L] is the concentration of lignin, t is time, and kj and k are probably 2nd order L  -50rate constants.  Ideally, we want a tracer dye that is inert, that is k =0. D  However, this is  unlikely for the dyes studied here since the chemical structure of lignin is so similar to the dye structure. Therefore, some reaction is expected.  Ideally, the reaction rate constant for the  fluorescent dye (kj) should be much smaller than the reaction rate constant of the lignin (kj because typical concentrations of bleaching chemicals were used (see Table 2 in section 3.5) and ppm levels of the tracers were used; therefore, the concentration of the bleaching chemical does not change much, and can be assumed to remain constant.  This yields equation (9) from  equation (8).  ^=k' [D] dt  ( 9 )  D  where k ' is given by D  k' =k lB] D  (10)  D  and [B] is assumed constant. Equation (9) can be integrated to give equation (11).  (ID  ,  [D ] is the initial concentration of the fluorescent dye at time 0 s. k ' is the slope of a ln [D] 0  D  versus t plot. k ' can be used to determine the reaction rate constant, k , by using equation (10). D  D  The reaction rate constant, k , is a strong function of temperature, and in the lab it is D  difficult to recreate exact temperature conditions of the bleach plant (see table 2 in section 3.5). However, it is possible to conduct experiments at lower temperatures, and use the Arrhenius equation to determine k at typical bleach plant temperatures. The Arrhenius equation is shown D  in equation (12).  -51(12)  k =Ae-  EJRT  D  where A (s ) is the preexponential factor or frequency factor, E is the activation energy in 1  a  J/mol, R is the gas constant which is equal to 8.314 J/mol, and T is the absolute temperature in degrees Kelvin. The activation energy, E , can be thought of as the minimum energy that a  must be possessed by reacting molecules before the reaction will occur. Therefore, by taking the natural logarithm of equation (12), a working equation can be used to determine the activation energy of a particular system. This equation is given by the following expression.  (13) From this equation a plot of In  versus 1/T should be a straight line where the slope is  proportional to the activation energy. Thus by conducting experiments at different temperatures, it is possible to determine both E and A, which allows k to be estimated at any temperature a  D  (Fogler, 1992).  4.4.2 Reaction Kinetics Between Fluorescent Dyes and Bleaching Chemicals The second order reaction rate constants (k ) were calculated for C10 , H 0 , and NaOH D  2  2  2  at various temperatures using concentrations typical of bleach plants, and the activation energy constants were determined for both mephenesin and 2-naphthalene sulphonic acid. For example, data was collected by the method described in section 3.5. The raw data for the decay of 2-NSA with H 0 is shown in Figure 15 (A). The reaction rate constants were found by taking the 2  2  natural log of the concentration. The data was then plotted on a In [D] versus time graph, which  -52should yield a straight line. The result was a straight line for most of the data. The slopes were calculated for each line, which is k '. Equation (10) was then used to determine k . For the D  D  reaction at 55°C, the slope was steep at first, and then tapered off, which could be due to poor mixing in the reaction vessel. For these cases the initial slope was taken. This is shown on Figure 15 (B). The natural log of the rate constants were then plotted against the inverse of the temperature in Kelvins.  This again yields a straight line, as shown in Figure 16, and the  activation energy was found from the slope. Similar tests were done with C10 and NaOH using both mephenesin and 2-NSA and it 2  is given Appendix III. The rate constants are summarized in Table 5, and the activation energy and the frequency factor are summarized in Table 6.  -53Table 5:  Summary for rate constants and activation energies for mephenesin and 2-naphthalene sulphonic acid.  Bleaching  [B]  Temp  Chemical  (mol/L)  (°C)  Mephenesin  2-NSA  0.0267  10  450  250  20  550  460  40  2500  590  20  0.42  0.55  30  0.49  1.2  40  2.6  4.7  55  11  12  20  0.016  0.0034  30  0.30  0.0081  40  0.90  0.013  55  1.8  0.041  C10  H 0 2  2  0  0.0353  2  0.105  NaOH  Table 6:  -k (s ) 1  D  Summary of activation energies and frequency factors for mephenesin and 2naphthalene sulphonic acid.  2  1  a  Mephenesin  2-NSA  Mephenesin  2-NSA  0.044  0.047  10  1.2xl0  7.2xl0  10  2  0.080  0.074  6.5xl0  13  7.7xl0  12  2  0.12  0.056  7.0xl0  C10 H 0  A (s )  E (MJ/mol)  Bleaching Chemical  NaOH  19  2.8xl0  8  -54Using the data in Table 6 it is possible to calculate the half life of mephenesin and 2-NSA under typical bleaching conditions summarized in Table 2 (section 3.5).  The half life is  calculated by utilizing equation (14).  ln(0.5) *l/2  Ae-  (14)  [E\  EJRT  The half lifes are summarized in Table 7.  Table 7:  Typical half life of mephenesin and 2-naphthalene sulphonic acid with each bleaching chemical.  Bleaching  Stage  2  Half Life (s)  Temp (°C)  (mol/L)  Chemical C10  [B]  Mephenesin  2-NSA  D at C =3%  0.0092  70  0.83  0.40  D at C =12%  0.0303  70  0.88  0.42  D,  0.0161  70  1.1  0.50  D  2  0.0060  70  2.8  1.4  P  0.0401  75  20  18  E  0.102  65  20  440  0  m  0  m  HO 2  z  NaOH  -55i  •  r  1  1  1  r-  (A)  I Temperature, ° C • 20 A 30 • 40  3v  •  55  • •  M.  •  • 0  '  i 100  '  1 200  1  1 300  >  A AAA A  A  •  1 400  1 500  1  , —  600  Time (Seconds)  Figure 15: Reaction between 2-naphthalene sulphonic acid and H 0 shown as (A) raw data and (B) determination of rate constants. 2  2  -56i  •  i  •  1  1  1  •  1  1  In(k) = lr<A)-E ^/R(1/IV\ A = 7.7E12s-i E = 7.4E4 J/mol  \ . "  N.  a  r2=0.9764 1 0.0030  1  \ 1 0.0031  >  1  >  0.0032  1 0.0033  •  1 0.0034  1  0.0035  l/T(K-i)  Figure 16: Detemunation of activation energy for 2-naphthalene sulphonic acid and H 0 . 2  2  From Table 7, it is evident that C10 reacts very quickly with both mephenesin and 22  NSA.  Consequently, the use of a fluorescent tracer can not be used in a C10 environment. 2  Furthermore, NaOH reacts very slowly (half life = 440 seconds) with 2-NSA, which shows that 2-NSA can be used in the extraction stage of the bleach plant with little interference from NaOH. H 0 reacts at about the same rate with mephenesin and 2-NSA. The rate of reaction 2  2  in the peroxide stage is quite slow (half life =18 seconds); however, there would still be some reaction even after a short period of time. In the presence of chemical reaction, there is a problem of using fluorescent dyes in a bleaching environment, since it is difficult to realize how much of the variation in mixing is due to the distribution of the dye, and how much is due to the reaction between the bleaching  -57chemical and the fluorescent tracer. Furthermore, since the reaction is so fast for the D , D,, 0  and D stage, the previously described method will not be useful because all the fluorescent dye 2  would have reacted by the time it reaches the probe. For the these bleaching stages perhaps the temperature tracer (Section 2.3) is a possible technique to use to assess mixing quality, since C10 is significantly colder than the pulp, and fast thermocouples can be used to sense the 2  variation in temperature. Another method that can be used is to still utilize the fluorescent tracer, but this time put the tracer into the pulp before the mixer, and then after the mixer measure the fluorescent dye that has not reacted. In this method, good mixing is evident when there is no signal. However, there are problems with this method; such as, it is impossible to ensure that the fluorescent dye is perfectly mixed with the pulp before addition of the C10 . Although the reaction of the 2  fluorescent dyes with bleaching chemicals does complicate the problem of assessing mixing quality, it is still possible to assess mixing in the bleach plant, especially in the extraction stage with 2-NSA.  -584.5 Assessing Mixing With Fluorescent Dyes as a Tracer 4.5.1 Relationship of Mixing Quality and Mixing Energy in Hand Mixing Hand mixing tests were conducted to give an indication on how the mixing index relates to energy dissipated. The results for the hand mixing test are shown in Figure 17. It was found that as energy increased the mixing index decreased as expected. A lower mixing index (see equation 1) corresponds to better mixing, so as the energy dissipated increases, the better the mixing. The relationship was exponential on both axis. The results from my hand mixing test yielded the following equation. M=0.2E-° i? =0.79 5  ( >  2  15  Where M is the mixing index and E is the energy dissipated in MJ/t. This is very similar to the equation obtained by Bennington et al. (1996 b) given by the following equation. Af=0.40£-°  51  fl =0.65 2  ( > 16  This equation was obtained based on work done on a series of batch-operated laboratory mixers (including a repulper, MC lab mixer, a high-shear mixer called a Fluidizer, and a Hobart mixer).  -59-  Figure 17: Relationship between mixing index and energy dissipated found by hand mixing mephenesin, and the error is calculated using the standard deviation.  -60-  4.5.2 Continuous On-line Mixing Tests of a Batch Mixer Continuous on-line mixing tests were performed with all three probes using the method described in section 3.8. The vertical movement of the fluid in the mixing vessel was noted, and it is shown in Figure 18. The fluid was also moving in a circular motion in the same direction as the impeller was rotating, and there was faster movement closer to the impeller. One result of the mixing test with the Perkin Elmer probe is shown in Figure 19. Similarly, one result with Probe 1 (see section 3.6) is shown in Figure 20, and one result with Probe 2 is shown in Figure 21. These three graphs show a sample of the mixing tests performed. Tests were done at 0%, 1%, and 2% consistency and repeated four times.  These are shown in  Appendix V, and a study of the amount of noise in the fluorescent signal is given in Appendix IV.  -61-  2) T  1  ^ ^ ^ ^ ^  I ^gggglgfBr  Figure 18: Observed movement of the fluid in the stirred tank.  0  200  400  600  800  Time (Seconds) Figure 19: On-line mixing test result using Perkin Elmer Probe at 1 % consistency and the rotor stirring at 150 rpm.  -62-  200  400  600  Time (Seconds) Figure 20: On-line mixing test result using Probe 1 at 1% consistency and the rotor stirring at 150 rpm.  200  400  600  Time (Seconds) Figure 21: On-line mixing test result using Probe 2 at 1% consistency and the rotor stirring at 150 rpm.  -63It is evident from this work that the change in dye concentration is being detected with the fibre optic probes. This is because the concentration of the fluorescent dye begins with high fluctuations, which corresponds to a high mixing index hence poor mixing. As time passes, the fluctuations in concentration evens out giving a lower mixing hence good mixing. This can be quantitatively assessed by mixing time ( T J . By observing Figures 19, the mixing time is found to be about 160 seconds. Similarly, the mixing times are 350 seconds and 370 seconds for Figures 20 and 21 respectively. Furthermore, for each trail the concentration of the fluorescent dye changed differently. The average mixing times are summarized in Table 8.  Table 8: Summary of mixing times. Probe  T (s) m  C =0%  C =l%  C =2%  Perkin Elmer  44 ± 4 . 8  154 ± 3 0  236 ± 1 4 0  Probe 1  50 ± 7 . 1  200 ± 1 0  328 ± 6 8  Probe 2  51 ± 6 . 3  383 ± 2 7 0  220 ± 5 4  ra  m  m  Each probe gave a slightly different resolution. The Perkin Elmer probe detected the highest intensity in fluorescence due to the larger surface area; however, the Probe 1 measured the intensity of the smallest area (0.1 cm). However, it may be advantageous to have a probe with a large surface area (in the floe size range of 1 cm), which will average out the effect of concentration fluctuations due to floes passing the tip of the probe. This is because the pulp fibres adsorb the fluorescent dye and quench fluorescence (see section 4.2). Another problem  -64encountered with the probes that were immersed (Probe 1 and Probe 2) in the 2% consistency mixtures was that quite often a floe would get caught on the tip of the probe, which would cause the fluorometer to read the same value of concentration until enough force was produced by the suspension flow to push the floe forward. This can be overcome by designing the tip of the probe at an angle in the direction of the flow, and with smooth edges so that no fibres will get caught on sharp edges. A schematic of this conceptual probe is shown in figure 22.  flow direction  glass cover probe tip stainless steel cladding  Figure 22: Conceptual probe designed to facilitate pulp floes to move over the probe tip easily.  -654.5.3 Comparison Between Fluorescent Tracer and Lithium Tracer In order to verify assessing mixing with a fluorescent tracer, it is important to compare the results with a method that is widely accepted such as the lithium tracer. This was done experimentally by the method that was described in section 3.9. The results are given in table  Table 9: Comparing the mixing index for the fluorescence method and the lithium method. Mixing Index, M  Trial 1  Trial 2  Trial 3  Trial 4  Trial 5  300rpm C =2% t=300s  Batch Fluorescence  0.27  0.19  0.29  0.31  0.17  0.09  0.10  Batch Lithium  0.29  0.14  0.04  0.08  0.27 ± 0 . 0 6 Averaged from all 4 runs (noise is taken into account, see appendix IV)  0.10  0.11  On-line tests with Probe 1  0.31 ± 0 . 0 8 Averaged from all 4 runs (noise is taken into account, see appendix IV)  0.11  0.14  On-line tests with Probe 2  0.30 ± 0 . 0 9 Averaged form all 4 runs (noise is taken into account, see appendix IV)  0.10  0.12  Method Used  On-line tests with Perkin Elmer probe  300 rpm; C =2%; t=100s m  m  150 rpm C =l% t=100s m  Average = 0.25 ± 0 . 0 6 0.15  0.28  0.30  Average = 0.23 ± 0 . 0 8  Table 9 shows that the fluorescence method and lithium method both gave very close results; for example, at 2% consistency and 100 s, the fluorescence method gave M=0.25 ± 0 . 0 6 and the lithium method gave M=0.23 ± 0 . 0 8 . The comparison between the fluorescence and lithium give very close results for each  -66-  trial. There was, however, variation in the mixing index between trials. This is due to slight differences in the movement of the fluid in the stirred tank. Table 9 also compares the mixing index obtained from the on-line tests discussed in the previous section. The mixing index was obtained from by taking 15 values before and 15 values after the particular point in time that is in question, and calculating the mixing index from these 30 values.  The values were then  corrected for noise in the signal by the method discussed in appendix IV, which shows that the variances due to background noise and floes were subtracted from the variance in the test. These values also correlated very closely with the mixing index becoming larger as the mixing volume became less. This was also found by Bennington et al. (1997). Furthermore, it verifies that the on-line mixing tests were indicative of mixing quality.  -675. CONCLUSIONS  The results of this study are summarized below.  1. Mephenesin and 2-NSA were selected as fluorescent dyes suitable for use in the mill. These two fluorescent dyes did not change the pulp, had a monotonically increasing relationship between concentration and intensity, were highly sensitive, non-toxic, water soluble, and relatively inexpensive.  2. Fluorescent dyes can be used with the presence of fibres giving a linear signal with aqueous concentration to a correlation coefficient of 0.95. However, the fluorescent intensity decreases as the consistency increases due to the combined effect of quenching and adsorption. The amount the fluorescent intensity decreases depends on the type of dye. It was found that 2-NSA's intensity decreased less than mephenesin's intensity.  3. It was found that both mephenesin and 2-NSA adsorbed on to the fibres, and it was concluded that the effects of adsorption could be corrected for by knowing q as a function of consistency, and then calculating the mixing index, or by e  sampling on a sufficiently large scale to avoid the effect of non-uniform mass distribution; that is, a scale larger than 100 mm, which is ten times the floe dimension.  -68-  4. Bleaching chemicals do react with both mephenesin and 2-NSA. C10 reacts very 2  quickly  with  both  mephenesin  and  2-NSA.  It  was  found  that  mephenesin and 2-NSA react at a moderate rate with hydrogen peroxide. Sodium hydroxide reacts with 2-NSA very slowly, which would make 2-NSA applicable to the extraction stage in mills.  5. The relationship between mixing index and total mixing energy was found using mephenesin and hand mixing.  The relationship was found to be a log-log  relationship, and the results agreed very closely to previous work done with a lithium tracer and various batch mixers.  6. Continuous on-line tests in a batch mixer verified that the spatial distribution of the fluorescent tracer was being measured. The results gave the mixing time, and the mixing index could be calculated at an given time.  7. The result of the on-line tests and sampling tests with the fluorescent tracer compared very closely to results with the lithium tracer. This verifies the use of a fluorescent tracer as a useful method to assess mixing quality.  In conclusion, the findings of these test verify that mixing can be assessed with the presence of fibres to assess mixing quality using a fluorescent tracer. This method can be useful  -69in assessing mixing in stock chests, in the headbox of the paper machine, and in the NaOH stage of the bleaching plant. The largest problem is posed by the bleaching chemicals, and this questions the applicability of this method in C10 and H 0 environments. 2  2  2  -70-  6. RECOMMENDATIONS FOR FURTHER WORK  Further work needs to be done on this topic to fully understand the applicability of this method in the pulp mill, which can be done in a two step process. First, a method needs to be investigated to combine the fluorescence probe with another set of optical fibres to be placed beside the fibres for fluorescence detection.  Half of these fibres will transmit an optical  wavelength, and the other half will take the reflected signal (at the same wavelength) to a photodetector.  This will measure the mass distribution of the fibres.  Then a computer  processing unit (CPU) could correlate this data to obtain a proper mixing index; that is, by fully compensating for the adsorption of the dyes on to the fibres, a proper mixing index can be found. The seconds step would be to do a mill trial. A proposed schematic of a mill installation is shown in Figure 18. This figure illustrates that the fluorescent dye would be added directly before the mixer. The sample port, which will insert the fibre optic probe, will be directly after the mixer. Furthermore, the probe will have both sets of fibres. One set of fibres will be for finding the spatial distribution of the fluorescent dye, and the second set is to find the mass distribution of the pulp fibres. The data is then correlated in a CPU, and the mixing index is found.  -71-  fluorescent tracer  pulp flow  mixer Xenon discharge lamp  AID  fluorometer  A/D  photodetector  M c t  Figure 23: Proposed schematic for mill installation of a fluorescent tracer method utilizing a fibre optic probe.  -72-  7. NOMENCLATURE  A  Preexponential factor [s ]  [B]  Concentration of the bleaching chemical [mol  C  m  Consistency of the pulp suspension [%]  C  e  Residual equilibrium concentration [mol/L]  1  bleachillg  [D]  Concentration of the fluorescent dye [mol /mol J  E  Energy exerted for mixing [MJ/t]  dye  ^^^/mol^  tot  E  a  Activation Energy [J/mol]  k  L  Reaction rate constant of the lignin [s ]  k  D  Reaction rate constant of the fluorescent dye [s ]  1  1  k' D  Reaction rate constant multiplied by the concentration of the bleaching chemical [molfcieachjng chemical/S molto^j]  K  p  Partition coefficient [mol  adsorbed  L/g mol] fibre  [L]  Concentration of Lignin [mol^/mol^,]  m  Mass of the sample, [kg]  M  Mixing Index  MM  Molecular Multiplicity  q  Moles of fluorescent dye absorbed per weight of the O. D. pulp fibre [mo^^/g^]  e  Q  Heat in the system [J]  R  Universal gas constant [8.314 J/mol]  -73s  Standard deviation  S  Spin quantum number  T  Temperature [K]  T  m  Mixing time [s]  Xj  Concentration at point i, or of sample i  x  Average concentration  -74-  8. REFERENCES  Andre, C , R. David, J. C. Andre and J. Villermaux, "A New Fluorescence Method for Measuring Cross-Fluctuations of Two Non-Reactive Conponents in a Mixing Device", Chem. Eng. Technol., 15, 182-185, (1992). Backhand, R., E. Bergnor, P. Sandstrom and A. Teder, "The Benefits of Better Mixing", Pulp and Paper Canada, 88(8), T279-T285 (1987). Bennington, C. P. J., "Mixing and Mixers ".Pulp Bleaching-Principles and Practice. C. W. Dence and P. W. Reeve, ed., TAPPI Press, 537-568 (1996). Bennington, C. P. J., "The Use of a Mixing-Sensitive Chemical Reaction for the Study of Pulp Fibre Suspension Mixing", The Canadian Journal of Chemical Engineering, Volume 71, October, 667-675 (1993). Bennington, C. P. J., C. M . Peters and R. MacLaren, " Characterization of Mixing Quality in Laboratory Pulp Mixers", PGRLR 651, Paprican, October (1996). Breed, D. B., "Discovering the Mechanisms of Pulp Mixing - A Pilot Approach to High Shear Mixing", 1995 Medium Consistency Mixing Seminar Notes, Atlanta, 33-37 (1985) Danckwerks, P. V . , "Theory of Mixtures and Mixing", Research, 6, 335-361 (1953 a). Danckwerks, P. V . , "The Definition and Measurement of some Characteristics of Mixtures", Appl. Sci. Res., Section A Vol. 3, 279-296 (1953 b). Elliott, R. G. and T. D. Farr, "Mill-Scale Evaluation of Chlorine Mixing", TAPPI, 5 6 ( 1 1 ) , 68-70 (1973). Felder, R. M . and R. W. Rousseau, Elementary Principles of Chemical Processes. John Wiley & Sons, Inc., 295-310 (1986). Fogler, H. S., "Elements of Chemical Reaction Engineering", Prentice Hall, Inc., 62-80 (1992). Francis, D. W. and R. J. Kerekes, "Measurement of Mixing in High-Consistency Pulp Suspensions", Journal of Pulp and Paper Science, 1 6 ( 4 ) , J130-J135 (1990).  -75Gaskey, S., P. Vacus, R. David, J. C. Andre and J. Villermaux, "Investigation of Concentration Fluctuations in a Continuous Stirred Tank by Space Resolved Fluorescence Spectroscopy", 6th European Conference on Mixing, Pavia, Italy, 129-136 (1988). Guilbault, G . , Practical Fluorescence: Theory. Methods, and Techniques. Marcel Dekker, Inc., New York, 22-323, (1973). Haam, S., R. S. Brodkey and J. B. Fasano, "Local Heat Transfer in a Mixing Vessel using a High-Efficiency Impeller", Ind. Eng. Chem. Res., 32, 575-576 (1993). Heidemann, R. A . , A. A. Jeje and F. Mohtadi, An Introduction to the Properties of Fluids and Solids. The University of Calgary Press, Calgary, AB, 150-160 (1987). Holmes, D. B., R. M . Voncken and J. A. Dekker, "Fluid Flow in Turbine-Stirred, Baffled Tanks", Chemical Engineering Science, Vol. 19, 201-208 (1964). Kerser, A. D., "Distributed and Multiplexed Fibre Optic Sensors", Fiber Optic Sensors: An Introduction for Engineers and Scientists. Eric Udd, Ed., John Wiley & Sons, Inc., New York, 325-368 (1991). Kouppamaki, Risto, "The Quality of Mixing Studied Using a Radiotracer Technique", 1985 Medium Consistency Mixing Seminar Notes, Atlanta, 13-17 (1985). Kouppamaki, R., O. Pikka and K Peltonen, "New High-Intensity MC Mixer - Direct Measurement of Mixing Efficiency", Proceedings the European Pulp and Paper Week, Bologna, Italy, 216-232 (1992). Lee, J. and R. S. Brodkey, "Turbulent Motion and Mixing in a Pipe", AIChE Journal, 10(2), 187-193 (1963). Mann, R., P. Knysh, E. A. Rasekoala and M . Didari, "Mixing in a Closed Stirred Vessel: Use of Networks-of-Zones to Interpret Mixing Curves Acquired by Fibre-Optic Photometry", Fluid Mixing III, I Chem E Symposium Series No. 108, Hemisphere Publishing Corp., 49-62 (1987). Mitro, P., "Diagnostic Retention Study of Bleaching Towers Using Trasar ", CPPA Pacific Coast Branch 1995 Technical Conference Preprints, (1995). R  Nagase, Y. and H . Yasui, "Fluid Motion and Mixing in a Gas-Liquid Contactor with Turbine Agitators", The Chemical Engineering Journal, 27, 37-47 (1983).  -76Olmstead, J.A., and D.G. Gray, "Fluorescence Spectroscopy of Cellulose, Lignin and Mechanical Pulp", MR 321, PAPRICAN, October (1995). Ouellet, D . , C. P. J. Bennington, J. J. Senger, J. F. Borisoff and J. M . Martiskainen, "Measurement of Pulp Residence Time in a High-Consistency Refiner", Journal of Pulp and Paper Science, 22(8), J301-J305 (1996). Paterson, A. H. J. and R. J. Kerekes, "Potentiometric Determination of Chlorine in Microliter Water Samples", J. Assoc. Off. Anal. Chem., 67(1), 132-136 (1984). Paterson, A . H . J. and R. J. Kerekes, "Fundamentals of Mixing in Pulp Suspensions: Measurement of Microscale Mixing of Chlorine", Journal of Pulp and Paper Science, 11(4), J108-J113 (1985). Paterson, A . H . J. and R. J. Kerekes, "Fundamentals of Mixing in Pulp Suspensions: Measurements of Microscale Mixing in Mill Chlorination Mixers", Journal of Pulp and Paper Science, 12(3), J78-J83 (1986). Patterson, G. K . , W. Bockelman and J. Quigley, "Measurement of Mixing Effects on Local Reaction Conversion in Stirred Tanks", Fourth European Conference on Mixing, 303-312 (1982). Pattyson, G. W., "Kamyr MC Mixer for Chlorine Dioxide Mixing at Great Lakes Forest Products", 70th Annual Meeting Preprints, Technical Section, CPPA, Montreal, A63-A68 (1985). Roy, B.P., "Fluorescence in Recycled Pulp", MR 317, PAPRICAN, June (1995). Shenoy, U . V . and H . L . Toor, "Unifying Indicator and Instantaneous Reaction Methods of Measuring Micromixing", AIChE Journal, 36(2), 227-232 (1990). Sinn, S., "State-of-the-Art- Chlorine Dioxide Mixer Installed at Weyerhaeuser", Pulp and Paper, 58(6), 119-121 (1984). Sculmann, S. G., Fluorescence and Phosphorescence Spectroscopy: Physicochemical Principles and Practice. Pergamon Press, Oxford, 213-220 (1977). Spillman, W.B., "Optical Detectors", Fiber Optic Sensors: An Introduction for Engineers and Scientists. E. Udd, ed., John Wiley & Sons, Inc., New York, 69-90 (1991). Torregrossa, L . O., "Effect of Mixing Efficiency on Chlorine Dioxide Bleaching", 1983 Pulping Conference Proceedings, TAPPI Press, Atlanta, 635-641 (1983).  -77Wamer, "Fluorescence and Phosphorescence", Instrumental Analysis, G. D. Christian and J. E. O'Rielly, Ed., Allyn & Bacon, 253-270 (1986). Weber, W. J., "Adsorption Theory, Concepts, and Models", Adsorption Technology: A Step-bv-Step Approach to Process Evaluation and Application. F. L . Slejko, Ed., Marcel Dekker, Inc., New York, 1-35 (1985).  -78-  APPENDIX I: DETAILS OF TECHNIQUES T H A T H A V E A T T E M P T E D T O ASSESS MIXING QUALITY  l.Mixing Assessment in Pulp Suspensions on a Laboratory Scale  1.1 Kuoppamaki (1985): Kuoppamaki used a pulse input of a radioactive tracer to assess mixing quality. He used a Ba-137m tracer because of its short half-life of only 2.6 minutes. The short half-life reduces some of the radiation risks, however all the risk is not removed since radiation is still dangerous to people even if they are only exposed to it for a short time. The quality of mixing is assessed by injecting a short pulse of a radioactive tracer into the inlet of the vessel and the radiation intensity response is measured by radiation detectors at the outlet. The response is converted into a residence-time distribution (RTD) curve by background substraction, possible half-life correction, and normalization. This method is advantageous since the relative concentration of the tracer can be measured through pipe and vessel walls, which means that there are no sample ports required. It is also relatively simple to mount the detector to the pipe or vessel wall. Furthermore, the tracer concentration can be measured with known accuracy without the interference of pressure and temperature change in the process. Radioactive tracers can be detected with high sensitivity.  -79However, there are some major disadvantages to this method; for example, this method is not continuous because he used a pulse input of a radioactive tracer. It is relatively simple, although expensive, to apply this method to a continuous addition of the tracer. There are also extreme dangers and hazards that exist when handling radioactive material, and it would not be possible to bring any radioactive material into a mill. Finally, radioactive material and the detectors are exceptionally expensive (in the order of $500 000).  1.2 Kuoppamaki et al. (1992): Kuoppamaki furthered his work by adding a continuous input of a radioactive tracer into a pilot plant. This was done by labelling the chemical fed into the mixer with a radioactive tracer (137m-Ba) solution, and the radiation intensity around the output pipe was registered at very short intervals (20 ms) using radiation detectors mounted symmetrically around the pipe. The inhomogeneity components were computed from the registered time series under fairly general assumptions concerning the inhomogeneity distribution. The tracer injection time in a measurement was about 2 minutes, and the time interval in the count registration was 20 milliseconds, which he found to give a mixing scale of 4-9 cm (this does not cover the fibre scale). An injection time of 2 minutes does not allow enough time to do a continuous mixing test; however, the high cost of the tracer could be a factor in allowing the test to run for a longer period of time. This has all the same advantages and disadvantages that were mentioned previously.  -801.3 Paterson and Kerekes. 1985: Paterson and Kerekes measured fibrescale mixing in the laboratory by measuring the residual chlorine concentration in samples ( 1-5 ml) extracted from the suspension at 2 mm intervals. A microsampler was constructed, which was designed to withdraw a row of very small samples spaced 2 mm apart because through investigation of chlorine diffusion through suspensions of unbleached pulp showed that the distance of diffusion was approximately 3-5mm (Paterson and Kerekes, 1984). The necessary close spacing was obtained by having 4 rows of 12 or 13 syringes, spaced 8 mm apart, intersect on a straight line. The sampler was designed to permit the volume withdrawn by each syringe to be varied over a range from 1 to 5 microlitres (in microlitre increments) by changing the position of a circlip on the plunger of each syringe. In a typical test, a container filled with a pulp suspension is placed on lab jack. The jack is raised until the tips of the syringes protrude well into the suspension. Microlitre-sized samples are then withdrawn from the pulp suspension by quickly releasing the spring loading of each row of the sampler in sequence. Then, each row of syringes is removed from the sampler, the tips wiped, and the contents of the syringes discharged into a row of glass tubes previously filled with 100 microlitres of acidified iodine reagent.  Finally the solutions are  stirred with the tips of the syringes, and the glass tubes sealed with paraffin film. The solution is allowed to stand 1 minute and then chlorine-demand-free water is added. A 100 microlitre portion of this solution is then injected into a fresh cap fitted over the Orion combination electrode (Model 97-70, Orion Research Corp., Cambridge, MA) probe, with holes aligned over the electrode. The maximum mL potential is recorded, and several samples from the parent  -81solution should be tested if possible. The replicate mV readings are averaged and an equivalent chlorine concentration is determined from a calibration graph (Paterson and Kerekes, 1984). There are two advantages to this method. First, there is no tracer that needs to be added to the process, and secondly, a precise measurement is obtained for fibre scale mixing. However, the disadvantages to this method are numerous. It is timely and tedious to obtain chlorine concentration values. Furthermore, this method can only be applied to the chlorine stage, it is difficult to apply this to an on-line technique.  1.4 Breed (1985): Breed analyzed mixer performance by measuring the distribution of NaCl (which is used as a tracer) concentration in the discharge of an industrial mixer in the pilot laboratory. Each mixing test was carried out with the injection of a salt solution into the circulating pulp stream ahead of or at the mixer. A background sample was collected prior to injection to ensure a uniform starting condition. When the injected salt had reached the sampling area but prior to any recirculation, the chemical injection pump and mixer were stopped. Immediately following shutdown, pulp samples were collected from the vertical section of the discharge pipe at the top of the tower.  Sampling in the vertical section eliminated error resulting from cross-flow  contamination. Samples were taken several inches below the pulp surface using a 22-sample cross-section template. Samples collection, including separation of 25-40 mm of liquid from each 300-400 mL pulp sample was accomplished within 5-8 min after system shut down. After each trial run, the 22 samples and between 5-22 background samples were titrated for chloride ion concentration.  -82This technique of assessing mixing is quite easily done, and can be applied to a mill by using lithium (since it is inert and it is not present in the mill) as the tracer. It can be used in any bleaching stage. She does not use a continuous method, however, and very large samples are collected. Therefore, only macroscale mixing is measured.  1.5 Francis and Kerekes (1990): Francis and Kerekes measured mixing in high consistency pulp suspensions using a tracer dye, which is extracted by a solvent after mixing. The tracer dye solution consisted of Astra Red P Liquid in water. The pulp was fed through the mixer at a uniform rate and the dye solution was injected at several locations before or at the mixer at a constant rate. For each test condition, a single 0.5 o.d. kg sample of mixed pulp was collected at the mixer discharge, which took about 5 seconds. For each mixing trial, 50 individual pulp floes were randomly selected, which was done by taking the entire sample and repeatedly dividing it in half until the desired number of floes remained. The floe size distribution was similar for each mixing trial with the mass of the pulp samples being log normal distributed with a log mean mass of the order of 1 mg. The dye concentration for each pulp sample was determined by, first, placing the pulp sample in a test tube and adding 4 mL of glacial acid. The solution is well mixed, and allowed to stand for 10 minutes at 20 °C, and then centrifuged. The absorbance at 280 nm and 595 nm for the decanted solution is measured by a spectrophotometer. Finally, the pulp sample is dried in an oven and its o.d. mass is determined. The absorbance at 545 nm is due to the dye extracted from the pulp sample, and therefore reflects the mass of dye absorbed. This divided by the pulp mass is a measurement of the dye concentration in the pulp sample. Separate  -83calibration experiments were performed to determine the proportionality constant for this relationship. The absorbance at 280 nm is used to measure the lignin concentration in the pulp sample. Although this method is highly tedious in extracting of the dye and weighing the pulp sample, it can be applied to an on-line technique with the use of fibre optics. This would yield a very fast response. However, this method can never be applied to a mill since it will colour the pulp, and a ultra-violet (UV) dye can not be used because lignin absorbs UV light.  1.6 Bennington and Thangavel (1993): Bennington and Thangavel assessed micromixing and turbulence intensity in pulp fibre suspensions by using the competitive, consecutive chemical reactions between 1-naphthol and diazotized suphanilic acid. My research work is not focused on measuring microscale mixing since microscale mixing is not improved by mechanical action; however, the method could shed some light on fibre-scale and macroscale mixing. Mixing quality was determined from the distribution between mono and bis substituted reaction products once a correlation was made for the adsorption of the product dyes onto the suspended fibres.  However, matters were  complicated by the fact that the chemicals adsorbed on the fibres.  Tests were conducted to  determine the extent of dye adsorption onto pulp fibres. The change in product distribution was determined for a number of fibre suspensions as a function of suspension mass concentration. The use of the azo coupling method for evaluation of pulp suspension mixing is greatly complicated by the fact that the product distribution depends on lignin content of the fibre, which means that a separate correction needs to be made for each pulp tested.  -84-  "This page left blank"  =3  s  T3  i  rt  co  G  ^  C O OS  co  o CaX § cu ^ c 'ts M<u 2g Sc E o  3  U  c CD  CX  CU u3  O  c o  3  co  3  c  T3  o ° CXfe -2 a S  G  bfl  JS  2 «  oOn  — to3 O  o3  o3  o  cu  c  •o c  JS  o3  u  J3 _03  _c 'C  OS  JU 03 O  O  <u  < eu  3  C  g bO  c c 'x E0 S CU  *-t—» Oj  O  "o o  1 (U  G  CO  CU '3 cr  CD 6  CU O  CU  (U J>  CX.£ ex  C C  G O  CU c o ocx 3 t/i c  oj co  co . -G >> c  Q  (U  »-l  ~  CO  cu CO  oO <2 i i  i  o\ ON  >n  r  oo TA ON  RT  1—i  CU  is ic5 E £ o3  o3 C X C X C X C X 3 3 O  o  CU Oj  <U  _> o  CX3  O  1— 3  •»—>  o3 1— o3  <+->  o  •*-> 3  C X G  co  cu  c  G  :b ctU o c— > '3 cu  +-•  cU - G .  •a I  bp  6  O'  C  a  "  I  <4_  cu M  03  •o c  C X C X oJS o3 03  CU 3  w  C  a* cu  c E  JS  <u  ou <  3  CO  o3 CU  P  cu  1« V o  CX <u .S2  O  "+H  CX  o —  00  ON  ON  21  U  -  CO  bO  <U CU  CU  — (  <u  lu  t—  00  c o o3  JS  o  73  -a  <3 2  c o3 C  JS  .9 4—1  • °  -5  C  c ^: o 1  G  ?3  3  ui G O  2  G  § CQ ON bfl  '> O  6  'o  c  _C  3 T3  o3  *  CQ  <U _G bO G O  ON  < ci _uu  OH  g  •0  o3  CU Oj  O  c  CO  CU  co  T3 o3 co 3  OO  3  <U  CX G  <u co 03.. O co E  ' 5> ..  O C  O .2  E CO CO  CU CO CO  o3 CU  -a 03  ^ c— •S § M  D a-  -86-  2. Mixing Assessment in Pulp suspensions on a Mill Scale  2.1 Elliott and Farr (1973) Elliot and Farr assesses mixing in a mill by determining the "free" available chlorine and oxidation reduction potential (ORP) measurements immediately after mixing. Plots of ORP vs time were made by collecting data taken at sufficient intervals to sketch the potentiometric curve. An ORP probe connected to a pH meter with a potentiometric scale was inserted into the samples of chlorinated stock to obtain these measurements. No further mixing or agitation was provided after insertion. Samples of chlorinated residual by iodometric titration. Titration for free chlorine residual was proceeded by carbon tetrachloride extraction.  This method is  advantageous because no substances need to be added to the pulp suspension and it is easily done. The disadvantage is that it is only applicable to the chlorine stage in a bleach plant, it is not a continuous method, and it is timely in terms of the titrations and creation of the ORP curves.  2.2 Backlund. Bergnor. Sandstrom. and Teder (1987) Backlund et al. added LiCl to the C10 (at a constant ratio of L i to C10 ) in the mill, +  2  2  and the L i concentrations were measured. The pulp samples were withdrawn from different +  positions on the surface of the bleaching tower because there were no sampling ports between the mixer outlet and the tower inlet. The pulp samples were about 200 mL and contained about 20 g of dry pulp. Lithium content was determined by flame spectroscopy. The measuring period  -87-  was five hours.  Pulp samples were withdrawn every 30 min from different positions.  Altogether nine pulp samples were withdrawn along both the diameter and the circumference of the tower. Pulp samples of 40 g (wet pulp) were treated with 40 mL of 0.1 mol/L hydrochloric acid for 15 min in order to turn the ionic groups in the pulp into their hydrogen form and to liberate lithium ions adsorbed on the fibres. The solution was separated form the pulp, and the lithium content was determined. The wet samples was dried and the pulp consistency was determined. The 40 g sample that were taken measure only macroscale mixing, and a sampling time of 30 minutes is too long, since there is such a long period in between that is unknown. This is a relatively easy method that can be applied to any bleaching stage. However, samples were still withdrawn; therefore, it is not continuous, although is can be continuous by use of an ion selective electrode. A fast electrode would still measure only the macroscale. 2.3 Torregrossa (1983): Torregrossa assessed mixing quality in a similar manner to Backlund et al.. He mixed the LiCl with the C10 solution. After the LiCl has been added to the C10 supply for enough 2  2  time to ensure that any channelling in the tower does not influence results, small pulp samples are removed after the bleaching tower and analyzed for consistency and lithium concentration. However, sampling should always be done before the bleaching tower since the effects of channelling cannot be removed from the effects of bad mixing. The average C10 charge for 2  these small pulp samples can then be calculated. Several positions in the tower were sampled. This method is problematic since the samples are withdrawn after the bleaching tower, and there is no way to isolate how much of the mixing occurred in the tower. There was nothing stated in the paper to indicate how large the sample were (that were withdrawn), and how the lithium  -88ion concentration was measured; therefore, it is difficult to determine what mixing scales were measured, and what the mixing time was. The same advantages and disadvantages apply to this that were mentioned in the previous section.  2.4 Torregrossa (1983). Sinn (1984). and Pattvson (1985) Mixing was assessed by all three of these people by taking temperature measurements at the surface of the discharge pipe, which can be an indicator of C10 distribution since C10 2  2  is considerably colder than the pulp. Measurements were made of the surface temperature at 6-8 positions around the mixer discharge piping. This can be done for the C10 stage since the 2  C10 comes in considerably colder (10°C) than the pulp (70°C). This is a continuous technique 2  that is easy and efficient.  However, even with the fastest thermocouples (1 Hz) will only  measure macroscale mixing. Another disadvantage is that it is only applicable to the C10 stage. 2  The use of a temperature tracer can be applied to other stages by adding a cold stream of water, but this would change process conditions. Finally, heat transfer coefficients (in the order of 0.1 W/m K) are larger than mass transfer coefficients (in the order of 10. m /s), which means that 10  2  mass distribution is not measured, but rather the distribution of heat (Heidemann et al., 1987).  2.5 Paterson and Kerekes (1986) Paterson and Kerekes applied their laboratory work to the mill in low consistency environments. They measured fibrescale mixing in the mill by using syringes spaced 2 mm apart, and measuring the residual chlorine concentration. A line sampler was devised (after the mixer but before the chlorination tower) to minimize mixing during sampling. The sampler  -89consisted of a 76 m diameter schedule 80 PVC pipe. This may be bolted onto a flanged port on the stock line. The other end of the sampler consists of a 38 mm diameter tee with two valves connected. One of these valves is attached to a high pressure hose (400-1400 kPa). The other valve opens to the atmosphere. First the line is purged with water, and then it is filled with pulp stock by opening and closing the appropriate valves, and then the line sampler is disconnected from the stock line. This is done to facilitate the syringe sampling method, and to find the variation in mixing in relation to the length of the sampler. A small portion of the pulp suspension is removed from the samples and placed under the micro sampler, and samples are taken and analyzed using the procedure described previously in this appendix in section 1.3. The advantages to this work are that no tracer is needed, fibrescale mixing is determined, and the mill trial proved that this method is useful in assessing mixing quality. There are several disadvantages. First, this technique is confined to the chlorination stage of the bleach plant. Furthermore, it is a sampling technique with no application as an online technique.  This leaves very little scope for trouble shooting causes of poor mixing.  Finally, it is hard to measure the extent of mixing or segregation that occurs in the line sampler.  CD  CD  .s O  o  •S <D  co  o  CD X3  C  3 CD T3 3 CD  ^  o  2 3 &  &  .3  CX  cd  CO  I- I-  CO  cd  cd  :>v.2 «g *  2  i OO .ei  C 3  "2  "° s. M  •a e o I  I  I  3  X,  ? 6  cn  4-»  _c >> 1-  co  Cd  •a 3  CD O  cD  ex  e  'fl  fl  —  _(D  CX CD  6  cd  3  1 -s O  CD  V  3  O i-J  cd  cd  OO  •O X3  CD CD  3  in  ID <=.  i  0  4-»  T3 CD 3 3  O  o  cd  3  T3 , 0  (4-1  (D I  o  cn " 3  6  <D  4-*  -B  CXI  6  cd CO  O 3  CD  6  <D 3 cd CD co  00  CD  3  <D  cd  a,  M  O  O  -3  CD cn > JD  O  «* 6  g:  3 flco <P  o  3  V  O cd  CD  i-<  4-»  .s  CD  CO  o > N u 3 0 = -^ O v-. cd 3 I  co  O  3  JZ  CD 00  3  to  X>  i  cd  cd  •2  CD  O  e  T3 •3  CO  1 / 5  3  O  cd  "3  T3  CD -O CD CD  u.  T3 CD <D 3  3  O  O cd  co  JZ  g  3  <D  <D X3  o o  S «SJ .a CO  CD  2 e  J2  3  X5  0  O  B .3  <D X J jd  .5  3  O cd  -2 cd O  3  CO  sum  cd  OO  cd 4 -*  c x:  are  O  he  CD  JD  no  _o •o x: o o  O  T3 T3 cd O 3  CD - 3  > 1+3  O  CX  T3 3  oo  cd 3 CX  ON ON  T3  CD  4-»  CX  B  CD  cd tX,  ON  CD co  cd  O  cn  t>0  82  c  oo ^  o  <D O .3 h «  co cd &.  3 3  _o  CD  o OQ H cd  w  co oo  ON  cd  CD  xs  oo  '-' .  m oo ON  r_I  CO  oo  CD -^ <D uCD  <*$ c  o  CO  t-i CX, ON  4-> Cd  •a o  X3 4-» CD  CD <D CD 60  *» §  •s-s <D  6  cd  6  3 C/3  •t;  cd  H  Ui 4 -4 o  XJ  O 3  cd  cr g CD  X)  C D O cd  "c  O  •5 a JD  CD  00  g o o  3  O cd  -a  <  00  3  00 cr CD  1-4 3  4-> cd u, CD  JZ  O  "cd 3 "3  <D  4->  00  CD cd  3  T 3 3 Q4  C  o  f  is  i_ •° s co  S  O  T3  CD  CD  a  3  IS  3  CD  •2 i c3 o b w §  50) 3Op  & £  cd cn  -913 Mixing Assessment in Other Industries  3.1 Nagase and Yasui (1983): Nagase and Yasui measured mixing by using an ionic tracer. This done on-line with electrodes.  Experiments were carried out by using deionised water and potassium chloride  solution. They place four electrodes at several points in the mixing vessel. These electrodes were made of platinum wires, and their diameter were 0.5 mm. 5 mL of a saturated potassium chloride solution were introduced either at the bottom or just above the impeller at the centre of a 12.3 L vessel. The final concentration of the solute was 1.3 x 10" g/(mol L). The output 3  signal from the electrode was linear with the potassium chloride concentration in the ranges studied. The output signal vs time was then plotted for each electrode to give an indication of how mixing was occurring for each electrode. This method is easily applied anywhere in the mill and it is continuous; however, the electrodes have a slow response (2 Hz), which means that only macroscale mixing is measured.  3.2 Shenov and Toor (1990): Shenoy and used a dye to measure micromixing optically. They present the chemical indicator method in which one follows the colour intensity of an indicator which is present and whose equilibrium is shifted by the acid-base reaction. The colour intensity is measured by a fibre optic light probe with a gap of about 2 mm and a fibre optic bundle diameter of approximately 0.7 mm. The chemical indicators were phenolphthalein because it is a one-colour indicator, and bromothymol blue because its colour transition in almost symmetrical around the  -92neutral point.  The colour density proved to be related to the concentration of the base.  Therefore, the colour intensity was measured using the fibre optic probe, which gave an on-line indication of the quality of mixing. This is a fast optical technique, and it gives a precise measurement; however, it would colour the pulp and could not be used in a mill.  3.3 Mann. Knvsh. Rasekoala. and Didara (1988): Mann et al. measured mixing by adding an inert tracer dye and measuring its concentration with an optical on-line technique by using fibre-optic probe at various locations in the mixing vessel. Tracer experiments were carried out on a 30 cm I.D. flat bottomed vessel of 20 litres capacity with a 10 cm Rushton Turbine 15 cm below the surface. The probe tip was securely positioned at the required point in the vessel and the photometer activated.  After  establishment of a steady mixing state, equal amounts of Nigrosine dye solution were simultaneously injected into the vessel from four syringes equally spaced around the shaft. At the instant of injection the computer sampling routine was activated (measuring the intensity of the tracer dye) and allowed to run to completion. This method has the same advantages and disadvantages that were mentioned in the previous section.  -933.4 Lee and Brodkev (1964): Lee and Brodkey studied turbulent mixing by injecting a dye solution to the centre of a pipe, and the concentration fluctuations was measured by means of a new light probe developed. The measurements of mean concentration and intensity of concentration fluctuations were made both along the axial distance and across the pipe. The dye used was gentian violet, and it was injected at a concentration of 100 mg/L at a rate of 0.822 L/min (the total flowrate is 2.7 L/s). As mentioned previously this method is a fast optical technique and yields a precise measurement; however, it can not be used in mill, since it would colour the pulp.  3.5 Haam and Brodkev (1992): Haam and Brodkey studied local heat transfer in a mixing vessel using heat flux sensors. The mixing vessel is a cylindrical stainless steel container that has an elliptical dished bottom. The vessel has heat flux sensors, mounted on the inside wall exactly over small local cooling regions designed to increase the heat flux. In this way the bulk of the fluid is predominately cooled slowly by natural convection. The heat sensor that used were Rdf Micro-Foil, which consist of a heat flux sensor part to measure heat flow per unit area, and a thermocouple part to measure the wall temperature. The heat flux and temperature can then be measured as a function of time. The use of a temperature tracer can be applied to other stages by adding a cold stream of water, but this would change process conditions.  Finally, heat transfer  coefficients (in the order of 0.1 W/m K) are larger than mass transfer coefficients (in the order of IO m /s), which means that mass distribution is not measured, but rather the distribution 10  of heat.  2  -943.6 Holmes. Voncken. and Dekker (1964) Holmes et al. measured mixing in baffled vessel by measuring the response to a pulse of ionic tracer, injected into the impeller zone, with a conductivity cell around the impeller. The circulation time (the time required for a fluid element to flow once around the average recirculation loop in the tank) was measured in tank of varying diameter by using a pulse injection technique. With the turbine running at a constant speed, a small amount of aqueous conduction solution (30 or 98 wt % H S0 ) was rapidly injected into the turbine centre. A 2  4  conductivity cell, composed of three copper rings concentric with the turbine, detected this tracer as it left the turbine. The change of the conductivity with time was recorded. This method has a fast response, it is continuous, and it is easily done; however, it is difficult to apply to a mill because of the high conductivity of the other ions present.  3.7 Patterson. Bockelman. and Ouiglev (1982) Patterson el al. introduced a method for measuring local concentration of reactants and/or products of chemical reactions in mixing vessels. The method utilizes intensity of fluorescence and its relationship to concentrations of the measured compound. These reactions must involve reactants or produce products which are strongly florescent with the emitted light spectrum well separated from the incident light spectrum. A quartz window was constructed at the side of a baffled vessel where the excitation light entered. The emission light was viewed by a telescopic pinhole at the top of the vessel. The emission light intensity was measured by a photomultiplier. In order to eliminate stray light and light scattered by particles in the measuring volume, a band pass filter is used which passes only a narrow band of the fluorescence spectrum. The reactants  -95are fed to the stirred tank from storage vessels by centrifugal pulps. This method describes the measurement of microscale mixing; however, the application of fluorescence in a pulp suspension has not been done before, and it appears to have tremendous potential as will be discussed in the next section.  3.8 Gaskev. Vacus. David. J.C. Andre, and Villermaux (1988) In 1988, Gaskey et al. investigated concentration fluctuations in a continuous stirred tank by space resolved fluorescence spectroscopy. A standard, baffled reactor, stirred by a Rushton turbine is fed by two liquid streams, one containing a fluorescent dye. A laser beam is focused at a given point in the reactor, exciting the fluorescent molecules. The emitted fluorescent light is monitored at a right angle to the incident beam. The concentration fluctuations due to the imperfect mixing lead to changes in the fluorescent intensity, which is monitored using a photomultiplier tube. The current from the tube is converted to a voltage, amplified, filtered, and digitized. The digital signal is stored for later analysis. The results are given for spatial resolution of 50 micrometers and frequencies of 200 Hz, although the operating limits of the equipment will allow spatial resolution to approximately 10 micrometers and frequency resolution up to 5 kHz. The details of their experiments include injecting the liquid at the height of the agitator and toward the bottom of the tank. Two quartz, optical quality windows, one in the bottom , the other in the wall (90°), allow the excitation and analysis beams to be precisely focused in the measurement plane. The sampling volume can be changed by adjusting the pin-hole from 1 mm to 10 micrometers.  -96This method has several advantages.  First, it has a fast response, and can measure  mixing on all scales. Fluorescent dyes can be non-toxic and hazard free. Fluorescence is very sensitive to a particular compound, and Nalco has used fluorescence in the mill already to determine channelling in a bleaching tower (Mitro, 1995); therefore, we know it is possible use fluorescence in the mill environment. However, the down side is that fluorometers cost about $15 000.  Fluorescence is subject to quenching quite easily; furthermore, lignin is known to  fluoresce, and bleaching chemicals might breakdown the fluorescent dye.  3.9 C. Andre. David. J.C. Andre, and Villermaux (1992) C. Andre et al. presented a novel method for measuring local deviation from perfect mixing of two non-reactive tracers to measure microscale mixing. This was done by adding a fluorescent dye (Fluorescien) and a quencher (I), and the fluorescent intensity was measured. Quenching occurs at the molecular level; therefore, microscale mixing was measured. An argon laser beam is focused on a given point within the mixer. The laser source operates at 488 nm. They used a stirred tank of 1125 cm equipped with a Rushton turbine. The feed streams are 3  injected at the level of the stirrer and towards the tank bottom. Two flat windows, one in the bottom, the other in the wall, allow the excitation and re-emitted beams to be precisely focused. The measurement point is determined by the intersection of the incident beam with the image of a pinhole of 50 micrometers diameter. Fluorescein was chosen as the fluorescent species and I" (from KI) as the quencher. Fluorescence intensity was measured by a photomultiplier, and converted into an electrical potential by an intensity-voltage converter after amplification. This method is also used to assess microscale mixing; however, the advantages and disadvantages  -97mentioned in the previous section apply here as well.  OO C  T3  cd 3 cd  CO  CD O o  CD  cx  CO  s  2 «  CX cd  2 £  £ >?  5  c  3 a 3 cd O  CO  Ct—  CD  fD  b  C "O  CD 4-» co  o _04  •s  c o  O O  CO  •*—»  CD  CX CD  G O %—» cd cd 3 _o "o o  CO CO  C  •a  CD CO  d  CX co  <D i  O CD cD  ° -3  CX X>'3 co CX^ 3 CX CD O ^  CD O  CD .>  o  :3 3  JZ  CX  3 CX  —  OO  cd j = r> ~ »CD O 2" 3 CD <D ~ x; +- o cd -3 b o ^ a CD <4—i T X) O  3  CO  r-  o  S  3  CD  «  CO  53  1  CD ^  3^  co CD  "3  2 S -  .2 «3 _  3  CD Q . s CD C CD 3 cd 3 CD § S o 3 •a > CD 3 cr CD 2 53 cd 1c=1 Q .  "5. | 8 i  CD  i  O  ^  00 £ 3 £ — O co  •s  H3  S? c3 O0.2 S S X3 »-i ~ i X5 •a £ c  6 cd  _3  8  •< &•*  CD l-i CD JZ  CD  CD JZ 4-» O  a  T3 CD 4-»  ta  a  CD  4—> 4—»  cd CO  •o  "5 <*>  oo ON  is jfoft  ta jft A  cd >< CD CO cd OO cd  3 O  2 3. 3 -3 3  cx <u fJ.SS <+-i  >•>.£  r3 +"* co 3 cd O  CX  CD  O  o  -o I-I cd  ON ON  8-S  H  CO  cd"  2 e 3 s  JZ  .5  *  O  M  CO  C  ° 35 W  O. CO CD  Q  3  O  £  52 ccj O cd <+- O CD  «  i  B  CD O cd 4-» o 'c _o  CO  3  cd  6  ON OQ ' - '  3  o  cT 3 cd CO  O vo > CO  CD C >b JL  3  cd CD  CD CO  3  -3 CD O cd  cd 00  .£  CO  CD JZ cd 00 c  ON  O 3  •a  3 •—' 3 co  cD 4-> 4-> cd 0-  cd <+-  C CD O co CD l-  C eo  fd CD b >  OO - r  cD  i—  »—  3  cd  3 CD  . £  §§ 1 i3 ON  % TZ:  CD  b 3  CO  i(D CD ^N  O  C  cd CD  co CD 1-4  3 - Q > O - cd CD X T3 CQ — Xj 3 3 c cd cd o .SP >^ 3 E  o  "3  1)  i_  OO CXJJ  1  -£ CD ^> Q  <u  CD  cW  <D  1?3C S £CD O  cd  ef  CD 00 cD co OO CD  b  CO rt  53  3  -a o  ^  O  b  O  ON  M  ON  rt  CD CD CD co l_ CD 3 CD O 3 co CX cd C*-4 3 co CD CD 1 /5 CD a •3  O JZ o  3  •3 m  3 O  CD CO  o  CX  £  •a _CD  CO  co • — cd ^  CX  •cd  X J (D o B CD  >,.£  3 3  .2 *4—» 1cd  CO  3 3 cr CD c  E?  rG 'co cd CD  _CD cd O  CD 4->  CX 3 & cd O 3  CO  <D co cd 3  cd  <  jd  U cd ^  4--  .2 £ 3  S 8.2  O 3 53 cd «+-  O u  CD CD DCD b co cd  CD' (N ON  -99-  APPENDIX II: FLUORESCENCE  1. INTRODUCTION During the process of absorbing ultraviolet or visible electromagnetic radiation, molecules are elevated to an excited electronic state. Most molecules will dissipate this excess energy as heat by collision with other molecules. Some molecules, however, will emit some of this excess energy as light of a wavelength different from that of the absorbed radiation. This process is called photoluminescence, which includes both fluorescence and phosphorescence. appendix will concentrate on fluorescence.  This  Basically, fluorescence is when a radiationless  process stops at an excited singlet electronic level, and the molecules will be able to return directly from there to the ground state by the radiation of a photon. Phosphorescence is when they shift to a metastable triplet level before emitting radiation.  2 . THE EXCITATION PROCESS A molecule that has absorbed electromagnetic radiation is in an excited state. It must use some mechanism to eliminate this excess energy. The present discussion covers the competitive processes that lead to elimination of this energy or deexcitation. The molecular multiplicity, M M , is defined as MM=2S+l  (1-1)  where S is the spin quantum number of the molecule, and is the sum of the net spin of the electrons in the molecule. For most organic molecules, S=0 because the molecules have an  -100even number of electrons and, thus the lowest energy state (ground state) must be one in which all electrons are spin paired. The multiplicity for these compounds is one and is referred to as a singlet state. The ground singlet state is designated as S , and the first and second excited 0  singlet states are denoted S and S , respectively (Warner, 1986). x  2  Qualitatively, the most effective approach to describing the absorption and emission processes is to use a Jabolonski energy-level diagram, such as that shown in Figure II-1.  "0-0" Transitions  Absorption  (10-^ec)  Fluorescence  (IO" to IO" sec) 9  7  FIGURE II-1: Jabolonski energy-level diagram depicting absorption and emission processes.  -101-  3. THE DEEXCITATION PROCESSES Consider excitation of the molecule to the excited state S . In condensed-phase systems, 2  the molecule can rapidly dissipate excess vibrational energy as heat by collision with solvent molecules through a process termed vibrational relaxation (VR). In addition, during this deexcitation process the molecule can pass from a low vibrational level of S , to an equally 2  energetic vibrational level of the first exited singlet, S . x  This process is called internal  conversion (IC). The energy-degradation processes of VR and IC occur rapidly (approximately 10" seconds) until the molecule reaches the vibrational levels of S . Boltzmann distribution is 12  t  then rapidly established among the vibration levels of Sj and the lowest vibrational level (V=0) is most likely to be occupied. Due to the rapid energy loss, emission from excited states higher than the first is rare. Only a few molecules such as azulene and some of its derivatives are found to violate this general rule. For many molecules, once the molecule reaches the first excited single, internal conversion to the ground state is a relatively slow process.  Thus,  emission of a photon from these molecules becomes more competitive with other decay processes. This process of emitting a photon for deexcitiation of S, to S is termed fluorescence. 0  Generally, fluorescence emission occurs very rapidly after excitation (in approximately 10" to 9  10" seconds). 7  Consequently, it is not possible for the eye to perceive fluorescence emission  after removal of the excitation source. As can be seen from the energy-level diagram in Figure II-1, fluorescence and absorption should have at least one electronic transition of the same energy. Since these transitions occur between the zero vibrational levels of S, and S , they are frequently called the "0-0" transitions. 0  -102In practice, there is a slight shift between the "0-0" bands due to the differences in solvent effects on the ground and excited states (Warner, 1986).  4. QUENCHING OF FLUORESCENCE A deactivation process that will compete with florescence is called quenching. Quenching can occur by a variety of mechanisms. However, the net result in all cases is deactivation of 5, state through a radiationless process involving interaction with some sort of quencher molecule. There are four common types of quenching observed in luminescence processes: temperature, oxygen, concentration, and impurity quenching (Warner, 1986; Guilbault, 1973). The first type of quenching is temperature quenching. fluorescence decreases.  As temperature increases  The degree of temperature dependence varies from compound to  compound. Temperature increases molecular motion and collisions, and hence robs the molecule of energy. The change in fluorescence is normally 1% per 1°C (Giulbault, 1973). The second type of quenching is oxygen quenching. Oxygen present in solutions at a concentration of 10~ M , normally reduces the fluorescence of a typical compound by 20%. 3  High concentration of the fluorescing compound is the third type of quenching. Fluorescence intensity is proportional to the molar absorptivity; the more highly absorbing the substance, the greater its fluorescence, but when the absorption is too large, no light can pass through to cause excitation. Thus, at low concentrations, when the absorbance is less than about 0.05, there is a linear relationship between fluorescence and concentration. At intermediate concentrations the light is not evenly distributed along the path of light. The portion of the solution nearest the light source absorbs so much radiation that less and less is available for the  -103rest of the solution. As a result, considerable excitation occurs at the front of the solution, but less and less occurs throughout the rest of the cell.  The overall effect of concentration  quenching is that as the concentration of the fluorescing molecule becomes too high the fluorescence intensity levels off and then decreases. The fourth and final type of quenching is impurity quenching. When impurities are present at moderate concentrations, interferences can result. This interference can be in the form of the inner-cell effect, collisional quenching, energy transfer, charge transfer, or the heavy-atom effect. (Giulbault, 1973)  5. FLUQROMETERS The fluorometer is used to detect fluorescence and typically consists of six major components as shown in figure II-2. The first component is the radiation source. Because of its relatively broad continuum, extending into the ultraviolet region, the xenon-arc lamp is usually the preferred radiation source for a grating fluorometer. However, for greater excitation energy at selected wavelengths, the mercury-arc lamp is often used. Excitation spectra are usually severely distorted with this lamp because a mercury arc consists of strop mercury lines superimposed on a continuum. Some fluorometers use a hybrid xenon/mercury-arc lamp as the source in order to combine the advantages of both types. The second component of the fluorometer is the lensing system for efficient transfer of the efficient transfer of the exciting and emitting radiation. Quartz lenses are used in the ultraviolet-wavelength range (200-380 nm), and glass lenses can be used in the visible-  -104wavelength range (380-700 nm). Isolation of selected wavelengths from the radiation source is the next necessary part of a fluorometer. This isolation is provided by the third component, the excitation-wavelength selector system. The function of this component is the selection of monochromatic or narrowband radiation for sample excitation. The excitation-wavelength selector system can be a filter (or series of filters) or a grating monochromator. Thefilterprovides more through-put, but the grating monochromator has the advantage or greater versatility and is usually the preferred system for luminescence research because broad-band excitation will increase the possibility of interferences. A sample cell (cuvet) is the fourth component, and provides sample containment and a uniform illumination surface. These cuvets are typically constructed of quartz to allow passage of ultraviolet radiation. These cells are similar to 1-cm ultraviolet absorption cells, except that all sides are polished because fluorometers usually use the 90° geometry rather than the 180° geometry common to absorption spectrophotometer. The fifth component, the emission-wavelength selector system, is similar in description to the excitation-wavelength selector system. However, this system which monitors emitted radiation, is generally placed at an angle of 90° with respect to the excitation axis to minimize interferences from transmitted and scattered exciting light. With regard to filter instruments, the bandwidths of the primary and secondary filters should not overlap. The final component of the fluorometer system is the detector, which is placed at the exit slit of the emission-wavelength selection system. The photomultiplier tube is usually used in most fluorometers because of its high gain and relatively broad spectral sensitivity to low  -105radiation levels. (Warner, 1986)  Figure II-2: Six major components of a basic fluorometer.  -106-  APPENDLX III D E C A Y O F FLUORESCENT DYES WITH T H E BLEACHING CHEMICALS  -107-  [Temperature  •  40°C  •  20°C  A  10°C  4H  i i  3  o  (A) -f-  10  ~r-  15  25  20  Time (seconds) 1  1  •  I  i  •  1  Temperature and Rate Constants  0  &  i  10  A  40°C k =-1620s-i  •  20°C k =-460s-i  •  10°C k =-250s-i  D  -  -  D  D  •  1  o \ A  X!  c o  H  'is  •  A  c co o c o  U  (B) 0.1 •  i 0  i  1  5  1  1 ——1  10  1  15  1  1  20  1—.—.  r 25  Time (seconds) Figure III-1: Reaction between 2-NSA and C10 (A) raw data (B) determination of rate constants. 2  -108-  Figure III-2: Determination of activation energy for 2-NSA and C10 . 2  -109I  '  1  '  1  • <  -1  1  1  •  1  •  1  5-\ -  •  & o •s  Temperature, °C  S3  •  <o  o c o U  2H  A  2 0 3 0  •  4  •  5 5  0  (A)  100  200  300  400  500  600  Time (Seconds) 10-  I o  •  2 0 oC k = -0.0034 s-i  *  3 0 °C k = - 0 . 0 0 8 1 s-i  •  40°Ck  D  43 c  D  = - 0 . 0 1 3 s-  1  = - 0 . 0 4 1 s-  1  D  CD O  •  l  55°Ck D  o'  U  — r  1  T e m p e r a t u r e a n d R a t e Constants  c o  c  T — 1 — 1 — ' — 1 — • — 1 —  n—'—1— —r 1  —I  0  1  1  100  r  1  200  1  1  300  1  1  1  400  1  500  1  I  600  1  I  700  Time (Seconds) Figure III-3: Reaction between 2-NSA and NaOH (A) raw data (B) determination of rate constants.  -110-  Figure III-4: Determination of activation energy for 2-NSA and NaOH.  -111-  Time (seconds) c  Temperature and Rate Constants  .2 o a  •  10°C k =-450s-i  t-i  •  20°C k =-550s-i  A  40°C k =-2481 s-i  D  D  D  O i—i X c o  t-l 4->  <L>  o c o  u  (B) o.i  T" 0  5  10  15  20  Time (seconds) Figure III-5: Reaction between Mephenesin and C10 (A) raw data (B) determination of rate constants. 2  -112-  Figure III-6: Determination of activation energy for Mephenesin and CICV  -113T — ' — r  5.0 H  (A)  AAATTTflll' A •  •  •  4.5  a a  4.0  es o  H  • •  • •  o O  O  3.0 H  2.5  Temperature, °C • 20 A 30 • 40 • 55  2.0 -50  -p—i 0  • • • «  1  1  1  50  100  '  1  —I  150  '  I  200  1  250  ' 300  Time (seconds) Temperature and Rate Constant  .2 *+ -» o ccj «-i  M 10 H "o  •  20°C k = -0.42 s-i  A •  30°C k = -0.49 sp 40«C k =-2.6s-'  •  55°C k =-11 s-  D  1  D 1  D f  <A«A»AMA  o  H  c o •••••  c CD o C! o O  •••••••  (B)  i— 100  I 300  200  400  Time (seconds) re III-7: Reaction between Mephenesin and H 0 (A) raw data (B) determination of rate constants. 2  2  -114-  1  •  1  1  1  '  1  1  1  •  10-  1  -  ln(k ) = ln(A) - E /R ( 1 / T ) \  on  D  a  -  Q 1 -  A = 6.46 x 10 s-  1  13  E - 80500 J/mol  •  ^  1  '  -  a  -  r 2=0.965 0.1 -  1  I  0.0030  1  '  1  I  0.0031  1  1  1  I  r  '  0.0032  I  0.0033  i  r  I  0.0034  1/T(K"1)  Figure III-8: Determination of activation energy for Mephenesin and H 0 . 2  2  1  i  1  0.0035  -11525  ***  T  T  I.20G O .•s  is 15 d <o o c  o • O  10  5H  Temperature, °C| • 50 • 40 A 30 T 20  (A)  ~l—  —r—  I  100  200  400  300  Time (seconds)  c o (X,  i o  c o oj u. +->  c  <D  o d o  Temperature and Rate Constants •  50°C k =-1.7s-i  •  40°C k = -0.90 s-i  A  30°C k = -0.30 s-i  T  20°C k =-0.016 5-'  D  D D  (B)  D  O  i 100  200  — I 300  —I— 400  500  Time (seconds) Figure III-9: Reaction between Mephenesin and NaOH (A) raw data (B) determination of rate constants.  -116-  1/T (K- ) 1  Figure 111-10: Determination of activation energy for Mephenesin and NaOH.  -117-  APPENDIX IV: STUDY OF T H E NOISE IN T H E FLUORESCENT SIGNAL  The problem of noise in an optical detector relates to the resolution accuracy, and dynamic range available in a detected signal. The noise is that portion of the signal that varies in an unpredictable manner (within the frequency band of interest) and is therefore not suitable for transmitting information. The background noise was studied in the background water and from pulp floe interference. The background noise was determined for each probe (that is the Perkin Elmer probe, Probe 1, and Probe 2), and it was found that each probe's resolution was slightly different. The variation was measured in intensity units, which are in arbitrary units on a scale of 1 to 1000. Perkin Elmer probe's intensity varied by ± 2.25 with no fibres. For Probe 1, the intensity varied by ± 0.15, and Probe 2's intensity varied by + 0.45. These intensity values can be converted into concentration by using the following equation Intensity=m(Concentration)  IV-1  where m is the slope and the values for m are shown on Table IV-1 for each dye and all three probes.  -118Table IV-1: The slope shown in equation IV-1 for each probe. Probe  2-NSA  Mephenesin  Perkin Elmer  44.63  4.97  Probe 1  4.95  0.525  Probe 2  14.75  1.63  Therefore, the variance can be calculated in terms of concentration for each probe with 2-NSA and Mephenesin. These are summarized in Table IV-2. In the presence of floes, the background noise increased. Accounting for the noise in pulp suspensions would minimize the effect of variation due adsorption of the dye into the pulp fibres and quenching due to the pulp fibres. The variance due to pulp floes was found in a well mixed 3 ppm solution of fluorescent dye (both 2-NSA and mephenesin were done). The dye and pulp suspension (C =10%) mixture was allowed to mix for 60 minutes, which is indicative of m  very good mixing (see mixing times given in section 4.5.2). It was found for the Perkin Elmer probe that the intensity varied by ± 1 4 for 2-NSA and ± 1 . 5 2 for Mephenesin. For Probe 1, the intensity varied by +1.42 for 2-NSA and ±0.156 for Mephenesin. For Probe 2, the intensity varied by ± 4 . 2 6 for 2-NSA and ± 0 . 5 5 for Mephenesin. By using equation IV-1, the variance in concentration can be determined. This was also done for 3 % consistency suspension. Again these are summarized in Table IV-2.  -119Table IV-2: Summary of variances for each probe. Probe  ± ^background (PP ) m  C =3%  C =10%  c =o%  m  m  m  2-NSA  Mephenesin  2-NSA  Mephenesin  2-NSA  Mephenesin  Perkin Elmer  0.05  0.75  0.31  0.31  0.29  0.30  Probe 1  0.03  0.28  0.29  0.30  0.32  0.32  Probe 2  0.03  0.28  0.29  0.34  0.30  0.31  The variance without fibres was slightly higher for the Perkin Elmer probe than the probes that I constructed. Furthermore, the variance is one order of magnitude lower with 2NSA than Mephenesin. The variance with fibres was about the same for both dyes and all three probes at about ± 0 . 3 ppm. This changes as the overall concentration of the dye changes by the following equation.  ±o=0.1(Overall  concentration)  1 V _ Z  Now that the amount of noise is known due to background noise, the mixing index needs to be corrected for this noise. This is done by utilizing the following equation.  -120-  ,, M =  / 2 _ 2  V°test •  °background  Average Concentration where:  jy_3  M=Mixing Index o =Variance of the mixing test 2  tast  lackground  a  =Variance  of background noise  This equation is applied when finding the mixing index for the on-line tests that were described in section 4.5.2. The values were obtained from the on-line test at a particular point in time by taking 15 values before and 15 values after the time in question.  From these values, the  standard deviation was obtained, which was squared to obtain the variance, and all these points were averaged for the average concentration.  -121-  APPENDIX V RESULTS O F CONTINUOUS MIXING TESTS  12^  •co o  o> CaO u  6  P  (uidd) uop&qusoucQ  |  1  !  •  1  '  I  1  1  - ccc  c o o  -  1o  to  O  II S  I  (uidd) UOI BJJU30U03 4  ,  (uidd)  1 _,  1  •  UOIJBJJUSOUOO  12.3  -a c o o  -a c o o CO  CO  e  to  H  H  •o c  n _1_  I I I  o  (tudd) uonBJiuaouoo  (iudd) uonBJjuaouoo  ^- o  o  «o  (l2dd)U01JBJJU90UOO ^  o «n ^ o  (uidd) UOIJBJJUSOUOQ  \1£  (uidd)  UOIJBJIU30U03  (uidd) uonBJjuaouoo  (Uidd) U0..BJ}U30UO3  (UKW) UO.JBJJU30U03  (uidd) uoueijuaouoo  (uidd) uoqBjjuaouoQ  130  (uidd) UOUBX)U30UO3  (uidd) uoijBJjuaoucQ  

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.831.1-0058521/manifest

Comment

Related Items