UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Chlorination effluent recycle in kraft pulp bleaching Farr, Gary Derek 1991

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

Item Metadata

Download

Media
831-UBC_1991_A7 F37.pdf [ 10.32MB ]
Metadata
JSON: 831-1.0058684.json
JSON-LD: 831-1.0058684-ld.json
RDF/XML (Pretty): 831-1.0058684-rdf.xml
RDF/JSON: 831-1.0058684-rdf.json
Turtle: 831-1.0058684-turtle.txt
N-Triples: 831-1.0058684-rdf-ntriples.txt
Original Record: 831-1.0058684-source.json
Full Text
831-1.0058684-fulltext.txt
Citation
831-1.0058684.ris

Full Text

CHLORINATION EFFLUENT RECYCLE IN KRAFT PULP BLEACHING  by GARY DEREK FARR Dipl.T., British Columbia Institute of Technology, 1982 B.Eng., Lakehead University, 1985  A THESIS SUBMITTED IN PARTIAL 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 August 1991 © Gary Derek Farr, 1991  In  presenting  degree freely  this  at the  thesis  in  partial  University  of  British Columbia, I agree  available for reference  copying  of  department  this or  publication of  thesis for by  his  fulfilment  and study. I further  of  or  her  representatives.  cng-Mc^u,  DE-6  (2/88)  %LSUfi£a-i*~Jbo*«~  requirements that the  agree  It  this thesis for financial gain shall not  fetKa^Se.ft.vaQt  The University of British Columbia Vancouver, Canada  Date  the  scholarly purposes may be  permission.  Department  of  ~* ,  i^^l  is  for  an  Library shall make  that permission for granted  by the  understood  be allowed  advanced  that  it  extensive  head  of  copying  my or  without my written  ABSTRACT When performed at an industrial scale, kraft pulp bleaching commonly utilizes recycled chlorination effluent for brown stock dilution.  Much of the information  regarding this practice pertains to pulp quality. Some of the previous research done in this field failed to separate its effect from the effect of countercurrent effluent recycle for pulp washing.  In addition, there are some shortcomings of the batch techniques  that have been used to simulate effluent recycle in the laboratory. Thus little is known about the effect of this practice on effluent quality. Accordingly, the objective of this research project was to determine the effect of chlorination effluent recycle for brown stock dilution on the quality of kraft pulp bleaching effluent.  Chlorination effluent recycle was investigated in conjunction with  the variables of multiple, mode, and chlorine dioxide substitution.  Effluent quality  was characterized in terms of toxicity, total chlorinated organic compounds, and chlorate. Pulp quality was also measured. A full 2 factorial design was carried out 4  with 24 bleaching runs. The experimentation was performed in the laboratory using a two-stage (CE) bleaching  sequence.  A  specially  designed  continuous  laboratory-scale pulp  chlorination apparatus, which was operated with an unbleached pulp throughput of 6 odg/min, was used to execute the first stage. By producing sufficient quantities of  ii  effluent at steady state in a short period of time, this apparatus enabled the efficacious use of trout bioassays.  The second stage was performed on a batch basis.  Results  from the laboratory experiments were compared to those from an industrial-scale bleach plant. The chlorination stage toxic emission factor for rainbow trout decreased by 314 m /adt as the level of recycle was increased from 4.5 to 22.5 m /adt. The effects 3  3  of multiple, mode, and substitution were also significant. Effluents from the extraction stage were, much less toxic than those from the chlorination stage. The first stage toxic emission factors for trout and luminescent marine bacteria were not correlated. In addition, the concentration of organochlorine compounds and the number of toxic units were not correlated. A novel method for determining the toxic emission factor from median lethal time data was also developed. Recycle had no significant effect on the total production of chlorinated organic compounds^ which were quantified together in terms of AOX. A regression equation, which includes the effects of multiple and mode as well as the interaction between multiple and substitution, was formulated for the AOX results. Recycle had no significant effect on the production of chlorate. A regression equation, which includes the variables of substitution and mode, was formulated for the chlorate results.  A hypothesis, which involves the reactions of chlorine and  chlorine dioxide with the phenolic hydroxyl groups of lignin, was developed to explain several characteristics of this equation.  iii  Recycle also had no effect on pulp quality, which was measured after the extraction stage with the kappa number and the viscosity. Mode and multiple were the only factors that significantly affected the kappa number.  A hypothesis, which  involves the reactions of chlorine and chlorine dioxide with dissolved organic material in the chlorination stage, was developed to explain the effect of mode on dehgnification. None of the variables had a significant effect on the viscosity.  iv  T A B L E OF CONTENTS  ABSTRACT  ii  TABLE OF CONTENTS  v  LIST OF TABLES  viii  LIST OF FIGURES  x  LIST OF DRAWINGS  xi  ACKNOWLEDGEMENTS  xiii  1. INTRODUCTION  1  2. LITERATURE REVIEW  6  2.1.  Toxicity  6  2.2.  AOX  20  2.3.  Chlorate  31  2.4.  Kappa Number  38  2.5.  Viscosity  43  3. EXPERIMENTAL 3.1.  47  Apparatus 3.1.1.  47 Industrial Comparison  57  3.2.  Experimental Design  61  3.3.  Materials and Methods  65  v  3.3.1.  Brown Stock Pulp Supply  65  3.3.2.  Peristaltic Pump Calibration  67  3.3.3.  Chlorine and Chlorine Dioxide Analysis  68  3.3.4.  Sampling at Steady State  68  3.3.5.  Residual Chemical Determination and Reduction  69  3.3.6.  Black Liquor Carryover  72  3.3.7.  Pulp Extractions  74  3.3.8.  Toxicity  74  3.3.9.  AOX  77  3.3.10. Chlorate  77  3.3.11. Kappa Number  78  3.3.12. Viscosity  79  4. RESULTS AND DISCUSSION  80  4.1.  Toxicity  80  4.1.1.  Trout Toxicity  82  4.1.2.  Bacterial Toxicity  89  4.2.  AOX  93  4.3.  Chlorate  101  4.4.  Kappa Number  110  4.5.  Viscosity  116  5. SUMMARY AND CONCLUSIONS  120  6. RECOMMENDATIONS  124  vi  REFERENCES  .  126  A. SPECIFICATIONS OF THE PULP CHLORINATION APPARATUS  135  A.l.  Stock Tank System  135  A.2.  Brown Stock Decker  136  A.3.  Chlorination Reactor  138  A.4.  Dilution Vat  140  A.5.  Chlorination Washer  140  A.6.  Vacuum System  141  A.7.  Heat Exchanger  142  A.8.  Peristaltic Pumps  142  A.9.  pH Measurement  143  A.10.  Temperature Measurement  143  A.11.  Control Panel  143  A.12.  Support Structure  144  B. DETERMINATION OF STEADY STATE  175  C. CHLORINATION STAGE OPERATING DATA  182  D. DETAILED EXTRACTION PROCEDURE  188  E. TOXICITY DATA  189  F. TOXIC EMISSION FACTOR DETERMINATION FROM LT50 DATA  192  vii  LIST OF TABLES  1.  Effluent flows from bleached kraft pulp mills  3  2.  Toxic emissions from industrial bleach plants  7  3.  Studies relating softwood bleaching effluent toxicity to specific chemical constituents  10  Studies of the effect of chlorine dioxide substitution on softwood bleaching effluent toxicity  16  4.  5.  The effect offiltraterecycle on the formation of organically bound chlorine  29  6.  Studies of chlorate formation in the first bleaching stage  35  7.  Equivalent levels of recycle  59  8.  Summary of variation in factor levels  62  9.  Coded settings of factor variables  64  10.  Toxicity production in the first two bleaching stages  81  11.  Correlation between the concentration of AOX and the number of toxic units  82  12.  Factorial analysis of chlorination effluent toxicity to trout  83  13.  Factorial analysis of chlorination effluent toxicity to bacteria  91  14.  AOX formation in the first two bleaching stages  94  15.  Chlorate formation in the first bleaching stage  102  16.  Kappa numbers of brown stock and bleached pulps  Ill  viii  17.  Factorial analysis of delignification  113  18.  Viscosities of brown stock and bleached pulps  117  19.  Qualitative summary of experimental results  121  20.  Peristaltic pump specifications  142  21.  Chemical concentrations and residuals  183  22.  Pulp consistencies  184  23.  Temperature and pH data  185  24.  Peristaltic pump flows  186  25.  Variation in factor levels  187  26.  Trout toxicities (LC50 procedure)  189  27.  Trout toxicities (LT50 procedure)  190  28.  Bacterial toxicities  191  29.  Slopes of the log C versus log LT50 plots  194  ix  LIST O F FIGURES  1.  Flowchart of chlorination as performed with the continuous laboratoryscale apparatus  48  2.  Continuous laboratory-scale pulp chlorination apparatus  49  3.  Brown stock decker and reactor inlet  50  4.  Decker headbox  50  5.  Chemical addition  51  6.  Reactor outlet  51  7.  Dilution vat  52  8.  Chlorination washer  52  9.  Washer seal tank  53  10.  Constant-head tank  53  11.  Stock tank system  54  12.  Control panel  54  13.  The effect of chlorine dioxide substitution on AOX production  97  14.  The effect of chlorine dioxide substitution on chlorate production  104  15.  Block diagram of the chlorination process model  176  16.  Response of the process model to a step change in chlorine addition . . . .  179  17.  Response of the process to a step change in chlorine addition  181  x  LIST OF DRAWINGS  CR-1.  Inlet section  145  CR-2.  Downflow section no.l  146  CR-3.  Downflow section no.2  147  CR-4.  Downflow section no.3  148  CR-5.  Downflow section no.4  149  CR-6.  2" flange detail  150  CR-7.  Horizontal section  151  CR-8.  Upflow section no.l  152  CR-9.  Upflow section no.2  153  CR-10.  Outlet section  154  CR-11.  4" flange detail  155  CR-12.  Discharge launder  156  CR-13.  Outlet cover  157  CR-14.  Mixer no.2 mount  158  CR-15.  Mixer no.l bushing  159  CR-16.  Mixer no.3 bushing  160  CR-17.  Reactor assembly  161  Dilution vat  162  DV-1.  xi  HT-1.  Constant-head tank  l  b  3  PW-1.  Pulp washer  1  6  4  PW-2.  Decker headbox  1  6  5  PW-3.  Washer headbox  1  6  6  PW-4.  Washer table  1  6  7  PW-5.  Motor mount  1  6  8  PW-6.  Motor housing  169  PW-7.  Drive-discharge roll  170  PW-8.  Press roll  PW-9.  Doctor mount  I  ST-1.  Decker seal tank  173  ST-2.  Washer seal tank  I  ••  xii  1  7  1  7 2  7 4  ACKNOWLEDGEMENTS  The supervisory efforts of Dr. L.R. Galloway, Dr. R.J. Kerekes, and Dr. K.L. Pinder during the course of this work are acknowledged with great appreciation. Financial support for this project from the following institutions is gratefully acknowledged: •  Natural Sciences and Engineering Research Council of Canada  • • •  Pulp and Paper Research Institute of Canada MacMillan Bloedel Limited Microbics Corporation  •  Province of British Columbia  The following individuals are sincerely thanked for their contributions to this research: •  Ron Zaitsoff; Celgar Pulp Company  • • •  Graham van Aggelen; British Columbia Ministry of Environment Ron Watts; Environment Canada Yuan-Shing Perng and Mark Martinez; Department of Chemical Engineering at UBC Jim Wearing, Dr. M.D. Ouchi, Doug Weddell, and Tibor Kovacs; Pulp and Paper Research Institute of Canada Ping Ma; Statistical Consulting and Research Laboratory at UBC Allan Spence; Department of Mechanical Engineering at UBC Scott Charban; Barrday Incorporated  • • • •  G.D.F.  xiii  Most of the cheap and simple inventions have, to put it bluntly and unpersuasively, been made.  John Kenneth Galbraith, American Capitalism: The Concept of Countervailing Power  CHAPTER 1 INTRODUCTION  Pulp bleaching is the process by which wood pulp fibers are brightened. Pulps produced by chemical means such as the kraft process are bleached by the removal of the lignin remaining in the cellulose fibers after the completion of the pulping process. A typical unbleached kraft pulp contains 5% lignin and has a brightness of 25%. Once the lignin is removed by bleaching, the brightness increases to 90%. At an industrial scale, kraft pulp bleaching is performed on a continuous basis typically using a five-stage sequence designated as (DC)EDED.  The first stage,  designated as (DC) in the bleaching sequence, involves the addition of both chlorine (C) and chlorine dioxide (D) to the unbleached pulp.  These chemicals alter the  structure of the polymeric lignin. After a given reaction time, the resulting pulp is washed to remove the dissolved fragments of lignin. This first stage is commonly referred to as the chlorination or C stage. The washed pulp then enters the second stage, designated as E in the sequence, where caustic soda is added. This chemical aids in the solubilization of additional chlorinated lignin. The pulp is washed again after a suitable reaction time. This stage is commonly called the extraction or E stage. The remaining three stages are similar to the first two in that chemical is added, bleaching reactions occur during a retention period, and the dissolved lignin is washed  1  2 from the pulp.  More detailed information covering virtually all aspects of pulp  bleaching is given by both Rydholm (1965) and Singh (1979). The filtrates resulting from washing the pulp after each stage eventually form bleach plant effluent which is discharged to the environment as part of the whole mill effluent usually after some form of treatment.  Table 1 gives the whole mill and  bleach plant effluent flows for two industrial softwood kraft pulp mills.  Despite a  wide variation in water usage, bleach plant effluent makes up the majority of the whole mill effluent flow.  Bleach plant effluent accounts for a large portion of the  BOD, color, and toxicity as well as all of the chlorinated organics and chlorate discharged in whole mill effluents from bleached kraft pulp mills. The majority of this pollution emanates from the first two stages of bleaching. While many mills treat their effluent, industry looks favorably upon process modifications which reduce the formation of pollutants. In the early 1970s the pulp and paper industry made a substantial effort to reduce the consumption of fresh water primarily in an attempt to conserve energy. Significant reductions in fresh water usage in bleach plants were accomplished by recycling bleaching filtrates for pulp washing. Two common methods developed for this purpose, in which filtrates are recycled countercurrently to the flow of pulp, are the direct and jump-stage techniques.  Direct countercurrent washing involves recycle  between every consecutive bleaching stage whereas jump-stage countercurrent washing involves recycle between every second bleaching stage. Thus, with respect to washing in the chlorination stage of a CED sequence, the direct method utilizes E stage filtrate  3 Table 1 Effluent flows from bleached kraft pulp mills Reference  Effluent flow (m /adt) 3  Whole mill Holmbom and Lehtinen (1980)" Tomlinson (1980)  C stage  a  E stage  60  18  14  162  58  53  'Water consumption and pollutant production are commonly rated on the basis of either bleached or unbleached pulp production. The figures of Tomlinson (1980) were specified on a bleached pulp basis while those of Holmbom and Lehtinen (1980) were unspecified. Because most of the data in the literature review pertain to partial bleaching sequences, they are mainly rated on an unbleached pulp basis. The experimental data presented here are also rated on unbleached pulp. Pulp moisture not specified.  b  whereas the jump-stage method utilizes D stage filtrate.  Another large reduction in  water usage occurred when chlorination filtrate was recycled for the dilution of unbleached pulp entering the bleach plant (see section 3.2).  Because chlorination is  the only stage operated at low consistency and most of the bleach plant pollution originates from the first two stages, the chlorination effluent represents a considerable amount of the effluent volume and the pollutant load from a bleaching sequence. Thus recycling a part of this effluent has the potential of significantly affecting the formation of bleach plant pollution.  Effluent recycle for brown stock dilution and  countercurrent pulp washing was implemented despite there being little knowledge of its effect on the quality of both pulp and effluent. Much of what is known about pulp bleaching has come from bleaching experiments carried out in the laboratory. Nearly all laboratory bleaching is performed  4 on a batch basis with equipment that has a wide range of sophistication. The simplest technique involves performing the various stages in plastic bags. This is in contrast to the batch apparatus of Reeve and Earl (1986) which features high-shear mixing, an adjustable reaction chamber, gaseous or liquid chemical addition, and computerized monitoring of temperature, pressure, color, and residual. The only laboratory-scale bleaching apparatus in which pulp flow is continuous is that of Jamieson (1970). This five-stage bleach plant utilized continuous stirred tank reaction vessels having mean residence times of 10 minutes at 4% consistency.  Althoughfiltrateswere recycled  countercurrently, effluent properties were not investigated as the motivation for the study was the development of a rapid bleaching process.  The conditions used were  vastly different from those in industry where plug flow reactors are used at 10% consistency (3% in the C stage) for a total residence time of approximately 8 hours. The rapid process was never utilized industrially making the lab-scale equipment obsolete. A considerable effort is required to produce effluents at a steady state using batch recycling because of the number of batch bleaching runs needed. The length of time taken to produce such effluents casts some doubt on their validity. This has not stopped researchers from building lab-scale equipment which enables a series of batch bleaching trials to be performed in a relatively efficient manner.  Sepall (1970)  constructed the first such apparatus. However, the equipment did not have a suitable means of mixing the bleaching chemical with the pulp.  As a result, there are no  published bleaching studies which utilized this apparatus. Histed and Nicolle (1973)  5 and Wartiovaara (1980b) also built lab-scale bleach plants designed for repetitive batch runs with filtrate recycle.  However, most of the work with these systems failed to  distinguish between the effects of chlorination effluent recycle for brown stock dilution and countercurrent effluent recycle for pulp washing. Another problem with these labscale systems was their inability to produce enough effluent to allow the use of trout bioassays. In light of previous work, there remains a need to study the effect of chlorination  filtrate  recycle  on the  characteristics  of  bleach  plant  effluent.  Accordingly, the primary objective of this research effort was to determine the effect of the recycle of chlorination effluent for brown stock dilution on the production of toxicity, total chlorinated organics, and chlorate from the first two stages of softwood kraft pulp bleaching. A secondary goal was to determine the effect of the recycle of chlorination effluent in conjunction with other parameters already known to affect the bleaching process in order to ascertain the presence of interaction.  While the  emphasis was placed on several environmental aspects of pulp bleaching, pulp quality was also monitored.  CHAPTER 2 LITERATURE REVIEW  2.1. Toxicity The toxicity of effluent from bleached kraft pulp mills stems from two major sources. The first is from processes associated with pulping and the second is from bleaching. With the advent of condensate stripping, closed brown stock washing, and reductions in hydraulic debarking the proportion of toxicity emanating from pulping related sources is decreasing. Contrarily, with the possible exception of cases where substantial chlorine dioxide substitution (see section 3.2), improved brown stock washing, oxygen delignification, or extended cooking are practiced, bleach plant effluent toxicity expressed as the toxic emission factor (see appendix F) has remained relatively constant.  Thus, as controls are implemented on pulping operations, bleach  plant effluent is making up a larger portion of a decreasing total toxic discharge. Table 2 shows the toxic emission factors of chlorination and extraction effluents in relation to the total toxic discharge from four bleached kraft pulp mills utilizing softwood furnishes. The toxicity of bleach plant effluents is caused by low molecular weight molecules.  Salkinoja-Salonen et al. (1981) found upon ultrafiltration of an industrial  birch extraction effluent that the majority of the toxicity to Daphnia magna resided in  6  7 Table 2 Toxic emissions from industrial bleach plants Reference  Toxic emission factor C stage (m /adt)  E stage (m /adt)  Proportion of total toxic discharge (%)  Holmbom and Lehtinen (1980)°  286  246  84  Tomlinson (1980)  193  106  69  Scroggins (1984) mill no. I*  280  Scroggins (1984) mill no.2  200  8  3  3  —  b  22  8  260  29  e  "Determined for rainbow trout. Sum of all individual toxic emission factors.  b  Tulp moisture not specified. Combined effluent.  d  Toxic emission factors from two mills in a dual mill complex having largely different production rates were incorrectly summed.  the fraction with molecular weights less than 1,000 u. Sameshima, Simson, and Dence (1979) found similar results when a sdftwood extraction effluent, produced in the laboratory, was separated by ultrafiltration. In addition, ether extractions of the same extraction stage effluent and of the corresponding chlorination effluent removed 86 and 109% of the respective toxicity to Daphnia magna. It has also been demonstrated that resin adsorption can retain nearly all of the toxicity to rainbow trout in chlorination and extraction filtrates (Leach and Thakore 1975, 1976).  8 With knowledge that organochlorine compounds are toxic to aquatic organisms, resistant to biological degradation, and accumulate in the fatty tissues of many organisms, the search for compounds responsible for the toxicity of bleaching effluent was narrowed to emphasize low molecular weight substances containing chlorine. Both Voss et al. (1980) and Salkinoja-Salonen et al. (1981) showed that the toxicity of several phenolic compounds increased with the degree of chlorine substitution on the phenolic nucleus. Of the 268 low molecular weight compounds identified in pulp mill effluents, nearly 80% contain chlorine (Suntio, Shiu, and Mackay 1988). Only a few of these compounds have been implicated as causing the toxicity of bleaching effluent. The toxicity of softwood chlorination filtrate is ascribed to chlorocatechols while chloroguaiacols are thought to cause the toxicity of softwood extraction filtrate. A group of extractives known as resin acids are generally held responsible for the toxicity of effluents from pulping operations. These low molecular weight compounds and their chlorinated analogues can cause some of the toxicity in bleaching effluent due to the carryover of black liquor into the bleachery. This effect tends to be most pronounced in extraction effluent as resin acids are largely insoluble at the low pH of chlorination. As a result, they are carried with the pulp fibers into the extraction stage where solubilization occurs at high pH. In addition, fatty acid extractives and their chlorinated analogues can cause some of the toxicity of bleaching effluent. Since adsorbable organic halogen (AOX, see section 2.2) measurements include substantial amounts of high molecular weight material they are not a good indicator of toxicity. Chlorine contained in chlorophenolic compounds typically accounts for only  9 2% of the A O X in bleaching effluent.  Liebergott et al. (1990a) failed to find a  relationship between the toxicity to fathead minnows and Ceriodaphnia affinis and the concentration of A O X in combined C(EO) effluents produced in the laboratory. However, the low molecular weight fraction of AOX can be quantified separately by extracting the effluent with a suitable organic solvent and then analyzing the resultant extract for AOX. Earl and Reeve (1989a) found a linear relationship between heptane extractable AOX in g/t, denoted as EOX, and AOX in kg/t given by EOX = 33AOX -100  (1)  Accordingly, it is unlikely that EOX measurements are causally related to toxicity. Evidence to support this was given by Axegard (1985) in measuring both ether extractable total organic clorine (TOC1, see section 2.2) and total chlorophenolics in effluents from bleaching a softwood kraft pulp at various levels of chlorine dioxide substitution. While the ether extractable TOC1 decreased linearly with substitution, the chlorophenolics showed a maximum at 50% substitution. Attempts to account for the toxicity of softwood bleaching effluents have involved the simultaneous measurement of toxicity and concentrations of chlorinated phenolics and resin and fatty acids. Table 3 outlines eight such investigations. Some of the studies were simply searches for correlations between the levels of toxicants and the resulting toxicity.  This approach is somewhat weak, as significant correlations  have been found even though the concentrations of constituents were well below levels required to cause acute toxicity. A better technique involves calculating the number of toxic units for each suspected toxicant (the ratio of a toxicant's concentration to its  10 Table 3 Studies relating softwood bleaching effluent toxicity to specific chemical constituents Reference  Effluent(s) tested  Bioassay organism(s)  Constituents measured  8  Leach and Thakore (1975)  E  rainbow trout  CG, RA, FA.  Holmbom and Lehtinen (1980)  C and E  rainbow trout  CP, CG, CC, RA, FA.  Tomlinson (1980)  C and E  rainbow trout  CP, CG, CC, RA.  Salkinoja-Salonen et al. (1981)"  CE combined  Daphnia magna  CP, CG, CC, CS.  Voss, Wearing, and Wong (1981)  C and E  rainbow trout  CP, CG, CC, CV, RA, FA.  Gergov et al. (1988)  C and E  Daphnia magna  CP, CG, CC, CV.  Liebergott et al. (1990b)  C(EO) combined  rainbow trout  CP, CG, CC, CV.  Liebergott et al. (1990a)  C(EO) combined  fathead minnows CP, CG, CC, CV, and Ceriodaphnia RA, FA. afftnis  "CP-chlorophenols, CG-choroguaiacols, CC-chlorocatechols, CV-chlorovanillins, CS-chlorosyringols (specific to hardwoods), RA-resin acids, FA-fatty acids. Hardwood results included.  b  11 LC50). The sum of the individual toxic units is then compared with the number of toxic units measured for the effluent by bioassay (100% divided by the effluent's LC50). Of the studies listed in table 3 only the work of Leach and Thakore (1975) on extraction effluent was able to account for the majority of the measured toxicity. This is indicative of the relative lack of progress in determining the toxic constituents of chlorination filtrate as compared to extraction filtrate. The other investigations had only limited success.  Included is the study by Holmbom and Lehtinen (1980) in  which the same toxicants were measured in extraction effluent as were measured by Leach and Thakore (1975). Surprisingly only 42% of the toxicity was accounted for. It is clear that much more work is required before a full understanding of the toxicity of bleach plant effluent is acquired. Of the low molecular weight compounds identified in pulp mill effluents only 21% are of known toxicity (Suntio, Shiu, and Mackay 1988).  The chlorovanillins and chloroacetones are examples of chemicals  which occur in significant quantities in bleach plant effluent but are of unknown toxicity.  In addition some toxicants may still remain unidentified.  Even the  sophisticated analytical methods used to identify and quantify toxicants can mask information.  For instance, the derivatization procedure used by Leach and Thakore  (1975) caused two isomers of trichloroguaiacol and an isomer of trichlorocatechol to be analyzed as a single trichloroguaiacol (NCASI 1981).  The use of constituent  toxicity values that have been overestimated due to improper bioassay techniques (see appendix F) may also have hampered progress.  12 The lack of progress in accounting for the toxicity of chlorination effluent may be due in part to a group of low molecular weight chlorinated organic compounds called chloroquinones. The first chemical investigation of bleached kraft chlorination effluent identified through several analytical techniques tetrachloro-o-benzoquinone as an important toxic constituent (Das et al. 1969). Analysis of the data in this study shows tetrachloro-o-benzoquinone approximately 0.4 mg/1.  to  have  an LC50  to  Atlantic salmon of  Leach and Thakore (1976) confirmed the presence of  chloroquinones in chlorination effluent.  Paradoxically none of the studies in table 3  included the chloroquinones to account for the toxicity of bleaching effluent.  This is  because the modern analytical methods used for the analysis of effluent constituents do not determine chloroquinones unless they are reduced to chlorocatechols. before  and after effluent  Analysis  reduction enables both the chloroquinones and the  chlorocatechols to be quantified. Nonni and Dence (1981) found the yields of four different chlorocatechols increased on average 3 times after a chlorination effluent was reduced with sodium borohydride.  Similarly Axegard (1987) found the yields of  chlorophenolic compounds increased on average 1.6 times after the effluents from four bleaching sequences were reduced with ascorbic acid. responsible for the increases.  Chlorocatechols were  Chloroquinones are not found in extraction filtrate due  to degradation to nontoxic entities by the prevailing caustic conditions (Braddon and Dence 1968). This may partly explain the success of Leach and Thakore (1975) in accounting for the toxicity of extraction effluent.  13 As the group of chemicals generally held responsible for a significant portion of the environmental impact of bleaching effluent, including acute toxicity, the chlorinated phenolics have been studied for their response to changes in the bleaching process. The substitution of chlorine dioxide for chlorine in the first stage is one of the more important process modifications.  Voss et al. (1980) investigated the total  amount of chlorophenolic compounds produced in the first two stages of softwood kraft pulp bleaching. The formation of chlorophenolics showed a maximum at 50% substitution with the amount at 100% substitution being considerably less than that at 0% substitution. This pattern was substantiated by Axegard (1985) and by Liebergott et al. (1990b). In addition Nikki and Korhonen (1983) found a substantial reduction in the production of chlorophenolics when the level of substitution went from 20 to 100%. Contrary to this body of work is the study by Hakulinen and Salkinoja-Salonen (1982) which showed that chlorophenolic compound formation decreased continuously with increasing chlorine dioxide substitution.  Corroborating this is the work of Earl  and Reeve (1989b) in which fewer chlorinated phenolics were found at 50% substitution than at 0% substitution. However, the lack of a maximum was probably caused by the exclusion of chlorovanillins in the determination of chlorophenolics in these two studies.  This is supported by the work of Liebergott et al. (1990a) which  showed that a maximum occurred only when chlorovanillins were included in the analysis of total chlorophenolics along with chlorophenols, chloroguaiacols, and chlorocatechols.  Additionally, since Earl and Reeve (1989b) used a lower multiple  (see section 3.2) at 50% substitution than at 0% substitution to reach a constant CE  14 kappa number their results are not directly comparable to those of the other investigations in which constant chemical charges were used. The effect of mode (see section 3.2) of addition of chlorine and chlorine dioxide on the formation of chlorinated phenolic compounds has been studied by Liebergott et al. (1990a). It was found that simultaneous addition gave similar results to those when the addition of chlorine was 1 minute before chlorine dioxide. However, when chlorine dioxide was added 1 minute prior to chlorine the production of chlorophenolics increased continuously with substitution up to 70%, the highest level studied.  The interaction between substitution and mode meant that at 70%  substitution DC addition gave nearly twice the amount of chlorinated phenolic compounds as CD addition. The results from two other studies are at odds with this result.  Nikki and Korhonen (1983) utilized DC addition with an interval of 30  seconds and found a substantial reduction in the production of chlorophenolics when the substitution went from 20 to 100%. Axegard (1985) also used DC addition with an interval of 2 minutes at 30 and 50% substitution and an interval of 4 minutes at 70 and 90% substitution.  The total chlorophenolics showed a maximum at 50%  substitution. The reason for these discrepancies is not known. Voss, Wearing, and Wong (1981) investigated the effect of several process variables on the production of chlorophenolic compounds from the first two bleaching stages.  The chlorophenolics showed a maximum at a first stage multiple of  approximately 0.2 %/ml. In addition, both decreasing the pH and increasing the temperature of chlorination were shown to increase the formation of chlorophenolics.  15 Nikki  and  Korhonen  (1983)  simulated  both  jump-stage  and direct  countercurrent washing techniques in a three-stage bleaching sequence at a water consumption of approximately 25 m /adt. This was done by recycling filtrates from a 3  laboratory-scale bleaching apparatus on a batch basis.  The production of  chlorophenolics from these tests were compared to that from open bleaching at a water consumption of roughly 280 m /adt. 3  At 20% substitution, direct countercurrent  washing produced slightly lower levels of chlorophenolics than the jump-stage method. However, both washing techniques  reduced the formation of chlorophenolics  substantially over open bleaching. The effects of washing technique and recycle were not distinguishable at 100% substitution.  Thus recycle and its interaction with  substitution were shown to be important factors in the production of chlorophenolic compounds. This study also investigated the effect of black liquor carryover. The addition of 19.5 kg Na^Oyadt as black liquor to the chlorination stage caused the formation of chlorophenolics to increase by approximately 50% at a substitution of 20%. The effect was not observed at 100% substitution. A considerable amount of research has gone directly into determining the effects of various process variables on softwood bleaching effluent toxicity.  The  majority of this effort has been spent investigating chlorine dioxide substitution in the first stage. Table 4 details thirteen such studies. Ten of the studies showed that the toxicity decreased as the level of chlorine dioxide substituted for chlorine increased. Three of the studies gave different results.  Hakulinen and Salkinoja-Salonen (1982)  found that the toxicity showed a maximum at 50% substitution with 100% substitution  Table 4 Studies of the effect of chlorine dioxide substitution on softwood bleaching effluent toxicity Reference  Effluent(s) tested  Bioassay organising)  Levels of C 1 0 substitution tested (%) 2  Betts and Wilson (1966)  C  Atlantic salmon  0, 30, 55, 70, 100.  Gall and Thompson (1973)  C  fish  0, 50.  Rapson, Anderson, and Reeve (1977)  C  fish  0, 70.  Wong et al. (1978)  CE combined  rainbow trout  0, 50.  Voss, Wearing, and Wong (1981)  C and E  rainbow trout  0, 58.  Belt, Joyce, and Chang (1982)  C and E  Daphnia pulex 0, 30, 50, 70.  Hakulinen and Salkinoja-Salonen (1982)  CE combined  Daphnia magna 0, 50, 70, 100.  Donnini (1982a)  C  Daphnia magna 0, 10, 70.  Donnini (1982b)  C and E  Daphnia magna 0, 4, 10, 20, 30, 50, 70, 100.  Kutney, Macas, and Donnini (1983)  CEDED combined  Daphnia magna 0, 10, 30, 50, 70.  Nikki and Korhonen (1983)  C and E  Daphnia magna 20, 100.  Liebergott et al. (1990b)  C(EO) combined  rainbow trout  10, 30, 50, 55, 60, 65, 70, 90.  Liebergott et al. (1990a)  C(EO) combined  fathead minnows and Ceriodaphnia  10, 30, 50, 70, 90.  a  "Hardwood results included.  17 being less toxic than 0% substitution. Kutney, Macas, and Donnini (1983) found that the toxicity increased with increasing substitution up to 70%, the highest level studied. Lastly, Donnini (1982b) claimed in comparison to 0% substitution that 4 and 10% substitution decreased the toxicity, 20 and 30% substitution gave the same toxicity, and 50, 70, and 100% substitution decreased the toxicity to the lowest values measured. However the conclusions of Kutney, Macas, and Donnini (1983) and Donnini (1982b) were based on improper data analysis. Firstly, confidence intervals for each toxicity value were based on the mean deviations of duplicate tests instead of the standard deviations.  Secondly, the mean deviations expressed as percentages of the  means of duplicate LC50 tests were applied to the data in the form of toxic units instead of LC50 values. As a result the true error is much larger than that shown by the researchers. The construction of statistically correct confidence limits applied to the LC50 data before conversion to toxic units shows that none of the toxicity values are significantly different.  Thus the conclusions of these two studies are not valid.  This leaves only the work of Hakulinen and Salkinoja-Salonen (1982) as being somewhat contrary to the consensus that chlorine dioxide substitution decreases the toxicity of bleach plant effluent. Another study listed in table 4 with questionable conclusions is that of Rapson, Anderson, and Reeve (1977).  These investigators used LT100  values to compare toxicities at two levels of chlorine dioxide substitution.  Such  measurements at the end of the mortality distribution are statistically unreliable  18 because of the distinct probability that the last organism to respond will be exceptionally deviant from the norm. Process variables other than chlorine dioxide substitution have received much less attention in regards to their effect on toxicity. Liebergott et al. (1990a) found that the use of DC addition at an interval of 1 minute caused the toxicity to remain constant regardless of the level of substitution, with 70% being the highest level studied.  Thus the benefits of substitution occurred only with simultaneous or CD  addition. Other studies failed to find this interaction between mode and substitution. Liebergott et al. (1990b) used DC addition at an interval of 1 minute and found the toxicity decreased with increasing substitution from 30 to 70%.  Belt, Joyce, and  Chang (1982), Donnini (1982a), and Gall and Thompson (1973) all found that the toxicity decreased with increasing substitution even with the addition of chlorine dioxide before chlorine. A logical explanation of this controversy has not been given. Donnini (1982b) investigated the effects of both mode and multiple on toxicity. However, due to previously discussed errors in data analysis the results were inconclusive. Voss, Wearing, and Wong (1981) found that the toxicity of chlorination filtrate showed a maximum at a multiple of 0.22 %/ml.  The toxicity of the  corresponding extraction filtrates remained constant regardless of the multiple applied in the first stage. In addition, the toxicity of both C and E stage effluents was found to increase with decreasing pH in the first stage. This effect was most pronounced below a pH of 1.5.  19 The carryover of black liquor into the chlorination stage with the unbleached pulp has a detrimental effect on toxicity. Thomas, Becker, and Gibney (1975) added 5.4 kg Na^O/adt as black liquor to a well-washed hemlock kraft pulp prior to bleaching with 100% chlorine. As a result the chlorinationfiltrate'sLC50 to rainbow trout went from 13% without carryover to 10%. The toxicity of the extraction filtrate was constant. Germgard et al. (1984) added 25 kg Na^Oyt as black liquor to a wellwashed softwood kraft pulp prior to bleaching with 100% chlorine in the first stage. The number of toxic units to luminescent marine bacteria (Microtox System) of the combined effluent from the CED sequence went from 24 without carryover to 26. Concurrently the ether extractable TOC1 increased by 34% as soluble lignin fragments in the carryover reacted with active bleaching chemical to form additional low molecular weight chlorinated organic compounds. Thus black liquor carryover results in higher concentrations of both extractives as well as chlorophenolics in bleaching effluent. Both factors increase toxicity. Nikki and Korhonen (1983) investigated the effect of black liquor carryover under different techniques of pulp washing and at two levels of chlorine dioxide substitution using a pine kraft pulp.  At 20% substitution, the addition of  19.5 kg NajSCyadt as black liquor caused the toxic emission factor determined with Daphnia magna for the three-stage sequence to increase by 8% using direct countercurrent washing and by 17% using jump-stage countercurrent washing. The toxic emission factor remained constant when black liquor was added using 100% substitution in the first stage.  20 Upon investigating the effect of different washing techniques on toxicity, Nikki and Korhonen (1983) showed that direct countercurrent washing gave lower toxic emission factors than jump-stage countercurrent washing at both 20 and 100% chlorine dioxide substitution. The effluents tested were produced in the laboratory by recycling filtrates from a series of 30 batch bleachings. Since roughly 1 month was required to produce each effluent, some uncertainty is cast on the applicability of the results. Unfortunately, a comparison to open bleaching was not made in terms of toxicity. Thus the effect of the degree of effluent recycle on toxicity has not been determined.  2.2. AOX Approximately 90% of the elemental chlorine used in pulp bleaching ends up as inorganic chlorine in the bleaching effluent. The remaining 10% is combined with lignin fragments in the form of organically bound chlorine. Most of this organically bound chlorine also ends up in the effluent.  The presence of large quantities of  chloride in bleach plant effluent means that these streams cannot be reclaimed using conventional kraft recovery techniques.  As a result, the bleaching effluent with its  associated chlorinated organic compounds is eventually discharged to the environment. In the 1970s and 1980s, much effort went into identifying individual organic compounds found in bleach plant effluent.  However, the emphasis has shifted to  determining the total amount of organically bound chlorine produced.  Due to the  potential environmental impact of chlorinated organic compounds, legislation has and will likely continue to be enacted to control discharges of such substances as a whole.  21 The first measurements of organically bound chlorine in pulp bleaching effluents were performed by Rapson and Anderson (1966). This method involves the determination of chloride ion by potentiometric titration before and after the mineralization of organically bound chlorine. The organic chlorine is thus calculated by difference.  Such measurements will accordingly be referred to as AOC1.  Nitric  acid was used for mineralization, but hydrogen peroxide or combustion can also be used. The method is limited because the mineralization is not completely efficient. In addition, because of the preponderance of inorganic chloride, a small error in its determination can have a large effect on the results of the organically bound portion. Despite these problems Rapson and Anderson (1966) clearly showed that chlorine dioxide substitution decreases the formation of AOC1. At 0% substitution, the AOC1 from the first two bleaching stages was 5.5 kg/adt. At 20 and 40% substitution, this decreased to 4.1 and 3.2 kg/adt respectively.  No organically bound chlorine was  found at substitution levels of 60, 80, and 100%. With the aim of improving accuracy and precision, two methods were developed for the direct determination of organically bound chlorine. Both procedures involve the separation of inorganic and organic chlorine followed by conversion of the organic chlorine to chloride by combustion. The first method, developed by Sjostrom, Radestrom, and Lindstrom (1982), yields the parameter total organic chlorine (TOC1). This technique uses ultrafiltration and resin adsorption for separation.  Chloride  produced from the organic chlorine is determined by potentiometric titration. The second technique, method 506 of APHA (1985), gives a value called adsorbable  22 organic halogen (AOX). Activated carbon adsorption is used for separation while the chloride from organic compounds is quantified by microcoulometric titration. The AOX procedure is becoming widely adopted as the basis for measurements and legislation as it is both simpler and less time consuming to perform than the TOC1 procedure. The three methods for the determination of organically bound chlorine in bleaching effluents do not yield identical results.  Liebergott et al. (1990b) related  AOX and AOCl measurements, both expressed as kg/odt, with  AOX = 0.897AOC7 +0.13  () 2  Equation 2 was developed for AOCl measurements between 2 and 7 kg/odt, which resulted from experiments performed with levels of chlorine dioxide substitution up to 70%. This result implies that AOCl values are slightly larger than AOX values and thus contradicts the results of some work done earlier. Since Rapson and Anderson (1966) found no AOCl above 40% substitution and not all organic chlorine is converted to chloride by the AOCl procedure, it would appear that AOCl values should be lower than AOX values. Axegard (1989) found that AOX measurements were on average 20% greater than TOC1 measurements.  Some researchers believe this  difference is caused by volatile chloro-organic compounds being excluded in the TOC1 analysis.  However, chloroform accounted for less than 1% of the AOX values of  Axegard (1989). Considering the finding of Hall et al. (1988) that chloroform makes up more than 99% of the volatile fraction of AOX in C and E stage effluents, it is  23 unlikely that the different between T0C1 and AOX measurements is caused by volatile compounds. While the majority of organochlorine bonds are formed in the first bleaching stage, significant quantities of chlorinated organics are found in the effluents from the first and second stages.  An industrial (DC)(EO)DED bleaching sequence utilizing  50% chlorine dioxide substitution on a softwood kraft pulp with a permanganate number of 19 ml produced 2.0 kg AOX/t (Hall et al. 1988). The chlorination stage accounted for 66% of this AOX while the extraction stage produced the remainder. Histed and Vega Canovas (1989) bleached a softwood kraft pulp with a kappa number of 27.8 ml using a (DQ(EO)(D/E/D) sequence in the laboratory.  Despite widely  varying conditions in the first stage, the amount of AOX produced in the last three stages was constant at 0.1 kg/adt.  This was not significant when compared to the  AOX discharged from the first two stages which was as high as 5.2 kg/adt. Molecules which comprise the organically bound chlorine in bleaching effluent range in size from low molecular weight compounds such as chloroform to complex chorolignin fragments having molecular weights greater than 10,000 u.  Pfister and  Sjostrom (1979a) found that 38% of the AOC1 in the C stage effluent from a CEH sequence was bound in compounds with a molecular weight less than 1,000 u. In contrast, the corresponding figure for the extraction effluent from the same sequence was only 3% (Pfister and Sjostrom 1979b). These results were largely corroborated by the work of Lindstrom, Nordin, and Osterberg (1981). It was found that 30% of the TOC1 in a C stage filtrate was associated with compounds with a molecular weight  24 less than 1,000 u while this fraction represented only 5% of the TOC1 in the corresponding E stage filtrate.  Pfister and Sjostrom (1979a, 1979b) also found that  organic chlorine (AOCl) accounted for 20% of the mass of all nonvolatile organics dissolved in chlorination effluent but only 8% in extraction effluent. Thus chlorination filtrate contains substantial amounts of low molecular weight chlorinated organics while extraction fdtrate is primarily comprised of high molecular weight chlorolignin. In addition, C effluent has a much higher degree of substituted chlorine than E effluent.  Most of the environmental impact caused by organochlorine  compounds from pulp bleaching comes from the low molecular weight fraction. However, it has been shown theoretically possible for chlorolignin to be degraded in receiving waters to low molecular weight chlorinated organics (Lindstrom, Nordin, and Osterberg 1981). The most important factor in determining the amount of chlorinated organic material produced during bleaching is the consumption of total elemental chlorine (TEC).  Germgard and Larson (1983) defined the TOCl produced from a bleaching  sequence in kg/t with  TOCl = k(ACL + - A C / O " + -AC7CU 2 5 2  2  (3)  where AC7 , ACIO', and AC10 are the consumptions of chlorine, hypochlorite, and 2  2  chlorine dioxide in kg/t, expressed as active chlorine, and A: is a constant for a given pulp.  The average value of k for an oxygen delignified kraft pulp (20 ml kappa  number), which was bleached industrially using a (C+D)(EO)DED sequence, was  25  found to be 0.089 kg TOCl/kg TEC (Axegard 1989). Not surprisingly a similar result was found for AOX production in kg/t as given by AOX = k(Cl + 0.526C/O ) 2  2  (4)  where Cl and C10 are the consumptions of chlorine and chlorine dioxide in kg/t, 2  2  expressed as chlorine and chlorine dioxide respectively (Earl and Reeve 1989a). A k value of 0.10 kg AOX/kg TEC was determined for a softwood kraft pulp (30 ml kappa number), which was bleached in the laboratory using a (DC)E sequence. Equations 3 and 4 are nearly identical since Earl and Reeve did not utilize hypochlorite as a bleaching agent, and both l/5AC/c9 and 0.526ClO are equivalent to the elemental 2  2  chlorine in chlorine dioxide. The only significant difference is the k value required for each of the two forms of chlorinated organics measurement. It is readily apparent then that chlorinated organics production can be decreased simply by using less bleaching chemical. Thus processes that decrease the amount of lignin entering the bleach plant, such as oxygen and extended delignification, produce less organically bound chlorine due to decreased active chlorine consumption. Another method of reducing the formation of organochlorine compounds is by replacing some of the chlorine used in the first bleaching stage with chlorine dioxide. This is commonly known as chlorine dioxide substitution.  Since, as a first  approximation, a constant amount of active chlorine or oxidizing power is required to perform a given bleaching task regardless of the chemical used, equation 3 shows that chlorine dioxide will produce one-fifth as much TOC1 as an equivalent amount of chlorine.  The Canadian pulp and paper industry plans to use chlorine dioxide  26 substitution with other process and treatment options to meet recently imposed AOX discharge limits. The one-fifth factor in equation 3 originates from the difference in oxidizing strength of chlorine and chlorine dioxide. One mole of elemental chlorine as chlorine dioxide possesses 5 oxidizing equivalents while 1 mole of elemental chlorine as chlorine has 1 oxidizing equivalent.  Alternatively, on an oxidizing equivalent basis,  chlorine dioxide introduces one-fifth as many chlorine atoms to the bleaching process as does chlorine.  Thus equation 3 suggests that the reduction in TOC1 with  substitution is entirely attributable to the decrease in the number of chlorine atoms added.  Correspondent with this is the suggestion that an atom of chlorine from  molecular chlorine has the same probability of forming an organochlorine bond as does an atom of chlorine from chlorine dioxide.  Yet it is well understood that  chlorine dioxide does not partake in substitution reactions that result in the formation of TOC1.  However, the reactions of chlorine dioxide with lignin result in the  production of significant quantities of hypochlorous acid (Bergnor et al. 1987). Chlorine is then formed in equilibrium with the hypochlorous acid. This secondary source of chlorine is likely responsible for the formation of chlorinated organics from chlorine dioxide. The effect of mode of addition on the production of chlorinated organics is small but measurable. Histed and Vega Canovas (1989) found different values for the constant in equation 4 for two modes of addition.  CD mode had a k value of  0.088 kg AOX/kg TEC while DC mode had a k value of 0.105. Thus a 30 ml kappa  27 number pulp bleached with a multiple of 0.20 %/ml using 20% chlorine dioxide substitution would yield 0.8 kg AOX/adt less if CD mode was used instead of DC mode.  The time span between chemical additions was 2 minutes, and the k values  were found to apply to both kraft and oxygen pulps. Liebergott et al. (1990a) found similar results.  CD mode was determined to yield between 0.3 and 0.7 kg AOX/adt  less than DC mode. Simultaneous addition was deemed to be equivalent to CD mode with respect to AOX formation.  A 1-minute delay was used between chemical  additions. It is not surprising that the addition of chlorine first produces the least AOX since it also produces the least dehgnification.  As was mentioned previously, the  production of chlorinated organics is a function of the amount of lignin removed in bleaching.  While industry would not consider using CD addition due to poorer  dehgnification, this may prove to be a simple way to reduce organochlorine discharges. Liebergott et al. (1990a) found that due to the decreased dehgnification of CD addition up to 4 kg/adt more chlorine dioxide was needed in the third stage to reach a target brightness of 90% ISO in afive-stagesequence. According to equation 4, this would produce an extra 0.2 kg AOX/adt from the last three stages of bleaching. In theory then, the use of CD mode would produce a net decrease in AOX formation between 0.1 and 0.5 kg/adt when compared to DC addition. The effect of bleaching effluent recycle on the production of organically bound chlorine has not been studied extensively. One study on this topic is that of Mattinen and Wartiovaara (1981).  A pine kraft pulp with a kappa number of 28 ml was  28 bleached to 89% ISO brightness in the laboratory with various levels and methods of effluent reuse. The (DQEDED sequence using 30% chlorine dioxide substitution was performed on a batch basis. The results are summarized in table 5. Two conclusions can be drawn from this work: 1.  Tighter closure produces less organically bound chlorine with the trial using jump-stage countercurrent washing at an effluent flow of 25 m /adt being a 3  significant exception. 2.  Direct countercurrent washing, in which alkalinefiltratesenter acidic bleaching stages,  produces  more  organically  bound  chlorine  than  jump-stage  countercurrent washing, in which acid and alkalinefiltratesare kept separate. However, these conclusions are much less clear when the following limitations are considered: 1.  Organically bound chlorine was measured using the AOCl method which has been shown to be inaccurate.  2.  All of the trials were performed at different chemical consumptions making comparisons difficult.  3.  Since 1 month of batch bleachings was needed to reach a steady state, the integrity of the effluents is questionable. A second study looked specifically at the recycle of chlorination filtrate for  brown stock dilution (Histed and Nicolle 1973). A kraft pulp was chlorinated in the laboratory on a batch basis using a constant charge of active chemical. After each batch chlorination, the filtrate generated by thickening the pulp from 3.5 to 10%  29 Table 5 The effect of filtrate recycle on the formation of organically bound chlorine (Mattinen and Wartiovaara 1981) Method of recycle  Effluent flow (m /adt)  Active Cl consumed (%)  426  8.9  4.9  50  9.1  3.3  75  8.5  3.2  40  8.6  2.5  25  9.3  3.5  27  9.8  5.8  21  10.6  3.6  3  open bleachery  jump-stage countercurrent direct countercurrent  2  AOC1 produced (kg/adt)  consistency was recycled to dilute the next batch of unbleached pulp.  Subsequent  washing filtrate was discarded. Nine cycles of this process were performed with the concentration of inorganic chloride being measured in the recycled filtrate after each cycle. The process was gradually reaching a steady state where the inorganic chloride leaving the system with the chlorinated pulp at 10% consistency would equal the chlorine added to the system as active chemical. However, the investigators calculated the difference in inorganic chloride between successive cycles and mistakenly called this value the inorganic chloride produced by the second of the two successive cycles. This erroneously labelled quantity decreased with increasing cycle number. Since a constant amount of chlorine was added at each cycle, and the amount of inorganic  30 chloride produced was thought to be decreasing, it was concluded that the amount of organically bound chlorine must increase with cycle number or the degree of recycle. Unfortunately the amount of chloride leaving the system with the chlorinated pulp was not accounted for. Had this been done it is unlikely that the same conclusion would have been made. Thus specific knowledge on the effect of chlorination filtrate recycle on the formation of organically bound chlorine is lacking. Histed and Nicolle (1973) also found that the pH of the chlorination filtrate dropped from 1.86 after the first cycle to 1.05 after the ninth cycle.  It was thus  reasoned that the supposed increase in chlorinated organics resulted from this change in pH. As pH decreases, molecular chlorine is favored in its equilibrium hydrolysis. Since only molecular chlorine reacts via substitution forming carbon-chlorine bonds, the formation of chlorinated organics might be expected to rise.  Chlorine dioxide  substitution might also have an effect due to pH on the production of organochlorine compounds.  Since only molecular chlorine produces hydrochloric acid both via  hydrolysis and as a byproduct of substitution reactions, the use of chlorine dioxide in its place causes the pH of chlorination to increase.  This increases the proportion of  active chlorine reacting via oxidation thereby producing less organic chlorine. The carryover of black liquor into the bleach plant with the unbleached pulp has a significant impact on the formation of chlorinated organic compounds. Germgard et al. (1984) found that TOCl production increased by 2.6 kg/t for a carryover of 25 kg NajSO/t in a CED bleaching sequence.  Johannes and Lowe  (1989) found that AOX production increased by 3.9 kg/adt for 22.5 kg Na^Oyadt of  31 carryover in a (C+D)(EO)DED sequence using 10% substitution.  With black liquor  carryover in Canadian softwood kraft pulp mills averaging 30 kg NajSO/adt (Byzyna, Clifford, and Billmark 1989), it is evident that substantial amounts of AOX come from this source. An interesting finding of both of the studies cited on black liquor carryover was that the quantity of organically bound chlorine formed with respect to black liquor was the same as that with respect to in situ lignin. This result is consistent with the view that black liquor is essentially a solution of lignin fragments which, aside from being more accessible than in situ lignin, reacts in a similar fashion with active chlorine. Apparently then the production of chlorinated organics due to black liquor carryover is included in equations 3 and 4 as part of the chemical consumption. Thus the production of organochlorine compounds is constant for a fixed chemical consumption regardless of the amount of carryover.  However, the extent of  delignification will decrease as the carryover increases due to the accessibility of the black liquor lignin fragments.  2.3. Chlorate One of the few disadvantages in the use of chlorine dioxide in the first bleaching stage is the formation of chlorate (C10 ). 3  Not only is chlorate toxic to  certain algae thus posing an environmental problem but its formation represents a nonreversible loss of available chlorine as it is unreactive to lignin. While chlorate is  32 produced in all bleaching stages utilizing chlorine dioxide, the primary interest here is in that formed in the first stage. Within several years of opening, a Swedish pulp mill, utilizing between 15 and 30% chlorine dioxide substitution, managed to kill all of the bladder wrack kelp (Fucus vesiculosus), a brown alga, in a 12-sq-km area next to its outfall in the Baltic Sea (Eckerman 1987).  The problem was traced to chlorate in the mill's effluent.  Research showed that by decreasing the dissolved oxygen in the first portion of the mill's aeration lagoons the chlorate could be biologically reduced to chloride. BOD removal was unaffected as the lagoons were oversized to begin with.  Germgard  (1989) reviews the development of this biological solution in addition to outlining another treatment method in which chlorate reacts with sulfur dioxide according to  C10" + 3S0 + 3H 0 3  2  2  CI' + 3H S0 2  4  (5)  This reaction is favored by high temperature and low pH. Thus the bleach plant acid sewer, where the chlorate is in its most-undiluted state, is a suitable stream in which the reaction can occur. If conditions are right, less than 1 hour of reaction time is needed. While effluent treatment is capable of substantially reducing chlorate discharges to the environment, an understanding of how the chlorate is formed in the first place might lead to bleaching process modifications in which less chlorate is produced. Thus an increase in the efficiency of chlorine dioxide bleaching reactions would  33 simultaneously help solve an environmental problem. Some progress has been made in this area. Lindgren and Nilsson (1975) studied the formation of chlorate from the oxidation of nine lignin model compounds with chlorine dioxide.  Compounds  containing phenolic hydroxyl groups produced the lowest yields of chlorate, between 3 and 5% of the oxidizing power of the chlorine dioxide charged.  Compounds  containing both phenolic hydroxyl and carbonyl groups gave between 10 and 15% chlorate. The highest yields, between 12 and 35%, came from compounds containing an aliphatic double bond.  It was also discovered through some related work that  chlorine dioxide preferentially attacks phenolic hydroxyl groups over other functional groups. Based partly on the above findings, Bergnor et al. (1987) summarized two pathways in which chlorine dioxide reacts with lignin. Path A involves the reaction with phenolic hydroxyl groups. One-half of the chlorine dioxide reacting via this path is reduced to chlorite.  In turn, small amounts of chlorite disproportionate to form  chlorate, chlorine dioxide, and chloride.  The other half of the chlorine dioxide  reacting via path A is reduced to hypochlorous acid with no chlorate formation. Path B involves the reaction with carbonyl groups and aliphatic double bonds. Here onehalf of the chlorine dioxide is reduced to hypochlorous acid with no resultant chlorate. The other half is directly oxidized to chlorate.  Thus path B produces most of the  chlorate generated from the reaction between chlorine dioxide and lignin. This two path scheme helps to explain why both oxygen and CE prebleached pulps produce  34 more chlorate per mole of chlorine dioxide than kraft pulp. Since kraft pulp lignin has a larger proportion of phenolic hydroxyl groups, more chlorine dioxide reacts via path A, and less chlorate is formed. Chlorate can also be produced by the spontaneous decomposition of chlorine dioxide according to  2C10 + H 0 — ClO-3 + ClCr + 2H 2  2  2  +  (6)  Knowledge on this reaction, which is catalysed by hypochlorite, has been summarized by Nilsson and Sjostrom (1974).  It was concluded that decomposition is only  significant if chlorine is present and the pH is greater than 5. A summary of studies of chlorate formation in the first bleaching stage is given in table 6.  It is common for investigators to show results graphically as plots of  chlorate production against chlorine dioxide substitution. for each mode of chemical addition.  A separate curve is drawn  In general, these curves show an increase in  chlorate production for increasing levels of substitution. However, in four of the six investigations a maximum in chlorate production occurred between 70 and 85% substitution.  In these instances production then decreased as substitution was  increased further. There is limited agreement as to the effect of mode of addition on chlorate production. The addition of chlorine dioxide before chlorine gave the lowest production in all cases except one at 90% substitution (Bergnor et al. 1987). Beyond this, effects of mode are unclear.  35  Table 6 Studies of chlorate formation in the first bleaching stage Reference  Unbleached kappa no. (ml)  Multiple (%/ml)  Mode(s) studied  Location (s) of relative maxim um(s)  (%C10,  Relative order of chlorate production  Maximum chlorate production (kg/adt)  substitution)  Kramer (1972)  323  0.20  C+D  —  Reeve and Rapson (1980)  32  0.16  CD C+D DC  70  C+D>CD>DC  Bergnor et al. (1987)  29.3  CD C+D DC  70 75  CD>C+D>DC C+D>CD>DC C+D>DC>CD  5.2  b  0.20  2.5"  —  3.2  C  d  £  e  Axegard (1987)  17.2  0.20  C+D DC  70 70  C+D>DC  3.0  Germgard and Karlsson (1988)  17.2  0.18  C+D DC  75 85  C+D>DC  2.2  Liebergott et al. (1990a)  31.2  0.22  CD C+D DC  —  CD=C+D>DC  4.6  "From graphical results at 25°C and pH 2. •Converted from a permanganate number of 21 ml. Tor C10 substitution below 60%. 2  Tor C10 substitution between 60 and 90%. 2  Tor 90% C10 substitution. 2  'Pulp moisture not specified.  f  f  36 However, in studying the effect on chlorate formation of small amounts of chlorine added to the first chlorine dioxide stage, Nilsson and Sjostrom (1974) found that less chlorate was formed when the chlorine was added before, as opposed to with, the chlorine dioxide. While not directly applicable to the chlorination stage, bleaching in a D stage with 2, 4, and 6% chlorine substitution as studied by Nilsson and Sjostrom is somewhat analogous to bleaching in a C stage with 98, 96, and 94% chlorine dioxide substitution. This tends to corroborate the finding that at high levels of chlorine dioxide substitution CD mode yields less chlorate (Bergnor et al. 1987) In terms of maximum chlorate production, table 6 shows a wide range of values.  However, much of this apparent discrepancy is readily explainable.  Firstly,  the maximum productions found by Axegard (1987) and Germgard and Karlsson (1988) are low because of low unbleached kappa numbers.  Axegard utilized an  oxygen delignified softwood kraft pulp while Germgard and Karlsson worked with a birch kraft pulp.  Since less lignin was present, less bleaching agent was applied  resulting in less chlorate.  Secondly, the results of Kramer (1972) are low because  iodometry was used to determine the chlorate.  Easty, Johnson, and Webb (1986)  analyzed two chlorination stage filtrates for chlorate using iodometry, a traditional titrimetric method, and ion chromatography, a modern instrumental technique. In both cases iodometry gave substantially lower results than ion chromatography. Iodometry is inherently less accurate because it requires the simultaneous determination of other forms of chlorine with the errors from each determination being additive. Lastly, the  37 results of Reeve and Rapson (1980) are low because of the use of a low multiple in addition to the use of iodometry. Other less extensively studied aspects of chlorate formation include the following observations. Kramer (1972) found that chlorate formation was independent of the reaction times used of 15, 30, and 60 minutes. This is consistent with the fact that bleaching reactions as carried out industrially are largely complete after only a few minutes.  It was also found that chlorate production increased slightly as the  temperature was increased from 25 to 60°C. In addition, chlorate formation decreased with increasing pH. At the same time however, the degree of dehgnification failed to improve significantly. Liebergott et al. (1990a) found similar results in their work on the effect of pH on chlorate formation and dehgnification. Bergnor et al. (1987) investigated the effect of small amounts of chlorine in the chlorine dioxide on the formation of chlorate. With 70% substitution and a DC mode of addition, having 10% of the available chlorine in the chlorine dioxide solution as chlorine caused a 0.5 kg/t decrease in chlorate formation. At 30% substitution, the difference between using pure chlorine dioxide and that containing some chlorine was not significant. In addition, an increase in the multiple from 0.17 to 0.23 %/ml caused chlorate production to increase by approximately 2 kg/t at 70% substitution.  This  effect was not detected at 30% substitution. The only study on the effect of effluent recycle on chlorate production was done by Reeve, Anderson, and Rapson (1980).  Chlorination effluent recycle was  simulated by adding first-stage filtrate from a pulp mill to a series of batch laboratory  38 bleaching trials.  Both C+D and DC modes of addition were investigated at 70%  substitution. Chlorate production was found to be unaffected by recycle levels up to 10% total organic carbon (TOC) on pulp. Approximate calculations show that the maximum recycle of chlorinationfiltratepossible industrially corresponds to 2% TOC on pulp. The mill effluent was concentrated by vacuum evaporation enabling this value to be surpassed.  2.4.- Kappa Number The primary purpose of pulp bleaching is to produce a pulp of suitable brightness. The first two stages of afive-stagebleaching sequence (CEDED) remove approximately 80% of the lignin remaining in the unbleached pulp.  Since little  brightening occurs in these two stages, they are often referred to as prebleaching. Brightness levels are increased in the last three stages as the final portions of lignin are removed. The extent of bleaching is followed by measuring the lignin content after the first extraction stage using either the kappa or permanganate number. Brightness measurements are then made after each of the chlorine dioxide stages. Utilization of this data helps allow the chemical charges to each stage to be adjusted so that the required brightness is achieved. Germgard and Karlsson (1984) investigated the effects of chlorine multiple and chlorine dioxide substitution in the chlorination stage on total chemical consumption to reach 90% ISO brightness in a (DC)EDED sequence.  The minimum first stage  multiple required to reach the target brightness decreased from 0.20 %/ml at 10%  39 substitution to 0.14 %/ml at 90% substitution. This resulted in an optimum CE kappa number that was also dependent on the level of substitution.  Thus, at 10%  substitution, the optimum CE kappa number was 6.4 ml. This decreased to 4.8 ml at 30% substitution before increasing to 8.7 ml at 90% substitution. Another significant finding was that the consumption of active chlorine in the two D stages to reach the final brightness was directly related to the CE kappa number and independent of the level of substitution used in the chlorination stage.  A more detailed look at the  relationships between total chemical usage, CE dehgnification, and final brightness level is beyond the scope of this review as only the prebleaching stages were investigated. Chlorine multiple has a large effect on the dehgnification attained in the first two bleaching stages. One of many studies involving this variable is that of Rapson, Anderson, and Reeve (1977). A softwood kraft pulp with a kappa number of 31.3 ml was bleached using a (D+C)E sequence at 70% substitution. The kappa number after extraction decreased linearly from 8.0 ml at a multiple of 0.13 %/ml to 4.5 ml at a multiple of 0.21 %/ml. It then approached an asymptotic limit of 2.5 ml at a multiple of 0.39 %/ml.  Thus, once the asymptotic portion of dehgnification is reached,  bleaching becomes increasingly less efficient. The effects of chlorine dioxide substitution and mode of addition have also been well studied. Reeve and Rapson (1980) bleached a softwood kraft pulp with a permanganate number of 21 ml using a constant chemical charge at 60°C.  The CE  permanganate number was plotted as a function of chlorine dioxide substitution for the  40 three modes of addition. The CE permanganate number at 0 and 100% substitution was 5.1 ml. Simultaneous addition gave similar values for all levels of substitution. The addition of chlorine dioxide first resulted in better delignification with a minimum permanganate number of 3.8 ml occurring at 50% substitution.  The addition of  chlorine first resulted in poorer delignification with a maximum permanganate number of 5.7 ml occurring at 50% substitution.  Thus the system exhibited interaction  between substitution and mode as well as curvature. Similar results were reported for bleaching trials performed at 25°C.  However, at this temperature all three modes of  addition showed worsening delignification as the substitution level was increased past 50%. This was attributed to a reaction tirrie not long enough for all of the chlorine dioxide to react. chlorine.  Lignin is known to consume chlorine dioxide more slowly than  Thus temperature has little effect on delignification as long as sufficient  time is allowed for all of the chemical to react. In addition, Reeve, Anderson, and Rapson (1980) showed that chlorine contamination of the chlorine dioxide used in the DC mode enhances delignification. The largest effect was a 0.5 ml decrease in CE kappa number when 8% of the available chlorine in the chlorine dioxide solution was chlorine. The effect of chlorination effluent recycle on CE kappa number is not as readily apparent from the literature as the effects of other variables. A weak trend of lower CE kappa number with increasing C stage pH was shown by du Manoir (1980). Presumably this was due to the shift towards hypochlorous acid at higher pH in the equilibrium hydrolysis of chlorine. A greater portion of the active chlorine would be  41 consumed by oxidation resulting in more efficient delignification.  Thus, as effluent  recycle decreases the first stage pH, delignification might be expected to decrease. It should be noted that the pH of the first stage was adjusted by the addition of H Q and NaOH. The influence of dissolved organic material due to effluent recycle was thus eliminated. The effect of chlorination effluent recycle on delignification was studied directly by Histed and Nicolle (1973).  Nine cycles of a (CD)ED sequence were  carried out on a softwood kraft pulp with a kappa number of 24.3 ml. Chlorination effluent recycle was performed on a batch basis and consisted of the filtrate produced when the chlorinated pulp was thickened from 3.5 to 10% consistency.  With the  chemical charge in the first stage held constant, the CE kappa number went from 4.2 ml after the second cycle to 4.4 ml after the ninth cycle. At the same time however, the chlorination stage pH dropped from 1.59 to 1.05.  Thus the small decrease in  delignification might have been caused by either dissolved organic material or pH. The influence of dissolved organic matter in the chlorination stage on CE kappa number can be very significant. The effect of recycling black liquor, extraction filtrate, and chlorination filtrate on delignification was investigated by Reeve, Anderson, and Rapson (1980).  A well-washed softwood kraft pulp with a kappa  number of 34.2 ml was bleached on a batch basis using a (DC)E sequence with a constant chemical charge and 70% chlorine dioxide substitution. The three different filtrates were added to the first stage at various levels expressed as the percentage of TOC on pulp.  Black liquor increased the CE kappa number by 9.3 ml for each  42 percentage increase in TOC on pulp. The corresponding values for the extraction and chlorination fdtrates were 3.2 and 0.3 ml respectively. Chlorination fdtrate consumed relatively little active chlorine as it had already been exposed to active chemical. Conversely, black liquor having never been exposed readily consumed active chlorine. Another investigation of the effect of effluent recycle on delignification is that of Wartiovaara (1980a). . A softwood kraft pulp with a kappa number of 28 ml was bleached using 30% substitution in a (DC)EDED sequence.  Seven different filtrate  recycling methods, including direct and jump-stage countercurrent washing, were used at various levels of fresh water consumption.  The filtrates were brought to  equilibrium through a series of 40 batch bleachings while the pulp was brought to 89% ISO brightness by adjusting the chemical charges.  The effectivity of the first  stage, defined as the difference in kappa number between the brown stock and the CE pulp divided by the active chlorine consumed in the first stage, was plotted against the TOC content of the unbleached pulp.  This plot gave a single line with a constant  slope of -0.38 ml/(%Cl »%TOC) for all of the methods of recycle employed. 2  Converting to the CE kappa number and substituting the average chemical charge used of 5.5% active chlorine, the slope becomes 2.1 ml/%TOC.  Thus the addition of 1%  TOC on pulp caused the CE kappa number to increase by 2.1 ml. The same increase in CE kappa number with effluent recycle occurred for each of the different methods of reuse. This contradicts the results of Reeve, Anderson, and Rapson (1980) when the compositions of the chlorination filtrates from direct and jump-stage countercurrent washing are considered.  C stage filtrate from direct  43 countercurrent washing contains a large amount of extractionfiltratewhereas that from jump-stage countercurrent washing contains none. Accordingly, direct countercurrent washing should result in a much higher increase in CE kappa number with increasing recycle than the jump-stage method. The reason for this discrepancy is not known. Black liquor contamination of the brown stock can be eliminated as a cause, as washing losses were 0.5 kg Na^Cyadt. However, it is interesting to note that the rate of increase in CE kappa number with recycled organic material of 2.1 ml/%TOC lies between the values for C and E effluents found by Reeve, Anderson, and Rapson (1980). This suggests a mixture of the two filtrates occurred even in the jump-stage case.  2.5. Viscosity The reactions of chlorine and chlorine dioxide in pulp bleaching are not limited to those involving lignin. Cellulose is degraded in the presence of chlorine resulting in chain cleavage.  This translates into a decrease in the degree of polymerization of  the cellulose. Once a critical level of degradation is reached, the strength properties of the resulting paper begin to decrease rapidly. A widely used method of observing the degradation is to measure the viscosity of a solution of the cellulose chains.  This  degradation, obviously undesirable for the production of paper pulps, is inevitable in the first bleaching stage. Current bleaching strategy simply minimizes it. In attempts to decrease water consumption, the pulp industry started to recycle bleach plantfiltratesextensively in the early 1970s. As a result the temperature in the  44 chlorination stage rose from ambient levels to over 40°C. This temperature increase caused the viscosity loss to increase to unacceptable levels.  Chlorine dioxide's first  use in the chlorination stage was to counteract the increase in cellulose degradation. Hypochlorous acid, formed from the hydrolysis of chlorine, is thought to be responsible for the oxidation via a free radical mechanism. Chlorine dioxide acts as a free-radical scavenger thereby protecting the cellulose. Many of the factors influencing the viscosity of CE pulps were investigated by du Manoir (1980). At a chlorination temperature of 30°C the CE viscosities at 0, 4, 10, 20, 50, 75, and 100% substitution were 12, 30, 28, 26, 28, 34, and 36 cp respectively.  The unbleached pulp viscosity was 37.7 cp. Thus, by far the largest  effect on viscosity occurred with the first small amounts of chlorine dioxide substituted. The same trials performed at 60°C gave viscosities approximately 5 cp lower than those at 30°C for all levels of substitution. All of the above trials utilized simultaneous addition except for the trials at 4% substitution, where the chlorine dioxide was added 10 seconds after the chlorine. The effect of mode of addition was also studied.  It was concluded that  chlorine dioxide had to be added so that it would be present during the later stages of chlorination, when most of the viscosity loss occurs. Thus, at 4% substitution, it had to be added 10 seconds after the chlorine. At this level of substitution, addition of the chlorine dioxide any sooner meant that it would be consumed rapidly by oxidizing lignin and thus would not be present later on to retard viscosity loss. As substitution was increased, the range of possible timings of addition became broader. Thus, at  45 30% substitution, chlorine dioxide could be added up to 10 seconds before the chlorine.  At 10 and 20% substitution, simultaneous addition was suggested as an  optimum.  The proposed timings for chlorine dioxide addition were based on both  viscosity protection and delignification efficiency. The effect of chlorination stage pH on viscosity was also investigated. As acidity was decreased past a critical pH of 1.8, viscosities decreased by approximately 20 cp for 4, 10, and 25% substitution added optimally at 30°C. Presumably this was caused by an increase in the concentration of hypochlorous acid due to the shift in chlorine hydrolysis with increasing pH. Since the recycle of chlorination filtrate decreases pH, it might also be expected to improve pulp viscosity at constant temperature.  This has not been  substantiated in the literature. A study by Wartiovaara (1980b) showed a very slight decrease in CE viscosity as filtrate recycle increased. A five-stage sequence using open washing and dilution gave a CE viscosity of 23.3 cp. Fresh water use was then decreased to levels of 33 and 26 m /adt by recycling filtrates on a batch basis. 3  Viscosities for these two trials were 22.7 and 21.2 cp respectively. However, it is not clear that this effect was caused by increased effluent recycle as the chemical charged to the first stage was varied to meet brightness requirements. Thus the multiple went from 0.19 %/ml for open washing to 0.20 %/ml at a water usage of 33 m /adt and to 3  0.21 %/ml at 26 m /adt. The first bleaching stage utilized 30% substitution added 30 3  seconds before the chlorine at a temperature of 45°C. The unbleached pulp viscosity was 26.3 cp. In a second study, Histed and Nicolle (1973) recycled on a batch basis 100% of the filtrate resulting from thickening a chlorinated stock from 3.5 to 10%  46 consistency.  Nine cycles of a three-stage bleaching sequence were performed using  5.7% substitution added just after the chlorine at a temperature of 25°C. The brown stock viscosity was 26.1 cp. After the second cycle the CE viscosity was 24.3 cp. This increased to 24.4 cp after nine cycles. Concurrently the pH decreased from 1.8 after the first cycle to 1.3 after the ninth cycle. The viscosity improvement, predicted by du Manoir (1980), did not occur under simulated industrial conditions because even open bleaching reaches a pH below the critical value, where little effect is seen on pulp viscosity. Filtrate recycle simply causes the pH to drop further below the critical pH. Lastly, it has been demonstrated by many investigators that viscosity decreases with increasing multiple.  One such study is that of Rapson, Anderson, and Reeve  (1977). For 70% substitution added simultaneously at 60°C, the CE viscosity dropped linearly from 36 to 27 cp when the available chlorine added to the first stage increased from 4 to 12% on pulp. The unbleached pulp viscosity was 36.9 cp.  CHAPTER 3 EXPERIMENTAL There is a distinct lack of information regarding the effect of filtrate recycle on the quality of bleach plant effluent.  Thus the primary goal of this research was to  determine the effect of the recycle of chlorination fdtrate for brown stock dilution on the production of toxicity, AOX, and chlorate from the first two stages of softwood kraft pulp bleaching.  This required the construction of a continuous lab-scale  chlorination apparatus. The variables of mode, substitution, and multiple were studied factorially in combination with recycle to determine the effect of interaction (see section 3.2).  In addition, the pulp was characterized in terms of both kappa number  and viscosity after the extraction stage, which was performed on a batch basis.  3.1. Apparatus A continuous laboratory-scale pulp chlorination apparatus was designed to enable: (1) the careful control of the factors affecting chlorination, (2) the recycle of chlorination fdtrate on a continuous basis thereby yielding effluent at steady state in a short period of time, and (3) the use of trout bioassays by the production of sufficient quantities of effluent in a short period of time. rendered  operational  is  represented  photographically in figures 2-12.  The apparatus as fabricated and  schematically  in  figure  1  and  shown  The accompanying 4-minute video presentation, 47  ©  S T O C K  ®  H E A D  ©  T A N K S T A N K  B L O W E R  ©  C l  ©  C I 0  ©  H  ©  H E A T  ©  B R O W N  ©  D E C K E R  ©  C H L O R I N A T I O N  ©  MIXER  2  C A R B O Y  2  C A R B O Y  2  0  T A N K E X C H A N G E R S T O C K S E A L  D E C K E R T A N K  ©  MIXER  N O . 2  R E A C T O R  ©  MIXER  N O . 3  ©  DILUTION  NO.1  Figure 1.  V A T  © ©  C H L O R I N A T I O N W A S H E R  S E A L  W A S H E R T A N K  oo  Figure 2.  Continuous laboratory-scale pulp chlorination apparatus  Figure 8. Chlorination washer  Figure 10. Constant-head tank  55 A Continuous Lab-Scale Pulp Bleaching Apparatus, documents the equipment in operation. Detailed specifications are given in appendix A. A brief description of the equipment and its operation is given here. Brown stock pulp, diluted with tap water to a consistency between 0.35 and 0.40%, is pumped from the no.l stock tank to a constant-head tank. The stock is then pumped from the head tank to the brown stock decker headbox at a rate of 6.0 odg/min (0.0096 adt/day). Excess stock in the head tank overflows back into stock tank no.l. The brown stock decker then thickens the stock to approximately 10% consistency with the aid of suction from the blower.  The pulp is discharged as a  coherent mat into the top of the downflow section of the bleaching reactor. Dilution filtrate, separated from the brown stock by the decker, is collected in the decker seal tank and discharged to the sewer. Upon entering the reactor, the unbleached pulp mat is disintegrated by a jet of heated dilution water composed of a mixture of distilled water and recycled chlorination filtrate. The pulp is then broken up further by the upper impellers of chlorination mixer no.l before reaching the top of the pulp slurry at a consistency of approximately 2.2%.  Solutions of chlorine and chlorine dioxide are then added  through the injection ports on the reactor.  These ports enable either simultaneous  addition or sequential addition with a delay of 40 or 60 seconds. The chlorine and chlorine dioxide are mixed with the pulp by the no.l mixer impellers located near the ports. The pulp, now at a consistency of 1.9%, then flows to the bottom portion of  56 the downflow section of the reactor under the influence of the lower impellers of mixer no.l. The reaction mixture then enters the horizontal section of the reactor controlled by chlorination mixer no.2. The temperature of the stock is measured in this section by a thermistor. The pulp then enters the upflow section regulated by chlorination mixer no.3. The stock spends most of the 24-minute residence time in this section as the inside diameter of the reactor increases from 2 to 4 inches. Once the pulp reaches the top of the upflow section, it is diluted to approximately 0.5% consistency with the injection of chlorination filtrate from the chlorination washer seal tank.  A second  thermistor measures the temperature of the stock at the reactor outlet. The chlorinated stock overflows through two weirs into the launder ring. This guides the stock into the dilution vat, where largerfloesare broken up by the dilution vat mixer. The dilution vat then discharges the stock onto the chlorination washer, where the stock is separated from the chlorination filtrate with the aid of suction generated by the blower. The resulting pulp sheet is washed with distilled water applied at a rate of 9.0 m /adt from the washer headbox located between the two press rolls of the 3  washer. The washed pulp is then discharged as a coherent mat at approximately 12% consistency. The chlorination filtrate is collected in the chlorination washer seal tank, where a third thermistor and an electrode measure temperature and pH. The filtrate is recycled to both the inlet and the outlet of the reactor. Excessfiltrateoverflows from the seal tank at a rate which is dependent on the level of recycle used for brown stock  57 dilution. At recycle rates of 4.5, 13.5, and 22.5 nrVadt the respective effluent flows are 42.9, 33.9, and 24.9 m /adt. 3  3.1.1. Industrial Comparison There are several important differences between chlorination as it is done industrially and as it is performed by the continuous lab-scale apparatus: 1.  Brown stock delivery.  At an industrial scale, unbleached pulp is stored at  approximately 12% consistency.  Prior to chlorination, this stock is partially  diluted and pumped to the chlorination stage using a centrifugal pump. At a flow of 6 odg/min, such technology is not feasible.  Thus the stock is stored  between 0.35 and 0.40% consistency at which it is pumped peristaltically at the required rate to the brown stock decker. The decker then thickens the stock to approximately 10% consistency before discharging it as a mat directly into the chlorination reactor. 2.  Chlorinated stock transport.  Industrial scale chlorination uses a centrifugal  pump with an open impeller to develop the head required for the pulp to flow through an up-flow chlorination reactor. Again, this technology is not feasible at a flow of 6 odg/min.  Thus the laboratory apparatus utilizes a series of  impellers in conjunction with a U-tube design reactor. The impellers, located throughout the reactor, keep the pulp from adhering to the reactor wall. This allows the pulp to move freely through the reactor in a plug flow manner due to the gravitational force acting on the pulp in the inlet side of the U tube.  58 Chlorine addition.  In industry, chlorine is usually added to the unbleached  stock as a dispersion of gas bubbles in water or recycled filtrate. For reasons of simplicity, a solution of chlorine gas in distilled water is used in the labscale apparatus. The use of dissolved chlorine is more efficient than gaseous chlorine because the solubilization process is already complete when the chemical is added to the pulp. Chemical mixing. Proper mixing is essential for effective chlorination. Earl and Reeve (1989b) found that poor mixing caused the production of certain chlorophenolic compounds from the first two bleaching stages to increase. Concurrently the amount of delignification decreased.  Both static and high-  shear mixers are commonly used in industry for mixing in the chlorination stage. Such technology is not feasible at a laboratory scale. Thus marine-type propellers are used for chemical mixing in the lab-scale apparatus.  The  fluidized zones around these impellers are indicative of good mixing. Chlorination consistency.  Because of the rapidity at which the chlorination  reactions occur, industrial chlorination is carried out at approximately 3% consistency. This level enables efficient mixing. Due to factors of scale, the lab-scale reactor operates at 1.9% consistency. While consistency affects the concentrations of substances dissolved in the filtrate, it is not believed to affect their production rates.  For instance, Earl and Reeve (1989a) found that the  production of A O X was unaltered when the chlorination consistency was increased from 4 to 10%.  59 The difference in consistency makes the comparison of filtrate recycle rates difficult.  Table 7 gives equivalent recycle rates for chlorination consistencies  of 1.9 and 3.0%.  In addition, the rates are expressed as percentages of the  maximum possible level of recycle.  The maximum level is the amount of  dilution required to bring 12% consistency brown stock to the consistency of chlorination.  Table 7 Equivalent levels of recycle 1.9% consistency  6.  3.0% consistency  (m7adt)  Recycle  Percentage of maximum recycle  (m7adt)  Recycle  Percentage of maximum recycle  4.5  11  2.8  12  13.5  34  8.4  37  22.5  56  14.1  63  Process control.  Due to the variability of the unbleached pulp entering an  industrial chlorination reactor, the use of automatic feedback control is vital in producing high-quality bleached pulp while minimizing chemical usage. Chlorination control is typically accomplished through color or residual chemical measurements taken within a few minutes of chemical addition. The residual after chlorination and the kappa number after extraction are used to supplement this data. scale apparatus.  Control as applied industrially is not used in the lab-  Instead, the control strategy relies on a constant flow of  60 brown stock at a constant kappa number.  In addition, both the flows and  concentrations of bleaching chemicals are required to be unvarying. Interstage connections. To conserve water and energy, industrial bleach plants recycle filtrates on a countercurrent basis for pulp washing and seal tank dilution.  Typical connections to the chlorination stage include the use of  filtrate from the first extraction stage as final wash water and the use of filtrate from the first chlorine dioxide stage as initial wash water and seal tank dilution.  Because the laboratory reactor is operated as an isolated bleaching  stage, these interstage connections and their effects on the chlorination stage are precluded. Thermal energy source. The average temperature of industrial chlorination of over 40°C is largely due to the use in the chlorination stage of filtrate that is recycled from the later stages of bleaching which are performed at approximately 70°C.  Because of the lack of interstage connections, the  laboratory chlorination stage uses a heat exchanger to warm both the recycled chlorination filtrate and the dilution water prior to entering the reactor. Pulp washing. Industrial bleach plants utilize rotary-drum washers to wash the pulp after each stage. Vacuum for these devices is usually generated by the action of the filtrate flowing down a sealed drop-leg.  The lab-scale  chlorination stage utilizes a horizontal-belt washer to wash the chlorinated pulp. Vacuum is supplied by a blower.  61  3.2. Experimental Design A  full 2  4  factorial design with three centerpoints was chosen for the  experimentation. Factor levels were selected to represent those achievable industrially and to allow for the possible formation of a central composite design with the inclusion of axial points. The four factors studied were: 1.  Recycle (R). The recycle of chlorinationfiltratefrom the chlorination seal tank to dilute the brown stock pulp was investigated at levels of 4.5, 13.5, and 22.5 m /adt. The level of recycle is the volumetric flow of recycled filtrate 3  rated on unbleached pulp throughput. 2.  Mode (m).  The mode of addition of chlorine and chlorine dioxide in the  chlorination stage was studied at levels of -40, 0, and 40 seconds. The level of mode corresponds to the difference in time between the addition of the two chemicals. Accordingly, a mode of -40 seconds, also referred to as DC mode, represents  the sequential  addition of chlorine dioxide before chlorine.  Simultaneous addition, sometimes known as D+C or C+D mode, is represented by a mode of 0 seconds. Lastly, a mode of 40 seconds, also referred to as CD mode, represents the sequential addition of chlorine before chlorine dioxide. 3.  Substitution (5).  The substitution of chlorine dioxide for chlorine in the first  bleaching stage was examined at levels of 20, 50, and 80%.  The level of  substitution is the percentage of total active chlorine added in the first stage that is supplied by chlorine dioxide.  62 4.  Multiple (M). The multiple was studied at levels of 0.18, 0.20, and 0.22 %/ml. The level of multiple represents the charge of active chlorine used in the chlorination stage expressed as the percentage on oven-dry unbleached pulp and divided by the kappa number of the unbleached pulp. A postrun analysis of deviations in peristaltic pump flows and bleaching  chemical concentrations enabled the variation in the levels of the four factors to be calculated (see appendix C, table 25).  This data is summarized in table 8 as 95%  confidence intervals for the levels of the four factors. Table 8 Summary of variation in factor levels Factor  95% confidence interval 8  recycle  ± 0.3 m /adt  mode  ± 2.7 seconds  substitution  ± 0.5 %  multiple  ± 0.003 %/ml  3  'Excludes data from run 0.  Parameters associated with the environmental impact of bleaching effluent were chosen as the primary response variables.  Thus the production in the chlorination  stage of chlorate and the productions in the first two stages of toxicity and AOX were measured.  Other response variables were indicative of pulp quality.  Thus kappa  number and viscosity were measured after the extraction stage (see section 3.3.7).  63 A total of 24 experimental bleaching runs, numbered 0-23, were performed. Table 9 gives the coded settings of the four factors for each run. comprised the 2* factorial design with three centerpoints.  Runs 5-23  Replicate runs, vital for  determining the reliability of the results, included runs 7, 12, and 18 representing the centerpoints. Replicated factorial points included runs 1 and 20, 2 and 22, 3 and 21, and 4 and 23. The factor levels and the standard operating conditions of the chlorination apparatus were not finalized until the completion of run 0. As a result, this run was performed at 17% chlorine dioxide substitution and at a consistency of 2.0%. The flow of effluent was thus 40.5 m /adt. 3  consistency.  Pulp flow difficulties were evident at this  This problem was alleviated by decreasing the consistency to 1.9%.  Another problem, which occurred during run 0, was the large decrease in the concentration of chlorine in the chlorine solution between the beginning and the end of the run (see appendix C, table 21).  This was solved by maintaining a large excess  volume of chlorine solution in the storage vessel.  This ensured that the loss of  chlorine due to vaporization was always small in comparison to the total amount of chlorine held in the vessel.  The same strategy was used with the chlorine dioxide  solution. A problem was encountered with the trout toxicity tests performed in four of the first five runs.  Inexplicably, mortality in these LC50 tests increased with  decreasing effluent concentration at the lower end of the concentration series used. A substantial effort was made without success to find the cause of this problem.  64  Table 9 C o d e d settings o f factor variables  Note:  Run  Recycle  Mode  Substitution  Multiple  0  -  +  -  -  1  -  +  -  -  2  -  +  +  +  3  -  -  -  -  4  -  +  -  +  5  +  -  -  -  6  +  -  -  +  7  0  0  0  0  8  +  +  -  -  9  +  -  +  +  10  -  -  +  -  11  -  +  +  -  12  0  0  0  0  13  +  +  +  -  14  -  -  -  +  15  -  -  +  +  16  +  +  +  +  17  +  -  +  -  18  0  0  0  0  19  +  +  -  +  20  -  +  -  -  21  -  -  -  -  22  -  +  +  +  23  -  +  -  +  +, 0, and - represent high, centerpoint, and low settings respectively.  65 Although such mortality patterns cannot be considered normal, at least one other study of bleaching effluent toxicity found similar behaviour (Betts and Wilson 1966). The LC50 test was thus replaced with the LT50 test starting with run 5.  Runs 1-4 were  replicated with runs 20-23 to enable the calculation of factorial effects using data obtained on an equivalent basis. The runs were performed in a randomized order according to run number with several exceptions. To minimize the duration of the experimentation, runs 10 and 11, 14 and 15, 20 and 21, and 22 and 23 were positioned so that each pair could be done in a single day.  In addition, runs 0-4 were performed first because they had low  levels of recycle. Thus the required effluent volumes were produced in the shortestpossible time. This enabled more time to be spent troubleshooting the problem which plagued the trout bioassays of these runs.  33. Materials and Methods 33.1. Brown Stock Pulp Supply An industrially produced softwood kraft pulp was used as the raw material for the 24 experimental bleaching runs. Roughly 600 kg of brown stock pulp at 12.5% consistency was sampled from the brown stock decker of the Celgar Pulp Company softwood kraft pulp mill at Castlegar, BC. Celgar Pulp, which began operations in 1961, produces fully bleached pulp at an approximate rate of 600 adt/day using standard kraft pulping practices and a five-stage (DC)(EO)DED bleach plant. The typical chip furnish at this mill is 35% hemlock, 30% fir and larch, 15% cedar, 10%  66 spruce and Balsam fir, and 10% Lodgepole pine.  Four sealable steel drums with  polyethylene liners were filled simultaneously in an attempt to dampen the effects of any process fluctuations. The pulp was received 4 days after sampling and was immediately stored at 4°C.  The filtrate associated with the pulp tended to collect in the bottom of each  drum. Thus, prior to use, the contents of each drum were unloaded into heat-sealable polyethylene bags.  Each bag, containing 5 kg of stock, was numbered according to  the position in the drum from which the pulp came. This system enabled the black liquor contained in the filtrate to be distributed equally between the bleaching runs. No degradation was revealed from visual inspections of the brown stock over the 9-month period between the sampling and the last bleaching run. A comparison was made between the levels of the response variables produced in the laboratory-scale experiments and those produced in an industrial-scale bleach plant.  Thus samples of effluent from the first two stages of bleaching were taken  from Celgar Pulp. The effluent samples, collected under stable operating conditions, were shipped refrigerated and were received the day after sampling. Sample treatment was identical to that used on the experimental effluents.  In addition, samples of  brown stock and CE pulp were taken for viscosity analysis. During effluent sampling, the 31 ml kappa number brown stock pulp entered the chlorination stage at a rate of 650 adt/day with 12 kg NajSO/adt as black liquor carryover. Chlorination was performed for 40 minutes at 43°C and 3.1% consistency with a multiple of 0.21 %/ml. Chlorine dioxide was added 30 seconds before the  67 chlorine at a level of 30% substitution. The recycle of chlorinationfiltratefor brown stock dilution was 7.2 m /adt. In addition, the residual chemical amounted to 2.3% of 3  the active chlorine added. The flow of chlorination filtrate at a pH of 1.5 was 26 m /adt. This represented 41% of the total bleach plant effluent flow. 3  Oxidative extraction was performed for 48 minutes at 74°C and 11% consistency.  The chemical application was 2.8% NaOH and 0.5% 0 on an oven-dry  pulp basis.  An extracted kappa number of 4.8 ml resulted. The flow of extraction  2  filtrate at a pH of 10.5 was 25 m /adt. This represented 39% of the total effluent flow 3  from thefive-stagesequence.  3.3.2. Peristaltic Pump Calibration Prior to each run, the peristaltic pumps were fitted with new tubing. The pumps were then operated near the setpoint flows for approximately 30 minutes to acclimatize the tubing to the various fluids being pumped.  Calibrations near the  setpoint flows were then carried out by measuring the time required for a known volume of fluid to be delivered. were checked.  Immediately following each run, the calibrations  All calibrations were performed in triplicate.  Setpoint flows and  postrun deviations are given in table 24 (see appendix C) for six of the seven peristaltic pumps used. A less-accurate calibration curve method was implemented for the pump used to dilute the chlorinated pulp at the reactor outlet.  This calibration,  applicable for flows from 800 to 1,000 ml/min, was defined by  7/ = 0.00777F+1.49  C) 7  68 where T is the number of turns of the vernier potentiometer and F is the flow in ml/min. The setpoint flow for this application was 850 ml/min.  33.3. Chlorine and Chlorine Dioxide Analysis Chlorine solutions were prepared by absorbing gaseous chlorine in distilled water. Chlorine dioxide solutions were prepared by passing gaseous chlorine through a bed of sodium chlorite and absorbing the product gas in distilled water.  The  concentrations of active bleaching agents in the chlorine and chlorine dioxide solutions were determined every 60 minutes during the experimental runs. Active chemical was titrated iodometrically with standardized 0.1 N sodium thiosulfate conditions to a starch endpoint. partially  performed  under  under acid  The analysis of the chlorine dioxide solutions was  neutral  conditions  contaminating chlorine to be determined.  enabling  the  concentration  of  This chlorine was not included in the  determination of the levels of either multiple or substitution.  The results of the  analyses done during effluent sampling are summarized in table 21 (see appendix C).  3.3.4. Sampling at Steady State Each run was started by ensuring a stable flow of pulp through the reactor without the addition of bleaching chemical.  Thus 45 minutes were allowed for the  brown stock pulp to form a coherent sheet on the chlorination washer. The addition of bleaching chemical was then begun, and 90 minutes were allowed for a steady state to be reached.  A dynamic model of the process indicated that this period would be  69 sufficient.  This was confirmed prior to the bleaching runs by following the response  of the process to a step change in chemical addition. Details of the model and its confirmation are given in appendix B. Once a steady state was reached, the effluent was collected in 25.6-1 polyethylene carboys.  Of the 24 experimental runs, 16  produced a single carboy of effluent each. Runs 1, 3, 4, 18, 19, and 23 each produced two carboys while runs 0 and 2 each produced three carboys. Temperature and pH data were recorded at 15-minute intervals throughout the entirety of each run. A summary of the data taken during steady state conditions is given in table 23 (see appendix C). The temperature in the horizontal section of the reactor averaged 33°C for the 24 runs with a range from 31 to 35°C. chlorination seal tank averaged 1.9 with a range from 1.6 to 2.2.  The pH in the  Feed pulp for the  alkaline extractions was sampled from the chlorination washer discharge at regular intervals during effluent collection. In addition, samples of pulp discharged from both the brown stock decker and the chlorination washer were analyzed immediately for consistency.  These results along with the stock tank consistencies are given in table  22 (see appendix C).  33.5.  Residual C h e m i c a l Determination and Reduction  The  extreme toxicity of both chlorine and chlorine dioxide as residual  bleaching chemical can mask the toxic action of both chlorophenolic and extractive compounds in bleaching effluent.  However, chlorine and chlorine dioxide are rarely  present in whole mill effluents because of their reactivity. Thus there is a consensus  70 that they should be removed from bleaching effluents prior to performing bioassays. Wong et al. (1978) did not neutralize the residual chemical and found the toxicity of various laboratory produced chlorinationfiltrateswas highly correlated with the level of residual. This made interpretation of the effects of several process modifications on bleaching effluent  toxicity  difficult.  Residual chemical is also important in  determining chemical consumption, which governs the extent of delignification. There are many analytical methods available for the determination of low levels  of  active  chlorine  in effluents  including potentiometry,  amperometry,  spectrophotometry, colorimetry, ion chromatography, and titrimetry.  Although ion  chromatography (Easty, Johnson, and Webb 1986) was the preferred method, the necessary equipment was not readily accessible. Thus an iodometric titration with a visual starch endpoint was utilized.  This method is used extensively in industrial  bleach plants. However, a problem with this method is the continual recurrence of the characteristic blue color due to the formation of iodine after the titration with thiosulfate is first complete.  Thomas, Becker, and Gibney (1975) documented this  phenomenon. Thus some preliminary research was performed to clarify the usefulness of this procedure. A synthetic effluent was made containing only residual bleaching chemical. The reaction of the residual in the starch-iodide system was immediate and nonrecurring. As a result, it is hypothesized that the recurring blue color is due to a slow reduction of dissolved organic compounds producing iodine.  Unsuspecting  analysts could titrate this response as if it were active bleaching chemical. Another potential problem with the iodometric titration is the effect of chlorate.  While the  71 reaction between chlorate and iodide is spontaneous as indicated by a negative free energy, nothing is implied about its rate. Thus a synthetic effluent containing only chlorate was tested with the titrimetric procedure. The rate of reaction was deemed slow enough not to interfere with the determination of residual chlorine and chlorine dioxide as 5 minutes were required for the blue color to appear. There are nearly as many methods to remove the residual as there are of quantifying it. Betts and Wilson (1966) added unbleached pulp to chlorination effluent in order to consume the residual chemical. The efficacy of this method is suspect because of the added formation of chlorinated organic compounds.  Nikki and  Korhonen (1983) removed residual chemical by sparging nitrogen through the bleaching effluent.  This method has the disadvantage that other volatile constituents,  such as chloroform, are also removed. The method used here involved the reduction of residual chemical to chloride with sodium thiosulfate, which was added in a carefully controlled manner to minimize the slower reductions of organic compounds and chlorate. Accordingly, the following method was used to determine and reduce the residual bleaching chemical in the chlorination effluents from the 24 experimental runs and Celgar Pulp. Upon completion of a run, the effluent was thoroughly mixed in the carboy used for sampling. A 100-ml aliquot of the effluent was then titrated with standardized 0.01 N sodium thiosulfate under acid conditions in the presence of starch and an excess of potassium iodide. The endpoint was deemed to be reached when the solution remained colorless for 10 seconds. Sodium thiosulfate was then added to the  72 carboy in an amount equal to 1.05 times the stoichiometric requirement. The contents were mixed again. A second aliquot was then tested to ensure the absence of residual chemical. The chemical residuals as determined by this procedure are given in table 21 (see appendix C).  33.6. Black Liquor Carryover Black liquor carryover affects the formation of both AOX and toxicity as well as the delignification in the first two stages of bleaching.  The design of the  laboratory-scale chlorination apparatus, which incorporated dilution and thickening of the brown stock pulp prior to chlorination, created low levels of carryover. This circumstance was ideal as black liquor carryover was not intended as a variable to be investigated.  The low carryover of black liquor was verified by measuring both the  sodium and extractives content of the brown stock pulp.  Sodium. Black liquor carryover, expressed as sodium sulfate, is indicative of the soluble lignin fragments transported with the brown stock pulp into the chlorination stage. Samples of the brown stock deckerfiltratewere taken for each of the 24 experimental runs and analyzed for sodium using a Perkin Elmer Model 306 atomic absorption spectrophotometer at a wavelength of 589 nm. Prior to analysis, the samples were stored at 4°C in 100-ml amber glass bottles with PTFE-lined caps. The black liquor carryover calculated from the soluble sodium averaged 0.16 kg NajSOyadt with a standard derivation of 0.01 kg Na^Oyadt.  The results were  73 corrected for the initial sodium content of the dilution water.  As expected, the  carryover was very low compared to industrial levels.  Extractives. Extractives form another component of black liquor carryover. To verify that the levels of extractives were low, samples of thickened brown stock were taken for runs 2, 8, and 17 from the brown stock decker and extracted with acetone (PAPRICAN 1986). The average extractives content was 1.5 kg/adt. Prior to analysis at the Vancouver Laboratory of PAPRICAN, the pulp samples were stored for a maximum of 6 months at 4°C in heat-sealed polyethylene bags. In addition, the extractives were derivatized (PAPRICAN 1989b) and quantified individually using gas chromatography (PAPRICAN 1989a).  The concentrations of  abietic, dehydroabietic, neoabietic, pimaric, isopimaric, levopimaric, sandaracopimaric, palustric, linoleic, and linolenic acids were thus determined in the brown stock entering the C stage. Further calculations gave the maximum-possible levels of these resin and fatty acids in the experimental bleaching effluents.  Comparisons between  these concentrations and the 96-hour LC50 values of the same compounds allowed toxic unit contributions to be calculated. The extractives accounted for at most 0.9 toxic units in the C stage effluents and 0.2 toxic units in the E stage effluents. Thus the carryover of extractives had relatively little effect on the toxicities of the experimental filtrates.  74 33.7.  Pulp  Extractions  The second stage of bleaching for the 24 experimental runs was carried out on a batch basis.  Extractions were performed at 70°C and 10% consistency for 90  minutes using a chemical charge of 3% NaOH on an oven-dry pulp basis.  Prior to  extraction, the chlorinated pulp was stored in heat-sealed polyethylene bags at 4°C after being sampled from the chlorination washer without further washing or residual chemical reduction. The maximum period of storage was 34 weeks (run 0) while the minimum was 3 weeks (run 23).  Typically 200 g of chlorinated pulp at the  chlorination washer discharge consistency was extracted producing 2.5 1 of fdtrate. The pH of the extraction effluents averaged 11.0 with a range of 10.6 to 11.2. The effluents were adjusted to pH 2.0 with concentrated HN0 before taking samples for 3  toxicity and AOX analyses. The extracted pulps were air dried at constant temperature and humidity prior to kappa number and viscosity tests.  The detailed extraction  procedure is given in appendix D. Thus a comparison can be made between the batch method, which is prevalent in laboratory bleaching work, and the continuous method, which is represented by the laboratory-scale equipment developed here.  33.8.  Toxicity  Separate toxicity tests were performed on both the chlorination and the extraction effluents at the Aquatic Toxicity Laboratory of Environment Canada in North Vancouver, BC. production.  All bioassays commenced within 24 hours of effluent  Prior to testing, the chlorination filtrates were stored at 4°C in sealed  75 25.6-1 polyethylene carboys after residual chemical reduction. The extraction fdtrates were stored at 4°C in 500-ml amber glass bottles with PTFE-lined caps after the pH was adjusted to 2.0.  These storage conditions have been shown to preserve the  toxicity of kraft pulp bleaching effluent (Guo, Bedford, and Jank 1976). All effluents were adjusted to a pH of 7.0 before testing. Caution was exercised in neutralizing the chlorination effluents so as to avoid highly alkaline conditions capable of abstracting chlorine from organochlorine compounds.  T r o u t tests.  The toxicity of the chlorination effluents towards rainbow trout  (Oncorhynchus mykiss) was measured according to standard methodology (British Columbia Ministry of Environment 1982). Static bioassays were used at a temperature of 14±2°C. Dilutions and controls were made with dechlorinated tap water from the City of North Vancouver, BC. Effluents from runs 0-4 were tested to determine the 96-hour LC50 using a test volume of 30 1. Fish data for these tests are given in table 26 (see appendix E). Median lethal concentrations were determined by plotting mortality versus log concentration followed by interpolation between the two largest concentrations straddling 50% mortality. Effluents from runs 5-23 and Celgar Pulp were tested at 100% concentration to determine the LT50 using a test volume of 25 1. Fish data for these tests are given in table 27 (see appendix E). Median lethal times were determined in most cases by the method of Litchfield (1949), which is a nomographical simplification of the method of Bliss (1937). Several cases not covered by the nomographs required the use of the original technique.  The only control  76 mortalities of this investigation occurred during the testing of run 5.  The three  mortalities were deemed not to have affected the LT50 result for two reasons. Firstly, the control mortalities occurred well after 100% mortality had occurred in the test effluent. Secondly, the cause of the control mortality was believed to be related to the use of stagnant control water.  Bacterial tests.  The toxicity of both the chlorination and the extraction  effluents towards luminescent marine bacteria (Photobacterium phosphoreum) were measured at 15°C using a Microtox Model 2055 analyzer. A slightly modified version (Western Canada Microtox Users Committee 1986) of the standard method (Microbics Corporation 1982) was used.  Version 5.0 of the Microtox Calculations Program  (Microbics Corporation, Carlsbad, CA) was used to calculate values for the 15-minute, color corrected EC50 (the median effective concentration causing a decrease in luminescence).  Confirmatory tests. Filtrate from the brown stock decker was collected during run 23 and analyzed for toxicity. At the inherent pH of 7.3, the sample was nontoxic to both trout and bacteria.  This gave additional evidence that the black liquor  carryover was low enough to have had no effect on the toxicity of the bleaching effluents produced in the laboratory. Because chlorate is known to be toxic to algae, its effect on Microtox bacteria was questioned. At 15°C, the 15-minute EC50 for sodium chlorate was determined to  77 be 33 g/1. Thus, at the levels found in the bleaching effluents, no interference was expected.  33.9. A O X  Total chlorinated organics were determined as A O X separately in both the chlorination and extraction stage filtrates. Analysis was performed by Econotech Services of New Westminster, BC, according to the microcolumn technique of method 506 of APHA (1985) using either a Dohrmann or a Mitsubishi AOX analyzer. C stage samples were taken from the chlorination effluent after residual bleaching chemical was reduced but before pH adjustment for bioassays. E stage samples were taken from the extraction effluent after the pH was adjusted to 2.0.  Samples were  stored refrigerated in 100-ml amber glass bottles with PTFE-lined caps.  33.10.  Chlorate  Chlorate was determined in the C stage effluents by ion chromatography in accordance with method T700 om-87 of TAPPI (1988). Analysis was performed by Quanta Trace Laboratories of Burnaby, BC, using a Dionex Autoion System 12 analyzer. Samples were taken from the chlorination effluent after residual bleaching chemical was reduced but before pH adjustment for bioassays.  Samples were stored  refrigerated in 100-ml amber glass bottles with PTFE-lined caps. Due to similarities in chemical structure, nitrate interferes in the ion chromatographic determination of chlorate. Although significant nitrate levels were  78 not expected in the effluent, this assumption was validated. Thus the nitrate content of the effluent from run 4 was determined using the cadmium reduction method in conjunction with a Hach DR/2000 spectrophotometer. Nitrate was not detected. Another concern with the production of chlorate was whether it originated from the reaction of chlorine dioxide with pulp or from the spontaneous decomposition of chlorine dioxide either prior to chemical addition in the chlorine dioxide solution or during chemical addition in the pulp slurry. Since chlorine was present in the chlorine dioxide solutions used (see appendix C, table 21), hypochlorite was available to catalyze the decomposition.  However, the same chlorine in the unbuffered chlorine  dioxide solutions would cause acidity thereby minimizing its progress.  This  assumption was verified when the chlorine dioxide solution used in run 23 was found to have a pH of 2.8. Also, the addition of recycled chlorination filtrate to the brown stock pulp before chemical addition ensured that any alkalinity from black liquor carryover was neutralized. Thus the pH of the stock during chemical addition was low enough to prevent chlorate formation via decomposition.  33.11. Kappa N u m b e r Kappa numbers of the brown stock and extracted pulps from the 24 experimental runs were determined in duplicate according to method T236 cm-85 of TAPPI (1988).  The kappa number is the number of milliliters of 0.1 N potassium  permanganate consumed by 1 gram of oven-dry pulp.  It is corrected to 50%  79 consumption and 25°C. The brown stock kappa numbers were determined on samples of the pulp in the stock tanks prior to each run.  33.12. Viscosity Viscosities of the brown stock and extracted pulps from the experimental runs and Celgar Pulp were determined according to method T230 om-82 of TAPPI (1988) using the closed bottle procedure. This method entails determining the viscosity of a 0.5% solution of pulp dissolved in 0.5 M cupriethylenediamine at 25°C using a capillary viscometer. The brown stock viscosity samples taken from the discharge of the brown stock decker and the Celgar pulp viscosity samples were stored in heatsealed polyethylene bags at 4°C prior to analysis.  CHAPTER 4 RESULTS AND DISCUSSION  4.1. Toxicity Chlorination effluent toxicities to rainbow trout are given as median lethal concentrations and median lethal times in tables 26 and 27 (see appendix E). The toxicities of both the chlorination and the extraction effluents to luminescent marine bacteria are given as median effective concentrations in table 28 (see appendix E). The corresponding toxic emission factors for the 24 experimental runs and Celgar Pulp are given in table 10. Considering the ranges and numbers of variables studied, the experimental toxic emission factors are in reasonable agreement with those in table 2 and from Celgar Pulp.  Appendix F gives some appropriate background  information on toxic emission factors including their calculation using median lethal time data. Chlorination stage toxic emission factors for trout were calculated for runs 0-4 using equation 17 in appendix F. Those for runs 5-23 and Celgar Pulp were calculated using equation 21 in appendix F. Toxic emission factors for bacteria were calculated using equation 17 in appendix F by substituting EC50 values for LC50 values. It should be noted that the decrease in pulp throughput due to delignification was not accounted for in the toxic emission factors of the extraction stage.  80  Table 10 Toxicity production in the first two bleaching stages Toxic emission factor (m /adt) 3  Run Chlorination filtrate  Extraction filtrate  Trout  Bacteria  Bacteria  0  319  355  122  1  364  282  140  2  57  373  95  3  482  258  150  4  352  340  171  5  329  290  102  6  394  488  167  7  ND*  278  174  8  191  346  171  9  238  252  94  10  299  279  181  11  609  557  153  12  275  265  148  13  130  198  157  14  914  300  151  15  851  178  98  16  204  147  94  17  132  165  94  18  325  234  127  19  431  389  173  20  346  320  186  21  658  442  149  22  295  670  130  23  585  412  169  Celgar  669  619  122  'Not determined.  82 Table 11 shows the correlations between the A O X concentrations and the number of toxic units for the 24 experimental runs. As expected, the correlations were all quite weak.  Celgar Pulp results were not included as their AOX concentrations  were substantially higher than those of the laboratory produced filtrates. As a result, they would have carried a disproportionate amount of significance in the analysis. Table 11 Correlation between the concentration of A O X and the number of toxic units Effluent tested  Bioassay organism  Correlation coefficient (r)  chlorination  trout  0.50  chlorination  bacteria  0.49  extraction  bacteria  0.40  a  'Run 7 excluded.  4.1.1. Trout Toxicity The factorial effects of the chlorination filtrate toxicity to trout are given in table 12. These were calculated using the toxic emission factors from runs 5-23. The pooled standard deviation from replicate runs was 120 m /adt with five degrees of 3  freedom. No distinct pattern in replicate results was noticed. Using 90% confidence intervals, all four single factor effects were found to be significant. In addition, the three factor interactions between multiple, mode, and substitution and between multiple, mode, and recycle were found to be significant indicating the complex nature  Table 12 Factorial analysis of chlorination effluent toxicity to trout Effect on toxic emission factor (m /adt)  90% confidence interval (m /adt)  recycle (R)  -314  ±121  multiple (M)  152  ±121  substitution (S)  -136  ±121  MSm  -132  ±121  mode (m)  -128  ±121  MRm  128  ±121  curvature  -113  ±181  Rm  94  ±121  Mm  -93  ±121  MRSm  80  ±121  Sm  58  ±121  MS  -48  ±121  RSm  -41  ±121  MR  -31  ±121  RS  -24  ±121  MRS  17  ±121  Factor  3  3  84 of the system. This complexity should be kept in mind when interpreting the effects of individual factors. Undoubtedly the most significant finding of this research was that the recycle of chlorination filtrate had the largest effect on toxicity. Increasing the recycle from 4.5 to 22.5 m /adt caused the C stage toxic emission factor to decrease on average by 3  314 m /adt. This result coincides well with the previous finding that increasing the 3  level of filtrate reuse reduces the formation of chlorophenolic compounds (Nikki and Korhonen 1983). It is clear that recycling filtrate in the first stage enables toxicants to be degraded by the action of fresh bleaching chemical. studied this indirectly.  Two investigations have  Strumila and Rapson (1979) found that chlorine dioxide  substantially degraded several chloroguaiacols. Nikki and Korhonen (1983) found that the  actions  of both chlorine and chlorine dioxide significantly  reduced the  concentration of chlorophenolics in chlorination and extraction effluents. A factor complicating the effect of recycle on toxicity was the amount of time the collected effluent spent in the presence of residual bleaching chemical. Since the flow of effluent was dependent on the level of recycle, runs at high recycle required more time to fill a carboy than runs at low recycle. In addition, it was not feasible to reduce the residual continuously.  Thus the average time that the collected effluent  spent in the presence of residual was 45 minutes at a recycle of 4.5 m /adt and 77 3  minutes at 22.5 m /adt. Accordingly, it could be argued that the effect of recycle was 3  caused by the effluent from runs at high recycle being in contact with residual chemical for longer periods of time than the effluent from runs at low recycle. The  85 difference in time would enable more degradation of toxic chlorophenolics with residual chlorine and chlorine dioxide to occur in runs at high recycle. In turn, this would cause the same runs to be the least toxic. However, there is considerable evidence to refute this argument. Firstly, some relevant information is provided from replicate runs 12 and 18. Since two carboys of effluent were produced in run 18 compared to only one carboy in run 12, the effluent from run 18 spent on average 113 minutes in the presence of residual while that from run 12 spent only 57 minutes. This time difference was of similar magnitude as the difference between runs at high and low recycle. Despite the extra time available for reaction, the effluent from run 18 had a toxic emission factor that was 50 m /adt 3  greater than that of the effluent from run 12. Secondly, there was no correlation for the 24 experimental runs between the level of recycle and the chemical residual expressed as a percentage of the total active chlorine added (r = -0.16). Thus it cannot be argued that because of the added time for reaction runs at high recycle had lower residuals than runs at low recycle.  Thirdly, because chlorophenolic compounds are  nearly always found in the presence of residual chemical in chlorination filtrate, the belief is supported that at the concentrations involved the two components are relatively nonreactive.  Thus it is unlikely that the effect of recycle on chlorination  stage toxicity was caused by differences in the time delay between effluent collection and residual reduction. There are several important ramifications of recycle having a large effect on the toxicity of chlorination filtrate. Firstly, industrial bleach plants, presently operating  86 with minimal levels of chlorinationfiltraterecycle, could reduce toxic discharges by increasing the level of recycle.  Currently the two primary uses of recycled C stage  filtrate are for brown stock dilution and for chlorine addition.  The feasibility of  utilizing chlorination filtrate as the absorption medium for gaseous chlorine dioxide should be investigated as a third use.  Secondly, it is clear that toxicity data from  chlorination effluents produced in the laboratory without recycle are not representative of the bleaching process as it is typically carried out industrially. Lastly, the model used by Wong et al. (1978) to predict toxicity at steady-state conditions of effluent reuse for bleached pulp dilution and washing is not extendable to the case of effluent reuse for brown stock dilution.  This model, based on a mass balance without  consideration for chemical reaction, was used to calculate the steady-state toxicity from the toxicity measured after two cycles of reuse on a batch basis. The model was never proven for toxicity although it was shown to be appropriate for measurements of both COD and color. Toxicity cannot be viewed as an inert substance that simply builds up to an equilibrium asfiltrateis recycled. The effect of multiple was also found to be statistically significant.  By  increasing the multiple from 0.18 to 0.22 %/ml, the chlorination stage toxic emission factor to trout increased on average by 152 m /adt. 3  Curvature was not significant.  Thus a maximum in toxicity with increasing multiple as was found by Voss, Wearing, and Wong (1981) did not occur. It is likely that the range of multiple studied was too narrow to encompass a maximum. The effect of multiple on toxicity can be explained in terms of delignification and the concurrent formation of chlorophenolics. As the  87  multiple is increased, more delignification occurs producing more chlorophenolics. This in turn causes increased toxicity. However, delignification can only be extended until readily accessible lignin becomes scarce.  At this point, excess bleaching  chemical builds up and destroys chlorophenolic compounds in the same manner as occurs with filtrate recycle. decreased toxicity.  The decreased concentration of chlorophenolics causes  Although overchlorination is discouraged from an economic  standpoint, it benefits effluent toxicity. The substitution of chlorine dioxide for chlorine in the first bleaching stage was found to affect toxicity.  The toxic emission factor decreased on average by 136  m /adt as the chlorine dioxide substitution increased from 20 to 80%. This result is in 3  agreement with the majority of the literature. Chlorine dioxide, which reacts primarily through oxidation as opposed to substitution, produces less low molecular weight chlorinated organics compared to an equivalent amount of chlorine thereby reducing toxicity. Considering the sizeable effort of the pulp and paper industry to implement substantial levels of chlorine dioxide substitution, partly to meet AOX requirements, the return in terms of decreased effluent toxicity is modest. A much greater reduction in toxicity could be made by encouraging mills to increase the recycle of chlorination filtrate. A recent survey of bleach plant operating practices in Canada found that only 11 of 19 mills utilized chlorination filtrate for brown stock dilution (Byzyna, Clifford, and Billmark 1989).  Certainly it is much less costly to implement chlorination  effluent recycle than substantial chlorine dioxide substitution.  88 Lastly, the mode of addition was also found to affect the toxicity of chlorination effluent.  The use of CD mode decreased the toxic emission factor by  128 m /adt on average over DC mode. This agrees with the finding of Liebergott et 3  al. (1990a). Although interaction between mode and substitution was not found to be significant, the three factor interaction between multiple, mode, and substitution might include this effect. The effect of mode on toxicity is probably related to its effects on delignification and the corresponding formation of chlorophenolic compounds. Thus CD mode produces the least amount of delignification and the lowest toxicity. Reasons for the effect of mode on delignification are discussed in section 4.4. Only three of the experimental filtrates had a toxic emission factor larger than that of the chlorination effluent from Celgar Pulp.  Undoubtedly the difference in  black liquor carryover between Celgar Pulp (12 kg NajSO/adt) and the 24 experimental runs (0.16 kg NajSO^adt) was partly responsible for this. The limited use of chlorination filtrate recycle at 7.2 m /adt for brown stock dilution was also a 3  likely cause of the effluent's high toxicity. This level of recycle represents only 33% of the total requirement to dilute the pulp from 12% consistency in storage to 3.1% consistency in the chlorination reactor.  The equivalent level of recycle at 1.9%  consistency is 11.8 m /adt, which represents 30% of the maximum-possible level. 3  Recycle is kept low due to the use of a blend tank prior to chlorination which is open to the mill's atmosphere. It is feared that additional use of C stage filtrate in this tank would cause the workplace to become contaminated with residual chemical vapor. Some inexpensive modifications to this system would enable a larger portion of the  89 chlorination filtrate to be safely recycled.  This in turn would decrease the toxic  emission factor. In addition, the use of DC mode at only 30% substitution contributed to the magnitude of Celgar's chlorination effluent toxicity.  4.1.2. Bacterial Toxicity It is evident from table 10 that the extractionfiltrateswere much less toxic to luminescent bacteria than the corresponding chlorination filtrates. While several of the factorial effects of the extraction effluents were found to be statistically significant, the magnitudes of these effects were much smaller than those of the significant chlorination effluent effects. For this reason, under the conditions which prevailed, a combined effluent toxicity would likely have followed the effects which emanated from the C stage. This ignores the detoxification which can occur when C and E stage effluents are mixed, primarily due to the abstraction of chlorine atoms from chlorinated organic compounds.  No pattern was evident from the toxic emission  factors of the extractionfiltratesfrom replicate runs. Thus there was no effect due to the time delay between chlorination and extraction, which varied directly with run order. Celgar Pulp's extraction stage toxic emission factor was very similar to those from the experimental runs.  Since the industrial extraction filtrate was produced  without any delay between chlorination and extraction, the claim that the pulp storage time had no influence on the experimental toxicities is supported.  Celgar Pulp's E  stage toxic emission factor was also much lower than from its C stage. This might  90 suggest that the mill's black liquor carryover was comprised mostly of dissolved lignin fragments as opposed to extractives. The factorial effects of the chlorination filtrate toxicity to bacteria are given in table 13. The toxic emission factors from runs 5-23 formed the basis for the analysis. Results from replicate runs had a pooled standard deviation of 104 m /adt with six 3  degrees of freedom and formed no distinct pattern. determined with the use of 90% confidence intervals.  Significant effects were The only significant single  factor effect was recycle, which decreased the toxic emission factor on average by 111 m /adt when increased from 4.5 to 22.5 m /adt. The interactions between recycle and 3  3  substitution; recycle and mode; and recycle, substitution, and mode were also found to be significant.  Thus substitution and mode were shown to be dependent factors.  Multiple was not found to be a significant factor. It is clear by comparing the chlorination filtrate factorial effects for the two bioassay organisms that the responses were not alike, aside from the agreement on the effect of recycle. Not surprisingly a logarithmic plot of toxic emission factors (trout versus bacteria) gave a correlation coefficient of 0.30. This analysis included the data from Celgar Pulp but excluded that from run 7. However, by dropping those points for which the ratio of the two toxic emission factors was greater than 2:1 or less than 1:2, the correlation coefficient increased to 0.80. By this analysis, the data from runs 2, 14, 15, and 22 were dropped. Thus, in general, the Microtox bioassay was able to predict the toxicity to rainbow trout relatively well (20 of 24 cases). However, there were several instances where the agreement was very poor (4 of 24 cases).  Table 13 Factorial analysis of chlorination effluent toxicity to bacteria Effect on toxic emission factor (m /adt)  90% confidence interval  RS  -120  ±101  recycle (R)  -111  ±101  Rm  -110  ±101  RSm  -101  ±101  Sm  94  ±101  MRm  -93  ±101  mode (m)  81  ±101  substitution (S)  -68  ±101  MR  39  ±101  MRS  -33  ±101  multiple (M)  30  ±101  Mm  20  ±101  MS  -18  ±101  curvature  8  ±127  MRSm  5  ±101  MSm  -1  ±101  Factor  3  (m7adt)  92 Somewhat contrary to this result is the work of Firth and Backman (1990). Both rainbow trout and Microtox bioassays were performed on 34 effluent samples from a bleached kraft pulp mill. A correlation coefficient of 0.91 was found between the 96-hour LC50 and the 15-minute EC50.  However, some of the individual data  pairs were not in good agreement. For example, one sample had an EC50 of 7% and an LC50 of 35%, and another sample had an EC50 of 18% and an LC50 of 67%. Considering the degree of correlation between the two bioassays, it was not explained why low molecular weight A O X measurements correlated well with LC50 values (r = 0.94) but not with EC50 values (r = 0.19). The study was not confined to bleach plant effluent and EC50 values were not corrected for color.  The absorption of a  portion of the bioluminescence in the Microtox bioassay due to effluent color causes a false increase in toxicity. Bioassay organisms should not be expected to behave identically. SalkinojaSalonen et al. (1981) determined the toxicities of several chlorophenolic compounds to both fish (Poecilia reticulata) and bacteria (Pseudomonas putida). No correlation was found between the responses of the two organisms. Sameshima, Simson, and Dence (1979) found that chlorination effluent actually stimulated the growth of a fungus (Aspergillus fumigatus) while being toxic to Daphnia magna. Since there is a significant chance for any single effluent tested that the Microtox bioassay will not closely agree with the trout bioassay, Microtox data should be regarded as being supplemental to trout data.  This has added significance  considering that nearly all regulatory standards for toxic discharges are based on trout  93 as the test organism. Thus the use of the chlorination apparatus developed here is important not only becausefiltraterecycle can be performed on a continuous basis but also because large effluent volumes can be generated, enabling the use of trout bioassays.  4.2. A O X The AOX data for the effluents from the 24 experimental runs and Celgar Pulp is given in table 14. The data are presented asfiltrateAOX concentration and AOX production based on unbleached pulp throughput. The total production is given as the arithmetic sum of the productions from the C and E stages. It should be noted that the decrease in pulp throughput due to delignification was not accounted for in the production figures of the extraction stage. The pooled standard deviation for AOX production, determined from replicate runs, was 0.20 kg/adt with six degrees of freedom. In addition, there was no discernible trend in E stage results from replicate runs. Thus it can be inferred that the time delay between chlorination and extraction, which varied directly with run order up to a maximum of 8 months, had no effect on AOX levels.  Fraser and Reeve (1989) found a similar stability of A O X in  prestabilized effluent samples, which were stored properly for up to 3 months. The average storage time between effluent sampling and analysis was 11 days for both C and E filtrates. According to the results of Sjostrom, Radestrom, and Lindstrom (1982), the chlorinated organic content of the filtrates would require less than 1 week of storage to become stabilized.  Table 14 A O X formation in the first two bleaching stages  Run  Chlorination filtrate  Extraction Citrate  Total AOX production (kg/adt)  AOX concentration (mg/1)  AOX production (kg/adt)  AOX concentration (mg/D  AOX production (kg/adt)  0  43  1.74  26  2.46  4.20  1  39  1.67  23  2.17  3.84  2  19  0.82  8.0  0.76  1.58  ,3  49  2.11  23  2.17  4.28  4  50  2.14  23  2.18  4.32  5  70  1.74  26  2.45  4.19  6  87  2.17  24  2.27  4.44  7  48  1.63  18  1.69  3.32  8  42  1.04  22  2.07  3.11  9  36  0.90  11  1.04  1.94  10  21  0.90  8.6  0.81  1.71  11  17  0.73  5.5  0.52  1.25  12  36  1.22  18  1.70  2.92  13  28  0.70  7.3  0.68  1.38  14  51  2.19  23  2.17  4.36  15  25  1.07  8.6  0.81  1.88  16  29  0.72  9.7  0.91  1.63  17  37  0.92  10  0.94  1.86  18  35  1.19  18  1.69  2.88  19  98  2.44  28  2.63  5.07  20  30  1.29  23  2.17  3.46  21  43  1.84  24  2.26  4.10  22  16  0.68  7.3  0.69  1.37  23  53  2.28  22  2.07  4.35  Celgar  149  4.06  69  1.72  5.78  95 In accordance with the work of others, a relationship between the total elemental chlorine consumed and the AOX produced was sought. Thus the value of k in equation 4 (see section 2.2) was determined to be 0.096 kg AOX/kg TEC for the data of runs 5-23. This regression, which gave a coefficient of determination (R ) of 2  0.92, utilized the total elemental chlorine added as residuals were low (see appendix C, table 21). As would be expected, the value of k representing a mixture of CD, DC, and C+D experiments lies between the values representing separate CD and DC experiments as determined by Histed and Vega Canovas (1989). This analysis was improved upon by regressing the data comprising the factorial design to give AOX = 0.96 + 19.4M - 0.199M5 -0.00595/n  (> 8  which defines the AOX produced from the first two bleaching stages in kg/adt. Equation 8 gave a coefficient of determination (R ) of 0.97. 2  The resultant residual  standard deviation of 0.21 kg/adt, having 14 degrees of freedom, compares well with the replicate standard derivation. Run 19 was deemed to be an outlier using internally studentized residuals and thus was not included in the analysis.  This trial had a  combination of decreasing brown stock flow and increasing chemical flows (see appendix C, table 24) that caused the multiple initially at 0.22 %/ml to increase to 0.235 %/ml by the end of the run. Thus the measured AOX of 5.07 kg/adt was much higher than expected. Otherwise, a normal plot of the raw residuals was not unusual, and the internally studentized residuals were not a function of the fitted values.  96 Figure 13 shows equation 8 as a plot of AOX production versus chlorine dioxide substitution at three combinations of multiple and mode.  Several 95%  confidence intervals are also shown for the case where chlorine dioxide is added 30 seconds before chlorine at a multiple of 0.20 %/ml. These were developed from  S.E.EST. = 0.214^-+ (M-0.199)*+ (MS-10.3) + (m+lllA* 2  [IS  0.00598  565  which defines the standard error of estimation in kg/adt.  23,900 J  (9)  Figure 13 also shows  equation 4 at the same conditions for a pulp having a kappa number equal to the average value of the brown stock used in the 24 experimental runs (29.2 ml). There is good agreement between the two results. Equation 8 has a significant interaction between multiple and substitution. This is shown graphically in figure 13 for the two cases with multiples of 0.22 and 0.20 %/ml and a mode of -30 seconds. It is readily apparent from the respective lines being nonparallel that the effect of multiple is dependent on the level of substitution. However, this interaction is not unexpected as equation 4 transformed from chemical consumptions to the variables of multiple and substitution becomes  AOX = Kxk\\QM-—MS^  25  (10)  where K is the unbleached pulp kappa number in ml. Thus the interaction term, which went unmentioned in previous work, is revealed. The inclusion of mode in equation 8 makes its effect on AOX production easy to assess. Figure 13 shows that the use of CD mode at an interval of 30 seconds  97  eq 8: multiple = 0.22 %/ml; mode = -30 sec eq 8: multiple = 0.20 %/ml; mode = -30 sec eq 8: multiple = 0.20 %/ml; mode = 30 sec eq 4: multiple = 0.20 %/ml; mode = -30 sec  I  l  l  l  I  I  l  l  l  l  I  0  10  20  30  40  50  60  70  80  90  100  Percent chlorine dioxide substitution F i g u r e 13.  The effect of chlorine dioxide substitution on AOX production j  98 produces nearly 0.4 kg/adt less AOX than the use of DC mode at the same interval. This finding is in close agreement with the literature. However, equation 8 describes a continuum of AOX with respect to the time delay between chemical additions. Thus simultaneous addition was not found to be equivalent to CD addition as Liebergott et al. (1990a) found. Since simultaneous addition delignifies less than DC addition but more than CD addition, it is logical to believe that AOX formation would follow a similar trend.  Undoubtedly there exists a certain time delay between chemical  additions beyond which an increase in the effect on AOX production is not seen. This has not yet been demarcated. While it is obviously important to know what factors influence the production of chlorinated organics, it is also helpful to know what factors are not significant. The recycle of chlorination filtrate had no effect on the formation of AOX as a whole. This result contradicts the somewhat dubious finding of Mattinen and Wartiovaara (1981), which was discussed previously.  Because recycle affects the production of  chlorophenolic compounds (Nikki and Korhonen 1983), it could be argued that it would also affect the production of AOX. However, because these compounds make up such a small percentage of AOX production, this effect was not noticed. Equation 8 substantially underestimates the A O X produced in the C and E stages of Celgar Pulp. The AOX measured was 5.8 kg/adt whereas that predicted by equation 8, which was applied at a multiple of 0.22 %/ml with 30% substitution that was added 30 seconds before the chlorine, was 4.1 kg/adt. Problems with the use of equation 3 (see section 2.2) to predict the production of AOX from industrial-scale  99 operations have also been reported (Hall et al. 1988).  Possible reasons for the  discrepancy of the Celgar mill result include: 1.  Chlorination of black liquor carryover.  The effect of this variable on AOX  production was negligible in the 24 laboratory trials as the carryover averaged 0.16 kg NajSO/adt.  However, the brown stock bleached at Celgar Pulp  contained 12 kg Na^O/adt. Utilizing the result of Johannes and Lowe (1989), this corresponds to an excess AOX production of 2.1 kg/adt. Most of the AOX generated from black liquor carryover would be discharged in the chlorination filtrate. Accordingly the C stage accounted for an average of 47% of the total AOX produced in the 24 experimental runs while in the Celgar Pulp case it accounted for 70%.  Contrary to this is the previously discussed evidence that  active chlorine responds in the same fashion to black liquor as it does to lignin. Thus the chlorination of black liquor carryover should be accounted for in the multiple.  Since Celgar's extracted kappa number of 4.8 ml was normal, it  cannot be concluded that delignification was sacrificed at the expense of chlorinating black liquor.  In summary, the multiple used at Celgar Pulp  appears to be much too low when both delignification and black liquor carryover are considered in relation to the AOX produced. 2.  Chlorination of extraction filtrate. This was eliminated in the laboratory work by performing chlorinations and extractions separately. However, 63% of the wash water used in Celgar's chlorination stage is filtrate from the second extraction stage.  Penetration of this filtrate through the pulp sheet into the  100 chlorination seal  tank would cause an increase  in A O X production.  Concomitantly an increased multiple would be required to reach a given level of delignification because of the consumption of active chlorine by the extraction filtrate. Again this does not appear to have occurred. 3.  Countercurrent flow of AOX from the DED sequence.  Celgar Pulp applies  chlorine dioxide at an approximate level of 14 kg/adt to the pulp in the final three bleaching stages. When applied at this consumption, equation 4 gives an AOX yield of 0.7 kg/adt.  Since 70% of the filtrate from the DED sequence  eventually enters the first two stages, an additional A O X production of 0.5 kg/adt would be expected beyond that predicted by equation 8.  Thus the  predicted total of 4.6 kg/adt is still well below the 5.8 kg/adt that was measured. It should be noted that the use of equation 4 in such a manner is somewhat suspect. Histed and Vega Canovas (1989) measured the AOX from the final three stages of a (DC)(EO)(D/E/D) sequence to be 0.1 kg/adt. Yet when equation 4 is applied at the relevant chlorine dioxide consumption of 7.7 kg/adt, an AOX value of 0.4 kg/adt is given. Some discussion on the application of equation 8 is warranted.  Firstly,  although the multiple applied to the pulp was used in the regression analysis because chemical residuals were low (see appendix C, table 21), it is the multiple consumed which is meant to be utilized. Secondly, since the actual amount of chemical applied is dependent not only on the multiple but also on the kappa number of the unbleached pulp, equation 8 can only be used rigorously at a kappa number of 29.2 ml, the  101 average value of the brown stock used in the 24 runs. A less-rigid approach enables its use for any standard kraft pulp having a kappa number near this average. In such instances, the actual multiple consumed should be prorated to a 29.2 ml kappa number. For example, in applying equation 8 to the Celgar Pulp case, a multiple of 0.22 %/ml was used. However, the actual multiple consumed was 0.21 %/ml at a kappa number of 31 ml. Lastly, the proper use of a regression equation is limited to the ranges of the variables investigated in its formulation.  Thus the constant of 0.96 kg/adt in  equation 8 is meaningless at zero chemical consumption.  4.3, Chlorate The chlorate data from the 24 experimental runs and from Celgar Pulp is given in table 15. The data is presented as effluent concentration, yield based on chlorine dioxide added, and production rated on unbleached pulp throughput.  The pooled  standard deviation for chlorate production, determined from replicate runs, was 0.54 kg/adt with six degrees of freedom. In each case, the production decreased for the second and third replicates when compared with the first. A logical explanation for this occurrence could not be found. However, no discernable pattern was revealed when chlorate production was plotted against the time order of observation for all 24 runs. The production data from runs 5-23 was regressed to give  log C/O; = -1.23 +6.16xl(T S-4.90xl0" 5 +9.26xl0' /n- l^xlO^S/n (11) 2  4  2  3  which defines the chlorate produced in the first bleaching stage in kg/adt. Equation 11  Table 15 Chlorate formation in the first bleaching stage  Run  Chlorate concentration (mg/1)  Chlorate yield (mole % of C10 added)  Chlorate production (kg/adt)  2  0  50  56  2.03  1  45  44  1.94  2  67  13  2.87  3  15  15  0.65  4  55  44  2.36  5  15  8  038  6  16  7  0.40  7  143  40  4.85  8  37  20  0.92  9  165  19  4.11  10  88  21  3.77  11  80  19  3.43  12  120  33  4.07  13  136  19  338  14  10  8  0.43  15  122  24  5.24  16  140  15  3.48  17  144  20  3.58  18  109  29  3.70  19  56  26  1.40  20  22  21  0.95  21  5  5  0.22  22  57  11  2.45  23  35  27  150  Celgar  45  36  1.23  103 gave a coefficient of determination (R ) of 0.96. A residual standard deviation of 0.49 2  kg/adt with 14 degrees of freedom resulted.  This compares well with the standard  deviation determined from replicates. No outliers were detected and a normal plot of the raw residuals showed nothing unusual. In addition, internally studentized residuals were not a function of the fitted data. Initially the untransformed production data was regressed with the same variables appearing in the RHS of equation 11. However, the residuals from this analysis increased distinctly as the fitted values increased. Thus the constant variance assumption of regression analysis was violated. The use of the log transformation proved to be a satisfactory solution.  With the highest chlorate  concentration (run 9) being 33 times larger than the lowest (run 21), it is not unreasonable to attribute the changing variance to the analytical method used, ion chromatography. Figure 14 shows equation 11 as a plot of chlorate production versus chlorine dioxide substitution for the three modes of chemical addition. Several 95% confidence intervals are shown for the curve representing DC addition. These were developed from the standard error of estimation in kg/adt given by  J_ (S-50) (S-3,260) m Sm V* 19 14,400 1.46x10 25,600 8.70xl0,  S.E.EST. =0.492  2  (  |  2  2  2  t  s  2  |  +  (12)  7  The use of equation 11 at 30% chlorine dioxide substitution and a mode of -30 seconds (DC) predicts the chlorate production from Celgar Pulp's chlorination stage to be 1.05 kg/adt. This compares well with the measured value of 1.23 kg/adt. Celgar's chlorate yield of 36 mole % of the chlorine dioxide added demonstrates the need to  105 decrease chlorate discharges through bleaching efficiency improvements as opposed to pollution control methods. A third method of decreasing chlorate discharges might be through the regeneration of chlorine dioxide from the chlorate in the effluent.  The  chlorate could be separated from the effluent matrix and returned to the chlorine dioxide generator. Alternatively some form of chlorine dioxide regeneration might be feasible within the bleaching process. The low concentrations involved as shown in table 15 probably render such approaches uneconomic. Chlorate yields as given in table 15 ranged from 5 to 56 mole % of the chlorine dioxide added. These values compare well with the data of Lindgren and Nilsson (1975). However, according to the two path mechanism of chlorate formation (Bergnor et al. 1987), the maximum possible yield of chlorate is 50%. This would occur if all of the chlorine dioxide were to react via path B. With a yield of 56%, run 0 was the only trial that was inconsistent with this theoretical maximum.  This  erroneous result was undoubtedly caused by the substantial reduction in chlorine concentration (see appendix C, table 21), which occurred during the course of this run. Thus the chlorine dioxide substitution increased from 17% at the start of the run to 42% at the end. Maximum chlorate productions near 5 kg/adt as shown in figure 14 are in good agreement with the data of Bergnor et al. (1987) and Liebergott et al. (1990a). The maxima occurred between 57 and 68% chlorine dioxide substitution. Literature values are slightly higher.  106 The effect of mode on chlorate production is readily seen in figure 14. Below 69% substitution, DC addition gave less chlorate than CD addition. substitution, CD mode gave less chlorate than DC mode. simultaneous  Above 69%  Chlorate production for  addition was between the two sequential modes except at 69%  substitution where all three modes had the same production. This work agrees with the literature in that DC addition gives the lowest chlorate production at low levels of substitution.  At high substitution, Bergnor et al. (1987) and Nilsson and Sjostrom  (1974) agree with this study that CD mode gives the least amount of chlorate. Beyond this, there is little agreement on the effect of mode on the formation of chlorate. The recycle of chlorination filtrate had no effect on the production of chlorate. This is in agreement with the finding of Reeve, Anderson, and Rapson (1980). Another variable that was found not to affect the formation of chlorate was multiple over the range investigated of 0.18 to 0.22 %/ml. However, Bergnor et al. (1987) found that chlorate formation was increased dramatically by increasing the multiple from 0.17 to 0.23 %/ml at 70% substitution. This increase was not noticeable at 30% substitution suggesting strong interaction between substitution and multiple. Contrarily this interaction was not found to be significant in this study. A possible explanation for these discrepancies is the existence of a critical multiple beyond which a much larger portion of the chlorine dioxide reacts via path B thus producing more chlorate than usual for a given increase in multiple. The data of Bergnor et al. (1987) tends to support such a theory as the curve of chlorate production versus multiple had a substantially increasing slope with increasing multiple. Differences in pulp type then  107 may have meant that the critical multiple was exceeded in the work of Bergnor et al. but not in this study. Any increase in chlorate as a function of multiple in this work was probably masked by the 0.54 kg/adt standard deviation. Accordingly, the use of equation 11 outside of the ranges of the variables investigated or for pulps having kappa numbers significantly different from 29.2 ml is strongly discouraged. Chlorine contamination of the chlorine dioxide used industrially is common. The chlorine dioxide solution used in run 10 contained 5.50 gjl chlorine dioxide with 0. 48 g/1 chlorine.  This contaminating chlorine, which represented 3.2% of the  available chlorine in the chlorine dioxide solution, was at the highest level of all the runs (see appendix C, table 21).  Considering that Bergnor et al. (1987) needed 10%  available chlorine contamination to cause a 0.5 kg/t effect on chlorate production, it is unlikely that the results of this study were influenced by the presence of chlorine in the chlorine dioxide. With reference to this study and the literature, there is a consensus on three qualitative characteristics of chlorate formation in the first bleaching stage: 1.  Chlorate production has a relative maxima with respect to chlorine dioxide substitution.  2.  At low levels of substitution, DC addition gives the least amount of chlorate.  3.  At high levels of substitution, CD addition gives the least amount of chlorate.  Little thought has been given to date in the literature regarding the reasons for these traits.  An understanding of these phenomena is dependent on knowledge of the  reactions of both chlorine and chlorine dioxide with lignin.  108 The extensively studied reaction between chlorine and lignin involves three steps (Rydholm 1965, 929): 1.  Chlorine is substituted for hydrogen at the 5- or 6-position of the benzene nucleus.  2.  The methoxyl group at the 3-position is instantly hydrolysed to form a hydroxyl group.  3.  The resulting phenolic hydroxyl group is oxidized by chlorine to form a cyclic ketone group. Any phenolic hydroxyl group in existence prior to the first two steps can be oxidized identically. As previously mentioned, the reaction of chlorine dioxide with lignin involves  two pathways with respect to chlorate formation (Bergnor et al. 1987): 1.  Path A involves the oxidation of phenolic hydroxyl groups to cyclic ketones. This path is rapid and results in the least amount of chlorate formation.  2.  Path B involves the oxidation of carbonyl groups and aliphatic double bonds. This path results in larger amounts of chlorate. The characteristics of chlorate production then are linked to the proportion of  the chlorine dioxide added that reacts according to path A. This is dependent upon each chemical's ability to react with the phenolic hydroxyl groups present in the lignin. Thus chlorate production depends not only on the mode of addition but also on the relative concentrations of chlorine and chlorine dioxide molecules. Upon first inspection, it seems illogical that chlorate production should show a relative maxima with respect to chlorine dioxide substitution. Closer study reveals it  109  to be quite plausible. At low levels of substitution, chlorine molecules far outnumber chlorine dioxide molecules.  Thus, in the competition for phenolic hydroxyl groups,  chlorine takes precedence.  Most of the chlorine dioxide therefore reacts via path B  resulting in chlorate formation. At this point, small increases in substitution mean that more chlorine dioxide is available to react via path B and produce more chlorate. Thus, as substitution is increased, the production of chlorate increases.  However,  further increases in substitution bring the molar concentrations of the two bleaching agents closer together.  A larger portion of the chlorine dioxide reacts with the  phenolic hydroxyl groups causing chlorate production to level off.  As substitution  levels are increased still further, chlorine dioxide molecules outnumber chlorine molecules.  Thus an increasing proportion of the chlorine dioxide reacts via path A  thereby decreasing chlorate production. In theory, if the various reacting forms of the two bleaching chemicals are ignored, chlorine and chlorine dioxide have equal molar concentrations at approximately 71% substitution.  This coincides well with the  observed locations of chlorate production maxima. Phenolic hydroxyl groups are not only formed by the demethylation of chlorinated lignin but they are also present naturally in unbleached kraft pulp. These prechlorination hydroxyl groups are important in explaining the effect of mode on chlorate production.  At low levels of substitution, less chlorate is formed when  chlorine dioxide is added before chlorine. The chlorine dioxide is allowed to react with the original phenolic hydroxyls before any chlorine is present. A larger portion  110 of the chlorine dioxide reacts via path A resulting in less chlorate than if chlorine was added first. At high levels of substitution, there are two competing effects which determine chlorate production.  While adding chlorine first causes fewer original phenolic  hydroxyl groups to be available for oxidation by chlorine dioxide, it also creates new hydroxyl groups through the chlorination and demethylation mechanism. The addition of chlorine dioxide first at high substitution probably results in considerable amounts of benzene ring cleavage that renders the chlorination and demethylation pathway to hydroxyl formation much less likely. Thus the presence of prechlorination hydroxyl groups becomes overshadowed by the ability of chlorine to produce new hydroxyl groups. This ability is only available when chlorine is added first. At high levels of substitution, the addition of chlorine first provides a net gain in the number of hydroxyl groups available to react with chlorine dioxide. Since more chlorine dioxide reacts via path A, less chlorate is formed.  4.4.  Kappa  Number  The brown stock and CE pulp kappa numbers are given in table 16.  The  pooled standard deviation for the bleached pulps from replicate runs was 0.64 ml with six degrees of freedom. Differences between replicate CE kappa numbers tended to coincide with the random variation in the multiple (see appendix C, table 25). No apparent effect was seen on CE kappa number as a result of the wide variation in pulp storage time between chlorination and extraction.  The average brown stock kappa  Table 16 Kappa numbers of brown stock and bleached pulps Run  Brown stock kappa number (ml)  CE pulp kappa number (ml)  0  28.0  6.1  1  28.8  6.6  2  28.7  7.6  3  29.0  5.2  4  28.6  5.9  5  29.6  5.2  6  30.3  4.5  7  28.9  4.5  8  29.8  7.3  9  29.2  4.2  10  28.9  4.9  11  28.9  7.1  12  28.8  4.4  13  29.4  6.6  14  29.4  4.3  15  29.4  4.1  16  30.2  6.6  17  28.8  5.0  18  29.7  4.6  19  29.1  4.7  20  29.2  5.9  21  29.2  4.5  22  29.9  6.2  23  29.9  4.5  112 number was 29.2 ml with a standard deviation of 0.55 ml having 23 degrees of freedom. If it is assumed that the deviation of the brown stock results is due solely to the kappa number test and the same deviation expressed as a percentage of the mean applies at all levels of lignin content, then an approximate standard deviation of the test used on CE pulps is 0.1 ml.  Thus the largest portion of the bleached pulp  variation came from sources other than the kappa number test.  The experimental  kappa numbers compare well with those from Celgar Pulp and the literature review (see section 2.4). The factorial effects, given in table 17, were calculated using the bleached kappa numbers from runs 5-23.  Using 95% confidence intervals the effects of  multiple and mode were found to be significant.  Thus the change in multiple from  0.18 to 0.22 %/ml caused the CE kappa number to decrease by 0.9 ml. The change in mode of addition from DC (-40 seconds) to CD (40 seconds) caused a 1.5 ml increase in the CE kappa number. These findings agree with the literature (Rapson, Anderson, and Reeve 1977; Reeve and Rapson 1980). The effects of substitution and curvature and the interaction between substitution and mode were not found to be significant. While these effects have been observed by others (Reeve and Rapson 1980), it is likely that their magnitudes are too small to have had any significant influence in this study. It should be noted that the factorial analysis used chemical addition, which was represented by the chlorine multiple, to calculate the effects.  A more-accurate  predictor of delignification is chemical consumption. However, the use of chemical addition was deemed appropriate for the following reasons:  Table 17 Factorial analysis of delignification Factor  Effect on C E kappa number (ml)  95% confidence interval (ml)  mode (m)  1.5  ±0.8  multiple (M)  -0.9  ±0.8  curvature  -0.9  ±1.0  Sm  0.6  ±0.8  substitution (S)  0.5  ±0.8  MSm  0.5  ±0.8  recycle (R)  0.3  ±0.8  MS  0.3  ±0.8  MRS  0.3  ±0.8  Mm  -0.3  ±0.8  RS  -0.3  ±0.8  MRSm  0.2  ±0.8  Rm  0.1  ±0.8  MR  -0.1  ±0.8  RSm  -0.1  ±0.8  MRm  0.0  ±0.8  114 1.  The chemical residuals were low (see appendix C, table 21).  2.  Some residual chemical remained unwashed from the pulp.  3.  The chlorine multiple had some inherent variation (see table 8).  4.  The variation in CE kappa numbers from replicate runs was large.  In addition, it is unlikely that the low levels of chlorine which contaminated the chlorine dioxide solutions (see appendix C, table 21) had any measurable effect on delignification. The recycle of chlorination fdtrate had no significant effect on the CE kappa number.  The possible interferences of black liquor and extraction effluent were  eliminated as carryover to the chlorination stage averaged 0.16 kg NajSO^adt and the extractions  were  performed separately  from  the  chlorinations.  Approximate  calculations show that at a recycle rate of 4.5 m /adt the addition of total organic 3  carbon on pulp was 0.1 %. This increased to 0.9% TOC on pulp at 22.5 m /adt. A 3  rate of increase in the CE kappa number with chlorination effluent recycle of 0.3 ml/%TOC (Reeve, Anderson, and Rapson 1980) gives a predicted effect of 0.24 ml for the levels of recycle used. This small effect could not have been detected in this work. Much speculation has been given to the reason for the improved delignification associated with the DC mode of addition.  One line of thought purports that the  addition of chlorine dioxide first oxidizes the lignin thereby eliminating some of the ability of chlorine to react via substitution (Hatton 1967).  This results in more  chlorine reacting via oxidation and thus better delignification. This unproven theory  115 suggests that the production of AOX per kappa unit of delignification is less for DC mode than for CD mode. The present work failed to substantiate this. Both CD and DC modes produced the same amount of AOX per kappa unit of delignification (0.12 kg/(adfml)). Another line of reasoning attempts to explain the difference in delignification through the amount of chlorate formed in the first bleaching stage (Reeve, Anderson, and Rapson 1980). Increases in chlorate formation would cause less chlorine dioxide to be available for bleaching resulting in less delignification. Since CD mode yields the highest CE kappa number and usually the most chlorate, this theory seems logical. However at high substitution levels, there is evidence to suggest that DC mode gives the most chlorate despite having a lower CE kappa number than CD mode. Thus the theory is contradicted. Bergnor et al. (1987) attempted to account for the differences in delignification between the three modes of addition by correcting the consumptions of active chlorine for chlorate formation. The CE kappa number was plotted against the corrected chemical consumption. CD pulp kappa numbers were still much higher for a given chemical charge than those of the DC pulps. The same result was found when a similar plot was made with the data from this study. Having  partially  refuted  the  two  theories  explaining  the  enhanced  delignification of the DC mode of addition, an alternative explanation was developed here.  This new theory concentrates on the reaction of active chemical with the  dissolved organic material as opposed to its reaction with the pulp. The first chemical added in sequential addition comes in contact with recycled filtrate and pulp. The  116 second chemical added comes in contact with recycled filtrate, pulp, and newly produced dissolved organic compounds from the action of the first chemical. Thus the second chemical added has a greater opportunity to react with dissolved organic substances.  The freshly liberated organic compounds may be crucial in determining  the efficiency of delignification.  Reeve, Anderson, and Rapson (1980) showed that  chlorine dioxide reacts 9% faster than chlorine with chlorination filtrate. Since the second chemical added contacts the highest concentration of dissolved organic compounds and chlorine dioxide reacts faster than chlorine with those compounds, a larger portion of the bleaching chemical is consumed by the filtrate when chlorine dioxide is added second. Thus less chemical is available for delignification resulting in a higher CE kappa number.  4.5. Viscosity The bleached and unbleached pulp viscosities are given in table 18.  The  pooled standard deviation for bleached pulps from replicate runs was 2.18 cp with six degrees of freedom. In each case, the bleached pulp viscosity increased for the second and third replicates when compared with the first. This was probably caused by the varying length of time that the chlorinated pulp samples spent in storage before being extracted. Even though the chlorinated pulp was washed with distilled water on the chlorination washer, it is likely that some residual chemical remained with the pulp causing continued degradation during sample storage. The storage time varied directly with run order. Thus the chlorinated pulp from run 0 had the longest storage of 34  Table 18 Viscosities of brown stock and bleached pulps Run  Brown stock viscosity" (cp)  CE pulp viscosity (cp)  0  —  25.2  1  —  26.1  2 3  36.6 —  24.0 27.6  4  35.3  24.9  5  35.0  28.3  6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22  —  34.3 —  35.7 —  33.7 —  34.8 —  34.5 —  34.7 —  33.8 —  34.4 —  28.2 23.8 28.0 25.2 23.7 26.6 25.4 27.8 28.7 25.8 27.6 26.2 27.1 27.1 27.6 28.8 30.5  23  33.8  25.5  Celgar  37.1  33.4  "Single determinations except for Celgar sample.  118 weeks while that from run 23 had the shortest storage of 3 weeks.  Despite the  variation in time between chlorination and extraction, no distinct trend was revealed when the CE pulp viscosities were plotted against the time order of observation for all 24 runs. The largest portion of the variation in the data came from sources other than the viscosity test as the standard deviation of the brown stock viscosities was 0.85 cp with 11 degrees of freedom. The average viscosity loss for the 24 runs was 8.0 cp compared with only 3.7 cp for the pulp bleached at Celgar Pulp. The larger degradation experienced in the laboratory was presumably due to the extended storage of the chlorinated pulps in the presence of residual chemical.  Differences between the two brown stocks as  evidenced by their viscosities may also have contributed. However, none of the CE pulp viscosities can be considered substandard. All of the values compare favorably with acceptable levels as noted in the literature review. The effects of the four variables, their interactions, and curvature were calculated using the bleached pulp viscosities from runs 5-23 comprising the factorial design.  None of the effects were found to be significant using 95% confidence  intervals. The most noteworthy discrepancy between these results and those of du Manoir (1980) is the effect of mode on viscosity.  According to du Manoir, the addition of  20% substitution 40 seconds before the chlorine at 30°C results in a severely degraded pulp of 12 cp. viscosities.  The same conditions in this study gave indistinguishable pulp  There is some evidence to support the finding that mode has no effect on  119 cellulose degradation under similar conditions.  Wartiovaara (1980b) used 30%  substitution added 30 seconds before the chlorine at 45°C and found no detrimental effect.  In addition, Celgar Pulp uses 30% substitution added 30 seconds before the  chlorine at 43°C without severe viscosity loss. A possible reason for this discrepancy could be the effect of chlorine contamination in the chlorine dioxide. Pure chlorine dioxide was used by du Manoir (1980) whereas that used here contained on average 0.22 g/1 chlorine, which represented 1.6% of the available chlorine in the chlorine dioxide. This contamination may have been sufficient to prevent the chlorine dioxide from being exhausted at an early stage thereby preserving the viscosity of the cellulose. The finding that chlorinationfiltraterecycle has no effect on pulp viscosity is in agreement with Wartiovaara (1980b) and Histed and Nicolle (1973). In addition, since substitution levels were well above the 4% needed for viscosity protection, it is not surprising that substitution had no effect.  Finally, the range of multiple studied  was probably too narrow for it to have influenced pulp viscosity.  CHAPTER 5 SUMMARY AND CONCLUSIONS  1.  A continuous lab-scale pulp chlorination stage was designed and fabricated. The apparatus was successfully operated over a wide range of conditions and produced results which can be directly compared to those from industry. As well, this special equipment allowed the continuous recycle of chlorination filtrate for brown stock dilution as is done industrially and made it feasible to produce sufficient effluent at steady state for use in trout bioassays.  2.  A qualitative summary of the results from the experimentation carried out in a factorial design with the apparatus is given in table 19.  3.  Of the four factors studied, the recycle of chlorination fdtrate for brown stock dilution had the largest impact on the toxicity to trout of the chlorination effluent. The chlorination stage toxic emission factor decreased on average by 314 m /adt as the level of recycle increased from 4.5 to 22.5 m /adt. 3  3  This  effect was more than twice as large as that caused when the chlorine dioxide substitution in the first stage was increased from 20 to 80%. This finding has two important implications.  Firstly, bleached kraft pulp mills currently  operating with low levels of chlorination filtrate recycle could substantially decrease toxic discharges by increasing the level of recycle. Secondly, toxicity  120  Table 19 Qualitative summary of experimental results Response variables Pulp quality  Effluent quality C stage toxic emission factor  Significant factors  trout  bacteria  Recycle (R)  /  y  Mode (m)  /  Substitution (S)  /  Multiple (M)  /  AOX from CE sequence  Chlorate from C stage  C E kappa number  /  /  /  / /  Curvature Interactions  Formulation of results  / MSm MRm  Regression equation Factorial effects  /  /  RS Rm RSm  MS  Sm  /  / /  CE viscosity  122 data from chlorination effluents produced in the laboratory on a batch basis without recycle are not representative of the bleaching process as it is performed industrially. 4.  The recycle of chlorination filtrate for brown stock dilution had no effect on AOX, chlorate, kappa number, and viscosity.  Thus mills can increase  chlorination effluent recycle in order to decrease the discharge of toxicity without detrimentally affecting pulp quality or the formations of AOX and chlorate. In addition, bleaching experiments performed in the laboratory on a batch basis without recycle are likely to produce results with respect to these four response variables which are in good agreement with those produced industrially using various levels of recycle. This conclusion must be tempered by stating that the interaction between the recycle of chlorination filtration for brown stock dilution and the countercurrent recycle of bleaching effluents for pulp washing and seal tank dilution could well affect AOX, chlorate, and kappa number. The cause of such an effect would most likely be the chlorination of extraction filtrate transported into the first bleaching stage. 5.  The toxicity of chlorination effluent to luminescent marine bacteria did not correlate well with the toxicity to rainbow trout. It is unrealistic to expect that two organisms with vastly different biological complexities should behave identically. Since most regulatory standards for toxicity are based on trout as the test organism, Microtox data should be viewed as being supplemental to  123 trout data. Accordingly, the use of the Microtox bioassay simply because of its ease of use is not warranted. 6.  A new method for determining the toxic emission factor to trout was developed.  The technique involves determining the LT50 at 100% effluent  concentration in conjunction with the slope of the concentration versus time function.  The method ensures a rapid response, which maximizes sensitivity  and minimizes toxicant depletion.  This procedure may be of use when  problems that are associated with standard bioassay methodology, such as test duration and toxicant depletion, are encountered.  CHAPTER 6 RECOMMENDATIONS  Future studies should be undertaken to determine the effects of black liquor carryover and countercurrent extraction filtrate recycle for pulp washing and seal tank dilution on the quality of both effluent and pulp.  Chlorination of  black liquor and extraction fdtrate could be easily performed with the continuous lab-scale bleaching apparatus. Black liquor could be added to the unbleached pulp at the point of dilution between the decker discharge and chemical addition. Extraction filtrate could be added either to the wash water headbox of the chlorination washer or to the chlorination seal tank. A detailed chemical study of the cause of the toxicity to trout of chlorination effluent should be carried out.  The ability of the continuous lab-scale  chlorination apparatus to produce enough effluent under steady-state conditions of recycle to enable the use of trout bioassays makes it an ideal tool for this task.  Significant advances may be made if the effects of the chloroquinones  and chlorovanillins are included. The assumption that LT50 values determined for effluent concentrations near the 96-hour LC50 follow the same slope when plotted against concentration as LT50 values determined for much higher concentrations should be tested. This  124  125 assumption formed the basis for the use of median lethal time data in the determination of trout toxic emission factors.  Theory predicts that all points  should lie on the same line provided that proper volume factors are maintained. The effect of residual chlorine and chlorine dioxide on low molecular weight chlorinated organics in chlorination effluent should be investigated.  This  knowledge would help to clarify any ambiguity between the effects on toxicity of the recycle of chlorination filtrate and of the action of residual chemical between effluent collection and residual reduction. The determination of chemical residual in chlorination effluents that are to be tested for toxicity or chlorophenolic compounds should be performed using ion chromatography. This would eliminate the possibility, which is associated with the iodometric titration used in this study, of detecting and reducing chlorophenolic compounds as if they were residual bleaching chemical. Pulp extractions should be performed immediately following chlorination. This would eliminate the combined effect of residual chemical and chlorinated pulp storage time on extracted pulp viscosity. In addition, any uncertainty regarding the validity of the extraction filtrates due to the storage of chlorinated pulp would be removed.  REFERENCES  American Public Health Association. 1985. Standard methods for the examination of water and wastewater. 16th ed. Washington, DC: APHA, AWWA, and WPCF. Axegard, P. 1985. Effect of C10 -substitution on bleaching efficiency and the formation of organically bound chlorine - Part II. Proceedings of the CPPAITAPPI International Pulp Bleaching Conference in Quebec, PQ, June 18-21, 1985, 169-175. 2  Axegard, P. 1987. C10 substitution reduces the load of TOCl. Proceedings of the TAPPI Pulping Conference in Washington, DC, Nov. 1-5, 1987, 105-111. 2  Axegard, P. 1989. Improvement of bleaching plant effluent by cutting back on Cl . Pulp and Paper Canada, vol.90, no.5, T183-T187. 2  Belt, P.B., T.W. Joyce, and H.-M. Chang. 1982. Some environmental aspects of alternative pulp bleaching technologies. Proceedings of the TAPPI Annual Meeting in Washington, DC, Mar. 8-11, 1982, 391-404. Bergnor, E., U. Germgard, J.J. Kolar, and B.O. Lindgren. 1987. Formation of chlorate in chlorine dioxide bleaching. Cellulose Chemistry and Technology, vol.21, no.3, 307-314. Betts, J.L., and G.G. Wilson. 1966. New methods for reducing the toxicity of kraft mill bleachery wastes to young salmon. Journal of the Fisheries Research Board of Canada, vol.23, no.6, 813-824. Bliss, CI. 1937. The calculation of the time-mortality curve. Annals of Applied Biology, vol.24, no.4, 815-852. Braddon, S.A., and C.W. Dence. 1968. Structure and reactivity of chlorolignin - Part I: Alkaline hydrolysis of chlorine-substituted lignin model compounds. Tappi, vol.51, no.6, 249-256. British Columbia Ministry of Environment. 1982. Provincial guidelines and laboratory procedures for measuring acute lethal toxicity of liquid effluents to fish. Victoria, BC: Province of British Columbia. 126  127 Byzyna, L., B. Clifford, and G. Billmark. 1989. Bleach plant operations. In Kraft pulping in Canada: A manual of kraft mill operating practices, edited by M. MacLeod, 89-118. Montreal, PQ: CPPA. Das, B.S., S.G. Reid, J.L. Betts, and K. Patrick. 1969. Tetrachloro-o-benzoquinone as a component in bleached kraft chlorination effluent toxic to young salmon. Journal of the Fisheries Research Board of Canada, vol.26, no.ll, 3055-3067. Donnini, G.P. 1982a. Bleach plant effluents: Chemical control of some environmental problems. Proceedings of the CPPA Annual Meeting in Montreal, PQ, Jan. 2629, 1982, B85-B92. Donnini, G.P. 1982b. The effect of chlorine dioxide substitution on bleaching effluent toxicity and mutagenicity. Proceedings of the CPPA Environment Conference in Halifax, NS, June 8-10, 1982, 73-81. du Manoir, J.R. 1980. Chlorination strategy for pulp viscosity protection. Transactions of the Technical Section (CPPA), vol.6, no.2, TR25-TR30. Earl, P.F., and D.W. Reeve. 1989a. Chlorinated organic matter in bleached chemical pulp production: The effect of chlorination-stage variables on chlorinated organic matter in effluent. Tappi Journal, vol.72, no.10, 183-187. Earl, P.F., and D.W. Reeve. 1989b. Chlorinated organic matter in bleached chemical pulp production - Part VI: Chlorinated compounds in effluents. Proceedings of the TAPPI Pulping Conference in Seattle, WA, Oct. 22-25, 1989, 31-37. Easty, D.B., J.E. Johnson, and A.A. Webb. 1986. Analysis of bleaching liquors by ion chromatography. Paperi ja Puu, vol.68, no.5, 415-417. Eckerman, P. 1987. Recovery of bladder wrack kelp expected at Odangla. Nordisk Cellulosa, vol.4, no.7, 48-49. Translated by Ivar Johnsson. Firth, B.K., and C.J. Backman. 1990. Comparison of Microtox® testing with rainbow trout (acute) and Ceriodaphnia (chronic) bioassays in mill wastewaters. Tappi Journal, vol.73, no.12, 169-174. Fraser, J.L., and D.W. Reeve. 1989. Chlorinated organic matter in bleached chemical pulp production - Part V: An interlaboratory comparison of techniques and test results for A O X analysis in effluents. Proceedings of the CPPA Spring Conference in Whistler, BC, May 24-27, 1989. Gall, R.J., and F.H. Thompson. 1973. The anti-pollution sequence: A new route to reduced pollutants in bleach plant effluent. Tappi, vol.56, no.ll, 72-76.  128 Gergov, M., M . Priha, E. Talka, and O. Valttila. 1988. Chlorinated organic compounds in effluent treatment at kraft mills. Tappi Journal, vol.71, no.12, 175-184. Germgard, U. 1989. Chlorate discharges from bleach plants - how to handle a potential environmental problem. Paperi ja Puu, vol.71, no.3, 255. Germgard, U., and R.-M. Karlsson. 1984. Optimised charge of bleaching agents in the prebleaching of softwood kraft pulp at different chlorine dioxide levels. Paperi ja Puu, vol.66, no. 11, 627. Germgard, U. and R.-M. Karlsson. 1988. Bleaching of birch kraft pulp with different fractions of chlorine dioxide in the prebleaching. Nordic Pulp and Paper Research Journal, vol.3, no.4, 166-171. Germgard, U., and S. Larsson. 1983. Oxygen bleaching in the modern softwood kraft pulp mill. Paperi ja Puu, vol.65, no.4, 287-290. Germgard, U., R.-M. Karlsson, K.P. Kringstad, F. de Sousa, and L.M. Stromberg. 1984. Oxygen bleaching and its impact on the environment. Proceedings of the TAPPI Oxygen Delignification Symposium in San Francisco, CA, Nov. 15-16, 1984, 99-103. Guo, P.H.M., W.K. Bedford, and B.E. Jank. 1976. Factors affecting activated sludge treatment of kraft bleachery effluent. Environmental Protection Service Report No.EPS 4-WP-76-1. Ottawa, ON: Environment Canada. Hakulinen, R., and M. Salkinoja-Salonen. 1982. Treatment of kraft bleaching effluents: Comparison of results obtained by Enso-Fenox and alternative methods. Proceedings of the CPPA/TAPPI International Pulp Bleaching Conference in San Diego, CA, May 9-12, 1982, 97-106. Hall, E.R., J. Fraser, S. Garden, and L.-A. Cornacchio. 1988. Organochlorine discharges in wastewaters from kraft mill bleach plants. Proceedings of the CPPA Environment Conference in Vancouver, BC, Oct. 25-27, 1989, 53-59. Hatton, J.V. 1967. Delignification of kraft pulp in the first bleaching stage using chlorine dioxide and chlorine: Comparison of sequential addition of chlorine dioxide and chlorine with treatments involving chlorine and mixtures of chlorine and chlorine dioxide. Pulp and Paper Magazine of Canada, vol.68, no.4, T-181. Histed, J.A. and F.M.A. Nicolle. 1973. Water reuse and recycle in kraft bleacheries. Pulp and Paper Magazine of Canada, vol.74, no.12, T386-T397.  129 Histed, J.A., and R. Vega Canovas. 1989. Decreasing the usage of chlorine and caustic in bleaching. Proceedings of the TAPPI Pulping Conference in Seattle, WA, Oct. 22-25, 1989, 85-93. Holmbom, B., and K.-J. Lehtinen. 1980. Acute toxicity to fish of kraft pulp mill waste waters. Paperi ja Puu, vol.62, no.ll, 673. Jamieson, A.G., 1970. A rapid bleaching process for kraft pulp. Tappi, vol.53, no.9, 1732-1736. Johannes, D.L., and R.W. Lowe. 1989. AOX testing of bleach plant effluents. Proceedings of the CPPA Spring Conference in Whistler, BC, May 24-27, 1989. Kramer, J. 1972. Delignification of kraft pulp with chlorine, chlorine dioxide, and their mixtures. Tappi, vol.55, no.6, 964-971. Kutney, G.W., T.S. Macas, and G.P. Donnini. 1983. The C stage for the 1980's. Proceedings of the CPPA Annual Meeting in Montreal, PQ, Feb. 1-4, 1983, A11-A18. Leach, J.M., and A.N. Thakore. 1975. Isolation and identification of constituents toxic to juvenile rainbow trout (Salmo gairdneri) in caustic extraction effluents from kraft pulpmill bleach plants. Journal of the Fisheries Research Board of Canada, vol.32, no.8, 1249-1257. Leach, J.M., and A.N. Thakore. 1976. Identification of the toxic constituents in kraft mill bleach plant effluents. Cooperative Pollution Abatement Research Project Report No.245-3. Ottawa, ON: Environment Canada. Liebergott, N., B. van Lierop, T. Kovacs, and A . Nolin. 1990a. A comparison of the order of addition of chlorine and chlorine dioxide in the chlorination stage: Comparison at constant chemical charge. Tappi Journal, vol.73, no.10, 207213. Liebergott, N., B. van Lierop, A. Nolin, M . Faubert, and J. Laflamme. 1990b. Modifying the bleaching process to decrease AOX formation. Proceedings of the CPPA Annual Meeting in Montreal, PQ, Jan. 30 - Feb. 2, 1990, B229B238. Lindgren, B.O., and T. Nilsson. 1975. Chlorate formation during the reaction of chlorine dioxide with lignin model compounds. Svensk Papperstidning, vol.78, no.2, 66-68.  130  Lindstrom, K., J. Nordin, and F. Osterberg. 1981. Chlorinated organics of low and high relative molecular mass in pulp mill bleachery effluents. Chapter 52 in Advances in the identification and analysis of organic pollutants in water, vol.2, edited by L.H. Keith, 1039-1058. Ann Arbor, MI: Ann Arbor Science Publishers. Litchfield, J.T., Jr. 1949. A method for rapid graphic solution of time-percent effect curves. The Journal of Pharmacology and Experimental Therapeutics, vol.97, no.4, 399-408. Mattinen, H., and I. Wartiovaara. 1981. The pollution load of a closed D/CEDED bleachery. Paperi ja Puu, vol.63, no.ll, 688. Microbics Corporation. 1982. Microtox system operating manual. Carlsbad, CA: Microbics Corporation. National Council of the Paper Industry for Air and Stream Improvement. 1981. Experience with the analysis of pulp mill effluents for chlorinated phenols using an acetic anhydride derivatization procedure. Stream Improvement Technical Bulletin No.347. New York, NY: NCASI. Nikki, M., and R. Korhonen. 1983. Chlorinated organic compounds in effluents of bleaching with countercurrent washing. Journal of Pulp and Paper Science, vol.9, no.5, TR123-TR128. Nilsson, T., and L. Sjostrom. 1974. Losses in chlorine dioxide as a result of chlorate formation during bleaching. Svensk Papperstidning, vol.77, no. 17, 643. Nonni, A.J., and C.W. Dence. 1981. The reactions of creosol and propioguaiacone with chlorine, chlorine dioxide and their combinations. Svensk Papperstidning, vol.84, no.3, R17-R24. Pfister, K., and E. Sjostrom. 1979a. Characterization of spent bleaching liquors Part 2: Composition of material dissolved during chlorination (CEH sequence). Paperi ja Puu, vol.61, no.4a, 220. Pfister, K., and E. Sjostrom. 1979b. Characterization of spent bleaching liquors - Part 3: Composition of material dissolved during alkali extraction (CEH sequence). Paperi ja Puu, vol.61, no.4, 367-370. Pulp and Paper Research Institute of Canada. 1986. A scheme for routine analysis of non polar extractives. Vancouver, BC: PAPRICAN.  131 Pulp and Paper Research Institute of Canada. 1989a. GC calibration and analysis. Vancouver, BC: PAPRICAN. Pulp and Paper Research Institute of Canada. 1989b. Methylation of extractives for GC analysis. Vancouver, BC: PAPRICAN. Rapson, W.H., and C.B. Anderson. 1966. Mixtures of chlorine dioxide and chlorine in the chlorination stage of pulp bleaching. Pulp and Paper Magazine of Canada, vol.67, no.l, T47-T55. Rapson, W.H., C.B. Anderson, and D.W. Reeve. 1977. The effluent-free bleached kraft pulp mill - Part VI: Substantial substitution of chlorine dioxide for chlorine in the first stage of bleaching. Pulp and Paper Canada, vol.78, no.6, T137-T148. Reeve, D.W., and P.F. Earl. 1986. Mixing gases, water, and pulp in bleaching. Tappi Journal, vol.69, no.7, 84-88. Reeve, D.W., and W.H. Rapson. 1980. Developments in chlorine dioxide bleaching. Proceedings of the TAPPI Pulping Conference in Atlanta, GA, Nov. 16-19, 1980, 403-409. Reeve, D.W., C.B. Anderson, and W.H. Rapson. 1980. The effluent-free bleached kraft pulp mill - Part X: Dissolved organic matter affects the first bleaching stage. Pulp and Paper Canada, vol.81, no.6, T142-T146. Rydholm, S.A. 1965. Pulping processes. New York, NY: Interscience Publishers. Salkinoja-Salonen, M., M.-L. Saxelin, J. Pere, T. Jaakkola, J. Saarikoski, R. Hakulinen, and O. Koistinen. 1981. Analysis of toxicity and biodegradability of organochlorine compounds released into the environment in bleaching effluents of kraft pulping. Chapter 56 in Advances in the identification and analysis of organic pollutants in water, vol.2, edited by L.H. Keith, 1131-1164. Ann Arbor, MI: Ann Arbor Science Publishers. Sameshima, K., B. Simson, and C.W. Dence. 1979. The fractionation and characterization of toxic materials in kraft spent bleaching liquors. Svensk Papperstidning, vol.82, no.6, 162-170. Scroggins, R.P. 1984. In-plant toxicity balances for a bleached kraft pulp mill. Proceedings of the CPPA Annual Meeting in Montreal, PQ, Jan. 31 - Feb. 3, 1984, B189-B194. Sepall, O. 1970. Laboratory apparatus for experimental work in pulping, bleaching, and chemical treatments of cellulosics. United States Patent 3,540,982.  132 Singh, R.P., ed. 1979. The bleaching of pulp. 3rd ed. Atlanta, GA: Technical Association of the Pulp and Paper Industry. Sjostrom, L., R. Radestrom, and K. Lindstrom. 1982. Determination of total organic chlorine in spent bleach liquors. Svensk Papperstidning, vol.85, no.3, R7-R13. Strumila, G.B., and W.H. Rapson. 1979. The destruction of toxic tri- and tetrachloroguaiacol by aqueous chlorine dioxide. Proceedings of the CPPA Annual Meeting in Montreal, PQ, Jan. 30 - Feb. 2, 1979, A133-A138. Suntio, L.R., W.Y. Shiu, and D. Mackay. 1988. A review of the nature and properties of chemicals present in pulp mill effluents. Chemosphere, vol.17, no.7, 12491290. Technical Association of the Pulp and Paper Industry. 1988. Tappi test methods. Atlanta, GA: TAPPI. Technical Association of the Pulp and Paper Industry. 1990. Tappi useful methods. Atlanta, GA: TAPPI.  Thomas, P., E. Becker, and K.B. Gibney. 1975. Identification and determination of the toxicity contribution of the toxic materials in each kraft and sulphite proces effluent prepared in a pilot plant. Cooperative Pollution Abatement Research Project Report No.360-1. Ottawa, ON: Environment Canada. Tomlinson, G.H. 1980. Toxicity to fish helps identify individual process streams. Pulp and Paper Canada, vol.81, no.12, T342-T349. Voss, R.H., J.T. Wearing, and A. Wong. 1981. Effect of softwood chlorination conditions on the formation of toxic chlorinated compounds. Pulp and Paper Canada, vol.82, no.2, T65-T71. Voss, R.H., J.T. Wearing, R.D. Mortimer, T. Kovacs, and A. Wong. 1980. Chlorinated organics in kraft bleachery effluents. Paperi ja Puu, vol.62, no.12, 809-814. Walden, C , D. McLeay, and D. Monteith. 1975. Comparing bioassay procedures for pulp and paper effluents. Pulp and Paper Canada, vol.76, no.4, T130-T134. Wartiovaara, I. 1980a. The chemical consumption of a closed D/CEDED bleachery. Paperi ja Puu, vol.62, no.5, 319. Wartiovaara, I. 1980b. Water re-use in bleachery. Pulp and Paper Canada, vol.81, no.7, T167-T171.  133 Western Canada Microtox Users Committee. 1986. Microtox interlaboratory comparison study: Standard assay procedure. File No.2600-BK3-13. Vegreville, AB: Alberta Environmental Centre. Wong, A., M. Le Bourhis, R. Wostradowski, and S. Prahacs. 1978. Toxicity, BOD and color of effluents from novel bleaching processes. Pulp and Paper Canada, vol.79, no.7, T235-T241.  APPENDICES  134  APPENDIX A SPECIFICATIONS OF T H E PULP CHLORINATION APPARATUS  A . l . Stock Tank System The function of the stock tank system, which is shown in figure 11, is to provide a uniform supply of unbleached pulp for the brown stock decker. The pulp is contained in two cylindrical, polyethylene tanks, which are 30 in. in diameter and 48 in. in height. Each tank sits on a welded structural steel stand and is equipped with a 1/3-hp mixer that operates at 1,725 rpm (Lightnin Model NS-1). The mixers, which are mounted on the stands, are each supplied with two triple-blade, marine-type propellers that are 3 in. in diameter. The tanks are connected to each other and to a drain with 2-in. PVC pipe and two ball valves. Figure 1 shows the valve locations. A centrifugal pump is used to transfer the pulp suspension from stock tank no.l to the constant-head tank. This stainless steel pump has a 3 1/2-in., open impeller (Liquiflo Equipment Model 62) that is driven by a 1/4-hp, 90-Vdc, variable speed motor (Dayton Electric Manufacturing Model 2M167C).  The motor, which has a  maximum speed of 1,725 rpm, is regulated to approximately 900 rpm with a controller (KB Electronics Model KB1C-120). The suction line, which is made of 1-in. PVC pipe, has its inlet 4 in. above the bottom of the stock tank. The pump discharges into the top of the constant-head tank through 3/4-in. PVC pipe.  135  136 Figure 10 shows the constant-head tank.  The suction of the brown stock  peristaltic pump is taken from a 3/8-in. o.d. brass tube that extends through the 1/2-in. NPT port on the side wall of the head tank (see drawing HT-1). A conical section, which increases in diameter from 3/8 in. to 2 1/4 in. over a distance of 1 1/2 in., is attached to the end of the tube inside the head tank. This prevents the tube entrance from becoming blocked with fiber floes.  A.2. Brown Stock Decker The function of the brown stock decker, which is shown in figure 3, is to provide the chlorination reactor with a constant supply of unbleached pulp at medium consistency.  A belt made from 40-mesh PVDF fabric is positioned around two rolls  (see drawing PW-7) located at opposite ends of the decker. The belt, which covers the width between the two walls of the decker, is sewn with a lap seam and sealed at the edges with vinyl. The drive roll is mounted in the 3/8-in. hole near the end of the decker (see drawing PW-1).  It is connected to a 1-rpm, 1/250-hp shaded-pole  gearmotor (Dayton Electric Manufacturing Model 3M095). The motor is mounted on one of the walls of the decker (see drawing PW-5) and is covered to avoid water damage (see drawing PW-6). The discharge roll is mounted in the 3/8-in. horizontal slot with a spring loaded device. This system causes the belt, which is placed under tension, to move across a perforated table (see drawing PW-4) located between the two rolls. Two suction boxes are situated beneath the table.  137 The pulp suspension is pumped from the constant-head tank to the 1/4-in. NPT port on the decker headbox (see drawing PW-2).  Four machine screws, which are  located in the 10UNF32 holes in the top of the headbox, contact the top edges of the decker walls.  This enables the headbox, which is shown in figure 4, to be leveled.  Another machine screw, which is located in the side of the headbox, holds it in place. The headbox discharges the suspension over the first suction box approximately 2 in. from the rear of the table. Most of the drainage occurs from the first suction box due to the action of gravity.  Two press rolls (see drawing PW-8) cause additional drainage from the  second suction box.  The press rolls are mounted in the 3/8-in. vertical slots with  spring loaded mechanisms that are similar to that used with the discharge roll. The first press roll is positioned approximately 5/16 in. above the table and is not loaded. The second press roll is positioned on the table and is loaded lightly. Filtrate that is collected in the suction boxes is discharged from the 3/4-in. NPT ports. It then flows via gravity through two lines, which are made of 3/4-in. i.d. PVC tubing, to the 3/4-in. o.d. tubes of the decker seal tank (see drawing ST-1).  The seal tank enables the  vacuum created by the blower to be maintained in the suction boxes. Under normal conditions, a vacuum of roughly 6 in. of water is developed in the second suction box whereas the first operates at atmospheric pressure. Thefiltrateis discharged from the lower 1/4-in. NPT port on the seal tank. The thickened pulp is discharged from the end of the decker with the aid of a doctor. The doctor, which is a U-shaped section of 3/16-in. o.d. stainless steel tube, is  138 fastened to the doctor mount from the 4UNC40 holes (see drawing PW-9). The doctor mount is coupled to the spring loaded device associated with the discharge roll. In addition, the drive and discharge rolls and the press rolls are mounted on 1/4-in. stainless steel shafts. Specially designed PTFE bushings are used between the o  shafts and the decker walls.  The decker walls are fastened together at the 3/16-in.  holes with the aid of acrylic spacers, which are the same width as the table.  A3.  Chlorination Reactor The chlorination reactor is the vessel in which the pulp is bleached. Drawings  CR-1 through CR-16 contain details of the individual reactor components.  The  assembly of the reactor is shown in drawing CR-17. Recycled chlorination filtrate and dilution water are added to the unbleached pulp through the 1/8-in. NPT port on the inlet section (see drawing CR-1). Bleaching chemical is added through the 1/8-in. NPT ports on the second downflow section (see drawing CR-3). This is shown in figure 5. Recycled chlorination filtrate is added at the reactor outlet through the 1/4-in. hole in the outlet cover (see drawing CR-13). This is shown in figure 6.  A 1-in. port is provided in the horizontal section (see  drawing CR-7) to allow the reactor to be drained. In addition, ventilation is supplied to exhaust the vapors of chlorine and chlorine dioxide from the vicinity of the reactor. Chlorination mixer no.l is driven by a 1/6-hp, 90-Vdc motor (Dayton Electric Manufacturing Model 4Z528). The motor, which has a maximum speed of 1,800 rpm, is regulated to approximately 450 rpm with a controller (KB Electronics Model KBIC-  139 120).  The 3/16-in. stainless steel shaft is held in place by a specially designed  bushing (see drawing CR-15).  Nine 1.8-in. diameter marine-type propellers are  mounted on the shaft. Four propellers, which are made of plastic, are located above the first chemical addition port. The other five, which are cast from an aluminummagnesium alloy and coated with PVDF, are located below the first chemical addition port. Chlorination mixer no.2 is driven by a 1/27-hp, 90-Vdc motor (Dayton Electric Manufacturing Model 4Z142). The motor, which has a maximum speed of 1,800 rpm, is  regulated to  approximately  1,350 rpm with a controller (Dayton Electric  Manufacturing Model 4Z828). The 3/16-in. stainless steel shaft extends to the elbow where the upflow portion of the reactor begins.  Three 1.8-in. diameter marine-type  propellers, which are made of PVDF coated alloy, are mounted on the shaft. The mixer is mounted directly on the reactor (see drawing CR-14). The motor and controller of chlorination mixer no.3 are identical to those of mixer no.l. This mixer is regulated to approximately 325 rpm. The 3/16-in. stainless steel shaft is held in place by a specially designed bushing (see drawing CR-16). Seven propellers are mounted on the shaft. Two of these are 2.2-in. diameter marinetype propellers made of PVDF coated alloy.  One is located where the reactor  diameter increases from 2 to 4 in., and the other is located in the dilution zone near the reactor outlet.  Four 3.7-in. diameter aviation-type propellers made of stainless  steel are positioned between these two. The seventh propeller is located just beneath  140 the bushing that holds the shaft in place. This 1.8-in. diameter marine-type propeller is made of PVDF coated alloy.  A.4. Dilution Vat The function of the dilution vat is to provide the chlorination washer with a constant supply of a dilute suspension of chlorinated pulp. The dilution vat is shown in figure 7 and detailed in drawing DV-1. The motor and controller of the dilution vat mixer are identical to those of chlorination mixer no.2.  This mixer is regulated to  approximately 900 rpm. A single 2.2-in. diameter marine-type propeller made of PVDF coated alloy is mounted on the 3/16-in. stainless steel shaft.  A.5. Chlorination Washer The purpose of the chlorination washer, which is shown in figure 8, is to thicken and wash the chlorinated pulp. Its design is identical to that of the brown stock decker. However, there are some important operational differences. The dilution vat discharges the chlorinated stock over the first suction box approximately 2 in. from the rear of the washer table. Small sections of the washer walls near the drive roll are removed to facilitate the location of the dilution vat. The first press roll is positioned roughly 3/16 in. above the table and is not loaded. The second press roll is positioned on the table and is loaded moderately. Wash water is added to the washer headbox through the 1/16-in. NPT port (see drawing PW-3). Felt, which is draped over the front face of the headbox, is used to  141 distribute the wash water evenly across the pulp sheet. The felt is held to the headbox with two machine screws, which are located in the 10UNF32 holes next to the 1/16-in. NPT port. The position of the washer headbox is controlled in the same manner as that of the decker headbox. Filtrate that is collected in the suction boxes is discharged to the washer seal tank, which is shown in figure 9. Recycled filtrate for dilution at the reactor outlet is pumped from the lower 1/4-in. NPT port of the seal tank (see drawing ST-2). Recycled filtrate for dilution at the reactor inlet is pumped in tubing which extends through the 7/32-in. hole. Chlorination effluent for testing is collected from the upper 1/4-in. NPT port. The levels of vacuum are the same as those attained in the brown stock decker.  A.6. Vacuum System Suction is supplied to both the brown stock decker and the chlorination washer to aid the separation of filtrate and pulp.  This suction is generated by a 1/3-hp  blower, which operates at 3,450 rpm (Gast Manufacturing Model R2103). A section of 1-in. PVC pipe serves as a header from which individual vacuum lines extend to the 3/8-in. NPT ports on the suction boxes of the decker and the washer (see drawing PW-1). These lines are made from reinforced 3/8-in. i.d. PVC tubing. Each line is equipped with a 250-ml filter flask, which separates any entrained filtrate, and a 3/8in. globe valve, which regulates the suction. A 1-in. globe valve on the inlet of the header is used as a bypass for coarse vacuum adjustment.  142 A.7. Heat Exchanger A counterflow, concentric-tube heat exchanger is used to heat the diluted filtrate that is pumped to the reactor inlet.  Hot water at approximately 50°C is  supplied to the 3/8-in. o.d. copper tube that forms the outer section of the exchanger. The inner section, which carries the process filtrate, is made of 1/8-in. o.d. stainless steel tube. A ball valve on the hot water inlet is used to control the exchanger, which is roughly 5 ft long.  A.8. Peristaltic Pumps Except for the brown stock flow to the constant-head tank, all fluids are pumped peristaltically. Details of these variable speed, 1/10-hp pumps are given in table 20. Table 20 Peristaltic pump specifications Application  Cole-Parmer Instrument Model  Tubing size  Tubing formulation  Brown stock  7523-00  18  Tygon  Chlorine  7523-00  14  Norprene  Chlorine dioxide  7523-10  14  Norprene  Filtrate recycle to reactor inlet  7523-00  16  Tygon  Dilution water  7523-00  16  Tygon  Wash water  7523-10  16  Tygon  Filtrate recycle to reactor outlet  7553-20  15  Tygon  143 A.9. pH Measurement The pH of the chlorination filtrate is measured with a combination electrode (Omega Engineering Model PHE-2114) that is located at the 1/2-in. keyway in the chlorination washer seal tank (see drawing ST-2). This pH is displayed using a digital transmitter with manual temperature compensation (Omega Engineering Model PHTX91).  A.10. Temperature Measurement The temperature of the process is measured with three stainless steel thermistors (Omega Engineering Model ON-410) that are located at the 1/8-in. holes in the horizontal section of the reactor (see drawing CR-7), in the reactor outlet cover (see drawing CR-13), and in the chlorination washer seal tank (see drawing ST-2). A rotary selector switch (Omega Engineering Model OSW3-10) connects the sensors to a digital display (Omega Engineering Model 651).  A.11. Control Panel The control panel serves as a central location for the distribution of electric energy, the control of electric equipment, and the measurement of pH and temperature. The panel, which is constructed from aluminum sheet and link tube, is shown in figure 12. Since it is designed for a three-stage bleaching sequence with an allowance for expansion to five stages, it is oversized for the single chlorination stage. The panel is fed with a 60-A line at 120 Vac. In the case of an emergency, a main relay enables  144 the complete shutdown of all electric energy serving the apparatus. Otherwise, each component is equipped with separate panel mounted switches, indicating lights, and fuses.  In addition, the controllers of the variable speed motors are housed in the  control panel. The only items not controlled from the panel are the peristaltic pumps.  A.12. Support Structure The support structure, which is shown in figure 2, is a frame on which all of the components of the chlorination apparatus except for the stock tank system and the control panel are suitably mounted or positioned. The structural aluminum frame is made in several welded sections that are bolted together.  It includes two 1 1/2-in.  aluminum pipes that are mounted vertically with 12 in. between their centers. Individually designed brackets are used to support the brown stock decker and seal tank, the chlorination reactor and mixers, the dilution vat and mixer, and the chlorination washer and seal tank.  These stainless steel brackets, which are made  primarily from 1-in. angle, are fastened to the pipes with U-bolts.  This system  provides great flexibility in positioning the main components of the chlorination apparatus.  N O T E S : 1.  NLET U B C  G.D.F.  R E F E R  T O  D R A W I N G  C R - 6  F O R  SECTION S C A L E :  1 ;  2  N O . R E Q D :  F L A N G E  D E T A I L  M A T L :  1  5/21/87  PMMA CR-1  18  —I — I  1 / 2  '  1 / 4  1 / 4  —  N O T E S : 1.  M A C H I N E  2.  R E F E R  DOWNFLOW U B C  G.D.F.  E A C H  T O  F L A N G E  D R A W I N G  C R - 6  SECTION S C A L E :  1 :  T O  2  A C C E P T F O R  # 2 2 8  F L A N G E  NO.1 NO.REQD:  0  RING.  D E T A I L  M A T L :  1  5/07/87  PMMA  CR-2  N O T E S :  DOWNFLOW U B C  G.D.F.  1.  M A C H I N E  2.  R E F E R  3.  M O U N T  E A C H  T O A T  D R A W I N G 4 5 °  SECTION S C A L E :  1 ;  F L A N G E  2  T O  T O  C R - 6 PIPE  A C C E P T F O R  C E N T E R  NO. 2 N O . R E Q D :  1  # 2 2 8  F L A N G E  0  RING.  D E T A I L .  LINE.  M A T L :  5/12/87  PMMA  CR-3  148  1 / 4  5  1 / 2  1 / 4  -LU.  N O T E S : 1.  M A C H I N E  2.  R E F E R  T O  O N E  F L A N G E  D R A W I N G  DOWNFLOW S E C T M A T E R I A L :  U B C  T O  C R - 6  A C C E P T F O R  NO.3  PMMA G.D.F.  3/15/89  # 2 2 8  F L A N G E  0  RING.  D E T A I L  S C A L E :  1:1  N O . R E Q D :  CR-4  1  149  i — r  1 / 4  1  3 / 8  N O T E S : 1.  M A C H I N E  2.  R E F E R  F L A N G E  T O  T O  D R A W I N G  DOWNFLOW S E C T M A T E R I A L :  U B C  A C C E P T C R - 6  F O R  NO.4  PMMA G.D.F.  # 2 2 8  5/07/87  0  RING.  F L A N G E  DETAIL.  S C A L E :  1 ;  N O . R E Q D :  CR-5  1  1  150  2"  FLANGE  PMMA  M A T E R I A L :  U B C  DETAIL  G.D.F.  5/12/87  S C A L E :  1 :  N O . R E Q D :  CR-6  1  —  151  1 / 8  0 2  N P T  1 / 2  1  1 / 4 U N F 1 2  N O T E  2 N O T E  1  N O T E S : 1.  M A C H I N E  2.  F A B R I C A T E INSIDE  F A C E  A C C E P T  T H R E A D E D  P I P E  W A L L  HORIZONTAL  A N D  P L U G T O  #221 T O S E A L  SECTION  PMMA  M A T E R I A L :  U B C  T O  G.D.F.  5/08/87  0  RING.  SIT O N  F L U S H 0  WITH  RING.  S C A L E :  1;  NO.REQD:  CR-7  1  1  152  1  / 2  2  1 / 2  N O T E S : 1.  R E F E R  2.  M A C H I N E  UPFLOW  T O  D R A W I N G F L A N G E  T O  SECTION  F O R  A C C E P T  G.D.F.  F L A N G E  # 2 4 4  NO.1  PMMA  M A T E R I A L  UBC  CR—11  5/05/87  0  D E T A I L RING.  S C A L E :  1 ;  N O . R E Q D :  CR-8  1  1  1 / 2  —  18  —  1 / 2  N O T E S : 1.  M A C H I N E  2.  R E F E R  UPFLOW U B C  G.D.F.  SCALE:  E A C H  T O  F L A N G E  D R A W I N G  C R - 1 1  SECTION 1;  2  T O  A C C E P T F O R  # 2 4 4  F L A N G E  0  RING.  D E T A I L .  NO.2  N O . R E O D :  1  MATL:  5/05/87  PMMA  CR-9  154  1  1  3/  4  i—r T  T—1  — r T L I  1  1  1 LJ  3 / 4  1  1  1 / 4  —  N O T E S ; 1.  R E F E R  OUTLET  T O  D R A W I N G  F O R  F L A N G E  D E T A I L  SECTION  S C A L E :  PMMA  N O . R E Q D :  M A T E R I A L :  U B C  C R - 1 1  G.D.F.  4/30/87  1 :  2  1  CR-10  155  4"  FLANGE  PMMA  M A T E R I A L :  U B C  DETAIL  G.D.F.  5/05/87  S C A L E :  1 :  N O . R E Q D :  CR-11  2  -  N O T E S : 1.  2  H O L E S  0 1 / 1 6  2.  L O W E R  3.  A S S E M B L Y  S U P P O R T  F L A N G E 4.  U P P E R  S C A L E :  IS  30'  TO  O U T L E T  D R A W I N G S E A L E D  V A C U U M  C E N T E R  R E M O V A B L E  O V E R  T R O U G H  DISCHARGE G.D.F.  FITS  ( S E E  S I L I C O N E  U B C  A T  2  F R O M  S E C T I O N  U P P E R A N D  T R O U G H . R E S T S  O N  C R - 1 0 ) . T O  O U T L E T  S E C T I O N  WITH  G R E A S E .  LAUNDER 1 :  LINE.  N O . R E Q D :  M A T L :  1  4/30/87  PMMA  CR-12  157  2X  0 3 / 4  05  4 X  3 / 4  0 7 / 6 4  E Q U A L L Y S P A C E D 2 X  1 / 8 N P T  0 1 / 8  0 1 / 4  OUTLET M A T E R I A L :  U B C  COVER PMMA  G.D.F.  3/09/89  S C A L E :  N O . R E Q D :  1 ;  2  1  CR-13  E Q U A L L Y  S P A C E D  N O T E S : 1.  M O U N T LIP  S T A I N L E S S  S E A L  2 5 1 3 C R  MIX- R N O . 2 U B C  G.D.F.  S C A L E :  S T E E L O N  B E A R I N G  O U T S I D E  O F  K L N J 0 0 4 — Z R S 0 3 / 4  X  VIOUN" 1 :  1  N O . R E Q D :  1 / 2  O N  INSIDE  D E E P  H O L E .  M A T L :  1  8/17/89  A N D  VITON  PMMA  CR-14  MIXER M A T E R I A L :  U B C  N0.1 1/8"  BUSHING P T F E SHEET  G.D.F.  6/04/87  S C A L E :  N O . R E Q D :  1 :  1  1  CR-15  160  MIXER M A T E R I A L :  U B C  NO.3 1/8"  BUSHING  PTFE SHEET  G.D.F.  3/09/89  S C A L E :  1:  N O . R E Q D :  CR-16  1  1  161 P A R T N O T E : 1.  ITEMS  IN  B E A D E D  LIST G L A S S  C O M P O N E N T S ; S C H O T T S Y S T E M S  A R E P I P E O - l  P R O C E S S INC.  /  D E S C R I P T I O N  A  2"  90*  B  2"  S A N I T A R Y  C  4 " - 2 "  D  2"  B E A D - P L A I N  E  2"  B E A D - B E A D  F  4"  B E A D - P L A I N  R E Q D 1  E L B O W T E E  1  R E D U C E R  1 C P L G C P L G C P L G  4 1 1  C R - 1  nn  C R - 1 3  C R - 1 2  C R - 1 0  C R - 2  C R - 9 C R - 3  C R - 1 4  C R - 7  REACTOR U B C  ASSEMBLY  G.D.F.  5/15/87  S C A L E :  N  CR-17  TS  DILUTION U B C  G.D.F.  SCALE:  VAT 1: 2  MATL: NO.REQD:  1  1/06/87  PMMA  DV-1  CONSTANT-HEAD U B C  G.D.F.  S C A L E :  1 :  TANK 2  N O . R E Q D :  M A T L :  1  5/11/89  PMMA  HT-1  4X  0 7 / 3 2  8 X  R 1 / 8  1  1  9 / 1 6  N O T E S : 1.  R E F E R  T O  2.  R E M O V A B L E  3.  R E F E R  T O  D R A W I N G W A L L  P W - 4 S E A L E D  D R A W I N G  P W - 5  F O R T O  T A B L E T A B L E  F O R  M O T O R  D E T A I L WITH  S I L I C O N E  M O U N T  V A C U U M  DETAIL.  G R E A S E .  j —  1 3 / 1 6  165  1  1/4  DECKER M A T E R I A L :  U B C  HEADBOX PMMA  G.D.F.  8/05/87  S C A L E :  1:1  N O . R E Q D :  PW-2  1  166  WASHER M A T E R I A L :  UBC  HEADBOX PMMA  G.D.F.  8/07/87  SCALE:  1: 1  N O . R E Q D :  PW-3  1  167 CD CD  \  CN  1  oo  cr  m  oo  oo  <  1  Q_ O O O O O O O  o.oooooc> oooooo : oooooO" O O O O O O O ) O O O O O O " O O O O O O O  : ooooocx  oo  oo  1  O O O O O O O  in  O ) O ) O  ro  )000000"  O O O O O O O O O O O O " O O O O O O O O O O O O " O O O O O O  I— <  )000000" OOOOOOO-  Olo o o o o o o O O O O O O " OOOOOOCi  CN  )000000" O O O O O O O )000000"  O O O O O O O ) O O O O O O o O O O O O O O  o  >O0OOO0"  LaJ  O O O O O O O  cr d  T \—  CN  CN  CN  <  $0000000  > O O O O O O " O O O O O O O ) O O O O O O " O O O O O O O >0 O O O O O " O O O O O O O ^ O O O O O O " O O O O O O O IP O O O O O " O O O O O O O ) O O O O O O " O O O O O O O  loopooo>  O ) O o )O ) O  00  1  UJ I  cr  ro  < o  CO  ro  O O O O O O O O O O O O " O O O O O O O O O O O O " O O O O O O O O O O O O " O O O O O C i  oo  CN  ci CD  oo \—  cr  o  pq  168  4 X  4 U N C 4 0  /  II II II II 1  1 / 8  —j  - A  i-  —  I  ^  I  -—  1  K  I  1 / 4  "  Y  II  II II  1 1 / 3 2  —  —I 1 / 2 LU 1  //TV  k5 / 1 6  1 / 2  0 3 / 8 3 / 8 1  1 / 2  3 X  1 0 U N F 3 2  5 / 1 6  h —  MOTOR M A T E R I A L :  UBC  1  1 7 / 3 2  MOUNT PMMA  G.D.F. •  7/31/87  S C A L E :  1:  N O . R E Q D :  PW-5  1  2  169  4X  0 5 / 3 2  S A M E A S  8 X  R H S  0 1 / 2  N O T E : 1.  MOTOR M A T E R I A L :  U B C  S P O T  W E L D  S E A M S .  HOUSING  24 GAGE 316 SS G.D.F.  8/02/87  S C A L E :  1;  N O . R E Q D :  PW-6  2  2  01/4  01  1/2  N O T E : 1.  C O V E R  DRIVE-DISCHARGE U B C  G.D.F.  S C A L E :  1:  1  WITH  R U B B E R  INNER  ROLL N O . R E Q D :  T U B E .  M A T L :  4  7/31/87  PMMA PW-7  N O T E : 1.  PRESS U B C  G.D.F.  S C A L E :  316  S S  C O R E  WITH  P T F E  ROLL 1;  1  S H E L L .  M A T L :  N O . R E Q D :  4  7/31/87  SEE NOTE PW-8  172  N O T E : 1.  R E Q U I R E R H  S H O W N ,  DOCTOR  R H LH  A N D  2  L H  G.D.F.  M O U N T S ;  O P P O S I T E .  MOUNT PMMA  M A T E R I A L :  U B C  2  8/01/87  S C A L E :  1;  N O . R E Q D : S E E  PW-9  1  N O T E  N  T O P  DECKER M A T E R I A L :  U B C  —  1  / 4 N P T  VIEW  SEAL  TANK  PMMA G.D.F.  2 X  1/20/87  S C A L E :  1 :  N O . R E Q D :  ST-1  1  1  174  T O P  VIEW  N O T E S : 1.  DRILL  2.  R E F E R  0 7 / 3 2 T O  U B C  A T  D R A W I N G  WASHER M A T E R I A L :  H O L E  4 5 ° .  ST-1  F O R  SEAL  C O M P L E T E  TANK  PMMA G.D.F.  1/20/87  D I M E N S I O N S .  S C A L E :  1:  N O . R E Q D :  ST-2  1  1  APPENDIX B DETERMINATION O F STEADY STATE  Approximate values of the time required for the apparatus to reach a steady state were determined with the use of a simple dynamic model of the process. The theoretical model was based on an unbleached pulp flow of 6 odg/min and the addition of chlorine at a rate of 0.3 g/min. It was assumed that all of the chlorine added forms soluble chloride ion immediately upon being mixed with the unbleached pulp. The model enabled the concentration of dissolved chloride to be predicted as a function of time from the initial addition of chemical. A schematic representation of the process as it was modeled is given in figure 15. There were six components of the model: 1.  Reactor. The 7.5-1 chlorination reactor was modeled with plug flow at 2.0% consistency. Thus it was represented with a dead time of 25 minutes.  2.  Dilution zone. The 0.7-1 dilution zone at the outlet of the reactor was modeled as a continuous stirred tank reactor at approximately 0.5% consistency.  The  dilution flow from the chlorination seal tank to the dilution zone was 840 ml/min. 3.  Dilution vat. The 1.0-1 dilution vat was also modeled as a continuous stirred tank reactor.  175  W A S H  W A T E R  U N B L E A C H E D P U L P  P U L P  DILUTION  DILUTION  Z O N E  V A T  W A S H E R  W A S H E R  S U C T I O N  S U C T I O N  B O X  #1  B O X  E  T O  S T A G E  #2  J__L S E A L T A N K  f  Figure 15.  E F F L U E N T  Block diagram of the chlorination process model  ON  177 4.  Washer suction box no.l.  Chloride approaching the first suction box was  simply proportioned according to the existing filtrate flows, which were dictated by a pulp discharge consistency of 10%. 5.  Washer suction box no.2.  The concentration of chloride in the washed pulp  leaving the second suction box at 10% consistency was assumed to be 25% of that in the thickened pulp leaving the first suction box (Singh 1979, 643). This was accomplished by the addition of pure wash water at a rate of 60 ml/min. 6.  Seal tank.  The 1.2-1 seal tank was assumed to function with plug flow.  However, this element was ignored because the dead time of 1 minute was very short when compared to that of the reactor. The model was developed from dynamic chloride balances performed around both the dilution zone and the dilution vat. This analysis gave  dC  v  -Fr  DZ  dt  v DV  D ST  °  d C  £  r  =F C DZ^DZ  r  -F r  (13)  DZ~DZ  r  C  ( ) 14  DV^DV  where V is the volume in 1, F is the flow in l/min, and C is the concentration in g/1. The subscripts DZ, DV, R, D, and ST stand for dilution zone, dilution vat, reactor, dilution, and seal tank respectively.  Further mass balances performed around the  reactor, the seal tank, and the washer suction boxes enabled both C and R  replaced by functions involving C , m  to be  178 The system of two simultaneous differential equations in the two unknowns, C  DZ  and C , was solved analytically using the Laplace transform. DV  The initial  condition of zero chloride concentration was used in each of two solutions of the system. The first case involved chlorinationfiltraterecycle for brown stock dilution at the maximum level used in the experimental runs of 22.5 m /adt (150 ml/min). This 3  resulted in  C  = (1.82+0.0703e- 2  S T  70(r  - -1.89e^ »- ^)iy(r-25) 25)  5  2  (15)  which gives the seal tank chloride concentration as a function of the time t in minutes from initial chlorine addition.  H ( t - 2 5 )  is the Heaviside unit step function. The second  case involved zero recycle of chlorination filtrate for brown stock dilution and resulted in  = (0.952+0.0789e - - <'" ) -1.03e 2 57  C  S  T  25  d ) 6  ^  W  ^  W  f  -  l  S  )  Both of these responses are shown in figure 16. If a steady state is approximated when the concentration of chloride reaches 95% of its ultimate value, then 56 minutes are required for the system to reach equilibrium under conditions of maximum recycle while 41 minutes are required without recycle. The results of the model at zero recycle were tested by measuring the response of the actual process to a step change in chlorine addition. The filtrate's concentration change was measured with the seal tank pH while the pulp's response was measured as changes in the CE microkappa number (TAPPI 1990, UM 246) at the outlet of the dilution vat.  The lab-scale bleaching apparatus was operated under the conditions  179 2  I  1  1  1  1  1  1  1  r  0  10  20  30  40  50  60  70  80  90  Time from start of chlorine addition (min) Figure 16.  Response of the process model to a step change in chlorine addition  180 previously specified for the model, and the procedures used were similar to those of the factorial experimentation. Both the filtrate and the pulp responses are shown in figure 17.  The pH  reached 95% of its ultimate value in 37 minutes while the microkappa number required 40 minutes. These values compare well with the 41-minute value obtained from the model at zero recycle. This agreement gives credence to the 56-minute result from the model at maximum recycle.  Accordingly, prior to sampling in the  experimental runs, a 90-minute period was allowed after the initial addition of bleaching chemical to enable a steady state to develop. Further analysis of the pulp response in figure 17 gave a 26-minute mean residence time for the volume composed of the reactor, the dilution zone, and the dilution vat. This is identical to the theoretical value calculated from the appropriate volumes and flows. In addition, the dimensionless variance of the pulp response was found to be 0.09. Thus the pulp flow in the reactor is closely approximated by plug flow. The strong affinity of pulp fibers for hydrogen ions prevented a similar analysis of the filtrate response from yielding meaningful results.  181  APPENDIX C CHLORINATION STAGE OPERATING DATA  182  183  Table 21 Chemical concentrations and residuals Run  Cl solution concentrations (g/D  C10 solution concentrations (g/»  2  High  Low  2  To fix flow  High  Low  To fix flow  CI,  Average residual (% of active  a, added)  0  6.18  1.69  6.18  3.26  2.98  3.26  0.16  0.45  1  5.24  4.97  5.17  3.74  3.69  3.74  0.08  0.23  2  7.13  7.05  7.09  5.58  5.43  5.53  0.10  4.60  3  6.37  6.26  6.33  3.05  3.02  3.04  0.03  0.24  4  5.54  5.13  5.45  2.89  2.73  2.85  0.09  0.27  5  6.68  6.49  6.56  5.35  5.22  5.32  0.20  0.24  6  7.11  7.03  7.07  5.27  5.18  5.20  0.18  0.20  7  6.92  6.88  6.90  5.34  5.31  5.32  0.26  0.21  8  6.90  6.88  6.85  5.38  5.11  5.31  0.25  0.20  9  7.19  7.06  7.15  6.97  6.84  6.85  0.46  0.97  10  6.97  6.86  6.97  5.50  5.44  5.50  0.48  0.36  11  6.97  6.90  6.97  5.55  5.44  5.50  0.45  3.05  12  7.77  7.75  7.76  5.32  5.14  5.32  0.42  0.28  13  7.65  7.55  7.65  5.16  5.10  5.16  0.41  1.72  14  7.01  6.88  7.00  5.73  5.62  5.65  0.18  0.33  15  7.01  6.97  6.97  5.71  5.68  5.71  0.21  1.68  16  6.40  6.34  6.37  5.57  5.40  5.57  0.23  3.58  17  6.50  6.43  6.50  5.45  5.35  5.41  0.12  0.24  18  6.44  6.28  6.44  5.35  5.31  5.35  0.13  0.22  19  6.09  5.98  6.00  5.14  5.05  5.10  0.14  0.55  20  6.69  6.66  6.68  4.87  4.69  4.75  0.10  0.28  21  6.67  6.61  6.66  4.90  4.86  4.88  0.16  0.25  22  6.14  6.11  6.14  4.73  4.63  4.70  0.10  7.38  23  5.88  5.83  5.85  4.33  4.16  4.20  0.28  0.46  184 Table 22 Pulp consistencies Stock tank (%)  Decker discharge (%)  Washer discharge (%)  0  0.388  10.0  12.3  1  0.368  10.8  12.5  2  0.367  9.8  12.4  3  0.376  9.9  11.0  4  0.397  10.3  11.9  5  0.388  10.5  11.4  6  0.364  10.1  11.8  7  0.353  9.8  12.2  8  0.362  9.9  11.3  9  0.359  9.8  13.3  10  0.365  9.5  12.2  11  0.365  10.6  11.5  12  0.364  10.0  12.5  13  0.361  9.9  12.1  14  0.363  9.8  11.6  15  0.363  9.9  13.3  16  0.383  10.0  12.7  17  0.381  9.5  12.6  18  0.391  9.3  11.7  19  0.366  11.1  12.2  20  0.389  10.2  12.7  21  0.389  10.4  13.3  22  0.362  10.7  13.2  23  0.362  10.5  12.6  Run  Table 23 Temperature and pH data Run  Average p H in seal tank  Average temperature (°Q Horizontal section  Reactor outlet  Seal tank  0  1.8  33  30  29  1  1.7  34  30  29  2  2.1  33  29  28  3  1.8  32  29  29  4  1.7  31  28  28  5  1.6  33  30  29  6  1.7  33  30  30  7  1.8  32  31  30  8  1.8  33  31  30  9  2.0  33  30  30  10  2.2  33  31  31  11  2.2  34  31  31  12  1.8  33  30  30  13  2.0  33  31  31  14  1.9  33  31  30  15  2.2  34  31  31  16  2.0  33  31  30  17  2.1  33  31  31  18  2.0  34  31  31  19  1.6  33  31  31  20  1.8  35  32  31  21  1.9  34  31  31  22  2.1  35  32  32  23  1.7  34  32  31  186  Table 24 Peristaltic pump flows Brown stock  Ru  Chlorine  Chlorine dioxide Setpaint Oow (ml/ min)  Setpoint Oow (ml/ min)  A after ran  -22  30  6.32  +0.2  20.8  +1.5  Setpolnt flow (ml/ min)  A after ran  -55  40.6  -9.1  5.99  1630  -0.8  48.1  +1.5  1635  0.0  10.7  +8.2  Setpoint Dow (ml/ min)  A after run  0  1546  1 2  (*)  (*)  Dilution  Recycle  A after ran  (*)  Wash Setpout Oow (ml/ min)  A after ran  -25  60  -15  172  +1.2  60  +2.4  194  +0.1  60  -02  179  -03  60  +0.2  Setpoint Oow (ml/ min)  A after ran  -33  163  30  +2.0  30  +0.3  (*)  (*)  (*)  3  1596  -03  39.6  0.0  7.83  +0.8  30  +1.0  4  1511  -13  55.4  -1.8  10.1  +0.5  30  0.0  161  0.0  60  +0.7  5  1546  +1.1  39.0  -1.8  4.60  +0.7  150  +0.1  62  +0.3  60  +0.7  6  1648  -1.6  453  -33  5.80  +0.5  150  -0.2  55  0.0  60  -0.2  7  1700  -0.9  25.1  -1.6  12.4  +0.1  90  -1.7  129  -0.2  60  -05  8  1657  -1.4  37.4  -2.2  4.61  +0.4  150  -25  64  +0.8  60  +0.2  9  1671  -0.1  10.8  -0.8  17.1  0.0  150  -12  78  +0.1  60  +0.8  10  1644  -2.4  8.90  +1.3  173  +0.9  30  +1.0  200  +0.3  60  -12  +1.0  200  +0.3  60  -1.2  11  1644  -2.4  8.90  +1.3  173  +0.9  30  12  1648  -3.8  223  -73  123  +0.5  90  +0.1  131  -2.0  60  0.0  13  1662  -2.9  8.30  +7.7  18.7  +1.9  150  -1.7  79  +0.8  60  +0.3  14  1653  -2.0  44.4  -23  5.20  0.0  30  +1.7  176  -0.4  60  0.0  15  1653  -0.8  11.1  -1.4  20.7  +0.3  30  +1.7  194  -0.4  60  0.0  16  1567  -2.7  125  +0.2  21.8  -0.2  150  -13  -72  -25  60  -05  17  1575  -23  9.60  +1.1  175  +2.2  150  -1.6  79  -0.6  60  0.0  18  1535  •42  27.7  -4.1  12.7  +2.5  90  +0.2  126  -1.0  60  +0.3  19  1639  -2.7  51.2  +4.1  5.70  +1.5  150  -0.2  49  +0.6  60  -0.8  20  1542  -2.4  37.8  +0.8  5.00  +1.2  30  +0.7  183  -1.2  60  +1.3  21  1542  -2.4  37.9  +0.8  4.90  +1.2  30  +0.7  183  -12  60  +1.3  22  1657  -1.6  12.9  +1.5  255  +2.3  30  +0.7  188  -03  60  +0.3  23  1657  -0.7  54.0  -1.2  7.10  0.0  30  +0.7  165  -03  60  +03  Table 25 Variation in factor levels Run  Recycle  (m7adt)  Mode (sec)  Substitution  (%)  Multiple (%/ml)  0  4.60  43.9  42.4  0.068  1  4.63  40.0  20.2  0.177  2  4.51  41.6  78.7  0.222  3  4.56  -44.6  20.2  0.178  4  4.56  40.8  20.8  0.207  5  22.28  -45.0  19.9  0.177  6  22.82  -45.9  20.8  0.216  7  13.39  0.0  50.6  0.199  8  22.25  40.2  19.8  0.178  9  22.25  -40.6  80.6  0.222  10  4.66  -39.8  80.2  0.184  11  4.66  41.9  80.1  0.184  12  14.05  0.0  51.1  0.198  13  22.78  41.9  79.1  0.188  14  4.67  -45.6  20.8  0.217  15  4.61  -40.6  80.3  0.220  16  22.82  42.7  79.6  0.220  17  22.66  -40.4  80.1  0.185  18  14.12  0.0  52.1  0.203  19  23.08  40.5  19.2  0.235  20  4.64  40.3  20.3  0.186  21  4.64  -45.4  20.1  0.184  22  4.61  43.2  79.9  0.225  23  4.56  40.5  20.5  0.221  APPENDIX D DETAILED EXTRACTION PROCEDURE  1.  Weigh accurately the chlorinated pulp sample of known consistency.  2.  Weigh an amount of NaOH equivalent to 3% of the oven-dry weight of the pulp sample.  3.  Dissolve the NaOH in a volume of distilled water which will yield a consistency of 10% when added to the pulp sample.  4.  Add the NaOH solution to the pulp sample, and heat seal the mixture in a polyethylene bag.  5.  Knead the mixture to disperse the chemical evenly throughout the pulp.  6.  Immerse the bag in a water bath at 70°C for 90 minutes.  7.  Dilute the pulp to 1% consistency with distilled water, and mix the suspension mechanically for 5 minutes in a suitable vessel.  8.  Separate thefiltratefrom the pulp with a Buchner funnel, a filtrate flask, and a qualitative grade filter paper. Apply suction until all of the free liquor is drained. Remove the suction.  9.  Gently add distilled water at 70°C to the pulp pad in the funnel. Use a volume equivalent to the volume of filtrate associated with the pulp sample assuming a consistency of 10%.  10.  Apply suction again until all of the free liquor is drained.  11.  Mix the combined filtrate, and measure the pH. For preservation, adjust the filtrate's pH to 2.0 with concentrated HN0 . 3  12.  Repeat steps 7 and 8 discarding thefiltrateproduced. Allow the extracted pulp to air dry at constant temperature and humidity prior to testing.  188  APPENDIX E TOXICITY DATA  Table 26 Trout toxicities (LC50 procedure) Run  96-hr LC50 (% volume)  Slope  Average fork length (cm)  Average blotted weight (g)  Volume factor 0/(g.day))  0  12.7  196  5.5  1.45  0.52  1  11.8  274  5.5  1.58  0.47  2  75.1  81  5.6  1.71  0.44  3  8.9  199  6.1  2.03  0.37  4  12.2  343  6.2  2.21  0.34  8  "Plot of % mortality versus log % volume.  189  Table 27 Trout toxicities (LT50 procedure) Concentration (%vol)  LT50 (min)  95% confidence limits (min)  Slope function*  Average fork length (cm)  Average blotted weight (g)  0/fe-day))  5  100  211  158-282  1.329  3.5  0.41  28  6  100  167  152-183  1.161  3.7  0.46  36  7  100  ND"  300-660  ...  3.7  0.50  11  8  100  424  387-464  1.157  4.0  0.65  11  9  100  319  277-367  1.236  4.1  0.70  11  10  100  478  429-532  1.189  4.3  0.81  6.7  11  100  192  126-292  1.883  4.4  0.91  1.8  12  100  393  375-411  1.077  4.7  1.07  7.3  13  100  693  522-921  1.499  4.9  1.30  1.7  14  100  114  96-136  1.325  4.3  0.92  17  15  100  125  99-158  1.442  4.6  1.10  5.2  16  100  387  250-598  2.016  4.5  1.06  3.8  17  100  679  634-727  1.117  5.3  1.70  2.2  18  100  318  280-361  1.212  5.4  1.69  4.7  18  62  729  688-773  1.099  5.2  1.55  2.9  18  38  1256  1132-1394  1.116  5.4  1.59  1.3  19  100  149  138-161  1.136  5.2  1.62  11  19  62  268  252-285  1.104  5.6  1.97  5.7  19  38  482  467-497  1.051  5.2  1.56  4.1  20  100  397  379-415  1.076  4.8  1.24  6.2  21  100  174  154-197  1.218  4.8  1.20  14  22  100  486  422-560  1.249  4.4  0.94  5.4  23  100  202  180-227  1.206  4.3  0.79  17  23  62  391  368-416  1.098  4.2  0.76  11  23  38  641  612-671  1.078  4.5  0.98  4.9  100  95  90-101  1.100  4.9  1.18  23  Run  Celgar  "Plot of % mortality (probability) versus log time. 'Not determined.  Volume factor  191 Table 28 Bacterial toxicities Chlwuiation filtrate  Run  Extraction filtrate  EC50" (% volume)  95% confidence limits (% volume)  Slope*  EC50" (% volume)  95% confidence limits (% volume)  Slope"  0  11.4  10.1-12.8  0.723  773  57.8->100  0.836  1  15.2  12.9-17.9  0.781  673  58.6-77.4  0.824  2  115  113-11.7  0.723  993  583->100  1.430  3  16.6  15.9-17.4  0.811  63.0  46.1-86.2  1.031  4  12.6  11.9-133  0.659  555  385-80.2  1.202  5  8.6  7.6-9.8  0.788  92.2  58.0->100  1.269  6  5.1  3.5-7.5  0.849  565  39.0-81.8  0.691  7  12.2  10.2-145  0.866  54.0  45.7-63.9  0.682  8  7.2  5.9-8.8  0.921  55.1  365-833  0.884  9  9.9  8.0-12.2  1.224  >100  _  1.160  10  15.4  14.2-16.6  0.697  52.1  447.-60.7  0.841  11  7.7  6.0-9.9  1.638  61.8  38.2-99.8  1.286  12  12.8  9.2-17.8  1.176  63.7  40.9-99.4  1.024  13  12.6  10.7-14.8  1.075  60.0  45.8-78.6  0.908  14  143  12.8-15.9  0.724  625  455-85.9  0.929  15  24.1  213-27.2  0.864  95.9  54.0->100  0.973  16  16.9  11.9-24.0  1.017  >100  —  1.400  17  15.1  115-19.8  0.991  >100  ...  0.880  18  145  12.1-17.4  0.840  74.1  52.8->100  1.129  19  6.4  4.9-8.3  0.709  543  47.7-61.9  1.089  20  13.4  105-17.1  0.892  50.6  27.8-922  1.041  21  9.7  8.2-115  0.821  63.4  35.6->100  1.151  22  6.4  2.3-17.8  0.445  725  40.8->100  0.951  23  10.4  8.8-123  0.859  55.6  49.9-61.8  1.139  4.4  4.1-4.8  0.701  20.4  152-27.4  0.775  Celgar  '15-minute, 15°C, color corrected values. "Plot of log gamma versus log % volume.  1  APPENDIX F TOXIC EMISSION FACTOR DETERMINATION F R O M LT50 DATA  The toxic emission factor is defined by  TEF = ^-xQ LC50  (17)  where LC50 is the median lethal concentration of the effluent in volume % and Q is the flow of effluent rated for pulp production in m /adt. When fish are used as the 3  test organism for toxicity, the 96-hour LC50 is usually applied. The toxic emission factor can be described as the total dilution required to make an effluent nontoxic. For instance, an effluent produced at 20 m /adt having an LC50 of 10% would become 3  nontoxic when diluted to 200 m /adt. In theory, the LC50 is directly related to the 3  volume at which an effluent is produced.  Thus, if an effluent's flow is doubled  through dilution, its LC50 would also be expected to double thereby maintaining a constant toxic emission factor. The use of toxic emission factors is desirable for the following reasons: 1.  They increase as toxicity increases, unlike LC50 values.  2.  They enable the toxicity of effluents produced at different rates of water consumption to be compared.  192  193 3.  When used as the basis for regulation, they make attempts to meet requirements through effluent dilution futile. The fundamental relationship between the concentration (C) of an effluent and  the corresponding median lethal time (LT50) is given by  logC = mlogLT50+b where m and b are the slope and intercept respectively.  < ) 18  When the concentration is  100%, the resulting median lethal time is denoted LT50'. Thus  100 = 10 i O 6  mlogL750  '  (19)  When the LT50 is 5,760 minutes, the concentration equals the 96-hour LC50. Thus LC50 = 10* lO"" ' 1085  760  ( °) 2  When equations 19 and 20 are substituted into equation 17, it becomes  M  ..ig-"' lQmlog5,760  ( 2 1 )  ^  Thus the toxic emission factor can be calculated from the LT50 at 100% concentration and m, the slope of the log C versus log LT50 plot. The slope was calculated for runs 18, 19, and 23 by determining median lethal times for effluent concentrations of 100, 68, and 32%. The results as calculated by linear regression analysis are given in table 29. The average slope of -0.78 was used in equation 21 to determine the toxic emission factors for the chlorination effluents of runs 5-23 and Celgar Pulp. average slope.  Some sensitivity was undoubtedly lost by using the  However, for the mean LT50 of 328 minutes, the toxic emission  194 Table 29 Slopes of the log C versus log LTSO plots Run  Slope  Coefficient of determination (r ) 2  18  -0.69  0.98  19  -0.82  >0.99  23  -0.83  0.99  factors that were calculated using the slopes farthest from the average (-0.69 and -0.83) were within the 90% confidence limits of the toxic emission factor that was calculated using the average value. The data of Leach and Thakore (1976) supports the idea that a single slope defines a particular type of effluent.  In this study of six industrially produced  chlorination effluents representing a wide range of pulp species, bleaching sequences, and washing techniques, the slope ranged from -0.60 to -0.80 and averaged -0.70. These values are in good agreement with the slopes found here. A comparison between this novel method of determining the toxic emission factor and the conventional approach is warranted. One of the difficulties with the 96hour LC50 test is the depletion of toxicant with time due to both uptake by fish and degradation. Thus guidelines that define a minimum effluent volume factor rated on the bases of both fish weight and fish exposure have been developed for pulp mill effluents.  For static bioassays, the factor was determined to be 0.5 l/(g»day) (Walden,  195 McLeay, and Monteith 1975). Typical experimental conditions make this requirement difficult to meet because of limitations on the availability of both effluent and suitably sized fish.  Continuous-flow bioassays allow the use of lower volume factors.  However, these techniques are only useful for overcoming the problem of fish size as they require more effluent than static bioassays. The determination of the median lethal time for concentrations significantly higher than the 96-hour median lethal concentration ensures that mortality occurs rapidly. This minimizes the total time required to assess toxicity while substantially increasing the volume factor.  With reference to this study, runs 0-4 which were  performed using 96-hour LC50 tests had an average volume factor of 0.43 l/(g«day). The use of LT50 tests at a concentration of 100% in runs 5-23 caused the factor to increase to an average of 9.2 l/(g»day).  This occurred even though the LC50 tests  were performed at a volume of 30 1 while the LT50 tests were performed at 25 1. The use of volume factors less than 0.5 l/(g»day) in static bioassays with fish causes the resulting 96-hour LC50 to be higher than the true value.  Because of  toxicant depletion, higher concentrations of effluent are required to maintain toxic levels of constituents.  Accordingly, extrapolated median lethal time data yields lower  LC50 values than median lethal concentration data when the critical level is not met. The work of Leach and Thakore (1975) supports this.  The toxicities of five  compounds, which are responsible for a significant portion of the toxicity of extraction filtrate, were determined statically using both LC50 and LT50 tests.  The volume  factor was 0.25 l/(g»day) for the LC50 tests and 1.0 l/(g»day) for the LT50 tests. As  196 a result, the 96-hour LC50& of the five compounds determined by extrapolation of LT50 data were on average 41% lower than the values determined by standard methodology. Similarly, three of the four replicate tests in this study had larger toxic emission factors from LT50 tests than from LC50 tests. In summary, estimation of the 96-hour LC50 through extrapolation of LT50 data appears to be a promising technique. Maximum sensitivity is ensured through the use of high effluent concentrations. This in turn causes the response to occur rapidly, which eliminates the problem of toxicant depletion. However, research either in the literature or in the laboratory needs to be conducted to confirm the assumption that LT50 values determined for concentrations near the 96-hour LC50 follow the same slope as LT50 values determined for much higher concentrations. Theory predicts they should as long as the volume factor is maintained at proper levels.  

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.831.1-0058684/manifest

Comment

Related Items