UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The kinetics zinc removal from cobalt electrolytes by Ion exchange Swami, Nathan S. 1993

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

Item Metadata


831-ubc_1994-0051.pdf [ 2.26MB ]
JSON: 831-1.0078447.json
JSON-LD: 831-1.0078447-ld.json
RDF/XML (Pretty): 831-1.0078447-rdf.xml
RDF/JSON: 831-1.0078447-rdf.json
Turtle: 831-1.0078447-turtle.txt
N-Triples: 831-1.0078447-rdf-ntriples.txt
Original Record: 831-1.0078447-source.json
Full Text

Full Text

THE KINETICS OF ZINC REMOVAL FROM COBALT ELECTROLYTES BY IONEXCHANGEbyNathan SWAMI S.A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE STUDIESDepartment of Metals and Materials EngineeringWe accept this thesis as conformingTHE UNIVERSITY OF BRITISH COLUMBIAOctober 1993to the required standard© Nathan Swami S., 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Cokimbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of hJ’The University of British ColumbiaVancouver, CanadaDate O VA3DE-6 (2/88)ABSTRACTThe removal of trace zinc concentrations from the INCO (Port Colborne) cobalt advanceelectrolyte by solvent impregnated ion exchange, was studied in column and batch tests. The solventimpregnated resins containing the extractants D2EHPA, Cyanex 272 and Cyanex 302, were compared in terms of zinc loading and selectivity.D2EHPA impregnated OC 1026 resin demonstrated superior zinc loading and selectivity characteristics, but retained objectionably high amounts of cobalt, which were lost in the zinc elutionprocess. Cobalt loading, was found to be closely related to the electrolyte pH drop across the columnand could be reduced by a resin pre-treatment with the advance electrolyte at a pH of 3 or by anincrease in the feed electrolyte pH to 5, along with operation at a temperature of 40°C and a flowrate of 10 BV/hr.; all of which act to diminish the pH drop.The kinetics of zinc loading on each of the resins was found to be comparable, and the rate controllingmechanism in batch tests was found to be particle diffusion in the first fifteen minutes, while filmdiffusion became rate controlling at later time intervals. Pre-treatment enhanced the diffusioncoefficient inside the resin phase by nearly an order of magnitude, improved exchange kinetics byallowing a lower pH reduction during the loading process, and improved Co2/Zn exchange in amathx of the cobalt complex.An analysis of the breakthrough curves for the resins was done to determine the mass transfercoefficients inside the column, and a range of other parameters useful in the design of ion exchangecolumns. The rate controlling regime in the column was a mixture of the particle and film diffusionsteps, with the former being the dominant control mechanism at the operating flow rate.Further work is needed in the X-ray microprobe analysis of resin samples from the top, middle andbottom portions of the columns, and of samples from batch tests, to aid in understanding themechanism of ion exchange. The use of zinc selective electrodes in batch tests could also beundertaken to obtain a more accurate estimate of the diffusion and mass transfer coefficients.11Table of ContentsAbstract iiTable of Contents iiiList of Tables viTable of Figures ViiList of Symbols ixAcknowledgements xii1 CHAPTER ONE - INTRODUCTION 11.1 Ion Exchange Purification 11.2 Zinc Removal from Cobalt Electrolytes 12 CHAPTER TWO - LiTERATURE REVIEW 32.1 Ion Exchange Process 32.2 History and Technological Development of Ion Exchange 32.3 Present Applications 42.4 Chemistry of Ion Exchange 42.4.1 The resin matrix 52.4.2 Functional groups Strong and weak acid ion exchangers Strong and weak base ion exchangers 82.4.3 Aqueous chemistry of metal ions 82.4.4 Guidelines for ligand selection 112.5 Ion Exchange Technology 132.6 Kinetics of Ion Exchange 152.6.1 Mechanism of ion exchange 152.6.2 Rate determining step 182.6.3 Rate laws of ion exchange Ion exchange systems not involving chemical reactions Ion exchange accompanied by chemical reactions Electric potential gradient and the Nernst-Planck equation Systems involving diffusion and reaction Progressive shell mechanism Further developments in kinetic studies 272.7 Experimental Methods in Ion Exchange Kinetic Studies 272.7.1 Selection of the apparatus for rate studies 272.7.l.lThebatchsystem 282.7,1.2 Limited bath method [49,50] 281112.7.1.3 Indicator method [51,52]. Shallow bed technique 302.7.1.5 Photographic method [36,37] 302.7.1.6 Cone model method [53] 312.7.1.7 Laboratory column 312.7.2 Monitoring techniques and operating conditions 322.8 Solvent Impregnated Resins 352.9 Cobalt Purification 362.9.1 Introduction 362.9.2 Cobalt Hydrate Plant 372.9.3 Electro-Cobalt Plant 372.10 Extractants for Zinc 402.10.1 Di (2-ethyihexyl) phosphoric acid 412.10.1.1 Chemistry 412.10.1.2 The Zn-D2EHPA system 432.10.1.3 Applications 442.10.2 Cyanex 272 - The phosphinic group 462.10.3 Cyanex 302 - The thio-phosphinic group 472.11 Scope and Objectives of this Work 493 CHAPTER THREE - EXPERIMENTAL 503.1 Resins 503.1.1 Structure of the resins 503.1.2 Particle size and distribution 503.1.3 Physical properties 503.2 Reagents 523.2.1 Cobalt advance electrolyte 523.2.2 Zinc - free cobalt electrolyte 533.2.3 Cobalt sulphate solution 533.2.4 Zinc sulphate solution 533.2.5 Sodium hydroxide 533.2.6 Sulphuric acid 533.3 Experimental Apparatus 533.4 Experimental Procedure 553.4.1 Column tests 553.4.2 Batch tests 573.5 Solution Analysis 57iv3.6 Reproducibility of the Results 584 CHAPTER FOUR - KINETIC MODELS 594.1 Fickian Film Diffusion Model 594.2 Nernst-Planck Film Diffusion 624.3 Fickian Particle Diffusion Model 624.4 Unreacted Core Model 634.5 Kinetic Modelling of Column Tests 644.6 Minimum Superficial Velocity of Solution in Column Operations 685 CHAPTER FIVE- RESULTS AND DISCUSSION 695.1 COLUMN TEST RESULTS 695.1.1 Zinc loading on untreated OC 1026 695.1.2 Zinc loading on pre-treated OC 1026 715.1.3 Effect of pre-treatment on zinc loading and selectivity 735.1.4 Cobalt and zinc retained on pre-treated resin 765.1.5 Effect of flow rate on zinc loading and selectivity 785.1.6 Effect of temperature on zinc loading and selectivity 795.1.7 Effect of increasing feed electrolyte pH 795.1.8 Performance of Cyanex 302 and Cyanex 272 815.1.9 Comparison of cobalt retained on the resins 815.2 BATCH RESULTS 845.2.1 Kinetics of zinc loading on OC 1026 845.2.2 Effect of pre-treatment 845.2.3 Effect of initial zinc content 865.2.4 Effect of temperature on the kinetics of metal uptake 865.3 Kinetic Model Fit for the Batch Tests 865.4 Kinetic Modelling of the Column Tests 945.5 Mechanism of the Exchange 996 CONCLUSIONS AND SUGGESTIONS FOR FURTHER WORK 1036.1 Summary of the results 1036.2 Conclusions 1056.3 Suggestions for further work 106REFERENCES 107Appendix A: Raw Experimental Data of Column Tests 113Appendix B : Raw Experimental Data of Batch Tests 127VList of TablesTable 2.1: Classification of ion exchange resins 7Table 2.2: Classification of Lewis acids 9Table 2.3 : Crystal field stabilization energy of d complexes 12Table 2.4 : Dependence of ion exchange rate on experimental conditions 20Table 2.5 : Comparison of experimental techniques for kinetic studies 32Table 2.6 : Plating conditions for cobalt electrowinning 40Table 2.7 : Analyses of process streams at various stages in the flowsheet 40Table 2.8 : Calculated equilibrium aqueous phase exiractant concentration 43Table 2.9 : Extraction constants of various metal complexes 44Table 3.1: Properties of the Levextrel resins OC 1026, Cyanex 272 and Cyanex 302 52Table 3.2: Analysis and properties of the advance electrolyte 52Table 5.1 : Zinc loading on untreated OC 1026 71Table 5.2 : Effect of pre-treatment on zinc loading and pH profiles 73Table 5.3 : Analysis of INCO electrolyte 75Table 5.4 : Effect of pre-treatment on selectivity of OC 1026 75Table 5.5 : Comparison of metal retained on resin after various effluent bed volumes 76Table 5.6 : Effect of flow rate on zinc loading and selectivity 78Table 5.7 : Effect of temperature on zinc loading and selectivity 79Table 5.8 : Effect of increasing feed electrolyte pH on zinc loading and selectivity 79Table 5.9 : Comparison of cobalt retained on the resins 81Table 5.10: Kinetic model fitted to determine the D and km values 89Table 5.11: Effect of temperature on D and k. values 91Table 5.12 : Determination of the rate controlling mechanism from the activation energy 94Table 5.13 : Comparison of characteristic parameters of the breakthrough curve for theresins 95Table 5.14: Composition of electrolyte and operating conditions of Strong et. al test 95Table 5.15 : Kinetic parameters in the design of ion exchange columns 95Table 5.16: Comparison of the mass transfer coefficients obtained by various methods.. 98viTable of FiguresFigure 1.1 Extractants in OC 1026, Cyanex 272 and Cyanex 302 2Figure 2.1(a) Preparation of a crosslinked resin polymer matrix 5Figure 2.1(b) Surface characteristics of ion exchange resins 6Figure 2.2 The ion exchange process 6Figure 2.3 The periodic table in steric perspective 10Figure 2.4 Characteristic rate constants for substitution of inner sphere ofH20 10Figure 2.5 Co-ordination copolymers- donor atoms in one ligand 12Figure 2.6 Crystal field splittings of the d-orbitals 12Figure 2.7 Fundamental questions in kinetic investigations 17Figure 2.8 Rate determining steps 17Figure 2.9 Electric transference acts to equalize the net flux 24Figure 2.10 Schematic diagram of the batch process 29Figure 2.11 Shallow bed apparatus for ion exchange kinetic studies 29Figure 2.12 The laboratory column 34Figure 2.13 A typical autodiagram 34Figure 2.14 A typical X-ray microprobe analysis 34Figure 2.15 Cobalt purification at Port Colbome 36Figure 2.16 The Cobalt Hydrate Plant 38Figure 2.17 The Electro-Cobalt Plant 39Figure 2.18 Sketch of the D2EHPA extractant 41Figure 2.19 The D2EHPA dimer 42Figure 2.20 Extraction order of various ions by D2EHPA 45Figure 2.21 Extraction order of various ions by Cyanex 272 45Figure 2.22 Extraction order of various ions by Cyanex 302 45Figure 2.23 Sketch of the Cyanex 272 extractant 46Figure 2.24 Sketch of the Cyanex 302 extractant 47Figure 2.25 Comparison of zinc extraction by the extractants 48Figure 2.26 Comparison of cobalt extraction by the extractants 48Figure 3.1 Surface structure of the fractured bead 51Figure 3.2 Interior structure of the bead 51Figure 3.3 Apparatus for column studies 54Figure 3.4 Apparatus for batch studies 54Figure 4.1 The Film Diffusion Model 61Figure 4.2 The Particle Diffusion Model 61Figure 4.3 The Unreacted Core Model 61viiFigure 4.4 Typical breakthrough curve with the representation of the characteristicparameters and movement of the zones 65Figure 4.5 Variation of the rate function with 13 65Figure 4.6 Determination of charecteristic parameters 66Figure 5.1 Zinc loading on untreated OC 1026 at 40°C 70Figure 5.2 Zinc loading on untreated OC 1026 at 40°C and 10 BV/hr 70Figure 5.3 Zinc loading on untreated OC 1026 at 60°C and 4.6 BV/hr 72Figure 5.4 Zinc loading on pre-treated OC 1026 at 60°C 72Figure 5.5 Zinc loading of OC 1026 pre-treated at pH=4 at 25°C 74Figure 5.6 Effect of pre-treatment on pH profile and Zn loading 74Figure 5.7 Stripping of zinc and cobalt from OC 1026 77Figure 5.8 Effect of pre-treatment and electrolyte pH increase on metal retained 77Figure 5.9 Effect of temperature on zinc breakthrough 80Figure 5.10 Effect of electrolyte pH increase on zinc breakthrough 80Figure 5.11 Zinc loading on pre-treated Cyanex 302 at 60°C 82Figure 5.12 Effect of resin poisoning on breakthrough of Cyanex 302 82Figure 5.13 Zinc loading on pre-treated Cyanex 272 83Figure 5.14 Comparison of cobalt retained on the resins 83Figure 5.15 Kinetics of zinc uptake by untreated OC 1026 85Figure 5.16 Effect of pre-treatment on the kinetics of zinc uptake by OC 1026 85Figure 5.17 Effect of temperature on kinetics of zinc uptake by OC 1026 87Figure 5.18 Effect of initial [Zn] on zinc uptake 87Figure 5.19 Effect of pre-treatment on kinetics of zinc uptake by Cyanex 302 88Figure 5.20 Effect of temperature on the kinetics of zinc uptake by Cyanex 302 88Figure 5.20 Kinetic model fit of the experimental data for pre-treated (pH=3) OC 1026operated at 60°C 90Figure 5.21 Kinetic model fit of the experimental data for the first 5000 seconds 90Figure 5.23 Arrhenius plot of the aqueous phase mass transfer coefficients 92Figure 5.24 Arrhenius plot of the resin phase diffusion coefficients 92Figure 5.25 Variation of the activation energy with time 93Figure 5.26 Comparison of the kinetics of zinc uptake by the resins 93Figure 5.27 Breakthrough curve for zinc loading on pre-treated (pH=3) OC 1026 96Figure 5.28 Breakthrough curve for zinc loading on pre-treated (pH=3) Cyanex 302 96Figure 5.29 Breakthrough curve for zinc loading on pre-treated (pH=3) Cyanex 272 97Figure 5.30 Breakthrough curve analysis to determine characteristic parameters 97Figure 5.32 Cobalt and zinc loading on untreated OC 1026 100Figure 5.33 Pre-treatment and loading processes during the ion exchange 100viiiList of Symbolsa = Mass transfer area (m2).a1 = Activity of species i.c = Concentration of the species (g/L).c0 Feed concentration of the species (g/L).D = Diffusion coefficient of the species (m2/s).DC = Distribution constant.DbS = Degree of bed saturation.Dq = Energy splitting parameter.e = Bed voidage (dimensionless).F = Faraday constant (Coloumb/equivalent).f = Fugacity of the species.h = Height of an individual zone (m)H = Height of the resin bed (cm).Hf = Height of the transfer zone (m).H. = Height of the unused bed (cm).J = Flux of the diffusing species (g/m2 s).lç, = Mass transfer coefficient in the aqueous film (mis).k1,k2 = Rate constants of forward and reverse reaction (mis).kL =Liquid phase mass transfer coefficient obtained from Carberry’s correlation (mis).N = Molar flux of the species (mollm2s).r0 = Radius of the resin particle (m).Re = Reynold’s number (dimensionless).Ri3 eq Equilibrium concentration of the extracted species in the resin (g/L)ixR = Concentration of fixed ionic group (g/L).Sh = Sherwood number (dimensionless).Sc = Schmidt number (dimensionless).t = Time elapsed (seconds).T = Temperature (SC).U = Superficial fluid velocity through the column (mis).U = Velocity of mass transfer zone through the bed (mis).X = Fraction extracted.z = Valency of the ion.Subscripts:A = Ion diffusing from the resin phase to the aqueous phase.B = Ion diffusing from the aqueous phase to the resin phase.Superscripts:overbar = Resin phase parameter.*= Resin/solution interface.b = Bulk solution parameter.Greek Symbols= Separation coefficient (dimensionless).= Shape factor of breakthrough curve (dimensionless).6 = Aqueous phase film thickness (m).C = Fractional pore volume of the ion exchanger.xO = Time elapsed (hrs.).00 = Breakthrough time (hrs.).em = Mean breakthrough time (hrs.).= Bed saturation time (hrsj.PB = Solid phase bulk density of the resin (kglm3).PL = Density of the liquid phase (kg/rn3).= Viscosity of the liquid phase (kg/rn s).= Equivalent conductance (mho/eq.m2).= Electric potential gradient (Volt/rn).xiAcknowledgementsI would like to take this opportunity to thank a few people who have helped me write this thesis,by lending me timely academic and emotional support.I thank my research supervisor Prof. David Dreisinger for the trust he placed in my abilities, whenhe chose me to come to the Univ. of British Columbia to work with him on my Master’s. I thankhim for his patience and understanding, especially during the early stages of my research work, andfor the encouragement he provided throughout my work. I thank Prof. W.C.Coopers for his insightfuldiscussions on electrochemistry which have influenced my understanding and approach tohydrometallurgy. Thanks are also extended to INCO Ltd. for their financial support to my work.My colleagues have been instrumental in enhancing my understanding of Metals and MaterialsEngg.. I thank them for working with me as a team and providing me with an environment whereI could discuss my problems, and avert many a “dead end”. I thank Ish, Brenna, Ben and Paul forthe help they provided me with, in the initial set-up of my experimental apparatus. I thank Mikeand Ben for always being around in the lab and not making me feel that I was the only one workingso hard. I thank Mukunthan, Vinay, Bernardo, Victoire, Anita, Charlie and Harold for the livelydiscussions on topics ranging from Material Science to spiritualism.xii1 CHAPTER ONE- INTRODUCTION1.1 Ion Exchange PurificationIon exchange is an important unit process in many metal purification operations, and has awide variety of applications in the fields of analytical chemistry, biotechnology, catalysis,decontamination and separation technology. Its application is of special importance in the purification of electrolytes and waste streams, by the removal of impurities present in extremely lowconcentrations. Ion exchange has undergone a great deal of technological advancement since the1950’s, from the development of newer types of selective resins and novel contacting techniques,to intensive studies on the thermodynamics and kinetics of the process to understand the exchangeequilibria and mechanisms. In recent years, the resurgence of interest in ion exchange has been inthe direction oftechnological development based on newer theoretical understanding of the process,applied to its under-utilized potential in the recovery of valuable materials from natural resourcesand waste, and to the prevention or correction of damage to the environment. One step in thisdirection was the introduction of a selective extractant in the resin matrix, rather than the functionalgroup, to extract the impurity, which greatly enhanced the efficiency and selectivity and led to thedevelopment of solvent impregnated resins for some specific separation processes. The subject ofthis thesis is the study of a class of such solvent impregnated resins, for their application to theINCO zinc removal process for cobalt electrolytes, using standard methods to study the kineticsand interpret the process mechanisms.1.2 Zinc Removal from Cobalt ElectrolytesThe removal of zinc impurities from cobalt electrolytes is of the utmost importance in theextractive metallurgy of cobalt, as zinc is a deleterious impurity in cobalt electrowinning. Solventimpregnate technology pioneered by the Bayer Chemical Company, offered the resin OC 1026,an ion exchange resin impregnated with di-2-ethyl-hexyl phosphoric acid (D2EHPA) extractant,which is currently being used at the INCO cobalt refinery at Port Colborne. This resin, however,1shows poor zinc selectivity, and a fair amount of cobalt is retained on the resin at the zincbreakthrough point. This cobalt is lost in the zinc elution process. It has been suggested thatphosphinic acids or sulphur substituted phosphinic acids could offer better zinc selectivity. Henceworking in collaboration with the University of British Columbia, the Bayer Chemical Companysynthesized special resins impregnated with the Cyanamid extractants Cyanex 272 and Cyanex302. The chemical structure of these extractants is compared to that of D2EHPA in the followingfigure:R—O 0 R 0 R SP P P/\ /\ /\R—O OH R OH R OHD2EHPA Cyanex 272 Cyanex 302pK=1.2 pK=6.37 pK=5.63For D2EHPA; R = 2 ethyl hexyl groupFor Cyanex 272 and Cyanex 302; R = 2,2,4 trimethyl pentyl groupFigure 1.1 : Organophosphorous extractants used in OC 1026, Cyanex 272 & Cyanex 302The main objective of this work was to compare the performance of these solvent impregnatedresins in terms of zinc loading, cobalt/zinc selectivity and metal exchange kinetics, and to suggestimprovements to the INCO zinc removal process. It is hoped that knowledge gained from thisinvestigation will make some useful contributions to the understanding and process developmentsat the INCO cobalt refinery, and in other processes for the removal of zinc from electrolytes orwaste streams.22 LITERATURE REVIEW2.1 Ion Exchange ProcessIon exchange is a chemical process in which ions are exchanged by a solid ion exchanger andan aqueous solution as follows:Cation Exchange: RmA + mB —-- mRB + mA (2.1)Anion Exchange: RmA+mW —--> mRB+mA (2.2)where RA is the solid ion exchanger which is usually an organic polymer (the bar represents theorganic phase) and is called a cationic or an anionic exchanger depending on whether it exchangescations or anions respectively. A andB are called counterions, R is thejlxed ion and m is the valenceratio.2.2 History and Technological Development of Ion ExchangeIon exchange has an illustrious history dating back to pre-historic times when Moses softenedthe bitter waters of Mara to make them potable for his flock in the desert [1]. Aristotle observedthat the salt content ofwater is diminished upon percolation through certain sands [2]. More recentlyin 1850, ion exchange was studied in a systematic manner by two British soil scientists Thompsonand Way [3,4]. With progressing industrialization, the focus, around the turn of the century, shiftedto the application of ion exchange to plant-scale water softening, first with natural and later withsynthetic ion exchangers [5]. The chance discovery that a shattered phonograph record exhibitsion exchange properties, by English scientists Adams and Holmes [6] led to the invention of ionexchange resins. In World War II, ion exchange contributed significantly to the Manhattan Projectby providing solutions to the challenge of separating rare earth elements and other fission products[7-9]. Other important advances after the war include the advent of more stable and reproducibleion exchange resins based on polystyrene [10], the development of strong-base anion exchangers(with quaternary ammonium groups) [11], commercial development of effective inorganic ion3exchangers in the form of synthetic zeolites and molecular sieves [121 and the development ofmacroporous ion exchange resins. Later developments include novel ion exchange contactmethods, and the development of chelating and solvent impregnated ion exchange resins. Today,a wide range of improved organic and inorganic ion exchangers with a great variety of propertiesis available for laboratory and plant-scale applications ranging from chemical analyses to preparative separations, from catalysis to organic synthesis, from biomedical uses to decontaminationand detoxification.2.3 Present ApplicationsIon exchange has been instrumental in the development of our science and technology, and itsthrust and emphasis have continuously evolved to keep up with the changing outlook and needsof the time. Ion exchange processes are actively utilized in the following applications:(1) Dialysis apparatus for biomedical engineering(2) Enhanced oil recovery processes(3) Chemical analyses as in chromatography(4) Manufacture of gasoline additives to provide an effective no-knock performance(5) Chemical separations of rare earths, transition metal cations and optical isomers(6) Decontamination of water used in cooling systems of nuclear reactors and other industrialplants(7) Recovery of uranium and thorium from dilute leach liquors2.4 Chemistry of Ion ExchangeIon exchange resins are solid, water insoluble, high molecular mass polyelectrolytes, whichconsist of a hydrophobic polymer resin matrix (prepared by emulsion co-polymerization), to whichspecific functional groups are attached. The ion exchanger must preserve its electro-neutrality at4all times, and this is accomplished by balancing the resin’s fixed ions with counterions. Thecounterions can be exchanged for a stochiometrically equivalent quantity of other ions of the samesign when the ion exchanger is in contact with an electrolyte solution.2.4.1 The resin matrixIn ion exchange the resin substrate is usually an organic polymer which is produced in beadform (0.5-2mm diameter). One common type ofresin is the polystyrene-divinylbenzene matrix,crosslinked to provide structural strength as shown below:CH=CH CH-CH CH-CH —2 2 2HC)Figure 2.1(a): Preparation of a crosslinked resin polymer matrixGel type resins (microporous) result when only styrene and divinylbenzene are used in thesynthesis, while if an inert material such as toluene is added during polymerization and evaporated after bead formation, then a macroporous bead results. Macroporosity tends to enhancekinetics (refer Fig. 2.1(b) for a comparison of the structures).Important physical properties such as swelling, mechanical strength, abrasion resistance,rate of exchange and exchange capacity, which impact the use of ion exchange resins, dependto some extent on the degree of crosslinlcing. Swelling is caused by the adsorption of water intothe resin structure, which hydrates the functional groups and lowers the exchange capacity, andis considerably increased by a lower degree of crosslinking. A resin with a low degree ofcrosslinking will also break or abrade more easily under induced mechanical or osmotic stresses.5Figure 2.1 (b): Surface characteristics of ion exchange resins [15].(i) Microporous resin.(ii) Macroporous resin.Initial stateMatrix with fixed charges C Counter ionsEquilibriume Co-ionsFigure 2.2 The ion exchange process [15].(i) 100 micronS (ii)I I000O \\ç/0 06In contrast, higher rates ofexchange are achieved with a low degree ofcrosslinking. As a trade-offof these factors, the amount of crosslinking for polystyrene-divinylbenzene resins is usually inthe range of 4-8% divinylbenzene (DVB).2.4.2 Functional groupsThe nature of ion exchange resins is determined by the functional groups attached to theresin matrix. Ion exchange resins may be classified as follows based on the functional groups:Table 2.1 : Classification of ion exchange resins based on functional groups.Nature Functional Range Applications Capacity Commercial resinsgroup meq/gStrong -SO3H pH <2 Water softening 5 Lewatit S 100,acid Amberlite IR-200Weak acid =PO2H pH >4 Rare earth 10 Amberlite JFC-84processing Duolite CC3Strong -NR3X 0 - 14 Uranium 4 Zerolit MPFbase processing Lewatit MP200Weak base -NH2,-NR2, 0 - 10 Uranium 5 Zerolite MPH-NHR processing Lewatit MP62Chelating Picolylamine- Cu & Ni 1.5 XFS 4195resin -N(CH2C5H4) removal2.4.2.1 Strong and weak acid ion exchangersCation exchange can be carried out using suiphonic acid type exchangers. These resinshave the functional groups -SO3H, and are called strong acid ion exchangers. They can extractcations at low pH but are relatively unselective. Weak acid ion exchangers have functionalgroups of the -CO2H and =PO2Htype, and can extract cations at moderate pH values (pH>4) only, but are somewhat more selective. The pH limitation of weak acid ion exchangers isdue to stability of the protonated weak acid group at high acidities. Strong and weak base ion exchangersWeak base resins are those with the functional groups -NR2, -NHR and -NH2’ and mustbe protonated to exchange ions. The useful pH range is therefore approximately 0-10. Strongbase resins have the functional group -NR3Xand extract anions efficiently over the entire pHrange.2.4.3 Aqueous chemistry of metal ionsThe reactivity of metal ions in solution may be related to the following:(1) electronic configuration of the ion(2) ionic radii of metal ion(3) position on the hard-soft acid base (HSAB) scaleIons with electrons in s andp orbitals (alkali and alkaline earth) show a limited degree of varietyin their chemical reactivity, whereas, transition, lanthanide and actinide metal ions with partiallyfilled d andforbitals, have a more complex behavior. Hard bases are usually small in ionic radiiwith localized charges, whereas the soft bases are large in size with diffuse charges and highpolarizabilities (see Table 2.2). A rule of thumb is that soft acids tend to interact preferentiallywith soft bases, and conversely, hard acids with hard bases. Hard bases tend to form complexeswith the following order of stabilities:Class A (hard bases): F > Ci> Br >1O>>S >Se>TeN>>P >As>Sb>BiClass B (soft bases): W, R, C2H4C6H,CN, RNC, COSCN, R3P, RS -, 520328Table 2.2: Classification of Lewis acids [19].Hard SoftH, Li, Na, K Cu, T1, HgBe2,Mg2,Ca2,Sr, M2Al3,Sc3,Ga3,J3 La3 Tl3,Tl(CH)BHGa(a lN, Gd3,Lu3 RS, RSe, RTeCr3,Co3,Fe3,As3,Ce3 r, Br, HOP, RO’Si4,Ti, Zr, Th, Pu‘2’ Br2, ICNUO2,(CH)nVMoO Tri-nitrobenzeneBeMe2,BF3,B(OR)3 Chioranil, quinonesAl(CH3),A1C13,AIH3 Tetra-cyano-ethyleneRPO2,ROPO2 0, Cl, Br, I, NRS0, ROS02SO3 M° (metal atoms)j7+ ,5-i- Cl7,Cr Bulk metalsRCO, C02,NC CH2, carbenesHX (hydrogen bonding molecules)With Class B (soft bases) metal ions, this order is reversed. Quantitative data for stabilityconstants of metallic ions and organic or inorganic ligands is given in reference [13]. Anexamination of the electronic structure of metal ions helps to screen them into alkali alike ions(AAI) and transition metal alike ions (TMAI). In the former group steric considerations arepredominant, while electronic considerations are predominant in the latter [14]. The formationof a metal complex from a solvated metal ion and a ligand may involve a simple ion pairingmechanism, where the d subshell electrons of the transition metal are paired by a simplesubstitution of the solvent molecule by the ligand, or a chelating mechanism, where more thanone solvent molecule is substituted at a time, and hence the metal is attached to the ligand atmore than one point. The complexation behavior of the AAI group may be best described interms of the host-guest complexation, where the complementary behavior between the ion (guest)and the ligand (host) is required for the complexation process. For this to occur, an importantcondition is the complementary sizes of ionic radii and ligand cavity. Fig. 2.3 shows the PeriodicTable in a steric perspective, pointing out the large differences in ionic radii of alkaline cations.993®©88 ‘0.88 0.7? O.?t) .66 ()6Y) 1.121.40 (36®p®® o.99€[43 rrjl.I t t0i5o.so 1.84 [Si) €)€t3( (25(29 (.26 t 128c o12( U7() ()I.69Cu Zn;Jc(41 (.36 (.3 (33 (34 138 L44.j4t 4(41t37()()I0093 (26 097 081 071Lo QQ )00000 OOOc21cQ(.43 (37 (37 134 135 (35L44 O’71d5t46(.4 1.4 2.21.37 Hg’ j’ Pb4110 0.95 0.84Figure 2.3: The periodic table in steric perspective (atomic and ionic radii in angstromunits) [14]100Figure 2.4: Characteristic rate constants (sec4)for substitution of inner sphere H20 of various aquo ions [14]10Figure 2.4 shows the rate constants for substitution of inner sphere H20 of various aquo-ions.Ions to the right of the diagram, are termed labile, because of faster exchange kinetics than thoseon the left. Figure 2.5 shows the different kinds of co-ordinate bonds possible for a co-polymer.In the TMAI group, electronic configuration, complex structure and stereochemistry areimportant. Crystal field splittings of the d-orbitals of an ion in tetrahedral, octahedral, tetragonaland square planar complexes are shown in Fig. 2.6. The energy difference between the degenerated levels of the cation and the stabilized d-orbitals of the complex is indicative of the thermodynamic stability of the complex. Crystal field stabilization energies for d’ complexes (n designates the number of electrons in the d-orbital) of common octahedral and square planarconfigurations, and for both weak and strong field ligands are shown in Table 2.3. It may beobserved from the table that d° and d’° ions (also d5 ions at weak field ligands) are not stabilizedby complex formation, while d7 and d8 ions benefit from complex formation of both octahedraland square planar configurations, particularly with strong field ligands.2.4.4 Guidelines for ligand selectionFor a given separation task it may be necessary to go through the following steps to develop aschematic approach to process development:(1) A study is made of the inorganic chemistry of the target metal ion under the definedworking conditions.(2) The predominant interactions between the metallic ion and ligands are determined.(3) A separation strategy is selected for the given ion based on the host-guest interaction,chelate formation or ion-pairing mechanisms.(4) The effect of the various separation strategies on other counter-ions and co-ions needs tobe considered.(5) Finally, the strategies for successful elution need to be considered.11Matrix of Co-ordinathig Co-polymerCo-o4dinale bondpolydenta*e bidentateI I6 .o ss NS NJl N!O .omonodentaleI IS N 0Figure 2.5 : Co-ordinating copolymers - donor atoms in one ligand [191Figure 2.6: crystal field splittings of the d-orbitals of a central ion in regular complexes ofvarious structures [14]Table 2.3 : Crystal field stabilization energy of d complexes [14]System Example Octahedral Square PlanarWeak field Strong field Weak field Strong fieldd° Ca2,Sc9 ODq ODq ODq ODqdt Tin, U 4 4 5. 14 5, 14d2 Ti2,V 8 8 10. 28 10,22d3 V,Cr 12 12 14,56 14,56d4 Cr. Mn 6 16 12,28 19,70d5 Mn, Fe3,Os 0 20 0 24, 84d6 Fe2,Co, fr3 4 24 5, 14 29, 12d1 Co2,Ni, Rh2 8 18 10, 28 26, 84d Ni2.Pd2,Pt 12 12 14, 56 24, 56d9 Cu2.Ag2 6 6 12, 28 12, 28d1° Cu,Zn,CdAg,HgGa 0 0 0 0Tetrahedral Free Ion Octahedral Tetragonat or Square planaraquare pyramid122.5 Ion Exchange TechnologyThe theory and practice of ion exchange are well documented [15-18], and only the essentialdefinitions and principles are elucidated here.An ion exchange reaction may be expressed in general as follows:mA + RmB mRA + B (2.3)The thermodynamic equilibrium constant of the above reaction can be defined as:(Am(aBK = 1-I (2.4)aA) aB)where the a is the activity of the ion i.Replacing the activities by concentrations we get:K [A]J([BiJ (2.5)K’B is known in ion exchange technology as the selectivity coefficient.The extent of exchange in the process is defined by the distribution coefficient for the ion A; DA:D— milliequivalents of ion A per gm. of exchanger (26)A— milliequivalents of ion A per ml of solutionWhen a number of counter-ions are being exchanged by the resin matrix, the ease with which theion exchange process can separate one of the ions from the rest is expressed as the separationcoefficient of ion A over ion C:o. = — (2.7)D13The exchange capacity is the number of ionogenic groups per specified weight of the resin. It isusually expressed in milli-equivalents per gram (meq.Ig) of dry resin or equivalents per litre (eq./L)of the wet settled resin. The following factors lead to a high exchange capacity [19](1) Little or no cross-linking.(2) Flexible crosslinks (alkyl rather than aryl crosslinks).(3) Maximum number of active functional groups on the co-polymer.(4) Maximum number of active donor atoms per active functional group eg. chelating ratherthan monodentate.(5) High metal to ligand affinity.(6) Strong metal-ligand bonds.(7) Low co-ordination number of the metal.When an ion exchanger is brought into contact with an aqueous solution, a considerable amountof water may enter the resin phase and cause swelling of the resin. Swelling is opposed by theelastic force in the matrix and attains an equilibrium value. Swelling hydrates functional groupsand reduces capacity. Excessive swelling in ion exchange units can cause bed movement and leadto undesirable channeling.Selectivity is the preference of a resin for any particular ion among many present in an aqueoussolution and depends on both the resin and the counterion for strong and weak cation (acid)exchangers as follows [19]:(1) Selectivity increases with the charge on the cation (Th(TV) > La(III) > Ca(ll) > Na(I))(2) Selectivity increases with decreasing hydrated radius of the cation.(3) Selectivity increases with increasing polarizing power (related to the effective nuclear chargeso that an ion with a high charge and a small radius like Al3 has a high polarizing power).(4) Selectivity is decreased by association and complexation in the aqueous phase as thesedecrease the activity of the free counterions available for exchange.142.6 Kinetics of Ion ExchangeInitial theories on the kinetics of ion exchange were based on the speculative consideration of theprocess as a chemical reaction of a kinetic order corresponding to its stoichiometric coefficients.Much later in the 1950’s the process was understood to be essentially a statistical redistributionof ions by diffusion, with a rate limited by mass transfer resistances in either the particle or theexternal fluid [211. The next advance was made in the 1950’s with the realization that ions, ascarriers of electric charges, are subject to the electrical field that their own diffusion generates,thus obeying Nernst-Planck equations more closely than Fick’s laws [15]. Shortcomings in theapplication of the Nernst-Planck equations arose due to variations in the swelling conditions andthe relaxation behaviors of resin networks, which resulted in changes in particle size and ionicmobility, and the diffusion coefficient becomes a function of time as well as concentration [311.A more fundamental problem arose later when the individual diffusion coefficient premise, presentin the Nernst-Planck equations, did not satisfy experimental conditions, but an array of interactioncoefficients as in Stefan-Maxwell equations provide more promising answers [43]. This led to thedevelopment ofion exchange kinetic models based on the thermodynamics ofirreversibleprocesses[46]. This section surveys the work done so far in these fields, leading up to the models used inthe formulation of rate laws in ion exchange kinetics.In general, in most kinetic studies of ion exchange processes, information on the following pointsis vital to process understanding and design [201:(1) The mechanism of the ion exchange process(2) The rate determining step of the process(3) The rate laws obeyed by the process2.6.1 Mechanism of ion exchangeAn ion exchange process is represented by:15RmB+mA —-* mRA+B .(2.8)and could involve the following steps:(1) Diffusion of ions A through the film surrounding the particle to the resin surface - filmdiffusion.(2) Diffusion of ions A through the resin to the exchange sites - particle diffusion.(3) Chemical exchange of ions A and B at the sites of functional groups of the exchanger.(4) Diffusion of the exchanged ions B through the particle to the surface of the resin.(5) Diffusion of the exchanged ions B through the liquid film surrounding the particle intothe bulk solution.However, the principle of electroneutrality requires that steps (1) and (5) occur simultaneouslyat equal rates and analogously steps (2) and (4) must occur at equal rates in opposite directions.As a result, the number of rate determining steps is reduced to the following three:- film diffusion- particle diffusion- chemical reactionThe diffusion and loading inside the resin particle may proceed through three possible mechanisms as shown in Fig. 2.7, along with the associated concentration gradients [20]. Apart fromthe more common gradient diffusion mechanism based on Fick’s theory, two other possiblemechanisms expected to occur are the homogeneous mechanism (or progressive conversion)where the concentration gradient in the resin is homogeneous as discussed by Wen [22] andLevenspeil [23], and the shell progressive mechanism (or unreacted core model), developed byYagi and Kunii [24].16i) Film diffusionA) Rate determining step’ ii) particledifFusiOniii) chemical reactionB) Diffusion mechanism ?RcSii) HomogeneousHHiii) Sheli_prOgreSSiVCrR6Figure 2.7 : Fundamental questions in kinetic investigations [20]Porticle-dtfusion controlFigure 2.8 : Rate determining steps [15]I) FickiariCGo 00//F4m-djjfuj0ncontrol172.6.2 Rate determining stepIon exchange processes are commonly controlled by the diffusion of counter-ions in theadherent film or in the resin matrix, rather than by the chemical exchange reaction. The ratecontrolling step often depends primarily upon the concentration of the external solution, and asa general rule, the film diffusion step is slow with solutions of the order of 0.01 N or less, whileparticle diffusion may become controlling with solutions of the order of 0.1 N or more [211 (Fig.2.8). Helfferich [15] developed the following criteria for predicting the rate determining step,by comparison of the half times of particle and film diffusion:RD8Particle Dffuszon Control : ,, , (5 +2cLA,B) << 1 (2.9). U r0Film Diffusion Control : (5+2clA,B) >> 1 (2.10)where R = concentration of fixed ionic group (equivalents)C = concentration of counter-ions in the aqueous solution(equivalents)D = diffusion coefficient in solution (overbar symbol for the resin phase) (m2Is)8 = film thickness (m)r0= radius of the resin bead (m)cxA,B = separation factor (dimensionless)The quantities R, C, r0 and aA/J3 are usually known or may be determined based on the propertiesof the resin [15]. The values of the diffusion coefficient in the aqueous and resin phases aredetermined as harmonic means of the individual diffusion coefficients of counter-ions A and Bas follows [91]:— DADB(zA+zB)D, =—— (2.11)ZADA + ZBDBD D (z +z)= ABA B (2.12)ZADA + ZBDB18where z1 = electrochemical valence of species i.In the case that D values are not available, it can be assumed that DID = 10 or other estimatesmay be obtained from Wheeler [25]:— D.cD =—.-- (2.13)or Mackie [26]— D.c2D. = (2.14)(2—c)where c = fractional pore volume of the ion exchangerThe film thickness ö is obtained through the Sherwood number as follows:2r 2kr0Sh = 0 = m (2.15)Dwhere Sh = Sherwood numberkm = mass transfer coefficient in the filmThe Sherwood number is a function of the Reynolds (Re) and Schmidt (Sc) numbers [88]:Sh = 0.37 (Re)°6(Sc)°33 for 20 <Re <150000 (2.16)Sh = 2 + 0.37(Re)°(Sc)°33 for Re <20 (2.17)Depending on the hydrodynamic conditions, the film thickness is usually in the order of 102 -10’ mm. Further distinction can be made on the rate determining step based on experimentaldata as follows:19Table 2.4 : Dependence of ion exchange rate on experimental conditions [15].Factor Rate of PDC Rate of FDCCounter-ion mobility:In particle oc D No effectIn aqueous film No effect oc DCo-ion mobility No effect No effectParticle size oc hr02 oc hr0Capacity of resin No effect oc h/XDegree of crosslinking Decreases with increase in No effectcrosslinkingConcentration of solution No effect oc CTemperature Increases 4-8% per °C Increases 3-5% per °CRate of agitation or flow No effect Increases with agitationrateNote: PDC = Particle Diffusion ControlFDC = Film Diffusion Control2.6.3 Rate laws of ion exchangeThe rate of ion exchange is a complicated function of physico-chemical and hydrodynamicparameters. All these parameters except the diffusion coefficient of the ions in the liquid filmand in the resin particle are known or readily obtainable. One method of conducting kineticstudies involves trying to fit the experimental rate data to various available rate equations fromliterature and computing diffusion coefficients from this analysis. This, however, creates thedanger of choosing a rate law which gives the best fit of experimental data, obtained underspecific conditions, but which does not necessarily reflect the actual ion exchange process. Analternative but more complex approach is based upon the fundamental physico-chemical processes occurring during the course of ion exchange. The necessary tools and equations forcarrying out both approaches are outlined in the following sections. Ion exchange systems not involving chemical reactionsKinetic analyses of this kind were initially worked out by Boyd and co-workers [21,27,28].In very dilute solutions, if the mobility of the counter-ions is equal, the diffusion flux ofspecies i may be adequately described by Fick’s first law:Film Diffusion : J = —D1 grad c. (2.18)Particle Diffusion : = —D1 grad (2.19)where J = flux of diffusing species i (kg/rn2 s)= diffusion coefficient of species i (rn2/s)= concentration of species i (kg/rn3)and overbar species represent the resin phase.(A) Particle Diffusion ControlThe time dependence of the concentration is related to the material balance by Fick’ssecond law:= —divJ1 (2.20)Combining equation (2.19) and (2.20) for spherical particles:(ö2c 2&—= + ——H (2.21)& iör2 rSr)This equation may be solved for this and various other diffusion geometries [31] given theinitial and boundary conditions. For a linear driving force relation, this was approximated byGlueckauf and Coates as follows [29]:dc1 D1k= —--(cc) (2.22)at r021where Z =average concentration of the resin phase in equilibrium with the externalsolutionk = empirical constantr0 = radius of the resinA closer approximation of equation (2.21) is obtained by the quadratic relationship ofVermeulin [30]:2dt = 2c,(2.23)where c is the mole fraction rather than concentration of species i.(B) Film Diffusion ControlFrom the material balance condition and the requirement of equilibrium at the particlefilm interface, the following equation was obtained by Boyd [21]:dc1—3cD(C_Cj*)224dt — r0 8 (. )where c* = concentration of species i at the particle-film interface(kg/m3)c1b concentration of species i in the bulk solution(kg/m3)c total counterion concentration in solution (and in the resin phase for theoverbar term) (kg/rn32.6.3.2 Ion exchange accompanied by chemical reactionsEven though most ion exchange processes are diffusion controlled as argued by Boyd[21], through a mathematical coincidence, a quantitative description of ion exchange as areversible second order reaction proved much more tractable than the diffusion rate laws whenapplied to calculations of ion exchange columns [32] and so the reaction model was retained22by engineers who formalized this convenience by relating the reaction rate coefficient to themass transfer coefficient [33,34]. However, recently, more rigorous models of reaction in ionexchange have surfaced once again.Ion exchange systems involving neutralization, hydrolysis and complex-formation reactionshave been analyzed from a theoretical point of view, by Helfferich [35]. For certain types ofion exchange processes occurring in a spherical resin bead, such as for example, the protonation of a weak base or weak acid resin or complex formation reactions, Heifferich postulatedthe likelihood of a sharp moving boundary. The existance of sharp moving boundaries hasbeen demonstrated by Dana and Wheelock [36] who studied elution of copper amine complexes from a sulfonic acid exchanger and by Holl and Sontheimer [37] who investigatedprotonation of a weak acid resin. Both these groups of investigators used resins with atransparent matrix in their studies. This permitted visual observations of sharp movingboundaries in the reacting resin beads. Electric potential gradient and the Nernst-Planck equationIons, as carriers of electric charge are subject to electric forces. Their diffusion generatesan electrical potential gradient, a diffusional potential, whose action on the ions has to betaken into account in their equation of motion. This is done by the Nemst-Planck equation:J = _Dgradc1 + zicigradP) (2.25)Flux Diffusion Electrical transferencewhere z1 = Valency of the ionF = Faraday’s constantR = Universal gas constant1 = Electric potential gradient23This proposes that an initial minute disparity of diffusion fluxes causes a minute deviationfrom electroneutrality, immeasurable except through observation of electric potential differences. The electric potential gradient produces electric transference of both ions, in thedirection of the diffusion flux of the slower ion, in effect retarding the faster ion andaccelerating the slower one as shown below:InterfaceCounter-ion 1 Counter-ion 2Diffusion 4TransferenceNet Flux(newly equal magnitude)Figure 2.9 : Electric transference acts to equalize the net fluxThe electric potential gradient thus is the mechanism which enables the system to maintaina state so close to electroneutrality, that the deviation is negligible in just about all cases ofpractical interest in ion exchange kinetics [391.Initial applications of the Nernst-Planck equations integrated for ideal systems, under theconstraints of electroneutrality and absence of an electric current, for the geometry of ahomogeneous sphere, and under the simplest conceivable initial and boundary conditionswere tried by Heifferich [39]. Nernst-Planck equations have been applied to the deviationfrom electroneutrality at the particle-solution interface and in and around extremely small ionexchanger particles [38]. Another problem requiring a more elaborate treatment is that of ionexchange with macroporous resins, for which the flux equations must be solved for a morecomplex geometry [39]. Along different lines, application of the Nemst-Planck equations toreal systems calls for the inclusion of an additional term accounting for the effect of gradientsof activity coefficients:24J1 —Dgradc1 + c.gradlnf + zicjjgradØJ .(2.26)and has been carried out for the case of zeolites [40,41]. Still another complication to beaccounted for, concerns co-ions which are transferred in the direction of the faster counterionwith the solvent being transferred in the opposite direction, and this adds a convective termto the Nernst-Planck equation [42]. It may also be necessary to account for concentrationgradients in both the film and particle, with non-linear equilibrium at the particle surface thendeserving special attention [43]. Systems involving diffusion and reactionInspite of the understanding of ion exchange as a statistical redistribution of ions, thereare, however, situations in which a reaction - i.e. the formation or dissolution of a chemicalbond, is undoubtedly involved. Among the reactions that often accompany ion exchange areacid-base neutralization, dissociation of weak electrolytes in solution or weak ionogenicgroups in an ion exchanger, complex formation, or combination of these, resulting in anabnormally low apparent interdiffusion coefficient in the ion exchangers. A few examplesare listed below:RSO + H + Na ÷ 0H— RSO + Na + H20 (2.27) ERCOQ + Na + H + C1 —, RCOOH + Na + C1 (2.28)RNH2 + H + C1— RNH + C1 (2.29)2RSO + Ni + 4Na + EDTA4— 2RSO + 2Na + 2Na + NiEDTA..(2.30)The distinguishing feature is that the basic material balance must now account for mass transferand reaction rate as follows:öc.—divJ1 + R. (2.31)25The effect of this on kinetic behavior can be significant. The rate may be decreased by severalorders of magnitude, the dependence on variables such as solution concentration is considerably altered, and quite distinct mechanistic features may appear [44]. If the reaction is slowcompared to diffusion, then in the limiting case the diffusion is fast enough to level out anyconcentration gradients within the ion exchanger particle, and hence with the reaction at thefixed site being the sole rate controlling factor, the rate is independent of the resin particlesize. Such a situation, however, has never been convincingly demonstrated for ion exchange[38]. The more common case is that of the reaction being faster than diffusion, so that localequilibrium exists at any point. The rate here is controlled by (slow) diffusion which, in turn,is affected by the equilibrium of the (fast) reaction leading to theprogressive sheilmechanism. Progressive shell mechanismIn this case a fast and practically irreversible reaction in the particle eliminates or generatesone of the exchanging ions [44] as in neutralization of a weak acid cation exchanger in freeacid form by a strong base:RCOOH + Na + OW - RCOO + Na + H20 (2.32)The base reacts immediately with acid groups as soon as it reaches the unconverted ionexchanger site of RCOOH. As a result, conversion to the Na form proceeds with a sharpfront from the surface towards the centre of the particle. At any time, an unconverted, shrinkingcore still completely in free-acid form is surrounded by a converted (to the Na form) shelland growing in thickness. Such behaviour variously called progressive shell, shrinking core,unreacted core or moving boundary mechanism was verified by Dana et al [36], and a modelfor the same is developed in a later section. Further developments in kinetic studiesThe inherent premise of the Nernst-Planck approach is that diffusion and transference areadditive phenomena, both characterized by a coefficient describing the mobility of therespective ion, and that the fluxes are coupled only through the electric potential gradient. Inrecent years, several more fundamental approaches to mass transfer in ion exchangers havebeen taken, with statistical considerations [45], thermodynamics of irreversible processes[46,47] or the Stefan-Maxwell equation [47,48] as the starting point, and a more complexpicture emerges in which fluxes are coupled not only through the electric potential gradient,and more than one coefficient per species is required to characterize kinetic behavior.2.7 Experimental Methods in Ion Exchange Kinetic StudiesKinetic studies may be carried out on ion exchange processes for a variety of reasons. Theapplied researcher may desire to know, at a certain stage of development of a new application(usually after equilibrium evaluation), how the kinetic features of the resin compare with itsselectivity performance, while a design engineer may carry out these studies to experimentallydetermine mass transfer coefficients, and other kinetic parameters to design an ion exchange plantor to compare the performance ofvarious resins. This section surveys the various types of apparatusused by researchers, and suggests investigation techniques and experimental conditions relevantto particular ion exchange processes.2.7.1 Selection of the apparatus for rate studiesThe general type of apparatus used for kinetic studies may be broadly divided as batchsystems (where the concentration of the counter-ions in the external solution changes with time)and shallow bed techniques (where the concentration of counter-ions in the external solutionremains constant during the experiment). The design engineer, however, finds the traditionallaboratory column studies more practical in terms of the relevant design information obtained,and hence the technique is also outlined here. The batch systemAppropriate volumes of the ion exchange resin and solution of known initial compositionare mixed and vigorously stirred (1000- 3000 rpm to minimize mass transfer resistancesthrough the film), while the time variation ofarepresentative property of the system (eg. pH,conductivity, electric potential, concentration of ion or radioactivity etc.) is recorded. Theexperimental method outlined by Kressman and Kitchener [49] uses a centrifugal stirrer (seeFig. 2.10), which encases the resins. Centrifugal action forces the inner solution to leave thestirrer through the radial holes in the encasing, being instantaneously replaced by fresh solutionentering the cage at the bottom. This also allows for the instantaneous separation of the resinsand solution by raising the stirrer out of the solution. The method used by Boyd and Saldano[27] contacts an exchanger equilibrated with a radioactive isotope, with a feed solutioncontaining the same counter-ion as the ion exchanger but free of the radio-isotope. Aliquotsof solution are withdrawn at various times and analyzed for the radioactive isotope using ascintillation counter. The radioactivity of the solution is proportional to the fractionalattainment of equilibrium. Limited bath method [49,50]A known quantity ofresin in A-form is contacted with the solution containing an equivalentquantity of counter-ion B. After the desired period of time the resin is quickly removed fromthe solution and one of the two phases analyzed. The procedure is repeated at various timeintervals, to obtain the necessary relationship of fractional attainment of equilibrium as afunction of time. Indicator method [51,52]This method is applicable to systems involving the transfer of hydrogen ions. It was usedfor measuring the rate of Na uptake by hydrogen-form suiphonic acid resins. Aqueoussolutions containing sodium chloride (larger amount), sodium hydroxide (smaller amount)28Figure 2.10: Schematic diagram of the apparatus for the batch Process [49]RE=centrifugal stirrer, E=potentiometer, R=recorder, M=stirrer, T=thermometer, EL=electrodes,PA=pneumatic piston, UT=ultra-thermostat, B=jacketed vessel, bb=baffles, S=supplementarystirrerRE-4-t.5PARadioactive InactiveWoter solution solution‘1Spherical jointWirescreens-Stainless-steel casingRubber 0 rngIon exchangerEffluentFigure 2.11: Shallow bed apparatus for ion exchange kinetics [21]29and bromocresol green indicator were contacted with a suiphonic acid resin. On the equivalentbasis, the quantity of resin was always greater than that of the sodium hydroxide but less thanthat of the sodium chloride. The time at which the indicator changed color (i.e. when all NaOHwas consumed by the liberated hydrogen ions) was recorded. A series of experiments wasperformed at constant phase ratio and Na concentration in the aqueous phase, but with variableconcentration of the sodium hydroxide in the feed solution. If the sodium chloride is addedin a large excess, the concentration of Na in the aqueous phase is virtually constant and thissimplifies the mathematical treatment of the data. Shallow bed techniqueIn this method a thin layer of ion exchanger beads is placed in a micro-column and aknown solution is forced to enter the layer as in Fig. 2.11. The variation of some representativeparameter is determined with time to give a description of the kinetic behaviour of the system[211. The main advantage of this technique over the batch technique is that if the solutionflow is high enough, the exchange occurs under infinite solution volume (ISV) conditions -i.e. negligible concentration of the ion released by the resin to the solution throughout theprocess; a situation that greatly simplifies the calculations (the same ISV condition may berealized for the batch technique by using an equivalent ratio - solution to resin of greater than100). This technique results in a plot of exchange flux as a function of time. This is convertedto fraction of ion extracted (X) by graphical integration of the flux function. Photographic method [36,37]This method is applicable to ion exchange systems involving sharp moving boundaries.In order to observe the position of an unreacted core as a function of time, the resin matrixmust be transparent. A monolayer of resin beads, initially in the A-form, is contacted in acolumn with a solution of the counter-ion B. The monolayer of the resin is photographed30using a microscope at different time intervals. This function curve gives the position of anunreacted core as a function of time, which can easily be converted to fraction extracted as afunction of time. Cone model method [53]A cone with a large ratio of the height to the base diameter is filled with an ion exchanger inthe A-form. The lower section of the cone is assembled from rings approximately 1 mm thick.The base of the cone is covered with a porous filter. The assembled cone cell is then placedinto an agitated vessel containing a solution of the counter-ion B. After a certain period oftime the cone is removed from the vessel and the resin is sliced into thin layers along therings. The resin from the individual layers is analyzed to provide a curve showing the uptakeof the counter-ion B as a function of the cone height. A number of such experiments areperformed for different contact times. Laboratory columnColumn studies provide more practical data of interest for the design engineer like resinexchange capacity, breakthrough capacity, exhaustion time etc. (Fig. 2.12), which comparewell with those obtainable in industrial columns, if experimental conditions are similar (i.e.the use of real solutions, similar flow rates etc.). Crucial parameters [54] are column bed depth(usually about 60 cm or more to permit the full development of the exchange zone within thecolumn) and diameter (at least 40 times the average resin diameter to minimize wall effects(a column diameter of 1.2cm or more with commercial resins).Effluent aliquots are withdrawnat various known time intervals to obtain the breakthrough curve, which is extremely usefulfor practical purposes such as over-all and dynamic resin exchange capacities, comparisonof different resins for a similar industrial application, evaluation of service time, throughputetc.. Further, breakthrough curves may also be used to determine the type of equilibrium bydetermining the pattern of the curve as the bed depth is increased [15], or to discriminate31between particle and film diffusion kinetics by determining the variation of the midpoint slopeof the curve with feed flow rate, which would be independent of the feed rate for particlediffusion, but would be dependent on the square root of feed spatial velocity for film diffusioncontrolled kinetics [55]. The following compares the features of the experimental techniques:Table 2.5 : Comparison of experimental techniques for kinetic studies [20].Batch system Shallow bed system Laboratory columnOutputs Extraction (X) vs t Flux vs. t Breakthrough curveAdvantages -minimum film -ISV easy -practical design inforresistance -direct diffusion coeff. mation-direct diffusion coeff. calculations -choices of analyticalcalculations parameters-choice of analyticalparameters-FSV and ISV possibleDisadvantages -indirect design infor- -indirect design infor- -FSV rather than ISVmation mation -indirect calculation of-ISV is difficult -radioactive diffusion coeff.measurement generally -high liquid resistancenecessary-high liquid ressistanceNote : FSV Finite Solution Volume; ISV = Infinite Solution Volume2.7.2 Monitoring techniques and operating conditionsThe selection ofmonitoring techniques and operating conditions is a crucial factor in carryingout kinetic tests with the objective of determining the diffusion coefficients, from a well definedtheoretical background. In the kinetic investigation, an electrochemical property such as pH orconductivity may be followed continuously and reproducibly when H or Off ions are exchangedby other, less conductive species. Potentiometric measurements may be usefully performedwhen selective electrodes for one of the species exchanged are available. By means of reference32curves, these properties are easily related to the actual concentration of the species exchangedso that the fractional attainment of equilibrium can be obtained. When the difference betweenthe electric properties of the exchanging counter-ions is small, or when the extent of exchangeis limited (such as in differential exchanges, where only 10-20 % of resin conversion takes placein an experiment) or more generally, whenever the ISV condition is to be applied, the use of aradioactive species appears extremely useful with precise and affordable measurements beingpossible.The techniques mentioned so far are indirect techniques. Direct monitoring techniques such asautoradiography and X-ray microprobe analysis may also be used. Autoradiography, whichrequires the use of at least one radioactive species, consists of thin sections of resin particle,loaded with appropriate amounts of the isotope, being contacted in the dark with special filmssensitive to isotopic radiation. The magnified picture of the section at different values of fractionconverted (X) permits visible verification of the mechanism of isotope distribution inside theresin particle (Fig. 2.13) [56,57]. X-ray microprobe analysis does not require the use of isotopes,as the energy of the electron beam of the instrument may be adjusted so as to reveal selectivelyalmost any element present in the section of the resin bead, and follow its disthbution in theresin phase (eg. Fig. 2.14). Another direct investigation technique devised by Hoell [58] usesthe variation of the refractive index of methacrylic resins when converted from the H form tothe metal-alkali form.The dependence of exchange kinetics on experimental conditions such as stirring rate, temperature, resin particle diameter etc., is determined by whether the operating mechanism isparticle diffusion controlled (PDC) or film diffusion controlled (FDC). In order to ensure infinitesolution volume (ISV) conditions, it is necessary to have the equivalents of counter-ions insolution of at least one hundred times that in the resin. In general low values of solution concentration, stirring speed, resin particle diameter, ion diffusion coefficient in solution and a highresin exchange capacity facilitate the film diffusion controlled mechanism.33Figure 2.12 : The laboratory columnUO.3O uo.6O . TJ1.OOr........-•••. -‘.‘.;‘:•-.. . .-:.. .-Figure 2.13 : A typical autodiagram (U=fraction extracted) [57]Figure 2.14: A typical X-ray microprobe analysis at 33% and 70% extraction [20]FeedElectrolyteUs-342.8 Solvent Impregnated ResinsIon exchange is today viewed as one unit process amongst several separation methods, and isclosely related to solvent extraction and membrane separation processes. Solvent extraction offersbetter selectivity, larger mass output, lower reagent costs, but greater operating costs and considerable risks. Membrane separation processes offer the highest dynamic output efficiencies,excellent safety features, though the reagent costs are high and the selectivity is not as good. Ionexchange performs well in concentrating metals from very dilute solutions with fairly goodselectivities, whereas for the similar operation, solvent extraction would pose the problem of highreagent losses, even though the selectivity would be better. There are, however, problems with ionexchange, such as hydrophobicity of the resin, insufficient mobility of the resin bound ligand, andmuch weaker complexation constants in the resin phase, which reduce extraction efficiencyconsiderably in very dilute solutions. This led to the development of special solvent impregnatedresins, which combine the selectivity and specificity of conventional liquid extractants, with theadvantages of a discrete polymer support material, thus tailor-making adsorbents for a specificseparation process. These resins may be prepared by physical impregnation of the reagent onto apolymeric or other porous support without chemical bonding of any sort. Alternatively, copolymerization of a monomer (eg. styrene)- crosslinking agent (divinylbenzene) in the presenceof a reagent (eg. tri-n-butyl phosphate or di-2-ethylhexyl phosphoric acid) will produce a polymerencapsulated product. Typical of these products are the Levextrel resins developed by Bayer [59].An exhaustive review of extraction with solvent impregnated resins has been published by Warshawsky [60]. The principal difficulty in the use of these materials is the slow diffusion of thereagent Out of the polymer matrix.352.9 Cobalt Purification2.9.1 IntroductionSuiphide ores mined by INCO Limited from the Sudbury basin, contain copper, nickel andsmall quantities of cobalt, as well as trace precious metals. During benefication, cobalt distribution to various streams is similar to nickel. Recovery of the metals is carried out in multi-stagemilling, smelting and refining processes. After smelting, nickel is selectively extracted bycarbonyl refining and the cobalt remains in the carbonylation residue. This residue is processedin a specialized hydrometallurgical plant [611 (the CRED plant at Copper Cliff, Ontario) toproduce precious metal residues, copper cathodes, and a nickel-cobalt carbonate intermediateproduct, containing about 20 % cobalt. This nickel-cobalt carbonate is shipped to the refineryat Port Colborne, Ontario, where the cobalt is separated to cobaltic hydroxide in the CobaltHydrate Plant and refined to metallic cobalt in the Electro-Cobalt Plant (refer Fig. 2.16 & 2.17for the process flowsheets).Nickel-cobaltcarbonate fromCopper CliffSulphuric acidChlorineSodium carbonate.Sulphur dioxideSulphuric acidSodium carbonateFigure 2.15 : Cobalt purification at Port Colborne• ChlorideSodiumNickelZincIronCopperLeadSodium sulphateCobalt metal362.9.2 Cobalt Hydrate PlantThe slurry of nickel-cobalt carbonate is leached in a batch process, in which pre-reducedspent electrolyte is mixed with sufficient nickel-cobalt carbonate, to replenish the spent electrolyte strength from 45 g/L to 95 g/L cobalt. Copper removal takes place initially by cementationand zinc removal by Bayer Lewatit CC 1026 resin, which is the focus of this study. Nickelremoval is then accomplished by the addition of chlorine and soda ash, which oxidises thecobalt(ll) to the cobalt(llI) state, and raises the pH to about 5.0, bringing about the hydrolyticprecipitation of cobaltic hydroxide. This upgraded cobaltic hydrate is now sent to the ElectroCobalt Plant.2.9.3 Electro-Cobalt PlantThe cobaltic hydrate purification circuit forms an integral part of the process because of thetwo-way transfer of some of the cobalt and impurities between the hydrate circuit and the refiningprocess (Fig. 2.15). The slurry of cobaltic hydrate is aerated to remove the Cl2 from the previousprocess and is then filtered and washed. Barium carbonate and sulphur dioxide are then addedwith NaOH. Barium carbonate is added to co-precipitate lead and the leach liquor copper, whilesulphur dioxide reduces the cobaltic hydroxide to the cobaltous form. At the endpoint of thisstep, a small amount of cobaltic hydrate is added to scavenge any excess sulphur dioxide andto oxidize the ferrous iron. The solids are returned to the cobalt hydrate plant for cobalt recoveryand rejection of impurities. Practically all the zinc and nickel, and a minor amount of copperremaining in the solution are polished down to required concentration levels by a series of ionexchange systems. Copper is removed using DOW XF 4195 resin in a single column. Nickel isalso removed by DOW XF 4195 resin with four columns in series. All ion exchange columnsoperate downflow in the flooded mode. The electrowinning process relies on several noveldevelopments and process conditions [63], which allow for electrowinning with a very highcobalt depletion from the all sulphate electrolyte, using bagged and hooded lead anodes of specialdesign and assembly. The37Figure 2.16 : The Cobalt Hydrate Plant [62]PORT COLBORNE REFINERYCOBALT HYDRATE PLANTI SLURRY LEACHING AND IRON REMOVAL38Figure 2.17 : The Electro-Cobalt Plant [62]PORT COLBORNE REFINERYELECTROCOBALT PLANTFEED RECEMNGCOBALT HYDRATE39electrolytic tankhouse atmosphere is made cobalt and acid$ree, and high purity cobalt roundsare produced on 304 stainless steel mandrels. Plating conditions are summarized in the followingtable:Table 2.6 : Plating conditions for cobalt electrowinning.Parameter ValueCurrent density (based on total cathode area) 200 A/rn2Temperature 60 °CCatholyte pH 3.5Cell voltage 4.0 VCurrent efficiency 92 %Plating time 5 daysThe one inch diameter rounds are washed and harvested through an automated system. Afterdrying, the rounds are processed on a continuous basis through a hydrogen degassing systemfollowed by storage and barrelling for sale in 250 kg. drums. The following table presents atypical analyses of process streams at various stages.Table 2.7 : Analyses of process streams at various stages of the flowsheet.Concentration at Stage (g/L) Cu Ni Fe Zn Pb Ca CoFeed Stock Slurry, g/L 8 45 0.8 0.03 0.00005 2 45After dissolution & Fe removal(g/L) 6 35 <0.003 0.02 0.00005 0.6 35After initial Cu & Zn removal(g/L) 0.005 35 <0.003 0.00 1 0.00005 0.5 35Redissolution and Co/Ni separation 0.000 1 1 <0.003 0.00 1 0.00002 0.5 95Feed to Electrowinning, g/L 0.0002 0.2 <0.003 0.0002 0.00002 0.5 90Electrocobak, ppm 1 500 3 5 5 3 99.94%2.10 Extractants for ZincThe primary recovery of zinc from leach liquors through solvent extraction or ion exchangeprocesses is presently limited to only a few plants. These processes are however used widely for40the removal of zinc impurities from various electrolytes and waste streams. The following sectionsdeal with the extractants used for zinc recovery or removal, and survey their chemistry, extractionequilibria and industrial applications.2.10.1 Di (2-ethyihexyl) phosphoric acidDi (2-ethyihexyl) phosphoric acid (hereafter referred to as D2EHPA in the text and as HLin equations) is an alkyiphosphoric acid which has become widely used in solvent extractionpractice, as in the separation and recovery of uranium, cobalt/nickel and rare earths. It hasadequate kinetics, moderate separation capability, low aqueous solubility, and is chemicallystable [66]. Its wide availability and low cost make it popular, compared to other reagents. ChemistryThe chemical formula of D2EHPA is as follows:R—O 0P/\R—O OHR = 2 ethyl hexyl groupFigure 2.18 : Sketch of the D2EHPA extractant (pK 1.2)It is a weak acid, and is relatively insoluble in water. The reaction between a metal cation andthe D2EHPA dimer may be expressed by the following generalized equation:Mm + n(HL)2 <—---* MLm(HL),, + mH . . .(2.33)41where M represents the metal being extracted, HL is the extractant (D2EHPA), m is thevalence of the metal cation, n represents the number of dimerized extractant molecules participating in the reaction. The D2EHPA molecule usually exists as a dimer in nonpolar media(i.e. in most aromatic and aliphatic solvents) and is monomeric in highly poiar media (i.e.alcohols, carboxylic acids and water). The D2EHPA dimer consists of two D2EHPAmonomers, joined by hydrogen bonds between adjacent P tO and P-OH groups, as shownbelow:R-—-OH RR NQH__Q NRFigure 2.19 : The D2EHPA dimerThe dimerization reaction is:— [(HL)2]2HL —-- (HL)2 ; K2= — (2.34)[HL]2Komosawa [70] reported K2 = 3.1 x 1Q4m3/kmol in heptane, and concluded that the dimerization was strong.D2EHPA exhibits significant interfacial activity, and Komosawa [70] studied the distributionof the monomeric extractant to the aqueous phase:HL <—-- HL ; Kd = (2.35)[HLIand determined that Kd = 1.6 x 1O for D2EHPA in heptane. The actual total amount ofextractant reporting to the aqueous phase is also affected by the acid dissociation constant Kawhich was reported [70] as ]Q151. Table 2.8 summarizes the equilibrium concentration ofextractant calculated to be present in the aqueous phase for the overall equilibrium:42(HL)2 ÷—--* 2HL E——* 2HL ÷-— 2H + 2U . (2.36)Table 2.8: Calculated equilibrium aqueous phase extractant concentrations.[(HL)2]=0.1 kmol/m3,K2=3.1 x i0m3/kmol, Kd1.6 X iO, Ka=10’pH [UI (kmol/m3) [HL] (kmollm3) [HL] + [U] (kmol/m3)1.0 1.1x107 3.55x107 4.65x1072.0 1.1 x 106 3.55 x i07 1.46 x 1063.0 1.1x105 3.55x107 1.14x1054.0 1.1x104 3.55x107 1.10x105.0 1.1x103 3.55x107 1.10x1032.10. 1.2 The Zn-D2EHPA systemThe extraction chemistry of zinc with D2EHPA has been studied by numerous workers[67-69]. Ajawin [67] reported the overall equilibrium as:Zn2 + 1.5(HL)2 <——* ZnL2H + 2H (2.37)At low metal loadings the equilibrium is reported as [68]:Zn2 + 2(HL) —---* ZnL2(HL) + 2H (2.38)Sastre and Muhammed [69] in a computer model treatment of the extraction equilibrium,showed that both the ZnL2H and ZnL2(HL) complexes are possible under variable loadingconditions. The cobalt and nickel equilibria are as follows [70]:Co2 + 2(HL) +—---. CoL2(HL) + 2H (2.39)Ni2 + 3(HL)2 —--. NiL2.2(HL) + 2H (2.40)Table 2.9 summarizes the extraction constants (Kex) reported (units are kmol 0.5 m .1.5 for theZn-complex and m3 kmol for the Ni-complex):43Table 2.9 : Extraction constants for various metal complexes.Metal ion Aqueous phase Diluent K, ReferenceZn (as ZnL2HL) (Na,FT)Cl04 kerosene 4.9 x 102 [691Zn (as ZnL(1-I )) (Na,H)Cl0 kerosene 7.6 x 102 [69]Co (Na,H)N03 heptane 4.0 x i0 [70]Ni (Na,H)N03 heptane 4.5 x i08 [70]Ritcey et al. [67] studied the D2EHPA system and concluded that the extraction order wasas follows (Fig. 2.20):Fe > Zn > Cu2 > Co > Ni2 > Mn2 > M? > Ca2.10.1.3 ApplicationsThe Zincex process [72,73] uses two solvent extraction circuits to recover zinc from pyritecinder leach solutions. An amine extractant is used in the first circuit to produce a purifiedzinc chloride solution. D2EFTPA is then used in the second circuit to extract zinc. The zinc isthen stripped with spent sulphuric acid electrolyte and recovered in a conventional electrowinning plant. Even after passing through the first circuit, some iron remains in the purifiedsolution, which is removed by bleeding some of the D2EHPA and treating it with strong HC1.The METSEP process [74,75] was developed to recover zinc and iron from galvanizingpickling solutions. Zinc is removed using an ion exchange column. The column is then elutedwith HC1 and the zinc is extracted with D2EHPA. Zinc is stripped using sulphuric acid, andthen electrowon. The Valberg process [73,75] is used to extract zinc from rayon manufacturingwaste waters. A plant in Sweden uses D2EHPA in kerosene in a two-step counter-currentprocess to reduce the zinc concentration from 0.2 g/L to less than 2 ppm. Sulphuric acid isused to strip the zinc from the organic solution. The resulting 80 g/L zinc solution is recycledback to the rayon spinning bath.44IIpHFigure 2.20 : Extraction order of various metals by D2EHPA [671C0UCw604020_t—n _.—V//o Vi/I IJI(‘I//• Co(Il)NUI)Cu(It)9 Zn(II)O Fe(III)2 3 4 5 6?Equiltbrium pHFigure 2.21 : Extraction order of various metals by Cyanex 272 [78]100C05oxLU0Figure 2.22 : Extraction order of various metals by Cyanex 302 [79]0 2 4 6 8Equilibrium pH452.10.2 Cyanex 272 - The phosphinic groupThe American Cyanamid Company in 1982, introduced Cyanex 272 analyzing 85% Bis(2,4,4-trimethylpentyl) phosphinic acid, which has the following chemical structure:R 0P/R OHCH3 CH3where R = CH -C-CH-CH-CH3 2 2CH3Figure 2.23 : Sketch of the Cyanex 272 extractant (pK = 6.37)Rickleton et al. [76] studied this reagent for the separation of cobalt and nickel, and determinedthe extraction order as in Figure 2.21, with the added ability to reject calcium. At low metalloadings, the extraction reactions form a tetrahedral Co-complex and an octahedral Ni-complexas follows:Co2 + 2(HL) <—— Co(HL2) + 2H (2.37)Ni2 + 3(HL)2 —-4 Ni (HL2)(HL)2 + 2H (2.38)Cyanex 272 extracts other transition metals - cadmium, copper and zinc from nitrate solutionsas follows [77]:M2 + (p +2)HL ——* ML2(H ) + 2H (2.39)The extracted zinc complexes in sulphate solutions are ZnL2(HL) and ZnL2(HL),with thepredominant species depending on pH and extractant concentration.462.10.3 Cyanex 302 - The thio-phosphinic groupCyanex 302, also an American Cyanamid product, analyzes 84% Bis (2,4,4-trimethylpentyl)monothiophosphinic acid and has the following chemical structure:R SP/\R OHCH3 CH3where R = CH -C-CH-CH-CH3 2 2CH3Figure 2.24 : Sketch of the Cyanex 302 extractant (pK 5.63)Its extraction order is shown in Fig. 2.22 [79] and has the advantage of being able to extractzinc at lower pH values. An added feature is that from a pH of 1.0 onwards, it extracts zinc inpreference to Fe(llI). In Figure 2.25, the zinc extraction performance of the three reagents iscompared, while Figure 2.26 compares the cobalt extraction. It is seen that Cyanex 302 performsbetter than D2EHPA which performs better than Cyanex 272 in terms of zinc extraction. Thecobalt extraction of Cyanex 302 and D2EHPA is comparable in the pH range 3.5 to 4.5, implyingthat Cyanex 302 could be considered to possibly replace D2EHPA based on Co/Zn selectivity.470UiI—0IUi0zt4Figure 2.25 : Comparison of zinc extraction by various organophosphorous derivatives. Organicphase: 20% extractants in kerosene, aq. lg/L zinc; A/O=1 [79]I1008060402002.5pHFigure 2.26 : Comparison of Cobalt extraction of the resins (Obtained from data in [67], [78]and [79], considered under similar conditions). Organic phase: 0.1 M extractant, aq. phase: 1 x10-3M.2 3 4pH3 3.5 4 4.5 5 5.5 6 6.5482.11 Scope and Objectives of this WorkFrom the review of the literature presented in the preceding sections, it seems that the reagentsCyanex 272 and Cyanex 302 perform a comparable zinc extraction to that of D2EHPA. Further,according to the Hard-Soft Acid-Base concept [14,78] (in Section 2.4.3), the complexation of asoft Lewis acid, such as Zn(II), Co(II), Cu(II) or Ni(ll), with a soft Lewis base should occur withhigh selectivity. The donor atoms of the common Lewis bases have electro-negativities increasingin the order S<Br<N<CkO<F. Sulphur substitution of the organophosphorous reagents shouldtherefore prove beneficial to the extraction of these metal ions.The project was initiated in consultation with INCO Ltd., to compare the resins OC 1026,Cyanex 272 and Cyanex 302 for their performance in the INCO zinc removal process. The researchobjectives may be summarized as follows:(1)To compare the performance of the resins in fixed bed column tests, on the basis of their zincbreakthrough points and the cobalt retained on the resins at this point.(2)To study the effects of resin pre-treatment, electrolyte pre-treatment, temperature, flow rate andloading cycles on zinc breakthrough and selectivity in column tests, and to explain the effectsbased on mechanisms operating in the column.(3)To compare the kinetics of zinc extraction from the electrolyte by the resins, in batch tests, andpropose a mechanism for the extraction.The methodology followed was to divide the experimental work into column and batch tests, andresults from these tests were fitted to kinetic models, to compare the resin performances and gainan insight into the process mechanism.493 EXPERIMENTAL3.1 ResinsThe resins used in this work- OC 1026, Cyanex 272 and Cyanex 302, collectively calledLevextrel resins, are styrene-divinyl benzene based co-polymers of a predominantly macroporousstructure that contain a selective extractant. Representative samples of these resins were obtainedfrom Bayer AG, Ltd.. Two of these resins, Cyanex 272 and Cyanex 302 were synthesized by BayerAG Ltd., for this work, and hence little information is available about their properties.3.1.1 Structure of the resinsFigures 3.1 and 3.2 show stereoscan photographs of OC 1026 [921. Each individual Levextrelbead consists of a large number of microbeads, which are firmly attached to one another. Thespace between them is filled with the active substance (extractants D2EHPA, Cyanex 302 orCyanex 272), which is retained by adsorption. The size of the microbeads and the space betweenthem depend on the type and volume of the extractantused, as well as on the degree ofcrosslinkingof the polystyrene beads.3.1.2 Particle size and distributionThe mean particle diameter of the respective resin fractions, was determined by photographing representative samples and measuring the diameter of each particle, and computingthe mean (Table 3.1). The particle size distribution, determined by wet screening of the resinsample is also reported in Table Physical propertiesThe physical properties of the resins were determined by conventional methods [93] and aresummarized in Table 3.1.50Figure 3.1: Surface structure of the fractured Levextrel OC 1026 bead [94]Figure 3.2 : Interior bead structure of the Levextrel OC 1026 bead [94]51Table 3.1 : Properties of the Levextrel resins OC 1026, Cyanex 272 and Cyanex 302.Resin Active Capacity Particle Size Bulk true Void Watersubstance (eq/L) size range density density fraction content(tm) (mm) (g/cc) (glcc) by wt.OC 1026 D2EHPA 0.5 700 0.3 - 1.0 0.5882 0.930 0.4883 0.3%Cyanex 272 Cyanex 272 0.41 450 0.4- 0.7 0.4882 1.023 0.5-Cyanex 302 Cyanex 302 0.44 510 0.4- 0.8 0.51 1.100 0.5-The bulk density of the resins is rather low and approximately in the range of those of themacroporous weakly basic anion exchangers. The specific gravity is lower than that of normalion exchange resins and they are hydrophobic. The beads also have a particularly low watercontent. The active substances are retained in the resin structure by adsorption, rather thanchemical bonding, and their losses to aqueous solutions in column operations are negligible,particularly in the acidic range. In the treatment of a 20% cobalt sulphate solution [89], OC 1026had an average loss of active substance of only 5 mg/L, and about 12 mg/L during the elutionof the loaded column, using 10% sulphuric acid. However, a greater loss of active substance isobserved in stir tests, especially at high stirring speeds or with magnetic stirrers.3.2 Reagents3.2.1 Cobalt advance electrolyteThe cobalt advance electrolyte was obtained from the INCO cobalt refinery at Port Colborne,Ontario. The analysis and properties of the electrolyte are as listed in Table 3.2:Table 3.2 : Analysis and properties of the cobalt advance electrolyte from INCO PortColborne.Co (g/L) Ni (g/L) Zn (g/L) Fe (g/L) Cu (g/L) Density kg/rn3 Viscosity kg/ms pH45 40 0.012 0.008 0.010 1284 at 25CC 5 x i0 at 25 C 3.34523.2.2 Zinc - free cobalt electrolyteZinc - free cobalt electrolyte, was used to pre-treat the resins in the batch tests. It was obtainedfrom the effluent solutions of the column tests, with zinc content of less than 0.5 ppm in theelectrolyte.3.2.3 Cobalt sulphate solutionThis was prepared by dissolving the appropriate amount of reagent grade cobalt sulphate(CoSO4.7H20),in deionized water and adjusting the pH to the required value.3.2.4 Zinc sulphate solutionThis was prepared by dissolving the appropriate amount of reagent grade zinc sulphate(ZnSO4.7H20),in deionized water.3.2.5 Sodium hydroxideSodium hydroxide was used in batch tests, for pH stabilization and maintaining the solutionat a particular pH value during pre-loading and loading. It was prepared by dissolving theappropriate amount of NaOH in deionized water, and then standardized against potassiumhydrogen phthalate (KHP), with phenophthalein as the indicator.3.2.6 Sulphuric acid10% sulphuric acid was used to strip the resins of cobalt and zinc during column elutiontests. This was prepared by adding 100 ml of bottle strength sulphuric acid to 900 ml of deionizedwater in a 1000 ml volumetric.3.3 Experimental ApparatusThe apparatus for column experiments (Figure 3.3), consisted of a chromatography column, aneedle valve to finely adjust the effluent flow rate to equal the influent flow, a Masterfiex pump53FeedElectrolytes-Jacket—packedresin bedaOZE”EHEZHeight of the column = 27 cm.Volume occupied by the resin = 25 mlFigure 3.3 : Apparatus for column studiesS = StirrerB = Auto-buretteE = pH ElectrodeVolume of vessel = 500 mlFigure 3.4 Apparatus for batch studiesPumpocogcoo oOOo°d’mThermostaticBathT = ThermocoupleM = Variable Speed Motor54(Cole Parmer model no. 7520-35), attached with a number of variable occlusion cartridges (modelno. 7519-60) to deliver multiple pumping at different flow rates, a Cole-Parmer polystat circulator,which maintained a thermostatic bath by circulating water of the particular temperature throughthe column jacket, and an Eldex Universal Fraction Collector (UFC 3462) to collect volumes ofeffluent at different time intervals. The chromatography column was 60 cm long and 1.5 cm indiameter, and was glass fritted at the bottom to hold the resins.Figure 3.4 shows the experimental set-up used for batch tests. The impellor (or stirrer S) wasmounted on a lab stand anchored at each end to prevent vibration. The beaker (500 ml capacity)was held thermostatic by the polystat circulator. The pH of the solution was maintained constantusing an autoburette B, which is part of an auto-titration system, consisting of a Radiometer PHM82 pH meter, an ABU8O autoburette and a TTT8O titrator. The volume of the titrant dispersed tomaintain constant pH was recorded. A lid with holes for the stirrer, pH probe, thermocouple andtitrant addition was used to seal the beaker and prevent any evaporation during the experiment.3.4 Experimental Procedure3.4.1 Column testsThe ion exchange resins used in this work were used as recieved, without rinsing withdeionized water (to prevent any extractant loss). The column was filled with 25 ml of therespective resin, keeping a low water level initially, so that the resin was not totally coveredwith water, as the resins tend to float due to the hydrophobic nature of the bead surface. Waterwashing (upflow) was used to remove all the air bubbles, and achieve a compact bed of 25 mlof resin. As the beads tend to float, the water was first drained out completely, and then a blockof glass wool was inserted from the top of the column on the bed, to maintain the compact bedand prevent any flotation of the beads during column operation. The bed was then filled withwater to just a few centimeters above the bed level.55The resin column was then attached to the externally mounted water jackets, and the polystatcirculator was set at the respective temperature and started. The Masterfiex pump was thencalibrated to deliver the particular flow rate for its variable occlusion cartridges. The pH meterwas then calibrated using buffers at pH of 1 and 7, at the operating temperature of the experiment,and the pH of about 100 ml of the cobalt advance electrolyte was then adjusted to 3, for useduring pre-treatment. The fraction collector was set so as to collect column effluent samplesevery 30 minutes.During pre-treatment, the cobalt advance electrolyte with pH adjusted to 3 was pumped throughthe resin bed at a flow rate of 4.67 BV/hr. (116 mi/hr.). The needle valve at the mouth of thecolumn was finely adjusted, such that the flow out of the column equalled flow in (through theMasterfiex pump). The pH of the effluent was measured continuously until it was equal to 2.At this point, the flow to the column was shut off, and the cobalt advance electrolyte wassimultaneously pumped into the column at the same flow rate to begin loading. The timer at thefraction collector was started, and effluent samples were collected every 30 minutes in test tubesat the fraction collector. The fraction collector was covered to prevent any evaporation. The pHof the effluent samples was determined, and the samples were then taken in acid washed sampletubes and diluted appropriately to analyze for cobalt, nickel, zinc, copper and iron, using atomicabsorption spectrophotometry.Before elution, the resin bed was first washed with 3 BV (75 ml) of deionized water, collectingsamples of 25 ml (every 12 minutes 50 seconds), and then the column was filled to a particularlevel with 10% sulphuric acid and left standing for two hours. The effluent was called the firststrip solution - Si. The bed was again washed with 3 BV (75 ml) of deionized water, collecting25 ml samples, and the stripping was carried out once again by filling the column with 10%sulphuric acid to the same level and analyzed as the second strip solution - S2. The bed waswashed again with 3 BV (75 ml) of deionized water, and the pre-treatment for the next loadingcycle was started.563.4.2 Batch tests500 ml of zinc-free INCO cobalt advance electrolyte (obtained from the effluent samples ofthe column tests, analyzing less than 0.5 ppm zinc), was transferred to the water jacketed beakershown in Figure 3.4. The polystat circulator was set to the particular temperature, and water atthat temperature was circulated to thermostat the beaker. A series of 1 mL syringes, with a mouthof less than 0.03 cm diameter (so that the resin particles could not enter it), were cleaned andarranged for ready use. The pH meter was then calibrated at that particular temperature and thenused along with the auto-titrator, to set the pH of the solution to 3 (or any other desired value),by the addition of the NaOTzl titrant. The lid was then placed on the beaker to seal it and thetemperature was maintained for some time. About 1 g of the respective resin was weighed outand added to the beaker. Simultaneously, the autotitrator and the timer were started and theamount of titrant added with time was noted. After about half an hour, the resins reached anequilibrium with the solution, and no further titrant addition was needed to maintain the pH atthe particular value. Now an appropriate amount of zinc pipetted from the zinc sulphate solutionwas added, so as to attain a zinc concentration of 12 mg/L in the beaker with 500 mL of theelectrolyte. The timer was reset and started again, and small aliquots of the electrolyte werewithdrawn after every minute, for the first ten minutes, and later every five minutes, in the 1mL syringes and analyzed for zinc.3.5 Solution AnalysisAs mentioned before, the NaOH solution was standardized by using potassium hydrogenphthalate; which was dried for 3-4 hours and then titrated against the unknown NaOH solution.The raffinates of the column tests and samples of the batch tests were analyzed for Co, Ni, Zn, Feand Cu, after appropriate dilution by atomic absorption spectrophotometry, with matched matrixstandards and simultaneous background correction. Results quoted here have a relative precisionof better than 2%.573.6 Reproducibility of the ResultsIn column loading tests, the reproducibility of the results was excellent, with almost the samebreakthrough points (interpreted as zinc effluent concentration of greater than 1 ppm) obtainedduring trial runs. In elution tests, the reproducibility was good if the elutions were run within anhour after the loading tests, and the volume of the wash water was kept constant during the runs.The reproducibility of batch test results was found to be greatly dependent on the impeller speed,the depth of the impeller head and the particle size of the resin samples. A slight change in any ofthese parameters resulted in quite different results than those reported. The qualitative trend of theresults was, however, similar. In particular, particle size and size distribution played a major role,as the reaction rate is inversely proportional to the square of the particle radius, and hence even asmall deviation in particle size between two resin samples, would show a measurable differencein their kinetic performance. Wet screening of the resin samples before their application in thebatch tests, could not be accomplished, as this resulted in loss of extractant, and hence the resultswould not be a true picture of the resin performance. As the exchange process was to some extentmass transfer controlled, small changes in impeller speed had an observable influence on the results.It was not possible to increase impeller speed sufficiently or to use baffles to bypass the masstransfer control region, as this resulted in breakage of the beads and spilling of the solution. Inspiteof these constraints, the reproducibility of the results was satisfactory. Even though the absolutenumbers varied by about 8%, the underlying mechanisms, determined from the trends in the diffusion and mass transfer coefficients, remained the same as those proposed here.584 KINETIC MODELSFrom a survey of the various rate laws that may be used to describe ion exchange kinetics inSection 2.6.3, the Fickian particle and film diffusion models and the progressive shell (unreactedcore) models, were found useful in describing the ion exchange kinetics of the system chosen. Thesemodels are briefly described here for purposes of this work. The reader is advised to refer to theliterature [15,20,80] for more detailed descriptions. A model is also outlined for kinetic studies oncolumn loading, based on the methods suggested by Doulah and Jafer [83] and Rodrigues and Costa[85].For the ion exchange equilibrium represented by:RA+B—---*RB+A (4.1)the general assumptions for all of the diffusion models are as follows:(1)The resins are spherical particles of uniform size.(2)There are no temperature gradients in the resin bead or the aqueous film.(3)The effective diffusivity in the resin phase and the mass transfer coefficient in the aqueous phasefilm are constant.(4)The concentration of the external solution is constant through the process (Infinite solution volumecondition).4.1 Fickian Film Diffusion ModelThis model assumes, in addition to the initially mentioned assumptions, that diffusion of ionstakes place through a quasi-stationary film (i.e. diffusion across the film is extremely fast whencompared to concentration changes at the film boundaries), and the film is treated as planar andone-dimensional (Fig. 4.1). The kinetic analyses in Section, equation (2.24) gives:59b *dcB 3Dc (CB — CB)4 2dt — r0 Swhere CB= concentration of B at the resin-solution interface.concentration of B at the solution bulk.c = total counterion concentration in the aqueous phase.= total counterion concentration in the resin phase.The term is evaluated by the ion exchange equilibrium:c = [Bb] (4.3)C (RB)eq[Bib concentration of B in the solution bulk.(RB )eq = concentration of B in the resin at equilibrium.In the initial state, the concentration of B in the bulk solution is equal to cb.,u, while that at theresin-solution interface is equal to zero:t0, rr, CBO (4.4)t=O, r>r0+S, CBC,.,B (4.5)The boundary condition is defined by the infinite solution volume condition, which states that thebulk solution concentration is maintained at a value defined by the exchange equilibrium, throughthe entire process:t > 0, r > r0 + 5, CB = [=JCiZB (4.6)The equation (4.2) is solved with the equilibrium condition (4.3) and initial conditions (4.4) &(4.5) and boundary condition (4.6) to give (X=fraction extracted):( 3D[BjbX(t) = 1 — exP[_rS(RB) jt] (4.7)60BulkSolutionFigure 4.1: The Film Diffusion ModelBulkSolution 181bFigure 4.2 The Particle Diffusion ModelLiquid SolidsFigure 4.3 : The Unreacted Core Model[RB][RB] -o[RA]eq614.2 Nernst-Planck Film DiffusionFrom Section and references [15,20j, the Nernst-Planck criterion for ions of equalmobility may be incorporated into equation (4.7) by including a selectivity parameter aBA for theexchange equilibrium in equation (4.1).CBCA==— (4.8)CA CBwhere the superscipt * represents the resin-solution interface. Hence:( i (3D[Blb 1ln(1—X) + I 1”j P( = —I — It (4.9)aA) r0ö(RB )eq aA)4.3 Fickian Particle Diffusion ModelDiffusion in the resin particle is rate controlling (Fig. 4.2), and hence from Section DB(j-r + (4.10)In the simplest initial condition, the concentration of B in the resin phase is zero and that at theresin-aqueous interface is a value of [Bib. Hence:t=0, O<r<r0, B° (4.11)t=0, r=r0, CBBb (4.12)The boundary conditions are defined by assuming infinite solution volume conditions, and hencethe resin-aqueous interface is at an equilibrium value defined by the exchange equilibrium:t>0, r=r0, B_(’)eq (4.13)Equation (4.10) is solved with initial conditions (4.11) and (4.12), and boundary condition (4.13),and along with the Vermeulen’s approximation (eq. 2.23) [30] to give:r ( [Bib EX(t) = I 1 — expi —____ —t I (4.14)L (RB)eq r )j624.4 Unreacted Core ModelThe theory of the unreacted core shrinking model (Fig. 4.3), has been presented in the literaturefor irreversible solid-fluid non-catalytic reactions [80-82]. For the equilibrium in equation (4.1),the material balances for the various processes of film diffusion, particle diffusion and chemicalreaction may be written as follows:Film Diffusion:1 dIVB 1 dN= km([B1b[B1s) (4.15)4itr0 dt 4itr0 dtDiffusion through the reacted resin layer:1dW— -(d[Bfl (416)4itr dt — e dr )rrChemical reaction:idN-- ——i= k[RA][B]— k2[RB][A] (4.17)4itr dtwhere dNjdt = infinitesimal change in the number of moles of component i.[II = molar concentration of the I-component.r0, r = radius of the particle and of the unreacted core respectively.De = effective diffusivity in the reacted resin layer.k1, k2 = rate constants of forward and reverse reactionsThese equations are solved to give the following rate expression for the fraction extracted (X):2De 3De 1 1 6De[Blb E—X + 1 + 2(1 —X) — 3(1 —X) +— I — 1 I = — t..(4.18)‘ k i FD ii (1 — Y\3 I 2FDD11r0L1J\‘ L) J roLiWjeqFor pure particle diffusion through the reacted layer:6D [Bib1 + 2(1 —X)— 3(1 —X)’3 = t..(4.19)r[RB]634.5 Kinetic Modelling of Column TestsKinetic data may be obtained from column tests, by the analysis of the breakthrough curves,which have a characteristic slope, time of appearance and mean time, that can be directly relatedto parameters influencing the performance of ion exchange equipment. The method is based onthe following assumptions [831:(l)The solute concentration wave in a solid bed of particulate sorbents, moves through the bedlength by forming zones of saturated particles, with a uniform solute concentration across thecrossection of the bed. Figure 4.4 shows the zone movement, for a typical breakthrough curve.(2)All zones in a bed are of equal length.(3)Fluid velocity in an operation is greater than the solid phase concentration phase wave velocity.(4)The fluid flows through the bed of particulate solids in a plug flow fashion, and longitudinalliquid diffusion is absent.Following the analysis in references [83-851, the parameters characteristic to a breakthrough curve;3, 0, and °m (defined below) may be related to solute concentration as follows:c 0—0= i— ex[_0m‘ I (4.20)where f3 = shape factor (dimensionless).00 = time of appearance of the breakthrough curve.Om = mean time of the breakthrough curve function.This may be written as:log[ln(lI(l—c/c0))] = 3log(O—00) — f3logO (4.21)Therefore from a log-log plot of ln(lI(l-c/c0))vs. (0-0), the slope of the straight line obtainedwould be f3, and the X-axis value of the point at which c/c0=0.632 (Figure 4.6) would be equal toOm ( as at c/c0=0.632, ln(l/(l-cIc0))=1 and from (4.21); O-0=9,,,). These characteristic parameters(f3, 0 and 9m) may be used to determine a range of kinetic parameters.64Figure 4.4: Typical breakthrough curve with the representation of the characteristic parameters and movement of the zonesFigure 4.5 : Variation of the rate function d(CIC0)/dO with 3ICo ICoExhaustione0 eE•11ri Sela — 12 Bela 23 Bela - 1.54 Bela - 2Bela - 36 Bela - 40-0 050 1-00 1-50Time65= 0.632slope beta0C)9Time DifferenceFigure 4.6 : Determination of characteristic parameters from a plot of equation (4.1).The overall mass transfer coefficient in the liquid phase (kf) is determined by:k1a 13/Om (4.22)where, a= mass transfer area, and for spherical particles:a = 1.5e2d (4.23)where, e= bed voidage (dimensionless); and d= resin particle diameter.The bed saturation time (Os) is given by:= Go + em (4.24)The number of exchange zones (Ni) is given by:Nz = (O + Om)/Gm (4.25)The height of an individual zone (hi) is given by: (where H = bed height)= HOm/(Go + Gm) (4.26)The velocity with which a mass transfer zone moves in a resin bed (Ui) is determined by:U = hz/Gm (4.27)If, U = superficial velocity of the fluid (mis); andp8=solid phase bulk density (kg/rn3), then theheight of a transfer Unit (Hf) is given by:HWf = UsOm/1366and the saturation capacity (q0) is given by:q0 UsC0(Oo+Om)/(HPB) (4.29)The distribution constant (DC) is given by:DC = q0/c = U(e0+ Om)I(PBH) (4.30)The degree of bed saturation (D1,) at the breakthrough point is:Db = = O/(o + em) (4.31)The unsaturatedfractional bed length (H,,) at the breakthrough point is:( e0H = H1— e + 8 J (4.32)Equations (4.21) to (4.32) may be used to determine the respective parameters, which are extremelyuseful in the design of the ion exchange columns. Figure 4.5 shows the variation of the rate functiond(c/c )—-‘- against time (0), and it is seen that determines the shape of the rate curve, which is directlyinfluenced by the mechanism of mass transfer. It was concluded [83] that for values of 3 less than1.5, particle diffusion is the rate controlling step; while for values greater than 3, film diffusion isthe rate controlling step. The value of the overall mass transfer coefficient, kfobtained from equation(4.22), may be compared to the value of kL (liquid phase mass transfer coefficient) obtained fromCarberry’s correlation [86]:/ ‘-2J3f‘kL = 1.151 PL U P-L (4.33)IlLe ) PL’LJ e}where PL liquid phase density (kg/m3).1k = liquid phase viscosity (kg/ms).d = diameter of the resin particle (m).superficial velocity of the fluid (mis).DL = liquid phase diffusivity (m2/s).e = bed voidage67The liquid phase diffusivity is determined by correlations with the liquid film thickness, 6 [15],which is defined empirically for fixed bed columns with spherical resin beads, operating at lowflow rates as [87]:0.2r6= (1+70rU) (4.34)Furthermore, breakthrough curves may be used to determine whether the exchange equilibrium isfavorable or not, depending on whether the curve remains constant or flattens respectively, withincreasing bed depth [15]. Finally, the variation of the mid-point slope of the breakthrough curvewith flow rate, would be independent of the feed spatial velocity for particle diffusion controlkinetics, while the variation would depend on the square root of the feed spatial velocity for a filmdiffusion control mechanism [20].4.6 Minimum Superficial Velocity of Solution in Column OperationsIn most column processes, the superficial velocity of the solution should be fixed at a value,at which particle diffusion kinetic control is the dominant mechanism. This is possible at such avalue of U5, at which:km >> /6 (4.35)From Perry’s Chemical Engineering Handbook, 4-th Edition [88], the following relation for masstransfer coefficient is obtained:5.45U(1 — e) (‘ DL °‘ (DLpLka = I I I I (4.36)mp iii Ir0 \r0l.J) flLThis may be rewritten as follows, to obtain the minimum superficial velocity for column operation(Umin), for the column to operate in the particle diffusion control regime.r 1 ‘-o.si ,, ‘\_O.1212.08u =Imczpr0 LIL LILPL. .. .(4.37)[5.45(l—e)2r0) ilL ) J685 RESULTS AND DISCUSSIONThis chapter presents the results of the work, and a discussion on the possible mechanismsoperating in these processes, and their outcome during industrial applications. Column and batch testresults are presented, followed by a general discussion on the exchange process. A greater effort ismade to the analyze data obtained from tests on OC 1026, as this was the more efficient resin, andis being used in the INCO zinc removal process at Port Colborne.5.1 COLUMN TEST RESULTS5.1.1 Zinc loading on untreated OC 1026Initial trial runs were carried out bypassing INCO electrolyte through a 25 ml bed ofuntreatedOC 1026 at various flow rates and temperatures, and the zinc concentration and pH of the effluentwere followed until about 250 bed volumes (BV) of the electrolyte had been passed. During theinitial feeding, a bright blue coloration, probably from the Co/HDEHP complex, spread slowlydownwards. This coloration, however, disappeared slowly to become a lighter color as thefeeding progressed.Fig. 5.1 shows the results obtained by using untreated OC 1026 at 40°C with a feed flow rateof 5 By/hr.. The electrolyte pH drops from the initial value of 3.7 at 40°C to a value of about2, after which it rises rapidly in the first 50 By, and then at a slower rate, until it reaches a valueof about 2.75 after 250 BV of electrolyte have passed. The loading of zinc on the resin seemsfairly efficient, as breakthrough (interpreted as a zinc effluent concentration greater than 1 ppm)takes place only after about 210 BV of electrolyte have been passed. Fig. 5.2 shows the resultsof the test carried out using untreated OC 1026 at 40°C, but with a flow rate of 10 By/hr.. Theprofiles seem similar, but the breakthrough takes place a little earlier in this case, after 190 By.Fig. 5.3 shows the profiles for a test carried out using untreated OC 1026 at 60°C with a flowrate of 4.67 BVIhr.. Zinc breakthrough takes place much later here, after about 320 BV, showing6913-J—.. 12Eu1o043.5C., ci)“ DNuJ2.521.5300Bed Volumes of EffluentFigure 5.1 : Zinc loading on untreated OC 1026 at 40°C and 5 BV/hr.. INCO electrolyte feed(pH 3.7 at 40° C, [Znj=12 mgfL), flow rate=5 BV/hr., OC 1026 bed =25 ml.0 50 100 150 200 25043.51312E— 10CC.)0 1.5300Figure 5.2: Zinc loading on untreated OC 1026 at 40°C and 10 BV/hr.. INCO electrolyte feed(pH 3.7 at 40°C, [Znj=12 mgfL), flow rate=10 BV/hr., OC 1026 bed = 25 ml.4-Cci)“ DN-NuJN02.520 50 100 150 200 250Bed Volumes of Effluent70that an increase in temperature enhanced zinc loading. The pH of the electrolyte falls from aninitial 3.75 to about 1.9 and then rises rapidly in the first 80 By, and then slowly up to a finalvalue of 2.94 after 500 BV of electrolyte have been passed. The results from the tests aresummarized in the following table:Table 5.1 : Results of Zn-loading on untreated OC 1026 from INCO electrolyte.Temp.°C Flow rate Initial pH pH after 1 pH after pH after Break(By/hr.) BV 250 BV 500 BV through40 10 3.7 2.25 2.94 - 19OBV40 5 3.7 2.1 2.77- 21OBV60 4.67 3.75 1.9 2.76 2.94 320 BVThus it seems from these results that a decrease in flow rate from 10 By/hr. to 5 By/hr. marginallyenhances zinc loading, while an increase in temperature from 40°C to 60°C shows a significantincrease in zinc loading. The loading of zinc and cobalt (or any other clivalent metal cation)proceeds through a reaction of the type:M2+n(HL) —-4 ML2.H+2Hwhere M = Zn or CoThus the extraction of the metal results in a decrease in pH. A lower pH may thus be interpretedas greater extraction of metal cations by the resin, and hence it may be concluded from theseresults that the extraction capacity of the resins is higher at a lower flow rate and at a highertemperature. This does not yet tell us anything about the selectivity of the resin under theseconditions.5.1.2 Zinc loading on pre-treated OC 1026Two kinds of pre-treatment were carried out - one by using INCO electrolyte at pH 3 forpreloading, as carried out at the Port Colborne Refinery, and one using CoSO4at pH 4 to7113121110987654321-Jc,)ECC04-’c5ICC)C)C0C)4-’CC)Dw43.5Ca)DU]02.5k20 1.5600Figure 5.3 : Zinc loading on untreated OC 1026 at 60°C and 4.67 BV/hr.. INCO electrolyte feed(pH 3.75 at 60°C, [Zn]=12 mgfL), flow rate=4.67 BV/hr., OC 1026 bed = 25 ml.0 100 200 300 400 500Bed Volumes of Effluent13-j— 120)EuCC80asL.4-CC)05C04C)4.-Cci)D2Wi43.5C3 a)DU]02.5k20 1.5500Figure 5.4 : Zinc loading on pre-treated OC 1026 at 60°C. INCO electrolyte feed (pH 3.75 at60°C, [Zn]=12 mgfL), flow rate=4.67 BV/hr., OC 1026 bed =25 ml, resin pre-treatment : INCOelectrolyte fed at pH 3 until effluent pH 2.0 100 200 300 400Bed Volumes of Effluent72preload cobalt. Preloading INCO electrolyte at pH of 3 (Fig. 5.4), seems to enhance the zincloading considerably as breakthrough takes place only at about 400 BY, while the fall in theeffluent electrolyte pH is smaller; from 3.75 to about 2.35, from which it rises very rapidly toabout 3.5 in 50 BY, after which it levels off at about 3.57. Pre-treatment at pH=4 and 25°C (Fig.5.5) was not very successful in improving zinc loading as breakthrough takes place as early as90 BY.5.1.3 Effect of pre-treatment on zinc loading and selectivityThe efficiency of zinc loading is judged by how much later the breakthrough of zinc takes placethrough the column, while the selectivity of the resin for zinc over cobalt is determined bystripping the loaded resin and computing the amount ofcobalt and zinc on the resin. This analysis,however, assumes that all the cobalt and zinc loaded on the resin, is completely stripped offduring the elution run. The values of cobalt retained on the resin are those determined using10% sulphuric acid as the stripping solution (the cobalt in the wash water was not significantlydifferent, and hence not taken into account). Table 5.2 summarizes the results obtained in thestudy of the effect of pre-treatment on zinc loading.Table 5.2 : Effect of pre-treatment on zinc loading and the pH profiles.Treatment Flow Initial pH after pH after 50 pH after pH after Break-rate pH 1BV BV 250 BV 500 BV throughUntreated 4.67 3.75 1.9 2.5 2.76 2.94 320 BYPretreated 4.67 3.75 2.35 3.5 3.56 3.57 408 BYFig. 5.6 shows a comparison of the zinc loading and pH profiles for the untreated and pre-treatedresins. The concentration of the INCO feed electrolyte is as follows:7313-Jj— 12Eu‘210C807c:s4-C05C0404-CCDw013-Jc,)‘1OC803.73.63.5CLU3.3:!.3.23.13120Bed Volumes of EffluentFigure 5.5 : Zinc loading on pre-treated (pH 4) OC 1026 at 25°C. INCO electrolyte feed (pH3.34 at 25°C, [Zn]=12 mgfL), flow rate=5 BV/hr., OC 1026 bed =25 ml, resin pre-treatment:CoSO4fed at pH 4 for 12 hours.00 20 40 60 80 10043.5CDwI,-.02.5k21.5600Figure 5.6 : Effect of pre-treatment on pH profile and zinc loading. INCO electrolyte feed (pH3.75 at 60°C, [Zn]=12 mg/L), flow rate=4.67 BV/hr., OC 1026 bed =25 ml, resin pre-treatmentINCO electrolyte fed at pH 3 until effluent pH 2.0 100 200 300 400 500Bed Volumes of Effluent74Table 5.3 : Analysis of INCO electrolyte.Metal Co Ni Zn Fe Cuconc (g/L) 45 40 0.0 12 0.008 0.01As the content of zinc in the electrolyte is low (12 mg/L), its extraction alone by the resin shouldnot result in much of a change in the pH of the effluent. Thus much of the decrease in pH of theeffluent observed is due to the extraction of cobalt and so a lower pH may be interpreted asgreater cobalt extraction. In the comparison of the pH profiles in Fig. 5.6, it is seen that the pHfalls to a lower value of 1.9 for the untreated resin, when compared to a value of 2.35 for thepre-treated resin, thereby indicating that the cobalt extraction is higher initially for the untreatedresin. As the loading progresses, the pH profile for the untreated resin rises only gradually upto 2.94 after 500 BV, while for the pre-treated resin the pH rise is extremely rapid and levelsoff at 3.6 after 500 By. Thus even as the feeding progresses, the cobalt retention on the untreatedresin is higher. Thus pre-treatment has an effect ofreducing the initial drop in pH, and maintainingthe pH of the column between 3.75 and about 3.6, from 50 BV until the end (500 By). Thislevelling of the pH possibly plays a role in delaying the breakthrough of zinc from 320 BV toabout 408 By, and in addition results in a reduction of cobalt retained on the resin as shown inTable 5.4.Table 5.4 : Effect of pre-treatment on cobalt retained on the resins(25 ml of loaded resin stripped with 10% sulphuric acid).Treatment Co (gms.) Zn (gms.) Zn/Co selectivityUntreated resin 0.26 0.17 0.65Pre-treated resin 0.23 0.20 0.87755.1.4 Cobalt and zinc retained on pre-treated resinThe loading of the resin pre-treated at pH 3, was carried out till 233.5 BV, 350.25 BV and 500BV of electrolyte were passed, followed by stripping, to determine the amount of cobalt andzinc retained on the resins after each of these runs. The results in Fig. 5.7 and Table 5.5 showthat there is a continuous decrease in the amount of cobalt retained on the resin and a continuousincrease in the zinc on the resin with increasing volume treated.Table 5.5: Comparison of metal retained on OC 1026 after various effluent bed volumes.BV of effluent Co (g) Zn (g) Zn/Co selectivity233.50 0.45 0.059 0.013350.25 0.37 0.116 0.314500.00 0.23 0.200 0.870These data seem to suggest that cobalt was initially loaded on the resin, and then was displacedby the zinc. This is substantiated by the following mass balance:Between 233.50 BV and 350.25 BV:Total reduction in cobalt retained on the resin = 0.080 gTotal increase in zinc loaded on the resin = 0.05 8 gFrom the breakthrough curve; amount of zinc loaded from thesolution on the resin = 0.060 gSimilarly between 233.5 BV and 500 BV:Total reduction in cobalt retained on the resin = 0.200 g760.60)-Dci)ci)ciE0.5E0.10Figure 5.7 : Stripping of zinc and cobalt from OC 1026. Comparison of metal retained in 25 mlof loaded OC 1026; after various effluent bed volumes, determined by stripping the loaded resinby 10% H2S04at 60°C.233.5 350.25 500Effluent Volume (BV)0.20__0.150.100.050Figure 5.8 Effects of electrolyte pH increase and pre-treatment on metal retained on the resin.Comparison of metal retained in 25 ml of loaded OC 1026, after resin and electrolyte pretreatment, determined by stripping the loaded resin by 10% H2S04at 60°C.Electrolyte pH increase Untreated Pre-treatedNature of treatment77Total increase in zinc loaded on the resin = 0.141 gFrom the breakthrough curve; amount of zincloaded from the solution on the resin = 0.150 gThus these data seem to confirm that the resin initially loads cobalt, which is displaced by zincas the feeding progresses. Correlating this with the observation that the electrolyte pH is low atthe early effluent bed volumes, while it increases for the later effluent bed volumes, it may beconcluded that cobalt loads on the resin initially at low pH values (between 1.8 - 2.5), while thiscobalt is displaced by the zinc at higher pH values (between 3.5 - 3.6). Resin pre-treatmentresults in the column being maintained between 3.5 and 3.7 for the major part of the loading(after the initial 50 BV of electrolyte have been fed), and thus zinc loading on the pre-treatedresin is enhanced as against the untreated resin where the column is at a pH range between 2.6and Effect of flow rate on zinc loading and selectivityResults of tests carried out to determine zinc loading at flow rates of 5 BV/hr. and 10 BV/hr.are summarized below:Table 5.6: Effect of flow rate on zinc loading and selectivity:(Loading of untreated resin at 40°C).Flow rate Initial pH at pH @ Break- gm. Co on gm. Zn on Zn/CopH 1BV 250BV through resin resin selectivity5 By/hr. 3.7 2.10 2.77 210 BV 0.150 0.140 0.931OBV/hr. 3.7 2.25 2.94 19OBV 0.105 0.136 1.30Thus it is seen that even though zinc loading is enhanced by operating at a lower flow rate, thecobalt retained on the resin is lower when operated at 10 BV/hr. flow rate. This may be becausethe electrolyte pH is higher than that of the range in which cobalt loads favorably (1.9-2.3),when the column is operated at the higher flow rate.785.1.6 Effect of temperature on zinc loading and selectivityThe effect of temperature on the loading process was studied at temperatures 60, 40 and 25°C,with the resin pre-treated at a pH of 3, and the results are shown in Fig. 5.9 and Table 5.7.Table 5.7 : Effect of temperature on zinc loading and selectivity of pre-treated OC 1026.Temperature Breakthrough gm. Co on resin gm. Zn on resin Zn/Co selectivity60°C 408 0.230 0.200 0.87040°C 250 0.144 0.145 1.00725°C 140 0.110 0.110 0.909Lowering the temperature hastens zinc breakthrough, but the amount of cobalt retained on theresin is reduced. Zinc selectivity is highest at 40°C. Hence, operating at 40°C could be consideredwhen it is of the utmost importance to improve selectivity.5.1.7 Effect of increasing feed electrolyte pHAs a test run, the pH of the feed electrolyte was raised to 5, and this was fed to an untreated bedof OC 1026 at 10 BV/hr. and 40°C. Zinc breakthrough took place only at 250 By, and this iscompared with the performance of the pre-treated and untreated resin (Fig. 5.10) as follows:Table 5.8: Effect of increasing feed electrolyte pH on loading (T=40°C).Treatment Flow rate Breakthrough Co retained Zn retained Zn/Co selectivityUntreated 10 BV/hr. 190.00 BV 0.150 0.14 0.900Pre-treated 5 BVfhr. 250.00 BV 0.144 0.145 1.007Elect. pH increase 10 By/hr. 250.25 BV 0.120 0.16 1.333Increasing electrolyte pH thus works as well as resin pre-treatment to enhance zinc loading,with significantly higher Zn/Co selectivity.7912-J0)CNC008642Bed Volumes of EffluentFigure 5.9 Effect of temperature on zinc breakthrough of OC 1026. INCO electrolyte feed (pH3.75 at 60°C, [Zn]=12 mgfL), flow rate=4.67 BV/hr., OC 1026 bed =25 ml, resin pre-treatmentINCO electrolyte fed at pH=3 until effluent pH 2.131211-Jg7CCPLu054.54S3.5C32.521.5600Figure 5.10 : Effect of electrolyte pH increase on zinc breakthrough of untreated OC 1026.INCO electrolyte feed pH raised to 5.0 with NaOH, flow rate=10 By/hr., temperature=40°C,OC 1026 bed =25 ml, no resin pre-treatment.0 100 200 300 400 5000 100 200 300 400 500Bed Volumes of Effluent805.1.8 Performance of Cyanex 302 and Cyanex 272The performance of Cyanex 302 and Cyanex 272 was studied by testing them under conditionssimilar to those at which tests on CC 1026 were carried out. The resins were pre-treated byfeeding INCO electrolyte at a pH of 3, and as the pH reduction for both resins was not as greatas that of CC 1026, the effluent pH in both cases was equal to 2.5. For the Cyanex 302 column,zinc breakthrough took place after 175 BV (Fig. 5.11), while for Cyanex 272 breakthrough tookplace after 160 BV of electrolyte (Fig. 5.13). However for Cyanex 302, the resin was poisonedby the loading of copper and a little iron, which formed a dirty brown band at the top 4 ml ofthe resin column, and was not stripped off easily even by 10% HC1. The effect of this poisoningon further loading cycles is shown in Fig. 5.12, where the breakthrough took place earlier aftereach loading cycle, and the poisoned brown band spread slowly but continuously downwardthrough the resin bed.5.1.9 Comparison of cobalt retained on the resinsTable 5.9 and Fig. 5.16 show a comparison of cobalt retained on the resins after various effluentbed volumes.Table 5.9: Comparison of cobalt retained on the resins.BV of effluent Co on Zn on Co on Zn on Co on Zn onCC 1026(g) CC 1026(g) Cy302(g) Cy302(g) Cy272(g) Cy272(g)233.5 0.450 0.059 0.150 0.055 0.185 0.050350.25 0.370 0.116 0.135 0.067 0.178 0.060500 0.230 0.200 0.125 0.073 0.170 0.070Zn/Co selectivity 0.870 0.5 84 0.411at 500BVThus, it is seen from these data that even though the cobalt retained on the resin after 233.5 BVof electrolyte was much higher for OC 1026 than that for the other resins, after 500 BV ofelectrolyte,813.532.52413-J-. 12C)Eli‘ElO04-,Cci)050404-,ci)D9-9-uJ0 ‘.5600Figure 5.11: Zinc loading on pre-treated Cyanex 302. INCO electrolyte feed (pH=3.75 at 60°C,[Zn]=12 mg/L), flow rate=4.67 BV/hr., Cyanex 302 bed = 25 ml, resin pre-treatment : INCOelectrolyte fed at pH 3 until effluent pH 2.5.0 100 200 300 400 500Bed Volumes of Effluent5-J0)E.—; 4CN0250Figure 5.12 : Effect of resin poisoning on zinc breakthrough of Cyanex 302. INCO electrolytefeed (pH 3.75 at 60°C, [Zn]t=12 mgfL), flow rate=4.67 BV/hr., Cyanex 302 bed = 25 ml, resinpre-treatment: INCO electrolyte fed at pH 3 until effluent pH 2.5.0 50 100 150 200Bed Volumes of Effluent8212 3.81o0.C.)3.63.4•1-Ca):2:2.82.6600Bed Volumes of EffluentFigure 5.13 : Zinc loading on pre-treated Cyanex 272. INCO electrolyte feed (pH 3.75 at 60°C,[Zn]=12 mgfL), flow rate=4.67 BV/hr., Cyanex 272 bed = 25 ml, resin pre-treatment : INCOelectrolyte fed at pH=3 until effluent pH 2.5.0 100 200 300 400 5000.60.59O.4a)0.3a)E 0.20.10Figure 5.14 : Comparison ofcobalt retained in 25 ml of loaded OC 1026, Cyanex 302 and Cyanex272 (for a similar pre-treatment), determined by stripping the loaded resin by 10% H2S04at60° C.233.5 350.25 500Flow Rate (in By/hr.)83the value of cobalt retained on OC 1026 is comparable to that of the other resins. Yet, in termsof amount of cobalt loaded, Cyanex 302 seems to perform better than OC 1026, even thoughzinc loading and selectivity are better in the latter. The reason for lower amounts of cobaltloading on pre-treated Cyanex 302 and Cyanex 272, may be that the pH never drops below 2.6during the loading of these resins, while the pH drops to about 2.3 for loading on pre-treatedOC 1026. From the comparison of the pH profiles in Figure 5.6, it was concluded that zinc loadson the resin only as the pH rises, while cobalt continues to load on the resin even at lower pHvalues, and hence cobalt retention by pre-treated OC 1026 is higher than that by Cyanex 272and Cyanex 302.5.2 BATCH RESULTS5.2.1 Kinetics of zinc loading on OC 1026The kinetics of zinc loading on OC 1026 was followed in batch loading up to 500 minutes bydetermining the zinc concentration in the withdrawn aliquots at the particular time intervals.Fig. 5.15 is a plot of fraction of zinc extracted with time. It shows an extremely small incubationtime followed by a rapid rise in the first 50 minutes, levelling off gradually. Studies were madeon the types of pre-treatment possible, temperature variations and variations in the initial zinccontent of electrolyte based on these kinetic results.5.2.2 Effect of pre-treatmentFig. 5.16 shows a comparison of the effects of various pre-treatments on the kinetics of zincloading on OC 1026. All pre-treatments seem to enhance the kinetics of zinc extraction, whilepre-treating at pH of 3.5, 3 and 2, provide the maximum improvement; the difference betweenthem being marginal (the curves almost overlap). For the case of Cyanex 302 (Fig. 5.19) pretreatment at pH 2 was the most favorable pre-treatment. In a comparison of the performance ofOC 1026 and Cyanex 302 based on pre-treatments in Figure 5.26,8410.8C.)CC•C 0.4C)0.210CC8CCdD6 C• —CN4120 100 200 300 400 50020 0600Time (minutes)Figure 5.15 Batch test study of zinc loading kinetics on untreated OC 1026; initial [Zn] = 12mg/L, pH 3.75, T = 6OC.10.8C)C0•- 0.40.20600Time (minutes)Figure 5.16 : Effect of pre-treatment on the kinetics of zinc uptake. Batch test study of zincloading on pre-treated OC 1026; initial [Zn] = 12 mg/L, pH 3.0, T = 60CC. Pre-treatmentEquilibrating resin in zinc free INCO electrolyte at pH 3.0 100 200 300 400 50085it was found that the performance of OC 1026 and Cyanex 302 are almost comparable whenpre-treatment was done at pH 3, while for pre-treatment at pH 2, Cyanex 302 performs betterthan OC 1026.5.2.3 Effect of initial zinc contentIn the study shown in Fig. 5.18, the kinetics of zinc extraction at 40 mg/L zinc is compared tothat at 12 mg/L. The increase in zinc content seems to make almost no difference in the zincloading on OC 1026 in the first 20 minutes, after which time the kinetics of zinc loading areconsiderably increased at the higher initial zinc content. This seems to suggest that the initialrate limiting step could be particle diffusion (where diffusion rate is independent of concentration), while the dominant mechanism after 20 minutes into the process is film diffusion control(which is dependent on concentration).5.2.4 Effect of temperature on the kinetics of metal uptakeThe kinetics of metal extraction was studied for OC 1026 in Fig. 5.17, for Cyanex 302 in Fig.5.20 and for Cyanex 272 in Fig. 5.26. In all cases the kinetics are slackened by a lowering oftemperature. In an attempt to quantify the influence of temperature and other parameters, andhelp determine the loading mechanism, the kinetic data was fitted to the models described inChapter 4.5.3 Kinetic Model Fit for the Batch TestsFigures 5.21 and 5.22 show one example of the fitting of the experimental results for the loadingof zinc on OC 1026, pre-treated at pH 3. No single model seems to be able to successfully fit theexperimental data. From Figure 5.22, it is seen that the data obey the Fickian Particle DiffusionControl (PDC) model, with a resin phase diffusivityD = 9.Ox102m2/s, for the first ten minutes,and then from 20 minutes onwards, obeys the Film Diffusion Control (FDC) model, with a masstransfer coefficient, k= 2.7x i07 rn/s. Finally after about 150 minutes860.800.60.40.20Time (minutes)600Figure 5.17 : Effect of temperature on the kinetics of zinc uptake. Batch Test Study of zinc loadingon pre-treated OC 1026; initial [Zn] = 12 mgfL, pH 3.0. Pre-treatment : Equilibrating resin in zincfree INCO electrolyte at pH=3.20,._. 150600Figure 5.18 : Effect of initial zinc concentration on the kinetics of zinc uptake. Batch Test Studyof zinc loading on pre-treated OC 1026, pH 3.0, T = 60CC. Pre-treatment : Equilibrating resin inzinc free INCO electrolyte at pH 3.0 100 200 300 400 500Time (minutes)87100. 100 200 300 400 500 600Time (minutes)Figure 5.19 : Effect of pre-treatment on kinetics of zinc uptake by Cyanex 302. Batch Test Studyof zinc loading on pre-treated Cyanex 302, Initial [Zn] = 12 mgfL, T = 60°C. Pre-treatment:Equilibrating resin in zinc-free INCO electrolyte at particular pH. 100 200 300 400 500Time (minutes)600Figure 5.20 : Effect of temperature on kinetics of zinc uptake by Cyanex 302. Batch Test Studyof zinc loading on pre-treated resin, Initial [Zn] = 12 mg/L, pH 3.0. Pre-treatment: Equilibratingresin in zinc free INCO electrolyte at pH 3.88(Figure 5.21), the data fall between the Fickian Particle Diffusion Control (PDC) model, withD = 1.5x 10”m2/s, and the unreacted core diffusion model with D = 7.3x102m2/s. Results forother batch tests seemed to almost follow the same trend. It was possible to determine accuratelythe diffusion and mass transfer coefficients only for the first 150 minutes, as after that the deviationsfrom the models are too large.Table 5.10 lists the D and km values obtained for the Fickian PDC and FDC models for the datafrom the three resins at a given pre-treatment. It is seen that the difference between the D and k,r,values, of the three resins for a similar pre-treatment is not significant. Hence the kinetic performance of the resins is comparable (also seen in Fig. 5.26).Table 5.10: Kinetic model fitted to determine D and km values.Resin (Operating temp.=60°C) D (m2/s) km (m/s)Untreated OC 1026 1.0 x 1012 1.6 xPre-ireated (pH=3) OC 1026 9.0 x 1012 2.7 x i07Untreated Cyanex 302 9.0 x iO 1.0 x i07Pre-treated (pH=3) Cyanex 302 8.5 x 1012 2.1 x i07Pre-treated (pH=3) Cyanex 272 8.0 x 1012 1.8 x i0Pre-treatment seems to enhance the D values of the resins by nearly an order of magnitude, whichleads to a slight increase in the k. values in the aqueous film. The effect of temperature on the Dand k., values is shown in Table 5.11. For infmite dilution, the temperature dependence of theliquid diffusivity is usually accounted for by assuming that the term DLTfT is constant (D=liquidphase diffusivity, =viscosity, T=temperature). For more concentrated solutions, the exponentialdependence of DL on T has been suggested [881:8910.8><0a)0.64-xwo 0.4U-0.20Time (minutes)600Figure 5.21 : Kinetic model fit of the experimental data for zinc loading on pre-treated (pH=3) OC1026 at 60°C. Model A=Fickian PDC with D =9.0 x 1012 mIs; Model B=FDC, lç1=2.7x iO mis;Model C=Fickian PDC, D=1.5 x 10h1 m2Is; Model D=Unreacted Core PDC.10.8-oG)30.6xwC04-0cU-0.20100Figure 5.22 : Kinetic model fit of the experimental data for zinc loading on pre-treated (pH=3) OC1026 at 60°C, for the first 5000 s. Model A=Fickian PDC with D =9.0 x 1012 mIs; Model B=FDC,k,=2.7 x io rn/s.0 100 200 300 400 5000 20 40 60 80Time (minutes)90RT.D. =. (5.1)IzJF2where ? = ionic equivalent conductance.z1 = valency of the i-ion.F = Faraday’s constant.. is a function of temperature:= A1(25°C) T (5.2)11where : r, = viscosity of water at T.As 2 and ri depend on temperature, it was concluded [88,91] thatD1depends exponentially on T.The temperature dependence of the resin phase diffusivity should also follow the same trend[27,88,91]. Arrhenius plots were therefore made of the diffusion coefficients and mass transfercoefficients for the three resins (Figures 5.23 and 5.24). The activation energy for diffusion, Ea,was determined from the slope of the line obtained, and is listed in Table 5.11.Table 5.11 : Effect of temperature on D and km values.Temp. OC 1026 Cyanex 302 Cyanex 272(UK) D km D km 13 km333 9.0 x 1042 2.7 x i07 8.5 x 1012 2.1 x i07 8.0 x 1012 1.8 x313 5.9 x 1O 1.9 x i07 5.3 x 1O 1.5 x i07 4.8 x 1012 1.3 x i07298 4.0 x 1012 1.2 x 3.5 x 10.12 1.1 iO7 3.0 x 1012 9.5 x 10Ea (keal.) 4.6 4.2 4.9 3.7 5.5 3.6Following the work of Boyd and Saldano [27], the activation energy (Ea) can be used to give ahint as to the nature of the operating mechanism.91-15-26.60.29-15.2-15.4-15.6O-15.8-16-16.2-16.40.29 0.34Figure 5.23 : Arrhenius plots for the aqueous phase mass transfer coefficients. Determination ofthe activation energy for diffusion in the aqueous phase : Slope=-EJR, for the resins under a givenpre-treatment.0.3 0.31 0.32 0.33(im-25.2-25.4C-25.6C-25.80-26C)C-26.2(0D-26.40)00.30 0.31 0.32 0.33(11)0.34Figure 5.24 : Arrhenius plots for the resin phase diffusion coefficients. Determination of theactivation energy for diffusion in resin phase Slope=-Ea/R, for the resins under a given pretreatment.9265210Time (minutes)Figure 5.25 : Variation of the activation energy for diffusion with time.-oa)wC0U-0.80.6OA0.20Time (minutes)600Figure 5.26: Comparison of the kinetics of zinc uptake by the resins. Batch test study of loadingon pre-treated resin; initial [Zn] = 12 mg/L, T=6OC. Pre-treatment : Equilibrating resin in zincfree INCO electrolyte at the particular pH.0 5 10 15 20 25 30 35100 200 300 400 50093Table 5.12 : Determination of the rate controlling mechanism from the Activation Energy.Activation Energy Rate Controlling Step> 12 kcal. Chemical reaction control5-9 kcal. Particle diffusion control0-4 kcal. Film diffusion controlFrom the values reported for the resins in Table 5.11, it seems that the rate controlling step is mostcertainly a combination of the particle and film diffusion control steps, with PDC prevailing in theinitial part of the process (first 15 minutes), while FDC is the dominant mechanism after 20 minutes.This may also be concluded from Figure 5.25, which plots the activation energy for diffusionobtained from D and /ç, values against time. It shows the activation energy values to be in theparticle diffusion regime in the first 15 minutes, while they fall in the film diffusion regime after20 minutes.5.4 Kinetic Modelling of the Column TestsThe complete breakthrough curves for zinc loading on OC 1026, Cyanex 302 and Cyanex 272,are presented in Figures 5.27, 5.28 and 5.29 respectively. The value of O (Section 4.5) was fixedat the point where breakthrough was first detected. This was at about 0.9 ppm for OC 1026,0.95ppm for Cyanex 302 and 0.6 ppm for Cyanex 272. Based on the method outlined in Section 4.5,a log-log plot is made of ln(1/(1-cfc0))vs. (0 - 0), as shown in Figure 5.30 for the resins. This iscompared to the analysis of the curve obtained by Strong and Henry [89], and the characteristicparameters were determined as follows:94Table 5.13 : Comparison of characteristic parameters of the breakthrough curve for theresins.Resin f3 0(hrs.) 9m(hrs.)OC 1026 1.75 87.5 65Cyanex 302 2.1 37.5 70Cyanex 272 2.2 25.0 68Strong et. al 0.9 24.0 40The breakthrough curve obtained by Strong and Henry (Fig. 5.31) [89] was determined by carryingout the test under conditions in Table 5.14:Table 5.14 : Composition of Electrolyte and Operating Conditions of Strong and Henry [89]test (Resin: OC 1026).Analysis (g/L) Operating conditionsCo Ni Zn Fe Cu T(°C) pH Flow(BV/h) Volume of resin40 0.25 0.04 0.004 0.01 40 5.0 10 200 mLUsing these characteristic parameters, the kinetic parameters useful in the design of ion exchangecolumns, outlined in equations (4.24)- (4.32), were calculated for the respective resins (Table5.15).Table 5.15 : Kinetic parameters in the design of ion exchange columns (Resin pre-treated atpH=3, Operating temperature = 60°C, H=27 cm., U=1.835 x 10 mis).Resin O(hrs) N h (cm) U (mis) Nf Hf (m) D1, H.OC 1026 152.5 2.34 11.50 4.91 x 10 2.45 26.50 0.574 11.50Cyanex 302 107.5 1.53 17.65 7.00 x io 2.45 23.78 0.35 17.55Cyanex272 93 1.36 19.85 8.10x107 2.45 23.11 0.27 19.719513120)Eli190i0 I I I0 200 400 600 800 1000Bed Volumes of EffluentFigure 5.27 Breakthrough curve for zinc loading on pre-treated OC 1026. INCO electrolyte feed(pH 3.75 at 60°C, [Zn]=12 mgfL), flow rate=4.67 BV/hr., OC 1026 bed = 25 ml, resin pre-treatmentINCO electrolyte fed at pH 3 until effluent pH 2.13120).10 -c 8-0Permissible concentration in effluent0.•• I I I I I0 100 200 300 400 500 600 700Bed Volumes of EffluentFigure 5.28 : Breakthrough curve for zinc loading on pre-treated Cyanex 302. INCO electrolytefeed (pH 3.75 at 60°C, [ZnJ=12 mg/L), flow rate=4.67 By/hr., Cyanex 302 bed = 25 ml, resinpre-treatment: INCO electrolyte fed at pH 3 until effluent pH 2.5.96(2)Bed Volumes of Effluent0.6 0.8 1 1.2 1.4 1.6 1.8log (time difference from breakthrough point)13120)E112b0‘9c80U00.1U)22Wi0700Figure 5.29 : Breakthrough curve for zinc loading on pre-treated Cyanex 272. INCO electrolytefeed (pH 3.75 at 60°C, [Zn]=12 mg/L), flow rate=4.67 By/hr., Cyanex 272 bed = 25 ml, resinpre-treatment: INCO electrolyte fed at pH 3 until effluent pH 2.0 100 200 300 400 500 600Breakthrough Curve Analysis0.50(0.5)C, (1.5)(2.5)2 2.2Figure 5.30 : Breakthrough curve analysis. Determination of characteristic parameters of thebreakthrough curve. Slope=J3; Atc/c0=0.632, mean breakthrough time= time difference.97The minimum superficial velocity of the solution in column processes, in order to be operating inthe particle diffusion control regime, as determined from equation (4.36) is in the order of i07rn/s. As the operating superficial velocity of 1.835 x io- rn/s is well above this value, attempt hasbeen made to operate in the particle diffusion control regime. Comparing the breakthrough curveobtained for a column with 25 ml OC 1026 (Fig. 5.27),(operated at 4.67 BV/hr., 60°C), with thatobtained by Strong and Henry (Fig. 5.31) [89] using 200 ml of untreated OC 1026 (operated at 10BV/hr., 40°C), it may be concluded that the curve does not flatten with increasing bed depth, andhence the exchange equilibrium is favorable [15].Using equation (4.22) the overall mass transfer coefficients for the respective resins in columnperformance were determined. These values are compared with those determined from the Car-berry’s correlation (eq. 4.33) [86], and the method in Perry’s Chemical Engineering Handbook[88] (equation 4.36), and also with those obtained from the analyses of the batch kinetic data, inTable 5.16.Table 5.16: Comparison of the mass transfer coefficients (mis) obtained by various methods.Resin kf (eq. 4.22) kL (eq. 4.33) km (eq. 4.36) km (batch tests)OC 1026 1.60 x 108 5.5 x i0 6.2 x iO 2.7 xCyanex 302 1.74 x 108 5.5 x i0 6.2 x 2.1 x i0Cyanex 272 1.87 x 108 5.5 x i07 6.2 x iO 1.8 x i0From the values of , it may be concluded that the rate controlling process for the three resins, ispossibly in the intermediate region between particle and film diffusion control. This propositionis substantiated by the comparison of the values of mass transfer coefficients; which shows the kvalues to be an order of magnitude smaller than the kL and km values. While, both kL (from equation(4.33)) and km (from equation (4.36)) are obtained by determining the diffusion coefficient froma calculation of the liquid film thickness 6, the value of kf (from equation (4.22)) is obtained froma more realistic estimate, based on the existing mass transfer conditions of the column. If the ratecontrolling mechanism were purely film diffusion control, the values of kf would compare well98with those of kL and km. while if the mechanism was purely particle diffusion control, the variationwould be greater [83]. An intermediate regime as is probably the case, with f and mass transfercoefficients in column processes, fairly close to particle diffusion control, while the activationenergy for diffusion in batch processes being at the intermediate particle-film diffusion regime,depending on the time range from start of the process, adequately defines the process kinetics.5.5 Mechanism of the ExchangeBased on the discussion in the preceding sections, a mechanism for the ion exchange separationof zinc from cobalt electrolytes is proposed in this section.For the case of the untreated resin, (Figure 5.32) the following processes occur during extraction:- Co2 and Zn2 ions diffuse inside the resin.- H ions diffuse out as a result of the exchange initiating a large initial pH reduction.At low metal loadings on OC 1026, the following reactions compete [68,70]:Co2 + 2(HL) E——* CoL2(HL) + 2H (5.3)Zn2 + 2(HL) —--* ZnL2(HL) + 2H (5.4)In the case of the pre-treated resin (Figure 5.33), the following steps occur during the pre-treatmentand loading cycles:(a) During pre-treatment, cobalt loads on the resin to displace the II ions from the resin:Co2 + 2(HL) <—---* CoL2(HL) + 2H (5.5)Zinc ions present in the solution during column tests, are prevented from loading on to theresins for the following reasons:-The electrolyte pH is initially reduced to about 3. At this pH Co2 loads more efficiently onthe resin than Zn2 (Section 5.1.4).-As the cobalt loading proceeds, the local pH of the electrolyte reduces to about 2, whichfurther discourages any zinc loading.99g/L Co/Zn60 50TIME (mm)TIME (HOURS)Figure 5.31: Loading and elution curves for Levextrel zinc removal from cobalt electrolytes. Feedelectrolyte :43 mg/L Zn, pH=5.O; elution with 5% sulphuric acid.Co2nd Zn 2+ 2+Co and ZnFigure 5.32: Cobalt and zinc loading on untreated resin from the INCO electrolyte.Figure 5.33 : Pre-treatment and loading processes during the ion exchange._mg/t ZINCIN RAFFINATE20--pHOFCOL1RAFFINATE5.016COLUMN 14.0ZINC3.02 LESStHAN 0.1 mg/I2.01.03220 80H HHH2*Co2*CoCo2*Zn2HH(a) Pre-treatment (b) Loading100-As cobalt concentration in the electrolyte is 40 g/L, while the zinc concentration is onlyabout 0.0 12 g/L, the Co2 ions far outnumber the Zn2 ions and thus reduce the competitionfrom the zinc ions.(b) During the loading cycle, Zn2 ions displace the Co2 ions on the resin by:Zn2 + CoL2(HL) —--* ZnL2(HL) + Co2 (5.6)This reaction is favoured during the loading cycle, probably due to the following reasons:-The equilibrium of reaction (5.6) is more favorable and the kinetics are faster than those ofreaction (5.4) under the operating conditions of the electrolyte.-The pH increase that occurs due to the pre-treatment (reaction (5.5)) attaining equilibriumaids reaction (5.6). This was seen in the column runs, where the effluent pH attained a valuevery close to the feed electrolyte pH within 50 effluent bed volumes, due to resin pretreatment (Section 5.1.3).-Reaction (5.6) results in no reduction in pH and hence, Zn2 loading, which occurs at pHvalues higher than 3.5, is favoured and proceeds continuously.-The diffusion of Zn2 ions into a resin matrix of CoL2(HL) is faster than the diffusion ofZn2 ions into the (HL)2matrix. This is supported by the observation that the D(diffusioncoefficient inside the resin matrix) values are enhanced by nearly an order of magnitudedue to pre-treatment (Table 5.10).Even in the case of the untreated resin, reactions (5.3), (5.4) and (5.6) occur. The function ofpre-treatment, however, is to break up these steps and allow cobalt loading (reaction (5.3)) duringthe pre-treatment cycle and directly begin the loading cycle with zinc displacement of the cobaltcomplex (reaction (5.6)). The methods used to enhance cobalt loading during pre-treatment andzinc extraction during the loading cycle have been discussed in the previous paragraph. In summary,the pre-treatment step enhances the overall zinc extraction and reduces the net cobalt retained onthe resin, due to the following reasons:(1)Zinc extraction during the loading cycle is greater because of reduced competition from cobalt101ions for the extractant.(2)The kinetics of exchange of Zn2 for Co2 ions by reaction (5.6) is faster than that of Zn2 forH ions by reaction (5.4). Pre-treatment initiates reaction (5.5) so that in the loading cycle,reaction (5.5) and (5.6) can occur concurrently to result in better extraction kinetics. The kineticsare improved, not directly due to the improved kinetics ofreaction (5.6), but because the diffusionof Zn2 ions into a resin matrix of CoL2(HL) is faster than the diffusion of Zn2 ions into the(HL)2matrix. This is reflected in the enhanced D values due to pre-treatment, and the activationenergy for diffusion lies in the region between particle and film diffusion control, rather thanchemical reaction control.(3)The pH reduction during the loading cycle for the pre-treated resin is smaller than in the caseof the untreated resin. From the discussion in Sections 5.1.3 and 5.1.4, zinc extraction is favoredat pH values between 3.4-3.7 while cobalt extraction is favored at lower values, and hencepre-treatment enhances loading.1026 CONCLUSIONS AND SUGGESTIONS FOR FURTHER WORK6.1 Summary of the resultsThe series of column and batch tests on the resins at various operating conditions has yieldedsome extremely useful results on the performance of the resins, the mechanism of zinc loadingand the suitability of their application to the INCO zinc removal process. The following is a criticalappraisal of the results.(1) The loading of zinc onto untreated OC 1026 is extremely efficient as breakthrough does nottake place until after 200 By. However, the cobalt retained on the resin may be in objectionablyhigh amounts.(2) The resins initially load cobalt at lower pH values, while the zinc loads by displacing cobaltas the pH of the column rises.(3) Pre-treatment of OC 1026 with INCO electrolyte at pH of 3, improves zinc loading and reducescobalt retention by maintaining the column at a pH between 3.5-3.7 for a major part of theprocess, where zinc loads by displacing the cobalt loaded on the resin in the first 50 By.(4) Reduction in temperature hastens zinc breakthrough for all of the resins, but reduces the cobaltretained on the resins. At 40°C, the zinc loading is satisfactory, while the cobalt retained isconsiderably reduced (when compared to the 60°C test).(5) Increasing the flow rate of the electrolyte through the OC 1026 column, results in a slightlyearlier zinc breakthrough, but there is an associated reduction in cobalt retained on the resin,thus improving the selectivity. This reduction in the cobalt retained can be attributed to a smallerpH reduction for the column operating at the faster flow rate.(6) Increasing the pH of the feed electrolyte enhances zinc loading on OC 1026, resulting in abreakthrough point similar to that of a column pre-treated at pH 3 with INCO electrolyte.However, the selectivity of the former is greater as the cobalt retained is reduced.(7) The performance of Cyanex 302 and Cyanex 272 is inferior to that of OC 1026 in terms of zincloading, under the operating conditions of the INCO plant. Yet Cyanex 302 presents the103advantage of lower cobalt retention with a satisfactory zinc loading (breakthrough after 175By). However, the resin is poisoned by copper and iron from the electrolyte which hastenszinc breakthrough in further loading cycles. Hence if copper removal was accomplished (byXFS-4 195 used for nickel separation) before the use of Cyanex 302 for zinc removal, then thisresin could be considered as an alternative because of lower amounts of cobalt retained.(8) A variety ofpre-treatments were tried out in batch tests. For OC 1026 pre-treatment with INCOelectrolyte at pH values of 2, 3 and 3.5 enhance kinetics to an equal level when compared tothe untreated resin. For Cyanex 302 however, pre-treatment at pH of 2, enhances the kineticsa little more than other pre-treatments. In fact it was found that the kinetics of zinc uptake byCyanex 302 pre-treated with INCO electrolyte at pH 2, were slightly better than that of OC1026 pre-treated at pH of 2 or 3. However in column performance, the Cyanex 302 columnnever reached a pH below 2.7, and hence probably OC 1026 performs better.(9) Increasing the initial zinc content of the INCO electrolyte makes almost no difference to thezinc loading on OC 1026 in the first 15 minutes, but enhances the kinetics considerably at latertime intervals.(l0)Decrease in temperature slackens the kinetics of zinc uptake for all the resins.(11)The kinetic data could not be fitted to any single model. The data however, obey the FickianParticle Diffusion Control (PDC) Model for the first 15 minutes, after which they obey theFilm Diffusion Control (FDC) Model until 150 minutes. After this, the data lie between modelsof Particle Diffusion Control (with a D value of an order of magnitude higher than the initialPDC D) and the Unreacted Core Model (where diffusion through the reacted resin layer israte controlling). The values of D and km were determined, based on these models for theexchange under different conditions.(12)Pre-treatment enhances the value of D by about an order of magnitude, in the initial PDCregion, and this enhances km to some extent in the FDC region.(13)The kinetic performance of the three resins is comparable and the D and km values are withinthe same order of magnitude.104(14)The Arrhenius plot (log vs. T’ and log km vs. T’ ) gives values of the activation energy fordiffusion to be in the range between particle and film diffusion control regimes, for all theresins.(1 5)The column breakthrough curves were used to determine the characteristic parameters 3 (meanslope of the curve), O (breakthrough time) and Om (mean time of the breakthrough function).These were used to determine a range ofkinetic parameters useful in the design of ion exchangeequipment for the three resins, such as saturation time, height of the ion exchange zone, velocityof the ion exchange zone, degree of bed saturation at breakthrough and height of unused bedat breakthrough.(16)The characteristic parameters were used to determine the value of the mass transfer coefficientin the column (kf), which was compared to the values obtained by other methods suggested inthe literature. Based on this comparison and the values, it was concluded that the processkinetics in the column were controlled by a mixture of particle and film diffusion mechanismsfor the three resins.6.2 ConclusionsThe major conclusions of this work may be summarized as follows:(1)The performance of OC 1026 was superior to that of Cyanex 272 and Cyanex 302, under theoperating conditions of the INCO zinc removal process plant at Port Colborne.(2)Column operating conditions such as resin pre-treatment, electrolyte pre-treatment, temperatureand flow rate have a significant influence on the zinc loading and the selectivity characteristicsof the resins. For OC 1026, it was found that a resin pre-treatment at a pH of 3, or an electrolytepH increase to 5, along with operation at a temperature of 40°C and a flow rate of 10 BV/hr.,provided the maximum improvement to zinc loading and selectivity.(3)The kinetics ofzinc loading on pre-treatedOC 1026, Cyanex 302 and Cyanex 272 are comparableat similar operating conditions.(4)In batch tests, the loading process operated in the area between particle and film diffusion105control, with PDC being rate controlling in the first 15 minutes, while FDC becomes importantlater. The column process is also a mixed particle-film diffusion controlled process, with thecharacteristic parameters being a little closer to particle diffusion control.6.3 Suggestions for further workThis work leaves a number of unexplained results and further work in the following directionscould shed light on the process mechanism and help in process optimization:(1)Studies need to be made on various other possible pre-treatments for the resins, to improve theirzinc loading and selectivity characteristics. For Cyanex 302 studies need to be made for specificpre-treatments or process changes in order to avoid the resin poisoning by copper.(2)The results of the column and batch studies from this work, need to be substantiated bySEM-WDX studies (X-ray microprobe analyses) of the resin cross-section, to gain an insightinto the process mechanism, and understand why cobalt loads on the resin initially, while zincloads by displacing the cobalt from the complex. This analysis may also aid in understandinghow resin pre-treatment exactly influences the process, and why it enhances the resin phasediffusivity by an order of magnitude. Similar studies can be done in column processes, on resinsamples taken from the top, middle and bottom portion of the bed, to explain the effect thatlocal pH has on zinc loading and selectivity.(3)The accuracy of batch studies should be greatly enhanced by using zinc selective electrodesand monitoring the Eh values, rather than the technique of determining zinc concentration atvarious time intervals, as done in this project. Work is thus needed in this direction, to accuratelydetermine the values of diffusion and mass transfer coefficients.106REFERENCES1. Exodus 15: 23-25.2. Aristotle, Works, (Clarendon Press, London, 1927), Vol. 7, p.933b.3. Thompson, H.S.; J. Roy. Agr. Soc. Engi., Vol. 11(1850), p.68.4. Way, J.T.; J. Roy.Agr. Soc. Engi., Vol. 11(1850), p.313; Vol. 13(1852), p.123.5. Harm, F. and A. Rumpler; 5th Intern. Congr. Pure Appi. Chem. (1903), p.59.6. Adams, B.A. and E.L.Holmes; .1. Soc. Chem. md. (London), Vol.54, (1935), p.16.7. Tompkins, E.R.; J.X. Khym, and W.E. Cohn; J. Am. Chem. Soc., Vol. 69 (1947),p2769.8. Ketelle, B.H. and G.E. Boyd, J. Am. Chem. Soc., Vol. 69 (1947),p.2800.9. Spedding, F.H.; E.I. Fulmer, T.A. Butler, E.M. Gladrow, M. Gobush, P.E. Porter, J.E. Powell, andJ.M. Wright, J. Am. Chem. Soc., Vol. 69 (1947), p.2812.10.D’Alelio, G.F. (General Electric Co.); US Patent 2,366,007 (1944).11.Wheaton, R.M. and W.C. Bauman; I&EC, Vol. 43 (1951), p1088.12.Breck, D.W.; “Zeolite Molecular Sieves : Structure, Chemistry, and Use” (Wiley, New York,1974), p.95.13.Christensen, J.J. and R.M. Izzat, “Handbook of Metal-Ligand Heats and Related ThermodynamicQuantities”, 3rd. Ed. (New York, Wiley & Sons 1980).14.Warshawsky, A.; Modem Research in Ion Exchange, in “Ion Exchange Science and Technology”,Edited by Rodrigues, A.; NATO ASI Series, Martinus Nijhoff Publ. 1986.1 5.Helfferich, F.; “Ion Exchange”, McGraw-Hill, N.Y. (1962).16.Vermeulen, T.; G. Klein and N.K. Hiester “Adsorption and Ion Exchange”, in Perry’s Chem.Eng. Handbook, 5-th Edition, Sec. 16, Mc-Graw Hill Book Co., NY (1973).17.Paterson, R.; “An introduction to ion exchange”, Heyden & Sons Ltd., London (1970).18.Marinsky, J.A.; Ed. of “Ion Exchange, Vol. I”, Marcel Dekker Inc., NY (1966).19.Hudson, M.J.; Co-ordination chemistry of selective ion exchange resins, in “Ion ExchangeScience and Technology”, Edited by : Aliro Rodrigues, NATO ASI Series 1985.20.Liberti Lorenzo; “Planning and interpreting kinetic investigations”, in “Ion Exchange: Science107and Technology”, Edited by : Aliro Rodrigues, NATO ASI Series 1985.21.Boyd, G.E.; A.W. Adamson and L.S. Myers Jr.; J. Am. Chem. Soc., Vol. 69 (1947),p.2836.22.Wen, C.Y.; md. Eng. Chem., Vol. 60 (1968), p.34.23.Kunii, P. and 0. Levenspeil; “Fluidization Enginnering”, Wiley, New York, NY, 1969.24.Yagi, S. and P. Kunii; Chem. Eng. (Tokyo), 19 (1955), 500.25.Wheeler, A.; “Advances in Catalysis”, 3, 249, (1951).26.Mackie, J.S. and P.Meares, Proc. Roy. Soc. (London), A232 (1955) 498.27.Boyd, G.E. and B.A. Saldano, J. Am. Chem. Soc., 75 (1953) 6091-9.28.Sillen, L.G. and A.E. Martell; “Stability of metal-ion complexes”, The Chemical Society, Burlington House, London, Spec. Publ. no. 17 and no.25.29.Gluckauf, E. and J.I. Coates; J. Chem. Soc., (1947) 1315-21.30.Vermeulen, T.; md. Eng. Chem., 45, (1953) 1664-70.31.Crank, J.; “The Mathematics of Diffusion”, (Clarendon Press, Oxford, 1975).32.Thomas, H.C.; Heterogenous Ion Exchange in a Flowing System, J. Am. Chem. Soc., 66 (1944)1664.33.Thomas, H.C.; The Kinetics ofFixed-Bed Ion Exchange, in “ION EXCHANGE”, ed. F.C. Nachod,p.29 (Academic Press, NY 1964).34.Gilliland, E.R. and R.F. Baddour; The Rate of Ion Exchange, md. Eng. Chem. 45 (1953) 330.35.Helfferich, F.; J. Phys. Chem., 69 (1965) 1178-87.36.Dana, P.R. and T.D. Wheelock; md. Eng. Chem. Fundam., 13 (1974) 20-6.37.Holl, W. and H. Sontheimer; Chem. Eng. Sci., 32, (1977) 755-62.38.Helfferich, F.; Ion exchange kinetics - evolution of a theory, in “Mass Transfer and Kinetics ofIon Exchange”, NATO ASI Series, Martinus Nijhoff Pub. (1982).39.Helfferich, F. and M.S. Plesset; J. Chem. Phys. 28 (1958) 418.40.Barrer, R.M.; R.F.Bartholomew and L.C.V. Rees; J. Phys. Chem. Solids, 24 (1963) 309.41.Brooke, N.M., and L.C.V. Rees; Trans. Faraday Soc. 12(1968)3383.42.Schlogl. R.; Stoffiransport durch Membranen (Steinkopff, Dermstadt, 1964).10843.Baypai, R.K.; A.K. Gupta and M. Gopala Rao; AIChE J. 20 (1974) 989.44.Helfferich, F.; J. Phys. Chem. 69 (1965) 1178.45.Shewnon, P.G.; “Diffusion in Solids” (McGraw Hill, New York, 1963).46.De Groot, S.R. and P. Mazur; Thermodynamics of Irreversible Processes, (North Holland,Amaterdam, 1962).47.Lightfoot, E.N. and E.M. Scattergood; AIChE J., 11(1965)175.48.Graham, E.E., andJ.S. Dranoff; Applications of the Stafan-Maxwell equations to diffusion in ionexchangers; IEC Fundamentals (1982).49.Kressman, T.R.E., and J.A. Kitchener; Discussions Faraday Soc., 7 (1949), 90.50.Tetenbaum, M., and H.P. Gregor; J. Phys. Chem., 58 (1954), 1156-63.51.Hale, D.K., and D. Reichenberg; Disc. Faraday Soc., 7 (1949) 79-85.52.Reichenberg, D.; J. Am. Chem. Soc., 75 (1953), 589-97.53.Semenovskaya, T.D.; V.T. Avgul, A.Ya. Pronin and K.V. Chmutav, Zhur. Fiz. Khim., 53 (1979)1850-2.54.Smith, M.; “Chemical Engineering Kinetics”, 2-nd Edition, Mc-Graw Hill, New York, N.Y.(1970).55.Wilke, 3. and O.A. Hougen; Trans. Am. Inst. Chem. Eng., 41(1945) 445.56.Bahu, R.E.; M.J. Craske and M. Streat, Proc. mt. Conf on Theory and Practice ofIon Exchange,SCI Cambridge, (1976).57.Boghetich, G., and D. Petruzelli; ICP, XI, 10 (1981).58.Hoell, W., and G. Geiselhart; Desalination, 25 (1978) 217.59.Kauczor, H.W., and A. Mayer; Hydrometallurgy, 3 (1978), 65.60.Warshawsky, A.; Extraction with solvent impreganated resins, in “Ion Exchange and SolventExtraction”, Eds.: Marinsky J.A., andY. Marcus, Vol. 8, Marcel Dekker, 1981.61.Agnew, R.; P. Garritsen and D. MacVicar; Production of electrolytic cobalt rounds at INCO’scobalt refinery in Port Colborne, Ontario, Canada, in “Extractive Metallurgy ofNickel and Cobalt”,Edited by G.P. Tyroler and C.A. Landolt, The Metallurgical Society Inc. Publication, 1988.10962.Agnew, R.G.; L.D. Kowal and D. Young; Process optimisation at INCO’s cobalt refinery, in“Electrometallurgical Plant Practice”, Pergamon Press 1990.63.Ettel, V.A.; “Process for recovering cobalt electrolytically”, US Patent no. 4274930. 1980.64.Renzoni, L.S. and Barker, W.W.; Production of electrolytic cobalt rounds at INCO’s CobaltRefinery in Port Colborne, in “Extractive Metallurgy of Copper, Nickel and Cobalt”, Queneau,P. (Ed.), AIME p.537 (1961).65.Kerfoot, D.G.E., and D.R. Wier; The hydro and electrometallurgy of nickel and cobalt, in “Extractive Metallurgy of Nickel and Cobalt”, Edited by G.P. Tyroler and C.A. Landolt, The Metallurgical Society Inc. Publication, 1988.66.”Developments in Solvent Extraction”, S. Algret ed., Ellis Horwood Ltd., Chichester, England(1988).67.Ajawin, L.A.; E.S. Perez de Ortiz and H. Sawistowski, Chem. Eng. Res. Des., 61(1983) 62.68.Cianetti, C., and P.R. Danesi; Solv. Ext. Ion Ex., 1, (1983) 9.69.Sastre, A.M., and M. Muhammed; Hydrometallurgy, 12, (1983), 177.70.Komasawa, I.; T. Otake andY. Hagaki, J. Inorg. Nuci. Chem., 43(12), (1981) 3351.71.Ritcey, G.; A. Ashbrook and L. Lucas; Development of solvent extraction process for the separation of cobalt from nickel, CIM Bull., 68(753)111, 1975.72.Nogueira, E.D.; J.M. Regife and P.M. Blythe; Zincex- the development of a secondary zincprocess, Chem. Ind.,(2) (1980),p.63-7.73.Ajawin. L.A.; E.S. Perez de Ortiz and H. Sawistowski, Kinetics of extraction of zinc by di(2-ethyihexyl) phosphoric acid in n-heptane, Proceedings ofISEC’80, Paper 80-112, (1980).74.Rice, N.M., and M.R. Smith; J. Appi. Chem. Biotechnol., 25 (1975)p.379-402.75.Flett, D.S.; J. Chem. Biotechnol., 29 (1979)p.258-72.76.Rickleton, W.A.; D.S. Flett, and D.W. West, Hydrometallurgy, 9, (1982) 115.77.Sastre, A.M.; N. Miralles and E. Figuerola, Solvent Extr. Ion Exch., 8 (1990) p.597-614.78.Sole, K.C. and J.B. Hiskey; Hydrometallurgy, 30 (1992)p.345-6.79.Caravaca, C., and F.J. Alguocil; Hydrometallurgy, 27 (1991), p.327-338.11080.Levenspeil, 0.; “Chemical Reaction Engineering”, John Wiley & Sons, Inc., NY (1972).81.Ishida, M., and C.Y. Wen; Chem. Eng. Sd., 26, 103 1-41 (1971).82.Smith, J.M.; “Chemical Engineering Kinetics”, Mc-Graw Hill Book Company, NY (1970).83.Doulah, M. S.; and S .A. Jafer; A method for the prediction of mass transfer parameters in an ionexchange process, in “Ion Exchange for Industry”, editor: Streat, M., Ellis Horwood Pub., (1988).84.Micheals, A.S.; md. Eng. Chem., 44(8), 1922-30 (1952).85.Rodrigues, A.; and C. Costa; Fixed bed processes : a strategy for modelling, in “Ion Exchange:Science and Technology”, Edited by: Aliro E. Rodrigues, NATO ASI Series, Martinos NijhoffPub. (1986).86.Carberry, J.J.; AJ.Ch.EJ., 6 (3), 460-63 (1960).87.Glueckauf, E.; in “Ion exchange and its applications”, edited by the Society of Chemical Industry,p. 34 (1955).88.Perry, R.; and D. Green; “Chemical Engineering Handbook”, 50-th edition, Mc-Graw Hill Pub.(1991).89.Strong, B.; and R.P. Henry; Hydrometallurgy, 1,311-317(1976).90.Cussler, E.L.; “Diffusion, Mass transfer in fluid systems”, Cambridge Univ. Press, Cambridge,1984.91.Babjak, G.; “A study of ion exchange kinetics involving chelating resin” Ph.D. thesis, Univ. ofWaterloo (1982).92.Kauczor, H.W.; and Meyer, A.; Hydrometallurgy 3 (1978) 65-73.93.Simon, G.P.; “Ion Exchange Training Manual”, Van Nostrand Reinhold, New York (1991).94.Lewatit OC 1026 brochure, Bayer AG. Leverkusen; Order no. OCTh 20208.111Appendix A : Raw Experimental Data from Column TestsTable A-i : Zinc loading on untreated OC i026Resin: 25 ml OC 1026Pre-treatment: NoneOperating Temperature: 40°CFeed Flow rate: 5 By/hr. and 10 BV/hr.Feed Electrolyte Composition: Co=40 g/L, Ni=35 g/L, Zn=0.012 g/L, Fe=0.001g/L, Cu=0.01 g/L.Feed Electrolyte pH: 3.7Flowrate: 5BV/hr. 1OBV/hr.Effluent [Zn] pH [Zn] pHVolume(BV) (mgfL) (mg/L)0 0.00 3.70 0.00 3.702 0.00 2.10 0.00 2.255 0.00 2.20 0.00 2.4010 0.00 2.25 0.00 2.4515 0.00 2.30 0.00 2.5020 0.00 2.35 0.00 2.5525 0.00 2.40 0.00 2.6030 0.00 2.45 0.00 2.6540 0.00 2.52 0.00 2.7250 0.00 2.55 0.00 2.7560 0.00 2.57 0.00 2.7770 0.00 2.60 0.00 2.8080 0.00 2.62 0.05 2.8190 0.10 2.63 0.12 2.82100 0.15 2.64 0.18 2.83120 0.25 2.65 0.27 2.84140 0.40 2.66 0.45 2.85150 0.45 2.67 0.50 2.86160 0.50 2.68 0.60 2.87170 0.60 2.69 0.70 2.88180 0.80 2.70 0.90 2.89190 0.85 2.71 1.10 2.90200 0.92 2.72 1.35 2.91210 1.10 2.73 1.65 2.92220 1.50 2.74 1.90 2.93230 1.75 2.75 2.25 2.93240 1.90 2.76 2.50 2.94250 2.00 2.77 2.70 2.94112Table A-2 : Effect of temperature on zinc loading of untreated OC 1026Resin: 25 ml OC 1026Pre-treatment: NoneOperating Temperature: 60°CFeed Flow rate: 4.67 By/hr.Feed Electrolyte Composition: Co=40 g/L, Ni=35 g/L, Zn=0.012 g/L, Fe=O.001g/L, Cu=0.01 g/L.Feed Electrolyte pH: 3.75Effluent [Zn] pHVolume(BV) (mg/L)0.000 0.00 3.751.000 0.00 1.902.335 0.00 2.004.670 0.00 2.1011.675 0.00 2.2023.350 0.00 2.3546.700 0.00 2.5070.050 0.00 2.6093.400 0.00 2.63116.750 0.00 2.64140.100 0.10 2.65163 .450 0.25 2.67233.500 0.50 2.75280.200 0.76 2.80326.900 1.06 2.85373.600 1.40 2.88420.300 1.90 2.91467.000 2.50 2.93500.000 2.80 2.94113Table A-3 : Zinc loading on pre-treated OC 1026Resin: 25 ml OC 1026.Pre-treatment: INCO electrolyte fed at pH=3 until effluentOperating Temperature: pH=2.Feed Flow rate: 60° C.Feed Electrolyte Composition: 4.67 BV/hr.Co=40 g/L, Ni=35 g/L, Zn=0.012 g/L,Feed Electrolyte pH: Fe=0.001 g/L, Cu=0.01 g/L.3.75.Effluent [Zn] pHVolume(BV) (mg/L)0.000 0.00 3.702.335 0.00 2.3511.675 0.00 2.9523.350 0.00 3.2046.670 0.00 3.5070.050 0.00 3.5293.400 0.00 3.54116.750 0.00 3.55140.100 0.00 3.55163.450 0.00 3.55186.800 0.00 3.55210.150 0.00 3.55221.825 0.00 3.55233.500 0.10 3.55245.175 0.18 3.55256.850 0.30 3.56268.525 035 3.56280.200 0.40 3.56291.875 0.45 3.56303.550 0.50 3.56315.225 0.55 3.56326.900 0.60 3.57338.575 0.70 3.57350.250 0.75 3.57361.925 0.80 3.57373.600 0.85 3.57385.275 0.90 3.57396.950 0.95 3.58408.625 1.07 3.58420.300 1.20 3.58431.975 1.27 3.58443.650 1.35 3.59455.325 1.42 3.59467.000 1.50 3.59114Table A-4 : Effect of temperature on zinc loading on pre-treated OC 1026Resin: 25mlOC 1026.Pre-treatment: INCO electrolyte fed at p11=3 until effluentOperating Temperature: p11=2.Feed Flow rate: 40°C, 25°C.Feed Electrolyte Composition: 4.67 BV/hr.Co=40 g/L, Ni=35 g/L, Zn=0.012 g/L,Feed Electrolyte pH: Fe=0.001 g/L, Cu=O.01 g/L.3.7.Temperature 40°C 25°CEffluent [Zn] pH [Zn] pHVolume(BV) (mgfL) (mgIL)0.000 0.00 3.70 0.00 3.342.335 0.00 2.40 0.00 2.254.670 0.00 2.80 0.00 2.8023.350 0.00 3.30 0.00 3.2046.700 0.00 3.50 0.00 3.2170.050 0.00 3.53 0.25 3.2293.400 0.00 3.55 0.60 3.23116.750 0.20 3.55 0.88 3.24140.100 0.40 3.55 1.07 3.24163.450 0.55 3.55 1.46 3.25186.800 0.70 3.55 1.76 3.25210.150 0.80 3.55 1.91 3.25233.500 0.95 3.55 1.95 3.25256.850 1.15 3.55 2.34 3.25280.200 1.50 3.55 2.53 3.25326.900 2.30 3.55 2.80 3.25373.600 3.20 3.55 3.40 3.25420.300 3.95 3.55 4.20 3.25467.000 4.50 3.55 4.75 3.25500.000 4.75 3.55 5.10 3.25115Table A-5 :Zinc loading on untreated OC 1026 with increased feed electrolytepHResin:. 25 ml OC 1026Pre-treatment: None.Operating Temperature: 40°C.Feed Flow rate: 10 BV/hr.Feed Electrolyte Composition: Co=40 g/L, Ni=35 g/L, Zn=0.012 g/L, Fe=0.001g/L, Cu=0.01 g/L.Feed Electrolyte pH: 5.0.Effluent [Zn] pHVolume(BV) (mg/L)0 0.00 5.005 0.00 2.2010 0.00 3.0050 0.00 3.25100 0.00 3.60150 0.00 4.00200 0.00 4.40210 0.21 4.41220 0.42 4.42230 0.65 4.43240 0.81 4.44250 1.05 4.45260 1.30 4.46270 1.70 4.47280 2.10 4.48290 2.60 4.49300 3.00 4.50400 3.75 4.53500 4.50 4.55116Table A-6: Zinc loading on pre-treated Cyanex 272Resin: 25 ml Cyanex 272Pre-treatment: INCO electrolyte fed at pH=3 until effluentpH=2.5Operating Temperature: 60°CFeed Flow rate: 4.67 By/hr.Feed Electrolyte Composition: Co=40 g/L, Ni=35 g/L, Zn=0.012 g/L, Fe=0.001g/L, Cu=0.01 g/L.Feed Electrolyte pH: 3.75Effluent [Zn] pHVolume(BV) (mgfL)0.000 0.00 3.702.335 0.00 2.804.670 0.00 2.907.005 0.00 3.029.340 0.00 3.0511.675 0.00 3.1023.350 0.00 3.3946.700 0.00 3.4970.050 0.00 3.5081.725 0.10 3.5093.400 0.24 3.51105.075 0.35 3.52116.750 0.50 3.53128.425 0.60 3.54140.100 0.70 3.55147.105 0.80 3.55151 .775 0.87 3.55158.780 0.95 3.55163.450 1.12 3.56168.120 1.30 3.56175.125 1.50 3.56182.130 1.75 3.57186.800 2.00 3.57191 .470 2.30 3.57198.475 2.60 3.57205.480 2.90 3.58210. 150 3.25 3.58214.820 3.50 3.58221 .825 3.75 3.59228.830 4.06 3.59233.500 4.28 3.60300.000 5.02 3.60350.000 5.80 3.61500.000 7.50 3.61117Table A-7: Zinc loading on pre-treated Cyanex 302.Resin: 25 ml Cyanex 302Pre-treatment: INCO electrolyte fed at pH=3 until effluentOperating Temperature: pH=2.5Feed Flow rate: 60°CFeed Electrolyte Composition: 4.67 By/hr.Co=40 g/L, Ni=35 g/L, Zn=0.012 g/L, Fe=0.001Feed Electrolyte pH: g/L, Cu=0.01 g/L.3.75Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5Effluent [Zn],mgIL pH [Zn],mg/L [Zn],mg/L [Zn],mg/L [Zn],mg/LVolume(BV)0.000 0.00 3.70 0.00 0.00 0.00 0.002.335 0.00 2.80 0.00 0.00 0.00 0.004.670 0.00 2.90 0.00 0.00 0.00 0.007.005 0.00 3.02 0.00 0.00 0.00 0.009.340 0.00 3.05 0.00 0.00 0.00 0.0011.675 0.00 3.10 0.00 0.00 0.00 0.0023.350 0.00 3.39 0.00 0.00 0.00 0.2046.700 0.00 3.49 0.00 0.00 0.15 0.3070.050 0.00 3.50 0.00 0.00 0.30 0.4081.725 0.00 3.50 0.25 0.20 0.40 0.6093.400 0.15 3.51 0.35 0.40 0.50 0.80105.075 0.30 3.52 0.50 0.55 0.60 0.90116.750 0.45 3.53 0.55 0.58 0.70 1.10128.425 0.60 3.54 0.62 0.65 0.85 1.15140.100 0.70 3.55 0.70 0.75 0.95 1.20147.105 0.80 3.55 0.72 0.85 1.10 1.30151.775 0.90 3.55 0.75 0.95 1.25 1.40158.780 0.93 3.55 0.85 1.05 1.37 1.50163.450 0.96 3.55 0.95 1.25 1.50 1.60168.120 0.98 3.55 1.28 1.40 1.60 1.70175.125 1.01 3.56 1.40 1.50 1.70 1.90182.130 1.03 3.56 1.50 1.60 1.85 2.10186.800 1.05 3.56 1.60 1.70 2.10 2.25191.470 1.08 3.56 1.70 1.85 2.25 2.50198.475 1.14 3.56 1.80 2.10 2.50 2.80205.480 1.20 3.56 2.10 2.25 2.80 3.10210.150 1.28 3.56 2.19 2.50 3.10 3.40214.820 1.35 3.56 2.50 2.75 3.40 3.70221.825 1.43 3.56 2.75 3.10 3.70 4.10228.830 1.50 3.56 3.00 3.40 4.10 4.50233.500 1.60 3.56 3.30 3.80 4.50 5.00118Appendix B : Raw Experimental Data from Batch TestsTable B-i : Effect of pre-treatment on zinc loading of OC i026Resin: 1.000 g OC 1026.Pre-treatment: Resin equilibrated in Zn-free electrolyte at a particular pH value.Operating Temperature: 60°C.Feed Electrolyte Composition: Co=40 g/L, Ni=35 g/L, Zn=0.012 g/L, Fe=0.001g/L, Cu=0.01 g/L.Feed Electrolyte Volume: 500 ml.Fraction Extracted: XPre-treatment Untreated atpH=3 atpH=2 atpfl=3.5 atpH=4Time(mintes) X X X X X0.000 0.000 0.000 0.000 0.000 0.0000.500 0.012 0.070 0.080 0.080 0.0501.000 0.090 0.160 0.180 0.170 0.1502.000 0.130 0.250 0.260 0.250 0.2003.000 0.150 0.290 0.300 0.300 0.2504.000 0.167 0.320 0.330 0.320 0.2905.000 0.175 0.350 0.365 0.350 0.3207.000 0.200 0.370 0.380 0.375 0.3508.330 0.230 0.390 0.400 0.395 0.37013.330 0.300 0.470 0.480 0.450 0.39016.670 0.360 0.520 0.540 0.530 0.43020.000 0.420 0.570 0.580 0.575 0.48025.000 0.470 0.600 0.610 0.600 0.53030.000 0.520 0.635 0.640 0.640 0.58041.670 0.560 0.680 0.680 0.680 0.63050.000 0.590 0.700 0.700 0.700 0.68066.670 0.625 0.742 0.750 0.745 0.72083.330 0.650 0.760 0.770 0.765 0.740116.670 0.710 0.770 0.780 0.775 0.760166.670 0.730 0.790 0.795 0.795 0.780250.000 0.750 0.820 0.820 0.820 0.800333.330 0.780 0.840 0.840 0.840 0.820416.670 0.800 0.860 0.860 0.860 0.830500.000 0.820 0.870 0.870 0.870 0.840119Table B-2 : Effect of temperature on zinc loading of OC 1026Resin: 1.000 g OC 1026.Pre-treatment: Resin equilibrated in Zn-free electrolyte atpH=3 .0.Operating Temperature: 60°C,40°C or 25°C.Feed Electrolyte Composition: Co=40 g/L, Ni=35 g/L, Zn=0.012 g/L, Fe=0.001g/L, Cu=0.01 g/L.Feed Electrolyte Volume: 500 ml.Electrolyte pH: 3.00.Fraction Extracted: XOperating temperature 60CC 40CC 25CTime (minutes) X X X0.000 0.000 0.000 0.0000.500 0.070 0.040 0.0251.000 0.160 0.120 0.1002.000 0.250 0.220 0.1953.000 0.290 0.260 0.2304.000 0.320 0.280 0.2505.000 0.350 0.300 0.2707.000 0.370 0.340 0.2908.330 0.390 0.350 0.3 1013.330 0.470 0.420 0.37016.670 0.520 0.490 0.42020.000 0.570 0.520 0.46025.000 0.600 0.550 0.50030.000 0.635 0.584 0.52841.670 0.680 0.631 0.57050.000 0.700 0.660 0.60066.670 0.742 0.690 0.64083.330 0.760 0.725 0.675116.670 0.770 0.750 0.710166.670 0.790 0.765 0.725250.000 0.820 0.785 0.750333.330 0.840 0.810 0.780416.670 0.860 0.825 0.800500.000 0.870 0.840 0.820120Table B-3 : Effect of initial [Zn] on loading of pre-treated OC 1026Resin: 1.000 g OC 1026.Pre-treatment: Resin equilibrated in Zn-free electrolyte atpH=3.0.Operating Temperature: 60CC.Feed Electrolyte Composition: Co=40 g/L, Ni=35 g/L, Zn=O.040 g/L, Fe=O.0O1g/L, Cu=0.01 g/L.Feed Electrolyte Volume: 500 nil.Electrolyte pH: 3.00.Fraction Extracted: X.Time(miriutes) mg extracted0.00 0.000.50 0.841.00 1.942.00 3.003.00 3.504.00 3.855.00 4.207.00 4.408.33 4.7013.33 5.3016.67 6.3020.00 6.8525.00 7.3030.00 7.9041.67 9.0050.00 10.0066.67 11.0083.33 12.00116.67 13.40166.67 15.00250.00 16.00333.33 17.00416.67 18.00500.00 19.00121Table B-4 : Effect of pre-treatment and temperature on loading of Cyanex302Resin: 1.000 g Cyanex 302.Pre-treatment: Resin equilibrated at a particular pH.Operating Temperature: 60°C, 40°C or 25°C.Feed Electrolyte Composition: Co=40 g/L, Ni=35 g/L, Zn=0.012 gIL, Fe=0.001g/L, Cu=O.01 g/L.Feed Electrolyte Volume: 500 mlFraction Extracted: XTemperature 60’C 60°C 60CC 40CC 25’CPre-treatment Untreated at pH=3 at p11=2 at pH=3 at p11=3Time X X X X X0.000 0.000 0.000 0.000 0.000 0.0000.500 0.010 0.070 0.080 0.040 0.0251.000 0.080 0.150 0.170 0.120 0.1002.000 0.120 0.250 0.270 0.220 0.1903.000 0.150 0.300 0.320 0.260 0.2304.000 0.160 0.330 0.360 0.280 0.2505.000 0.170 0.350 0.390 0.300 0.2707.000 0.185 0.370 0.410 0.330 0.2908.330 0.200 0.390 0.440 0.350 0.31013.330 0.250 0.440 0.500 0.400 0.37016.670 0.340 0.490 0.550 0.450 0.41020.000 0.410 0.550 0.580 0.490 0.44025.000 0.460 0.590 0.610 0.520 0.47030.000 0.500 0.614 0.640 0.550 0.50041.670 0.550 0.660 0.680 0.600 0.55050.000 0.580 0.690 0.700 0.625 0.58066.670 0.620 0.730 0.730 0.660 0.63083.330 0.660 0.750 0.760 0.695 0.670116.670 0.710 0.770 0.780 0.730 0.700166.670 0.730 0.790 0.800 0.750 0.720250.000 0.760 0.830 0.840 0.780 0.750333.330 0.780 0.850 0.870 0.800 0.780416.670 0.800 0.860 0.880 0.820 0.800500.000 0.810 0.870 0.890 0.840 0.810122Table B-5 : Effect of temperature on loading of Cyanex 272Resin: 1.000 g Cyanex 272.Pre-treatment: Resin equilibrated at pH=3.Operating Temperature: 60°C, 40°C or 25°C.Feed Electrolyte Composition: Co=40 g/L, Ni=35 g/L, Zn=0.012 g/L, Fe=0.001g/L, Cu=0.01 g/L.Feed Electrolyte Volume: 500 ml.Electrolyte pH: 3.0.Fraction Extracted: X.Temperature 60CC 40CC 25CCTime X X X0.000 0.000 0.000 0.0000.500 0.070 0.040 0.0251.000 0.150 0.120 0.1002.000 0.250 0.220 0.1903.000 0.290 0.250 0.2204.000 0.320 0.270 0.2405.000 0.350 0.290 0.2607.000 0370 0.330 0.2808.330 0390 0.350 0.30013.330 0.440 0.390 0.35016.670 0.480 0.430 0.39020.000 0.510 0.460 0.42525.000 0.550 0.500 0.45030.000 0.570 0.530 0.48041.670 0.620 0.570 0.53050.000 0.650 0.600 0.56066.670 0.690 0.640 0.60083.330 0.720 0.675 0.630116.670 0.760 0.720 0.680166.670 0.790 0.750 0.710250.000 0.820 0.780 0.750333.330 0.850 0.800 0.780416.670 0.860 0.820 0.800500.000 0.870 0.830 0.8 10123


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items