Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Aspects of phase separation in an experimental mixer-settler using two solvent extraction systems Eckert, Norbert L. 1987

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

Item Metadata

Download

Media
831-UBC_1987_A7 E24.pdf [ 10.21MB ]
Metadata
JSON: 831-1.0058830.json
JSON-LD: 831-1.0058830-ld.json
RDF/XML (Pretty): 831-1.0058830-rdf.xml
RDF/JSON: 831-1.0058830-rdf.json
Turtle: 831-1.0058830-turtle.txt
N-Triples: 831-1.0058830-rdf-ntriples.txt
Original Record: 831-1.0058830-source.json
Full Text
831-1.0058830-fulltext.txt
Citation
831-1.0058830.ris

Full Text

ASPECTS OF PHASE SEPARATION IN AN EXPERIMENTAL MIXER-SETTLER USING TWO SOLVENT EXTRACTION SYSTEMS By NORBERT L. ECKERT B.A.Sc, The University of British Columbia, 1984 SUBMITTED IN PARTIAL FULFILLMENT OF REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Chemical Engineering) We accept this thesis as conforming to the required standard A THESIS THE THE UNIVERSITY OF BRITISH COLUMBIA 1987 © Norbert L. Eckert, 1987 4 6 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia 1956 Main Mall Vancouver, Canada Department V6T 1Y3 i i ABSTRACT An experimental investigation was undertaken to study the factors affecting phase separation in a specially constructed, laboratory scale mixer-settler. Two phase systems were used: 1. A laboratory prepared HSLIX64N-copper phase system, similar to that used in commercial copper solvent extraction processes. 2. A phase system obtained directly from the uranium extraction circuit of the Key Lake Mining Corporation, Sask., millsite. A settler scale-up criterion relating dispersion band thickness to specific settler flow of dispersed phase, was found to have considerable merit. Besides being dependent on specific settler flow, the dispersion band thickness was found to be a function of the phase ratio (for system 2 only), dispersion introduction level (for system 2 only), and temperature. Mixing intensity had no appreciable effect with either system. Microscopic examination of the dispersion produced with system 2 revealed the existence of double dispersions; that is, drops within drops. A photomicroscopic technique was used to undertake a drop size investigation of the dispersion produced with system 2. Drop size was found to be independent of dispersion throughput, a weak function of impellor speed and a relatively strong function of the phase ratio. With the exception of organic continuous operation with system 1, i t was possible, based on the holdup and drop size profile within the dispersion band, to distinguish two horizontal sublayers within the dispersion band. In the "even concentration sublayer", comprising the majority of the dispersion band, the holdup and average drop size is nearly constant throughout. In the "dense concentration sublayer", both the holdup and the average drop size increase sharply as the coalescence front is approached. iv Table of Contents ABSTRACT i i LIST OF TABLES v i i LIST OF FIGURES v i i i ACKNOWLEDGEMENT xi CHAPTER 1 - INTRODUCTION 1 1 . 1 General 1 1.2 Research objectives 4 CHAPTER 2 - BACKGROUND 7 2.1 General settler operation 7 2.2 Settler scale up 10 2.3 Separation mechanism in a deep layer settler 14 2.4 Entrainment and phase continuity 18 2.5 Effect of mixer operation on settler performance 21 2.6 Temperature ...23 2.7 Surface effects 24 2.8 Effect of mass transfer on coalescence 25 2.9 Installation of horizontal plates .26 CHAPTER 3 - DESCRIPTION OF TYPICAL PROCESSES 28 3.1 Introduction 28 3.2 Description of KLMC solvent extraction process 28 3.3 Description of NCCM copper extraction plant 37 CHAPTER 4 - EXPERIMENTAL PHASE SYSTEMS 42 CHAPTER 5 - EXPERIMENTAL APPARATUS 43 5.1 Introduction 43 5.2 Mixer-settler 43 5.3 Interphase regulator 53 5.4 Storage vessels 53 5.5 Pumps 55 V 5.6 Temperature regulation 55 5.7 Flow metering and regulation 56 5.8 Panel board 57 5.9 Safety considerations 57 CHAPTER 6 - EXPERIMENTAL PROCEDURE 59 6.1 Properties of the phases 59 6.2 Mixer-settler start-up procedure 59 6.3 Time for dispersion band to reach steady state 62 6.4 Sampling for dispersion band holdup 65 6.5 Determination of dispersion drop size 66 6.6 Orientation of microscope 69 CHAPTER 7 - RESULTS AND DISCUSSION 72 7.1 General observations of mixer-settler operation 72 7.1.1 Dispersion band appearance « 72 7.1.2 Crud 76 7.1.3 Operation at desired phase continuity 78 7.1.4 Drift in physical behavior of phases -^ 80 7.1.5 Effect of introduction level on AH 81 7.2 Qd/A - AH relationship 88 7.2.1 Case 1 - HSLIX64N-copper system (org. cont.) 89 7.2.2 Case 2 - HSLIX64N-copper system (org. cont.) 93 7.2.3 Case 3 - Key Lake system (org. cont.) 95 7.2.4 Qd/A-AH data at Key Lake, Sask 97 7.3 Effect of operating temperature 99 7.4 Effect of impellor speed 101 7.5 Dispersion band holdup profiles 105 7.6 Dispersion drop size distribution 110 7.6.1 General observations 110 7.6.2 Double and triple dispersions 114 7.6.3 Effect of O/A ratio and Qd on mixer drop size 118 vi 7.6.4 Effect of impellor speed on mean drop size 122 7.6.5 Drop size profile within dispersion band 124 7.7 Theoretical terminal velocity 129 CHAPTER 8 - SUMMARY AND CONCLUSIONS 136 NOMENCLATURE 141 REFERENCES 143 APPENDIX I-PLOTS OF CHAPTER 7 148 APPENDIX 11-CHARACTERISTIC CURVE STATISTICAL ANALYSIS 166 APPENDIX III - ERROR ANALYSIS; GENERAL COMMENTS 177 v i i LIST OF TABLES TABLE DESCRIPTION PAGE 1 Key Lake pregnant solution composition. 30 2 Key Lake typical product analysis. 35 3 Dimensions of NCCM solvent extraction process 41 mixer-settlers. 4 Summary of equipment specifications and 45 ranges in capability. 5 Physical properties of the experimental 60 phases. 6 Effect of dispersion introduction level on 83 dispersion band thickness (system: HSLIX64N-copper; temperature: 24°C). 7 Effect of dispersion introduction level on 84 dispersion band thickness (system: Key Lake; temperature: 35°C). vi i i L I S T OF F I G U R E S FIG. DESCRIPTION PAGE 1 Wedge and deep-layer dispersion bands. 8 2 Theoretical dispersion band holdup profiles. 15 3 Flow diagram of KLMC solvent extraction 31 process. 4 KLMC mixer-settler design and dimensions. 36 5 Flow diagram of NCCM solvent extraction 38 process. 6 Schematic arrangement of experimental 44 apparatus. 7 Photograph of mixer-settler Gore-tex seal. 47 8 Dimensions of experimental mixer-settler. 48 9 Impellor geometry. 50 10 Photograph of mixing chamber. 51 11 Interphase regulator. 54 12 Photograph of panel board. 58 13 Photograph of mixer-settler operation with 63 Key Lake phase system. 14 Time for dispersion band thickness to achieve 64 steady-state after introduction of dispersion into settler (HSLIX64N-copper phase system). 15 Design of glass-walled c e l l used for 67 determination of dispersion drop size. 16 Photograph of arrangement of microscope and 71 c e l l . 17 Effect of dispersion introduction level for 85 Key Lake phase system for cases identified in Table 7. ix 18 Dispersion band thickness versus specific 90 settler flow of dispersed phase (effect of O/A ratio and settler area), HSLIX64N-copper system, organic continuous operation, 24°C, 300rpm. 19 Dispersion band thickness versus specific 94 settler flow of dispersed phase (effect of O/A ratio and settler area), HSLIX64N-copper system, aqueous continuous operation, 24°C, 300rpm. 20 Dispersion band thickness versus specific 96 settler flow of dispersed phase (effect of O/A ratio and settler area), Key Lake phase system, organic continuous operation, 35°C, 300rpm. 21 Dispersion band thickness versus specific 100 settler flow of dispersed phase (effect of operating temperature), HSLIX64N-copper phase system, organic continuous operation, 300rpm. 22 Dispersion band thickness versus specific 102 settler flow of dispersed phase (effect of impellor speed), HSLIX64N-copper phase system, organic continuous operation, 24°C. 23 Dispersion band thickness versus specific 103 settler flow of dispersed phase (effect of impellor speed), HSLIX64N-copper phase system, aqueous continuous operation, 24°C. 24 Dispersion band thickness versus specific 104 settler flow of dispersed phase (effect of impellor speed), Key Lake phase system, organic continuous operation, 35°C. 25 Dispersion band holdup profiles, 106 HSLIX64N-copper phase system, organic continuous operation, 24°C, 300rpm. 26 Holdup profiles within dispersion band, 107 HSLIX64N-copper system, aqueous continuous operation, 24°C, 300rpm. 27 Holdup profiles within dispersion band, Key 108 Lake phase system, organic continuous operation, 35°C, 300rpm. X 28 Change in arithmetic mean drop size of 113 dispersion sample with time after withdrawal from mixer, Key Lake phase system, organic continuous operation, 35°C, 300rpm. 29 Photograph of double dispersion sample 115 withdrawn from settler inlet, Key Lake phase system, organic continuous operation, 0/A=1.5, 35°C, 300rpm. 30 Effect of O/A ratio and dispersed phase 119 flowrate on arithmetic mean drop diameter at mixer overflow, Key Lake phase system, organic continuous operation, 35°C, 300rpm. 31 Photographs showing effect of O/A ratio on 121 drop size of dispersion withdrawn from mixer overflow, Key Lake phase system, organic continuous operation, 35°C. (a) O/A=0.67, (b) O/A=3.0. 32 Effect of impellor speed on arithmetic mean 123 drop diameter at mixer overflow, Key Lake phase system, organic continuous operation, 35°C. 33 Drop size profile within dispersion band for 125 O/A=0.67, Key Lake phase system, organic continuous operation, 35°C, 300rpm. 34 Drop size profile within dispersion band for 126 0/A=1.5, Key Lake phase system, organic continuous operation, 35°C, 300rpm. 35 Drop size profile within dispersion band for 127 O/A=3.0, Key Lake phase system, organic continuous operation, 35°C. 36 Theoretical terminal drop velocity versus 132 drop diameter for different dispersion concentrations, HSLIX64N-copper phase system, organic continuous operation, 24°C. 37 Theoretical terminal drop velocity versus 133 drop diameter for different dispersion concentrations, HSLIX64N-copper phase system, aqueous continuous operation, 24°C. 38 Theoretical terminal drop velocity versus 134 drop diameter for different dispersion concentrations, Key Lake phase system, organic continuous operation, 35°C. xi ACKNOWLEDGEMENT I would like to express my thanks to Dr. Lynton Gormely for his guidance and patience throughout the course of this work I would also like to thank Key Lake Mining Corporation for their contribution of liquid samples for my experiments. Financial support from the Natural Sciences and Engineering Research council is gratefully acknowledged. CHAPTER 1 - INTRODUCTION 1.1 GENERAL Solvent extraction i s a process which has achieved widespread application in the recovery of metals from leached ores and concentrates. The process consists of an extraction section and a stripping section, with additional operations being required in s p e c i f i c circumstances. The purpose of the extraction section is to purify and concentrate the metal of interest by selective extraction. The aqueous solution containing the metal of interest i s intimately contacted with an immiscible organic phase containing an organic extraction reagent. The extraction reagent reacts with the metal of interest, s e l e c t i v e l y transferring i t across the aqueous/organic interface into the organic phase. Undesirable constituents are l e f t behind. An e f f e c t i v e organic extraction reagent i s now known for v i r t u a l l y every metal. The purpose of the stripping section i s to transfer the extracted metal into a second aqueous phase at a concentration where further processing i s economical. The desired metal is now in an aqueous phase at a higher purity and at a higher concentration than i t was in the o r i g i n a l aqueous solution. The metal can then be recovered from solution by methods such as e l e c t r o l y s i s or p r e c i p i t a t i o n . 2 A f u l l description of two typical solvent extraction processes is provided in Chapter 3. In the hydrometallurgical industries, mixer-settlers are generally used for contacting and separating the immiscible liquid phases. Mixer-settlers consist of a mixing chamber and a settling chamber. The function of the mixer is to disperse the two entering phases by means of an agitator, thus promoting mass transfer by creation of a large interfacial area (typically 500-1000 m2/m3 of mixer (Hanson,1983)), and to provide the residence time required for the desired approach to equilibrium. The dispersion generated in the mixer then enters the settling chamber and disengages by means of a gravitational settling/coalescence mechanism. The settler typically constitutes 80-90% of the total mixer-settler volume. Due to the reversibility of the organo-metallic reactions, a cascade of several mixer-settlers providing countercurrent contacting of the phases is generally required in order to achieve the desired extraction. In the hydrometallurgical industries, mixer-settlers are generally chosen over other contacting devices (such as column extractors) due to the following characteristics: 1 . Mixer-settlers have the capacity for handling large liquid throughputs, such as those found in hydrometallurgical processes. Mixer settlers can typically handle flows up to 800 m3/hr compared to a maximum of 100 m3/hr for columns. 3 2. A close approach to equilibrium between phases can be expected in mixer-settlers, reducing the number of stages required. 3. Mixer-settlers are rugged and are r e l a t i v e l y easy to construct and operate. Column fabrication involves more complex mechanical engineering techniques. 4. Liquid systems containing some suspended s o l i d s , such as those commonly processed in the hydrometallurgical industries, can be treated in mixer-settlers, since shut down for solids removal is r e l a t i v e l y simple or even unnecessary (At Key Lake so l i d s are removed during operation by 'vacuuming' them o f f ) . Solids are d i f f i c u l t to handle with columns. 5. Phases can be easily r e c i r c u l a t e d from the s e t t l e r to att a i n the desired phase r a t i o in the mixer. 6. Process changes are possible at any time by adding or removing sidestreams. 7. E q u i l i b r i a in these systems are such that r e l a t i v e l y few stages are required, with s i g n i f i c a n t mass transfer in each stage. Each stage can therefore afford to be more c a p i t a l intensive. There i s considerable incentive for reducing the s e t t l e r size required to achieve phase separation. Increased size represents higher c a p i t a l cost for equipment, increased plant area, and higher organic inventory (reagents are expensive so that the cost of increased organic inventory can be high). Nevertheless, s e t t l e r design procedures have frequently proven 4 to be unreliable, resulting in the use of large safety factors, and consequent settler overdesign. The performance of a settler is influenced by a variety of parameters including throughput, temperature, phase flow ratios and continuity, and mixing conditions. Furthermore chemical composition influences the settler performance and thus different systems may behave differently. A host of sophistications such as flow distributors and baffles have been introduced into some designs to enhance performance. In order to improve the r e l i a b i l i t y of settler design procedures so that equipment can be confidently and conservatively designed, without the use of excessive safety factors, the effects of these parameters should be understood. Also, inherent in accurate design is an understanding of the basic mechanism of phase separation in the settler. It is the aim of this thesis to develop an understanding of these factors affecting settler performance and to investigate the separation mechanism. 1.2 RESEARCH OBJECTIVES This project involved the experimental study of phase separation in a laboratory scale mixer-settler. The following areas were investigated: 1. The effect of settler throughput on dispersion band thickness. The data obtained was compared with previously established scale-up c r i t e r i a to confirm or question their validity. A comparison was made between the laboratory settler performance and the reportedly enhanced performance obtained with the KREBS designed mixer-settlers used at the Key Lake, Saskatchewan, uranium extraction plant. Experimentation with liquid phases obtained from Key Lake facilitated this direct comparison. The effect of dispersion type (organic or aqu-eous continuous) and the organic to aqueous ratio on settler performance. The effect of temperature. The effect of mixing intensity on phase separation. The effect of settler inlet configuration; i.e., method of introduction of dispersion into the settler. The mechanism of phase separation. This was studied through a detailed examination of the dispersion band structure, as follows: visual and photographic examination of the dispersion band including an investigation of the dispersion drop size distribution. - sampling to determine the organic to aqueous ratio at different points in the dispersion band. - observation of the flow pattern in the dispersion band. 6 Based on this examination, the applicability of the separation mechanism proposed by Barnea and Mizrahi (1975) to the two industrially relevant systems investigated in this study could be assessed. 7. A comparison of the phase separation behavior of the 2 experimental phase systems. 7 C H A P T E R 2 - B A C K G R O U N D 2.1 G E N E R A L S E T T L E R O P E R A T I O N The product of the mixer consists of droplets of one liquid phase dispersed in a continuum of the other. Upon entering the settler, this dispersion behaves as a third liquid phase of a density intermediate to the two separated bulk phases. Consequently i t spreads out horizontally in the settler, between the two bulk phases,, producing a "dispersion band". Depending on the flowrate of dispersion to the settler, the dispersion band may take either of two forms. When the flow rate of dispersion to the settler is relatively low, the time required for phase separation is less than the time required for horizontal spreading and the dispersion band does not extend over the entire length of the settler. The dispersion band adopts the form of a wedge (Fig. 1(a)). As the settler throughput is increased, the length of this wedge increases until the dispersion reaches the end of the settler. Beyond this point, the settler is said to be operating with a "deep-layer" dispersion band of near uniform thickness (Fig. 1(b)). Further increases in settler throughput result in an increasing thickness of this layer. With increasing settler loading the thickness of the dispersion band eventually increases to the point where i t exits through either or both the organic overflow weir and the 8 ORGANIC OUTLET AQUEOUS OUTLET a) Wedge-shaped band ORGANIC OUTLET A Q U E O U 8 OUTLET b) Deep-layer band Figure 1. Wedge and deep-layer dispersion bands. 9 aqueous d i s c h a r g e p o r t . T h i s u n d e s i r e a b l e s i t u a t i o n i s known as f l o o d i n g . F l o o d i n g c o n d i t i o n s i n a d e e p - l a y e r l i q u i d / l i q u i d s e t t l e r were s t u d i e d by Barnea (1978). The i n t e r f a c e between the d i s p e r s i o n band and the s e p a r a t e d d i s p e r s e d phase i s known as the " c o a l e s c e n c e f r o n t " or a c t i v e i n t e r f a c e , s i n c e the d r o p l e t s must c o a l e s c e i n o r d e r t o j o i n the b u l k homophase. The i n t e r f a c e s e p a r a t i n g the d i s p e r s i o n band from the b u l k c o n t i n u o u s phase i s r e f e r r e d t o as the " s e t t l i n g f r o n t " . S i n c e d r a i n a g e of the c o n t i n u o u s phase from the d i s p e r s i o n band r a t h e r t h a n c o a l e s c e n c e o c c u r s h e r e , t h i s boundary may a l s o be c a l l e d t he " p a s s i v e i n t e r f a c e " . D i l u t e d i s p e r s i o n s ( t h o s e w i t h a d i s p e r s e d phase volume f r a c t i o n l e s s than a p p r o x i m a t e l y 1 5 %), g e n e r a l l y do not form d i s p e r s i o n bands. Here, s e t t l i n g of l i q u i d d r ops c o n s t i t u t e s the s e p a r a t i o n mechanism, whereas i n the case of a d i s p e r s i o n band, phase s e p a r a t i o n i s l i m i t e d by i n t e r d r o p l e t c o a l e s c e n c e . U n t i l r e c e n t l y , o p e r a t i o n of s e t t l e r s i n the wedge-shaped d i s p e r s i o n band mode was f a v o r e d by some i n v e s t i g a t o r s ( J e f f r e y s , D a v i e s , P i t t , 1970). I t was reasoned t h a t sudden changes i n o p e r a t i n g c o n d i t i o n s (a sharp i n c r e a s e i n t h r o u g h p u t , f o r example) would not be as l i k e l y t o r e s u l t i n f l o o d i n g because of the g r e a t e r e x c e s s s e t t l e r c a p a c i t y a v a i l a b l e . I t i s g e n e r a l l y a c c e p t e d t o d a y , however, t h a t o p e r a t i o n i n t h i s mode r e p r e s e n t s a case of se v e r e s e t t l e r o v e r d e s i g n . T h i s i n v o l v e s a p e n a l t y due t o i n c r e a s e d c o n s t r u c t i o n c o s t s and due t o i n c r e a s e d o r g a n i c i n v e n t o r y 10 requirements. Provided the dispersion band depth is not too great, it is possible to operate deep-layer settlers and s t i l l provide a sufficient safety margin to ensure that flooding can not occur under usual operating circumstances. Despite the advantages of operating with a deep-layer dispersion band, some industrial settlers do operate in the wedge-shaped mode. For example, the settlers used in the extraction circuit of the world's largest copper extraction plant, at Chingola, Zambia, operate with a wedge-shaped band occupying only one half to two thirds of the settler length (Orjans et a l . , 1977). The overdesign of these settlers is a result of uncertainty regarding the r e l i a b i l i t y of scale-up methods and illustrates the need for scale-up c r i t e r i a which can be used with confidence. 2.2 S E T T L E R S C A L E - U P To maintain reasonable capital cost, practical settler design should be based on use of a deep-layer dispersion band. Therefore this discussion w i l l be restricted to scale-up procedures applicable to this mode of operation. For a deep-layer settler, the thickness of the dispersion band is a measure of the approach to flooding and can therefore be used to characterize the flow capacity of the settler. Ryon, Daley and Lowrie (1959) working with the Dapex process for uranium extraction from sulfuric acid leach 11 solution 1, observed that a straight line was produced if this thickness was plotted against the dispersed phase throughput per unit horizontal cross-sectional area on log-log coordinates. Therefore, the following relationship was proposed: AH=k(Qd/A)y (1) The authors' data gave the value of y as 2.5 and the correlation was found to apply over a scale-up ratio of 1000:1. The dispersion band thickness was found to be dependent only on the flow of dispersed phase per unit area, and not on that of the continuous phase. Furthermore, the thickness was observed to be independent of the power input per unit mixer volume. Some investigators dispute Ryon1s contention that the O/A ratio is not important in determining AH (Mizrahi and Barnea, 1973). Gondo and Kusonoki (1969) obtained a correlation of the following form: AH = 0«. 9N 2. 8(Q/A) 3. 1 (2) This correlation indicates that the phase ratio and mixer power do affect settler capacity and is therefore in disagreement with Ryon. Other experimental data confirms the importance of the phase ratio (Mizrahi and Barnea, 1975, Part IV). However, recent pilot plant studies for the copper tailings leach plant at Chingola, Zambia (Orjans et a l . , 1977) indicate that dispersion band depth depends only on the flow of dispersed phase, thus lending some support to Ryon's contention. 1 (using di(2-ethyl-hexyl) phosphoric acid dissolved in kerosene modified with tributyl phosphate) 1 2 The effect of power input per unit mixer volume is also unclear. For some systems, the dispersion band thickness has been shown to increase with impellor speed. For others however, including that tested by Ryon, experiments indicate that dispersion band thickness may stay relatively constant with increasing impellor speed over a wide range of operating values. For systems used in uranium extraction, phase separation is reported to be uninfluenced by the intensity or the duration of the mixing (Merritt, 1971). Equation (1) remains the most commonly used correlation for scale-up of settlers. The value of the exponent y is typically in the range 1-7 depending on the phase system. However, i t is generally acknowledged that the value of the constants k and y may change with phase ratio or mixing intensity as well as phase continuity and temperature, and that this should be investigated experimentally for each phase system. As an estimate for sizing cylindrical horizontal settlers, with a length/diameter ratio of 4, Treybal (1963) recommends the following rule of thumb: D=8.4(Q C+Q d)°. 5 (3) where Qcand Q^are given in m3/s and D is given in metres. If the horizontal cross-sectional area (A) of the settler is approximated as D(4D), this equation can be modified to give: (Qc+Qd)/A = 0.0035 m3/m2/s =12.6m3/m2/hr (4) 1 3 Thus, equation (3), in effect, specifies a specific settler flow based on total dispersion flow of 12.6 m3/m2/hr for this settler geometry. Equation (3) is intended only for a preliminary estimate in the absence of any actual operating data for the system of interest. A rule of thumb of 1-1.5 USGPM/ft2 (2.5-3.7 m3/m2/hr) is used in the copper industry. This lower specific flow, compared to that recommended by Treybal would be partly explained by the different settler geometries to which they pertain. In the copper industry, settlers are rectangular and large in area (100 - 400 m2, typically) and relatively shallow (less than 1.2 m). Since the horizontal, cylindrical settler geometry specified by Treybal would have a greater depth for a given area than a typical settler used in the copper industry, the higher settler loading recommended by Treybal would be expected. Correlations between a standardized separation time obtained from batch separation tests and the dispersion throughput per unit settler area in a continuous settler have been proposed by several investigators (Barnea and Mizrahi, 1975, part IV), (Golob and Modic, 1977). However, such procedures, using batch separation results have led to large errors when applied to the design of continuous deep-layer settlers. Despite the inadequacies of such methods, they can be useful for obtaining a general impression of the settling behavior of various systems prior to conducting continuous flow 14 tests. 2.3 SEPARATION MECHANISM IN A DEEP LAYER SETTLER Phase separation in a deep-layer settler involves both settling and coalescence phenomena. A general description of the separation mechanism in a deep-layer settler was presented in a series of papers by Barnea and Mizrahi (1975), based on an experimental investigation using a phase system relevant to the industrial process for the extraction of magnesium bromide and chloride from Dead Sea end brine using butanol (1975). Observations led to a proposed qualitative model for the separation process. Barnea and Mizrahi sampled the dispersion band at various distances from the coalescence front to determine the local volume fraction of dispersed phase (holdup). It was found that, for constant settler feed composition, plotting holdup against the dimensionless vertical location in the dispersion band (obtained by dividing the actual distance from the coalescence front by the dispersion band thickness) gave a constant curve for any dispersion thickness (Fig. 2). Based on this observation, two main sublayers were identified within the dispersion band. The "dense sublayer", in the vicinity of the coalescence front typically occupied 10 to 20% of the total volume of the dispersion band. It was structured like foam and was characterized by a hold-up level much higher than that of the 15 * from Barnea and Mizrahi, 1975, Parti. Figure 2. Theoretical dispersion band holdup profiles. 16 settler feed. The holdup was unity at the coalescence front and decreased sharply with increasing distance from the coalescence front. The "even concentration sublayer" comprised the remainder of the dispersion band. The holdup in this sublayer was nearly constant at a level slightly below that of the feed, decreasing slightly and nearly linearly towards the settling front where it dropped sharply to zero. Investigation of both continuous and batch separations and observation of the flow pattern of the continuous and dispersed phases within the dispersion band led to the development of a simplified picture of phase separation in mixer-settlers and of the roles played by the two sublayers in the process. The dense sublayer is characterized by a high rate of interdroplet coalescence and consequently increasing average drop diameters as the coalescence front is approached. Near the coalescence front, significant drop deformation is noted. It was demonstrated in batch tests that the rate of coalescence at the coalescence front i t s e l f is proportional to the effective weight of droplets in the dense sublayer. Pressure transmission between layers of drops increases the force available for "squeezing out" the thin film of continuous phase separating the layer of drops at the coalescence front from the separated bulk phase. Any change in settler feed rate results in an adjustment of dense sublayer thickness and holdup gradient and the establishment of a new equilibrium, so that the effective 17 dense s u b l a y e r weight i s changed i n d i r e c t p r o p o r t i o n t o the f e e d r a t e and s u b s e q u e n t l y c hanging the r a t e of c o a l e s c e n c e a t the i n t e r f a c e . Thus, the p r o g r e s s i o n of d r o p l e t s i n the s u b l a y e r i s dependent on the o c c u r r e n c e of a c o a l e s c e n c e event and the f l o w i s e s s e n t i a l l y u n i d i r e c t i o n a l ( p l u g f l o w ) . In c o n t r a s t t o t h e dense s u b l a y e r , i t was o b s e r v e d t h a t the d r o p l e t s w i t h i n the even c o n c e n t r a t i o n l a y e r moved i n d i f f e r e n t d i r e c t i o n s . Drops whose h i n d e r e d s e t t l i n g v e l o c i t y was l e s s than t h a t of the d r a i n i n g c o n t i n u o u s phase were e n t r a i n e d towards the s e t t l i n g f r o n t . A f t e r a p e r i o d of time the s e d r o p l e t s were o b s e r v e d t o r e v e r s e t h e i r d i r e c t i o n . T h i s o b s e r v a t i o n was a s c r i b e d t o a c o l l i s i o n / c o a l e s c e n c e o c c u r r e n c e w i t h a n o t h e r d r o p l e t , p r o d u c i n g a drop whose di a m e t e r i s g r e a t e r than the c r i t i c a l d i a m e t e r n e c e s s a r y f o r f l o w i n the d i r e c t i o n of the c o a l e s c e n c e f r o n t . From these o b s e r v a t i o n s i t was p o s t u l a t e d t h a t t h e even c o n c e n t r a t i o n s u b l a y e r a d j u s t s t o the t h i c k n e s s which w i l l p r o v i d e the r e s i d e n c e time which a l l o w s the s m a l l e s t d r o p l e t s t o grow to the c r i t i c a l s i z e by t h i s c o l l i s i o n / c o a l e s c e n c e mechanism. S i n c e the even c o n c e n t r a t i o n s u b l a y e r o c c u p i e s most of the d i s p e r s i o n band, Barnea and M i z r a h i c o n s i d e r e d i t t o be the r a t e d e t e r m i n i n g s u b l a y e r . I t was t h e r e f o r e c o n c l u d e d t h a t the c a p a c i t y of a deep l a y e r s e t t l e r i s l i m i t e d by h i n d e r e d s e t t l i n g and c o a l e s c e n c e i n the even c o n c e n t r a t i o n l a y e r r a t h e r than c o a l e s c e n c e a t the c o a l e s c e n c e f r o n t . 18 This model suggests that the capacity of a s e t t l e r could be increased by production of a dispersion having a r e l a t i v e l y narrow size d i s t r i b u t i o n curve, thus eliminating the " t a i l " commonly exhibited by drop size d i s t r i b u t i o n s . 2.4 ENTRAINMENT AND PHASE CONTINUITY If a g i t a t i o n i s stopped during batch operation, i t i s often found that disengagement of the immiscible phases occurs in two stages. In the f i r s t stage (primary break), the vast majority of the dispersion coalesces and separates very rapidly. However, the continuous phase generally remains clouded to some degree by very fine droplets of the dispersed phase. There may be a much lesser quantity of continuous phase droplets in the separated dispersed phase as well. Separation of these fine droplets (secondary break) occurs very slowly. An analogous situation exists for a continuously operated mixer-settler. The "primary break" i s achieved in the dispersion band, leaving fine droplets suspended in the separated continuous phase. As indicated by the scale-up correlations of section 2.2, s e t t l e r s are designed to f a c i l i t a t e the occurrence of only the primary break since i t i s impractical to u t i l i z e s e t t l e r s large enough to enable the secondary break to occur completely. Consequently, suspended droplets exit the s e t t l e r and are referred to as entrainment. In designing mixer-settlers, i t i s important to consider the implications a design may have with 19 regards to entrainment. Entrainment can constitute a significant source of solvent loss or can result in contamination of process streams. If excessive, i t can become a c r i t i c a l factor in the economic fe a s i b i l i t y of a process. As such, measures must be taken to minimize i t . It is generally accepted that entrainment is primarily generated as a result of the formation of fine droplets in the mixer. It can typically be minimized by the avoidance of impellors which produce high shear rates (small diameters and high speeds). A second factor, of much lesser importance, is the reported formation of very fine droplets of either phase during drop-drop coalescence and coalescences at the phase boundary or at solid surfaces (Davies et a l , 1970). This mechanism could partially account for the fact that there is a small amount of entrainment of continuous phase in the separated dispersed phase leaving most commercial settlers. Air ingestion into the mixer has also been observed to increase entrainment (Orjans et al.,1977). Through a suitable mixer start-up procedure, i t is generally possible to disperse either of the two contacting phases within a certain range of flow ratios. This range is referred to as the ambivalence region and typically l i e s in the O/A range of 1/3 to 3/1 (McClary and Mansoori, 1978). Although aqueous continuous operation wi l l generally result in narrower dispersion bands (Orjans et a l . , 1977), (Henkel, Lott et a l . , 1972), industrial mixer-settlers are generally operated organic 20 continuous. This is done since organic continuous operation gives the lowest organic entrainment and loss of organic is the dominant consideration. Entrainment of aqueous phase is usually not as serious. Fluctuations in flow rates or changes in the physical properties of the phases (viscosity, density, interfacial tension) can result in phase inversion. The increase in organic entrainment which may result makes this an undesirable phenomenon. Although i t may be d i f f i c u l t to visually determine which phase is continuous, phase continuity can be established by a conductivity probe since an aqueous continuous dispersion is conductive while an organic continuous one is not. Orjans et a l . (1977) u t i l i z i n g N'Changa pilot plant studies, found that for organic continuous operation, entrainment of aqueous phase may increase dramatically i f a certain c r i t i c a l linear velocity of the organic phase above the dispersion band is exceeded. Therefore, in order to minimize this velocity, for a given settler area and depth, long and narrow settlers should be avoided in favor of wider, shorter ones. The width of the settler would, however, be limited by problems associated with non-uniform distribution of dispersion into a wide settler and uneven flow within i t . Although entrainment is an important factor affecting mixer-settler performance, attempts to predict the level of 21 entrainment encountered in plant-size equipment from laboratory experiments have proven largely unsuccessful due to the small scale of the lab equipment (DeMent, 1979). Entrainment was not explicitly examined in this study. However, the possible effects of various designs and operating conditions on entrainment should be kept in mind. 2.5 EFFECT OF MIXER OPERATION ON SETTLER PERFORMANCE The function of the mixer is to provide the residence time and the degree of agitation necessary for the desired extraction efficiency. Although extraction efficiencies can be improved by increasing the mixing intensity, consideration must also be given to the possible effects this may have on the phase separation process. Increased mixing intensity can produce a dispersion which is more d i f f i c u l t to separate, resulting in greater settler size requirements and higher entrainment levels. Consequently, the mixer-settler operation must be optimized as a whole. One would expect, based on the phase separation mechanism proposed by Barnea and Mizrahi, that the performance of the settler would be influenced by the dispersion drop size distribution produced in the mixer. The drop size produced by the mixer can be considered to be the result of an equilibrium between droplet break-up and interdroplet coalescence (Glasser 22 et a l . , 1976). Depending on the relative effect of agitation intensity on these opposing mechanisms, phase systems can be classified into either of two general and somewhat subjective categories; "sensitive" or "insensitive" (Mizrahi and Barnea, 1974). With sensitive systems, increases in agitation energy result in decreased phase separation rates (thicker dispersion bands). However, an asymptotic value is eventually approached where further increases in mixing intensity do not affect separation rates appreciably. This asymptotic behavior can be explained as follows: Increasing mixing intensity simultaneously affects both the droplet break-up mechanism (as a result of higher shearing forces) and the re-coalescence mechanism (due to increased co l l i s i o n rates between droplets). Near the asymptotic value, increases in mixing intensity affect these two opposing mechanisms such that the drop size distribution produced by the mixer changes only minimally. With insensitive systems, i t is suggested that this asymptotic value is reached at a relatively mild mixing intensity. Phase separation rates do not change appreciably with energy input in the normal range of mixer operation. Experimentation with relatively insensitive systems is considerably simplified since the operation of the settler can be investigated independently of the mixer. In contrast, for sensitive systems i t is d i f f i c u l t to analyze the mixing and settling operations separately. Even if the mixing regime is 23 kept constant, changes in flow rates would result in a varying residence time within the mixer, and therefore a variation in the energy input per unit of dispersion volume. This could affect the drop size distribution and consequently, the settler performance. 2.6 TEMPERATURE Temperature is an important parameter in settler design. A rule of thumb states that a 20°C temperature increase can typically result in a doubling of settler capacity. The increase in settler capacity with temperature can be attributed to a decrease in liquid viscosities. Decreased viscosities result in higher coalescence rates due to easier drainage of the continuous phase film trapped between the droplets. Also, with respect to the separation mechanism proposed by Barnea and Mizrahi, the decrease in continuous phase viscosity would result in a smaller c r i t i c a l droplet size. The equilibrium between mechanical shearing and re-coalescence in the mixer could also be affected so as to produce a dispersion of a greater or lesser average drop size. Because of the appreciable effect of temperature, a mixer-settler test unit should include a temperature control system. 24 2.7 S U R F A C E E F F E C T S Surface effects can have a significant effect on coalescence rates and phase continuity. Ryon (1959) observed that coalescence was promoted if the dispersed phase wet the mixer settler surfaces. Thus, glass walls would promote coalescence of organic continuous dispersions and plastic walls would promote coalescence in aqueous continuous dispersions. Although installation of packing material wet by the dispersed phase could improve settler capacity, such coalescence elements are rarely incorporated into large industrial mixer-settlers. This is due to the high cost posed by such sophistications and due to possible plugging of the elements by suspended solids (present to some extent in any industrial plant) and the corresponding requirement for periodic removal and cleaning. Urban (1972) found that i t was considerably easier to maintain a particular phase dispersed i f i t did not preferentially wet the mixer walls (ie: the ambivalence region was shifted). This applied for a vessel diameter of approximately 100 drop diameters. For small test equipment where the surface area is large relative to the volume, surface effects could become significant and might invalidate the extrapolation of small 25 scale results to large scale equipment. Therefore, experimental test units should be as large as possible. 2.8 EFFECT OF MASS TRANSFER ON COALESCENCE Observations of single drops suspended in a continuous liquid phase (Groothius and Zuiderweg, 1960) or rising to a flat interface (MacKay and Mason, 1961) indicate that mass transfer may promote or hinder coalescence depending on the direction of mass transfer. If mass transfer is from the liquid drops to the continuous phase, coalescence increases. Transfer from the continuous to the dispersed phase decreases coalescence. This behavior is attributed to gradients in interfacial tension caused by non-uniform mass transfer along the interface. It has been pointed out that higher solute concentration reduces interfacial tension (Treybal, 1963). Mass transfer from the droplet to the continuous phase produces a region of low interfacial tension in the film separating the droplets, resulting in enhanced drainage and coalescence. Transfer in the opposite direction produces a high regional interfacial tension which tends to draw in the continuous phase thus opposing film drainage. These observations are based on studies involving single drops. It is not clear whether the effect is equally significant for the case of mixer-settlers, where there are 26 swarms of d r o p l e t s under the i n f l u e n c e of mass t r a n s f e r , as t h e r e i s no d a t a a v a i l a b l e . For swarms of d r o p l e t s , the i n t e r f a c i a l g r a d i e n t s a l o n g t h e s u r f a c e may be reduced, t h u s l e s s e n i n g the e f f e c t . In the e x p e r i m e n t a l equipment of t h i s s t u d y , the phases were r e c i r c u l a t e d and t h e r e f o r e mass t r a n s f e r d i d not o c c u r . T h i s r a i s e s the q u e s t i o n of whether comparison w i t h i n d u s t r i a l m i x e r - s e t t l e r s (such as thos e a t Key L a k e ) , where mass t r a n s f e r i s , of co u r s e o c c u r r i n g i s v a l i d even i f the same phase system i s used. In t h i s s t u d y , i t i s assumed t h a t s i n c e mass t r a n s f e r i s e s s e n t i a l l y completed i n the mixer w i t h l i t t l e o c c u r r i n g i n the s e t t l e r , c o a l e s c e n c e r a t e s i n the s e t t l e r s h o u l d not be a f f e c t e d . I t i s p o s s i b l e , however, t h a t s e t t l e r performance c o u l d be a f f e c t e d i f the dro p s i z e d i s t r i b u t i o n r e s u l t i n g from the d i s p e r s i o n / c o a l e s c e n c e mechanism i s changed a p p r e c i a b l y by mass t r a n s f e r . T h i s p o s s i b i l i t y , however, was i g n o r e d i n t h i s work. 2.9 INSTALLATION OF HORIZONTAL PLATES In o r d e r t o more e f f i c i e n t l y u t i l i z e a s e t t l e r ' s volume, u n i t s have been d e s i g n e d c o n t a i n i n g h o r i z o n t a l p l a t e s which, i n e f f e c t , p a r t i t i o n a s i m p l e box-type m i x e r - s e t t l e r i n t o a s e t of t h i n s e t t l e r s superimposed one on t o p of a n o t h e r . Improvements i n s e t t l e r c a p a c i t y of s e v e r a l - f o l d have been c l a i m e d f o r such d e s i g n s ( M i z r a h i and Barnea, 1973). 27 Such designs are based on the principle that the volumetric separation capacity of thick dispersion bands is less than that for thinner bands. If the scale-up criterion of Ryon (equation 1) is assumed to be applicable,this can be shown mathematically: AH=k(Q/A)y y»1 (5) Q/(AAH) = Q/V = l / ( k l / y A H 1 " l / y ) (6) Equation (6) indicates that the settler capacity per unit volume of dispersion band (Q/V) decreases with increasing dispersion band thickness and therefore supports the utilization of superimposed, thin settlers. Due to the higher capital cost and to problems which would be encountered as a result of crud, such designs have not, however, achieved widespread application in metallurgical extraction industries. 28 CHAPTER 3- DESCRIPTION OF TYPICAL SOLVENT EXTRACTION PROCESSES 3.1 INTRODUCTION Solvent extraction is a unit operation used for purification and concentration of metals. In order to illustrate the scope of the process and the role played by mixer-settlers in i t , two commercial processes will be described in some detail. These processes are those corresponding to the two systems used in the experiments of this study, namely; (1) The solvent extraction process of Key Lake Mining Corporation (KLMC) u t i l i z i n g Alamine 336 as the reagent for extraction of anionic uranium complexes. (2) The process of N'Changa Consolidated Copper Mines (NCCM) at Chingola, Zambia, using LIX-64N for extraction of copper(II) from sulphate solution. 3.2 DESCRIPTION OF KLMC SOLVENT EXTRACTION PROCESS I n i t i a l steps in the production of yellow cake from ore involve crushing and grinding of the ore to expose the uranium minerals, followed by sulphuric acid leaching of this ground ore slurry to extract over 97% of the uranium from the ore (Neven, Steane, Becker, 1985) (Neven, Gormely, 1982). The solution from acid leaching (known as the pregnant solution) contains suspended solids (50-150 ppm of colloidal 29 s i l i c a ) w hich would c o n s t i t u t e a c r u d problem i n the s e t t l e r s , a c c u m u l a t i n g a t the i n t e r f a c e and r e s u l t i n g i n poor phase disengagement ( R i t c e y , 1980). A d d i t i o n of P o l y o x (water s o l u b l e h i g h m o l e c u l a r weight n o n i o n i c polymers of e t h y l e n e o x i d e ) t o f l o c c u l a t e the s i l i c a , f o l l o w e d by s o l i d - l i q u i d s e p a r a t i o n i n a t h i c k e n e r and f i l t r a t i o n on a p r e c o a t p r e s s u r e f i l t e r , reduces the s i l i c a c o n t e n t t o <10 ppm. T h i s p r e t r e a t e d pregnant s o l u t i o n c o n t a i n s i n a d d i t i o n t o uranium, a number of i m p u r i t i e s such as a r s e n i c , n i c k e l , i r o n , magnesium and aluminum. The c o m p o s i t i o n of the pregnant s o l u t i o n i s g i v e n i n T a b l e 1. D i r e c t p r e c i p i t a t i o n would r e s u l t i n a low grade and unmarketable uranium p r o d u c t . T h e r e f o r e , i m p u r i t i e s a r e e l i m i n a t e d by the s o l v e n t e x t r a c t i o n p r o c e s s b e f o r e p r e c i p i t a t i o n of uranium. A s i m p l i f i e d f l o w diagram of the s o l v e n t e x t r a c t i o n p r o c e s s i s g i v e n i n F i g u r e 3. The e x t r a c t i o n s e c t i o n c o n s i s t s of f o u r m i x e r - s e t t l e r s o p e r a t e d c o u n t e r c u r r e n t l y . The pregnant s o l u t i o n ( f l o w = l 2 0 m 3/hr) i s c o n t a c t e d w i t h an o r g a n i c phase c o n s i s t i n g of 3-6% Alamine 336 d i l u t e d i n a h i g h grade kerosene and a l s o c o n t a i n i n g 3% i s o d e c a n o l . Alamine 336 i s a t e r t i a r y amine manufactured by Henkel ( f o r m e r l y G e n e r a l M i l l s ) . T h i s reagent e x t r a c t s the a n i o n i c uranium complexes from the aqueous a c i d l e a c h s o l u t i o n i n a c c o r d a n c e w i t h the f o l l o w i n g r e a c t i o n s where R 3N r e p r e s e n t s 30 Table 1. Key Lake pregnant s o l u t i o n composition COMPONENT CONCENTRATION (g/1) u3o8 6 -10 AS 2.2 - 4.4 Ni 2.0 - 4.0 Fe 2.5 - 3.5 Mg 0 . 8 - 1 . 5 A l 0.5 - 1.5 R a 2 2 6 800 - 1300 Bq/1 H 2S0 4 15 - 20 TSS <10 ppm * from Neven, Steane, Becker, 1985. P R C O N A N T SOLUTION < R A F F M A T E EXTR. s < S P E N T RE GEN. SOLUTION Ul K a ui SOLUTION Figure 3. EXTR. EXTR. i EXTR. < W A T E R 8 »- 3 X -J Ul O ft. • S C R U B i" STRB* ^2 S C R U B S C R U B wurre HjSoT\_ SOLUTION / STRB* h i STRB* Flow diagram of process. L O A D E D STRB* S O L M KLMC solvent extraction > 32 Alamine 336 (Merritt, 1971): Formation of amine sulphate and bisulphate salts: 2R 3N o r g+ H 2S0 4 = (R 3NH) 2SO U o r g (R3NH)2S04 + H2SO„ = 2(R 3NH)HS0 4 extraction: 2(R 3NH) 2S0 a o r g+ UOzfSOja-" =(R 3NH)„U0 2(S0«) 3 o r g+ 2S0«"2 Alamine 336 is highly selective for uranium in preference to other constituents. Over 90% of the organic phase is a high quality kerosene (either Napoleum 470B or suitable replacement is used, manufactured by Kerr-McGee Corporation) which acts as a carrier or diluting medium for the very viscous extractant. Besides giving the desired flow properties to the organic phase i t has a relatively high flash point to avoid potential fire hazards (Murray and Bouboulis, 1973), (Ritcey and Lucas, 1974). Addition of isodecanol as a modifier improves the solubility of amine salts in kerosene, thus preventing the formation of a third phase and inhibiting the formation of stable emulsions (Ritcey and Lucas, 1974). The organic flow rate is set to achieve maximum uranium loading on the organic, typically 7.5 gpl U 30 8, while maintaining the raffinate U 30 8 at <0.001 gpl. Recycle of organic from the settler to the mixer keeps this phase in the majority in the mixer to f a c i l i t a t e organic phase continuity. 33 Despite the fact that settler capacity would be greater for aqueous continuous operation (Merritt, 1971), the extraction section, as well as a l l other sections are run organic continuous. This is done to minimize organic entrainment and because the tendency for forming stable emulsions is increased with aqueous continuity (Merritt, 1971). Following the four extraction stages, the uranium bearing organic phase (loaded organic) is scrubbed with a dilute sulphuric acid solution (10-20 gpl H2SOft) in three countercurrent mixer-settler stages. This is done to remove quantities of impurities, especially arsenic, which may have been co-extracted with the uranium. Some loading of arsenic onto organic will occur, particularly i f i t is not fully loaded with uranium. The spent scrub aqueous phase is then fed back to the extraction circuit to recover any uranium that may also have been stripped off during scrubbing. Uranium is stripped from the scrubbed, loaded organic phase into an ammonium sulphate strip solution. The pH is carefully regulated between 4.0 and 4.2 by adding vaporized anhydrous ammonia to the strip mixers. At pH values below this range, stripping is poor while at values above this range phase separation may be poor or emulsion formation could result due to precipitation of uranium yellowcake. In the neutral scrub, barren organic from the stripping section is scrubbed with clean water in a mixer-settler to remove entrained ammonium sulphate prior to recycle of the 34 organic phase back to the extraction c i r c u i t . This prevents ammonia from entering the raffinate and subsequently being released to the final effluent. Molybdenum is strongly extracted by Alamine 336, yet is not effectively removed in the stripping circuit (Merritt, 1971). Therefore, a bleed of the barren organic may be contacted with a 5% sodium carbonate solution to remove molybdenum, preventing i t s accumulation in the organic. Final recovery of uranium from the uranium solution is achieved conventionally by chemical precipitation (Merritt, 1971). The strip solution pH is raised to 7.2 using ammonia gas to precipitate the uranium. The yellowcake precipitate is dewatered and roasted at 800 °C to produce the calcine product. A typical analysis of the final yellowcake product is shown in Table 2. Comparison of this analysis with that of the pregnant solution (Table 1) shows the effectiveness of the solvent extraction circuit in removing contaminants. The KLMC solvent extraction plant mixer-settlers are of the KREBS design and dimensions shown in Figure 4. The KREBS design is purported to allow for a 50-75% reduction in settler size over that normally required. This improvement is attributed primarily to two features which result in partial separation of the dispersion prior i t s entrance into the main settling chamber; a conical pump and a superimposed launder. 35 Table 2.Key Lake typical product analysis COMPONENT CONCENTRATION (%) U 30 8 91 - 99 S 0.1 - 4.00 As 0.01 - 0.04 Fe 0.01 - 0.03 MO 0.02 - 0.07 Na 0.01 - 0.05 K <0.01 V <0.01 Ca <0.01 Mg <0.01 B <0.01 Si 0.03 - 0.30 Ni <0.01 Ti <0.01 Zr <0.01 * Impurities expressed on a percent U basis. * from Neven, Steane, Becker, 1985. AQUEOU8 INLET KEY LAKE MIXER SETTLER DIMENSIONS PARAMETER EXTRACTION STAGES (m) SCRUBBING, STRIPPING, REGENERATION STAGES (m) MH (mixer height) MD (mixer diameter) MS L ( l a u n d e r / s e t t l e r length) LH (launder height) SH ( s e t t l e r height) s e t t l e r width launder width 90 44 30 45 50 15 3 .96 2.44 90 44 30 12 50 15 20 40 O J F i g u r e 4 . KLMC m i x e r - s e t t l e r design and dimensions. 37 The c o n i c a l pump t r a n s f e r s the d i s p e r s i o n from the mixer t o the s e t t l e r and ind u c e s i n i t i a l c o a l e s c e n c e as a r e s u l t of the c e n t r i f u g a l f o r c e s d e v e l o p e d . The l a u n d e r f e eds the d i s p e r s i o n t o the o p p o s i t e end of the s e t t l e r and f a c i l i t a t e s f u r t h e r phase s e p a r a t i o n . At t he end of the l a u n d e r , the l i g h t and heavy p a r t i a l l y s e p a r a t e d phases a r e f e d , r e s p e c t i v e l y , t o t h e upper and lower p a r t s of the s e t t l i n g chamber by means of a d j u s t a b l e o v e r f l o w and u n d e r f l o w w e i r s . A s e r i e s of p a r a l l e l p l a t e s e x t e n d i n g a c r o s s the f u l l w i d t h of the s e t t l e r s e r v e s t o e v e n l y d i s t r i b u t e t he p a r t i a l l y s e p a r a t e d d i s p e r s i o n i n the s e t t l e r . I n t he s e t t l e r , a d i s p e r s i o n band forms and f i n a l s e p a r a t i o n i s a c h i e v e d . 3.3 DESCRIPTION OF N'CHANGA CONSOLIDATED COPPER MINE SOLVENT  EXTRACTION PLANT, CHINGOLA, ZAMBIA The NCCM s o l v e n t e x t r a c t i o n p l a n t i s the w o r l d ' s l a r g e s t , y e t i s t y p i c a l of copper s o l v e n t e x t r a c t i o n p r o c e s s e s . I t t r e a t s a t o t a l pregnant l i q u o r f l o w of 3200 m 3/hr i n f o u r p a r a l l e l l i q u i d streams (Hanson, 1983). A s i m p l i f i e d f l o w diagram i s shown i n F i g u r e 5 f o r one stream. The pregnant l i q u o r from s u l p h u r i c a c i d l e a c h i n g of the ore c o n t a i n s 3-4 g/1 Cu a t pH 1.9-2.0. I r o n s a l t s c o n s t i t u t e the major c o n t a m i n a n t . 38 PREGNANT SOLUTION EXTRACTION 1 n J 1 RAFFINATE ELECTRO -WINNING STRIPPING COPPER CATHODE 99%+Cu F i g u r e 5. Flow diagram of NCCM solvent e x t r a c t i o n process. 39 Each pregnant liquor stream is fed to an extraction circuit consisting of a battery of 3 countercurrently operated mixer-settlers and is contacted with an organic phase containing 22 vol% extractant in kerosene. No modifier is required as third phase formation is not a problem with this system. LIX64N (manufactured by Henkel) serves as the extractant. (Note: LIX64N has recently been replaced by HSLIX64N which is concentrated by a factor of 1.51). Copper extraction is achieved by shifting to the right the equilibrium existing between copper ions, the extractant (RH) and the acid solution: (2RH) o r g + (Cu** + SO„- 2) a q= ( R 2 C u ) o r g + <2H* t S 0 a " 2 ) a q Copper is thus selectively transferred from the aqueous to the organic phase, leaving iron and other impurities behind. The loaded organic is then passed to a series of 2 countercurrently operated mixer-settler stages (constituting the stripping section) and contacted with high acid strength spent electrolyte from the electrowinning c i r c u i t . The high acidity (^ 150 g/1 H 2S0 4) shifts the equilibrium to the l e f t , thereby stripping the copper values from the organic. The extractant is thus regenerated and is recycled to the extraction c i r c u i t . 40 Stripping results in a concentrated copper solution (25-60 g/1) which is suitable for electrowinning. At the cathode of the electrowinning c e l l , copper ions are reduced to copper metal of >99% purity. Oxygen is evolved at the lead anode. The mixer-settlers are constructed of stainless steel lined concrete. Mixers are square in cross section and are fitted with a f u l l width horizontal baffle to prevent vortex formation. Dispersion is discharged from the mixer by way of a flared launder leading to a f u l l width entry slot into the settler at the level of the dispersion band. Each settler is provided with two picket fence type baffles. The function of these devices is to ensure even distribution and to reduce settler turbulence. A fence baffle consists of two rows of vertical bars, spaced such that the bars of the second row cover the gaps between the bars of the f i r s t row. Dimensions of the N'Changa mixer settler units are provided in Table 3. 41 Table 3. Dimensions of NCCM solvent extraction process mixer-settlers APPARATUS PARAMETER EXTRACTION STRIP mixers width 5.9 m 4.0 m depth 3.7 m 3.2 m impellor diameter 2.7 m 2.3 m settlers width 12.2 m 12.2 m length 36.5 m 26.2 m depth 0.76 m 0.76 m f i r s t picket fence distance from inlet 13.1 m 9.1 m second picket fence distance from inlet 3Q.5 m 21.9 m * from Hanson, 1983. 42 CHAPTER 4 ~ EXPERIMENTAL PHASE SYSTEMS Two phase systems were used in experiments. Key Lake Mining Corporation provided samples of r a f f i n a t e and loaded organic from the extraction section of the solvent extraction plant. The composition of these samples thus approximated that of the l i q u i d s in the extraction section mixer s e t t l e r s , the plant's largest, and could be expected to possess similar physical properties. Comparison of experimental results and plant operating data was thus v a l i d except for mass transfer e f f e c t s , i f any. A second phase system was prepared by disso l v i n g 8 parts by volume of HSLIX64N (Henkel, l o t 4G14212W) in 92 parts diluent (90% Nap470B and 10% Kermac500T, manufactured by Kerr-McGee) to produce the organic phase. The aqueous phase was prepared by disso l v i n g 3 g/1 of copper, as copper(II)-sulphate, in water and adjusting the pH to 1.9 with H2SO,,. The general behavior of th i s phase system could be compared with that of the Key Lake phase system, providing some indication of the extent to which d i f f e r e n t systems' behavior could be predicted. However, the v a l i d i t y of quantitative comparisons between the experimental HSLIX64N-copper system data and copper solvent extraction plant data would be limited due to the absence of contaminants in the experimental system. 43 CHAPTER 5 - EXPERIMENTAL APPARATUS 5.1 INTRODUCTION Equipment was constructed to f a c i l i t a t e experiments to be conducted allowing for the development of s e t t l e r scale-up c r i t e r i a , for determining the ef f e c t s of operating parameters (temperature, phase r a t i o , phase continuity, mixing intensity and conditions) on s e t t l e r performance, and for studying the effect of s e t t l e r configuration ( s p e c i f i c a l l y dispersion introduction geometry) on phase separation. Since the study was not concerned with mass transfer, the experimental phases were rec i r c u l a t e d . The equipment was designed to be as large as possible while s t i l l being p r a c t i c a l and could handle a maximum flow of 24 1/min of either phase. The schematic arrangement of the entire apparatus is shown in Figure 6 and i t s ranges in c a p a b i l i t y are tabulated in Table 4. Its chief components are described below. 5.2 MIXER-SETTLER The mixer-settler was constructed with 1/8 inch, 316 stainle s s steel with two f u l l , wire reinforced glass walls to f a c i l i t a t e v i s u a l and photographic observation of the mixing and s e t t l i n g operations. A continuous 1/4 inch wide s t r i p of Gore-tex joint sealant (W.L Gore and Assoc.,Inc) followed the edges of the sta i n l e s s ADJUSTABLE HEIGHT ROTAMETERS VALVES INTERPHASE REGULATOR Figure 6. Schematic of experimental apparatus. 45 Table 4. Summary of equipment specifications and ranges in capability. PARAMETER impellor speed aqueous flow organic flow dispersion band thickness temperature O/A ratio mixer dimensions mixer volume settler dimensions settler cross-sectional area RANGE 0-360 rpm 4-24 litres/min 4-24 litres/min 0-30 cm 20-45°C 6/1-1/6 length=30.5cm width=30.5cm depth=41 cm 38 l i t r e s length=l20cm width=30.5cm depth=38cm (length can be reduced with adjustable partition) 0-0.36 m2 46 steel and acted as a gasket between the steel and the glass walls. Gore-tex joint sealant is a soft, foamed form of pure PTFE. It is chemically inert and therefore w i l l not deteriorate or be extracted by the aromatic organic or the acidic aqueous phases. This eliminates the possibility of contaminating the systems with plasticizers, thus altering the physical behavior of the phases such that i t would not be representative of that found under plant conditions. For this reason, plasticized flexible tubing was avoided as well, despite i t s convenience. A leakproof seal was effected by applying pressure to the glass walls with 40 screws tapped into a steel "clamping frame". A relatively light clamping force was sufficient to compress the Gore-tex to a thin ribbon,, thereby ensuring a reliable seal. Sections of steel square tubing with rubber pads on one face served to distribute the force evenly along the joints and prevented chipping of the glass by the clamping screws. An end view of the settler showing the Gore-tex seal and the method of assembly of the mixer-settler is shown in Figure 7. Dimensions of the mixer-settler are given in Figure 8. The mixing chamber was 30.5 cm square and 41.2 cm deep to the overflow. The square cross-section eliminated the need for vertical baffles. Figure 7. Photograph of mixer-settler Gore-tex seal. 24.8 Glass PLAN 316 S.S. 00 CO co b CM j z k - A q . outlet "3- Org. outlet V3.8<*> csl u T w CM 124 Drain Org. outlet 3-Aq. outlet ± Drain ELEVATION F i g u r e 8. Dimensions of experimental mixer-settler (cm). 49 Agitation was provided by a 15.2 cm diameter open-style flat blade impellor having the proportions shown in Figure 9. The impellor axis was centrally located in the mixer-settler 15 cm from the mixer bottom. The vertical position of the impellor could be altered, if desired, by changing impellor shafts. The shaft was driven by a 190 W electric motor (GK Heller, model 25P4225). The speed could be varied between 0 and 360 rpm to an accuracy of ±1% by means of a motor controller equipped with a di g i t a l rpm display (GK Heller, model HS30-MD3). A 16.5 cm diameter disc was fitted onto the impellor shaft 16.5 cm above the impellor to prevent vortex formation. The aqueous and organic mixer inlet ports were constructed of 2.54 cm stainless steel pipe and arranged diagonally opposite each other in the mixer at equal distances from the impellor shaft. The dispersion exited the mixer via a f u l l width overflow weir to fa c i l i t a t e even distribution into the settler. A photograph of the mixing chamber is shown in Figure 10. The relatively large settler size was chosen to minimize wall effects yet s t i l l be practical for laboratory work. The length and therefore the effective cross-sectional area of the settler could be changed by moving an adjustable partition. The volume downstream of this partition was a "dead space" containing the two separated phases. This space could serve to allow entrainment to separate. The partition was sized so as to minimize the liquid depths at the partition underflow and 50 1 15 -*j .2 cm ELEVATION PLAN Figure 9. Impellor geometry. Figure 10. Photograph of mixing chamber. 52 overflow to eliminate backflow from the dead space to the test section. Experiments testing the eff e c t of s e t t l e r i n l e t location r e l a t i v e to the dispersion band could be conducted by adjusting a removable entrance b a f f l e just below the mixer overflow. I n s t a l l a t i o n of this b a f f l e created a passageway for the flow of the dispersion to the desired introduction l e v e l . The phases exited the s e t t l e r through 1 1/4 inch outlet ports. This diameter ensured that a l l flows up to the maximum could be e a s i l y accepted. F u l l width overflow and underflow weirs at the exit ports ensured that any flow disturbances r e s u l t i n g from the r e l a t i v e l y high flow v e l o c i t i e s in the withdrawal region were not conducted through the s e t t l e r . The s e t t l e r outlet nipples were connected by short lengths of f l e x i b l e t eflon tubing to the remainder of the apparatus. This tubing section served 2 purposes: (1) Tubing eliminates the need for a f i t t i n g requiring wrench tightening, the action of which might result in the creation of stresses in the mixer-settler assembly. This could lead to s e t t l e r leakage or glass fracture. A clamp was used to secure the t e f l o n tubing to the outlet nipples. (2) The tubing minimizes the transmission of vibrations (which could a f f e c t coalescence) from the pumps to the mixer-settler vi a the r i g i d piping network. 53 Both the mixing and the s e t t l i n g compartments contained drains so that the mixer-settler contents could be emptied into the storage vessels. 5 . 3 I N T E R P H A S E R E G U L A T O R The l e v e l of the nominal organic/aqueous interface in the s e t t l e r (that which would be present i f there were no dispersion band) could be precisely controlled by an interphase regulator (Figure 11). This device consisted of an inner tube of 316 stai n l e s s steel sheathed by a larger outside tube. A t e f l o n cylinder f i t t e d between the two tubes prevented leakage between them yet permitted the inside tube to be e a s i l y raised or lowered in a telescopic fashion by means of a rod which extended out of the top of the interphase regulator. Aqueous phase from the mixer-settler entered the outer casing of the interphase regulator and flowed over the l i p of the telescopic section which acted as an adjustable weir. Closing or extending this telescopic section altered the hydraulic balance in the s e t t l e r and resulted in r a i s i n g or lowering of the nominal s e t t l e r interface. 5 .4 S T O R A G E V E S S E L S Two 200 l i t r e s tainless s t e e l vessels were used, one for storage of each of the phases. These vessels were large enough t ' Z Z Z Z Z Z Z Z ^ Aqueous inlet Adjusting handle Outer casing (14.6cm O.D. x 46cm) Inner (Adjustable) tube (5.7cm O.D. x 25cm) Teflon cylinder seal Outlet Figure 11. Interphase regulator. 55 to hold the entire contents of the mixer-settler and to ensure that a sufficient liquid level was maintained above the pump suction during mixer-settler operation. When the equipment was not in use, the mixer-settler contents were drained into these vessels for reasons of safety. 5.5 PUMPS The phases were recirculated by means of 2 - 760 W centrifugal pumps (Hayward Gordon model 62RSIC-3.5) capable of providing 15 m of head (water, 20°C) at a flow capacity of 38 1/min. Construction materials were 316 stainless steel (casing and shaft) with teflon gaskets. 5.6 TEMPERATURE REGULATION The temperature of the circulating liquids could be controlled at any desired temperature in the range 20-45°C by passing each phase through a 1.9 cm stainless steel c o i l immersed in a 75 l i t r e capacity constant temperature water bath. Separate baths were provided for each stream to allow for changes in the O/A flow ratio. Each bath was fitted with a 4.5 kW, copper sheathed, screw plug type immersion heater (Chromalox, model MT-345). The heaters could be independently cycled on and off by a temperature controller (Barber-Coleman, model 121M) thereby maintaining the bath water and hence the phases at the desired 56 s e t p o i n t . The i n p u t s f o r the c o n t r o l l e r s were c h r o m e l / c o n s t a n t i n t h e r m o c o u p l e s , chosen because of t h e i r h i g h m i l l i v o l t o u t p u t i n the 20-45°C o p e r a t i n g range. These were l o c a t e d i n the aqueous and o r g a n i c r e c y c l e l i n e s upstream of the mixer i n l e t s . A f l o w of c o l d t a p water c o u l d be f e d t o the baths i f c o o l i n g r a t h e r than h e a t i n g was r e q u i r e d . The b a t h s were p r o v i d e d w i t h a 95 W, 1800 rpm c o n s t a n t speed s t i r r e r ( P r e s t o l i t e - L e l a n d , model 913507-01) f o r improvement of heat t r a n s f e r t o the c o i l s . The immersed a r e a of each c o i l was 0.25 m2. Thermometers l o c a t e d i n the m i x e r - s e t t l e r i t s e l f c o u l d be used t o see i f s i g n i f i c a n t t e m p e r a t u r e d i f f e r e n c e s e x i s t e d between the mixer i n l e t and s e t t l e r o u t l e t . 5.7 FLOW METERING AND REGULATION The f l o w r a t e s of the i n d i v i d u a l phases c o u l d be measured i n the approximate range 4-24 1/min by means of 2 g l a s s tube r o t a m e t e r s ( B r o o k s , model 1307). The r o t a m e t e r s were e x p e r i m e n t a l l y c a l i b r a t e d f o r each phase a t s e v e r a l t e m p e r a t u r e s i n the range 20-45°C and had a c a l i b r a t e d a c c u r a c y of ±1%. Globe v a l v e s upstream of the r o t a m e t e r s were employed f o r r e g u l a t i o n of the O/A r a t i o and the r e s p e c t i v e f l o w r a t e s . 57 5.8 PANEL BOARD A l l c o n t r o l l e r s and motor switches were mounted in a steel panel enclosure situated adjacent to the apparatus. The panel board display i s shown in Figure 12. 5.9 SAFETY CONSIDERATIONS Due to the potential hazard associated with flammable organic phases, a l l motors were explosion proof. Fumes were exhausted through a duct positioned over the mixer-settler. The o i l - t i g h t panel enclosure containing e l e c t r i c a l devices was maintained at p o s i t i v e pressure by connection to an a i r l i n e . The KLMC-supplied samples contained 7.5 g/1 U 30 B in the organic phase and trace quantities of R a 2 2 6 in the aqueous phase. The extremely low l e v e l of radiation associated with t h i s small quantity of radioactive material did not constitute a s i g n i f i c a n t radiation hazard or require protective shielding. Nevertheless, the laboratory was regularly tested for radiation with a geiger counter. The equipment was decontaminated upon completion of experiments. The mixer-settler was positioned over a s p i l l tray, of s u f f i c i e n t capacity to contain the entire mixer-settler contents, in case of leakage or glass fracture. 58 Figure 12. Photograph of panel board. 59 CHAPTER 6 - EXPERIMENTAL PROCEDURE 6.1 PROPERTIES OF THE PHASES The measured physical properties of the phases are tabulated in Table 5. V i s c o s i t y was measured using the c a p i l l a r y tube method with a Cannon-Fenske viscometer (procedure described in ASTM D 445-79) and i n t e r f a c i a l tension was measured using the ring method with a Du Nuoy tensiometer (procedure described in Central S c i e n t i f i c Company, B u l l e t i n 101). A s p e c i f i c gravity bottle was used to determine the density. 6.2 MIXER-SETTLER START-UP PROCEDURE With the HSLIX64N-copper system, the bulk l i q u i d phase which i n i t i a l l y covers the impellor at the time at which agitation i s i n i t i a t e d almost invariably becomes the continuous phase. Therefore, i t was necessary to adopt an appropriate mixer start-up procedure to produce a dispersion of the desired phase continuity. The procedure for mixer-settler start-up with the HSLIX64N phase system was as follows: Assume that the mixer-settler i s empty and that a l l motors are o f f . 1. The s e t t l e r p a r t i t i o n was positioned as required to give the desired s e t t l e r cross-sectional area and the i n l e t b a f f l e was positioned as desired. 2. With the impellor turned o f f , the 2 phases were pumped from the storage tanks into the mixer-settler u n t i l the s e t t l e r was f i l l e d to the l e v e l of the 60 Table 5. Physical properties of the experimental phases. PHASE PHASE TEMP. DENSITY VISCOSITY INTERFACIAL SYSTEM (8C) (g/cm3) (cp) TENSION (dyne/cm2) HSLIX64N-copper org. 24 0.824 2.356 -HSLIX64N-copper aq. 24 1 .002 0.917 -HSLIX64N-copper org. 35 0.816 1 .872 -HSLIX64N-copper aq. 35 0.995 0.726 -HSLIX64N-copper org-aq. 24 - - 14.2 Key Lake org. 35 0.818 2. 150 -Key Lake aq. 35 1 .034 0.836 -Key Lake org-aq 35 12.6 61 overflow weir with approximately equal volumes of organic and aqueous phases. The interface l e v e l in the mixer was adjusted so that the phase to be continuous covered the impellor by a minimum of 5 cm. This was done by adding (by pumping) and removing (by draining) the necessary quantities of each phase from the mixer. The impellor was turned on and set to the desired speed. A conductivity probe could then be used to confirm that the proper phase continuity had been achieved. The pumps were' started and the throughput of each phase, measured by means of ca l i b r a t e d rotameters, was adjusted to give the desired phase r a t i o and to t a l throughput. The agitators in the constant temperature baths were started and the temperature c o n t r o l l e r s were set to the desired set point. The two phases were now being recycled through the system. A dispersion was being formed in the mixer and separated in the s e t t l e r . The l i q u i d s were ci r c u l a t e d u n t i l the desired temperature was reached (as measured by a thermometer immersed in the mixer). The interphase regulator was positioned to maintain the nominal interface in the s e t t l e r at the proper l e v e l . It was necessary to monitor a l l flows and levels p e r i o d i c a l l y and to make continued adjustments. 62 8. No data was recorded u n t i l i t had been established that steady-state had been reached. With the Key Lake phase system i t was necessary to adopt a s l i g h t l y modified start-up procedure. Since this system exhibited a tendency for forming a stable, aqueous continuous dispersion (section 7.1.3), i t was necessary to start up the mixer-settler with no aqueous phase in the mixer to avoid t h i s . The aqueous phase flow to the mixer was then slowly increased to the desired rate. A photograph of the mixer-settler operating with the Key Lake phase system i s shown in Figure 13. 6 . 3 TIME FOR DISPERSION BAND TO REACH STEADY STATE The time required for the dispersion band thickness to achieve steady state i s presented in Figure 14 for various operating conditions for the HSLIX64N-copper phase system. The data shows that at r e l a t i v e l y high flow rates, when a thick dispersion band i s produced, 45 minutes may be required while at low flow rates (thin dispersion bands) as l i t t l e as 10 minutes would be adequate. Based on t h i s , when conducting experiments, at least 45 minutes were allowed after an operating condition adjustment had been made before a reading was taken. In addition, to further guarantee that steady-state had been achieved, data was only recorded after two successive dispersion band thickness readings, taken 5-10 minutes apart, agreed within 5%. The Figure 13. Photograph of mixer-settler operation with Key Lake phase system. 64 40 Legend A 0/A-.67, Od/A-2.43, A- .22 , ORG.C. X O / A - 6 7 . Qd/A-1.91. A- .22 . ORG.C. • O/A-1.5. Od/A-2.9 . A- .22 . ORG.C. O/A-1.5. Q d / A - 7 7 . A - 3 6 , ORG.C. 0/A-.67. Qd/A-2.44. A-.13. AQ.C. X O/A-1.6. Qd/A-3.22. A- .22 . AO.C. 40 60 t i m e ( m i n u t e s ) 80 100 Figure 14. Time for dispersion band thickness to achieve steady-state after introduction of dispersion into settler (HSLIX64N-copper phase system). 65 throughput was then changed for a subsequent run. It was realized that, although the dispersion band thickness may apparently have reached steady-state, the holdup profile and drop size distribution within the dispersion band might continue to change slightly for a period of time beyond this. Therefore, at least 90 minutes were allotted before holdup or drop size data were taken. 6.4 S A M P L I N G F O R D I S P E R S I O N B A N D H O L D U P Dispersion samples were withdrawn from the mixer-settler using a 100 ml pipet, connected via a stopcock and a length of tygon tubing to the suction side of an ejector. The pipet, inverted so that the larger end (0.5 mm ID) served as the tip to more readily admit the sample, could be supported in the mixer-settler by means of a clamp and stand. When the tip of the pipet was immersed at the desired level in the dispersion band and the stopcock was opened, the vacuum quickly drew a sample into the pipet. Once 100 ml had been withdrawn, the stopcock was closed and the sample was transferred to a graduated cylinder in which the phases were allowed to disengage so that their volumes could be determined. A minimum of 90 minutes was allotted to achieve steady state before dispersion samples were withdrawn. Sampling at various vertical locations within the dispersion band was performed successively on a random basis. Data was obtained only for large dispersion band thicknesses to minimize the error associated with positioning of the pipet tip at the 66 desired level. In order to test the accuracy of this technique, samples were taken from the mixer, which contained a known ratio of organic and aqueous phases. The sampled O/A ratio agreed within 2% in a l l cases, indicating that neither phase was preferentially withdrawn. 6 . 5 DETERMINATION OF DISPERSION DROP SIZE In order to obtain quantitative information regarding the dispersion drop size distribution, an attempt was made to photograph the dispersion directly through the glass walls of the mixer-settler using a tripod mounted SLR camera equipped with a macro-zoom lens, and from the photographs to determine the average drop size. This method failed because the magnification was insufficient and also because excessive light reflection off the glass walls rendered the photos indistinct. In order to circumvent these problems, i t was necessary to withdraw a sample of dispersion into a small, specially constructed, glass-walled c e l l mounted on the stage of a 40X light microscope, and to photograph the sample through the microscope's ocular system. The c e l l into which the samples were charged is illustrated in Figure 15. Using 6 bolts, two circular glass windows (1/8" thick) could be clamped to a 3/8" thick brass cylinder to form a 4" diameter X 3/8" cavity. Strips of Gore-Tex joint sealant ensured that the cavity was leakproof. 67 Glass window xJSjj. ntn oia j l v ' / / / //////////////////'/'////// / Vi / ' / / / > s s\ HA — ( L ^ - - Brass cylinder W (0.9cm x 10cm I.D.) Cell cavity OOcmjrf x 1.1cm) SECTIONAL VIEW ^ ° e i e c t o r suction Approx. scale FRONT VIEW Figure 15. Design of glass-walled c e l l used determination of dispersion drop size. for 68 Holes d r i l l e d through the cylindrical brass section at two diametrically opposite points were fitted with nipples which were connected to lengths of tygon tubing. One length of tubing was connected to the suction side of an ejector so that a vacuum could be applied to the c e l l when a stopcock on the line was opened, thereby drawing a sample of dispersion from the mixer-settler into the c e l l via the second length of tubing. The c e l l was fixed to the microscope stage such that the sample inlet connection was at the bottom and so that light from the microscope lamps could be admitted through one glass window while photographing the c e l l interior through the other. The tygon sampling line was connected to a smooth-tipped glass sampling tube which could be immersed to withdraw dispersi-on samples from any desired location within the mixer-settler. This probe and the connecting tygon tubing had a relatively large inside diameter of over 0.7cm (approximately 10-30 drop diameters) to minimize redispersion of the sample in the line. No coalescence was observed in this transparent line during sampling. The microscope's f i e l d of view at 40X magnification was approximately 2500 urn. Because of the apparently rapid movement of individual droplets in the sample when viewed at this magnification, a shutter speed of 1/1000 sec. at f 2.8 was utiliz e d as well as a fast film (Fujichrome P1600D slide film) pushprocessed to 1600 ASA. The microscope lamp was adjusted to its maximum intensity. These conditions resulted in sharp 69 images being produced of Key Lake phase system dispersions. However, because of the opacity of the HSLIX64N organic phase, light penetration through samples of this phase system was insufficient to permit photography. Therefore, only the Key Lake phase system (organic continuous) was analyzed for drop size distribution using the microscope. Slides of the dispersion samples were projected onto a digitizer board logging to a computer and 3 points on the circumference of each projected drop were touched with the digitizer probe. A computer program stored the coordinate values, calculated the diameter of each drop and s t a t i s t i c a l l y analyzed the data to determine the mean drop diameter (both arithmetic average and volume average) for each sample (slide). 6.6 ORIENTATION OF MICROSCOPE When the organic continuous Key Lake sample was charged into the c e l l , batch separation commenced and the coalescence front developed at the bottom of the c e l l . With the c e l l positioned on the stage of the microscope and with the microscope maintained in an upright position, photographs of the sample would be of the coalescence front because of the camera's location below the c e l l . Observation of the sample in the c e l l with the unaided eye revealed that, as a result of the coalescence process, droplets at the coalescence front were considerably larger on average than those in the remainder of the sample. This indicates that photographs taken with this setup were not representative of the original sample. 70 In order to avoid the problem of obtaining photographs of the coalescence front, the microscope was tipped on its side (it had been modified to permit this) and the c e l l was carefully positioned on the microscope stage so that the coalescence front would not ( i n i t i a l l y ) occupy the f i e l d of view of the microscope (Figure 16). About 10 seconds normally elapsed between the time sample withdrawal was initiated and the time at which the c e l l was f u l l and the f i r s t photograph could be taken. ? 1 Figure 16 Photograph of arrangement of microscope and c e l l . 72 C H A P T E R 7 - R E S U L T S A N D D I S C U S S I O N 7 .1 G E N E R A L O B S E R V A T I O N S O F M I X E R - S E T T L E R O P E R A T I O N 7 . 1 . 1 D I S P E R S I O N B A N D A P P E A R A N C E For the HSLIX64N-copper system at pH 1.9, the loaded organic phase was black in color and nearly opaque; light penetration was only a few millimeters. This contrasted with the clear gold appearance of the organic phase prior to being loaded with copper. The aqueous phase was blue-green in color and clear when free of organic entrainment. Because of the opacity of the organic phase i t was necessary to u t i l i z e an intense (500 W) spotlamp in order to observe the dispersion band and to identify the location of the interface between the dispersion band and the bulk organic phase. Unlike the HSLIX64N-copper system, both phases in the Key Lake system were clear. The loaded organic phase was golden yellow in color and the aqueous phase was blue-green. The dispersion band interfaces were clearly visible to the unaided eye. The cla r i t y of the Key Lake phase system facilitated photographic analysis of the dispersion. The HSLIX64N - copper system, however, did not lend i t s e l f to this because of the opacity of the organic phase. 73 For both systems, deep layer dispersion bands existed for a l l experimental runs. The dispersion band thickness varied between 1 cm and 30 cm and was essentially uniform in thickness between the front and back of the settler. Only when the throughputs were reduced almost to zero was i t possible to obtain a wedge-shaped band; in general the wedge was very thin and long, tapering slowly from a thickness of less than 1 cm at the settler inlet. Although the coalescence front was sharp in a l l cases, the settling front sharpness depended on the O/A ratio of the dispersion and on the specific settler flow. For concentrated dispersions the settling front was easily distinguishable, while for "dilute" dispersions (i.e.: O/A = 3, organic continuous) the settling front became indistinct. Increased throughputs (larger AH's) tended to further obscure the settling front. It was found that, if no inlet baffle was utilized, the dispersion band would oscillate in a wavelike fashion, especially with thicker dispersion bands. Operation with the inlet baffle essentially eliminated this. The use of the baffle also reduced entrainment, as could be observed by the degree of haziness of the clear aqueous phase during runs. Close observation of the. coalescence front indicated that a significant source of entrainment was the ejection of very small continuous phase droplets, possibly produced during the coalescence of two larger droplets at the coalescence front, 74 into the separated phase. This Davies (1970) and was thought forces. had previously been observed by to be a result of hydrodynamic It was generally possible, to a limited extent, to v i s u a l l y i d e n t i f y the two sublayers described by Barnea and Mizrahi. In the "dense sublayer" near the coalescence front, the droplets were cl o s e l y packed regardless of the concentration of the dispersion entering the s e t t l e r . The even concentration sublayer consisted of more loosely packed droplets. Although the o v e r a l l movement of these drops was towards the end of the s e t t l e r and towards the coalescence front, individual drops, of a l l sizes, appeared to move somewhat randomly. The d i s t i n c t i o n between the two sublayers was not sharp; i. e . , i t appeared, v i s u a l l y at le a s t , to be a somewhat gradual t r a n s i t i o n . Although there i s no sharp boundary, the concept of two d i s t i n c t sublayers is a useful one since rea d i l y observable structural differences do exist between the dispersion near the coalescence front and the dispersion more distant from the coalescence front. Visual observation indicated that various flow patterns existed within the dispersion band. Besides the primary horizontal flow from s e t t l e r i n l e t towards the end of the s e t t l e r , sometimes there was also a backflow above or below the primary horizontal flow. 75 For other cases, i t was observed that large v e r t i c a l c i r c u l a t i o n patterns developed, whereby the droplets moved in one di r e c t i o n along the glass walls and the other d i r e c t i o n in the i n t e r i o r of the s e t t l e r . For example, with an organic continuous dispersion, where the coalescence front was the lower surface of the dispersion band, a r e l a t i v e l y rapid and unexpected movement of droplets upwards along the glass walls was sometimes observed. Such c i r c u l a t i n g flow patterns were usually, but not always, associated with high throughputs, producing thick dispersion bands. Generally, thinner dispersion bands (less than 10-15 cm) did not exhibit such c i r c u l a t i o n patterns. Despite t h i s generalization, i t was not possible to develop any r e l i a b l e c r i t e r i a by which the flow pattern in the s e t t l e r could be predicted. The s e t t l e r frequently operated in an unpredictable manner with respect to flow patterns. A p r a c t i c a l conclusion which may be drawn from these observations i s that the flow pattern existing in the s e t t l e r may wield a considerable e f f e c t on the thickness of the dispersion band by upsetting i t s normal structure. Since c i r c u l a t i o n patterns usually are associated with thicker dispersion bands (the i n d u s t r i a l l y s i g n i f i c a n t ones) flow patterns could conceivably have a profound effect on the operation and performance of i n d u s t r i a l s e t t l e r s . 76 Close observation of the coalescing behaviour of the dispersion band seemed to indicate, as contended by Ryon (1959), that the majority of the coalescence did indeed occur at the coalescence front, or at least within a few centimeters of the coalescence front. Observation of the remainder of the dispersion band revealed that the droplets moved randomly and collided frequently, resulting in some coalescence, but that the amount of coalescence occurring here was small in comparison. This does not necessarily contradict the view of Barnea and Mizrahi that the coalescence in the even concentration sublayer is the limiting mechanism determining the dispersion band thickness since they contended only that its role was to coalesce droplets to the c r i t i c a l size permitting flow countercurrent to the draining continuous phase. This does not necessarily suggest that the number of coalescences or the volume of dispersed phase coalesced is greater in the even concentration sublayer. 7.1.2 CRUD For both phase systems, over an extended period of time, crud would accumulate at the dispersion band interfaces, primarily at the back of the settler. Crud was silky and fibrous in appearance and possessed the color of the organic phase. It hindered coalescence and also restricted visual observation of the interior of the settler by deposition on the glass walls. 77 Crud was periodically removed (every 4-6 hours) from the operating mixer-settler with a strainer before accumulating to a level which appreciably hindered coalescence. This was done so that various experimental runs were not conducted with significantly differing degrees of crud buildup. The mixer-settler was also occasionally drained and cleaned to remove crud adhering to the glass walls. Crud formation is a complex and poorly understood phenomenon, and the reader is referred to other authors for a detailed discussion of its causes (Ritcey, 1980). In general, crud can be loosely defined as the material resulting from the agitation of an organic phase, an aqueous phase and fine solid particles, commonly siliceous in nature. Living organisms existing in the aqueous phase, which die upon entering the organic phase have also been cited as a possible cause ( H i l l , Evans, Davies, 1967). Crud formation appeared to be minimized by operation at higher temperatures, possibly because of greater solubility of crud-causing constituents. Also, the rate of crud accumulation with time decreased slightly as the phase liquids aged. This may simply be a result of the cumulative reduction in crud-causing constituents arising from the periodic removal of crud with the strainer. 78 7.1.3 OPERATION AT DESIRED PHASE CONTINUITY With the HSLIX64N - copper system i t was possible to operate either organic or aqueous continuous by adopting an appropriate start-up procedure, as described in section 6.1. For either continuity, the dispersion was stable ( i . e . , phase inversion did not occur) provided that the O/A ra t i o was maintained in the range corresponding to dispersed phase concentrations of 0% to 80% . Below 20%, however, no dispersion band was formed in the s e t t l e r . Instead the dispersion entering the s e t t l e r was "hazy" and limited primarily by gra v i t a t i o n a l s e t t l i n g rather than coalescence. With the Key Lake system however, only organic continuous operation was possible since when aqueous continuous operation was attempted, a stable emulsion was formed in the mixer-settler which took many hours to several days to separate batchwise. Mechanical agitation reduced this time to perhaps 1/2 hour. This phenomenon occurs in i n d u s t r i a l i n s t a l l a t i o n s , including the Key Lake uranium extraction m i l l . Furthermore, when running organic continuous with the Key Lake system i t was not possible to operate with a dispersed phase concentration exceeding 65% without the occurrence of phase inversion and production of a stable emulsion. In other words, the ambivalence region for the Key Lake system (organic continuous) was s i g n i f i c a n t l y narrower than that for the HSLIX64N-copper system. 79 In order to ensure that a stable emulsion did not form, i t was necessary to take special precautions during startup. The impellor was started with the mixer containing only the organic phase. Thereafter, the organic and aqueous phase flowrates could be adjusted as required. It was not s u f f i c i e n t to simply start up the mixer with the impellor immersed in the phase to be continuous. If a s i g n i f i c a n t volume of aqueous phase was present in the mixer at startup, then some stable emulsion was formed even i f the dispersion eventually became organic continuous. During organic continuous operation with both the HSLIX64N-copper phase system and the Key Lake phase system, i t was sometimes observed that r e l a t i v e l y large aqueous continuous slugs of dispersion (2-5 cm in diameter) entered the s e t t l e r along with the organic continuous dispersion. It was estimated that the volumetric flowrate of the slugs constituted roughly 5-10% of the t o t a l dispersion flow to the s e t t l e r . With the HSLIX64N-copper phase system, th i s "slug" phenomenon was also sometimes observed during aqueous continuous operation. Slugs of organic continuous dispersion would issue from the mixer along with the majority aqueous continuous phase. The appearance of slugs was linked to two operating conditions:. 1. Inadequate mixing intensity in the mixer. Use of an impellor speed below 200-250 rpm, approximately, increased the tendency for slug formation. 80 2. Operation with a high dispersed phase concentration (greater than roughly 60% with the Key Lake system and 75% with the HSLIX64N-copper phase system). This may be attributed to the higher e f f e c t i v e v i s c o s i t y of a concentrated dispersion compared to a d i l u t e dispersion. In order to adequately disperse the phases, a higher impellor speed would be required with a more concentrated dispersion. Slug formation frequently preceded t o t a l phase inversion which could be triggered i f the dispersed phase concentration was further increased or the impellor speed was decreased. The range of O/A ratios and impellor speeds used during experiments was such that slug formation was avoided. No data was taken under conditions for which the phenomenon was observed. 7.1.4 DRIFT IN PHYSICAL BEHAVIOR OF PHASES It was observed for the HSLIX64N-copper phase system that the dispersion band thickness obtained for i d e n t i c a l organic continuous runs, performed roughly 2 weeks and 3 months after the newly-prepared phases had been introduced into the mixer-settler, d i f f e r e d s u b s t a n t i a l l y . Invariably, the value of AH obtained at the later date was greater (roughly 10-20%). A s h i f t for aqueous continuous operation was much harder to detect (AH increased perhaps 5-10% for a given run). The s h i f t may be attributed to a change in the physical properties of the phases as they aged, possibly due to the evaporation of the 81 more v o l a t i l e components of the organic phase. A r e s u l t i n g increase in organic phase v i s c o s i t y with time would account for the increased values of AH obtained at later dates. Since the v i s c o s i t y of the aqueous phase would not be expected to change appreciably with time, t h i s could explain why no large s h i f t could be distinguished for aqueous continuous operation. To minimize the effect of t h i s s h i f t and to ensure that v a l i d comparison of d i f f e r e n t experimental runs could be ca r r i e d out, a l l data reported for the HSLIX64N-copper system was taken within an 8 week period (starting about 2-3 months after the phases had been introduced into the equipment). With the Key Lake system no s h i f t in phase separation c h a r a c t e r i s t i c s with ageing was observed. However, since experimentation with t h i s phase system was conducted after experience in operating the equipment had been obtained with the HSLIX64N-copper system, a r e l a t i v e l y short (4-6 week) period was s u f f i c i e n t to complete Qd/A-AH experiments. 7.1.5 EFFECT OF INTRODUCTION LEVEL ON AH The entrance to the s e t t l e r was f i t t e d with an entrance b a f f l e which served to minimize turbulence and ensure even dispersion d i s t r i b u t i o n into the s e t t l e r . The b a f f l e was secured to the glass walls of the s e t t l e r by clamps and could be positioned to form a 5 cm X 30 cm passageway for the flow of the dispersion to the desired introduction l e v e l . By varying the v e r t i c a l position of the b a f f l e and the setting of the 82 interphase regulator, the eff e c t of the l e v e l of introduction r e l a t i v e to the dispersion band interfaces was determined. The results for the HSLIX64N-copper system are tabulated in Table 6 for various operating conditions (for both organic and aqueous phase con t i n u i t y ) . Analysis of t h i s data does not reveal an appreciable e f f e c t of introduction l e v e l on s e t t l e r performance. For the Key Lake system, however, the le v e l of introduction was found to have a major impact on dispersion band thickness. Examination of Table 7 reveals that in general, for a given throughput, introduction of the dispersion at or near the coalescence front resulted in the poorest s e t t l e r performance (largest AH). This i s i l l u s t r a t e d graphically in Figure 17. In some cases (cases 1,2,3,4,9) the thickness of the dispersion band was as much as 50% greater when the dispersion was introduced at the coalescence front rather than at the s e t t l i n g front or i n t e r i o r of the dispersion band. For other cases, the magnitude of the ef f e c t was much less (cases 5,11,12,13,14,15). Close scrutiny of Table 7 and Figure 17 shows that the primary parameter influencing the magnitude of thi s e f f e c t appears to be the s e t t l e r area. From the table, i t can be seen that for the largest s e t t l e r area 0.36m2, the introduction l e v e l did not have a major ef f e c t on dispersion band thickness since AH remained constant regardless of the i n l e t location. This contrasts sharply with the data taken for the smallest 83 TABLE G. EFFECT OF DISPERSION INTRODUCTION LEVEL ON DISPERSION BAND THICKNESS (SYSTEM : HSLIX64N-•copper; TEMPERATURE : 24 "C) CASE SETTLER SPECIFIC O/A ' INTRODUCTION A H AREA SETTLER LEVEL (cm) (m2) FLOW Qd/A (m 3 /mVhr) (cm from coalescence f ront ) A) ORGANIC CONTINUOUS: 1 . 0. 13 2.76 1 .5 3. 14. 23. 27 .5 27 . 27.5 2. 0. 13 1.17 1 .5 0. 3 . 6.5 7. 6.5 6.5 3. 0. 13 2 . 16 3.0 2. 5.5 15. 19.5 19.5 20. 4 . 0. 22 1 .97 0.67 2. 6.5 16 . 18 . 5 18. 18.5 5. 0. 22 1 .80 1 .5 0. 5. 10. 5 1 1 . 10. 5 10.5 6 . 0. 36 2.15 0.67 2 . 5.5 12. 13.5 13.3 13.5 7 . 0.36 2.22 1 .5 3. 1 1 . 18. 22. 21 . 21.5 8 . 0. 36 1 .33 3.0 0. 3. 6 . 6.2 6.3 6.4 B) AQUEOUS CONTINUOUS: 1 . 0. 13 2.77 0.33 0. 6.5 12. 12 . 12 . 12. 2 . 0.13 3.28 0.67 1 .5 7 . 14 . 15.5 15. 15. 3 . 0. 22 1 .80 0. 33 0. 3. 5. 6 . 6. 5.5 4 . 0. 22 3.27 1 . 5 2. 5.5 18. 20.5 19.5 19.5 5. 0.36 2.71 1 . 5 1 . 6 . 13 . 12.5 12.6 13. 6 . 0.36 2. 18 1 .5 0. 5. 8. 8 . 5 8.5 8. 84 TABLE 7. EFFECT OF DISPERSION INTRODUCTION LEVEL ON DISPERSION BAND THICKNESS (SYSTEM : KEY LAKE; TEMPERATURE: 35 "O. CASE SETTLER SPECIFIC O/A INTRODUCTION A H AREA SETTLER LEVEL (cm) (m2) FLOW,Qd/A (m 3 /m 2 /hr ) (cm from coalescence f ront ) 1 . 0.13 2.95 1 .5 3. 17.5 24. 34 . 27 . 27. 2. 0. 13 2.67 1 .5 6. 19.5 24. 19.5 3. 0.13 2.53 1 .5 2. 7 . 15 . 22. 14.5 15. 4 . 0.13 2.34 1 .5 0. 4.5 11.5 16.5 12.5 11.5 5. 0.13 2.11 1 .5 0. 5 . 10. 10. 5 9.5 10. 6 . 0.13 2. 15 3.0 1 6.5 15.5 18 . 5 16. 15.5 7. 0.22 2. 13 0.67 0. 8 . 8.5 8. 8 . 0. 22 2.25 1 .5 0. 5 . 5 10.5 14.5 1 1 . 10. 5 9. 0.22 2. 18 1 .5 0. 6 . 10.5 18.5 1 1 . 10.5 10. 0. 22 2 . 18 3.0 0. 10.5 20. 27 . 21 . 20. 1 1 . 0. 36 2.83 0.67 0. 9. 18. 18 : 18 . 18 . 12. 0.36 2.33 0.67 0. 7.5 14.5 14 . 15. 14.5 13. 0. 36 2. 17 0.67 0. 5 . 10.5 • 10.5 9.5 10. 5 14 . 0. 36 1 .30 0.67 0. 3. 3 . 3. 15. 0.36 2.45 1.5 0. 8. 16. 20. 16. 16. 85 Legend • C a s e 1, A—. 1 3 m 2 • C a s e 2 . A— . 1 3 m 2 A C a s e 3 , A— 1 3 m 2 T C a s e 4 . A- . 1 3 m 2 • c a s e 5 , A— 1 3 m 2 C a s e 6 , A"* 1 3 m 2 O C a s e 7 , A™ . 2 2 m 2 • C a s e 8 , A" 2 2 m 2 o C a s e 9. A" . 2 2 m 2 © C a s e 1 0 . A - . 2 2 m 2 + C a s e 1 5 . A™.38m2 Note: cases 11,12,13,14 (for area=0.36m ) not shown since deviation from minimum AH<5% 0.4 0.5 0.6 h/AH Figure 17. Effect of dispersion introduction level for Key Lake phase system for cases identified in Table 7 . 86 s e t t l e r area, 0.13m2, where the effect is very considerable. For these cases, (cases 1-6) the dispersion band thickness may vary by as much as 50% from the optimum depending on the introduction l e v e l . Although the data indicates that locating the introduction l e v e l at the coalescence front should be avoided, there does not appear to be any s i g n i f i c a n t advantage in locating the introduction l e v e l at the other extreme, the s e t t l i n g front rather than the dispersion band i n t e r i o r . For example, in case 3, locating the introduction l e v e l at the s e t t l i n g front results in a 15cm thickness as opposed to 14.5cm when the dispersion introduction l e v e l i s 7cm from the coalescence front. (The thickness when introduced at the coalescence front i s 22cm.) The same observation can be made for cases 1, 6, 10 and 11. A r e l a t i v e l y simple explanation exists for the observed trends i f one considers the structure of the dispersion band in terms of the dense and even concentration sublayer concept. The dense concentration sublayer, i t may be reca l l e d , occupies the region close to the coalescence front and consists of larger and more compacted droplets than those in the even concentration sublayer. F i n a l coalescence occurs here. The even concentration sublayer, occupying the majority of the dispersion band, consists of randomly moving, loosely packed droplets and has a density roughly equal to that of the dispersion issuing from the mixer. When the dispersion was 87 introduced at the l e v e l of the coalescence front, the dispersion was observed to flow towards the even concentration sublayer, as would be expected because of the density of the feed r e l a t i v e to the two sublayers. This flow (observed by Barnea and Mizrahi, 1976) and referred to as the "chimney flow" displaces the dense concentration sublayer mear the s e t t l e r i n l e t , e f f e c t i v e l y reducing the horizontal cross-sectional area of the s e t t l e r (defined as the cross-sectional area of the coalescence front) and results in a thicker dispersion band. If the dispersion i s introduced in the even concentration sublayer, by contrast, the volume of the chimney becomes small and the dense concentration sublayer i s not disturbed. The lev e l of dispersion introduction had no observable effect on flow patterns provided that i t was within the even concentration sublayer. This observation is re f l e c t e d in the curves of F i g . 17; the l e v e l of introduction has no appreciable effect on dispersion band thickness provided that the dispersion is not introduced close to the coalescence front. From t h i s , i t can be concluded that the natural or optimum le v e l of introduction i s anywhere within the even concentration sublayer. It had been noted that for large s e t t l e r areas, the negative impact of introduction at the coalescence front was greatly reduced. This can be attributed to the chimney flow displacing a smaller area (on a percentage basis) of the coalescence front as s e t t l e r size i s increased. This suggests 88 that for large i n d u s t r i a l s e t t l e r s , the introduction l e v e l may not be c r i t i c a l . With the HSLIX64N-copper system, introduction l e v e l had no appreciable e f f e c t on AH, even for small s e t t l e r areas. It i s suggested that this is due to differences in hydrodynamic behavior between phase systems; The chimney flow for the HSLIX64N-copper system does not disturb the dispersion band structure as seriously with t h i s phase system. For runs conducted to determine the AH-Qd/A relationship (section 7.2) for both the Key Lake phase system and the HSLIX64N-copper system, i t was ensured that the dispersion introduction l e v e l was maintained in the i n t e r i o r of the dispersion band, well away from the coalescence front. In t h i s way i t was ensured that the recorded values of AH were the optimum values. Furthermore, t h i s practice ensured that data was taken consistently. Consistency errors which would be introduced by locating the introduction l e v e l at the coalescence front in one case and in the s e t t l e r i n t e r i o r for another case were thereby avoided. 7.2 Qd/A - AH RELATIONSHIP By measuring the dispersion band thickness as a function of the s p e c i f i c s e t t l e r flow i t was possible to determine the v a l i d i t y of the s e t t l e r scale-up c r i t e r i a proposed by Ryon et al.(1959). This is expressed mathematically in equation (1), rewritten below: 89 AH = k(Q d/A) y (1) Equation (1) implies that i f AH i s plotted against Q^on log-log coordinates, a straight l i n e (referred to as the c h a r a c t e r i s t i c curve for the phase system) should be produced. The experimental data has been presented in t h i s conventional graphical fashion for three main cases; 1. Organic continuous operation with the HSLIX64N-copper system (Figure 18). 2. Aqueous continuous operation with the HSLIX64N-copper system (Figure 19). 3. Organic continuous operation with the Key Lake system (Figure 20). Furthermore, for each of the above cases, data was obtained in three d i f f e r e n t sized s e t t l e r s (with horizontal cross-sectional areas of 0.13m2, 0.22m2 and 0.36m2) at three d i f f e r e n t O/A rati o s (corresponding to dispersion concentrations of 25%, 40%, and 60%). The mixer impellor speed was held constant at 300 rpm throughout. 7.2.1 CASE 1- HSLIX64N-COPPER SYSTEM (ORGANIC CONTINUOUS  OPERATION) The data for thi s case is plotted in figure 18 showing the best f i t l i n e , for a l l data points, obtained by least squares analysis to determine the constants k and y. The individual k and y values for each p a r t i c u l a r operating condition ( i . e . : operating at one part i c u l a r phase r a t i o for one s e t t l e r size) are tabulated in Appendix 2 for Figure 18 and a l l subsequent 90 i i O/A=0.67 0/A=1.5 O/A=3.0 S E T T L E R A R E A ( m " 2 ) 0.13 0.22 0.36 O © • • A ffl V O 0.5 Qd/A (m**3/m**2/hr) 10 Figure 18. Dispersion band thickness versus s p e c i f i c s e t t l e r flow of dispersed phase (effect of O/A ra t i o and s e t t l e r area), HSLIX64N-copper system, organic continuous operation, 24°C, 300rpm. 91 Qd/A-AH graphs. Close scrutiny of Figure 18 shows that although there i s some scatter of the points around the best f i t l i n e , the data obtained for t h i s phase system f i t s the basic concept of Ryon et a l . For any value of /A, regardless of the O/A r a t i o , the experimentally determined dispersion band thickness i s within 30-40% of that predicted by equation (1). This indicates that, with the HSLIX64N-copper phase system (organic continuous), the dispersion band thickness depends only on the flow of dispersed phase and not on that of the continuous phase. A single c h a r a c t e r i s t i c curve can be used to adequately f i t the data for a l l three phase r a t i o s . This i s shown s t a t i s t i c a l l y in Appendix 2. This conclusion i s supported by p i l o t plant data for the t a i l i n g s leach plant at Chingola, Zambia, for a similar phase system. This system employed 23 V/V LIX64N (an e a r l i e r and less concentrated form of HSLIX64N) in Escaid 100 as the organic phase. The . aqueous phase consisted of 1 g/1 H 2SO„ and had an o r i g i n a l copper concentration of 3 g/1. This testwork had also concluded that the dispersion band depth is dependent only on the flow of dispersed phase and thus independent of the phase r a t i o . Phase rat i o s of O/A=1.0 and 0/A=1.5 were employed in th i s study. The Chingola p i l o t plant data c h a r a c t e r i s t i c curve, however, d i f f e r e d substantially from the author's data. It could only be described by either a linear r elationship or by an exponential relationship of lower exponent (y less than 2). 92 The effect of the O/A r a t i o i s very important in i n d u s t r i a l operation since the phase r a t i o can be adjusted by intern a l recycling between the mixer and the s e t t l e r . The data of Figure 18 exhibits considerable, apparently random scatter around the b e s t - f i t c h a r a c t e r i s t i c curve. Several sources of experimental error may have contributed to t h i s : 1. After the dispersion band thickness had reached steady state, there was s t i l l some minor o s c i l l a t i o n of the value of AH with time (contributing a maximum estimated error of roughly 5% to the recorded value of AH). 2. There was some . inaccuracy involved in precisely locating the l e v e l of the s e t t l i n g front, especially for low dispersion concentrations. 3. The rotameter readings fluctuated s l i g h t l y necessitating constant monitoring. 4. Unpredictable and varying flow patterns developed for dif f e r e n t runs. This may a l t e r the structure of the dispersion band somewhat and affe c t the dispersion band thickness produced. 5. There was a s i g n i f i c a n t d r i f t of physical behavior with time as the phase system aged, as discussed in section 7.1.4 93 7.2.2 CASE 2- HSLIX64N-COPPER SYSTEM (ORGANIC CONTINUOUS  OPERATION) As was the case for organic continuous operation, with aqueous continuous operation a single c h a r a c t e r i s t i c curve adequately predicts the s e t t l i n g behaviour for d i f f e r e n t phase r a t i o s . (See Appendix 2 for a s t a t i s t i c a l analysis v e r i f y i n g this) This further supports the contention of Ryon that dispersion thickness i s independent of continuous phase flow. Figure 19 plots a l l the experimental data points and shows the best f i t l i n e , obtained by considering a l l data points. Comparison of Figure 18 and Figure 19, shows that, although the curves are nearly p a r a l l e l , for a given value of Qd/A, aqueous continuous operation produces a dispersion band which is roughly 50% narrower than that produced with organic continuous operation. Considered alone, t h i s would seem to suggest that, to maximize s e t t l e r capacity, aqueous continuous operation should be employed. However, v i s u a l observation revealed that a high l e v e l of haze in the aqueous phase was associated with aqueous continuous operation. This haze, resu l t i n g from organic entrainment, would constitute excessive and economically unacceptable loss of organic in a commercial plant, and explains why i n d u s t r i a l i n s t a l l a t i o n s operate organic continuous. 94 E CO CO CD C u 1c -t-> c CO J Q c o CO CD Q . CO Qd/A (m**3/m**2/hr) Figure 19. Dispersion band thickness versus s p e c i f i c s e t t l e r flow of dispersed phase (effect of O/A ra t i o and s e t t l e r area), HSLIX64N-copper system, aqueous continuous operation, 24°C, 300rpm. 95 7.2.3 CASE 3- KEY LAKE SYSTEM (ORGANIC CONTINUOUS OPERATION) Because of the s t a b i l i t y of the emulsion formed during aqueous continuous operation with t h i s phase system, only organic continuous operation could be studied. The re s u l t s , plotted in Figure 20, d i f f e r from those obtained for ( the HSLIX64N-copper system in one very important respect; the dispersion band thickness i s here a function of both organic and aqueous flowrate ( i e : i t i s dependent on the O/A r a t i o ) . It can be seen that a family of p a r a l l e l curves are produced for diff e r e n t phase r a t i o s . This has been confirmed by the s t a t i s t i c a l analysis of Appendix 2. Although the difference in the location of the ch a r a c t e r i s t i c curves for O/A=0.67 and 0/A=1.5 i s barely perceptable, a considerable s h i f t e xists for the O/A=3.0 curve, representing a less concentrated dispersion. For O/A = 3.0, the dispersion band thickness for a given value of Qd/A i s roughly 50% greater than that for O/A = 0.67 and O/A = 1.5. In the uranium industry an O/A r a t i o of roughly 1.0 to 1.5 is generally employed in the extraction c i r c u i t . This, according to Figure 20, cl o s e l y corresponds to the optimum with respect to phase separation. Recycling of the organic phase to raise the O/A phase r a t i o would result in poorer phase separation for t h i s phase system, whereas for the HSLIX64N-copper system, s e t t l e r performance would not be changed s i g n i f i c a n t l y . 96 E CO CO CD c o T J c CO c o CO k_ <D Q . CO T J •A-fflV Estimated Operating Range of Key Lake Settlers (0/A=1-1.5) SETTLER AREA (m •*2) 0.13 0.22 0.36 O/A=0.67 o © • 0/A=1.5 • ffl • 0/A=3 0 A V O Qd/A (m**3/m**2/hr) Figure 20. Dispersion band thickness versus specific settler flow of dispersed phase (effect of O/A ratio and settler area), Key Lake phase system, organic continuous operation, 35°C, 300rpm. 97 Another c h a r a c t e r i s t i c of thi s phase system i s the low degree of scatter of the points around the best f i t curve, r e l a t i v e to the HSLIX64N-copper system. No obvious sat i s f a c t o r y explanation can be forwarded for t h i s difference in behavior, although i t i s possible that the behavior of th i s system did not d r i f t with time to the same extent as that of the HSLIX64N-copper system, thereby minimizing t h i s error source. The shorter duration over which the relevant data for the Key Lake system was obtained may have been a factor in t h i s . 7.2.4 Qd/A-AH DATA AT KEY LAKE, SASK. The s p e c i f i c s e t t l e r flow, Qd/A, in the KLMC extraction stage s e t t l e r s can be calculated to range from 3.08 m3/m2/hr to 4.27 m3/m2/hr. This i s based on the following data: aqueous phase flowrate: 90-125 m3/hr se t t l e r width: 3.96 m se t t l e r length [distance between i n l e t d i s t r i b u t o r and ba f f l e ( f i g . 4)]: 7.4m Although accurate measurements of dispersion band thickness in the Key Lake s e t t l e r s are not conducted by plant personnel, AH has been estimated to range from 5-20 cm (Neven, 1987). The dispersion band i s of the "deep-layer" type and i s of near uniform thickness. The O/A rat i o i s maintained between 1 and 1.5 and the operating temperature i s approximately 35°C. The s e t t l e r operating range corresponding to these values of Qd/A and AH i s represented by the shaded area of Figure 20. 98 Comparison with the data taken from the experimental s e t t l e r shows that the Key Lake s e t t l e r s operate with a considerably narrower dispersion band thickness than that which would be predicted from experimental data. The data supports the claims that the Krebs s e t t l e r design allows a reduction in the s e t t l e r area of one-quarter to one-half of that required by conventional s e t t l e r s Two primary features of the Krebs design (Fig. 4) are purported to f a c i l i t a t e t h i s enhanced performance; 1. A conical pump generates ce n t r i f u g a l forces which result in some coalescence of dispersion prior to discharge into the launder. 2. The launder allows some phase separation to occur before the dispersion enters the s e t t l i n g chamber. It was shown mathematically in section 2.9 that i f the scale up c r i t e r i o n of Ryon applies (as was shown to be the case in section 7.2.3), thin, superimposed s e t t l e r s w i l l provide improved s e t t l e r capacity over a non-partitioned s e t t l e r . This may partly explain the improved performance of the Krebs s e t t l e r design; the launder may, in fact act as a thin s e t t l e r , superimposed over the main s e t t l i n g chamber. No data was found to provide an indication of the magnitude of the improvement due to the conical pump's cen t r i f u g a l action. Therefore, the r e l a t i v e importance of the launder and the conical pump on the s e t t l e r performance i s d i f f i c u l t to gauge. 99 7.3 E F F E C T O F O P E R A T I N G T E M P E R A T U R E The e f f e c t of operating temperature on the location of the c h a r a c t e r i s t i c curve was investigated for the HSLIX64N-copper system. The re s u l t s , graphed in Figure 21 , show that a higher operating temperature re s u l t s in an increased s e t t l e r capacity, the c h a r a c t e r i s t i c curve being shifted downwards without perceptible change of slope. This can be attributed primarily to the decrease in the continuous phase v i s c o s i t y with increased temperature which does the following: 1. increases the rate of drainage of the continuous phase between 2 drops, f a c i l i t a t i n g more rapid coalescence. 2. increases the terminal s e t t l i n g v e l o c i t y of the droplets. The Key Lake system was studied only at 35°C because of i t s increased tendency for stable emulsion formation at lower temperatures. In i n d u s t r i a l operation, the increased s e t t l e r capacity f a c i l i t a t e d by operation at higher temperatures must be weighed against any extra costs associated with heating of process solutions. 100 I I S E T T L E R A R E A (m* # 2) 0.13 0.22 0.36 24 celcius o © • 35 celcius • ES • 0.5 Qd/A (m**3/m**2/hr) 10 Figure 21. Dispersion band thickness versus s p e c i f i c s e t t l e r flow of dispersed phase (effect of operating temperature), HSLIX64N-copper phase system, organic continuous operation, 300rpm, 0/A=1.5. 101 7.4 E F F E C T O F I M P E L L O R S P E E D In order to establish the degree of " s e n s i t i v i t y " or " i n s e n s i t i v i t y " of the experimental phase systems to mixing intensity, runs were conducted at impellor speeds of 250, 300 and 350 rpm (the maximum speed of the mixer). No runs were conducted with impellor speeds under 200 rpm since visual observation showed that the dispersion produced was sometimes not homogeneous in the sense that r e l a t i v e l y large slugs of a single phase would exit the mixer without a dispersion being produced (section 7.1.3). Sampling of the l o c a l holdup in the mixer confirmed the homogeneity of the dispersion above 250 rpm. Above th i s speed the O/A r a t i o was constant throughout the mixer, even at points just centimeters from the i n l e t ports. Results of this study are tabulated in Figures 22, 23 and 24 for the cases of organic continuous operation (HSLIX64N-copper system), aqueous continuous operation (HSLIX64N-copper system), and organic continuous operation (Key Lake system), respectively. In a l l cases, the impellor speed had no discernable e f f e c t on the c h a r a c t e r i s t i c curve within the range studied, indicating that the systems were insensitive to agitation intensity. Had the systems instead been found to be sensitive, then comparison of runs in which the residence time of the phases in the mixer was d i f f e r e n t ( i . e . , d i f f e r e n t flow rates) would have become very d i f f i c u l t . With a sensitive system, i f the flowrate were changed, the change in the power input per unit volume would also change the dispersion 102 50 E cn co <D c C J TJ c co c g to 0 Q. c/> XJ 10-1-r 0.5 • • / o Legend O RPM=250 • RPM=300 • RPM=350 Qd/A (m**3/m**2/hr) -i 1 r—r 10 Figure 22. Dispersion band thickness versus s p e c i f i c s e t t l e r flow of dispersed phase (effect of impellor speed), HSLIX64N-copper phase system, organic continuous operation, 24°C. 103 E co co (D C o !E +•* TJ C CO -O c o co L _ CD Q. CO Qd/A (m**3/m**2/hr) Figure 23. Dispersion band thickness versus specific settler flow of dispersed phase (effect of impellor speed), HSLIX64N-copper phase system, aqueous continuous operation, 24°C. • 1 04 50 to CO CD c 10 c CO c g CO k_ CD Q. CO XJ 0.5 L e g e n d O RPM=250 • RPM=300 • RPM=350 Qd/A (m**3/m**2/hr) 10 Figure 24. Dispersion band thickness versus specific settler flow of dispersed phase (effect of impellor speed), Key Lake phase system, organic continuous operation, 35°C. 105 c h a r a c t e r i s t i c s . Fortunately, the systems being investigated did not exhibit strong s e n s i t i v i t y to mixing, thereby making exact duplication of mixing regimes for d i f f e r e n t runs less c r i t i c a l . 7.5 DISPERSION BAND HOLDUP PROFILES In order to gain insights into the mechanism of phase separation, an experimental investigation of the structure of the dispersion band was undertaken by withdrawing samples from various v e r t i c a l locations within the dispersion band to determine the dispersion band holdup p r o f i l e . It was then possible to compare these results with those of Barnea and Mizrahi (section 2.3). The following cases were studied: 1.. organic continuous operation (HSLIX64N-Copper System), F i g . 25. 2. aqueous continuous operation (HSLIX64N-Copper System), F i g . 26. 3. organic continuous operation (Key Lake System), F i g . 27. The results for these three cases indicate that the concept of the existence of two d i s t i n c t sublayers cannot be extended equally to a l l phase systems. For the case of organic continuous operation with the HSLIX64N-copper system (Figure 25), i t can be seen that for low feed concentrations (0.3, 0.4, 0.5 v o l . % dispersed phase) i t was not possible to dis t i n g u i s h the two sublayers. The 1 0 6 CD CO CO T 3 CD CO <5 C L CO Q . 0.9-0.8 0.7 0.6 0.5-0 . 4 -o , o - o - - o — o - n Q ^ - A A A A -— 0.3H co o o 0.2 0.1-0 Legend V H O L D U P - 0 . 3 0 . A H - 2 6 c m . A - 3 6 m 2 • H O L D U P - 0 . 4 0 . A H - 3 0 c m . A - 2 2 m 2 • H O L D U P - 0 . 6 0 . A H - 2 6 c m . A - 3 6 m 2 A H O L D U P - 0 . 6 0 . A H - 2 7 c m . A - . 2 2 m2 O H O L D U P - 0 . 7 0 . A H - 2 8 c m . A - 3 6 m 2 • H O L D U P - 0 . 8 0 . A H - 2 6 . 6 c m . A - 2 2 m 2 I 0.1 0.2 0.3 0.4 0.5 h/AH 0.6 0.7 0.8 0.9 Figure 25. Holdup profiles within HSLIX64N-copper phase continuous operation, 24°C, dispersion band, system, organic 300rpm. 1 07 0 . 9 -0 . 8 -0 .7 52 0 . 6 -0 . 5 -0 . 4 — 0 . 3 -0 . 2 -0.1 O -O. "O" • o — o - - o - o --A-•o- _ o _ A • — • — • — • — • — • — • — • . A ~ • V - v v -Legend 1 3 m 2 V H O L D U P - 0 . 3 2 . A H - 3 3 c m , A -• H O L D U P - 0 . 4 0 . A H - 3 3 c m . A - 2 2 m 2 • H O L D U P - 0 . 6 2 . A H - 3 3 c m . A - 2 2 m 2 A H O L D U P - 0 . 6 0 . A H - 3 3 c m . A - . 1 3 m 2 O H O L D U P - 0 . 7 0 , A H - 3 0 c m , A - 2 2 m 2 0.1 0.2 0 . 3 0 . 4 0 .5 h/AH 0 . 6 0 .7 0 .8 0 .9 Figure 26. Holdup p r o f i l e s within dispersion band, HSLIX64N-copper system, aqueous continuous operation, 24°C, 300rpm. 1 08 0.9-0.8-0.7 XL 0.6 0.5-0.4-0.3-0.2-0.1-Legend V H O L D U P - 0 . 2 9 . A H - 2 9 c m . A - . 2 2 m 2 • H O L D U P - 0 . 4 1 . A H - 3 0 c m , A - . 2 2 m 2 • H O L D U P - 0 . 4 9 . A H - 2 6 c m , A - . 2 2 m 2 A H O L D U P - 0 . 6 0 . A H - 2 7 c m . A - . 2 2 m 2 0.1 0.2 0.3 0.4 0.5 h/AH 0.6 0.7 0.8 0.9 Figure 27. Holdup p r o f i l e s within dispersion Lake phase system, organic operation, 35°C, 300rpm. band, Key continuous 109 concentration increased smoothly and evenly from the s e t t l i n g front to the coalescence front. For these cases, the results support the observations of Gondo and Kusonoki (1969) who found that, regardless of the concentration of the feed to the s e t t l e r , the o v e r a l l (average) holdup in the dispersion band was roughly 0.55. For higher feed concentrations, however, (0.6, 0.7, 0.8 v o l . % dispersed phase) the concentration was found to be r e l a t i v e l y constant throughout the majority of the dispersion band (i d e n t i f y i n g the even concentration sublayer), with a discernable increase at the coalescence front ( i d e n t i f y i n g the dense sublayer). Figure 26, for aqueous continuous operation with the HSLIX64N-copper system, shows closer agreement with the studies of Barnea and Mizrahi. For a l l feed concentrations, the concentration p r o f i l e showed an even concentration throughout the dispersion band extending e s s e n t i a l l y to the coalescence front. For some curves, no increase in concentration near the concentration front could be detected, suggesting that the dense concentration layer, i f present, must be very narrow to avoid detection with the sampling probe, occupying less than perhaps 5% of the dispersion band. Figure 27, for the Key Lake system (organic continuous) exhibits the best agreement with the predictions of Barnea and Mizrahi. For a l l holdup values studied, both the dense concentration sublayer and the even concentration sublayer are 1 10 c l e a r l y d i s t i n g u i s h a b l e . The dense c o n c e n t r a t i o n s u b l a y e r f o r t h i s case o c c u p i e s r o u g h l y 20-25% of the d i s p e r s i o n band. The c o n c e n t r a t i o n i n the even c o n c e n t r a t i o n s u b l a y e r i s r o u g h l y t h a t of the incoming f e e d . The v a r i a t i o n i n the r e l a t i v e s i z e s of the two s u b l a y e r s and the a c t u a l l a c k of d i s t i n g u i s h a b l e s u b l a y e r s i n some ca s e s i l l u s t r a t e s the tremendous v a r i a t i o n i n the b e h a v i o r of v a r i o u s phase systems and the d i f f i c u l t i e s e n c o untered i n d e v e l o p i n g a u n i v e r s a l l y a p p l i c a b l e concept of the phase s e p a r a t i o n mechanism. 7.6 DISPERSION DROP SIZE DISTRIBUTION 7.6.1 GENERAL OBSERVATIONS V i s u a l o b s e r v a t i o n of d i s p e r s i o n drops t h r o u g h the g l a s s w a l l s of the s e t t l e r showed t h a t the d r o p l e t s near t h e c o a l e s c e n c e f r o n t were, i n g e n e r a l , n o t i c e a b l y l a r g e r than t h o s e i n the remainder of the d i s p e r s i o n band and grew l a r g e r as they approached the c o a l e s c e n c e f r o n t . I n a d d i t i o n , t h e s e d r o p l e t s appeared t o i n c r e a s e i n s i z e towards the end of t h e s e t t l e r . W i t h the un a i d e d eye, however, f o r thos e d r o p l e t s o c c u p y i n g the remainder of the d i s p e r s i o n band, no a p p r e c i a b l e d i f f e r e n c e i n drop s i z e i n e i t h e r the h o r i z o n t a l or v e r t i c a l d i r e c t i o n c o u l d be d e t e c t e d . Because of the s m a l l d i a m e t e r of the d r o p l e t s (50-500 urn), the use of the unaide d eye a l o n e was inadequate f o r d e t a i l e d a n a l y s i s of the d i s p e r s i o n drop s i z e d i s t r i b u t i o n and c o u l d 111 provide only the most general information regarding drop size. In order to obtain more quantitative information, the technique described in section 6.5 was used. Dispersion samples were charged to a glass c e l l , mounted on the stage of a microscope and then photographed through the microscope's occular arrangement. The drop size d i s t r i b u t i o n was then determined from the photographs. When a sample was charged to the c e l l , batch separation would commence; the coalescence front and s e t t l i n g front would develop and would advance towards each other u n t i l , 20 to 40 seconds after the sample had been charged to the c e l l , disengagement of the 2 phases would be complete. For each sample, i t was v i s u a l l y observed through the microscope, that the average drop size remained substantially constant with time u n t i l approximately 30-50 seconds after i n i t i a l sample withdrawal, the coalescence front had advanced to b r i e f l y occupy the microscope's f i e l d of view. Droplets in the v i c i n i t y of the coalescence front were considerably larger and more compacted than those in the remainder of the sample and coalescence occurrences were readily observable. In contrast, no s i g n i f i c a n t coalescence was observed before the coalescence front had advanced to f i l l the f i e l d of view. The v i s u a l observation that the drop size did not change substantially with time except at the coalescence front, was v e r i f i e d photographically. Samples, representative of the dispersion entering the s e t t l e r for 5 d i f f e r e n t mixer operating 1 12 conditions, were withdrawn from the mixer overflow. Through analysis of photographs taken successively after the c e l l had been charged, the change in sample arithmetic mean drop size prior to the coalescence front occupying the f i e l d of view, as observed through the microscope, was determined (figure 28). Examination of Figure 28 shows that the arithmetic mean drop diameter changed minimally (less than 20 % ) , between the time of the f i r s t photograph, taken 15 seconds after sample withdrawal, and the time of successive photographs. An exception is sample 1 with an O/A r a t i o of 0.67. In this case, the last photograph of the series, corresponding to t=30 seconds represents a mean drop diameter of over 600 urn, compared to 375 urn for the f i r s t point of the series (t=10 seconds). This large drop size can be attributed to inadvertently photographing the coalescence front as i t passed through the microscope's f i e l d of view immediately before completion of phase disengagement. The observation that the drop size changes minimally with time i s an important one. It establishes the v a l i d i t y of the assumption that the size d i s t r i b u t i o n determined from a photograph taken after the c e l l had been charged, about 10 seconds after i n i t i a l sample withdrawal, was representative of that of the dispersion taken from the mixer-settler. Had the drop size changed s i g n i f i c a n t l y with time, photographic analysis for drop size d i s t r i b u t i o n would have been complicated in that i t would have been necessary to extrapolate drop size 1 1 3 600 E 3 t 500 E CO CO +-> CD E CO TD CL O 400-300 c co CD E CD £ CO 200-100-0 Legend • Sample 1: 0/A=.67, Qd=19.1 l/min O Sample 2: 0 /A= 67. Qd=13.2 l/min • Sample 3: C7A=1.5, Qd=14 l/min • Sample 4: G7A=1.5, Qd=5 l/min V Sample 5: C7A=3.0, Qd=5 l/min 10 20 30 40 t i m e a f t e r w i t h d r a w a l o f s a m p l e ( sec . ) 50 Figure 28. Change in arithmetic mean drop size of dispersion sample with time after withdrawal from mixer, Key Lake phase system, organic continuous operation, 35°C, 300rpm. 1 1 4 data back to zero time in some manner, in order to approximate the actual mean drop size of the o r i g i n a l sample. Data points were obtained in a l l cases from single photographic s l i d e s , t y p i c a l l y showing about 30-60 droplets. Although t h i s sample size i s r e l a t i v e l y small and could result in a s i g n i f i c a n t s t a t i s t i c a l error in mean drop siz e , i t was thought to be acceptable since the purpose of the drop size determinations was merely to esta b l i s h general trends in dispersion mean drop size with changing parameters, not to obtain precise drop size values. In any case, determining the precise drop size would have no p r a c t i c a l application since impellor geometries and mixer-settler arrangements and operating conditions vary between i n d u s t r i a l i n s t a l l a t i o n s and would result in varying drop size s . 7.6.2 DOUBLE AND TRIPLE DISPERSIONS An interesting revelation of the photography was the appearance of double and, to a lesser extent, t r i p l e dispersions; that i s , drops within drops within drops. The outer, primary drop (aqueous phase) was spherical in a l l cases while the secondary drops within t h i s drop (organic phase) were somewhat deformed (Figure 29). The phenomenon, although s i g n i f i c a n t for a l l samples under a l l operating conditions, appeared to be somewhat less prevalent for samples withdrawn close to the dispersion band coalescence front. This indicates that the encapsulated 115 1000qm Figure 29. Photograph of double dispersion sample withdrawn from s e t t l e r i n l e t , Key Lake phase system, organic continuous operation, 0/A=1.5, 35°C, 300rpm. 1 16 droplets were expelled to some degree as the droplets aged, possibly as a result of coalescence. Note; In analyzing the s l i d e s , secondary and t e r t i a r y drops were not d i g i t i z e d and therefore they did not influence the calculated drop size d i s t r i b u t i o n . These drops were omitted since each assembly moved as a single unit. The appearance of a double and t r i p l e dispersion has several implications. 1. The secondary organic droplets could be c a r r i e d close to the coalescence front by the encapsulating primary droplet and then be ejected when the primary (aqueous) droplets coalesce. This would be expected to increase organic entrainment and losses in the mixer-settler. Because of this adverse e f f e c t of multiple dispersions, i t would be worthwhile to s t r i v e to design i n d u s t r i a l mixers to minimize the occurrence. 2. Since the e f f e c t i v e density of primary droplets varies depending on the number and size of encapsulated droplets, mathematical analysis of the s e t t l i n g process i s complicated: individual droplets have d i f f e r e n t terminal v e l o c i t i e s and thus exhibit d i f f e r e n t hydrodynamic behaviors. During observation of the dispersion band in the mixer-settler, i t was noted (section 7.1) that the coalescence of two droplets at the coalescence front was frequently 117 accompanied by the ejection of a small continuous phase droplet into the s e t t l e d dispersed phase. This was attributed to hydrodynamic forces. However, i t i s also possible that the ejected droplets may actually have been secondary (encapsulated) droplets, released upon the coalescence of the primary droplet. Because of the small sizes and the d i f f i c u l t y of observing the phenomenon, the mechanism was not c l e a r l y established. Key Lake Mining Corporation has indeed reported organic losses greater than expected; two to three times higher than guaranteed (Neven, 1987). Since no microscopic examination of the dispersion i t s e l f has been car r i e d out at the plant to date, i t cannot be confirmed whether these high organic losses could be attributed to double dispersion formation. Observations at Key Lake have revealed large "slugs" of aqueous continuous dispersion entering the s e t t l e r when the c i r c u i t is upset. This slug phenomenon was also observed in the experimental mixer-settler, p r i n c i p a l l y when the concentration of the dispersed phase was high or when the mixer impellor speed was i n s u f f i c i e n t to adequately disperse the phases (sect ion 7.1.3). This s i m i l a r i t y in behavior, coupled with the fact that the experimental phase system should be representative of that e x i s t i n g in the Key Lake c i r c u i t , allows one to conclude that i t i s highly l i k e l y that the double dispersion phenomenon occurs at Key Lake. It is possible that t h i s may be 118 contributing to the high solvent losses. No references to the appearance of double and t r i p l e dispersions in i n d u s t r i a l phase systems were found. Pike and Wadhamen (1971) observed the phenomenon using a phase system consisting of water and a commercial a l i p h a t i c petroleum fraction (AMSCO 460 solvent, by the American Mineral S p i r i t s Company, Palatine, 111.). The small size of the secondary and t e r t i a r y droplets makes i t d i f f i c u l t to d i s t i n g u i s h them with the unaided eye, possibly explaining the l i m i t e d reports on the subject. It would be reasonable to assume, however, that the phenomenon does occur to some degree in i n d u s t r i a l mixer-settlers, at least with the Key Lake phase system, since i t was observed experimentally under a l l operating conditions. 7 . 6 . 3 EFFECT OF O/A RATIO AND Qd ON MIXER DROP SIZE A series of samples were taken at the mixer overflow in order to determine the e f f e c t of mixer O/A r a t i o and dispersed phase flowrate on the arithmetic mean diameter of the droplets entering the s e t t l e r . The r e s u l t s , presented in Figure 30 show that, for a given O/A r a t i o , the dispersed phase flowrate has no appreciable ef f e c t on the drop si z e . This indicates that residence time in the mixer at a given phase r a t i o , does not c r i t i c a l l y a ffect the mean drop size of the dispersion produced. 1 19 500 E CD X E _c i— CD CD E CO X J Q. O C CD CD E o CD E CO 4 0 0 -300 A. A 2 0 0 -100 Legend O 0/A=0.67 • 0/A=1.5 A 0/A=3.0 5 10 15 20 25 dispersed phase flowrate (l/min) Figure 30. Effect of 0/A ratio and dispersed phase flowrate on arithmetic mean drop diameter at mixer overflow, Key Lake phase system, organic continuous operation, 35°C, 300rpm. 120 The O/A r a t i o i t s e l f , however, can have a c r u c i a l effect on the drop size produced. Figure 30 shows that, although there is no s i g n i f i c a n t difference in the dispersion produced at O/A=0.67 and 0/A=1.5, the mean drop size produced for the more di l u t e dispersion (O/A=3.0) i s considerably smaller than that produced for the more concentrated dispersions. Photographs of dispersions with O/A rati o s of 0.67 and 3.0, shown in Figures 31(a) and 31(b) respectively, i l l u s t r a t e t h i s difference. This result provides an explanation for the differences between the c h a r a c t e r i s t i c curves observed for the three d i f f e r e n t phase rat i o s with the Key Lake system (Figure 20). It had been determined that, although the c h a r a c t e r i s t i c curves for O/A=0.67 and 0/A=1.5 are e s s e n t i a l l y i d e n t i c a l , the curve for O/A=3.0 was shi f t e d upwards, indicating that for a given dispersed phase throughput, a thicker dispersion band was required. This would be expected from the drop size data obtained, since a finer dispersion (such as that produced for O/A=3.0) would require a greater dispersion band thickness for phase disengagement to occur. The e s s e n t i a l l y i d e n t i c a l arithmetic mean drop diameter as well as the i d e n t i c a l c h a r a c t e r i s t i c curves produced for O/A = 0.67 and O/A = 1.5 has implications for i n d u s t r i a l operations. Since, in general, i n d u s t r i a l mixer-settlers for uranium extraction maintain O/A rati o s deviating l i t t l e from 1.0, the generalization that phase separation with such phase systems does not depend on the O/A r a t i o ( i . e . : on the continuous phase 121 (b) Photographs showing e f f e c t of O/A r a t i o _ on drop s i z e of d i s p e r s i o n withdrawn from mixer o v e r f l o w , Key Lake phase system, o r g a n i c c o n t i n u o u s o p e r a t i o n , 3 5 ° C . (a) O/A=0.67, (b) O/A=3.0. 122 flow) appears to be v a l i d , p r a c t i c a l l y speaking, over th i s normal, limited range of operation. However, as can be seen from Figures 20 and 30, the generalization i s no longer correct when extended to more d i l u t e dispersions (O/A=3.0); the mean drop size appears to be s u f f i c i e n t l y smaller to s h i f t the c h a r a c t e r i s t i c curve upwards. 7.6.4 EFFECT OF IMPELLOR SPEED ON MEAN DROP SIZE The e f f e c t of impellor speed on the arithmetic mean drop size of the dispersion produced appears to depend on the O/A r a t i o of the dispersion (Figure 32). For O/A = 0.67 and O/A = 1.5, increasing impellor speed tends to have a moderate e f f e c t , producing a somewhat finer dispersion. From t h i s , one would expect that the c h a r a c t e r i s t i c curve at these phase ra t i o s would be shifted upwards with increasing impellor speed. However, from Figure 22, th i s was not observed, suggesting that the decrease in drop diameter with impellor speed over the range tested i s not s u f f i c i e n t to appreciably change the s e t t l i n g c h a r a c t e r i s t i c s of the dispersion. For O/A=3.0, the mean drop size does not appear to be sensitive to impellor speed. 1 23 500 400-3 +-> E % 300 a o 200-c CD CD E o CD E x: co 100 -v Legend O 0/A=.67. Qd=13.2 l/min • 0/A=1.5, Qd=9.5 l/min V 0/A=3.0, Qd=13.6 l/min 200 250 300 350 impellor speed (rpm) 400 Figure 32, Effect of impellor speed on arithmetic mean drop diameter at mixer overflow, Key Lake phase system, organic continuous operation, 35°C. 1 24 7.6.5 DROP SIZE PROFILE WITHIN DISPERSION BAND Dispersion samples were withdrawn from various points in the dispersion band to obtain the dispersion band drop size p r o f i l e , r e l a t i n g arithmetic mean drop size to the v e r t i c a l location in the dispersion band (expressed as the reduced dimensionless distance from the coalescence front; O=coalescence front, 1=settling f r o n t ) . The dispersion band drop size p r o f i l e was obtained for s e t t l e r feed O/A ratios of 0. 67. 1.5 and 3.0 (Figures 33, 34 and 35, respectively). Furthermore, for each of these three cases, the p r o f i l e was determined at two horizontal locations within the dispersion band; 1. At a point one t h i r d of the distance between the se t t l e r i n l e t and the end of the s e t t l e r . 2. Approximately 5-10 cm from the end of the s e t t l e r . Constant parameters were dispersion type (organic continuous), temperature (35°C) and s e t t l e r size (0.22 m 2). Comparison of Figures 33, 34 and 35 reveals a d e f i n i t e s i m i l a r i t y in the shapes of the p r o f i l e s for these three cases. Furthermore, the general shape of the drop size p r o f i l e corresponds closely to that for the concentration (holdup) p r o f i l e of the dispersion band. In fact i t i s possible to a r b i t r a r i l y distinguish the two sublayers, the even concentration sublayer and the dense concentration sublayer based on drop size data alone. With respect to drop size, the even concentration sublayer, comprising roughly 80 % of the 1 25 600 500-E 3 a> 4 0 0 CD E CD a o c CD CD E o 300 I 200 CD 100-Sample withdrawal point O 1/3 distance between settler inlet and end • 10cm from end of settler O r 0 —I ' r-0.4 0.6 h/AH 0.2 0.8 Figure 33. Drop size profile within dispersion band for O/A=0.67, Key Lake phase system, organic continuous operation, 35°C, 300rpm. 1 26 700 c CD CD E o +-» CD E 600-500 H 3 CD •«—> E ^ 400 CL O 300 CD 2 0 0 -100-Sample withdrawal point O 1/3 distance between settler inlet and end • 10cm from end of settler 0-J 1 | 1 1 1 1 1 1 1 0 0.2 0.4 0.6 0.8 1 h/AH Figure 34. Drop size p r o f i l e within dispersion band for 0/A=1.5, Key Lake phase system, organic continuous operation, 35°C, 300rpm. 127 600 500-© 400 E CO TJ 300-| 200 100 Sample withdrawal point O 1/3 distance between settler inlet and end • 10cm from end of settler 0.2 0.4 0.6 0.8 h/AH Figure 35. Drop size p r o f i l e within dispersion O/A=3.0, Key Lake phase system, continuous operation, 35°C. band for organic 1 28 dispersion bands' volume consists of droplets of approximately constant mean diameter. Near the coalescence front, however, a d e f i n i t e increase in the mean diameter i s detectable, i d e n t i f y i n g the dense concentration sublayer. Comparison of the approximate mean drop size of the even concentration layer with that of samples taken at the mixer overflow reveals that the drop size i s approximately equal to that issuing from the mixer. This i s analogous to the situ a t i o n encountered when comparing the concentration in the even concentration sublayer with that in the mixer; they too are approximately equal. Comparison of the arithmetic mean drop size p r o f i l e e x i s t i n g near the i n l e t of the s e t t l e r with that near the outlet shows that no substantial difference between the two can be detected. This agrees with vi s u a l observation of the dispersion band except in one respect; although no change in drop size with horizontal position was v i s u a l l y detected in the even concentration sublayer, i t was possible to notice an increase in drop size with distance from the s e t t l e r i n l e t at the coalescence front. It i s l i k e l y that the trend observed at the coalescence front was not detected photographically due to lim i t a t i o n s in the sampling technique. That i s , i t was not possible to prec i s e l y withdraw a sample within 1-2 cm of the coalescence front. The fact that no substantial change in drop size occurred between the front and the back of the s e t t l e r appears to 129 confirm that most of the coalescence occurs at the coalescence front. Otherwise, in the even concentration layer, an increase in drop size would be detectable as the dispersion advanced towards the back of the s e t t l e r and aged. The two sublayers had also been distinguishable in observations of batch separation in the c e l l (section 7.6.1); drop size remained r e l a t i v e l y constant with time as long as the "even concentration" sublayer was in the microscope's f i e l d of view. When the coalescence front occupied the f i e l d of view, t h i s was marked by larger more compacted drops. Figures 33, 34 and 35 confirm the results obtained from sampling at the mixer overflow. The mean drop size in the even concentration sublayer was about equal at 400 urn for O/A=0.67 and 0/A=1.5.. For O/A=3.0, however, the mean drop size was about 300 um, considerably smaller. 7.7 THEORETICAL TERMINAL VELOCITY In order to gain some insight into the role played by hindered s e t t l i n g in the phase separation process, i t is desirable to have available information on the effect of drop size, v i s c o s i t y , density, and the volumetric concentration of dispersed phase on the terminal v e l o c i t y of the droplets. Barnea and Mizrahi (1975, part III) developed a general semitheoretical c o r r e l a t i o n , supported by experimental results from fourteen independent sources, which could be used to 130 estimate the terminal s e t t l i n g v e l o c i t y of l i q u i d droplets in a concentrated dispersion. This correlation states that the r e l a t i o n s h i p established for a single spherical droplet, between the drag c o e f f i c i e n t , C D and the Reynolds number, Re, i s v a l i d for the case of concentrated dispersions provided that C D and Re are modified as follows: 4d(p,-p )g(1 - c6) C = - - (7) 0 0 3 p c V { 1 + 0 > 3 3 ) Re - * c (8) where i s the e f f e c t i v e v i s c o s i t y of the dispersion. For the case of l i q u i d spherical droplets with unretarded internal r e c i r c u l a t i o n (which results in surface motion, increasing the terminal v e l o c i t y ) , M ^ can be expressed as follows: M ( 2 M , 1 + 3/z,) _c ± d_ ( 9 ) 0 ( 3 M / + 3M d) 5M d+ 2M 0 where u. 1 = M exp — (10) <p c 3M d + 3M C 1-0 The above relationships are based on the following assumptions: - the p a r t i c l e s are spherical - the p a r t i c l e s ' spacial positions are completely random - there i s no wall effect - interactions between p a r t i c l e s are hydrodynamic only 131 The well known relationship between Reynolds number and the drag c o e f f i c i e n t can be approximated by the following three equations (Perry V, p. 5.61). 24 Re .< 0.3 (11) C = * D < t > Re , (Stoke's Law) 0 18.5 0.3< Re ,< 1000 (12) C = ^ D * (Re,) 0- 6 (Intermediate Law) C^= 0.44 - 1000< Re,< 200000 (13) uq> <p (Newton's Law) Using equations 7-13 and the property data of table 5, terminal v e l o c i t i e s for drops of various diameters and in dispersions of various concentrations, have been calculated. The results are graphed in Figures 36,37 and 38 for the following three cases: 1. HSLIX64N-copper phase system (org. cont.); Figure 36. 2. HSLIX64N-copper phase system (aq. cont); Figure 37. 3. Key Lake phase system (org. cont.); Figure 38. The p r a c t i c a l information which may be gleaned from the terminal v e l o c i t y data i s rather limited, as i t does not consider the coalescence aspect of phase separation. Furthermore, the flow patterns and eddies existing in the dispersion band are also disregarded. The t h e o r e t i c a l data would also have limited v a l i d i t y for the case of double dispersions, where drop densities vary. Nevertheless, some very general observations may be made. For example: 1 32 10 drop diameter (urn) e 36 Theoretical terminal drop v e l o c i t y versus drop diameter for d i f f e r e n t dispersion concentrations, HSLIX64N-copper qphase system, organic continuous operation, 24°C. 1 33 10 drop diameter (urn) Figure 37. Theoretical terminal drop v e l o c i t y versus drop diameter for d i f f e r e n t dispersion concentrations, HSLIX64N-copper phase system, aqueous continuous operation, 24°C. 134 10 drop diameter (urn) Figure 38. Theoretical terminal drop velocity versus drop diameter for different dispersion concentrations, Key Lake phase system, organic continuous operation, 35°C. 135 Figures 36,37, and 38 i l l u s t r a t e how hindered s e t t l i n g i s retarded as drop size decreases (lower terminal v e l o c i t i e s ) . Comparison of Figures 36 and 37 reveals the increased s e t t l i n g v e l o c i t y associated with aqueous continuous operation, which contributes to the smaller dispersion band thickness expected in t h i s mode of operation. 1 36 CHAPTER 8 - SUMMARY AND CONCLUSIONS An experimental investigation was undertaken to study the factors a f f e c t i n g phase separation in a laboratory scale mixer-settler. Two phase systems were tested: 1. The laboratory prepared HSLIX64N-copper phase system, similar to that used in commercial copper solvent extraction processes. 2. A phase system obtained d i r e c t l y from the uranium extraction c i r c u i t of the Key Lake Mining Corporation, Sask., m i l l s i t e . From experiments conducted in a specially-constructed, glass-walled, mixer-settler test unit, the following key observations and conclusions were derived. 1. Depending on the phase system, the l e v e l of introduction of dispersion into the s e t t l e r could have a c r i t i c a l e f fect on the thickness of the dispersion band produced. With the HSLIX64N-copper phase system (both organic and aqueous continuous operation) introduction l e v e l had no appreciable e f f e c t . However, with the Key Lake system, introducing the dispersion at or near the l e v e l of the coalescence front gave a larger AH than i f the dispersion was introduced elsewhere. This was attributed to the incoming dispersion occupying a portion of the coalescence front, thereby reducing 1 37 the e f f e c t i v e area of the s e t t l e r . The scale-up c r i t e r i o n of Ryon ( i e . AH=k(Qd/A)^) was tested by p l o t t i n g the c h a r a c t e r i s t i c curves for d i f f e r e n t O/A ratios for both phase systems. The relationship provided a good approximation for the Qd/A-AH data obtained. However, the following was observed: For the HSLIX64N-copper system, a single c h a r a c t e r i s t i c curve could be used to f i t the data taken for a l l O/A r a t i o s , indicating that for t h i s system AH depends only on the dispersed phase throughput and not on that of the continuous phase For the Key Lake phase system, separate c h a r a c t e r i s t i c curves were produced for each O/A r a t i o , indicating that with t h i s system AH depends on both the dispersed and the continuous phase throughput. For a given value of Qd/A, operation with a low dispersed phase concentration gave the thickest dispersion band. Comparison of the estimated dispersion band thickness ex i s t i n g in Key Lake's Krebs mixer-settlers, with the value of AH projected from the experimental data, shows that the Krebs mixer-settlers' performance exceeds that predicted from the experimental data. This may be due to the conical pump and superimposed launder features of the Krebs design, as claimed by Krebs. 1 38 Phase separation i s enhanced at higher operating temperatures. This may be attributed primarily to the lower continuous phase v i s c o s i t y associated with higher temperatures. Within the range tested, mixer impellor speed did not a f f e c t the dispersion band thickness, supporting the contention of Ryon that power input does not affect AH. From the holdup p r o f i l e of the dispersion band, i t was generally (but not always) possible to d i s t i n g u i s h the "dense" and "even" concentration sublayers i d e n t i f i e d by Barnea and Mizrahi. - For the Key Lake phase system the sublayers could be e a s i l y i d e n t i f i e d for a l l O/A ra t i o s studied. For the HSLIX64N-copper phase system (aq. continuous) the sublayers were also c l e a r l y distinguishable. However, for organic continuous operation at high O/A r a t i o s , i t was not possible to d i s t i n g u i s h the two. The holdup decreased uniformly, nearly l i n e a r l y , from the coalescence front to the s e t t l i n g front. In these cases, the results c l o s e l y correspond to those of Gondo and Kusonoki (1969) who found that, regardless of the concentration of the s e t t l e r feed, the average holdup in the dispersion band was constant at roughly 0.55. Microscopic examination of the mixer-settler dispersion (Key Lake system) revealed the existence 139 of double dispersions; that i s , drops within drops. This would be expected to increase continuous phase losses and may account for the higher than anticipated organic losses being experienced by Key Lake Mining Corp. Determination of the average drop size of the dispersion, for the Key Lake system, showed that the O/A r a t i o can dramatically affect the average drop size produced. The average drop size of concentrated dispersions was considerably greater than that of more d i l u t e dispersions. This provides an explanation for the separate c h a r a c t e r i s t i c curves obtained for d i f f e r e n t O/A r a t i o s with t h i s phase system. Because of the opacity of the HSLIX64N-copper phase system, no drop size data could be obtained for t h i s system. However, since a single c h a r a c t e r i s t i c curve could be used to f i t the data for a l l phase r a t i o s , i t may be inferred that the nature of the HSLIX64N-copper phase system i s such that the O/A r a t i o does not change the dispersion drop size d i s t r i b u t i o n to the extent necessary to noticeably s h i f t the c h a r a c t e r i s t i c curve. The average dispersion drop size was shown to decrease somewhat with increasing impellor speed. The e f f e c t was most pronounced for d i l u t e dispersions. Since i t had also been observed that, for 0/A=1.5, impellor speed did not s h i f t the c h a r a c t e r i s t i c curve appreciably, i t was concluded that the change in drop size was not substantial enough to s i g n i f i c a n t l y a f f e c t phase separation. 10. The drop size p r o f i l e for the Key Lake dispersion band allowed a d i s t i n c t i o n between the dense and even concentration sublayers to be made. In the even concentration sublayer, occupying 80-90% of the band, the average drop size was approximately constant throughout. However, near the coalescence front the mean drop diameter increases sharply, i d e n t i f y i n g the dense concentration sublayer. As indicated by the variations in behavior of the two phase systems, there does not appear to be a simple, universally applicable model of dispersion band structure, or of the separation mechanism in a deep layer s e t t l e r . It was possible to extend the "dense concentration-even concentration" sublayer concept to most, but not a l l cases. S i m i l a r l y , although the scale-up c r i t e r i o n of Ryon appears to have considerable merit, the diff e r e n t e f f e c t of phase r a t i o on the Qd/A-AH relationship for the two phase systems, i l l u s t r a t e s the unp r e d i c t a b i l i t y of the behavior of di f f e r e n t phase systems. Phase separation in a mixer-settler i s a complicated process which remains to some extent unpredictable and not f u l l y understood. It i s l i k e l y that design of s e t t l e r s w i l l continue to depend on empirical data taken for each s p e c i f i c phase system. 141 NOMENCLATURE A horizontal cross-sectional area of s e t t l e r L 2 C D drag c o e f f i c i e n t C D ^ modified drag c o e f f i c i e n t d drop diameter L D diameter of c y l i n d r i c a l horizontal s e t t l e r L g g r a v i t a t i o n a l constant ML/T2 h distance from coalescence front L AH dispersion band thickness L k experimental constant (equation 1) N impellor speed (rpm) 1/T O/A org./aq. volume r a t i o Q t o t a l dispersion flowrate = Q c +Q (j L 3/T Q c continuous phase flowrate L 3/T dispersed phase flowrate L 3/T Re Reynolds number Re^ modified Reynolds number U^ r e l a t i v e v e l o c i t y between drops and L/T continuous phase V dispersion band volume L 3 y experimental constant (equation 1) 7 i n t e r f a c i a l tension M/T2 M v i s c o s i t y M/LT M continuous phase v i s c o s i t y M/LT 142 Mg* dispersed phase v i s c o s i t y M/LT e f f e c t i v e dispersion v i s c o s i t y M/LT v kinematic v i s c o s i t y L 2/T p density M/L3 P c continuous phase density M/L3 p^ dispersed phase density M/L3 c6 holdup of dispersed phase 1 43 REFERENCES Bailes,P.J., J.C. Godfrey, and M.J. Slater. Designing Liquid-Liquid Extraction Equipment. The Chemical Engineer, July 1981, pp. 331-333. Barnea,E.. Liquid/Liquid Contacting-Art or Science? Part I I . Design Methods for Mixer S e t t l e r s . Hydrometallurgy, 5 (1980), pp. 127-147. Barnea,E.. Flooding Conditions in a Deep-Layer Liquid/Liquid S e t t l e r . Trans. I. Chem. E., 56 (1978), pp. 73-76. Barnea,E. and J . Mizrahi. Separation Mechanism of Liquid/Liquid Dispersions in a Deep Layer Gravity Settler-Part I. Trans. Instn. Chem. Engrs., 53 (1975), pp. 61-69. Barnea,E. and J . Mizrahi. Separation Mechanism of Liquid/Liquid Dispersions in a Deep Layer Gravity Settler-Part I I . Trans. Instn. Chem. Engrs., 53 (1976), pp. 70-74. Barnea,E. and J. Mizrahi. Separation Mechanism of Liquid/Liquid Dispersions in a Deep Layer Gravity Settler-Part I I I . Trans. Instn. Chem. Engrs., 53 (1976), pp. 75-82. Barnea,E. and J . Mizrahi. Separation Mechanism of Liquid/Liquid Dispersions in a Deep Layer Gravity Settler-Part IV. Trans. Instn. Chem. Engrs., 53 (1976), pp. 83-92. Central S c i e n t i f i c Company. The Ring Method for Surface and I n t e r f a c i a l Tensions with the Cenco-duNouy Tensiometers. B u l l e t i n 101. Davies,G.A., G.V. Jeffreys, F. A l i . Design and Scale Up of Gravity S e t t l e r s . The Chemical Engineer. London, no. 243, Nov. 1970, pp. 378-385 Davies,G.A. et a l . . The Formation of Secondary Droplets in a Dispersion at a Phase Boundary. Can. J. Chem. Engng., 48 (1970), pp. 328-329. DeMent,E.R. and N. Lynn. The Liquid Ion Exchange Process. TMS-AIME Short Course, New Orleans, Louisiana, Feb. 16-18, 1979. Selection of Equipment and Materials: A Special Workshop on Hydrometallurgical Plant Design, pp. 9.1-9.18. Flett,D.S. Some Recent Developments in the Application of Liquid Extraction in Hydrometallurgy. The Chemical Engineer, July 1981, pp. 321-324. Flett,D.S., D.N. Okuhara, and D.R. Spink. Solvent Extraction of 1 44 Copper by Hydroxyoximes. J. Inorg. Nucl. Chem., 35 (1973), pp. 2471-2487. Gardner,S.A. and G.C.I. Warwick. Pollution-free Metallurgy: Copper via Solvent Extraction. E/MJ-April 1971, pp. 108-110. Glasser,A., D.R. Arnold, A.W. Bryson, and A.M.S. V i e l e r . Aspects of Mixer Settler Design. Minerals Science Engineering, v. 8, no. 1, (1976), pp. 23-44. Godfrey,J.C., C. Hanson, M.J. Slater,and S. Tharmalingam. Studies of Entrainment in Mixer-Settlers. AIChE Symposium Series, v.74, no. 173. Fundamental Aspects of Hydrometallurgical Processes, 1978, pp. 127-133. Golob,J., and R. Modic. Coalescence of Liquid/Liquid Dispersions in Gravity S e t t l e r s . Trans. I. Chem. Eng., 55 (1977), pp. 207-211. Gondo,S., and K. Kusonoki. A New Look at Gravity S e t t l e r s . Hydrocarbon Processing, Sept. 1969, pp. 209-210. Groothius,H., and F.S. Zuiderweg. Influence of Mass Transfer on Coalescence of Drops. Chemical Engineering Science, 12 (1960), pp. 288-289. Hanson,C. Handbook of Solvent Extraction. John Wiley and Sons., 1983. Hanson,C. Recent Advances in Liquid-Liquid Extraction. Pergammon Press Ltd., 1971. H i l l , E . C , D.A. Evans, and I. Davies. The Growth and Survival of Micro-organisms in Aviation Kerosene, J. Inst. P e t r o l . , v. 53, no. 524, (1967), pp. 280-284. Hudson,M.J. An Introduction to Some Aspects of Solvent Extraction Chemistry in Hydrometallurgy. Hydrometallurgy, 9 (1982), pp. 149-168. Jeffreys,G.V., G.A. Davies, and K. P i t t . Rate of Coalescence of the Dispersed Phase in a Laboratory Mixer-Settler Unit: Part I, A. I. Ch. E. Journal, 16 (1970), pp. 823-831. Johnstone,R.E., and M.W. Thring. P i l o t Plants,Models, and Scale-up Methods in Chemical Engineering. McGraw H i l l , 1 957. Keeney,M.E., and K. Osseo-Asare. Thermal S t a b i l i t y of LIX63 Oxime. Hydrometallurgy, 12 (1984), pp. 255-265. Kordosky,G.A.. The Chemistry of Metals Recovery Using LIX® 145 Reagents. General M i l l s Chemicals Inc. (Henkel Corporation), 1979. Lewis,I.E.. Design of Mixer-Settlers to Achieve Low Entrainment and Reduce Capital Costs. Proceedings. International Solvent Extraction Conference ISEC71, Toronto, Sept. 1977, pp. 325-333. Lott,J.B., G.I. Warwick, and J.B. Scuffham. Design of Large Scale Mixer S e t t l e r s . Trans. Soc. Min. Engrs., AIME, 252 (1972), pp. 27-35. MacKay,G.D.M., and S.G. Mason. The Marangoni E f f e c t and Liquid/Liquid Coalescence. Nature, 191 (1961), pp. 191-191. Manchanda,K.D., and D.R. Woods. S i g n i f i c a n t Design Variables in Continuous Gravity Decantation. Industrial and Engineering Chemistry. Process Design and Development., v o l . 7, no. 2 (1968), pp. 182-187. McClary,M.J., and G.A. Mansoori. Factors Affecting the Phase Inversion of Dispersed Immiscible Liquid-Liquid Mixtures. AIChE Symposium Series no. 173, volume 74 (1978) , pp. 134-139. McGarr,H.J.. Liquid Ion-Exchange Recovers Copper from Wastes and Low Grade Ores. E/MJ, October 1970, pp. 79-81. Merritt,R.C.. The Extractive Metallurgy of Uranium Colorado School of Mines Research I n s t i t u t e . (1971). Mizrahi,J., and E. Barnea. Compact Settler Gives E f f i c i e n t Separation of Liquid/Liquid Dispersion. Process Engineering, Jan. 1973, pp. 60-65. Murray,D.J.. Equipment Development in Solvent Extraction. (1979) . p. 367. Neven,M., and L. Gormely. Design of the Key Lake Leaching Process. Twelfth Annual Hydrometallurgical Meeting, Uranium '82. Toronto, Canada. Aug 29-Sept 1, 1982. Neven,M., R. Steane, and J . Becker. Arsenic Management and Control at Key Lake. CIM Met a l l u r g i c a l Society, Hydrometallurgy Section. F i f t e e n t h Annual Hydrometallurgical Meeting (Proceedings). Vancouver,Canada, Aug. 18-22, 1985, pp. 6.1-6.21. Oldshue,J.Y., and A.T. Gretton. H e l i c a l C o i l Heat Transfer in Mixing Vessels. Chem. Eng. Progr., 50 (1954), pp. 615-621. Orjans,J.R., CW. Notebaart, J.C. Godfrey, C. Hanson, and 1 46 M.J. Slater. The Design of Mixer Se t t l e r s for the Zambian Copper Industry. Proceedings. International Solvent Extraction Conference ISEC77, Toronto, Sept. 1977, pp. 340-346. Paatero,E.Y.O.. Determination of the Composition of the Hydroxyoxime Extractant LIX65N. Hydrometallurgy, 13 (1984), pp. 193-201. Pike,F.P., and S.E. Wadhamen. Properties of an Emulsion Band in a Mixer-Settler Contactor. Proceedings. International Solvent Extraction Conference ISEC71, The Hague, 1971, pp. 113-124. Rachuk.,W. Manual of Radiation Hazards Control at the University of B r i t i s h Columbia, 1980. Rice,N.M.. Recommended Nomenclature for Solvent Extraction (Liquid-Liquid D i s t r i b u t i o n ) . Hydrometallurgy, 7 (1981), pp. 177-199. Ritcey,G.M.. Crud in Solvent Extraction Processing-A Review of Causes and Treatment. Hydrometallurgy, 5 (1980), pp. 97-107. Ritcey,G.M., and B.H. Lucas. Diluents and Modifiers - Their Ef f e c t on Mass Transfer and Separation. In Proceedings of the International Solvent Extraction Conference, Lyon 1974 (London: Society of Chemical Industry, 1974), pp. 2437-2481. Ryon,A.D., F.L. Daley, R.S. Lowrie. Scale Up of Mixer-Settlers. Chemical Engineering Progress, v o l . 55, no. 10 (1959), pp. 70-75. Scuffham,J.B.. CMS: A Davy McKee Concept. The Chemical Engineer, July 1981, pp. 328-330. Shaw,D.J., Introduction to C o l l o i d and Surface Chemistry. Butterworth, London, 1980. Skelland,A.H.P., W.K. Blake, J.W. Dabrowski, J.A. U l r i c h , and T.F. Mach. Heat Transfer to Coils in Propellor Agitated Vessels. A.I.Ch.E. Journal, v o l . 11, no. 5, (1955), pp. 951-954. Sokal,R.R., and F.J. Rohlf. Biometry. Freeman, 1981. Stonner,H.M., and F. Wohler. An Engineer's Approach to a Solvent Extraction Problem. In: Hydrometallurgy. Davies,G.A., and J.B. Scuffham (eds.), London, I. Chem. E. Symp. Series, no. 42, 1975, pp. 14.1-14.11. Treybal,R.E.. Liquid Extraction, McGraw H i l l , New York, 147 1963. Urban,M.R.. Factors Aff e c t i n g the Formation of the Dispersed Phase in Liquid-Liquid Systems. M.Sc. Thesis, Johannesburg, University of the Witwatersrand, 1972. Warwick,G.I., J.B. Scuffham, and J.B. Lott. Considerations in the Design of Large Scale Solvent Extraction Plants for the Recovery of Metals. Proceedings. International Solvent Extraction Conference, The Hague ISEC71, 1971, pp. 1373-1385. Whewell,R.J., H.J. Foakes, and M.A. Hughes. Degradation in Hydroxyoxime Solvent Extraction Systems., Hydrometallurgy, 7 (1980), pp. 7-26. 1 48 APPENDIX I - PLOTS OF CHAPTER 7 1 49 Figure 14. Time for dispersion band thickness to achieve steady-state after introduction of dispersion into settler (HSLIX64N-copper phase system). Settler operating condition Time (min) Dispersion Band thickness (cm) 1. Organic continuous operation 0 0 O/A ratio=0.67 1 5. Qd/A=2.43 m**3/m**2/hr 3 12.5 Area=0.22 m**2 7 17.5 13 22. 18 24.5 23 26. 28 27.5 2. Organic continuous operation 0 0 O/A ratio=1.91 1 5.5 Qd/A=L91 m**3/m**2/hr 2 7.5 Area=0.22 m**2 4 9. 6 9.5 1 1 10.5 21 11.5 31 12.5 40 13.5 49 14. 3. Organic continuous operation 0 0 O/A ratio=1.5 2 10.5 Qd/A=2.90 m**3/m**2/hr 5 15.3 Area=0.22 m**2 6 18.5 9 23.5 13 27. 17 28.5 22 31 . 36 31 . 56 30. 4. Organic continuous operation 0 0 O/A ratio=1.5 4 1 Qd/A=0.77 m**3/m**2/hr 8 1 .5 Area=0.36 m**2 1 5 1 .4 20 1.5 5. Aqueous continuous operation 0 0 O/A ratio=0.67 9 4.5 Qd/A=2.44 m**3/m**2/hr 15 6.2 Area=0.13 m**2 22 7.2 30 7.3 37 7.4 6. Aqueous continuous operation 0 0 O/A ratio=1.5 10 12. Qd/A=3.22 m**3/m**2/hr 13 14.5 Area=0.22 m**2 18 14.5 25 16. 30 16. 38 16. Figure 18. Dispersion band thickness versus specific settler flow of dispersed phase (effect of O/A ratio and settler area), HSLIX64N-copper system, organic continuous operation, 24°C, 300rpm. est O/A Settler Organic Aqueous Qd/A AH no. ratio area flow flow (m**2) (1/min) (1/min) (m**3/m**2/hr) (cm) 1 0.67 .22 4.7 7.0 1 .91 10.8 2 0.67 .22 7.1 10.5 2.87 24. 3 0.67 .22 6.0 8.9 2.43 27. 4 0.67 .22 4.7 7.0 1.91 13.9 5 0.67 .22 4.5 6.8 1 .86 14.5 6 0.67 .22 5.5 8.0 2.20 25. 7 0.67 .22 4.9 7.2 1 .97 18. 8 0.67 .22 3.9 6.0 1 .64 14. 9 0.67 .36 8.7 13.0 2.15 13.3 10 0.67 .36 12.7 18.9 3.14 31 . 1 1 0.67 .36 4.8 7.1 1.18 4.9 12 0.67 .36 6.7 10.0 1 .66 10.5 1 3 0.67 .36 10.7 16.0 2.65 30.0 1 4 0.67 .36 4.5 6.7 1.11 7. 15 0.67 .36 5.7 8.5 1 .41 1 1 . 16 0.67 .36 7.9 11.8 1 .96 18. 17 1 .5 . 1 3 8.8 5.9 2.76 27. 18 1.5 .13 3.8 2.5 1.17 6.5 19 1.5 .22 9.9 6.6 1 .80 10.5 20 1.5 .22 14.4 9.6 2.62 19.5 21 1.5 .22 15.9 10.6 2.90 30.0 22 1.5 .22 3.8 2.5 0.68 3.5 23 1 .5 .22 7.1 4.7 1 .28 7.0 24 1 .5 .36 8.8 5.9 0.97 4.3 25 1.5 .36 19.5 13.0 2.15 12.0 26 1.5 .36 13.8 9.2 1 .52 8.0 27 1.5 .36 23.5 15.7 2.60 20.0 28 1 .5 .36 10.5 7.0 1.16 6.5 29 1 .5 .36 18.4 12.3 2.04 14.5 30 1.5 .36 7.2 4.9 0.82 4.2 31 1.5 .36 15.0 10.0 1 .66 12.0 32 1.5 .36 19.9 13.3 2.22 21 . 33 3.0 .13 13.8 4.6 2.16 19.5 34 3.0 .13 8.5 2.9 1 .35 9.0 35 3.0 .22 16.5 5.5 1 .50 8.8 36 3.0 .22 24.0 8.0 2.20 23.5 37 3.0 .22 12.0 4.2 1.15 6.0 38 3.0 .22 14.4 4.8 1 .31 9.0 39 3.0 .36 15.9 5.3 0.88 3.5 40 3.0 .36 20.0 6.6 1.10 5.4 41 3.0 .36 14.1 4.7 0.78 2.8 42 3.0 .36 24. 1 8.0 1 .33 6.3 43 3.0 .36 14.4 4.8 0.80 4.0 44 3.0 .36 18.0 6.0 1 .00 5.0 45 3.0 .36 21.0 7.0 1 .20 6.5 151 Figure 19. Dispersion band thickness versus specific settler flow of dispersed phase (effect of O/A ratio and settler area), HSLIX64N-copper system, aqueous continuous operation, 24°C, 300rpm. 'est O/A Settler Organic Aqueous Qd/A AH no. area flow flow (m**2) (1/min) (1/min) (m**3/m**2/hr) (cm 1 0.33 .13 4.6 13.9 2.17 6.9 2 0.33 .13 6.0 18.0 2.77 12. 3 0.33 .22 4.6 13.9 1 .26 2.8 4 0.33 .22 5.7 17.1 1 .55 5. 5 0.33 .22 6.6 19.5 1 .80 6. 6 0.33 .22 5.1 15.3 1 .39 3.5 7 0.33 .36 5.5 16.4 0.91 1 .8 8 0.33 .36 5.1 15.3 0.85 1 . 1 9 0.33 .36 6.2 18.6 1 .03 1 .5 10 0.33 .36 7.5 22.5 1 .25 2.5 1 1 0.33 .36 4.7 14.1 0.77 0.7 1 2 0.67 .13 4.4 6.6 2.03 4.0 1 3 0.67 .13 5.3 7.9 2.44 7.4 1 4 0.67 .13 6.1 9.1 2.81 10.0 1 5 0.67 .13 7.1 10.6 3.28 14.7 1 6 0.67 .13 7.1 10.6 3.28 15.0 17 0.67 .13 8.0 12.0 3.69 21.0 18 0.67 .22 4.9 7.3 1 .33 3.3 1 9 0.67 .22 6.0 9.0 1 .64 3.5 20 0.67 .22 8.0 1 1 .9 2.18 6.3 21 0.67 .22 10.1 15.1 2.75 12.5 22 0.67 .22 11.5 17.2 3.14 16. 23 0.67 .22 13.3 19.8 3.64 25. 24 0.67 .22 14.7 22. 1 4.01 32.5 25 0.67 .36 4.5 6.7 0.81 1 . 1 26 0.67 .36 5.8 8.6 0.95 1 .7 27 0.67 .36 7.2 10.7 1.18 3.1 28 0.67 .36 8.8 13.1 1 .48 3.6 29 0.67 .36 13.2 19.7 2.22 6.3 30 1 .5 .13 4.2 2.8 1 .97 5.5 31 1 .5 .13 6.6 4.4 3.05 10. 32 1 .5 .13 7.0 4.7 3.23 12.5 33 1 .5 .13 7.6 5.1 3.52 21.5 34 1 .5 .13 8.0 5.3 3.69 19.0 35 1 .5 .13 9.5 6.3 4.39 33. 36 1.5 .22 12.6 8.4 3.43 25. 37 1 .5 .22 4.2 2.8 1.15 2.8 38 1 .5 .22 7.0 4.7 1 .92 5.7 39 1 .5 .22 8.8 5.7 2.35 7.0 40 1 .5 .22 10.0 6.7 2.73 13. 41 1 .5 .22 12.0 7.9 3.27 19.5 42 1 .5 .36 4.6 3.2 0.85 1 .5 43 1 .5 .36 7.2 4.8 1.21 4.4 44 1 .5 .36 16.0 11.0 2.71 12.5 45 1 .5 .36 7.3 4.9 1 .21 3.2 46 1 .5 .36 5.1 3.4 0.85 1 .5 47 1 .5 .36 4.2 2.8 0.70 .8 48 1 .5 .36 9.4 6.3 1 .56 3.7 49 1 .5 .36 13.1 8.7 2.18 8.5 50 1 .5 .36 10.6 7.1 1 .77 5.0 152 Figure 20. Dispersion band thickness versus specific settler flow of dispersed phase (effect of O/A ratio and settler area), Key Lake phase system, organic continuous operation, 35°C, 300rpm. 'est O/A Settler Organic Aqueous Qd/A AH no. area flow flow (m**2) (l/min) (l/min) (m**3/m**2/hr) (cm) 1 .67 .13 4.0 6.1 2.95 24.5 2 .67 .13 4.6 6.8 3.19 32. 3 .67 .22 5.3 7.8 2.13 8. 4 .67 .22 4.8 7.3 1 .99 8. 5 .67 .22 5.7 8.6 2.35 13. 6 .67 .22 6.3 9.4 2.56 15.5 7 .67 .22 7.0 10.5 2.86 19.8 8 .67 .22 8.1 12.1 3.30 26.5 9 .67 .36 9.3 14.0 2.33 15. 10 .67 .36 11.3 17.0 2.83 18. 1 1 .67 .36 12.3 18.7 3.12 28. 12 .67 .36 6.8 10.2 1 .75 5. 13 .67 .36 8.7 13.0 2.17 9.5 1 4 .67 .36 5.2 7.8 1 .30 '3. 15 1 .5 .13 7.5 5.0 2.34 12. 16 1 .5 . 1 3 8.1 5.4 2.53 14.5 1 7 1 .5 . 1 3 9.1 6.3 2.95 27. 18 1 .5 .13 6.8 4.5 2.10 9.5 19 1 .5 .13 8.6 5.7 2.67 19.5 20 1 .5 .22 12.0 8.0 2. 18 1 1 . 21 1 .5 .22 7.9 5.3 1 .45 3.7 22 1 .5 .22 9.1 6.0 1 .64 5.5 23 1.5 .22 15.0 10.0 2.73 20. 24 1 .5 .22 7.1 4.7 1 .28 3.3 25 1 .5 .22 16.5 11.0 3.00 30. 26 1 .5 .36 8.4 5.6 0.93 2.0 27 1 .5 .36 14.2 9.0 1 .50 5.2 28 1 .5 .36 20.2 13.5 2.25 13. 29 1.5 .36 22.0 14.7 2.45 16. 30 1.5 .36 13.0 8.0 1 .33 4. 31 1.5 .36 17.4 11.6 1 .93 8.5 32 3.0 .13 13.7 4.6 2.11 17. 33 3.0 .13 15.9 5.3 2.48 27. 34 3.0 . 1 3 15.0 5.0 2.36 22. 35 3.0 .13 15.5 5.2 2.39 20.5 36 3.0 . 1 3 17.3 5.8 2.67 31 . 37 3.0 .22 13.5 4.5 1 .23 4.5 38 3.0 .22 15.9 5.3 1 .45 7. 39 3.0 .22 18.0 6.0 1 .64 9.5 40 3.0 .22 21.3 7.1 1 .94 18. 41 3.0 .22 24.0 8.0 2.18 20. 42 3.0 .36 19.2 6.4 1 .07 3. 43 3.0 .36 21 .6 7.3 1.22 5. 44 3.0 .36 13.5 4.5 0.75 1 .7 45 3.0 .36 17.4 5.8 0.97 2.2 1 53 Figure 21. Dispersion band thickness versus specific settler flow of dispersed phase (effect of operating temperature), HSLIX64N-copper phase system, organic continuous operation, 300rpm. Test O/A Settler Organic Aqueous Qd/A AH no. ratio area flow flow (m**2) (1/min) (1/min) (m**3/m**2/hr) (cm) Temperature = 24°C: 1 1.5 .36 8.8 5.9 0.97 4.3 2 1.5 .36 19.5 13.0 2.15 12.0 3 1.5 .36 13.8 9.2 1 .52 8.0 4 1 .5 .36 23.5 15.7 2.60 20.0 5 1.5 .36 10.5 7.0 1.16 6.5 6 1.5 .36 18.4 12.3 2.04 14.5 7 1 .5 .36 7.2 4.9 0.82 4.2 8 1.5 .36 15.0 10.0 1 .66 12.0 9 1 .5 .36 19.9 13.3 2.22 21 . 10 1 .5 .22 9.9 6.6 1 .80 10.5 1 1 1.5 .22 14.4 9.6 2.62 19.5 12 1.5 .22 15.9 10.6 2.90 30.0 13 1.5 .22 3.8 2.5 0.68 3.5 1 4 1 .5 .22 7.1 4.7 1 .28 7.0 15 1 .5 . 1 3 8.8 5.9 2.76 27. 16 1 .5 . 1 3 3.8 2.5 1.17 6.5 Temperature = 35°C: 17 1 .5 .13 6.9 4.6 2.16 6.8 18 1 .5 .13 10.1 6.7 3.14 16.5 19 1 .5 .13 12.0 8.0 3.75 28.5 20 1 .5 .22 6.9 4.6 1 .26 4.1 21 1 .5 .22 10.1 6.7 1 .83 6.1 22 1 .5 .22 13.4 8.9 2.43 10.5 23 1 .5 .22 17.7 11.8 3.22 16.5 24 1 .5 .36 12.0 8.0 1 .33 4.8 25 1 .5 .36 16.5 11.0 1 .82 6.6 26 1 .5 .36 20.1 13.9 2.31 9.4 27 1 .5 .36 7.5 5.0 0.83 2. 1 28 1 .5 .36 6.9 4.6 0.77 1.5 1 54 Figure 22. Dispersion band thickness versus s p e c i f i c s e t t l e r flow of dispersed phase (effect of impellor speed), HSLIX64N-copper phase system, organic continuous operation, 24°C. Test Impellor O/A Settler Organic Aqueous Qd/A AH no. speed rat io area flow flow (rpm) (m**2) (l/min) (l/min) (m**3/m**2/hr) (cm) 1 250 1 .5 .36 7.5 5.0 0.83 3.5 2 250 1 .5 .36 12.0 8.0 1 .32 6.5 3 250 1 .5 .36 13.6 9.0 1 .95 11.5 4 250 1.5 .36 15.1 10.0 2.15 12.5 5 300 1 .5 .36 8.8 5.9 0.97 4.3 6 300 1 .5 .36 19.5 13.0 2.15 12.0 7 300 1 .5 .36 13.8 9.2 1 .52 8.0 8 300 1 .5 .36 23.5 15.7 2.60 20.0 9 300 1 .5 .36 10.5 7.0 1.16 6.5 10 300 1 .5 .36 18.4 12.3 2.04 14.5 1 1 350 1 .5 .36 10.1 6.7 1.11 5.7 1 2 350 1 .5 .36 15.7 10.4 1 .75 11.0 13 350 1 .5 .36 22.5 15.0 2.48 19.5 1 4 350 1 .5 .36 13.8 9.2 1 .52 8.5 1 55 Figure 23. Dispersion band thickness versus specific settler flow of dispersed phase (effect of impellor speed), HSLIX64N-copper phase system, aqueous continuous operation, 24°C. Test ImpellorO/A Settler Organic Aqueous Qd/A AH no. speed ratio area flow flow (rpm) (m**2) (1/min) (1/min) (m**3/m**2/hr) (cm) 1 250 0.67 .22 4.4 6.6 1.19 3.0 2 250 0.67 .22 7.6 11.3 2.07 7.5 3 250 0.67 .22 10.5 15.7 2.86 12.0 4 250 0.67 .22 14.7 22.1 4.01 26.0 5 300 0.67 .22 4.9 7.3> 1 .33 1.3 6 300 0.67 .22 6.0 9.0 1 .64 3.5 7 300 0.67 .22 8.0 1 1.9 2.18 6.3 8 300 0.67 .22 10. 1 15.1 2.75 12.5 9 300 0.67 .22 11.5 17.2 3.14 16.0 10 300 0.67 .22 13.3 19.8 3.64 25.0 1 1 300 0.67 .22 14.7 22.1 4.01 32.5 12 350 0.67 .22 6.2 9.3 1 .70 4.3 1.3 350 0.67 .22 8.2 12.2 2.23 7.5 14 350 0.67 .22 11.0 16.4 3.00 15.0 1 5 350 0.67 .22 14.7 22.1 4.01 24.5 156 Figure 24. Dispersion band thickness versus specific settler flow of dispersed phase (effect of impellor speed), Key Lake phase system, organic continuous operation, 35°C. 'est Impellor O/A Settler Organic Aqueous Qd/A AH no. speed ratio area flow flow (rpm) (m**2) (l/min) (l/min) (m**3/m**2/hr) (cm) 1 250 1.5 .22 12.8 8.5 2.32 1 1 . 2 250 1.5 .22 12.0 8.0 2.18 10. 3 250 1 .5 .22 7.0 4.7 1 .28 3.3 4 250 1.5 .22 15.1 10.1 2.80 21 . 5 300 1 .5 .22 12.0 8.0 2.18 1 1 . 6 300 1.5 .22 7.9 5.3 1 .45 3.7 7 300 1.5 .22 9.1 6.0 1 .64 6. 8 300 1.5 .22 9.1 6.0 1 .64 5.5 9 300 1.5 .22 15.0 10.0 2.73 20. 10 300 1 .5 .22 7.0 4.7 1 .28 3.3 1 1 300 1 .5 .22 16.1 10.8 3.00 30. 12 350 1.5 .22 7.4 5.0 1 .36 3.5 13 350 1.5 .22 15.9 9.9 2.68 19.5 14 350 1 .5 .22 13.3 8.9 2.45 16. 157 Figure 25. Holdup profiles within dispersion band, HSLIX64N-copper phase system, organic continuous operation, 24°C, 300rpm. Operating condition Distance, h, of h/AH Local sampling point holdup from coalescence front (cm) Mixer holdup=0 .3 2 .5 0 . 1 0 .75 H=26 cm 5 .0 0 . 2 0 .70 Sett ler area=.: 36 m**2 7 .5 0 .29 0 .65 10 .0 0 .39 0 .60 12 .5 0 . 49 0 . 58 13 .0 0 .5 0 .60 15 .5 0 .59 0 .54 16 .0 0 .62 0 .55 18 .0 0 .69 0 . 40 20 .5 • 0 .78 0 .37 23 .5 0 .9 0 . 34 25 .5 0 .98 0 .31 Mixer holdup=0 .4 2 .5 0 .08 0 . 77 H=30 cm 5 .0 0 . 17 0 .70 Sett ler area=0 . 22 m**2 5 .0 0 . 17 0 . 75 7 .5 0 .25 0 .67 10 .0 0 .33 0 .62 10 .0 0 .33 0 .64 12 .5 0 .42 0 .59 15 .0 0 .50 0 .58 17 .5 0 .58 0 .56 20 .0 0 .66 0 .45 22 .5 0 .75 0 .47 25 .0 0 .83 0 .41 27 .5 0 .92 0 .41 Mixer holdup=0 .5 2 .5 0 . 10 0 .82 H=26 cm 5 .0 0 . 19 0 .74 Sett ler area=0 .36 m**2 7 . 5 0 .29 0 .67 10 .0 0 .39 0 .63 12 .5 0 .48 0 .60 15 .0 o . 58 0 .56 17 .5 0 .67 0 .53 20 .0 0 .77 0 .51 22 .5 0 .87 0 .50 25 .0 0 .96 0 .22 Mixer holdup=0 .6 2 .5 0 .09 0 .81 H=27 cm 4 .0 0 . 15 0 . 77 Sett ler area=0 .22 m**2 5 .0 0 . 19 0 .72 7 . 5 0 .27 0. .65 9 .0 0. .33 0 .66 10 .0 0. . 37 0. .65 11 . .0 0. .41 0. .61 12. .5 0. .46 0. ,59 15. .0 0. .55 0. .58 17 .5 0. .65 0. .56 20. .0 0. . 74 0. 56 22. .5 0. .83 0. .58 25. .0 0. 93 0. .57 Mixer holdup=0, , 7 2 .5 0. .09 0. .83 H=28 cm 5. .0 0. . 18 0. 81 Sett ler area=0, 22 m**2 7 .5 0. 27 0. . 76 g. .0 0. 33 0. 69 10. .0 0. . 36 0. 67 11. ,0 0. 40 0. 70 12. 5 0. 45 0. 70 15. 0 0. 54 0. 71 18 . 0 0. 65 0. 70 20. 0 0. 71 0. 70 22. 5 0. 80 0. 71 25. 0 0. 89 0. 66 26. 5 0. 95 0. 68 Mixer holdup=0. 8 1 . 0 0. 04 0. 92 H=25.5 cm 2. 5 0. 10 0. 88 Sett ler area=0. 22 m**2 5. 0 0. 20 0. 85 7. 5 0. 29 0. 84 10. 0 0. 39 0. 83 12. 5 0. 49 0. 82 15. 0 0. 59 0. 79 17. 5 0. 69 0. 78 22. 5 0. 88 0. 76 24 . 0 0. 94 0. 75 158 Figure 26. Holdup profiles within dispersion band, HSLIX64N-copper system, aqueous continuous operation, 24°C, 300rpm. Operating cond i t ion Distance, h, of h/AH Local sampling point holdup from coalescence front (cm) 1. Mixer holdup=0.32 H=33 cm S e t t l e r area=.13 m**2 2. Mixer holdup=0.40 H=33 cm S e t t l e r area=.22 m**2 3. Mixer holdup=0.52 H=33 cm S e t t e l r area=.22 m**2 4. Mixer holdup=0.60 H=33 cm S e t t l e r area=.13 m**2 5. Mixer holdup=0.70 H=30 cm S e t t l e r area=.22 m**2 .5 0. .02 0. ,53 3, .0 0. .09 0. 38 5, .5 0. . 17 0. 37 8 .0 0. .24 0. ,35 10. .5 0. ,32 0. 38 15, .5 0. ,47 0. ,35 16, .0 0. ,49 0. ,35 18 .0 0. .55 0. ,35 20, .5 0. 62 0. 35 23 .0 0. 70 0. 36 25, .5 0. 77 0. 35 28. .0 0. 85 0. 34 30, .5 0. 92 0. 34 3, .0 0. ,09 0. 60 5, .5 0. , 16 0. 50 8 .0 0. 24 0. 43 10. .5 0. 32 0. 45 13, .0 0. ,39 0. ,44 15, ,5 0. ,47 0. 44 20, ,5 0. .62 0. 44 21 , .5 0. .65 0. 43 23 .0 0. ,70 0. ,43 25. .5 0. , 77 0. 43 28 .0 0. 85 0. 44 30, . 5 0. .92 0. 42 0, .5 0. .02 0. .56 3, .0 0. ,09 0. 63 5, .5 0. . 17 0. 54 8 .0 0. ,24 0. 57 10, . 5 0. ,32 0. 53 13, .0 0. 39 0. 53 15 . 5 0. ,47 0. .53 20. .5 0. ,62 0. 53 23. .0 0. ,70 0. 54 25 . , 5 0. 77 0. 54 28. .0 0. 85 0. 53 30. ,5 0. 92 0. 52 3. .0 0. 09 0. 59 3. .0 0. 09 0. 62 5. ,5 0. 16 0. 62 10. .5 0. 32 0. 62 1 1 . ,0 0. 34 0. 61 13. .0 0. 39 0. 62 15. .5 0. 47 0. 62 18 . 0 0. 55 0. 61 20. ,5 0. 62 0. 61 23. .0 0. 70 0. 58 25. .5 0. 77 0. 61 28 . 0 0. 85 o. 60 30. 5 0. 92 0. 59 2. ,5 0. 08 0. 69 5. .0 0. 17 0. 71 7. ,5 0. 25 0. 70 10. .0 0. 33 0. 68 12 .5 0 .42 0. ,68 15 .0 0 .50 0. .68 18 . 5 0 .62 0 .69 20 .0 0 .66 0 .70 22 .5 0 .75 0 .68 25 .0 0 .83 0 .69 1 59 Figure 27. Holdup profiles within dispersion band, Key Lake phase system, organic continuous operation, 35°C, 300rpm. Operating cond i t ion Distance, h, of h/AH Local sampling point holdup from coalescence f ront (cm) Mixer holdup=0.29 2. .5 0. 09 0. 59 H=29 cm 5, O O. 17 0. 51 S e t t l e r area=.22 m**2 7 , .5 0. 26 0. 33 10. .0 0. 35 0. 30 15. .0 0. 52 0. 28 20. .0 0. 69 0. 28 25. .0 0. 86 0. 28 27. .0 0. 93 0. 25 Mixer holdup=0.41 2 .5 0. 08 0. 52 H=30 cm 5, .0 0. 17 0. 49 S e t t l e r area=.22 m**2 7. .5 0. 25 0. 41 10 .0 0. 33 0. 38 12 .5 0. 42 0. .41 15 .0 0. 50 0. ,40 17, .5 0. 58 0. 39 20 .0 0. S7 0 39 25 .0 0. 83 0. . 38 Mixer holdup=0.49 2 .5 0. 10 0. 67 H=25 cm 5 .0 0. 20 0. .53 S e t t l e r area=.22 m**2 7, .5 0. 30 0. 51 10 .0 0. 40 0. 46 12, .5 0. 50 0. ,47 15 .0 o. 60 0. ,45 17 .5 0. ,70 0. ,42 20 .0 0. ,80 0. .43 22 .5 0. 90 0. , 40 Mixer holdup=0.60 2 .5 0. ,09 0. ,94 H=27 cm 5 .0 0. , 19 0. .67 S e t t l e r area=.22 m**2 7. .5 0. 28 0, .60 10 .0 0. , 37 0. .59 12 .5 0. .46 0, .58 15 .0 0, ,56 0. 58 17 .5 0. 65 0 58 20 .0 0. 74 0. .57 22 .5 o. 83 0. 56 25 .0 0. 93 0. .55 1 60 Figure 28. Change in arithmetic mean drop size of dispersion sample with time after withdrawal from mixer, Key Lake phase system, organic continuous operation, 35°C, 300rpm. Sample Operating Time after No. of Mean drop Standard no. condition withdrawal drops diameter deviation of sample in (urn) (s) sample Arith. Vol. Arith. Vol 1. Qd=l9.1 1/min 10 40 373 399 100 129 O/A =0.67 15 33 456 512 166 190 20 35 445 509 167 254 30 23 618 766 325 372 2. Qd=l3.2 1/min 15 40 372 412 127 145 O/A =0.67 20 38 410 469 165 168 25 45 339 400 147 221 30 37 460 559 221 305 3. Qd=14.0 1/min 15 39 395 410 78 90 O/A = 1.5 35 35 448 467 95 102 50 32 456 499 141 173 4. Qd=5.0 1/min 1 5 35 412 423 71 71 O/A = 1.5 30 34 396 415 93 85 40 34 487 51 1 1 10 130 50 31 471 504 127 1 56 5. Qd=5.0 1/min 15 51 284 308 92 70 O/A = 3.0 25 55 302 329 94 98 35 57 289 310 85 66 161 Figure 30. Effect of O/A ratio and dispersed phase flowrate on arithmetic mean drop diameter at mixer overflow, Key Lake phase system, organic continuous operation, 35°C, 300rpm. O/A Qd NO. Of Mean drop Standard ratio. (l/min) drops diameter deviation in (urn) sample Arith. Vol. Arith. Vol. 0.67 13.2 40 372 412 127 144 0.67 19.1 39 427 481 155 202 0.67 19.1 40 373 399 100 130 0.67 11.5 42 365 401 123 155 1 .5 5.0 35 412 423 71 71 1 .5 9.5 42 394 412 86 95 1 .5 14.0 39 392 410 78 90 3.0 5.0 51 284 308 92 70 3.0 5.0 60 253 292 100 105 3.0 5.2 57 273 296 102 79 1 6 2 Figure 32. Effect of impellor speed on arithmetic mean drop diameter at mixer overflow, Key Lake phase system, organic continuous operation, 35 0* Operating Impellor No. of Mean drop Standard condition Speed drops diameter deviation (rpm) in (urn) sample Arith. Vol. Arith. Vol. 1 . Qd=13.2 l/min 250 32 470 540 184 213 O/A =0.67 300 40 372 412 127 144 350 42 351 364 99 108 2 . Qd=9.5 l/min 250 40 416 435 104 140 O/A =1.5 300 42 394 412 86 95 350 37 331 348 79 74 3 . Qd=l3.6 l/min 250 46 258 287 94 84 O/A = 3.0 300 51 284 308 92 70 300 60 253 392 100 175 350 42 260 297 103 121 1 63 Figure 33. Drop size profile within dispersion band for O/A=0.67, Key Lake phase system, organic continuous operation, 35°C, 300rpm. Operating conditions: Qd/A=3.14 m**3/m**2/hr Dispersion band thickness (AH)=20 cm Area=0.22 m**2 Organic flow=7.8 1/min Aqueous flow=11.5 1/min S a m p l e D i s t a n c e h/AH No. of Mean drop Standard location from drops diameter deviation coalescence in (urn) front (cm) sample Arith. Vol. Arith. Vol. 1 2 .10 25 472 609 274 321 1 5 .25 36 374 448 171 230 1 1 1 .55 29 417 499 1 97 219 1 1 7 .85 29 371 406 1 20 1 1 7 2 2 .10 28 517 528 74 79 2 5 .25 26 451 469 90 101 2 1 1 .55 29 415 459 141 156 2 14 .70 30 422 449 148 161 2 1 7 .85 32 380 438 140 305 Note on sample location: 1. 1/3 distance between settler inlet and 2. 10 cm from end of settler end 164 Figure 34. Drop size p r o f i l e within dispersion band for 0/A=1.5, Key Lake phase system, organic continuous operation, 35°C, 300rpm. Operating conditions: Qd/A=2.50 m**3/m**2/hr Dispersion band thickness (AH) =16 cm Area=0.22 m**2 Organic flow=13.7 l/min Aqueous flow=9.2 l/min mple Distance h/AH No. of Mean drop Standard at ion from drops in diameter deviation coalescence (urn) front,h (cm) sample A r i t h . Vol. A r i t h . Vol 1 3.0 .20 31 485 528 144 201 1 5.5 .35 33 417 429 691 81 1 6.5 .40 47 344 419 1 66 221 1 6.5 .40 40 380 408 1 10 106 1 9.5 • .60 44 354 393 1 23 1 35 1 13.0 .80 46 307 347 1 18 125 1 13.0 .80 43 345 367 89 96 1 14.5 .90 38 380 390 69 67 2 1 .5 . 1 0 19 641 716 231 241 2 3.0 .20 38 447 470 102 119 2 6.5 .40 34 381 447 1 54 291 2 10.5 .65 39 396 415 86 102 2 13.0 .80 46 313 338 94 90 2 14.5 .90 36 383 395 66 71 2 15.0 .95 45 284 291 46 44 Note on sample locati o n : 1. 1/3 distance between s e t t l e r i n l e t and end 2. 10 cm from end of s e t t l e r 165 Figure 35. Drop size p r o f i l e within dispersion band for O/A=3.0, Key Lake phase system, organic continuous operation, 35°C. Operating conditions: Qd/A= 2.40 m**3/m**2/hr Dispersion band thickness (AH) =23 cm Area=0.13 m**2 Organic flow=l5.6 1/min Aqueous flow= 5.2 1/min S a m p l e D i s t a n c e h / A H N o . of Mean d r o p S t a n d a r d location from drops diameter deviation coalescence in (urn) front,h (cm) sample A r i t h . Vol. A r i t h . Vol 1 2 .09 23 466 508 150 135 1 5 .22 45 300 321 84 78 1 10 .43 43 262 275 60 61 1 15 .65 38 325 357 105 1 1 3 1 20 .87 48 304 336 101 118 2 1 .05 28 472 541 182 253 2 5 .22 41 361 380 86 84 2 10 .43 42 280 301 80 87 2 1 5 .65 38 284 305 77 88 2 20 .87 41 309 342 1 07 104 Note on sample locati o n : 1. 1/3 distance between s e t t l e r i n l e t and end 2. 10 cm from end of s e t t l e r 166 APPENDIX II - CHARACTERISTIC CURVE STATISTICAL ANALYSIS 167 Part 1: k and y values for the c h a r a c t e r i s t i c curves Values of the empirical constants, k and y, for the ch a r a c t e r i s t i c curves of figures 18-24 were calculated using linear regression and are tabulated below: Figure Case k y 18 O/A=0.67 5.16 1 .64 0/A=1.5 5.17 1 .47 O/A=3.0 4.77 1 .84 a l l data pts. 5.01 1 .62 19 O/A=0.33 1 .61 2.08 O/A=0.67 1 .65 1.91 0/A=1.5 1 .95 1.81 a l l data pts. 1 .73 1 .90 20 O/A=0.67 1 .34 2.61 . 0/A=1.5 1 .92 2.32 O/A=3.0 2.86 2.42 21 T=24 °C 5. 17 1 .47 T=35 °C 2.54 1 .63 22 a l l data pts. 4.68 1 .45 23 a l l data pts. 1 .29 2.21 24 a l l data pts. 1 .64 2.48 1 68 Part 2: s t a t i s t i c a l analysis of c h a r a c t e r i s t i c curves In order to determine whether the functional relationships described by two or more regression equations are the same, regression l i n e s were compared using an analysis of covariance procedure. A preliminary step in t h i s procedure i s testing the two (or more) l i n e s for homogeneity of slope. The procedure i s described f u l l y elsewhere (Sokal and Rohlf, 1981) and only the r e s u l t s are summarized here, in the form of an anova table. An F r a t i o (the r a t i o of one mean square, MS, to another) i s used to determine the p r o b a b i l i t y of two variances being from the same population. The f i n a l step i s an analysis of covariance (ancova) to test for homogeneity of the intercepts of the regression l i n e s . This step requires that the slopes of the regression l i n e s be equal, as determined in the f i r s t step. An F r a t i o of the appropriate mean square values i s calculated and can be used to determine whether the assumption of a single regression l i n e for two sets of data may be accepted or must be rejected. Comparisons (between the three c h a r a c t e r i s t i c curves, taken at three di f f e r e n t O/A ratios) were made for; 1. Figure 18 - HSLIX64N-copper phase system, organic continuous. 169 2. Figure 19 - HSLIX64N-copper phase system, aqueous continuous. 3. Figure 20 - Key Lake phase system, organic continuous. In this way, i t could be determined whether the three characteristic curves for each figure are s t a t i s t i c a l l y the same (to a 99% level of confidence). The f i n a l anova and ancova tables for the s t a t i s t i c a l analysis follow. 1. Figure 18 - HSLIX64N-copper phase system, org. continuous. a) Comparison of O/A=0.67 and 0/A=1.5. Anova table (for comparison of slopes) source of variation degree of sum of mean F ratio freedom squares square df SS MS variation among regressions weighted average of deviations from regression 1 .00536 .00536 .8134 28 .18448 .00659 F 01(1 28) = 7 * ^ 4 , s * n c e F < F 0 1 ( 1 28) w e c o n c ^ u c ^ e that the two characteristic curves have the same slope. Ancova table (for comparison of intercepts) source of variation df SS MS F adjusted means error (deviations from a common slope) 1 .01015 .01015 1.551 29 .18984 .00655 F.01 (1 ,29) = 7 - 6 0 « Since F<F 2g) w e c o n d u d e that the two characteristic curves have the same intercept. 170 b) Comparison of O/A=0.67 and O/A=3.0. Anova table (for comparison of slopes) source of variation df SS MS F variation among regressions weighted average of deviations from regression 1 . 0 0 4 8 3 . 0 0 4 8 3 . 7 6 0 6 2 5 . 1 5 8 8 6 . 0 0 6 3 5 F r\,t, O C A = 7 ' 7 7 ' Since F<F n. . „ . we conclude that the two .U1 l 1 r £9 J . U l ( 1 ,25) characteristic curves have the same slope. Ancova table (for comparison of intercepts) source of variation df SS MS F adjusted means error (deviations from a common slope) 1 . 0 0 0 0 1 3 . 0 0 0 0 1 3 0 . 0 0 2 2 6 . 1 6 3 6 9 . 0 0 6 3 0 F 0 1 ( 1 2 6 ) = 7 " 7 2 * S * N C E F < F Q 1 (1 2 6 ) W E C O N C L U D E T N A T T N E T W O characteristic curves have the same intercept. * Comparison of 0/A= 1 . 5 and O/A=3.0. Anova table (for comparison of slopes) source of variation df SS MS F variation among regressions weighted average of deviations from regression 1 . 0 0 2 4 4 . 0 0 2 4 4 6 . 0 8 2 2 5 . 1 0 0 2 4 . 0 0 4 0 1 F n , / . o c \ = 7 « 7 7 « Since F<F rt</. we conclude that the two . U l v l , i s ; • « ! 1 i , 2 5 ) characteristic curves have the same slope. Ancova table (for comparison of intercepts) source of variation df SS MS F adjusted means error (deviations from a common slope) 1 . 0 0 8 3 9 . 0 0 8 3 9 0 . 1 7 5 2 6 . 1 2 4 6 3 . 0 0 4 7 9 171 F.01(1,26) = 7' 7 2- S i n c e F < F . 01(1,26) we conclude that the two characteristic curves have the same intercept. 2. Fiqure 19 - HSLIX64N-copper phase system, aq. continuous. Comparison of O/A=0.33 and O/A=0.67 Anova table (for comparison of slopes) source of variation df SS MS F variation among regressions 1 .00656 .00656 .8482 weighted average of deviations from regression 25 .19330 .00773 F.01(1,25) = 7- 7 7- S i n C C F < F . 01(1,25) we conclude that the two characteristic curves have the same slope. Ancova table (for comparison of intercepts) source of variation df SS MS F adjusted means 1 .00286 .00286 0.372 error (deviations from a common slope) 26 .19986 .00769 F01(1,26) = 7' 7 2- S i n c e F < F . 01(1,26) we conclude that the two characteristic curves have the same intercept. ) Comparison of O/A=0.33 and 0/A=1.5. Anova table (for comparison of slopes) source of variation df SS MS F variation among regressions 1 .01777 .01777 2.235 weighted average of deviations from regression 28 .22267 .00795 1 72 F01(1 28) = 7 # ^ 4 * S i . n c e F<F ^  ^ ^gj w e conclude that the two characteristic curves have the same slope. Ancova table (for comparison of intercepts) source of variation df SS MS F adjusted means error (deviations from a common slope) 1 .00866 .00866 1.551 29 .24032 .00829 F01(1 29) = 7 , ^ u * s i n c e F < F 01(1 29) w e c o n c ^ u c ^ e that the two characteristic curves have the same intercept. Comparison of O/A=0.67 and 0/A=1.5. Anova table (for comparison of slopes) source of variation df SS MS F variation among regressions weighted average of deviations from regression 1 .00432 .00432 .5058 35 .30045 .00858 F Q 1 ^ 1 3 5 ) = 7 , 3 1 , S i n c e F < F 01(1 35) w e c o n c l u d e that the two characteristic curves have the same slope. Ancova table (for comparison of intercepts) source of variation df SS MS F adjusted means error (deviations from a common slope) 1 .01624 .01624 1.918 36 .30479 .00847 F «^\ = 7«31. Since F<F n, . „. we conclude that the two .01(1,36) .01(1,36) characteristic curves have the same intercept. 173 3. Figure 20 - Key Lake phase system, org, continuous. > Comparison of O/A=0.67 and 0/A=1.5. Anova table (for comparison of slopes) source of variation df SS MS F variation among regressions weighted average of deviations from regression 1 .00995 .00995 3.478 27 .07725 .00286 F Q 1^ 1 27) = 7 « 6 8 . Since F < F # o i ( i , 2 7 ) w e conclude that the two characteristic curves have the same slope. Ancova table (for comparison of intercepts) source of variation df SS MS F adjusted means error (deviations from a common slope) 1 .01976 .01976 6.346 28 .08720 .00311 F o o \ = 7 . 6 4 . Since F<F , „ we conclude that the two •01(1,28) .01(1,28) characteristic curves have the same intercept. b) Comparison of O/A=0.67 and O/A=3.0. Anova table (for comparison of slopes) source of variation df SS MS F variation among 1 .00449 .00449 1 .604 regressions weighted average of 24 .06721 .00280 deviations from regression F « W i <->>i\=7.82. Since F<F n, , , 0.\ we conclude that the two .01(1,24) .01(1,24) characteristic curves have the same slope. 174 Ancova table (for comparison of intercepts) source of variation df SS MS F adjusted means error (deviations from a common slope) 1 .37302 .37302 130.6 25 .07140 .00286 F 0 1 ( 1 , 2 5 ) = 7 , 7 7 , Since F > F Q 1 ( 1 2 5 ) w e conclude that the two characteristic curves have different intercepts. Comparison of 0/A=1.5 and O/A=3.0. Anova table (for comparison of slopes) source of variation df SS MS F variation among regressions weighted average of deviations from regression 1 .00227 .00227 .9154 30 .07446 .00248 F01(1 30) = 7 > 5 6 * Since F<F . 0 1 ( 1 , 3 0 ) w e conclude that the two characteristic curves have the same slope. Ancova table (for comparison of intercepts) source of variation df SS MS F adjusted means error (deviations from a common slope) 1 .29535 .29535 116.7 28 .07071 .00253 F 0 1 ( 1 2 8 ) = 7 * ^ 4 * S * N C E F > F 0 1 ( 1 2 8 ) W E C O N C ^ - U D E T N A T T H E T W O characteristic curves have different intercepts. 1 75 Conclusions of s t a t i s t i c a l analysis. The s t a t i s t i c a l analysis shows that for figures 18 and 19 the characteristic curves for the three O/A ratios are the same. However, for figure 20, although data for O/A=0.67 and 0/A=1.5 could be f i t t e d with a single characteristic curve, a separate ch a r a c t e r i s t i c curve existed for O/A=3.0. 176 APPENDIX III - ERROR ANALYSIS; GENERAL COMMENTS 177 The accuracy with which AH was measured depended on several factors. When the dispersion feed was d i l u t e , the s e t t l i n g front was not as sharp as when a concentrated dispersion was used. In addition, there was some cyclin g of the dispersion band thickness with time, es p e c i a l l y with thicker dispersion bands. Because of these factors, the error in a AH measurement was estimated to vary from a few millimeters (concentrated dispersions with thin dispersion bands) to roughly 2 cm (dilute dispersions with thick dispersion bands). An additional error source was in determining the flowrate of the phases, since some fluctuations in flowrate occurred and because inaccuracies in reading the rotameters would be expected. When a par t i c u l a r run was repeated several hours after i t had i n i t i a l l y been performed, the values of AH recorded t y p i c a l l y agreed within 5 to 10%. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0058830/manifest

Comment

Related Items