Browsing by Author "Lewis, Alison"
Now showing 1 - 11 of 11
Results Per Page
Sort Options
- ItemRestrictedAn investigation into the precipitation of Fe(III) oxyhydroxides in lime neutralization processes(2019) Mangunda, Cledwyn Tawanda; Lewis, Alison; Petersen, JoachimAcid Mine Drainage (AMD) and hydrometallurgical wastewater are products of extractive activities that pose environmental risks as they contain heavy metals. The contents of these streams, which is dominated by iron, necessitates an effective treatment protocol. The tendency of iron to precipitate as a sludge has drawn attention to the mechanisms and the associated kinetics of Fe(III) oxyhydroxide precipitation and the post precipitation solidliquid separation processes within the high density sludge (HDS) treatment process. The HDS process is affordable to operate and has the potential benefit of the Fe(III) oxyhydroxide precipitates removing other pollutants through coprecipitation and adsorption, but it has its own shortcomings. These include, but are not limited to, the tendency of lime to be affected by armouring which hinders further lime dissolution, the decreased solubility of lime at high pH and elevated temperature and the formation of a gelatinous low-density sludge comprised of minuscule particles of Fe(III) oxyhydroxides and gypsum. The sludge is characterized by poor dewatering behaviour and postneutralization instability which increases the risk of the adsorbed or encapsulated dissolved metals to re-leach from the disposal sites, thus lessening the likelihood for metal value recovery, waste material reuse or recycle. These negative attributes result in considerable water losses and a large land area being required for sludge disposal to landfill. The literature reviewed indicates that researchers have not reached a consensus on the actual causes of sludge formation and transformation in the course of Fe(III) oxyhydroxide precipitation, with reported reaction kinetics and mechanisms being significantly different. In light of the above difficulties and lack of sufficient progress in body of knowledge, the objectives of the current research were - to gain an in-depth understanding of the dewatering behaviour of primary Fe(III) oxyhydroxide precipitates through investigating the effect of Fe(III) concentration; [Fe(III)] and OH/Fe molar ratios on the performance (settling and filtration rates and specific resistance to filtration - SRF) of products generated in the precipitation process so as to establish the mechanistic relationships that exist, and - to understand the mechanisms and kinetics through which ageing of secondary Fe(III) oxyhydroxide precipitates occurs at elevated temperature and prolonged ageing time and to infer the effects of ageing on the dewatering behaviour of secondary Fe(III) oxyhydroxide precipitates. The mineral phase identification, though acknowledged to be valuable in iron precipitation, was seen to have negligible value in relation to describing dewatering behaviour, hence was not focused on. In order to achieve the objectives above, experiments were conducted in a 1.5 L glass mixed-suspension-mixed-product-removal (MSMPR) reactor under ambient laboratory temperature at 21 ± 1 °C. The agitated MSMPR was equipped with a lid containing ports for pH and temperature probes and four uniformly spaced stainless steel baffles to maximize mixing of the reagents. A synthetic acid mine drainage (SAMD) solution made from aqueous Fe2(SO4)3 was treated in the MSMPR reactor with a reagent grade Ca(OH)2 aqueous stream, with the Ca(OH)2 solution being used for pH adjustment thereby effecting neutralization and precipitation. A residence time of 15 minutes, as estimated from the reactor volume and feed solution flow rate, was used. The number of residence times for steady state attainment were identified as 16 from the particle size distribution (PSD) measurements, although the reactor was allowed to run up to 19 residence times (19 τ = 575 min) before sampling commenced. The ferric concentration ([Fe(III)]) in the feed was kept constant at six different levels of 50, 100, 200, 300, 400 and 800 mg L-1 (± 5 mg L-1 ) whilst adding the corresponding OH, at a molar ratio (OH/Fe - referred to as R) of 3, from the Ca(OH)2 solution. The reaction conditions were such that no gypsum was precipitated. The experiments to investigate the effect of molar ratios were conducted at OH/Fe molar ratios ranging from 3/1 (R = 3), 2.5/1 (R = 2.5) and 2/1 (R = 2) whilst the [Fe(III)] in the feed was kept constant at 100, 300, and 800 mg L-1 (± 5 mg L-1 ). The experiments to aid the understanding of ageing effects on the product characteristics (particle size, PSD, morphology and structure) and dewatering behaviour were conducted under two regimes, the continuous and the batch system. Under continuous conditions and a residence time of 15 minutes, the feed [Fe(III)] into the primary reactor (MSMPR) was set at 100, 300, and 800 mg L-1 and the aqueous stream of reagent grade Ca(OH)2 was added at a molar ratio of OH/Fe of 2.5. The primary reactor was set up on a secure elevated platform and was operated such that the overflow was continuously fed into a 1 L jacketed glass MSMPR reactor (ageing reactor). The ageing reactor was supplied with heated deionized water, operated at 30, 40 and 50 (± 1) °C and agitated at a rate of 230 rpm by a four bladed 45° pitched blade turbine. The batch ageing experiments were conducted in a 1 L jacketed reactor which was allowed to attain the setpoint temperature of 40 ± 1 °C before being filled with fresh Fe(III) oxyhydroxides from the primary reactor precipitated at a molar ratio of OH/Fe of 2.5 (R = 2.5) and [Fe(III)] of 300 mg L-1 . The agitated (230 rpm) reactor was run for 24 hours with samples collected at set times of 1, 2, 4, 6, 7, 8, 9, 10, 16 and 24 hours. The PSD results of the samples were used to establish the mechanisms and kinetics of the system by using the method of moments. The samples taken from the conditions investigated, as outlined above, were analysed for [Fe(III)] in the feed, precipitate and supernatant using spectrophotometry and the 1–10 phenanthroline method as the basis. Spectrophotometry was also used to determine the calcium and sulphate concentrations in the feed, precipitate and supernatant. The PSD and aqueous specific surface area were measured by means of laser diffraction that used static light scattering. Surface charge inferences were made through zeta potential measurements (ζ-potential) using electrophoresis to determine the isoelectric point (IEP). The pH dependence of the ζ-potential was investigated using multi-purpose titrations (MPT-2) with the titrants being 0.01 M H2SO4 and 0.01 M and NaOH. The crystal habit was analysed using scanning electron microscopy (SEM) whilst a combination of SEM with EDX was used for composition determination. Transmission electron microscopy was used to determine the morphology and size. The slurry viscosity was measured using a rheometer operating with the vane and cup arrangement. Iron mineral phase identification was done through Raman microscopy and XRD in order to verify if the mineral phases from a slurry based analysis and a solid sample analysis were the same. The initial studies on the dewatering behaviour indicated that the [Fe(III)] in the feed influenced (to varying extents) the dewatering trend of the precipitation product, via particle size, particle size distribution (PSD), surface charge, morphology and structure. It was established that there was a corresponding change in product characteristics and performance as the [Fe(III)] was increased. A transition was observed at [Fe(III)] of 300 mg L-1 where the settling rate (2.4 ± 0.04 mm min-1 ), particle size (D4,3 of 27 µm) and supersaturation passed through a maximum whilst the filtration rate (3 ± 0.46 cm3 /cm2 /min) and zeta potential approached a minimum. These conditions resulted in enhanced agglomeration, with the improved dewatering behaviour being attributed to a dense network of interlocked and interconnected bridges in the crystal habit of the particles. The results did show that there was a complicated relationship between the underlying micro- and nano-scale properties and the product performance from which dewatering behaviour is inferred. The OH/Fe molar ratio variations were described as low (R = 2.5) and high (R = 3) due to contact nucleation prompting the exclusion of the R = 2 results from further analysis. These molar ratios had an effect on product performance which was improved at the low OH/ Fe molar ratio. A low R positively affected the product characteristics which led to the fastest settling rate of 3.3 ± 0.2 mm min-1 , an increase in the filtration rate from 3.7 ± 0.4 to 5.1 ± 0.5 cm3 cm-2 min-1 and a decrease in the specific resistance to filtration (SRF) from 1.42 x 1012 to 6.71 x 1011 m kg-1 at R = 2.5 and a [Fe(III)] = 300 mg L-1 . The improved product performance was attributed to the dominance of large (D4,3 of 48.3± 0.05 µm and D3,2 of 30.8 ± 0.05 µm), regularly shaped particles which had a unimodal distribution and a notable decrease in the peak number of fine particles from 3.5 × 1013 (1 µm size) to 1.7 × 1012 (3.90 µm size). The Fe(III) oxyhydroxides formed at [Fe(III)] of 100 and 800 mg L-1 displayed insignificant product performance changes with variation in R with the exception of the 800 mg L-1 case which produced the largest change in settling rate (threefold – from 0.8 to 2.4 ± 0.2 mm min-1 ) at R = 2.5. These results reinforced the change in settling and dewatering behaviour observed at [Fe(III)] of 300 mg L-1 . It was noted that the lower R value (R = 2.5) resulted in a lower pH in solution, leading to decreased nucleation. However, higher agglomeration rates were observed, contrary to other studies. This was attributed, in the present study at low R, to a reduction in the surface charge related electrostatic repulsion, whilst van der Waals attractive forces had been favoured by the decrease in supersaturation. The product performance of the precipitates which underwent structural maturation in a continuous system were found to improve after elevated temperature treatment. The highest settling rate was found to have been reached at 40 °C and at [Fe(III)] of 300 mg L1 . This was ascribed to the increase in agglomeration which yielded a high D4,3 (26.3 ± 0.05 µm) mean particle size at this temperature; hence bigger particles implied faster settling. The filterability of the aged precipitates, at a given [Fe(III)], passed through a maximum as the temperature increased. Another transition was observed at Fe(III) of 300 mg L-1 where there was a decrease in surface charge with an increase in temperature as evidenced by the low zeta potential values and a steady state pH which was close to the isoelectric point. This was evidence of a circum-neutral surface charge that favoured agglomeration. The change in micro and nano-structure was seen in the image analysis results. However, these imaging results show evidence of deagglomeration at 50 °C which probably explains the improved dewatering behaviour. The batch ageing system revealed that structural maturation in the ageing process, at elevated temperature, has multiple particle rate processes that occur simultaneously but that which one dominates varies with time. The ageing process results identified four broad time intervals, namely 0 to 6, 6 to 8, 8 to 10 and 10 to 24 hours, through which the different mechanisms were active. The settling trend for the samples collected at the different ageing times displayed a high initial settling rate within the free settling zone of the settling curve. The solids from interval 1 and 2 (0 to 6 and 6 to 8 h) displayed the fastest settling rates within the hindered settling zone with interval 2 solids dominating the compression settling zone. It was noted, however that no significant improvements in settleability were observed with an increase in ageing time except at 8 hours. The filterability on the other hand experienced improvements with ageing especially in the interval 3 and 4 (8 to 10 and 10 to 24 h) where the specific resistance to filtration (SRF) was at its lowest value of 8.78 × 1011 m kg-1 at 8 hours. This was attributed to a combination of an increase in particle size, a decrease in the number of fine particles and a transition in the crystal habit and morphology of the particles as evidenced in the image analysis findings. The product performance followed a trend where it decreased at the onset of interval 1 and remained fairly constant throughout the interval. The onset of interval 2 saw an increase in product performance which reached a distinct maximum at 8 hours after which it reverted back, interval 3 and 4, to levels similar to those in interval 2. The mechanisms that were inferred from the moments of the distribution alluded to the nucleation of new particles (interval 1), the agglomeration and limited growth of those particles which involved structural incorporation (at 8 hours in interval 2) and transformation (of ferrihydrite into goethite particles via a dissolution and reprecipitation process which led to the goethite being sparsely distributed on top of the ferrihydrite base) (interval 3 and 4). The graphical illustration of these processes and their impact on settling and dewatering behaviour is illustrated below. The findings of this work on primary precipitation and ageing of iron during lime neutralization have implications on how iron is treated in AMD and hydrometallurgical processes. A slight elevation in post-precipitation operating temperature under reduced lime usage offers an attractive alternative to improving dewatering behaviour of the precipitated sludge with the potential of reducing metal leachate from landfills.
- ItemOpen AccessCrystal Engineering In Antisolvent Crystallization Of Neodymium Sulfate(2023) Sibanda, Jonathan; Lewis, AlisonAntisolvent crystallization is a separation technology that separates a solute from the solvent by the addition of another solvent, in which the solute is sparingly soluble. High yields are achieved by using higher antisolvent-to-solvent ratios, but this generates higher supersaturation, which causes excessive nucleation. This results in the production of smaller or finer particles, which are difficult to handle in downstream processes such as drying and filtration. Therefore, this study aimed at investigating the effect of varying the organic (antisolvent)-toaqueous ratio and seed loading on the yield, particle size distribution, and morphology of neodymium sulfate octahydrate product, during its recovery from an aqueous leach solution using antisolvent crystallization. A batch crystallizer was used for the experiments, while ethanol was used as the antisolvent. Neodymium sulfate octahydrate [Nd2(SO4)3.8H2O] seeds obtained from antisolvent crystallization at a lower organic-to-aqueousratio of 0.4 were used to investigate the effect of seed loading. The crystals exhibited a plate-like morphology, with the plates becoming thinner, rounded, and layered at higher organic-to-aqueous ratios. This was attributed to enhanced physical interactions between the ethanol and the growing crystals, which possibly caused the ethanol to be adsorbed onto the crystalline surfaces, thereby inhibiting growth from one plane to another. The final particle size distribution curves shifted showing an increase in the particle sizes as the organic-to-aqueous ratio was increased from 0.8 to 1.4. The mean particle size increased from 106.1µm at an organic-to-aqueous ratio of 0.8 to 141.4 µm at an organic-to-aqueous ratio of 1.4. This was attributed to the agglomeration of smaller or fine particles that formed at high supersaturation into bigger particle sizes. The total number of particles decreased while the total volume increased as the organic-to-aqueous increased from 0.8 to 1.4. As the organic-to-aqueousratio increased from 0.8 to 1.4, the final product yield increased from 44% to 90%. The increase in yield was attributed to the increased interaction of ethanol molecules with the solvent molecules, which reduced the solubility of neodymium sulfate. Increasing the seed loading above the critical seed loading of 2.98% to 20% resulted in smaller final particle sizes of 101.5 µm compared with 165.8 µm obtained when the system was not seeded. These smaller sizes obtained at seed loadings above the critical seed loading had narrow particle size distributions. The span of the particle size distribution for the unseeded case was 1.54 while for PSDs obtained at seed loadings above the critical seed loading was less than 1.30. Seeding improved the filtration performance by 47%. This was due to the narrow particle size distribution and improved crystal morphology. It is recommended that to obtain higher yields and particles with unimodal particle size distribution, higher organic-to-aqueous ratios above 1.2 be used, and seeding be conducted at seeding loadings above the critical seed loading. In addition, if crystals of well-defined or faceted morphology are required it will be reasonable to use a lower organic-to-aqueous ratio such as 0.6 to 0.8 at batch times greater than 2.5 h. The filterability of neodymium sulfate at higher organic-to-aqueous ratios needs to be investigated for future studies.
- ItemOpen AccessCrystal growth and nucleation kinetics of diethylenetriammonium hexachlororhodate (III) salt(2019) Engelbrecht,Edmund; Hagemann, Justin; Lewis, AlisonAt Anglo American Platinum's Precious Metal Refinery, rhodium separates from a base metal-rich solution by precipitation. Hexachlororhodate (III) ions and cationic protonated diethylenetriamine ions exchange to form diethylenetriammonium hexachlororhodate (III) crystals, a rhodium metal precursor. The objective of this work is to determine nucleation and growth kinetics of diethylenetriammonium hexachlororhodate (III) salt. Two reactor configurations, namely a transient continuous stirred tank reactor (CSTR) and a t-mixer plug flow reactor (PFR), were used to determine nucleation and growth rates. The objective of the configurations was to eliminate kinetic biases that may be caused by mixing at the mesoscale. Transient saturation in the CSTR ranged up to 43, and in the PFR saturation was varied between 2 and 64. Precipitation kinetic parameters were estimated through data fitting concentration and volume average crystal size profiles to a mass and population balance model. Temperature dependence of kinetic parameters was found to be universal between reactor configurations. Both growth through interfacial attachment and agglomeration, as defined in this work, were exothermic processes with activation energies of -192.9kJ/mol and -656.1kJ/mol respectively. Nucleation was found to be an endothermic process with an activation energy of 50.9kJ/mol in accord with the observed heat of crystallisation. No evidence of heterogeneous primary nucleation in the form of crystals adhering to the side walls of the reactor or the agitator blades was observed. The experiments were not explicitly designed to distinguish between primary and secondary nucleation mechanisms, therefore excluding one over the other is not justified. However, considering the presence of nucleated crystals in each system combined with the good model fit using a nucleation rate expression typically associated with secondary nucleation, it is likely that the dominating nucleation mechanism is secondary in nature. In the PFR configuration, nucleation and growth occur at a faster rate compared to the CSTR under the experimental conditions in this work. This observation is inferred from the fitted temperature independent kinetic parameters and is linked to a much higher mixing intensity achieved in the PFR relative to the CSTR. Flow conditions, described by the Reynolds number, can limit conversion in the PFR configuration by a mixing limitation at the micro- or mesoscale. Micro- and mesoscale mixing were represented by a characteristic length scale that was empirically related to temperature for the microscale and fluid velocity for the mesoscale. Under conditions where the Reynolds number is below the point where conversion is independent of the Reynolds number, either micromixing or mesomixing can become a rate limitation. At a sufficiently high Reynolds number neither micromixing nor mesomixing limits conversion and the system equilibrium becomes the limitation. In the CSTR, the system equilibrium limited reaction conversion as the micro- and mesoscale mixing zones were sufficiently small relative to the reactor volume. Parameters related to mixing were found to differ between the configurations, which was caused by different flow patterns within each configuration. Scanning electron microscopy (SEM) photographs suggest that crystals in the PFR configuration collide both in the radial and axial direction, giving rise to a feathery flat and elongated agglomerated crystal cluster. In contrast, the crystals in the CSTR configuration collide in a chaotic but consistent pattern, giving rise to a desert rose-like agglomerated crystal cluster. The derived model used to describe agglomeration is based on the agglomeration principles proposed by Von Smoluchowski coupled with Fick's law of diffusion and gives a good representation of crystal size. The PFR growth rate supersaturation exponent was 1.13, suggesting a Burton-Cabrera-Frank type growth model, and is indicative of crystal growth from screw dislocations that is limited either through mass transfer to the crystal surface, or surface integration. Thus, in this instance, the rate of aqueous hexachlororhodate (III) conversion to crystal would be responsive to mixing conditions on the micro- or mesoscale, as was experimentally found in the PFR configuration. In comparison, the CSTR growth rate supersaturation exponent was 2.31 and is more in line with polynuclear growth that appears to be limited by interfacial attachment kinetics, as the system equilibrates in the bulk. Lastly, a key finding of this work is the ability to manipulate the crystal morphology by changing reactor configuration. By creating elongated flat crystal structures in the PFR configuration as opposed to a desert rose crystal structure in the CSTR, it may be possible to reduce impurities within the crystal by entraining less mother liquor.
- ItemOpen AccessDeveloping methods to access sensitive industrial wastewater information in South Africa (with treatment in mind)(2018) Harding, Genevieve E; Lewis, Alison; Chivavava, JemitiasSouth Africa is a water stressed country, therefore it is important to understand water use and wastewater generation. Previous research and workshops have identified gaps in the characterisation and remediation of wastewaters in South Africa. Wastewater management can take advantage of wastewater as a valuable resource. However, treatment is required to recover this value, while characterisation is required to develop treatments. Yet wastewater characterisation information is often poorly reported. The nature of industrial wastewaters (in terms of volume, location and composition), and the norms of wastewater characterisation reporting (in terms of quality and accessibility) formed the basis for two research questions. A major component of this research was developing methods to access sensitive wastewater information. Relational approaches were based on building relationships through phone calls, emails, meetings and site visits. Formal, legal requests for were made with application in terms if the Promotion of Access to Information Act (PAIA). Even though wastewater information is not confidential, it is not readily accessible. 87 people from 42 companies or institutions were contacted; 14% of interactions lead to shared data or a meeting, and 12% shared resources. Key industries of interest were: pulp and paper, fish processing, power generation, mining and petroleum. Previous estimates of South African industrial wastewater volumes ranged from 70 – 350 Mm3 /annum. The pulp and paper industry contributed between 28 and 43% of this volume; petroleum contributed 9 to 26%. Both industries were located inland and in coastal regions of South Africa. These industries were most concerned with COD. Mining and power generation contributed 10 – 15% and 7 – 14% respectively. These industries were located inland, and were concerned with total dissolved solids, and specifically sulphate, sodium and chlorides. The fish processing industry contributed between 0 and 23% of volumes, depending whether wastewaters released to a marine environment were included. Seven parameters were reported for over half of the streams considered (65 in total). These parameters were: pH, volume, electrical conductivity, nitrogen, sulphate, sodium and COD. Sulphate and sodium were dominant ions. Calcium was not measured, even though discharge limits were listed in environmental licenses. Characterisation information was reported for compliance and not for treatability. The parameters measured should be expanded to include important parameters for treatability. Industry, research institution and governmental bodies can work together to identify such parameters and develop locally relevant treatments. It is recommended that possible synergies between these groupings be enhanced to improve wastewater management. But an atmosphere of trust and transparency is required to facilitate synergistic relationships. The legal framework in South Africa can be used to motivate for transparency with respect to wastewaters.
- ItemOpen AccessGypsum seeding to prevent scale formation and improve separation efficiency(2021) Chagwedera, Taona Malvin; Chivavava, Jemitias; Lewis, AlisonEutectic freeze crystallization (EFC) is a novel separation technique that can be applied to treat brine solutions such as reverse osmosis retentates. These are often a mixture of different inorganic solutes. The treatment of calcium sulphate-rich brines using EFC often results in gypsum crystallization before any other species. This results in gypsum scaling on the cooled surfaces of the crystallizer, which is undesirable as it retards heat transfer rates and hence reducing the yield of other products. Gypsum crystallizes in small quantities compared to ice and mostly as fines because it is at lower concentrations. This also results in the entrapment of gypsum by the ice during gravity separation of the crystallization products. The aim of this study was to firstly investigate and understand gypsum crystallization and gypsum scaling in the presence of gypsum seeds. Secondly, it was to investigate the effect of gypsum seeding on gypsum separation efficiency. Synthetic brine solutions were used in this research because they allowed an in-depth understanding of the gypsum bulk crystallization process and scaling tendency without the complexity of industrial brines. Cooling crystallization experiments were conducted to investigate the effect of gypsum seeding on gypsum scaling, and EFC experiments were conducted to investigate the effect of gypsum seeding on gypsum separability from the ice. Gypsum seeding was effective in decreasing the mass of scale that formed on the heat transfer surface. Gypsum seeds improved gypsum crystallization kinetics in the bulk solution which resulted in an increase in the mass of gypsum product. Gypsum seeding marginally increased melted ice purity, although the proportion of gypsum which crystallized as fines in the suspension decreased significantly. This was because gypsum deported to the ice fraction mainly through mother liquor entrainment which was contrary to the expectation at the start of the research. The hypothesis was that gypsum reported to the ice fraction mainly through entrapment. It was recommended that the use of silica as a seed material to prevent gypsum scaling should be investigated in future studies. In addition, the treatment of calcium sulphate-rich brines using EFC should be done in stages with the first stage focusing on maximum gypsum removal using a residence time of 5 hours or more.
- ItemOpen AccessImproving the settleability of a metal sulphide suspension by the application of a magnetic field(2015) Gqebe, Sibongiseni Yamkela; Lewis, Alison; Rodriguez-Pascual, MarcosGravitational sedimentation of suspensions in various precipitation processes is hindered by colloidal stability. More especially in sulphide precipitation, where high levels of supersaturation dominate and nucleation is favoured. This results in a large number of colloidal particles with a highly negative surface charge, which remain suspended in solution. The high surface charge of the suspension results in strong attraction/interaction between the ions on the particle surface and counter-ions in solution. Moreover, this strong interaction between ions on the particle surface and counter-ions in solution results in a charge build-up that renders the suspension stable. In order to induce gravitational sedimentation of these particles, a redistribution of ions close to the particle surface is required. This study therefore seeks to redistribute ions close to the particle surface by applying a magnetic field. This results in the reduction of inter-particle electrostatic repulsive forces and subsequent increase in the zeta potential of a suspension. For the purposes of this study, a copper sulphide suspension was used. A T-mixer was used as the reaction zone for the precipitation of each suspension. Subsequent to this, the suspended copper sulphide particles were exposed to a range of field strengths for set exposure times and their zeta potential was measured before and after exposure to the magnetic field. The effect of magnetic field strength, exposure time and particle speed on the zeta potential were tested. All particles had an initial zeta potential value equal to or less than - 40 mV prior to magnetic field exposure. A significant increase in zeta potential was observed with values reaching a maximum of - 16.5 mV when exposed to a 2 T field strength for 40 minutes. An increase in the zeta potential corresponds to a reduction in repulsive electrostatic forces between suspended particles due to the Lorentz force exerted by a magnetic field on the particle surface. DLVO plots were used to quantify this reduction in repulsive electrostatic forces.
- ItemOpen AccessInvestigating the Effect of Heat Transfer Driving Force and Seed Loading on the Batch Eutectic Freeze Crystallization of a Dilute Brine(2023) Spencer, Anotidaishe; Lewis, Alison; Chivavava JemitiasEutectic Freeze Crystallization (EFC) is a separation technology that separates solute from solvent. It is often used as a wastewater treatment process for industrial brines. In EFC, the solution is cooled to temperatures below the eutectic point of the brine, such that ice and salt simultaneously crystallize out of solution. Solid-solid-liquid separation of the resulting suspension is achieved by exploiting differences in the densities of the phases. EFC has been found to perform well when applied to concentrated brines, but not with dilute brines. Previous studies have shown that the chosen operating conditions for the crystallizer can result in phenomena that limit achievable yields and product purities. These phenomena are more severe when dilute brines are treated. One of these limiting phenomena is heat transfer, which is the principle that makes EFC possible. Understanding how operating conditions are related to the production of ice and salt through heat transfer can lead to better control of the EFC process. This would allow for consistent production of ice and salt at high yields and product purities, even for dilute brines. The heat transfer driving force (ΔTLMTD) and ice seed loading (SL) are two operating conditions that are of interest in this work as they are linked to heat transfer; ΔTLMTD, directly by affecting the heat transfer rate and in turn the production capacity of a crystallizing system, and SL indirectly by forming the initial magma density (solid content in the crystallizer by mass) which affects crystallization location and the hydrodynamics of the system. These two operating conditions were chosen because their effect on production of ice and salt from dilute brines has not been studied before. The aim of this work was to understand how SL and ΔTLMTD affect the yield of ice and salt and the purity of ice, in batch EFC. A synthetic Na2SO4 solution was used for crystallization experiments as its eutectic composition is dilute (4wt.%). Ice seed loadings between 0 and 12 wt.% and ΔTLMTD values between 2 and 10℃ were investigated. The yields of ice and salt, as well as the purity of ice, were measured to determine the relationship between ΔTLMTD and the yield of ice and salt, ΔTLMTD and ice purity, SL and the yield of ice and salt, and SL and ice purity. It was found that as the ΔTLMTD increased, the yield of ice and salt increased, due to the higher heat transfer rate. The ice yield was a sum of ice harvested from the bulk and ice formed on the wall (scale). The yield of ice in the bulk decreased with increasing ΔTLMTD, whilst the yield of scale increased. As ΔTLMTD increased, the cooled wall became colder, therefore the fluid closest to the wall had the most supersaturation. This led to the formation of a scale layer. The combination of increased driving force and shortened scale formation time resulted in the observed increase in the scale yield. This increase in the scale yield increased heat transfer resistance between the coolant and the bulk, which produced less ice in the bulk. It was also found that the purity of ice decreased as ΔTLMTD increased. The salt particles were much finer than the ice particles (ice was >13 times larger than salt). The large difference between the ice and salt crystals, combined with increased salt production and intense mixing resulted in an increase in salt entrainment. Increased salt entrainment resulted in poor separation, and a reduction in ice purity was observed. It was found that increasing the ice seed loading increased the yield of ice in the bulk and decreased the scale yield. A maximum in ice in the bulk and a minimum in the scale resulted from a seed loading of 9 wt.%. Solids in the system provided sufficient surface area to promote secondary nucleation and growth in the bulk such that heterogeneous nucleation at the wall was minimised. This reduced the propensity for scale formation. Once the seed loading increased to 12 wt.%, the mixing efficiency of the system decreased. This resulted in localised supersaturation at the wall, which in turn resulted in a scale layer. The reduced mixing efficiency increased the resistance to heat transfer within the bulk, resulting in reduced crystallization within the bulk, therefore the yield of ice in the bulk decreased. It was also found that the salt yield decreased as the seed loading increased. This was due to poor separation caused by the increase in magma density and increase in salt entrainment. The entrained salt in the ice product was washed off rather than harvested, explaining the low salt yield obtained. The salt product was small (80µm
- ItemOpen AccessInvestigating the Effect of Surface Properties on Ice Scaling in Eutectic Freeze Crystallization(2021) Motsepe, Lerato; Chivavava, Jemitias; Lewis, AlisonEutectic Freeze Crystallization (EFC) is an innovative technology that can be applied to treat reverse osmosis (RO) waste streams (brines), to produce pure salt and water. Scaling of the heat exchanger (HX) surface by both ice and salt is currently one of the major drawbacks in the industrial implementation of EFC. At present scaling is controlled by the use of mechanical scraping, which is susceptible to mechanical breakdown, thus reducing the overall process efficiency. Previous studies have shown that lower surface energy materials delay the onset of freezing, and that smooth surfaces reduce nucleation and adhesion sites, thereby reducing the probability of scale formation. Therefore, this study aimed to investigate how the HX surface properties affect ice scaling in EFC, without the influence of mechanical scraping. Copper, Aluminium, Stainless Steel 316 and Brass were the selected HX materials. Ice scaling on the HX materials was investigated using a near eutectic 4 wt.% Na2SO4 aqueous solution, in a crystallization test cell uniquely designed to mimic the region near the HX wall of a crystallizer. The Differential Interference Contrast (DIC) technique was used to study the formation of the initial ice scale layer on the HX material used in the test cell. This method of observation was effective, asfor the first time in a continuous system, the crystallization of the initial ice scale layer was observable in-situ and in real-time. Therefore, with this method, it was possible to investigate the evolution of the predominantscaling modes(nucleation and growth), which differed for the different HX surfaces. The difference was proposed to be due to their distinct surface free energies and surface topographies. The effect of surface free energy and topography on the scaling induction time was investigated while operating at similar heat fluxes (similar cooling rates) for all the metals. The scaling induction time decreased with an increase in the surface free energy, with the Aluminium as an outlier. The recorded scaling induction times for Brass, primary-SS316 and Copper were 92.54, 70.95 and 54.06 min, respectively. Aluminium recorded the longestscaling induction time of 134.74 min. Both the polytetrafluoroethylene (PTFE) coated-SS316 and the primary-SS316 HX surface were used to investigate further the effect of surface free energy on the scaling induction time. The PTFE-coated-SS316 was found to increase the scaling induction times 2.79-fold at a coolant temperature of -15°C, compared to that of the primary-SS316. However, at -20°C and -25°C, the scaling induction times on both surfaces were comparable, which indicated that the benefit of using a low surface free energy material was limited by the cooling rate of the system. It was also found that the scaling induction times were shorter when using a rough-SS316 HX plate, compared to the primary-SS316, because of the larger surface area available for heat transfer. The end of the scaling induction time was characterised by the heterogeneous nucleation and subsequent growth of the ice on the HX surfaces. There was no direct correlation between the HX surface free energy and the nucleation and growth rates. This was because the Brass, Aluminium, SS316 and Copper plates all consist of different surface topographies which also influenced the nucleation and growth rates. However, the nucleation rates consistently increased when the scaling induction times were longer, regardless of the HX material used. The presence of deep sharp crevices on the primary-SS316 also enhanced nucleation rates. These deep sharp crevices created regions of high local supersaturation, where heterogenous nucleation predominated. It was, therefore, reasonable to conclude that the ice scaling induction time was increased by using smooth materials and those of lower surface free energy. The scaling mode was dependent on the surface topography and length of the ice scaling induction time, as longer ice scaling induction times resulted in heterogenous nucleation dominated scaling mode and vice versa. Materials that had a low surface free energy and were smooth minimised the nucleation rate, resulting in a reduced overall scaling rate.
- ItemMetadata onlyThe TCLP and its applicability for the characterization of worst case leaching of wastes from Mining and Metallurgical operations(Elsevier, 1999) Cohen, Brett; Lewis, Alison; Petersen, Jochen; von Blottnitz, Harro; Drews, S C; Mahote, S ILaboratory batch extraction tests, such as the Toxicity Characteristic Leaching Procedure (TCLP), are widely used to classify industrial solid wastes destined for disposal in landfills. Whilst the ease and speed of such tests makes their use an attractive option, it is also immediately clear that the physical and chemical mechanisms dominating in such tests do not bear much resemblance to those expected in a landfill situation. Neither can tests of such short duration be expected to identify long term effects which, if present, are likely to dominate leachate generation behaviour once the waste is disposed. This paper discusses the mechanisms involved in leaching. A series of tests on a metallurgical waste (an EAF dust) is reported which aimed to identify the appropriateness of the TCLP in characterising worst case leaching for this type of material. The results are discussed together with some more theoretical aspects of leaching to assess the applicability of the TCLP for wastes from the mining and metallurgical operations. It is concluded that some of the physical and chemical parameters defining the TCLP require some flexibility in order to create adequate worst case scenarios for each particular waste material.
- ItemOpen AccessThe hydrated lime dissolution kinetics in acid mine drainage neutralization(2021) Mgabhi, Senzo Mntukhona; Petersen, Joachim; Lewis, Alison; Rampai, TokolohoHydrated lime, Ca(OH)₂, has been rediscovered as an environmentally sustainable product, which could be of help in the remediation of acid mine drainage (AMD), especially in the AMD neutralization process. This is due to its ease of acquisition, affordable price and unique versatile properties such as reactivity and neutralization efficiency. AMD is an acidic wastewater containing high concentrations of sulphates and dissolved heavy metals mainly ferrous iron. The dissolution of Ca(OH)₂ in aqueous solution is complex, which make its kinetics in AMD neutralization difficult to understand. The aim of this study was therefore to understand the Ca(OH)₂ kinetics in simplified solutions such as de-ionized water and CH₃COOH. The neutralization process is an acid-base reaction; therefore, pH was used as a critical parameter in determining Ca(OH)₂ dissolution rate. The determination of the dissolution rate was attempted in two ways – measurement of dissolved calcium and determining change of particle size distribution. There were two methods of determining calcium assays investigated, that is EDTA-EBT titration method and OCPC spectrophotometric method. Both methods worked successfully for a Ca(OH)₂-H₂O system. The EDTA-EBT titration method worked better even at higher concentrations of calcium (up to 100 ppm) while the complexometric spectrophotometric method was consistent with Beer-Lambert Law for a narrow calcium concentration range of 1 to 2 ppm, when a small amount of magnesium was introduced. However, both methods failed in the presence of appreciable quantities of magnesium, sulphates and ferric ion. The investigation for particle characterization found that image analysis of SEM images was a better particle-size characterization option than laser diffraction measurement, which tended to cause blinding of the instrument window, but still yielded only qualitative results. There were four reactor configurations used, that is batch reactor for determining the effect of the hydrodynamics (stirring rate and powder addition) and three types of slurry CSTRs. The jacketed chemostat was found to be the optimal reactor configuration while the other two were used as base cases. The Ca(OH)₂ dissolution rate in de-ionized water decreased from 4.0×10⁻⁵ to 1.6×10⁻⁵ mol‧L⁻¹‧s⁻¹ when the temperature was increased from 26 °C to 42 °C. Correspondingly, the pH decreased with Ca(OH)₂ dissolution rate from 11.89 to 11.6. The dissolution rate expression was first order and consistent with the Nernst-Brunner Equation, with the dissolution rate constant of 2.34×10⁻³ s⁻¹ and the activation energy of 18.1 kJ mol ⁻¹ respectively. The overall Ca(OH)₂ dissolution rate in CH₃COOH solution decreased from 2.6×10⁻⁴ to 1.7×10⁻⁴ mol‧L⁻¹‧s⁻¹ when the temperature was increased from 25 °C to 44 °C. At constant ambient temperature (22°C), the Ca(OH)₂ dissolution rate increased with the decrease in pH from 12.1 to 4.38, then decreased with the decrease in pH from 4.38 to 3.5. Using pH to correlate dissolved calcium data and then to determine the rate of reaction, it was found that the dissolution rate is zeroth-order to hydrogen proton and first-order with respect to calcium concentration with the dissolution rate constant of 1.2×10⁻² s⁻¹ and the activation energy of 5.7kJ mol ⁻¹ respectively. These results confirmed that the dissolution of Ca(OH)₂ in DI water and the acetic acid solution is complex. The lower values of the activation energies (5.7 – 18.1 kJ mol ⁻¹), signifies that the kinetics of the Ca(OH)₂ dissolution are mass transfer controlled. Furthermore, these results were confirmed by the weak dependence of the dissolution rate to temperature. However, it was found that slurry CSTR is an efficient reactor system to study the effect of pH on the kinetics of hydrated lime at steady-state conditions.
- ItemOpen AccessTreatment of a multicomponent mining effluent using calcium hydroxide in a fluidized bed crystallizer(2016) Maharaj, Chiara; Lewis, Alison; Chivavava, JemitiasWastewater which primarily emanates from mining operations and manufacturing industries, has the potential for re-use if treated effectively. These wastewaters, which are typically characterized by high concentrations of dissolved inorganic salts are often disposed in evaporation ponds, which promotes the risk of ground water pollution and land wastage. Moreover, this forfeits the potential benefits of valuable salts recovered. The aim of this project was to investigate the treatment of multicomponent saline wastewater rich in sodium and magnesium sulphates, since these salts are prevalent in most wastewater streams. The intention was to treat the wastewater with a calcium hydroxide (Ca(OH)2) suspension in a laboratory scale seeded fluidised bed crystallizer, thereby precipitating gypsum and magnesium hydroxide. The objectives of this study were to investigate how the chosen reactor configuration, feed stream and reagent characteristics affect the conversion and recovery of gypsum and magnesium hydroxide over a range of wastewater concentrations. Particular focus was on reducing the formation of fines through the use of seeds and to get an insight into the possible precipitation mechanisms. It was important that the resulting precipitate product quality favoured effective separation from the treated water stream for re-use. Preliminary experiments were conducted over a feed concentration ranging from 1.5 g/L - 120 g/L (total sulphate salts) which was contacted with a stoichiometric amount of calcium hydroxide with respect to the sulphates in the stream, that is a Ca:SO4 ratio of 1:1 in the fluidised bed crystallizer. These experiments identified a feasible feed concentration range for operation (8 000 -35 000 mg/L). High inlet concentrations (≥ 50 000 mg/L) were not feasible due to rapid formation of a large mass of precipitates which disrupted fluidisation and caused the reactor contents to be elutriated. These high concentrations resulted in high rates of accumulation which necessitated the need for frequent intermittent product removal.