Reliable Mine Water Technology” IMWA 2013

Golden CO; USA “Reliable Mine Water Technology” IMWA 2013

Mineral recovery from Lake Katwe brines using isothermal evaporation

Hillary Kasedde¹,², Matthäus U. Bäbler³, John Baptist Kirabira², Anders Tilliander¹, Stefan Jonsson¹

¹KTH Royal Institute of Technology, Brinellvägen 23, SE-100 44 Stockholm, Sweden. ²School of Engineering, College of Engineering, Design, Art and Technology, Makerere University, P.O. Box 7062, Kampala, Uganda. ³KTH Royal Institute of Technology, Teknikringen 42, SE-100 44 Stockholm, Sweden.

Abstract Lake Katwe is a saline lake within the East African Rift system in Western Uganda, with a rich source of mineral salts. The present work aims at evaluating possibilities of future salt ex- traction from the lake deposit. An isothermal evaporation experiment was conducted on the lake brines. The precipitated salts were characterized by X-ray diffraction (XRD) and scanning electron microscopy (SEM) methods. Various economic salts such as thenardite, gypsum, mirabilite, burkeite, hanksite, anhydrite, trona, halite, nahcolite, thermonatrite, and soda ash precipitate from the lake brines. The experiments also reveal the sequence of mineral salt pre- cipitation in the order sulfates→chlorides→carbonates.

Keywords Lake Katwe; Brine; Isothermal Evaporation; Sulfate salts; Chloride salts; Carbonate salts; XRD; SEM.

Introduction from the surface brine resources of the lake. Lake Katwe is a closed saline lake on the north- These salt extracts are composed of halite ern side of fresh water Lake Edward within the mixed with other impurities. western branch of the East African Rift valley Since the 1960s, studies have been devel- system, and about 15 km below the equator in oped to exploit and utilize the lake’s mineral western Uganda. The lake lies at an elevation resources. Besides the study of the geological of about 885 m, with a maximum area of setting of the region and chemistry of the lake 2.5 km², depth of less than 1.5 m, and measur- brines (Arad and Morton 1969), investigations ing 9 km in circumference. The natural salt were extended to studying the feasibility of lake brines are highly alkaline and rich in Na⁺, salt extraction through estimation of the salt K⁺, Cl⁻, CO₃²⁻, SO₄²⁻, and HCO₃⁻ with lesser reserves (Morton and Old 1968, Dixon and amounts of Mg²⁺, Ca²⁺, Br⁻, and F⁻. The surface Morton 1970, Morton 1973, UDC 1997), and the brines are hydro-chemically of a carbonate mineralogical composition of the evaporites type and represent an important source of (Nielsen 1999). Further studies involved devis- mineral salts that are of great economic value. ing techniques and concepts of improving salt The salinity and density of the lake brines mining and extraction from the lake resources varies from 140 to 150 g/L and 1.15 to 1.23 g/mL (Kirabira et al. 2013) and characterization of the respectively and depend on seasonal varia- mineral salt raw materials from the salt lake tions (Kasedde et al. 2012 submitted). The me- deposit (Kasedde et al. 2012 submitted). teorological conditions in this region are gen- In the present investigation, an isother- erally semi-arid, with little rainfall, and a great mal evaporation experiment was performed capacity for evaporation. Because of these con- to determine the nature of mineral salts that ditions, traditional solar pond evaporation can be recovered from Lake Katwe brines, and techniques are currently used to extract salts to study the sequence of their precipitation.

Wolkersdorfer, Brown ( Figueroa (Editors) 855 IMWA 2013 “Reliable Mine Water Technology” Golden CO; USA

The study is essential for evaluating the possi- in October 2012. The brine had not been satu- bilities for future comprehensive and sustain- rated since no solid precipitates were ob- able utilization of the salt-lake brine resources. served. For the isothermal evaporation exper- iment, one liter of the natural brine was filled Materials and methods in a glass beaker. The beaker was placed in a Apparatus and reagents thermostatic water bath which was main- An FP12 thermostatic water bath (Julabo tained at 30±1 °C. The evaporation conditions Labortechnik, GmbH, Seelback, Germany) was were close to those existing at Lake Katwe. The used for the isothermal evaporation experi- brine was left to evaporate without stirring in ment. The precision of temperature control a ventilated environment at 30±1 °C with a was ±1 °C. The density of the liquid phase was continuous air flow of 1 m/s. The evolution of measured by a portable densito-meter (DMA the brine evaporation was monitored on a 35 Anton Paar, Graz Österreich, Austria) with daily basis for newly precipitated solid salts. an accuracy of ±0.001 g/cm³. The salinity and When a sufficient amount of the solid salts ap- electrical conductivity were measured by an peared, they were separated from the solution electrode probe meter (HANNA instruments by filtration. The salts were then dried and HI 98360, Woonsocket, RI, USA) with an accu- stored in small plastic sample bags and subse- racy of ±0.5 %. The pH value of the liquid phase quently characterized with X-ray diffraction was determined by a PC Titrator (Mantech). and scanning electron microscope techniques. The mineralogy of the precipitated salts at At the same time, a 10 mL brine sample was each stage of evaporation was identified by X- taken from the liquid phase, diluted with dis- ray diffractometry (XRD) using a D2 Phaser tilled water to a final volume of 50 mL, then benchtop XRD system (Bruker Corporation, measurement of its physical properties were Massachusetts, USA) with a copper Kα radia- taken. The isothermal evaporation experiment tion (λ = 1.5405 Å) operating at a voltage and was repeated for each sample until all the current power of 30 kV and 10 mA. A diffrac- brine dried up. tion interval between 2θ-10°-80° with step in- crements of 0.01° and a scan speed of 0.5 sec- Results and Discussion onds were used. The morphology of the salts The brine sample used in the evaporation ex- was examined by a Field Emission Gun Scan- periment was the original brine from Lake ning Electron Microscope (FEG-SEM) using a Katwe. It’s chemical composition in g/L was 137 LEO 1530 Gemini (Zeiss, Oberkochen, Ger- Na⁺, 39.1 K⁺, 0.00143 Mg²⁺, 0.005 Ca²⁺, 124 Cl⁻, many) with settings at a voltage of 15 kV and 43 SO₄²⁻, 3.39 HCO₃⁻, 61 CO₃²⁻, and 0.082 F⁻. In aperture 60 µm. The images were taken with a the course of the isothermal evaporation ex- secondary electron detector. A gold sputter periment, twelve liquid and twelve solid sam- coater (Emitech K550) was used to prepare the ples were collected. The physico-chemical pa- mineral salt samples before SEM analysis to rameters of the original brine (sample Lo) and make them electrically conductive. Double dis- the mother liquors (sample Ln, n representing tilled water having a conductivity of the corresponding evaporation stage) are pre- 0.0182 S/m at 25 °C was used in the experi- sented in Table 1. ment. The evolution of density and conductivity as given in Table 1 is shown graphically in Fig. Experimental methods 1. The original brine (sample L₀) is undersatu- The dry season surface brine was sampled rated and hence, upon evaporation of water from Nambawu salt pans at Lake Katwe in Au- from the original brine, both density and con- gust 2012 and was stored in plastic bottles at ductivity increase until the first salt precipitate room temperature prior to the present study is harvested at sample L₁. Thereafter, density

856 Wolkersdorfer, Brown ( Figueroa (Editors) Golden CO; USA “Reliable Mine Water Technology” IMWA 2013

Sample Evaporation Electrical pH Salinity Density Time(days) Conductivity (g/) (g/cm3) (mS/cm) Lo 0 140.1 10.2 70.0 1.306 L1 6 147.3 10.7 73.8 1.315 L2 12 138.3 10.7 69.2 1.316 L3 15 133.5 10.7 66.6 1.322 L4 20 132.3 10.7 66.1 1.326 L5 25 134.5 10.8 67.4 1.320 L6 29 130.5 10.9 65.3 1.325 L7 33 123.6 10.9 62.1 1.336 L8 36 116.6 10.8 58.4 1.343 L 40 99.6 11.1 49.8 1.360 9 Table 1 Physico-chemical pa- L10 43 95.2 11.1 47.4 1.361 L11 47 93.0 11.2 44.0 1.355 rameters of the liquid brine L12 50 - - - - samples. and conductivity follow a different trend: increase is observed. This second period of in- while density tends to increase, conductivity crease culminates in a local maximum at sam- gradually decreases. This indicates a change in ple L₆ and L₇, followed by a local minimum at the ionic composition of the brine as different sample L₈. Thereafter, a third increase is ob- ionic species contribute differently to density served leading to plateau at sample L₉ and L₁₀. and conductivity. Furthermore, the local max- Likewise to the local extrema in density and ima of density, respectively the local minima conductivity, the different regions in the evo- of conductivity, observed for sample L₄ indi- lution of pH relate to the precipitation se- cates a change in the precipitation sequence, quence. as confirmed by analysis of the precipitates as The mineralogical composition of the re- outlined below. covered solids were analyzed by XRD and SEM. The inset in Fig. 1 shows the evolution of These measurements indicate the presence of pH during evaporation. After a sharp increase various solid phases in each sample. A typical when moving from the original brine to the XRD measurement showing several character- first sample, the pH assumes a constant value istic peaks is shown in Fig. 2. Analyzing the that lasts until sample L₄ after which a further XRD spectra allowed for identifying the solid

Fig. 1 Evolution of brine density (square symbols), Fig. 2 X-ray diffraction (XRD) results for salt sam- conductivity (circles), and pH (inset). ple S1 (Then-Thenardite, An-Anhydrite)

Wolkersdorfer, Brown ( Figueroa (Editors) 857 IMWA 2013 “Reliable Mine Water Technology” Golden CO; USA

Mineral phase Sample S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 Thenardite (Then) Na2SO4 X X X X X X X X X X X X Mirabilite (Mir) Na2SO4·10H2O X X X X X Burkeite (Bur) 2Na2CO3·Na2SO4 X X X X X X Hanksite (Han) 9Na2SO4·2Na2CO3·KCl X X X X X X X Gypsum (Gy) CaSO4·2H2O X X X X X Anhydrite (An) CaSO4 X X Trona (Tr) Na3(CO3)(HCO3)·2H2O X X X X X X X Nahcolite (Nah) NaHCO3 X X Thermonatrite (Th) Na2CO3·H2O X Soda ash (S) Na2CO3 X Halite (Ha) NaCl X X X

Table 2: Mineralogical composition of the solid salts phases in each sample. Table 2 gives an lizing as halite precipitates together with sul- overview of the identified solid phases. fate minerals. In the physico-chemical param- From the data in Table 2 we can identify eters of the brine shown in Fig. 1, this stage is several precipitation stages. A first stage lasting characterized by a strong increase in liquid from sample S₁ to S₄ is dominated by the pre- density combined with a weak decrease in con- cipitation of sulfates only, i.e. the XRD of the ductivity. The pH during this second stage first four samples indicated the presence of the goes through a local maximum as seen in the sodium sulfates thernardite and mirabilite, the inset of Fig. 1. SEM micrographs are shown in carbonate containing sulfates burkeite and Fig. 3 (S6 to S8). Small hexagonal crystals to- hansite, as well as the calcium sulfates gypsum gether with needles and some less well devel- and anhydrite. The corresponding mother oped particles are observed. The typical cube liquors (sample L₁ to L₄ in Fig. 1) had a constant like crystal shape of pure halite was not ob- pH of 10.7, while the density and conductivity served. were increasing, respectively decreasing. A third precipitation stage, lasting from SEM micrographs of the salt sample in sample S₉ to S₁₂, is characterized by the forma- this first stage are shown in Fig. 3 (S1 to S4). tion of sodium carbonates precipitating to- Well faceted crystals with either elongated or gether with sulfates. Several sodium carbon- platelet like shape together with long tubular ates were identified by XRD, namely trona, crystals are seen in sample S1. Similar shape nahcolite, thermonatrite, and soda ash. The and sizes are seen in sample S2 which addi- mother liquors at this precipitation stage as- tionally contains some smaller less developed sumed fairly high densities, i.e. around structures. Comparing the XRD measure- 1.36 g/cm³ (Table 1), which can be explained by ments of the first two samples, the additional the relatively high solubility of the involved structures seen in the SEM of sample S2 can be salts. On the other hand, conductivity was low related to the presence of hanksite, burkeite, while pH assumed values larger than 11. SEM and mirabilite in this sample. Considering micrographs of this last precipitation stage are sample S3 and S4, it is observed that the crystal shown in Fig. 3 (S9 to S12). While sample S₉ still sizes decreases while well faceted and elon- shows many small crystals of various shape, gated crystals together with platelets are also sample S₁₀ shows few large crystals together seen in these samples. with some bulky material, identified as In a second precipitation stage, lasting thenardite and trona, respectively. Sample S₁₁ from sample S₆ to S₈, sodium chloride crystal- shows small prismatic crystals together with

858 Wolkersdorfer, Brown ( Figueroa (Editors) Golden CO; USA “Reliable Mine Water Technology” IMWA 2013

Fig. 3 SEM micrographs of the recovered salts (S₁-S₁₂) during the isothermal evaporation experiment. Scale bars = 10 µm. large indistinguishable crystals which we re- oration process and its crystals were observed late to nahcolite and trona, respectively. to be present in all samples. The mineral salts precipitating from the Conclusion lake brines follow the sequence sulfates→chlo- From the isothermal evaporation experiment rides→carbonates. The evaporation pathway of Lake Katwe brine, the following conclusions of Lake Katwe brines thus differs from that of are drawn: modern sea water evaporation sequence. Precipitates show a rich variety of differ- The salts produced at Lake Katwe by tradi- ent mineral salts, i.e. upon evaporation of lake tional and artisanal techniques are composed brines thenardite, anhydrite, mirabilite, bur - of several mineral phases with limited produc- keite, hanksite, gypsum, trona, halite, nahcol- tion rates. The results from the present study ite, soda ash, and thermonatrite are formed. can provide an important reference in the de- Thenardite precipitates during the entire evap- velopment of technologies for the extraction

Wolkersdorfer, Brown ( Figueroa (Editors) 859 IMWA 2013 “Reliable Mine Water Technology” Golden CO; USA of various pure mineral salts from the natural Kasedde H, Kirabira JB, Bäbler MU, Tilliander A, Jonsson brines of Lake Katwe. Understanding the se- S (2012) Characterization of brines and evaporites of quence of salt precipitation from the brine Lake Katwe, Uganda. Manuscript submitted to Jour- helps to control its evolution during concen- nal of African Earth Sciences tration and hence will lead to an improved op- Kirabira JB, Kasedde H, Semukuuttu D (2013) Towards erating design scheme of the current extrac- the improvement of salt extraction at Lake Katwe. tion process. However, to fully exploit the International Journal of Scientific and Technology lake’s brine resource, further work is required Research 2: 76–81 in studying its thermodynamics and the re- Morton WH, Old RA (1968) A composition and tonnage lated phase equilibria. survey of the salt reserves of Lake Katwe. Uganda Geological Survey. Unpublished Report No. WHM Acknowledgements 6/RAO 7 This study was funded in part by the Swedish Morton WH (1973) Investigation of the brines and evap- International Development Cooperation orite deposits of Lake Katwe, western Uganda. Over- Agency (Sida) and Makerere University. seas Geology and Mineral Resources, No. 41, Insti- tute of Geological Sciences, 107–118. References Nielsen JM (1999) East African magadi (trona): fluoride Arad A, Morton WH (1969) Mineral springs and saline concentration and mineralogical composition. lakes of the Western Rift Valley, Uganda Geochimica Journal of African Earth Sciences 29:423–428. et Cosmochimica Acta 33:1169–1181 UDC (1997) Feasibility study of the rehabilitation of Dixon CG, Morton WH (1970) Thermal and Mineral Lake Katwe Salt Project. Uganda Development Co- Springs in Uganda. U.N Symposium on the Devel- operation, Republic of Uganda. opment and Utilization of Geothermal Resources, Pisa 1970. Vol. 2, Part 2

860 Wolkersdorfer, Brown ( Figueroa (Editors)