Development of an enhanced calcination, optimised leaching process for the improved recovery of Rare Earth Elements from South African discard coal. Prepared by: Hamza Harrar 1429563 Submitted to: A Dissertation Submitted to the School of Chemical and Metallurgical Engineering, Faculty of Engineering and the Built Environment, University of the Witwatersrand, Johannesburg, in Fulfilment of the Requirements for the Degree of Master of Science in Engineering. Supervisor: Prof. Samson Bada 17th May 2022 i DECLARATION I declare that this dissertation is my own unaided work, unless otherwise stated and acknowledged. It is being submitted for the degree of Master of Science in the School of Chemical and Metallurgical Engineering, to the University of Witwatersrand, Johannesburg. It has not been submitted before for any degree or examination to any other university. Signature of candidate: …………………………………. on the 17th day of: May, Year: 2022. ii ABSTRACT Globally there has been a rapid surge in demand for rare earth elements (REE) due to their numerous high-tech applications. China holds more than 43% of global conventional REE reserves and supplies 97% of the world’s REE. Research into other sources of REE is important in responding to society demands, coupled with China’s REE export restrictions and unique commodity applications. Mechanised mining in South Africa produces about 60 million tonnes of coal discard each year that is left in discard dumps. Thus, there is an innovative need to reuse discard coal to reduce its pollution and impact on the health of society and the environment. In this study, a Run of Mine (ROM) and discard coal were evaluated for their REE contents, distribution and association with the inorganic mineral and organic coal constituents. The Tescan Integrated Mineral Analyser was used to investigate the REE associations and distributions. X-ray diffraction provided information on mineral phases and Inductively coupled plasma-mass spectrometry was used to quantify the amount of REE in the coals and subsequent leachate samples. Varying lixiviates under the optimised leaching parameters were investigated using the response surface methodology (RSM) to recover REE from the coal samples. The data attained was used to optimise the leaching process to improve the REE leaching recovery. The same optimised leaching approach was applied to coal samples calcined between 500 oC to 800 oC. The Total REE (TREE) content of each of the two medium rank C bituminous coal samples exceeded 225 ppm. In addition, kaolinite, pyrite and hematite were the main REE-bearing minerals in the discard and ROM coal samples. Heavy REE (HREE) showed a weak ion- adsorbed association with clay minerals (kaolinite) finely dispersed in the organic matrices and fractures of both samples. Furthermore, HREE displayed a strong affinity for the organic macerals and were slightly enriched in the ROM coal, with the discard coal containing higher concentrations of TREE. The encouraging results of this study suggest that both coal sources contain more Critical REE than Uncritical REE, which are in greater demand internationally. The optimised leaching experiments for both coals indicated that an increase in lixiviate concentration (0.5 M to 2 M) and leaching temperature (30 oC to 50 oC), along with a decrease in solid:liquid ratio (40 g/l to 10 g/l), improved the percentage (%) recovery of REE. The RSM and statistical analysis of the leaching data were satisfactorily indicated by an error % of less iii than 1.26%. The optimised leaching parameters (2 M, 10 g/l and 50 oC) manifested a 18.95% REE recovery and 41.35% TREE recovery in the discard and ROM samples, respectively. The low recovery of REEs on the raw samples had indicated that significant quantities of the weekly ion adsorbed REEs were recovered. The study indicated that the best lixiviate was HCl, as it achieved a higher REE recovery than HClO4 and is relatively inexpensive compared to HNO3. The % REE leaching recovery increased as the calcination temperature increased from 500 oC to 700 oC, with optimal calcination at 700 oC. At this temperature (700 oC), the REE leaching recovery achieved was 94.73% (ROM) and 98.17% (discard). Calcination also increased the concentration of REE for ROM sample from 225 ppm to 347 ppm and discard sample from 245 ppm to 363 ppm at 700 oC. At 800 °C, the REE concentration increased to 362 ppm (ROM) and 390 ppm (discard). Leaching time was reduced as the majority of the REE were recovered under the optimised leaching conditions in the first 15 minutes of the process. The significant effect of calcination on REE recovery suggests that REE-bearing minerals were solubilised and oxidised during calcination. The discard coal used in this study had a significantly higher potential than the ROM coal for REE recovery, as it had higher REE abundance and greater recovery. It also establishes a potentially economically viable secondary REE source as no mining is required and contributes to the international pollution reduction target. iv ACKNOWLEDGEMENTS First and foremost, my gratitude and thanks to the Almighty Allah (God), the most merciful, the most beneficent who has given me the capacity and good health to complete this study. The journey was a beautiful bumpy challenge through which my belief in Allah continues to help me overcome the challenges. To my mother, Khadija, and my father, Abderrahim Harrar, thank you for all the support, encouragement and prayers. Thank you to my five sisters (Fatima, Aa’ishah, Zainab, Saudah and Boushrah) who motivated me, inspired me, supported me on tedious days, for humour and happiness during difficult times, coupled with the editing of the English grammar in the study. I truly appreciate all the assistance, love, encouragement and support. I acknowledge with immense pleasure and appreciation the assistance, guidance and contributions provided by my principal supervisor, Prof. Samson. O Bada. Thank you for always providing me with drive, energy and good work ethic, together with all the suggestions and motivations. I would like to thank Dr Jibril Abdulsalam for his contributions and the University of Witwatersrand's Clean Coal Technology Group for its support. Special thanks to Mr. Kingsley for becoming a close friend and research colleague who provided me with complex advice and support throughout my study. I wish to extend my appreciation for the immeasurable contributions to the completion of my research to the following persons: Prof. Rosemary Falcon, Mr. Phillip Pieterse, Prof. Nicola Wagner, Mrs. Petra Dinham, Mr. Paul Den Hoed, Mr. Motlatsi Phali, Mr. Bruce Mothibedi, Miss Ntokozo Dube, Mrs. Reghana Burns, Miss Sibongile Maswanganye and Mr. Phathutshedzo Sikhwari, Mr. Rodney Gurney. I extend my gratitude to the National Research Foundation of South Africa’s SARChI Clean Coal Technology Grant (Grant Number: 86421) for their financial support for this project. Lastly, the opinions, findings, and conclusions expressed and are not necessarily to be attributed to the National Research Foundation (NRF) as they accept no liability in this regard. v PUBLICATIONS The following journal articles were developed from this study: I. Orevaoghene Eterigho-Ikelegbe, Hamza Harrar, Samson Bada. (2021) ‘Rare earth elements from coal and coal discard – A review’, Minerals Engineering. Pergamon, 173, p. 107187. doi: 10.1016/J.MINENG.2021.107187. II. Hamza Harrar, Orevaoghene Eterigho-Ikelegbe, Samson Bada. ‘Mineralogy and distribution of rare earth elements in the Waterberg Coalfield discard coal’ (Accepted May 2022). III. Hamza Harrar, Orevaoghene Eterigho-Ikelegbe, Samson Bada, Jibril Abdulsalam ‘Development of an optimisation leaching process for the improved recovery of Rare Earth Elements from South African discard coal and Run of Mine coal’ (Prepared). IV. Hamza Harrar, Orevaoghene Eterigho-Ikelegbe, Samson Bada. ‘Development of an enhanced calcination optimised leaching process for the improved recovery of Rare Earth Elements from South African discard coal and Run of Mine coal’ (In preparation). vi TABLE OF CONTENTS DECLARATION ........................................................................................................................ i ABSTRACT ............................................................................................................................... ii ACKNOWLEDGEMENTS ...................................................................................................... iv PUBLICATIONS ....................................................................................................................... v LIST OF FIGURES .................................................................................................................. xi LIST OF TABLES .................................................................................................................. xiv NOMENCLATURE ................................................................................................................ xv CHAPTER 1: INTRODUCTION .............................................................................................. 1 1.1 Background and motivation ........................................................................................ 1 1.2 Problem statements ..................................................................................................... 3 1.3 Research questions ...................................................................................................... 3 1.4 Aim and research objectives ....................................................................................... 4 1.5 Hypothesis ................................................................................................................... 4 1.6 Dissertation layout....................................................................................................... 5 CHAPTER 2: LITERATURE REVIEW ................................................................................... 6 2.1 Introduction ................................................................................................................. 6 2.2 What are Rare Earth Elements .................................................................................... 6 2.2.1 Application and demand for REE ........................................................................ 8 2.2.2 Occurrences of REE ........................................................................................... 11 2.2.2.1 Bastnaesite ...................................................................................................... 13 2.2.2.2 Monazite ......................................................................................................... 13 2.2.2.3 Xenotime ........................................................................................................ 13 2.2.3 Alternative REE-bearing minerals. .................................................................... 14 2.2.4 REE in coal ........................................................................................................ 17 2.2.4.1 The abundance of coal and REE in coal-related sources ............................... 18 2.3 Background of coal ................................................................................................... 21 vii 2.3.1 Formation of coal ............................................................................................... 21 2.3.2 Origin of coal macerals (organic matter) ........................................................... 23 2.3.3 Coal microlithotypes .......................................................................................... 23 2.3.4 Mineral matter in coal ........................................................................................ 24 2.3.4.1 Formation of mineral matter in coal ............................................................... 25 2.3.5 Rank of coal ....................................................................................................... 27 2.3.6 Coal grade .......................................................................................................... 29 2.4 Overview of South African coals .............................................................................. 29 2.4.1 South African coal reserves ............................................................................... 29 2.4.2 Geological setting of coalfields in South Africa ................................................ 30 2.5 Occurrence of REE and REE-bearing minerals in South African coal reserves ....... 31 2.5.1 Modes and formation of REE minerals present in coal ..................................... 33 2.6 Physicochemical separation and physical separation of REE minerals .................... 34 2.6.1 Froth flotation .................................................................................................... 34 2.6.2 Magnetic separation ........................................................................................... 36 2.6.3 Spiral, shaking table and multi-gravity separators ............................................. 37 2.6.4 Electrostatic separation ...................................................................................... 38 2.7 Chemical beneficiation and thermal pre-treatment of REE ...................................... 39 2.7.1 Hydrometallurgical techniques for REE recovery ............................................. 39 2.7.1.1 Ion-exchange leaching of REE ....................................................................... 40 2.7.1.2 Sequential digestion/leaching of REE ............................................................ 40 2.7.1.3 Acid leaching of REE ..................................................................................... 42 2.7.1.4 Alkaline pre-treatment before leaching .......................................................... 45 2.7.2 Thermal pre-treatment and leaching of REE (pyrometallurgy) ......................... 47 2.8 Summary ................................................................................................................... 50 CHAPTER 3: METHODOLOGY ........................................................................................... 52 3.1 Introduction ............................................................................................................... 52 3.1.1 Origin and collection of samples ....................................................................... 52 viii 3.1.2 Sample preparation ............................................................................................ 52 3.1.3 Sample relative densities.................................................................................... 52 3.2 Analytical techniques ................................................................................................ 53 3.2.1 Particle size distribution analysis ....................................................................... 53 3.2.2 Proximate analysis ............................................................................................. 53 3.2.3 Ultimate and total sulphur analysis .................................................................... 53 3.2.4 Forms of sulphur analysis .................................................................................. 54 3.2.5 Petrography ........................................................................................................ 54 3.2.6 X-ray diffraction analysis .................................................................................. 54 3.2.7 Fourier transform infrared spectroscopy ............................................................ 55 3.2.8 Tescan Integrated Mineral Analyzer analysis .................................................... 55 3.2.9 Inductively coupled plasma mass spectrometry ................................................ 55 3.3 Experimental procedure ............................................................................................ 56 3.3.1 Summary of Experimental Procedures .............................................................. 57 3.4 Beneficiation and recovery of REE ........................................................................... 58 3.4.1 Part 1: Leaching of the non-calcined samples ................................................... 58 3.4.1.1 Stage 1: varying concentration, solid:liquid ratio and temperature ............... 58 3.4.1.2 Stage 2: leaching of non-calcined samples under DOE optimal conditions .. 61 3.4.2 Calcination method and apparatus ..................................................................... 61 3.4.3 Part 2: Leaching of the calcined samples at optimal parameters ....................... 62 3.5 List of apparatus and equipment required for the leaching tests ............................... 62 CHAPTER 4: RESULTS AND DISCUSSION ....................................................................... 63 4.1 Introduction ............................................................................................................... 63 4.2 Coal characterisation results...................................................................................... 63 4.2.1 Physico-chemical properties and petrography results ........................................ 63 4.2.2 Mineralogical composition of the coals ............................................................. 65 4.2.3 Fourier transform infrared spectroscopy results. ............................................... 72 4.3 REE contents ............................................................................................................. 74 ix 4.3.1 Comparative quantitative results of REE ........................................................... 74 4.3.2 Comparative quantitative REE results for the two calcined coal samples......... 75 4.3.3 Estimation of the quality of RC1 and DC1 as an REE source ........................... 77 4.4 Depositional environment of REE ............................................................................ 80 4.5 Hydrometallurgical treatment of RC1 and DC1 for REE recovery .......................... 83 4.5.1 Particle size distribution of samples used in the leaching tests ......................... 83 4.5.2 Part 1 and stage 1 of the leaching results from non-calcined RC1 and DC1 ..... 84 4.5.3 Leaching mechanisms of the REE-bearing minerals ......................................... 86 4.6 Experimental design and statistical analysis results from DOE ................................ 88 4.6.1 Optimisation of the REE leaching recovery from DC1 and RC1 ...................... 88 4.6.1.1 Suitability and adequacy of the model (ANOVA) ......................................... 90 4.6.1.2 Linear model equations for predicting % REE leaching recovery ................. 92 4.6.1.3 Response plots for determining the effect of HCl concentration on the % REE leaching recovery .......................................................................................................... 95 4.6.1.4 Response plots for determining the effect of solid:liquid ratio on the % REE leaching recovery .......................................................................................................... 96 4.6.1.5 Response plots for determining the effect of leaching temperature on the % REE leaching recovery ................................................................................................. 97 4.6.1.6 Response surface methodology plots for the interactions between the leaching factors and the % REE leaching recovery ..................................................................... 98 4.6.1.7 Determination of the optimal leaching parameters ...................................... 103 4.7 Percentage REE recovery under optimised leaching conditions for discard and ROM coals ................................................................................................................................. 106 4.7.1 Effect of HCl on the recovery of REE from RC1 and DC1 coals ................... 106 4.7.2 Effect of acid type on the percentage REE recovery under optimal leaching parameters ...................................................................................................................... 110 4.8 The effect of calcination pre-treatment on the subsequent % REE leaching recovery… ………………………………………………………………………………..112 4.8.2 The effect of calcination pre-treatment on the subsequent % HREE and LREE leaching recovery ........................................................................................................... 115 x 4.9 Summary ................................................................................................................. 117 CHAPTER 5: CONCLUSION AND RECOMMENDATIONS ........................................... 118 5.1 Overview ................................................................................................................. 118 5.2 Key findings of the research.................................................................................... 118 5.3 Recommendations for future studies ....................................................................... 120 REFERENCES ...................................................................................................................... 122 APPENDICES ....................................................................................................................... 139 Appendix A1 ...................................................................................................................... 139 xi LIST OF FIGURES Figure 2-1: REE industrial market demand per element during 2015 (adapted from Goodenough et al., 2018)……………………………………………………………………..10 Figure 2-2: Forecast of REE market demand estimate per element for 2025 (adapted from Goodenough et al., 2018)……………………………………………………………………..11 Figure 2-3: SEM-EDX analysis of a non-calcined coal sample from Western Kentucky USA (Zhang and Honaker, 2020)…………………………………………………………………..18 Figure 2-4: Variation in the key coal properties for rank advancement (adapted from Ward and Suárez-Ruiz (2008))………………………………………………………………………….29 Figure 2-5: Illustration of the 19 coalfields situated in South Africa (adapted from Snyman (1998))………………………………………………………………………………………..31 Figure 2-6: XRD spectrum for raw coal from the Tutuka Power Station (Akinyemi et al., 2012)………………………………………………………………………………………….32 Figure 2-7: Variation of the minerals by depth in the Grootegeluk formation based on XRD and petrographic analysis (Faure, 1993)……………………………………………………...33 Figure 2-8: The sequential extraction procedure for REE recovery from coal (adapted from Zhang and Honaker, 2019b)…………………………………………………………………42 Figure 2-9: Representation of REE recovery based on their mode of occurrence in sequential extraction tests (Zhang and Honaker, 2019b)…………………………………………………42 Figure 2-10: Acid leaching of coarse refuse coal and middlings coal at different calcination temperatures (Zhang and Honaker, 2019a)…………………………………………………...49 Figure 3-1: Overview of the calcination and leaching methodology applied in the study……57 Figure 3-2: Image of the experimental leaching set-up……………………………………….59 Figure 3-3: Image of the microfilter-syringe employed for leachate extractions……………...60 Figure 4-1: Petrography images 1a, b, c, d (RC1) and 1e, f, g, h (DC1)……………………….67 Figure 4-2: XRD diffractograms for RC1 and DC1…………………………………………..69 Figure 4-3: TIMA image results of REE-bearing minerals for RC1 (BSE image a and c, and SE image b and d) and DC1 (BSE image e and g, and SE image f and h)……………………...71 Figure 4-4: FTIR spectrum for the ROM (RC1) and discard coals (DC1)…………………….73 Figure 4-5: Categorisation of the RC1 (representing the ROM coal) (denoted by the red star) and DC1 (representing the discard coal) (denoted by the red star as well, similar position on chart) by employing the Coutlook in conjunction with the REYdef, rel % (Seredin and Dai, xii 2012)………………………………………………………………………………………….78 Figure 4-6: Normalised REE:UCC for RC1 and DC1…………………………………….…..79 Figure 4-7: REE chondrite (Schmitt et al., 1963) normalised plot for ROM coal (represented by RC1), discard coal (represented by DC1), USA coals (Finkelman, 1993) and Chinese coals (Dai et al., 2008)……………………………………………………………………………...82 Figure 4-8: PSD analysis curve for RC1 and DC1…………………………………………….83 Figure 4-9: The % REE recovery of RC1 (A, B and C) and DC1 (D, E and F) after leaching…85 Figure 4-10: Normal probability plots of the % REE leaching recovery for RC1 (A) and DC1 (B)…………………………………………………………………………………………….93 Figure 4-11: Residuals vs predicted plot for % REE leaching recovery for RC1 (A) and DC1 (B)…………………………………………………………………………………………….94 Figure 4-12: Linear regression plots for the predicted vs actual % of REE leaching recovery for RC1 (A) and DC1 (B)……………………………………………………………………..95 Figure 4-13: % REE leaching recovery vs HCl concentration for RC1 (A) and DC1 (B)……..96 Figure 4-14: % REE recovery leaching vs solid:liquid ratio for RC1 (A) and DC1 (B)……….97 Figure 4-15: % REE leaching recovery vs leaching temperature for RC1 (A) and DC1 (B)…..98 Figure 4-16: Contour plots for the relationship between HCl concentration and leaching temperature with respect to % REE leaching recovery for RC1 (A) and DC1 (B)……...……..99 Figure 4-17: 3D surface plots of the relationship between temperature, HCl concentration and their impact on the % REE leaching recovery for RC1(A) and DC1(B)……………………..100 Figure 4-18: Contour plots for the relationship between solid:liquid ratio and leaching temperature with respect to % REE leaching recovery for RC1 (A) and DC1 (B)…………..101 Figure 4-19: 3D surface plots of the relationship between leaching temperature, pulp density (solid:liquid ratio) and their impact on the % REE leaching recovery for the RC1(A) and DC1(B) coals………………………………………………………………………………..101 Figure 4-20: ROM (RC1) (A) and Discard (DC1) (B) coals contour plots for the relationship between solid:liquid ratio (g/l) and HCl (M) concentration with respect to REE recovery…..102 Figure 4-21: 3D surface plots of the relationship between leachate concentration, pulp density (solid:liquid ratio) and their impact on the % REE leaching recovery for the RC1(A) and DC1(B)……………………………………………………………………………………...102 Figure 4-22: Contour plots of desirability for varying solid:liquid ratios and HCl concentrations for RC1 (A) and DC1 (B)……………………………………………………………………104 Figure 4-23: Contour plots of desirability for varying solid:liquid ratios and leaching temperatures for RC1 (A) and DC1 (B)……………………………………………………...105 xiii Figure 4-24: RC1 and DC1 % REE leaching recovery under the optimised leaching parameters (HCl concentration of 2 M, solid:liquid ratio of 10 g/l, leaching temperature of 50 oC, D90 = - 106 µm, stirrer speed 550 rpm)……………………………………………………………...107 Figure 4-25: RC1 and DC1 % HREE (A) and % LREE (B) leaching recovery under the optimised leaching parameters (HCl concentration of 2 M, solid:liquid ratio of 10 g/l, leaching temperature of 50 oC, D90 = - 106 µm, stirrer speed 550 rpm)………………………………..109 Figure 4-26: RC1 and DC1 % REE leaching recovery under the optimised leaching parameters (HClO4/HNO3 concentration of 2 M, solid:liquid of 10 g/l, leaching temperature of 50 oC, D90 = - 106 µm, stirrer speed 550 rpm)…………………………………………………………..110 Figure 4-27: RC1 and DC1 % HREE (A) and % LREE (B) leaching recovery under the optimised leaching parameters (HClO4/HNO3 concentration of 2 M, solid:liquid ratio of 10 g/l, leaching temperature of 50 oC, D90 = - 106 µm, stirrer speed 550 rpm)……………………111 Figure 4-28: The effect of calcination of RC1 (A) and DC1 (B) on the % REE leaching recovery under the optimised leaching parameters (HCl concentration of 2 M, solid:liquid of 10 g/l, leaching temperature of 50 oC, D90 = - 106 µm, stirrer speed 550 rpm)……………………...113 Figure 4-29: Calcined RC1 (A) and DC1 (B) % HREE leaching recovery under the optimised leaching parameters of (HCl concentration of 2 M, solid:liquid ratio of 10 g/l, leaching temperature of 50 oC, D90 = - 106 µm, stirrer speed 550 rpm)………………………………..115 Figure 4-30: Calcined RC1 (A) and DC1 (B) % LREE leaching recovery under the optimised leaching parameters of (HCl concentration of 2 M, solid:liquid ratio of 10 g/l, leaching temperature of 50 oC, D90 = - 106 µm, stirrer speed 550 rpm)………………………………..116 Figure A1-1: PSD curve for RC1 coal sample calcined at 500oC……………………………139 Figure A1-2: PSD curve for RC1 coal sample calcined at 600oC. ……………...………...…139 Figure A1-3: PSD curve for RC1 coal sample calcined at 700oC...…………….………....…139 Figure A1-4: PSD curve for RC1 coal sample calcined at 800oC……………………………140 Figure A1-5: PSD curve for DC1 coal sample calcined at 500oC. …………………………140 Figure A1-6: PSD curve for DC1 coal sample calcined at 600oC……………………………140 Figure A1-7: PSD curve for DC1 coal sample calcined at 700oC……………………………141 Figure A1-8: PSD curve for DC1 coal sample calcined at 800oC……………………………141 xiv LIST OF TABLES Table 2-1: List of 17 REE, illustrating their valence states and ionic radii (Shannon, 1976)…...7 Table 2-2: Applications and uses of REE in the industry (Naumov, 2008)……………………..9 Table 2-3: Common and predominant, economic or potentially economically feasible REE- bearing minerals (Long et al., 2012)………………………………………………..……..15-16 Table 2-4: Minerals in coal deposits (adapted from Saxby (2000) and Xiuyi (2011))………...27 Table 3-1: Experimental design matrix (from DOE) for the leaching of non-calcined RC1…..60 Table 3-2: Experimental design matrix (from DOE) for the leaching of non-calcined DC1…..61 Table 4-1: Summary of the characteristics of the ROM coal (RC1) and coal discard (DC1)….64 Table 4-2: TIMA modal liberation analysis results for the quantities of the various mineral species present in RC1 and DC1……………………………………………………………...70 Table 4-3: ICP-MS results for individual REE contents and TREE, outlook coefficient (Coutlook), together with important REE ratios………………………………………………...75 Table 4-4: ICP-MS results for individual REE contents and the TREE, outlook coefficient (Coutlook) and REE ratios for the calcined coal samples………………………………………..77 Table 4-5: Enrichment patterns and anomalies for individual REE of RC1 and DC1…………79 Table 4-6: Experimental vs predicted % REE leaching recovery from RC1………………….89 Table 4-7: Experimental vs predicated % REE leaching recovery from DC1………………...89 Table 4-8: ANOVA table of the % REE recovery for RC1 with respect to the various leaching factors………………………………………………………………………………………...91 Table 4-9: ANOVA table of the % REE recovery for DC1 with respect to the various leaching factors………………………………………………………………………………………...91 Table 4-10: Leaching parameter settings for the optimised response of RC1 and DC1……...103 Table 4-11: Leaching conditions of the optimised solution for RC1 and DC1………………103 Table 4-12: Point prediction table for the optimised selected solutions……………………...105 xv NOMENCLATURE List of abbreviations Abbreviations and Acronyms Full meaning ASTM American Society for Testing and Materials BSE Backscattered Electron CFA Coal fly ash CREE Critical Rare Earth Elements DC1 Discard coal sample DOE Design of Expert EDX Energy-dispersive X-ray FTIR Fourier transform infrared H2SO4 Sulphuric acid H3PO4 Phosphoric acid HCl Hydrochloric acid HClO4 Perchloric acid HGMS High gradient magnetic separator HHS Hydrophobic hydrophilic separation HNO3 Nitric acid HREE Heavy Rare Earth Elements ICP-MS Inductively coupled plasma mass spectrometry IEA International energy agency ISO International Organization for Standardization LREE Light Rare Earth Elements MGS Multi-gravity separator NaOH Sodium hydroxide NdFeB Neodymium-iron-boron alloy PSD Particle size distribution RC1 Run of Mine coal sample REE Rare Earth Elements REO Rare earth oxide ROM Run of Mine RSM Response surface methodology xvi SE Secondary Electron SEM Scanning Electron Microscopy SG Specific Gravity TEM Transmission Electron Microscopy TIMA Tescan Integrated Mineral Analyzer TREE Total Rare Earth Elements UCC Upper continental crust UCREE Uncritical Rare Earth Elements USA United States of America WEEE Waste Electrical and Electronic Equipment XRD X-ray diffraction Units of measure Symbols Meaning oC degrees Celsius g grams g/l grams/litre > greater than hrs hours kg kilograms < less than l litre kA/m magnetic field intensity (kilo amp per metre) T magnetic field strength (Tesla) µg/g microgram/gram µg/ml microgram/millilitre µm micrometres ml millilitres mm millimetre min minute M molar concentration (mol/litre) No. number xvii - particle size and below the specific particle size ppm parts per million % percentage 𝐶𝑅𝑆 REE/elemental concentration rpm revolutions per minute s second SG specific gravity CL TREE/elemental concentration VL volume of leachate vol. % volume percentage 𝑀𝑅𝑆 weight of the coal % wt. weight percentage 1 CHAPTER 1: INTRODUCTION 1.1 Background and motivation The international demand for Rare Earth Elements (REE) was approximately 256 000 tonnes in 2021 (Statista, 2022). As of 2019, China had an estimated 44 million metric tonnes of rare- earth oxide, with South Africa owning 0.79 million metric tonnes (Garside, 2021). The demand for REE is based on its high-technology applications, as well as its benefits and economic value (Reid, 2018). The research into the mining, beneficiation and utilisation of this commodity will continue to gain more attraction due to its importance in the 4th industrial revolution. REE have unique intrinsic properties that are used in high-tech applications, including phosphors and electronic displays, high strength permanent magnets and the generation of renewable energy sources (wind and solar power) (Humphries, 2013). For many years REE have been mined conventionally from the ion-adsorbed clay deposit pre- dominantly found in China, Mongolia and Australia (Zaimes et al., 2015; Reid, 2018). The high concentration of these elements in one location or region is rare, apart from in China. This has led China to be the dominant global market for this commodity (JackLiftonReport, 2019). The Light Rare Earth Elements (LREE) are obtained from minerals like Bastnaesite and Monazite (LREE: Cerium (Ce), Lanthanum (La), Neodymium (Nd)), whilst the Heavy Rare Earth Elements (HREE) are obtained from Xenotime (HREE: Yttrium (Y), Dysprosium (Dy), Erbium (Er), Ytterbium (Yb) and Holmium (Ho)) (Akdogan et al., 2019). With the economic exploitation of conventional ores in many countries, as well as the limited supply and importance of the commodity, there is a growing interest globally to find alternative sources for extraction and enrichment of REE. Approximately 20 years ago, the first detection of REE from coal deposits was conducted in the Russian coal basin by Seredin (1991). Subsequently, other sources, such as fly ash (Kolker et al., 2017), the roof and floor of coal seams (Seredin and Dai, 2012), coal discard and acid mine drainage effluent (Zhang and Honaker, 2018) have been identified as potential sources for beneficiating REE. Significant research has been conducted on extracting and recovering REE from Run of Mine (ROM) coarse and fine coal discard (Huang et al., 2018; Zhang and Honaker, 2018) . With global coal production estimated at 8.13 billion tonnes in 2019 (Prime, 2019), it is expected that production of discard coal will increase and, thus could be a potential supply of REE. This 2 is significant for South Africa, where 60 million tonnes of coal waste or discard is generated annually (Nogzina, 2001). REE are also present in post-combustion fly ash, in which they are bound with other minerals or in an ion-adsorbed/ion-substitution form (Kolker et al., 2017). This form of REE require large amounts of energy for grinding and concentration of the REE at a fine particle size less than 10 µm (Huang et al., 2018). The distribution of REE at fluctuating concentrations was observed in phosphorite, shells, corals, coal and carbonatites. According to Schofield and Haskin (1964), phosphorite was found with the highest concentration of lanthanides (160 ppm), whilst coal was at 49.9 ppm, followed by shells at 3.3 ppm, corals at 0.13, 0.08 and 0.052 ppm, and carbonatites within 76 and 181 ppm. A shift in the application of coal solely for power generation is required if coal will continue to retain its dominance as the cheapest and most abundant fossil fuel. Using coal for non-energy applications such as the synthesis of high-value products is the only way that coal will continue to contribute to the South African economy for decades to come. A recent study by Wagner and Matiane (2018) revealed that Witbank and Highveld coalfields were high in REE with Total REE (TREE) levels for both samples exceeding 111 ppm. Thus, a potential source for REE recovery. Seventeen other coalfields in South Africa require investigation in terms of their REE-bearing content. Wagner and Matiane (2018), as well as others (Akinyemi et al., 2012; Akdogan et al., 2019; van Breugel et al., 2019), investigating REE in South Africa, have only focused on the distribution and acid leaching of REE in ROM coal and fly ash, respectively. The current study conducted, employed the integration of calcination before an optimised leaching process for improving REE recovery from a South African discard and ROM coals. South African coal discard has not been previously investigated for REE extraction using this procedure. Samples from discard and ROM coal were analysed for their physicochemical coal and mineralogy properties, REE distribution and abundance. Hydrochloric acid (HCl) leaching was applied to the raw coal samples according to conditions attained from Design of Expert (DOE) to optimise the leaching parameters (concentration, temperature and solid:liquid ratio) and to improve REE recovery. The pre-treatment method to increase the solubility of REE by calcination was adopted to mitigate the characteristic low REE recovery from raw coal. The REE contents of the leachate samples were analysed using Inductively coupled plasma mass spectrometry (ICP-MS). 3 1.2 Problem statements I. Due to the rapid rise in global demand for REE and declining conventional reserves, the supply of high-value REE has become a concern that needs to be addressed. Furthermore, China, the major international supplier of REE, has now imposed restrictions on REE export, driving the recovery of REE from other sources such as coal and coal discard. II. Coal is South Africa’s primary electricity generation resource. The mechanised mining of coal and whole coal seam rather than mining the high-quality seam has led to the generation of tonnes of fines and coarse discard tailing streams from beneficiation plants. These coals or tailings dumps are an environmental concern for mining companies and surrounding communities. These dumps may contain REE that can be mined economically and used as a source of income for the community and the mining sector and may be employed for rehabilitation costs. III. Another predominant problem is the ultrafine dispersion of REE within coal sources. It was noted that even with extreme particle liberation finer than 10 µm, physical separation methods manifest poor REE recovery (Jordens et al., 2013). Hence, acid leaching has been identified as a better method for REE recovery, but the recovery rate depends on the coal constituents (Jordens et al., 2013). Because the types of bonds and mineral associations differ from one coal to another, the recovery rate is not expected to be the same for all coals. Also, the acid concentration required for maximising REE recovery is expected to differ from coal to coal. Improving REE recovery requires optimising the acid leaching process in terms of concentration, temperature and solid:liquid ratio. IV. The recovery of REE also depends on the mineralogy of the discard and ROM coals in terms of the distribution of REE in the coal constituents. 1.3 Research questions The following questions were addressed in this study: I. Will a South African discard and ROM coal provide a substantial concentration of REE before pre-treatment? Thus, indicating whether these coals are a good source for REE extraction. 4 II. What does the mineralogy reveal about the occurrence of REE, their associations with the mineral matter and organic matter, and is HREE or LREE more dominant in this South African ROM and discard coal? III. Based on the association of REE with different minerals, can REE be recovered via leaching? IV. Which acid lixiviate will achieve the highest REE recovery? V. What effect will different leaching conditions (concentration, temperature and solid:liquid ratio) have on the percentage (%) REE leaching recovery? VI. Will the calcination step improve the % REE leaching recovery of the two coal samples, and will this treatment technique reduce leaching time? 1.4 Aim and research objectives This research aimed to develop an enhanced calcination and optimised acid leaching process route to recover REE from a South African discard and ROM coals. The objectives of this study were: I. To identify the particle size distribution (PSD) of the REE present in discard and ROM coals. II. To understand the REE associations with mineral and organic matter in the two coal samples. III. To determine the effect of acid concentration, temperature and solid:liquid ratio on the % REE leaching recovery. IV. To derive a statistically optimised leaching model for REE recovery. V. To establish the most effective lixiviate for improving REE recovery. VI. To determine the effect of calcination at optimal leaching conditions on the % REE leaching recovery from calcined discard and ROM coals. 1.5 Hypothesis Acidic leaching of coal samples is expected to recover REE, though not in high quantities. Furthermore, with the integration of calcination before an optimised leaching process, it is expected that the % REE leaching recovery will improve for both discard and ROM coal. 5 1.6 Dissertation layout This dissertation comprises of five chapters: I. Chapter 1: The background and motivation for this study are outlined. The research problems, questions, aim, objectives and hypothesis evaluated in this study are presented. II. Chapter 2: This chapter provides an extensive review of REE; their demand, applications, occurrences in coal and conventional minerals and associations with organic and mineral coal fractions. The chapter also presents the background of coal and its properties, the physical, chemical and thermal beneficiation methods for REE recovery from coal and other sources. III. Chapter 3: This chapter provides in-depth methodology, characterisation techniques and apparatus used in the leaching and calcination experiments. The chapter further describes the experimental design for optimising modelled leaching parameters (acid concentration, temperature and solid:liquid ratio, and lixiviate) to determine the optimal parameters for improving REE recovery. IV. Chapter 4: The chapter describes the coal characterisation, mineralogy and REE associations. In addition, the chapter presents the experimental and optimised results of the leaching parameters achieved, as well as the effect of calcination on improving the % REE leaching recovery. V. Chapter 5: This section highlights the overall findings of the study, coupled with recommendations for future research. 6 CHAPTER 2: LITERATURE REVIEW 2.1 Introduction The chapter provides a broader perspective on REE, their occurrences in conventional ores and coal sources, and their applications and benefits. Furthermore, it includes information on the association of REE with minerals in coal, together with the methods of extracting and increasing their recovery from various conventional resources and coal-related sources. 2.2 What are Rare Earth Elements REE, in the past, were only known to chemists and material scientists, as their applications and intrinsic qualities were not well known. At the beginning of the 21st century, REE received more recognition due to the media and development of the internet (Long et al., 2010). Their increasing recognition is due to their intrinsic and specialised properties, making them useful in many high-tech applications (Van Gosen et al., 2017). There are 17 REE, depicted on the periodic table as the lanthanide series from atomic number 57 to 71, including scandium (Sc, atomic number 39) and Y (atomic number 21) (Humphries, 2013; De Lima and Leal Filho, 2015). Y and Sc are considered REE due to their physical and chemical similarities, along with their affinity and association with REE minerals in nature (Dushyantha et al., 2020). However, promethium (Pm, atomic number 61) does not occur naturally and thus is not regarded in the group of REE (De Lima and Leal Filho, 2015). REE spontaneously occur together in nature as they share a common trivalent charge (+3) and primarily have similar ionic radii (Long et al., 2010; Table 2-1). Cerium (Ce; atomic number 58) and Europium (Eu; atomic number 63) are exceptions, as they can also occur in different valence states (Ce4 + and Eu2 +) (Shannon, 1976). Elements on the periodic table typically increase in ionic radii as their atomic weights increase (Long et al., 2010). However, with REE, their ionic radii decrease with increasing atomic weight, termed the lanthanide contraction (Gupta and Krishnamurthy, 2005; Jordens et al., 2013). 7 Table 2-1: List of 17 REE, illustrating their valence states and ionic radii (Shannon, 1976). Atomic number Symbol Name Valence state/ Ionic radii (Å) 21 Sc Scandium Sc3 +: 0.87 39 Y Yttrium Y3 +: 1.02 57 La Lanthanum La3 +: 1.16 58 Ce Cerium Ce3 +: 1.14, Ce4 +: 0.87 59 Pr Praseodymium Pr3 +: 1.13 60 Nd Neodymium Nd3 +: 1.11 61 Pm Promethium 62 Sm Samarium Sm3 +: 1.08 63 Eu Europium Eu2 +: 1.25, Eu3 +: 1.07 64 Gd Gadolinium Gd3 +: 1.05 65 Tb Terbium Tb3 +: 1.04 66 Dy Dysprosium Dy3 +: 1.03 67 Ho Holmium Ho3 +: 1.02 68 Er Erbium Er3 +: 1.00 69 Tm Thulium Tm3 +: 0.99 70 Yb Ytterbium Yb3 +: 0.99 71 Lu Lutetium Lu3 +: 0.98 REE are classified according to their geochemical properties and economic stance (Seredin and Dai, 2012; Zhang et al., 2015). The geochemical classification separates REE into LREE and HREE based on their atomic number (Gupta and Krishnamurthy, 1992). LREE are elements with an atomic number from 57 to 63, which include the following elements: La, Ce, Praseodymium (Pr), Nd, Samarium (Sm) and Europium (Eu) (Gupta and Krishnamurthy, 2005). Whilst HREE are elements with an atomic number of 21, 39 and from 64 to 71: Sc, Gadolinium (Gd), Terbium (Tb), Dy, Ho, Erbium (Er), Thulium (Tm) Yb, Lutetium (Lu) and Y (Gupta and Krishnamurthy, 2005). Even though Y (atomic number 39) is light, it is still in the HREE category due to the chemical and physical affiliation that it shares with HREE (Van Gosen et al., 2017). LREE are found more commonly in nature than HREE, reflecting why HREE are more valuable than LREE (Reid, 2018). The United States of America (USA) Department of Energy also categorised REE into Critical REE (CREE) and Uncritical REE (UCREE) based on their importance to global development and for supply within the next five to twenty years (Reid, 2018). According to the USA Department of Energy, the CREE include Nd, Tb, Y, Dy, Er and Eu based on their high demand and specialised characteristics (Ekmann, 2012). 8 2.2.1 Application and demand for REE The surge in global demand for REE is due to their increased use in high-tech applications (Dushyantha et al., 2020). In the 1890s, the first documented application of REE was the use of lanthanum oxide as a commercialised gas mantle in Vienna (Neary and Highley, 1984; Szabadvary, 1988). Since then, many scientists have been evaluating the characteristics and intrinsic properties of REE for application in different industrial sectors (Humphries, 2013; Van Gosen et al., 2017). This has led to the industrial use of these elements in many different chemical forms, such as oxides, chlorides and metal alloys (Kołodyńska et al., 2019). Modern societies are becoming increasingly reliant on REE due to their strong magnetism, high thermal stability, high electrical conductivity and optical properties (Goonan, 2011; Dushyantha et al., 2020). These characteristics are used in manufacturing products for primary end-consumer usage such as fluorescent light bulbs and light-emitting diodes (Humphries, 2013). Glass industries use cerium oxide for polishing and as additives to provide colour and special optical properties to glass (Goonan, 2011; Van Gosen et al., 2017). Manufacturing camera lenses requires REE such as La (Van Gosen et al., 2017). La and Ce are also used as catalysts for petrol refining and in automotive catalytic converters (Haque et al., 2014; Mancheri et al., 2019). Worldwide, the rapidly growing demand for permanent magnets has led to the development of neodymium-iron-boron alloy (NdFeB) magnets, which are the most powerful magnets known today (Humphries, 2013). NdFeB magnets are used in several products such as hard drives, cell phones, windmills (Morris, 2011), hybrid vehicles, and as actuators in aircraft (Long, 2011), and approximately 75% of these magnets are manufactured in China (Humphries, 2013). Furthermore, the rise in demand for renewable energy resources has, in turn, increased the demand for REE (Haque et al., 2014). Solar panels make use of REE as they absorb lower- energy photons, thereby increasing the efficiency of the solar panels (Strümpel et al., 2007; Dushyantha et al., 2020). Wind turbines require REE-based strong permanent magnets to design more reliable lightweight turbines with reduced maintenance costs (Heier, 2014). REE are also used in the manufacturing of military equipment such as actuators, missiles, and smart bombs (Humphries, 2010). Agricultural activities use REE as they are vital in biological structural and functional molecules (Abdelnour et al., 2019; Dushyantha et al., 2020). In the 9 agricultural sector, they are fed as additives to farmed animals to assist in fattening and enhancing the growth of livestock (Abdelnour et al., 2019). The many major end uses of REE (Table 2-2) indicate their important role in developing more efficient equipment and systems and are regarded as vitamins and a corner stone for future technologies. Table 2-2: Applications and uses of REE in the industry (Naumov, 2008). Chemical symbol Name Applications Sc Scandium High-strength Al–Sc alloys, electron beam tubes Y Yttrium Capacitors, phosphors, microwave filters, glasses La Lanthanum Glasses, ceramics, car catalysts, phosphors, pigments Ce Cerium Polishing powders, ceramics, phosphors, glasses, catalysts Pr Praseodymium Ceramics, glasses, pigments Nd Neodymium Constant magnets, catalysts, IR filters, pigments for glass Pm Promethium Sources for measuring devices, miniature nuclear batteries Sm Samarium Constant magnets, microwave filters, nuclear industry Eu Europium Phosphors Tb Terbium Phosphors Dy Dysprosium Phosphors, ceramics, nuclear industry Ho Holmium Ceramics, lasers, nuclear industry Er Erbium Ceramics, dyes for glass, optical fibres, nuclear industries Tm Thulium Electron beam tubes, visualisation of images in medicine Yb Ytterbium Metallurgy, chemical industry Lu Lutetium Single-crystal scintillators, catalysts REE applications in end products have increased drastically over the last ten to fifteen years due to its vast applications in industry. In 2015, Ce has the highest market demand (Figure 2- 1) (Goodenough et al., 2018). In the next six years, the use of REE in hybrid electric vehicles will lead to their increased demand (Roskill, 2017). By 2027 it is expected that approximately 10.1 million hybrid vehicles will be manufactured by various automotive manufacturers (Goodenough et al., 2018). The increase in demand for Nd (Figures 2-1 and 2-2) is attributed to the extensive use of NdFeB magnets (Roskill, 2017; Goodenough et al., 2018). This magnet will be the driving force behind the increase in the value of REE internationally due to its vast applications (wind turbines, automotive hybrid vehicles etc.). Furthermore, a 5% increase in REE demand growth rate is forecasted for 2021, with global REE demand continuing to grow in the years to follow (Zhou et al., 2017). Thus, new conventional REE deposits or alternate REE-bearing secondary resources need to be explored 10 to meet the future commercial extraction and production demands of these resources (Seredin and Dai, 2012; Goodenough et al., 2018). Given the growing demand and potential future economic benefits of REE, this literature review also explores other non-conventional sources of REE, especially coal-related sources. Figure 2-1: REE industrial market demand per element during 2015 (adapted from Goodenough et al., 2018). 11 Figure 2-2: Forecast of REE market demand estimate per element for 2025 (adapted from Goodenough et al., 2018). 2.2.2 Occurrences of REE The occurrence of REE are, in fact, not as scarce or rare as the name suggests, but rather are found in abundance in ultra-dispersive forms, and thus they are seldom concentrated in easily recoverable forms or in economically mineable deposits (Yang, 2019). Although REE are considered rare, they are originally found in Earth’s entire crust and are found in greater abundance than gold, platinum and silver (de Wit, 2015; Yang, 2019). In nature, REE do not originate in their elemental metallic forms but are associated with mineral compounds (Jordens et al., 2013). To date, approximately 250 REE-bearing mineral forms have been discovered (Gupta and Krishnamurthy, 2005; Jordens et al., 2013) in carbonate, silicate, oxide, phosphate, and halide forms (Van Gosen et al., 2017; Al-Ani et al., 2018; Dushyantha et al., 2020). Nonetheless, only a few minerals (monazite, xenotime and bastnaesites) have been established for commercial production of REE due to their significant economic content and recovery (Seredin and Dai, 2012; Jordens et al., 2013; Van Gosen et al., 2017). REE can be recovered as primary products or secondary by-products from REE-bearing mineral deposits and mining wastes. In addition, 12 they can be recovered as by-products from heavy mineral extractions of rutile, ilmenite and pyrite, which typically comprise REE-bearing minerals (Dushyantha et al., 2020). The REE- bearing mineral ores are found worldwide, with frequently large deposits in China, Australia, Mongolia, Russia, Finland, Brazil, India and Malaysia (Seredin and Dai, 2012; Van Gosen et al., 2017; Al-Ani et al., 2018). The fact remains that in 2018 China accounted for approximately 71% of REE production internationally (Shen et al., 2020) and since the late 1990s China contributed approximately 90% to the world REE production on average (Van Gosen et al., 2017). The world REE reserve is estimated to be over 130 million tonnes (Van Gosen et al., 2017), with China holding 43% (Ekmann, 2012) and Mongolia accounting for a further 17% (Kim et al., 2016). In 2016, China’s export restrictions affected the global demand for this resource, as out of the 210 000 tonnes produced in 2016, China supplied approximately 130 000 tonnes (Humphries, 2013). This restriction and the indispensable applications of REE have led to new efforts by several scientists (Van Gosen et al., 2017). The efforts of these scientists and scholars (such as Zhang et al., 2015; Honaker et al., 2016; Dushyantha et al., 2019) include exploring alternative sources, new deposits, secondary sources and REE wastes for the extraction and recycling of REE (Long et al., 2010; Seredin and Dai, 2012). An alternate conventional resource to REE-bearing minerals is ion-adsorbed clay ores, mined exclusively in China (Bao and Zhao, 2008) and first detected in the 1970s (Dushyantha et al., 2020). These ores exhibit extremely low concentrations of REE (0.035 % to 1 % Rare Earth oxides (REO)) but are advantageous in terms of manifesting easily leachable forms of REE (Bao and Zhao, 2008). Ion-adsorbed clay ores, also identified as weathered crust elution deposits, consist of low-grade Rare Earth ions that are adsorbed onto negatively charged alumina-silicate minerals and exhibit high economic potential consisting predominantly of HREE in easily extractable forms (Moldoveanu and Papangelakis, 2013). Many of these deposits in China are close to being exhausted, citing the need to explore alternate REE-bearing deposits (Seredin and Dai, 2012). Today, nearly all REE come from the weathered crust elution deposit and three viable REE-bearing minerals, namely bastnaesite, monazite and xenotime (Seredin and Dai, 2012). These REE-bearing minerals are discussed in the following section. 13 2.2.2.1 Bastnaesite Bastnaesite [REE(CO3)F] contains approximately 70 % of REO and is the largest source of REE available for economic extraction (Jordens et al., 2013; Dushyantha et al., 2020). Bastnaesite is a flouro-carbonate mineral that consists mainly of LREE and comprises 50 % Ce, 20 % La and does not contain any toxic minerals such as thulium and uranium (Gupta and Krishnamurthy, 2005). Bastnaesite deposits are distributed mainly in the USA Mountain Pass, California region and Mianning, Sichuan Province, China (Gupta and Krishnamurthy, 2005; Dushyantha et al., 2020). Bastnaesite bearing deposits are currently one of the main sources of LREE (Jordens et al., 2013). 2.2.2.2 Monazite Monazite [(REE)PO4] tend to be dispersed in South Africa, Australia, Sri Lanka, China, Brazil, Thailand and the USA (Kumari et al., 2015; Pandey, 2011). Monazite is ubiquitous in beach sands, metamorphic and granitic rocks and is a significant mineral present in carbonatite, iron oxide, aluminosilicates and apatite related deposits (Chen et al., 2017; Dushyantha et al., 2020). Monazite is a phosphate-dominant mineral comprising LREE predominantly and typically 10 % to 38 % La, 10 % to 30% Ce and 8 % to 15 % Nd (Thompson et al., 2011; Chen et al., 2017). REE-bearing monazite ores contain 50 % to 70% REO prior to physical treatment (Jordens et al., 2013). Nonetheless, monazite ores contain significantly elevated amounts of radioactive thorium (4 % to 12 %) and uranium (Thompson et al., 2011; Kumari et al., 2015), thus limiting the viability of REE extraction from monazite. 2.2.2.3 Xenotime Xenotime, an additional key REE-bearing mineral highly enriched in HREE and Y, comprises 67% REO (Gupta and Krishnamurthy, 2005; Jordens et al., 2013). Xenotime [(Y, HREE)PO4], a phosphate-dominant mineral, is commonly found with monazite and contains lower amounts of radioactive waste (Jordens et al., 2013). Nonetheless, xenotime is a mineral containing significant amounts of HREE like ion-adsorbed clay ores, despite its scarcity (Jordens et al., 2013). Xenotime placer deposits are distributed across the globe in China, Mongolia, Indonesia, USA, Scandinavia, South Africa and Australia (Gupta and Krishnamurthy, 2005; Lolon and Rahman, 2014; Mudd and Jowitt, 2016). 14 These three main REE-bearing minerals are presently being mined in a few countries for REE extraction, and the techniques used in processing these minerals and other REE mineral deposits are discussed in Sections 2.6 and 2.7. 2.2.3 Alternative REE-bearing minerals. As stated earlier, there are approximately 250 identified minerals associated with REE (Gupta and Krishnamurthy, 2005; Jordens et al., 2013; Dushyantha et al., 2020). These REE-bearing minerals, such as allanite, euxenite, loparite, cheralite, phosphorites, eudialyte, samarskite and others, are further described in Table 2-3 (Long et al., 2012; Jordens et al., 2013). These minerals occur in the forms of oxides, carbonates, phosphates, silicates, hydroxides and halides (fluorides) that are observed and identified in commercial or potentially commercial REE deposits (Van Gosen et al., 2017; Dushyantha et al., 2020). 15 Table 2-3: Common and predominant, economic or potentially economically feasible REE-bearing minerals (Long et al., 2012). Mineral Chemical formula REO % ThO2 % UO2 % Mineral Chemical formula REO % ThO2 % UO2 % Oxides and hydroxides Carbonates Aeschynite (Ce,Th,Ca…)[(Ti,Nb, Ta)2O6] - - - Ancylite Sr(Ce,La)(CO3)2(OH)(H2O) 46–53 0–0.4 0.1 Brannerite (U,Ca,Y,Ce)(Ti,Fe)2 O6 4 4 63 Bastnäsite (Ce, La,Y)CO3F 70–74 0–0.7 – Cerianite (Ce4+,Th)O2 82 5 – Parisite Ca(Ce,La)2(CO3)3F2 59 0–0.5 0–0.3 Euxenite (Y,Er,Ce,U,Pb,Ca)(N b,Ta,Ti)2(O,OH)6 – – – Synchysite Ca(Ce,Nd,Y,La)(CO3)2F 51 – – Fergusonite YNbO4 42–52 0–0.9 1–2.5 Tengerite Y2(CO3)3⋅n(H2O) – – – Loparite (Ce,Na,Ca)(Ti,Nb)O3 32–34 0.8 – Phosphates and fluorides Perovskite (Ca,REE)TiO3 <37 0–2 ≤0.05 Britholite (Na,Ce,Ca)5(OH)[(P,Si)O4]3 33–61 0.5– 21 0.2– 1.5 Pyrochlore (Ca,Na,REE)2Nb2O6 (OH,F) 2.6 0.2 0–10 Brockite (Ca,Th,Ce)(PO4)⋅H2O 7–24 24–45 3 Samarskite (Y,Er,Fe,Mn,Ca,U,Th ,Zr)(Nb,Ta)2(O,OH)6 – – – Cheralite (Ca,Ce,Th)(P,Si)O4 27–43 28–32 4 Uraninite (U,Th,Ce)O2 0.9–5 0.2–14 70–91 Churchite YPO4⋅H2O 50–53 – – 16 Mineral Chemical formula REO % ThO2 % UO2 % Mineral Chemical formula REO % ThO2 % UO2 % Phosphates and fluorides Silicates Britholite (Na,Ce,Ca)5(OH)[(P, Si)O4]3 33–61 0.5–21 0.2– 1.5 Allanite Ca(Ce,La,Y,Ca)Al2(Fe2 +,Fe3+)(SiO4)(Si2O7)O( OH) 3–51 0–5 0–3 Brockite (Ca,Th,Ce)(PO4)⋅H2 O 7–24 24–45 3 Chevkinite (Ca,Ce,Th)4(Fe2+,Mg)2( Ti,Fe3+)3Si4O22 40–45 0.7–0.8 – Cheralite (Ca,Ce,Th)(P,Si)O4 27–43 28–32 4 Eudialyte (Na,Ca,REE)5(Fe2+Mn) (Zr,Ti)[(Si3O9)2](OH,Cl ) 0.4–7 – ≤0.0 9 Churchite YPO4⋅H2O 50–53 – – Gadolinite Y2Fe2+Be2Si2O10 51–55 0–0.9 – Crandallite CaAl3(PO4)2(OH)5⋅H 2O – – – Gerenite (Ca,Na)2(Y,REE)3Si6O1 8⋅2H2O – – – Florencite (La,Ce)Al3(PO4)2(O H)6 18–32 1.4 – Iimorite Y2(SiO4)(CO3) 69 – – Fluocerite (La,Ce)F 83 1.6 – Kainosite Ca2(Ce,Y)2(SiO4)3CO3⋅ H2O 38 0.03 – Gagarinite NaCaY(F,Cl)6 55–57 – – Sphene (Ca,REE)TiSiO5 0–4.5 – 0.06 Gorceixite (Ba,REE)Al3[(PO4)2( OH)5]⋅H2O – – – Steenstrup ine Na14Ce6Mn2Fe2(Zr,Th)( Si6O18)2(PO4)7⋅3H2O 30–31 2 – Goyazite SrAl3(PO4)2(OH)5⋅H2 O – – – Thalenite Y2[Si2O7] 63–64 – – Monazite (Ce,La,Th,Nd,Y)PO4 35–71 0–20 0–16 Thorite (Th,U)SiO4 ≤3 72–82 8–16 Rhabdophane (Ce,La)PO4⋅H2O 58–69 0.7 0.4 Zircon (Zr,REE)SiO4 0–10.5 0–2 0–5 Xenotime YPO4 52–57 0.4 0–5 17 2.2.4 REE in coal Over the past three decades, several researchers have observed that REE are abundant in different coal types and grades (Seredin and Dai, 2012; Dai and Finkelman 2018). Coal naturally contains low concentrations of REE that are typically two and half times less in concentration related to the Earth’s upper continental crust (Seredin and Dai, 2012). In the past two decades, the existence, mineralogy and geochemistry of REE in coal have been well studied (Dai et al., 2012; Zhang et al., 2015; Hoshino et al., 2016). This had led to further developments of modern analytical techniques (Thompson et al., 2011; Dai et al., 2012; Dai et al., 2020) in detecting elevated concentrations of REE in coal. Ketris and Yudovich (2009) reported an average REE concentration of 68.5 ppm for world coals, whilst Seredin (1996) and Zheng et al. (2007) estimated that the average REE concentration in coal globally is 46 ppm. Seredin and Dai (2012) and Karayigit et al. (2000) reported an average REE content of 101 ppm and 116 ppm in Chinese and Turkish coal, respectively. Some coals in the USA have been reported with a concentration of 62 ppm. Nonetheless, the differences in the concentration of REE in world coals arise as a result of the variation in coal geochemistry due to environmental factors prevailing during its formation (Zhang et al., 2015). According to Zhang et al. (2015), the cut-off grade for REE in coal-related sources is around 115 ppm to 130 ppm, which depicts whether an ore is appropriate or suitable for REE recovery. A few scholars have suggested an outlook coefficient for REE ores to evaluate the market value and quality of the ore (Seredin and Dai, 2012; Zhang et al., 2015). The outlook coefficient (depicted in equation 2-1) employs a criterion that portrays the CREE for supply divided by excessive and UCREE of the ore of interest (Seredin and Dai, 2012; Dai and Finkelman 2018). 𝐶𝑜𝑢𝑡𝑙𝑜𝑜𝑘 = ( (∑ 𝐶𝑅𝐸𝐸) (∑ 𝑅𝐸𝐸) ) ( (∑ 𝑈𝐶𝑅𝐸𝐸) (∑ 𝑅𝐸𝐸) ) (2-1) From the equation 2-1, a higher 𝐶𝑜𝑢𝑡𝑙𝑜𝑜𝑘 would dictate an ore rich in CREE, meaning more CREE can be recovered at an increased profit. Ketris and Yudovich (2009) estimated the average 𝐶𝑜𝑢𝑡𝑙𝑜𝑜𝑘 for world coals to be 0.64. Moreover, an increase in REE content of coals will reflect a higher 𝐶𝑜𝑢𝑡𝑙𝑜𝑜𝑘. 18 2.2.4.1 The abundance of coal and REE in coal-related sources The deposition of REE within coal-bearing strata is understood to be an outcome of sedimentation mechanisms that include chemical and physical rock weathering (Li et al., 2019). This is followed by the deployment of REE with humic matter during coalification and diagenesis (Pourret et al., 2007). Many investigations reported considerably elevated concentrations exceeding 300 ppm REE in individual sites situated in the Illinois, Northern Appalachian and Central Appalachian Basins in the USA (Reid, 2018). Likewise, numerous scholars, for example, Zhang and Honaker (2019a), Montross et al. (2020) and Yang and Honaker (2020), have employed different analytical techniques such as SEM-EDX (scanning electron microscopy energy-dispersive X-ray spectroscopy), TEM (transmission electron microscopy) together with petrography to view the REE size distribution, site and association with mineral or organic matter in coal. These techniques determine the occurrence of REE in coal, as seen in Figure 2-3 (Zhang and Honaker, 2020). The SEM-EDX and TEM micrographs (depicted in Figure 2-3) indicate that REE occur in ultrafine particle size of 0.5 µm to 5 µm (Zhang and Honaker, 2020) and are predominantly concentrated in coal minerals such as monazite, xenotime, allanite, zircon and clay minerals like kaolinite and illite. Figure 2-3: SEM-EDX analysis of a non-calcined coal sample from Western Kentucky USA (Zhang and Honaker, 2020). In 2019, first-world developed countries in the European Union observed a decline in coal production, whereas countries like Indonesia and other Asian and African countries experienced increased coal production (IEA, 2019). China has been the dominant coal producer for the last three and half decades, and by May of 2020, China mined 3 960 million tonnes of raw coal (TheCoalHub, 2020). 19 South Africa, a dominant player in coal production, is the 7th largest coal producer globally, producing 254 megatonnes of coal in 2019 (Wood, 2019). Furthermore, South Africa holds the 6th largest recoverable coal deposits of 53 gigatonnes, 5% of the world’s total coal reserves and 97% of Africa’s hard coal (Mills, 2010). Coal generates 90% of the country’s electricity, provides a feedstock for Sasol’s gasification processes and is also used as coking coal for the production of steel in the country (Mills, 2010). The total reserves of REE in coal sources worldwide exceeds 50 million tonnes, which is nearly 50% of the conventional REE-bearing mineral reserves available (Zhang et al., 2015; Honaker and Groppo, 2016); thus, illustrating the economic potential of coal as a source of REE. The target REE resource from coal production and mining can be found in many different coal resources and by-products. These resources include acid mine drainage, which occurs from coal mining and discard coal stockpiles and hence assists in the natural leaching of REE (Reid, 2018; Zhang and Honaker, 2018). Coal fly ash (CFA), another resource generated as a by- product of coal combustion, is highly enriched in REE (Kolker et al., 2017). CFA manifests substantially high concentrations surpassing 400 ppm due to combustion, which leads to the consumption of the organic matter and hence leaves behind REE-rich mineral matter (Seredin and Dai, 2012; Kolker et al., 2017; Reid, 2018). Alternative target REE resources include coal tailings (discard coal), which is the by-product of the coal beneficiation plants (Reid, 2018). The clean product from the beneficiation plant is enriched in carbon, whilst the discard contains significant amounts of mineral matter, thus a potential high-grade resource of REE relative to the raw coal. Extracting REE from coal wastes forms part of an overall international goal to reduce land pollution and contamination from coal discards. Many factors indicate why extracting REE from discard coal is more advantageous than extracting from post-combustion CFA (Reid, 2018). By concentrating only on CFA, most of the REE associated with the discard coal would not have been accounted for, as this fraction would have been deposited in the discard dump before the cleaned coal was used for combustion. The fact is that the bulk of REE in coal (75%) is located in the mineral matter rather than the organic fraction, which would have been removed at the coal washing plant (Reid, 2018). The REE in CFA requires high-temperature thermal processes such as sintering for separating REE from their original co-anions, potentially improving the ease of leaching acid recovery (Zhang et al., 2020). The main challenge in concentrating REE from CFA is that REE silicates and phosphates in CFA are 20 fused into alumina-silicate glass matrices, thus resisting acid leaching and leading to a very costly process for the extraction of REE (Reid, 2018; Yang, 2019). Another reason for using discard coal is that the surrounding rock located on the roof and floor of coal seams are rich in REE, leading to optimised REE recovery during beneficiation into the high relative density fraction (Faure et al., 1996; Seredin and Dai, 2012; Reid, 2018; Yang et al., 2019). The occurrence and existence of REE in coal are either associated with the mineral matter or organic (maceral) matter, determining the beneficiation route for REE recovery (Zhang et al., 2015). There has been much discussion regarding whether REE are associated with mineral matter or organic macerals. Karayigit et al. (2000) reported that LREE in thirteen Turkish coal samples displayed a strong positive correlation with high ash yield, whereas HREE had no real correlation with ash yield. Furthermore, it was noted that HREE are enriched in coals with low clay content, thus depicting coals with low ash content have enriched levels of HREE (Seredin, 1996; Moldoveanu and Papangelakis, 2013). Furthermore, Karayigit et al. (2000) used an organic solvent extraction procedure to extract REE from coal. The result depicted a low LREE:HREE ratio; more HREE extracted implies HREE have a strong affinity for associating with organic matter. Based on these outcomes, it is evident that HREE possess a stronger affinity for the organic fraction of coal than LREE. Nevertheless, an investigation conducted by Wang et al. (2008) used the concept of organic solvent extraction of coal using carbon disulphide and N-methyl-2-pyrrolidone and reported that less than 5% of the TREE are associated with the organic matter. The bulk of the literature shows that REE distribution in coal is associated with the incombustible mineral matter, which fills in cracks and cleats within the organic matrices, rather than the organic matter (Zhang et al., 2015). Previous characterisation studies conducted on seams in the Southern Appalachian Basin in the USA found elevated concentrations of REE exceeding 700 ppm (Van Gosen et al., 2017). Nonetheless, previous research conducted in 2013 on coal from 20 preparation plants situated in the central and Northern Appalachian coalfields showed TREE in combined feed to all 20 plants at 9900 tonnes annually (Luttrell et al., 2016). The total amount of REE in the discard coal to these 20 plants was 63% (6825 tonnes), and these REE were always in the coarse discard coal after beneficiation (Honaker et al., 2017). The study showed that the REE within the coal fed into the coal preparation plants are separated and mainly concentrated into the coarse discard coal fraction. Based on this, the most promising economic potential avenue for REE recovery is the use of coal discard as a feedstock. 21 2.3 Background of coal 2.3.1 Formation of coal Coal is a sedimentary complex heterogeneous rock material consisting of different organic entities known as macerals together with variable amounts of inorganic matter (Taylor et al., 1998; Kossovich et al., 2020). Coal is classified as a fossil fuel due to its combustible properties, thus providing a source of energy in many industries (Epshtein et al., 2015). The organic structure of coal is primarily composed of carbon, hydrogen, nitrogen, oxygen and sulphur elements (Taylor et al., 1998; Maledi, 2017). Coal forms over millions of years when vegetation dies and is subsequently buried due to the build-up of silt and other sediments (World Coal Association, 2010). When vegetation is buried below ground level, the plant matter is altered due to the combined effects of high temperatures and pressures. This results in chemical and physical alterations of the plant matter, transforming the vegetation initially into peat and later into coal (World Coal Association, 2010). Approximately 360 million to 290 million years ago, during the Carboniferous Period (recognised as the first coal age), coal formation began in the Northern Hemisphere (i.e., the Laurussian region) (Taylor et al., 1998; World Coal Institute, 2005). During the Permian Period, about 250 million years ago, coal development began in the Southern Hemisphere (i.e., the Gondwana region) (World Coal Institute, 2005). The global coal deposits vary in characteristics and properties (organic maturity and maceral composition) because of different burial conditions experienced by the dead vegetation due to tectonic plate movements, climatic conditions, type of vegetation and the subsequent temperatures and pressures after coalification (Falcon, 1986). Initially, over millions of years, peat is converted to lignite, which is coals exhibiting a poor organic maturity (World Coal Institute, 2005). Lignite is a soft coal and varies in colour from brown to dark black. The continuous and progressive effects of variations in temperatures and pressures further transform lignite coal into sub-bituminous coal of increased organic maturity. Furthermore, progressive physical and chemical changes result in the formation of bituminous (hard) coals. Subsequently, the gradual increase in organic maturity establishes the formation of anthracite, which is coal with the highest carbon content (World Coal Institute, 2005; Zeng, 2008). 22 Southern African coals and coals from the Gondwana region vary significantly in rank (which reflects the organic content of the coal), host an increased amount of minerals and are relatively tricky to beneficiate (Pickel et al., 2018; Falcon, 1986). These features are responsible for the differences between Southern Hemisphere Gondwana coals and Northern Hemisphere Carboniferous coals. These differences are attributed to variations in conditions during coal formation and the subsequent historic different geological events occurring. Coal in the Northern Hemisphere was formed under humid, hot coastal forming swamps. Coals in the Gondwana region were formed under cold temperate conditions ascribed to the massive Ice Age (Falcon and Ham, 1988). A study conducted by Zeng (2008) reported that all coal-bearing deposits were formed in peat swamps. Coal-bearing deposits accumulated over time occurred in continental depressions along margins of glacial valleys, lakes and shallow seas (Falcon and Ham, 1988). With the burial of vegetation matter, highly diverse topographic and sedimentary environments existed, thus leading to numerous degrees of degradation of plant matter. Some plant matter accumulated in-situ, whilst seasonal floods carried other forms of vegetation into rivers, lakes and shallow seas (Zeng, 2008). The latter resulted in large amounts of mineral matter and smaller sediments being swept from mountainous regions into lower-lying lakes, low tidal seas and lagoons (Zeng, 2008; Ward, 2002). Plant matter in the Gondwana region was typically composed of cold-cool-temperate deciduous (probably hardwood) forests to warm savannah woodlands with reed-infested swamps (Falcon and Ham, 1988). Vegetation in swamps in the Laurussian region mainly arose from abundant ferns, lycopods, horsetails and fleshy-barked trees (Falcon and Ham, 1988). As time evolves, under the effect of temperature and pressure, vegetation increases in rank and organic maturity (Falcon and Ham, 1988). Northern hemisphere coals are higher in rank and maturity due to the deep burial of coal-bearing strata and the effects of elevated geothermal temperatures for long periods. Whereas Southern Hemisphere coal-bearing strata are predominantly buried in shallow environments and were exposed to the intrusion of hot volcanic lava, resulting in uneven rank and maturity (Falcon and Ham, 1988). The important constituents of coal are classified into three fundamental categories: organic matter, mineral matter and rank. These are discussed in Section 2.3.2. 23 2.3.2 Origin of coal macerals (organic matter) Macerals are optically homogeneous organic material with partially decomposed organic remains from vegetative material (Taylor et al., 1998; Zeng, 2008). All macerals are categorised into three groups: vitrinite, liptinite and inertinite (Zeng, 2008). Vitrinite, a significantly important maceral group that frequently occurs in bituminous coals, is obtained from the plant's woody tissues (Falcon and Ham, 1988; Zeng, 2008). Vitrinite has a medium reflectance with a higher carbon and hydrogen content than the other two groups of macerals. The important sub-maceral groups of vitrinite include collinite, tellinite and pseudo-vitrinite (Scott, 2002). Vitrinite constituents present in South African coals vary with an average amount of approximately 40% wt. (Falcon and Ham, 1988). Liptinites are derived from resins, pollen coats, cuticles, algae and other fatty excretions and are distinguished from vitrinite due to a higher hydrogen content and increased volatile matter (Falcon and Ham, 1988; Scott, 2002). Liptinite sub-maceral groups include sporinite, alginate, suberinite, cutinite, resinite and liptodetrinite (Pickel et al., 2018; Scott, 2002). Typically, South African coals reflect a minimum of 10% wt. and a maximum of 15% wt. of liptinites (Falcon and Ham, 1988). Inertinites are formed from the partial carbonisation of various flora tissues in the peat swamp stage by fire (Zeng, 2008). They are differentiated from the other macerals by their characteristic optical properties making them highly reflective. The inertinite maceral group includes fusinite, semi-fusinite, inertodetrinite, micrinite and sclerotinite (Zeng, 2008). A study conducted by Holland et al. (1989) illustrated that South African coals constitute 20% to 80% wt. of inertinites. Vitrinite and liptinite are reactive groups as they readily ignite under combustion conditions (Falcon and Ham, 1988). However, inertinites are unreactive as they require more oxygen and heat to ignite. The macerals play an important role in the application of coal, and much research has identified minute amounts of REE associated with coal’s organic macerals. Therefore, it is important that this study provided a review of the coal’s constituents. 2.3.3 Coal microlithotypes Microlithotypes, or the types of maceral association, are regarded as the natural assemblages of macerals at a microscopic level (Taylor et al., 1998). They contain a minimum of 5% maceral and more than 20% mineral matter. Together with the maceral content, microlithotypes are composed of 20 vol. % to 60 vol. % of silicate or carbonate minerals or 5% sulphide, thus they are regarded as carbominerite irrespective of the macerals present (Suárez-Ruiz and Ward, 24 2008). Microlithotypes play an integral role in the density, liberation, gasification, combustion, strength and coking behaviour of coal (Falcon and Ham, 1988). The chemical characteristics of microlithotypes are similar to their prevailing macerals, whereas their physical properties exhibit a combined effect of their occurrence with macerals (Falcon and Falcon, 1987). The nature and extents of the organic forms establish the coal type (Falcon and Ham, 1988). 2.3.4 Mineral matter in coal Improvements in highly efficient analytical techniques have led to identifying various minerals associated with coal strata. Approximately 200 different minerals have been identified in coal seams globally (Finkelman et al., 2020). According to Ward (2002) and Xiuyi (2011), mineral matter in coal is referred to as the inorganic matter distributed and associated with the coal, which typically occurs in five different forms: I. Crystalline mineral particles II. Non-crystalline inorganic particles III. Inorganic elements associated with organic macerals IV. Inorganic elements dissolved in pore water and surface water in coal V. Dissolved salts Analogous to organic matter, mineral matter in coal is formed via complicated geological processes and events (Pickel et al., 2001; Maledi, 2017; Wagner et al., 2018). These events are linked to detritus deposits in peat-forming swamps, amount of sediment weathering, temperature, pressure, pH, source rock and the migration of components within coal-bearing seams (Maledi, 2017). Mineral matter in coal can be macroscopic or microscopic (Xiuyi, 2011). Macroscopic minerals occur in round pellets, lenticels, bands, nodules and sometimes in dispersed crystals. Microscopic minerals are seen as fractured crystalline fragments, isolated euhedral minerals, microscopic nodules and sub-microscopic crystal accumulates (Xiuyi, 2011). Furthermore, minerals tend to be hosted in cracks, fractures and cleats within the maceral matrices. It is important to note that some minerals occur in extremely fine grains of particles not exceeding several microns and are also dispersed in organic macerals, which can only be viewed under an electron microscope (Xiuyi, 2011). Minerals in coal arise from sedimentary rocks such as shales, siltstones and sandstones interbedded within fractures and between coal seams (Falcon and Ham, 1988). Understanding the properties and the characterisation of coal provide 25 information on the inherent size and distribution of minerals interbedded in coal for the beneficiation and recovery of REE from coal. 2.3.4.1 Formation of mineral matter in coal The evolution regarding the incorporation of mineral matter in coal may be categorised into the following groups based on their origin (mineral matter) and mode of formation (Xiuyi, 2011; Wagner et al., 2018): I. Plant-derived minerals II. Terrigenous detrital minerals III. Chemical and biochemical minerals (authigenic minerals) IV. Minerals created through diagenetic alteration Minerals are accumulated in coal at different times, which define their genesis. The three significant roots (origin) of mineral matter in coal are epigenetic, syngenetic and detrital (Ward, 2002), which are discussed below. The syngenetic formation of minerals occurred from plant-derived coal forming matter (Maledi, 2017; Wagner et al., 2018). These minerals fall into two categories regarding the time of formation. Early syngenetic minerals are formed during the initial stages of peatification, soon after the burial of peat or other sediments, when major solidification has not taken place yet (Ward, 2002). The late syngenetic mineral formation occurs during the humidification- gelification of coal, associated with the deepened buried advanced peat formation stages of coalification (Ward, 2002). Late syngenetic mineral formation is because of chemical processes occurring between coal constituents and the environment. As a result, syngenetic minerals occur in ultra-dispersive forms and are closely distributed with the organic matter, thus presenting difficulties when beneficiating syngenetic minerals (Maledi, 2017). Important syngenetic minerals found in coal include siderite nodules, microcrystalline fragments of pyrite, and void filling compounds such as carbonates, quartz, kaolinite, sulphates, clays and phosphate minerals (Ward, 2002). Epigenetic minerals are deposited into coal seams and fractures after coal compaction or after the coal source has reached its present rank (Xiuyi, 2011; Wagner et al., 2018). Epigenetic minerals are also regarded as adventitious or excluded mineral matter found in void fillings and coal fissures. Typical mineral matter in coal via epigenetic mineralisation consists of silica, 26 clays, illite, dolomite, apatite, pyrite, ankerite, siderite, marcasite, dawsonite and chlorites (Ward, 2002). Epigenetic minerals tend to be large and are easily liberated during the milling of coal and beneficiation processes (Maledi, 2017). Detrital mineralisation typically occurs when minerals were deposited into the peat by either moving water, wind or are introduced into the peat by epiclastic and pyroclastic processes (Ward, 2002). Epiclastic procedures involve lithic clasts and minerals liberated by ordinary weathering processes from pre-existing merged rocks (Xiuyi, 2011; Wagner et al., 2018). Traditionally, epiclastic minerals are derived from the erosion of volcanoes or ancient volcanic terrains. Pyroclastic minerals are formed via the breakdown of magma, as gases are released from volcanic vents in the air or beneath water (Ward, 2002). Detrital minerals include sporadic bands of clay minerals, feldspar, tonsteins and quartz fragments, which are all thoroughly mixed matter (Ward, 2002). Nonetheless, it is typically difficult to determine the origin of each mineral, as many minerals of different origins can occur in the same coal seam (Wagner and Matiane, 2018). A fraction of some of the mineral species occurring in coal reserves is presented in Table 2-4, and a portion of these minerals cannot be viewed by the naked eye (Maledi, 2017; Wagner et al., 2018). Nonetheless, a few mineral groups and some individual minerals such as silicates, pyrite and calcite can be seen with the naked eye if granted the correct specimen (Ward, 2002). This is the core reason for the evolution of enhanced analytical techniques to view microscopic amounts of mineral matter in coal. Minerals in coal can be abundant, rare, common or frequently seen in most coals (Falcon and Ham, 1988; Ward, 2002; Xiuyi, 2011). Likewise, analytical techniques such as X-ray diffraction (XRD), SEM-EDX, petrographic analysis and X-ray fluorescence (XRF) assist in determining and viewing the minerals present in coal. Mineral matter can affect the grade of the coal as low-grade coals tend to have high amounts of mineral matter compared to their counterpart high-grade coals (Maledi, 2017). Although the geological occurrence of minerals in coal is regarded as having deleterious effects on coal operations as they release toxic elements and contribute largely to CFA and discard coal stockpiles, they do have beneficial features to the industry (Honaker and Groppo, 2016). CFA and discard contain REE and are future reserves for the recovery of CREE. Likewise, many of these minerals found in coal (Table 2-4) are known to bear REE such as apatite, illite, kaolinite, xenotime, monazite, muscovite, biotite, calcite and other clay minerals (Faure, 1993; Akinyemi et al., 2012; Wagner and Matiane, 2018). 27 Table 2-4: Minerals in coal deposits (adapted from Saxby (2000) and Xiuyi (2011)). Mineral groups Minerals Arsenide Lollingite Sulphides Pyrite, marcasite, galena, sphalerite, melnikovite, cooperite, siegenite, pyrrhotite, chalcopyrite, bornite, tetrahedrite, tennantite, argentite, chalcocite, realgar, ferroselite, ullmannite, platarsite, arsenopyrite, glaucodot, stibnite, bismuthinite, millerite, cinnabar, orpiment, herzenbergite, covellite Selenides Penroseite, clausthalite, hakite Oxides Quartz, chalcedony, hematite, cristobalite, uraninite, ilmenite, corundum, spinel, chromite, magnetite, hetaerolite, anatase, rutile, brookite, cassiterite, columbite Hydroxides Limonite, bauxite, goethite, lepidocrocite, diaspore, boehmite, gibbsite, brucite, portlandite, chalcophanite Silicates Kaolinite, illite, sericite, montmorillonite, mixed-layer clay minerals, dickite, halloysite, chlorite, ferrihalloysite, muscovite, hydromuscovite, feldspar, zircon, leucite, zeolite, tourmaline, coffinite, andalusite, garnet, olivine, kyanite, topaz, staurolite, sphene, epidote, allanite, pyroxene, actinolite, hornblende, chloritoid, talc, serpentine, pyrophyllite, glauconite, biotite, gadolinite, barytolamprophyllite, fenaksite, canasite Carbonates Calcite, dolomite, siderite, ankerite (ferroan dolomite), dawsonite, strontianite, aragonite, magnesite, barytocalcite, azurite Vanadate Carnotite, tyuyamunite Phosphates Apatite, phosphorite, xenotime, monazite, svanbergite, berlinite, florencite, crandallite, gorceixite, autunite, torbernite, vivianite, goyazite Tungstate Scheelite Sulphates Barite, gypsum, anhydrite, celestite, linarite, alunite, natrojarosite, jarosite, alunogen, epsomite, kieserite, melanterite, szomolnokite, mirabilite, polyhalite Nitrates Nitronatrite Halides Halite, sylvine, bischofite, fluorite Amorphous mineral Opal, volcanic glass 2.3.5 Rank of coal The rank is the degree of maturation or metamorphism of coal achieved by process variations such as temperature, time, pressure as a result of peat burial or vicinity to heat following peat accumulation (Falcon and Ham, 1988; Ward and Suárez-Ruiz, 2008). The transformation of peat into different coal ranks passes through progressive stages, i.e., peat → lignite (brown coal) → sub-bituminous → bituminous → semi-anthracite → anthracite → graphite (Falcon and Ham, 1988). Coal’s position in these coalification levels is regarded as its rank. The maturity is enhanced or affected by significant deviations in the macerals’ physical, chemical and technological properties, particularly in the vitrinite and liptinite groups (Falcon and Ham, 28 1988). The inertinite retains its original structure, chemistry, form and, for the most part, undergoes minimum change throughout coalification. The quantitative determination of coal rank in reactive macerals (liptinite and vitrinite) can be achieved by determining its proportion of volatile matter, carbon content, moisture content and hydrogen (Falcon and Ham, 1988). However, using volatile matter as a parameter to determine coal rank is not reliable for establishing the rank of coals depleted in volatile matter. This is because the thermal dissociation of minerals accounts for up to 60% of volatile matter in coal (Falcon and Ham, 1988). Carbonate minerals such as calcite contribute to carbon dioxide, whilst clay minerals contribute to water vapour and basic oxides. In addition, the coal inertinite contributes significantly higher carbon content with lower volatile matter, thus indicating another source of carbon achieved by rank alone. Coincidently the presence of high proportions of liptinite indicates a low carbon content with increased amounts of volatile matter (Falcon and Ham, 1988). Thus, volatile matter and carbon content do not produce reliable indicators for coal rank (Falcon and Ham, 1988). Hence a more uniform standard approach is used to determine the rank of coal by measuring the light reflected from a polished vitrinite particle surface under oil immersion (Falcon and Ham, 1988; Ward and Suárez-Ruiz, 2008). This reliable standard technique referred to as vitrinite reflectance (SANS 7404-5) is preferred because the optical reflectance of vitrinite increases with an increase in the rank of coal, as depicted in Figure 2-4 (Ward and Suárez-Ruiz, 2008). As the carbon content increases, the optical reflectance of vitrinite and the coal’s C:H ratio increases, whereas the volatile matter decreases (Falcon and Ham, 1988). Low-rank coals (lignite, sub-bituminous and bituminous) are high in volatile matter, moisture content, porosity and reactivity and are thus more easily oxidisable than higher rank coals (semi-anthracite, anthracite and graphite). High-rank coals are more difficult to ignite and require higher temperatures for combustion. Due to rank being only measured by optical microscopy, the only requirement for determining the rank is the presence of vitrinite (Falcon and Ham, 1988). 29 Figure 2-4: Variation in the key coal properties for rank advancement (from Ward and Suárez- Ruiz (2008)). 2.3.6 Coal grade The grade of coal is an assessment of the amount of mineral matter and organic matter present in coal. Furthermore, it reflects the extent to which the accumulation of plant vegetation has been kept free of contamination by inorganic mineral matter during and after peat burial and during the coal’s rank advancement (Ward and Suárez-Ruiz, 2008). High-grade coal consists of a low amount of mineral matter and an increased amount of organic matter (Suárez-Ruiz and Ward, 2008). 2.4 Overview of South African coals 2.4.1 South African coal reserves South Africa’s main source of energy generation is dominated by coal-powered stations, and the country is a major participant in the global coal market (Mills, 2010). South Africa exports about 25% (Wagner and Matiane, 2018) of the coal produced annually and is the dominant player in the African coal market as it holds 95% of Africa’s coal reserves (Mills, 2010). The recoverable coal reserves in South Africa are estimated to be 58 billion tonnes (Mills, 2010). Coal exports generate more income than any other commodity in South Africa (Ratshomo, 2016). 30 Currently, the Witbank and Highveld coalfields are the main source of coal production in South Africa and represent 75% of the country's coal production (Ratshomo, 2016). The Waterberg coalfield is regarded as the most promising coalfield for future coal mining in the country (Faure et al., 1996; Moumakwa, 2010) as the Witbank and Highveld coalfields are close to exhaustion (Jeffrey, 2005; Pooe and Mathu, 2011). South Africa hosts relatively shallow coal mines with thick coal seams close to the surface, thus reducing mining costs (Eberhard, 2011). More than 50% of the coal mining occurs in opencast mines, with the balance being mined underground (Eberhard, 2011). Coal in the South African region generally contains a variable ash content and can reach up to 65% in ash content underground (Eberhard, 2011). Thus, export-grade coal and coal required for power generation in South Africa requires prior beneficiation (Eberhard, 2011). The beneficiation of the coal generates discard coal. The estimated discard coal available in stockpiles in South African is more than one billion tonnes, and more discard is constantly being added at a rate of approximately 60 million tonnes per annum (Eberhard, 2011; van Breugel et al., 2019). Discard coal provides many environmental challenges but can also provide a potential source for economic gain if it can be recycled in other processes. As reported by Luttrell et al. (2016), the amount of discard coal available from coal preparation plants, as well as the fines generated during mining activities, provide a potential source for REE recovery. 2.4.2 Geological setting of coalfields in South Africa South African coals were formed during the Permian Period to the mid-Triassic Period and are located in the Gondwana Karoo Basin (Falcon and Ham, 1988). The formation occurred under low cool-climatic temperatures to warmer climatic regimes (Falcon, 1986). South Africa hosts 19 coalfields, with 18 coalfields belonging to the Karoo Basin and the remaining coalfield lying in the Kalahari Basin (Jeffrey, 2005). These 18 coalfields are predominantly situated in Mpumalanga, KwaZulu-Natal, Limpopo and the Free State, with reduced amounts in Northwest, Gauteng and the Eastern Cape (Jeffrey, 2005). The stratigraphy of the Karoo Basin is composed of Dwyka, Ecca, Beaufort and Stormberg groups. The majority of coalbeds in South African occur in the Ecca Group of the Karoo Basin, with the Waterberg coalfield lying in the embayment of the Kalahari Basin (Faure et al., 1996). Figure 2-5 represents the 19 coalfields situated in the country, which further emphasises the abundance of coal reserves in S