Flotation of Yxsjöberg historical tungsten ore tailings

(1)

Flotation of Yxsjöberg historical tungsten ore tailings

Princess Rochelle O. Gan

Natural Resources Engineering, master's level (120 credits) 2019

Luleå University of Technology

Department of Civil, Environmental and Natural Resources Engineering

(2)

Flotation of Yxsjöberg historical tungsten ore tailings

Princess Rochelle O. Gan

Supervisors: Jane Mulenshi, Saeed Chehreh Chelgani Examiners: Jan Rosenkranz, Lev Filippov

Division of Minerals and Metallurgical Engineering

Department of Civil, Environmental and Natural Resources Engineering Luleå University of Technology

August 2019

(3)

i

Acknowledgements

I would like to give my utmost gratitude and respect for the following people, for whom without their presence and help I would not be able to finish my work:

To Professor Jan Rosenkranz, Professor Lev Filippov, and Saeed Chehrehlgani for guiding me during my thesis work and sharing with me their expertise on flotation. Thank you for the wisdom and inspiration to do good with my study.

To Ms. Jane Mulenshi, my supervisor, for keeping her doors open for my numerous questions. For worrying with and for me during the most stressful moments of my thesis. May my findings aid your pursuit for a doctorate and may you mentor more students in the future.

To Malin Johansson, who has been nothing but a ray of sunshine during very dark days in the lab. Who never failed to cheer me on and brighten my day with a few minutes of chitchat.

To Isabella Johanesson, Erlinda Olsson, and Ligaya Åström, the lovely ladies of Luleå who became family and gave me a home. Thank you for welcoming me into your lives with open arms, and making Luleå a warmer place for me to live in.

To Chryselle Mancenido and Carol Mafra, for always taking time out of their schedules to check up on me to make sure I’m okay, and for letting me ask questions any time of the day. Thank you for your care and support. You girls served as a safe haven even with the distance and I will always be grateful for my connection with you. To Chris Soto and Javad Ghanei, my fellow EMerald students who were also based in LTU, for their company and help. Working in the lab was more enjoyable with you two in it. And to the rest of my EMerald classmates, who have all been like my sisters and brothers for the last 2 years. It was a pleasure riding this wave with you all!

To Karlo Flores, for being my sponge and always lending me a listening ear, for always being there to remind me that I am strong when I was nothing but doubtful. Thank you for your patience and love.

To my dearest parents, Romeo and Cecilia, for their unwavering support and unconditional love.

Without them I would not have had the courage to go on and face the challenges I was thrown everyday.

Thank you for always believing in me. To my siblings Precious, Patricia, and Prinz, and my beloved maltese Angel: thank you for the calls when I’m down, the endless laughs, the inspiration, and for the love. You all are the rock that keeps me grounded and the wings that lift me up. To Yaya, thank you for watching over me from Heaven. As always, this is for all of you.

And lastly, to the creator of the Universe, for without Whom this wouldn’t have been possible. Thank you for allowing me to have the strength, grit, and capability to persevere and finish my work. I owe this all to You.

(4)

ii

Summary

Tailings dams, due to their sheer size and metal content, pose safety and environmental hazards. The Småltjarnen repository, which hosts material from the largest known tungsten mineralization in Sweden named the Yxsjöberg deposit, is estimated to contain 2.2 million tons of material from previous operations when recovery rates of scheelite, chalcopyrite, and fluorite were low. The repository is also observed to contain at least four of the listed critical raw materials by the European Commission in 2017, namely tungsten, fluorite, beryllium, and bismuth. The amenability of this tailings repository as secondary sources for valuable minerals and metals supports the Raw Material Initiative and the drive of the European Union towards a more circular economy.

This masters thesis is part of the REMinE (Improve Resource Efficiency and Minimize Environmental Footprint) project that aims to evaluate the amenability of historical mining waste for re-processing from the technical, economical, and environmental perspectives. The study focuses on work package 3 of the project: Identification of new processing methods for mine waste.

Previous work on this repository includes geochemical characterization and physical separation through magnetic and gravity separation tests. Since scheelite, the main mineral of the deposit, is commonly recovered through combinations of gravity separation and flotation methods, it is imperative to investigate the response of the material to flotation tests.

Further characterization work and flotation tests were rendered on samples from sampling location 6 (60°02'33.6"N 14°46'30.8"E). Previous work was done on samples from and near sampling location 1, which is southwest from location 6. Characterization methods performed on the material included elemental analysis through ICP-SFMS, X-ray diffraction measurements, and mineral liberation analysis as well as physical characterization through particle size distribution analysis and determination of specific gravity. Based on MLA, the material from the main samples showed good liberation by free surface which is important for flotation processes. Mineral association also showed low percentages between scheelite and other Ca-bearing minerals which is a main concern for scheelite flotation.

Comparison of characterization work between the two sampling locations allows information on the repository at a larger scope.

The common scheelite flotation collector sodium oleate, as well as novel formulations Atrac 2600 and Berol 8313 from Nouryon were tested in combination with the depressant sodium silicate. Based on mass recovery, grade and recovery, selectivity, required dosage, and the degree to which it is environmentally safe, Atrac 2600 at 400 g/t is deemed to produce the most positive results.

Based on characterization of the material and flotation tests, an estimated 222,200 tons of -75 µm can go directly into the flotation circuit with an average grade of 0.2768% WO3 and 0.195% Cu.

Approximately 15,000 tons at 0.50% WO3 and 14,000 tons at 0.224% Cu can be recovered at a single flotation. Rough mass balance of the process flowsheet indicates a Cu concentrate at 30% Cu will produce 880.50 tons, while a saleable scheelite concentrate at 65% WO3 will yield 672.16 tons from only the original -75 µm fraction of the material. Including the gravity separation tails intended to join the flotation circuit after grinding, an estimated 1,205.54 tons of scheelite at 65% WO3 and 1.860.20 tons of 30% Cu concentrate can be produced. Aside from this, the gravity separation circuit will still be able to yield 188,000 tons of 0.92% WO3 concentrate, which can be further studied if it can be beneficiated to a saleable product.

Value estimation of the products for the recommended flowsheet indicate a total of US$ 17 million for the WO3 and Cu concentrates to be produced from the readily amenable -75 µm fraction to be treated directly by flotation. Inclusion of the gravity separation tails mass into the flotation feed yields a value estimated at US$32 million.

(5)

1

List of Figures

Figure 1. Location of the Smaltjärnen historical tungsten tailings repository in Yxsjöberg, Sweden (outline based on Mulenshi et. al. (2019), white arrow indicates sampling location 6, yellow arrows

for sampling location 1) ... 7

Figure 2. Flowsheet of the Flotation division of the Yxsjöberg dressing mill (Rothelius, 1957) ... 13

Figure 3. Flotation of scheelite and calcite at constant sodium oleate concentration as shown in (Patil & Nayak, 1985) ... 18

Figure 4. (L)Effect of sodium silicate depressant on floatability of minerals (1-scheelite, 2-calcite, 3- fluorite, 4-garnet, 5-quartz) with NaOl collector from Yongxin & Changgen, (1983); (R) scheelite flotation porfermance depending on sodium silicate dosage ... 20

Figure 5. Effect of NaOl dosage on scheelite and calcite particles on flotation recovery, source: (Gao et al., 2016) ... 21

Figure 6. New flowsheet for tungsten mineral extraction in Shizhuyuan dressing plant (Han et al., 2017) ... 26

Figure 7. Visual representation of drill cores in sampling location 6, relative to depth and separation into layers ... 28

Figure 8. Actual sample material (drill cores sub-sampled based on differences in texture and color) ... 29

Figure 9. (A) Clausthal flotation cell, (B) side profile, (C) back width, (D) cell height, (E) mechanical rotor ... 32

Figure 10. Flotation flowsheets for pure scheelite mineral and Yxsjöberg tailings material ... 33

Figure 11. Flotation testing scheme ... 34

Figure 12. Molecular interpretation of sodium oleate, source: U.S. National Library of Medicine (2019) ... 34

Figure 13. Particle size distribution of samples in location Latitude 60.043, Longitude 14.775 ... 37

Figure 14. Moisture vs. D80 of samples (without identified bottom layers) ... 38

Figure 15. Density values of samples differentiating upper layers to assumed bottom layers (6-3, 6D- 2, 6E-2, 6F-2) ... 39

Figure 16. Elemental distribution by sub-sample of drill cores 6 and 6F... 39

Figure 17. Elemental distribution by size fraction of drill cores 6 and 6F... 40

Figure 18. Distribution of oxides in drill core 6 and 6F by sub-sample and size fraction ... 40

Figure 19. Elemental concentration and distribution of 6-1 ... 41

Figure 22. Elemental concentration and distribution of 6F-1 ... 42

Figure 23. Elemental concentration and distribution of 6F-2 ... 43

Figure 24. Elemental concentrations of all samples in location 6 (Latitude 60.043, Longitude 14.775) ... 43

Figure 25. Elemental S concentration of samples in location 6 ... 44

Figure 26. Elemental distribution of different layers in remaining drill holes of location 6 ... 45

Figure 27. Cross-plot of WO3 and CaO content of all samples... 45

Figure 28. Modal distribution of scheelite in different size classes of samples from drill core 6 and 6F ... 48

Figure 29. Modal composition of elements in different size classes of 6-1 and 6-2 ... 48

Figure 30. Modal composition of -75 µm of samples 6A-6E ... 49

Figure 31. Boxplot of element residuals from element-to-mineral conversion ... 50

(8)

4

Figure 32. Elemental distribution of W based on MLA... 52

Figure 33. Elemental distribution of Cu based on MLA ... 52

Figure 34. Elemental distribution of Zn based on MLA ... 53

Figure 35. Elemental distribution of S based on MLA ... 54

Figure 36. Elemental distribution of main economic minerals... 54

Figure 37. Modal Mineralogy by MLA ... 55

Figure 38. Modal mineralogy of different size fractions of 6F-1 ... 56

Figure 39. Composition of locked and liberated scheelite particles of 6F-1... 57

Figure 40. Composition of locked and liberated scheelite particles of 6-1 and 6-2 ... 57

Figure 41. Mineral locking of minerals of interest in -75 µm ... 58

Figure 42. Composition of minerals associated to the minerals containing elements of interest ... 59

Figure 43. Classification of particles based on composition and free surface with corresponding ease of recovery. Source: (Cropp, 2013) ... 60

Figure 44. Mineral Liberation by composition and free surface of scheelite ... 61

Figure 45. Mineral Liberation by composition and free surface of fluorite... 62

Figure 46. Mineral Liberation by composition and free surface of chalcopyrite ... 62

Figure 47. Mineral Liberation by composition and free surface of danalite ... 63

Figure 48. Mineral Liberation by composition and free surface of bismuthinite ... 63

Figure 49. Mineral Liberation by composition and free surface of pyrrhotite ... 64

Figure 50. Mineral Liberation by composition and free surface of cassiterite ... 64

Figure 51. Mineral Liberation by composition and free surface of sphalerite ... 65

Figure 52. Mass recovery of flotation tests based on different collectors and dosages ... 66

Figure 53. Elemental distribution in concentrate product of preliminary flotation tests of Yxsjöberg tailings ... 67

Figure 54. W grade and recovery comparison for different collectors and dosages ... 68

Figure 55. (L) Cu and (R) CaO grade and recovery comparison for different collectors and dosages . 69 Figure 56. Recovery curves of CaO and W for (L) Berol 8313 and (R) Atrac 2600 at different dosages ... 69

Figure 57. Difference of CaO and W recovery in products from Berol 8313 and Atrac 2600 at 400 g/t ... 70

Figure 58. Behavior of all elements of interest based on different dosages of Berol 8313 and Atrac 2600 ... 71

Figure 59. Comparison of scheelite composition in flotation concentrate and tail product ... 71

Figure 60. Comparison of (L) fluorite and (R) pyrrhotite compositions in flotation concentrate and tail product ... 72

Figure 61. Mass recovery of concentrate product of other drill cores floated with Atrac 2600 at 400 g/t ... 73

Figure 62. Elemental concentrations of flotation concentrate of upper layers 6A to 6E ... 74

Figure 63. Elemental assay grade and recovery for flotation concentrate of 6-1 and 6A-1 ... 74

Figure 64. Correlation plot from R of CaO, W, and Cu feed and concentrate assays and recovery of all flotation samples from location 6 ... 75

Figure 65. Mass recovery of concentrate product of pure mineral flotation tests ... 76

Figure 66. Recommended flowsheet for Yxsjöberg historical tungsten ore tailings ... 79

(9)

5

List of Tables

Table 1. Yearly output and corresponding production value of Yxsjöberg mine (Rothelius, 1957)... 8

Table 2. Description of the different work packages included in the REMinE Project ... 10

Table 3. Average mineral grades of Yxsjöberg ore; source: (Gräsberg & Mattson, 1979) ... 11

Table 4. Indicated reasoning behind variation in mineral composition of the repository as enumerated in Hällstrom (2018)... 11

Table 5. Grade and recovery of material under novel process implemented by AB Statsgruvor (Gräsberg & Mattson, 1979) ... 15

Table 6. Maximum allowable content of impurities ... 17

Table 7. Similar properties shared by scheelite, fluorite, and calcite that can complicate separation by flotation ... 19

Table 8. List of collectors and depressants studied for scheelite flotation ... 23

Table 9. Set-up of seach peak, background parameters, and elemental restrictions based on Khavari (2018) implemented on XRD analysis of samples ... 30

Table 10. Description of rounds for Element-to-Mineral conversion of Yxsjöberg samples ... 31

Table 11. Corresponding D80, density, and moisture values of all samples ... 37

Table 12. Weighted average concentration of -75 µm of all samples in location 6 (% for S, SiO2, Al2O3, CaO, and Fe2O3) ... 46

Table 13. Main minerals containing elements of interest in the Yxsjöberg tailings ... 46

Table 14. Summary of detected mineral phases in 6 and 6F-1 by XRD analysis ... 47

Table 15. Average difference in elemental assays between EMC and chemical analysis ... 50

Table 16. List of minerals identified from samples in location 6 (Latitude 60.043, Longitude 14.775) 51 Table 17. Mineral grouping for minerals detected by MLA ... 55

Table 18. Summary of mineral locking of minerals of interest ... 58

Table 19. Free surface percentage of minerals containing elements of interest ... 59

Table 20. Summary of mineral liberation percentages ... 65

Table 21. WO3, Cu, and CaO assays of feed material for flotation tests ... 66

Table 22. Calculated grade and recovery of W, Cu, and CaO in preliminary flotation tests ... 67

Table 23. Calculated grade and recovery of Be, Bi, Sn, and Zn in preliminary flotation tests ... 70

Table 24. Criteria for choosing ... 72

Table 25. Difference in % between values of 6-1 and 6A-1 ... 74

Table 26. Expected mineral route in proposed flowsheet ... 79

Table 27. Rough mass balance of recommended flowsheet for Yxsjöberg tailings ... 80

Table 28. Rough mass balance for recommended flowsheet including gravity separation tails in flotation circuit ... 80

Table 29. Average elemental grades of Yxsjöberg tailings material based on location and size fraction ... 81

Table 30. Value estimation for re-processing of Småltjarnen repository material of the -75 µm (80% recovery) ... 82

Table 31. Required number of stages to enrich scheelite concentrate to a saleable product ... 83

(10)

6

1 Introduction

Continuous increase in demand for metal and mineral products are expected due to population growth and urbanization, as well as increase in the material standard of living. This strong demand coupled with decreasing mineral grades and less deposits that are easily accessed has put pressure on the global raw material market. Ore deposits have been the primary source of raw materials from the economic interest. However, the focus on historical tailings being considered as potential secondary sources due to their inherent high mineral and metal content and advances in mineral processing methods are becoming more and more common. This characteristic of historical tailings containing considerable amounts of minerals may be brought about by inefficient extraction processes, low demand, or low metal prices at the time of deposition. Aside from being a potential resource, repositories holding high amounts of these materials pose environmental risks such as acid mine drainage requiring attention.

Mine waste comes from the extractive industry as products from exploration, mining, and processing activities governed by the different legislations. Waste rock pertains to non- or low-grade mineralized rocks that was moved during the extraction processes. On the other hand, processing waste or what is commonly called “tailings” material, is the product that is left after applying separation processes on the ore to obtain a concentrate. A rigid definition of the grade and components of tailings material cannot be given as it varies from mine site to mine site, depending on the deposit and what minerals are being recovered, what the benefication method is used, what the recovery rates are, what is the cut-off grade of what mineral, which all varies over time.

All over the world structures called tailings dams are constructed to contain these tailings materials.

Even post-closure these dams usually remain and the number of tailings repositories continue to grow and take up large land spaces as more mines start to operate. Throughout history, several failures of these structures have happened. Re-processing the tailings material will not only reduce safety hazards significantly and lower the environmental risks and foot print, but will also potentially contribute to decreasing the dependence of bodies such as the European Union by providing an alternative source of metals and minerals.

The list of critical raw materials for the European Union is updated at least every three years. A list of 14 critical raw materials was established in 2011, with a revision for 20 critical raw materials in 2014.

The list presented in 2017 is the result of a third asessment carried out for 78 raw materials, which contains 27 critical raw materials. Beryllium, Bismuth, Fluorspar, and Tungsten were considered by Hällström (2018) as elements for potential concern in her study “Geochemical Characterization of Historical W, Cu, and F Skarn Tailings at Yxsjöberg, Sweden”. These same elements are listed as critical raw materials in 2017 by the European Commission (European Commission, 2017). Elements that provide economic value such as Copper, Tin, and Zinc but are not critical based on the list of the European Commission but are among the main components of the Yxsjöberg tailings and will be considered in the following discussions. Pyrrhotite was determined to be the main sulfide mineral in the tailings by Hällström (2018) and with the potential to cause acid mine drainage and will be included in the elements that will be looked into in this study.

By studying possible methods to reprocess and extract valuable material from tailing repositories, the historical mine sites can be potentially “cleaned up” to decrease environmental impact risks, as well as provide another resource for minerals and metals. This supports the mission of the European commission for the Raw Materials Initiative and their drive towards a circular economy. Previous physical separation methods for magnetic and gravity separation has been done on Yxsjöberg historical tungsten tailings. However, scheelite, the main mineral of interest in the tailings material is usually recovered through flotation. Hence, the need to study the amenability of the material to be reprocessed through this method is presented.

(11)

7

1.1 Yxsjöberg tungsten deposit

The Yxsjöberg historical mine is hosted in the region of Bergslagen in central Sweden (Figure 1). It is the largest known tungsten mineralization in the country where 90% of the total tungsten produced by Sweden came from. The deposit is identified as a skarn-hosted tungsten-copper-beryllium-fluorite style, with the mineralization thought to be brought about by hydrothermal solutions from a late Svecokarelian granite. The deposit consists of three mineralized bodies, namely Kvarnåsen, Nävergruvan, Finngruvan.

The three bodies all lie in the same horizon and contain 0.24 – 0.32 wt.% of W, 0.16 wt.% of Cu and 5 – 6 wt.% of fluorite (Romer & Öhlander, 1994; Rothelius, 1957). The main skarn minerals according to Hällström (2018) are pyroxenes, amphiboles, garnets, and fluorite. Based on Mattson (1982), the ore material consists of hedenbergite-diopside skarn, hornblende skarn, and grossalurite-andradite skarn.

Scheelite occurs as coarse grains unevenly disseminated in the diopside skarn, but spread evenly in finer sizes in the hornblende skarn (Mattson, 1982). Chalcopyrite, fluorite, and pyrrhotite, pyrite, and small amounts of magnetite are also present (Ohlsson, 1979). Other accessory minerals present in the deposit include calcite, helvite, molybdenite, wolframite, sphalerite, apatite, titanite, chlorite, epidote, allanite, zircon, and hematite (Romer & Ohlander, 1994).

Figure 1. Location of the Smaltjärnen historical tungsten tailings repository in Yxsjöberg, Sweden (outline based on Mulenshi et. al. (2019), white arrow indicates sampling location 6, yellow arrows for sampling location 1)

The earliest records of mining in Yxsjöberg is in the year 1728, through small scale mining of copper which continued intermittently until the 19^th century. The earliest recorded production of tungsten was in 1918. A new concentrator plant was then built in 1937, and in 1951 a roasting furnace and gravity separator was added. A circuit to recover fluorite and chalcopyrite through flotation cells was added in 1956. At this point the recoveries of scheelite, fluorite, and chalcopyrite were 75% (previously at 50%

before the addition of the flotation cells), 50%, and 25%, respectively (Rothelius, 1957)^. Due to the fall of the tungsten price in the early 1960s, the mine was closed in 1963 and flooded. Towards the end of the 1960s, tungsten price picked up renewing interest in tungsten production. Swedish state-owned mining company AB Statsgruvor acquired the mine in 1969 and constructed a new concentrator for gravity separation, which was then converted to selective flotation in 1977 (Tasman Metals Ltd., 2016).

Approximately more than 5 million tonnes of ore material with an average of 0.35% WO3 with additional copper and fluorite was extracted during the life of mine.

(12)

8 The dressing mill of the mine was constructed in 1936 and had a capacity of 23 tons per hour of crude ore, operating at 18 hours per day. In a year this translated to 105,000 tons crude ore processed, with a concentrate output shown in Table 1. Production value in the dressing mills of Yxsjöberg was 80% for scheelite, 15% for fluorspar, and approximately 5% for chalcopyrite. The main consumers of these products were manufacturers of hard alloys for scheelite, manufacturers of welding electrodes and the iron and steel market for fluorspar, and lastly the Rönnskär Smelting works for the copper concentrate.

Fluorspar and copper concentrates were trucked to Hörken to be later on transported via railway while scheelite concentrate was transported directly to the consumers via truck to avoid re-loading of the material.

Table 1. Yearly output and corresponding production value of Yxsjöberg mine (Rothelius, 1957)

Concentrate Yearly output, tons Production value, %

Scheelite 350 80

Fluorspar 3000 15

Copper 200 5

All the beneficiation processes including crushing and grinding for the ore material were all done on- site. The tailings produced from these processes were then pumped into two tailings repositories identified as Smaltjärnen (1897-1963) and Morkulltjarnen (1969-1989) (Höglund, Jones, & Lindgren, 2004). The two large tailings dams were left mostly untouched since the final closure in 1989, and are estimated to contain approximately a total of 4.6 million metric tonnes of material (Tasman Metals Ltd., 2016). The tailings in the Smaltjärnen repository has been stored there for more than 100 years, and have been exposed to the atmosphere for more than 30 years due to a lack of a complete cover.

According to the findings of Hällström (2018), the uppermost layer tailingswere observed to have intensive pyrrhotite oxidation and carbonate dissolution. This exposure resulted in an acidic environment with pH <4 and with release of elements and formation of secondary minerals such as gypsum and hydrus ferric oxides (HFOs) (Hällström, 2018). The low pH conditions caused by the tailings indicate environmental concern and a need for further investigation on how to handle the repository.

Based on the paper of Höglund et al. (2004), the Smaltjärnen repository was classified as Risk 1 according to the MIFO, or the Swedish Environmental Protection Agency’s classification system.

MIFO in Swedish translates to Methods for Inventories of Contaminated Sites. The system takes into account hazard assessment, the contamination level, migration potential, and the sensitivity protection value of the surrounding environment. The classification of Smaltjärnen as Risk 1 indicates that it is of the highest environmental risk.

1.2 Tungsten

Tungsten was the main product that came out of the Yxsjöberg operations. Tungsten, or otherwise known as Wolfram from which it hails its atomic symbol W from, is a silvery-white metal that shines blue-white under UV light, although it can become more yellow depending on the percentage of Mo substitution. It is known to have the highest melting point of 3410°C compared to other pure metals.

The name Tungsten hails from the Swedish words “tung” which means heavy, and “sten” which means stone, due to its dense nature having a density of approximately 19.3 g/cm³, which is approximately close to that of gold. This similarity in density allows tungsten to be a substitute for gold or platinum (RankRed, 2018).

The vital properties of tungsten include very high melting point and density, high strength and wear resistance, as well as high thermal and electrical conductivity. These properties make it attractive for industrial applications such as in steel alloying. A patent for the first tungsten-containing steels were accepted in 1858 which led to the development of self-hardening steels in 1868 (International Tungsten

(13)

9 Industry Association, 2011). Tungsten is also an important component in electrodes, as well as in the filaments of incandescent bulbs which replaced the carbon filament lamps since its patent in 1904.

These kinds of lightbulb are being currently phased out in different countries, but tungsten still plays a major role in the lighting industry through the use of calcium and magnesium tungstates in fluorescent lighting (Emsley, 2012). Tungsten also has the lowest vapor pressure of all metals, very high thermal creep resistance, and high thermal and electrical conductivity. Due to this, tungsten is considered the most important metal for thermo-emission applications (Yang X. , 2018). Tungsten carbide, an example of hardmetal or cemented carbide, is produced by combining carbon and tungsten and is the hardest artificial substance. Its major use is as a cutting tool due to its diamond-like hardness and strength.

Approximately 54-72% of globally-produced tungsten is used in hardmetals making it the most important application of tungsten. Aside from this, tungsten is an important material in the X-ray industry, as well as in microchips and liquid crystal displays (Lenntech, 2019).

Tungsten resources are geographically widespread, being found in the United States, Vietnam, Portugal, and Russia, to name a few. China dominates the world supply market, ranking first in terms of tungsten resource and reserves and owning at least 80% of the global production. In 2016 and 2017, China imposed production quotas in an effort to reduce illegal mining and improve environmental conditions.

This resulted to limited availability of tungsten concentrates in their local market. This, coupled with the closure of Canada’s sole tungsten mine in 2015 led to a major decrease in tungsten world production.

China is also the leading consumer of tungsten, wherein they increased their consumption during the first 6 to 9 months in 2017, leading to an increase of the global tungsten consumption. Due to these events, the price of tungsten concentrates as well as the downstream materials continued to increase, following the upward trend that started in late 2015 to early 2016 (U.S. Geological Survey, 2018).

Tungsten is listed as one of the 20 critical materials in Europe due to its economic importance and high supply risk.

Processing of tungsten operates with the aim of producing a saleable concentrate of tungsten trioxide, WO3 of at least 65% to be acceptable for the international market. (Rao, 1996) Known beneficiation processes of tungsten include gravity, magnetic, and electrostatic separation (to separate cassiterite from non-conducting scheelite-wolframite ores), as well as froth flotation, with the choice of beneficiation process governed by the characteristics of the ore. Despite the numerous occurences of tungsten minerals, only scheelite (CaWO4) or wolframite ((Fe,Mn)WO4) are considered to be of economic importance. Scheelite is an important ore of tungsten. It is more common than wolframite, being found in approximately 65% of all tungsten deposits (Pitfield, Brown, Gunn, & Rayner, 2011). Scheelite has the mineralogical formula of CaWO4, composed geochemically of 80.52% WO3 and 19.48% CaO. Its bulk density is 6.10 with a hardness of 4.5-5 based on the Mohs scale. Its appearance is usually white, yellowish, brown, or grey and is neither magnetic or conductive (Elsner, 2010) The mineral scheelite usually occurs in metamorphic skarns, high temperature veins, but less commonly in granite pegmatites and medium-temperature hydrothermal veins or possibly in alluvial settings (Handbook of Mineralogy).

Scheelite is the most abundant tungsten mineral present in roughly two-thirds of known tungsten deposits (British Geological Survey, 2011).

Both major ores of tungsten, namely scheelite and wolframite, are characterized with high density and brittleness. Due to the high density of tungsten, gravity separation for coarser particles are a logical option for beneficiation. Magnetic separation has also been implemented for wolframite and pyrrhotite ores utilizing their paramagnetic properties. The combination of magnetic and gravity separation is often used for wolframite ores, while the combination of gravity separation and froth flotation is implemented for scheelite ores, as only scheelite is considered readily amenable for flotation. The liberation size of tungsten minerals is found in a wide range, from several mm to 10-20 µm. The mineralization of tungsten ores is usually complex, containing different minerals and metals as well as ore textures, causing flowsheets for processing this material to be complex (Yang X. , 2018).

Because of the increasing exploitation of primary tungsten resources, grades are expected to become lower and the ore material to be finer in size and have more complex mineralogy. This will ultimately result to more difficult processing of tungsten. (Yang X. , 2018)

(14)

10 The material of focus in this masters thesis came from the Smaltjärnen repository which was operational between the years 1897 to 1963. The time when material was being deposited in this repository, operations were intermittent as previously mentioned, and recoveries were at maximum 75% for scheelite for the last 7 years, 50% for fluorite, and 25% for chalcopyrite. This further indicates high mineral and metal content in this repository pointing towards good potential for the repository to be a secondary resource.

2 Research topic and purpose

The Flotation of the Yxsjöberg historical tungsten ore is a masters thesis project that is part of the REMinE (Improve Resource Efficiency and Minimize Environmental Footprint) project. The REMinE project is a joint effort of Luleå University of Technology (LTU) in Sweden, University of Porto in Portugal, and the Research and Development National Institute for Metals and Radioactive Resoruces (INCDMRR) in Romania in order to evaluate the amenability of historical mining waste for reprocessing from the technical, economical, and environmental perspectives. The project consists of historical mines from three European countries: Portugal, Romania, and Sweden. The historical mine that is studied in Sweden is Yxsjöberg. Based on Alangkas et al. (2016), the REMinE project is comprised of 5 work packages which are enumerated below:

Table 2. Description of the different work packages included in the REMinE Project

WP Description

1 Project management

2 Detailed characterization and risk assessment of the mine wastes selected 3 Identification of new processing methods for mine waste

4 Characterization and risk assessment of the remaining residuals

5 Outlining business opportunities and environmental impact in a conceptual model for sustainable mining

Previous work on the area includes geochemical characterization by Hällström (2018) and physical separation work through gravity and magnetic methods by Khavari (2018). However, scheelite is most commonly recovered through combined gravity and flotation methods. The overall aim of this masters thesis is to determine the feasibility of applying flotation on the Yxsjöberg historical tungsten tailings as a reprocessing method, which is under WP3 of the REMinE project. Focus on scheelite and consequently the element W was given due to the special interest of re-extracting this metal based on economic and environmental perspectives. The behavior of the different drill cores and layers of the selected area will also be assessed to determine potential problems in the flotation processes.

Specific objectives targeted in this thesis consist of the following:

• Further characterization of feed material

• Determine feasibility of using flotation as a reprocessing method for the Yxsjöberg historical tungsten ore tailings

• What collector can be used? – what is the maximum grade and recovery

• Will the different layers react similarly to the chosen flotation conditions?

3 Review of Related Literature

3.1 Geological Background

The Yxsjöberg ore deposit is classified as a skarn-type which commonly hosts elevated concentrations of certain metals such as W, Cu, Zn, Sn, Be, and Bi, as well as fluorite, carbonates, and sulfides. Skarn type deposits have a typical grade of 0.3-1.4 % WO3 with a size of <10⁴ to 5x10⁷ tons (Yang & Niemistö, 2017). Tungsten primarily occurs in scheelite, which is the mineral of economic interest in the former

(15)

11 operations of the Yxsjöberg mine, aside from chalcopyrite and fluorite. As previously mentioned, Be and Bi minerals exist in the deposit but was not extracted previously.

The main part of the ore comes from pyroxene skarn where hedenbergite is the dominant mineral. The skarn exhibits high fluorspar occurrence and pegmatite material. Hence, pegmatite minerals from the feldspar group such as albite and oligoclase, quartz, as well magnetite, pyrrhotite, bismuthinite, helvite, and sphalerite were detected because of the country rock. Aside from these, the main gangue minerals in the deposit were from the carbonate, pyroxene, amphibole, garnets, mica, and chlorite groups, as well as apatite and titanite.

Table 3. Average mineral grades of Yxsjöberg ore; source: (Gräsberg & Mattson, 1979)

Mineral Scheelite Chalcopyrite Fluorite Calcite Quartz Biotite Apatite Feldspar Sulfides Skarn Minerals

Grade

(%) ^0.5 ^0.7 ^7.5 ⁵ ^5.2 ¹⁰ ^0.1 ¹⁵ ^6.1 ⁵⁰

Mining methods changed during the time of operation, with underground mining occurring for three separate periods, specifically 1918-1920, 1935-1963, and 1969-1989. Throughout the operations, the minerals being mined varied. The particular extraction periods of these cannot be pinpointed due to limited sources and differences in the information available. Table 4 shows the reasoning in Hällstrom (2018) why the mineral composition of the tailings repository were observed to be varied both in horizontal and vertical directions.

Table 4. Indicated reasoning behind variation in mineral composition of the repository as enumerated in Hällstrom (2018)

1 Variation of the mineral assemblages within the three ore bodies mined

2 Change in mineral processing methods implemented during the mining operations 3 Change in commodities (W, Cu, CaF2) being mined

4 Movement of pipeline that doposits the tailings into the repository (change of point of deposition)

5 Frequent smoothening (hence disturbing) of tailings material

6 Removal of tailings from the repository to be used as backfill for the mine (Hoglund et al., 2004)

3.2 Previous operations at Yxsjoberg

Hedenbergite (Ca(Fe,Mn)Si2O6) was the dominating skarn mineral in the Yxsjöberg mine. The ore contained approximately 0.3% WO3 as scheelite, 5-6% CaF2, and 0.16% Cu. The skarn was found to have a rather high content of sulphides of which were mostly pyrrhotite, with some chalcopyrite and pyrite. The grain size of scheelite found homogenously distributed in the skarn is approximately 0.2- 4.0 mm. According to previous operations, the liberation size of the grains of the ore at 90% by weight is 0.4 mm. (Rothelius, 1957)

Based on accounts from Rothelius (1957), the Yxsjöberg mine had two main divisions for its mineral processing process: the wet gravity and magnetic separation division, and the flotation division. Ore material first went through the wet gravity and magnetic separation division. After, the tailings then proceeded to the flotation division. Prior to scheelite flotation, the material first went through a copper flotation circuit. The bottom product of the rougher in the copper flotation then served as the feed for fluorspar flotation which was the eventual source of the WO3 concentrate.

Beneficiation started with the material passing through a Blake-type jaw crusher with a setting of 90 mm. The primarily crushed material was brought to a vibrating screen with 65 mm-diameter opening via conveyor belt. The coarse fraction was ground again in a jaw crusher with a 65 mm setting, while the -65 mm passing product continued to a vibrating screen with 10x10 mm square openings. Symons cone crusher and rolls were in closed circuit with screens for fine crushing. This was implemented upon

(16)

12 choosing a crushing and grinding method that would minimize slime production since scheelite and fluorspar tend to enrich in the finer grain sizes due to their friability. After being crushed to -3 mm, the material proceeded to go through jigs and shaking tables. The coarse liberated scheelite grains were then recovered before the bulk material was brought into the ball mill.

The crushed ore then went through jigs and a series of slime and shaking tables for pre-concentration.

The material was then ground to -0.5 mm of 80% by weight and sent to flotation. The concentrate from the flotation stage was further treated by Krupp and Humboldt-type slime tables while the finest fraction went to tilting tables. This process resulted in a low total recovery of maximum 60% by weight.

Modifications in the wet gravity method has increased the recovery up to 70% by weight, by replacing slime tables with diagonal shaking tables.

The flotation circuit comprised first of sulphide flotation, where tailings from the wet gravity division was first dewatered and then conditioned with 30 g/t xanthate and 15 g/t pinol. Before entering the rougher, 60 g/t xanthate and 15 g/t pinol was added. The rougher concentrate was conditioned with 200 g/t of sodium silicate and proceeded to the cleaning circuit where the end float product is the Cu- concentrate. The flotation middlings from the cleaning stage were re-circulated in the sulphide flotation circuit and conditioned with 50 g/t of sodium silicate.

From the rougher of the sulphide flotation circuit, the bottom product served as feed for the fluorspar flotation. This circuit started with pumping the material to a large conditioner with a volume of 5.6 cubic meters. The pulp was then conditioned with 90-100 g/t of quebracho and 200 g/t sodium silicate and the pulp heated to +24°C. The pulp was then brought to another conditioner where 240 g/t of tall oil was added and sent to a rougher flotation circuit consisting of 8 cells to float fluorspar. Although studies for recovery of fluorspar has been in effect since 1941, full-scale production for fluorspar only commenced in 1956, requiring fine-grinding of the material and a pulp that is heated to +24°C. During the time of writing of Rothelius (1957), intent of doing test to float the beryllium mineral helvin has existed, however no further studies regarding this has been found.

The rougher concentrate was then forwarded to a rod mill and then another conditioner of 2.5 cubic meter volume. This was where the pulp was steamed to 26°C before being cleaned in three 2-cell flotation machines in series. The cleaner tailings were then pumped into a hydrocyclone where the underflow proceeded into a ball mill. The cleaned concentrate was sent to a 1m³ condition where the pulp was steamed to 28°C. To this, 20 g quebracho and 100 g/t of sodium silicate were added, and continued into four 1-cell cleaner machines in series. In the second machine, 10-15 g/t quebracho was added. In the third, 10 g/t of the same depressant, and 5-10 g/t in the fourth machine. Starting from the first machine, tailings were sent to two slime tables where tailings were again pumped into a hydrocyclone. Middlings of the last three cells return to the previous cell. The cleaned concentrate was fed onto a slime table resulting into a WO3 concentrate and acid grade CaF2 concentrate. Metallurgical grade CaF2 and a WO3 concentrate was recovered in the last cleaning step of the rougher tailings circuit.

The whole process produced WO3 concentrates through 5 outlets, that were ultimately collected through the slime tables.

The rougher scheelite concentrate recovered from the shaking tables were dried in a rotating tube. The pyrrhotite content was recovered through weak field magnetic separators. The non-magnetic product was heated to 600°C and subjected to high intensity magnetic separation to remove chalcopyrite, pyrite, garnet, and other skarn minerals from the scheelite concentrate. The tailings from the process was pumped through a 300-meter pipe line with a diameter of 400 mm. The material left then flowed into extensive bogs and swamps where most of the tailings sedimented.

Recovery of scheelite was low in 1942, at only 50% by weight. Through modifications in the mineral processing methods applied, recovery was increased to approximately 70% by weight in 1956. The produced scheelite concentrate has a grade assaying between 66-74% WO3 with very low amounts of contaminants making it suitable for production of hard alloys.

(17)

13

Figure 2. Flowsheet of the Flotation division of the Yxsjöberg dressing mill (Rothelius, 1957)

(18)

14 Fluorspar concentrate recovery was measured to be less than 50% by weight with quality that was found to be satisfactory. Chalcopyrite recovery was low, at only 25% by weight and with grades that were not looked into. No information regarding pyrrhotite and pyrite was available since it was considered to have no economic importance. (Rothelius, 1957)

Material that directly ended up in the tailings based on these processes mainly come from the overflow of the Dorr thickener from the gravity separation division, and slime table 3 of the flotation division.

Although the entire process was complex and forced the material to re-circulate, the reported low recoveries are an indication of the abundance of these minerals still reporting into the repository.

It must be noted however that the flotation processes for the Yxsjöberg deposit was implemented at the tail end of the operational life of the Småltjarnen repository. The recovery of scheelite was only increased up to 70% in 1956, only 7 years before the repository was declared at the end of its life.

Hence, the majority of tailings material considered in the masters thesis is processed through earlier operations, and looked into from this perspective.

3.3 Novel Process

AB Statsgruvor, a member of the Swedish state-owned LKAB group, took over Yxsjöberg in 1969 when it was out of production. The operations were restarted in 1972 with conventional processing, specifically with gravity concentration and magnetic separation. The method involved gravity separation with low and high intensity magnetic separation then magnetizing roasting. Both copper and fluorspar concentrates were recovered through flotation although the production was considered unsuccessful due to lack of material and inefficient mineral processing. Hence, a new process has been developed to treat the material however it must be noted that it is no longer 100% from Yxsjöberg but composed of material from other deposits as well as some old tailings to provide a reasonable lifetime for the operations.

The main skarn minerals in Yxsjöberg are hedenbergite, hornblende, biotite and some garnet. The hedenbergite skarn occurs with fairly coarse-grained scheelite, while the hornblende skarn contains evenly distributed small-grained scheelite in the matrix.

Focus on scheelite flotation was brought about to secure “long-term profitability” after numerous attempts in improving the processes. Due to the presence of minerals with similar flotation behavior such as fluorite and calcite, the concentrates were either of low recovery or low grade. Different tests have been done for scheelite flotation, both either through bulk flotation of calcium minerals, or selective flotation of scheelite. Bulk flotation showed difficulties in pilot plant tests brought about by varying calcite content. This prevented the system from achieving steady-state separation conditions resulting to difficulty in having a concentrate with constant quality. Selective flotation, on the other hand, although requires high consumption of chemicals due to the use of a fatty acid collector, were proven reliable between laboratory scale tests and pilot plant tests, and was easy to upsize into full scale operations. (Gräsberg & Mattson, 1979)

The process implemented for treating the Yxsjöberg material between January to September 1978 started with open circuit rod milling with D80 of 160 µm which then proceeds to a closed ball mill grinding circuit with a 250 mm hydrocyclone. The hydrocyclone overflow is kept at 50% weight-solids with 50% -74 µm. However, when dealing with old tailings, the ball mill grinding stage is skipped to prevent overgrinding. Two flotation circuits were implemented, one each for chalcopyrite and scheelite.

As earlier mentioned, the material treated in this process was not exclusively from Yxsjöberg. However, only Yxsjöberg ore was passed through the copper flotation circuit, while other material by-passed this stage and is directly brought to scheelite flotation. At this point in time, fluorspar was not floated.

(19)

15

Table 5. Grade and recovery of material under novel process implemented by AB Statsgruvor (Gräsberg & Mattson, 1979) Source Stream Mass (t) %Cu %WO3 %CaF2 %CaCO3 Recovery (%)

Yxsjöberg ore

Feed 92,649.0 0.21 0.384 6.2 5.9

Cu concentrate 678.0 22.60 0.010 79.0 WO3 concentrate 405.3 0.01 69.900 1.8 10.7 79.4

Tailings 91,566.0 0.05 0.080

Old Tailings

Feed 5,735.0 0.05 0.123 3.6 4.5

WO3 concentrate 6.0 66.600 7.3 7.6 56.4

Tailings 5,729.0 0.05 0.054

Fatty acids are commonly used as well as frothers, hence for scheelite flotation, fatty acid serves as the collector and the frother. (Wills & Finch, 2016) This is performed with the aid of intense flocculation, requiring special attention during the cleaner stage to ensure release of contaminants from the scheelite grains. Scheelite floats considerably faster than most sulphide minerals. It is observed that Si, Fe, As, Sn, Bi, and S in the concentrate is also lowered through scheelite flotation than in magnetic and gravity methods. However, Molybdenite has not shown difference in grade and recovery with the flotation process. According to Table 5, implementation of this method allowed the concentration ratio of WO3

for old tailings to reach 500:1 from 0.123% WO3 to 66.6% WO3 at 56.4% recovery. (Gräsberg &

Mattson, 1979)

3.4 Flotation

Froth flotation is considered to be one of the most important mineral separation techniques. More than 80% of the total metallic minerals in the world is reported to be recovered through flotation processes.

By definition, it is a physicochemical separation method utilizing the differences in interaction of solid particles to an aqueous solution and the gaseous phase. Therefore, it is a 3-phase process that involves the mineral particles (solid) and the water (liquid) that make up the pulp, as well as the air bubbles (gas) that lift the hydrophobic materials to the froth, leaving the hydrophilic materials in the pulp. This operation is governed by factors such as bubble-particle and bubble-bubble interaction. The main objective of flotation is bringing the valuable material to the surface by the froth while the gangue material is left in the pulp. In the case wherein gangue is explicitly floated to the surface, the method is referred to as reverse flotation since the valuable minerals are what is left in the pulp. (Wills & Finch, 2016)

Generally, it is claimed that a particle must be at least 20% hydrophobic to be floated. This can be either the total surface area of the mineral particle being composed of 20% hydrophobic minerals, or have at least 20% of the mineral surface be amenable for collector adsorption. Different indications for what size range is most favorable for flotation varies between references. In some, the size range of 20-300 µm is considered to be typical for flotation. Below this, less collision between particles and bubbles can be expected, leading to a less efficient system. Based on other sources, fairly good recovery of 10-100 µm particles can be expected in a mechanical flotation cell. At 1-10 µm, lower collision probability can be expected, while 100-1000 µm will have a higher probability of detaching. (Otsuki, 2018) Others claim that it can be between 5-500 µm, while below 50 µm is considered too fine and would require higher reagent consumption due to higher surface area exposed as well as still lower collision probability. Higher than 500 µm is said to be hard to recover due again to high detachment (Fornasiero

& Filippov, 2016).

Most minerals are naturally hydrophilic and are rendered hydrophobic through the use of collector reagents. An essential prerequisite in order to separate minerals by flotation is selective collector adsorption onto the minerals of interest. The hydrophobicity of mineral surfaces is required to form stable bubble-particle aggregates which can be recovered by flotation. Collectors adsorbed on to the mineral surface can enhance the hydrophobicity and consequently their flotation rate and recovery.

Aside from collectors, there are also other flotation reagents such as depressants, frothers, and

(20)

16 regulators. These reagents contribute in producing selectivity between minerals by changing the characteristics of the particle surfaces. (Fornasiero, 2018)

The interactions between the particles and bubbles, or particle-bubble interaction, relies on the mineral surface chemistry. Hydrophobic surfaces are commonly built with unbroken bonds, while hydrophilic surfaces contain broken bonds making way for hydrogen bonding with water molecules. Mineral surfaces have a surface charge that may be due to different reasons such as broken bonds, adsorption of ions, or ion dissolution. As mentioned earlier, these can be brought about by breakage of bonds through comminution resulting to electron imbalances, or change in pH that alters the mineral surface charge.

However, due to mineral dissolution or masking brought about by adsorption of species in solution onto mineral surface, the surface chemistry may change once in contact with water during grinding and conditioning before flotation. Masking is a concern as it can inhibit the preferred characteristics of the mineral surface to be recognized in the flotation system (Ralston & Fornasiero, 1999).

Selective flotation can be achieved through pH control. The electrokinetics of the mineral suspension affects the dispersion or aggregation of fine mineral particles in aqueous suspensions. The zeta potential affects the interaction between similar and dissimilar particles, as well as the particles and flotation reagents in a flotation system. By changing the pH of the system, the surface charge (hence, the Zeta Potential) of the mineral is shifted (Wills & Finch, 2016). Reagents are used to adsorb selectively on the target minerals to either make them hydrophobic with collectors, or hydrophilic with depressants.

The adsorption of mineral particles on to the gas molecules occur due to collision of the two, governed by the kinetic gas theory. However, not all mineral particles will stay onto the gas molecule, which is then predicted by the “sticking probability”. Either chemisorption or physisorption can occur between the interface and are the main binding mechanisms for the adsorption on the solid-gas interface.

Chemisorption involves direct chemical bond between the adsorbate (gas molecules) and the substrate (mineral surface) with an electron exchange, resulting to a strong bond (Wills & Finch, 2016).

Physisorption, on the other hand, is characterized by van der Waals attraction. No electron exchange is observed creating a weaker bond than chemisorption. In physisorption, binding between adsorbate to substrate is in the same magnitude between adsorbate to adsorbate, allowing multilayer formation of adsorbates on the substrate whereas only a monolayer is observed for chemisorption. The rate of solid- gas interactions where in the bubbles collect the mineral particles not only depend on the hydrophobicity of the mineral surface, but on the conditions of the flotation cell as well. (Fornasiero, 2018)

Sulphide mineral flotation differs from non-sulphide mineral flotation generally through what collectors can be used. Sulphide minerals are usually dense, and have a well-crystallized and defined composition that can be easily differentiated from gangue. However, non-sulphides are usually porous, have variable compositions, and usually require higher collector dosage since it is more similar to gangue resulting to less selective collector adsorption. These differences contribute to characteristics that make collectors commonly selected for sulphide mineral flotation to not be amenable for non-sulphide mineral flotation.

(Otsuki, 2018)

Fatty acids are common anionic collectors for non-sulphides, specifically used for P, ZnO, Ni, Sn, and W (Finch & Riggs, 1986). The adsorption mechanism of anionic collectors is through physisorption due to electrostatic attraction. The electrostatic interactions are due to weak attraction with minerals. In recent times, the direction of studies for scheelite flotation have focused on using mixed collectors to create selectivity between the main mineral and other Ca-bearing minerals. Mixed collector systems pertain to those with different properties (e.g. ionic and non-ionic) to drive better collector adsorption, decrease in collector dosage, and ultimately improvement of flotation performance.

The properties that make up an ideal collector include high molecular stability (i.e. long shelf life, resistance to decomposition in oxidizing aqueous solutions), good aqueous stability (for ease of dissolution into flotation pulps and to minimize non-selective adsorption of collector onto mineral surfaces), and ability to form specific metal ion chelates or complexes (Fornasiero & Filippov, 2016).

Generally, the trade-off between these traits are strong collectors usually portray low selectivity due to

(21)

17 longer hydrocarbon chains. Hence, weak collectors are more selective but have shorter chains. Oxides need collectors with long hydrocarbon chains, which are more hydrophobic, generally with n>10 where n is the number of hydrocarbons. The general trend is that the weaker the collector is, the higher the dosage needed to float the same amount of minerals.

When the collector adheres to the mineral of interest, the stability of this complex then plays an important role in the success of collecting it. Metal-collector stability is affected by the nature of donor atoms, central metal atom, pKa of the collector, susbtituents, ring size where a bigger ring size leads to better stability, and the number of rings. A lower solubility product of a collector-metal indicates greater stability and therefore more favourable complex formation. In selecting collectors, a lower solubility product is then selected with regards to the minerals of interest and the other minerals that is not wanted in the concentrate.

Aside from collectors, depressants are a reagent also used in the flotation system. Contrary to the collector, the role of a depressant is to make the mineral surface hydrophilic therefore allowing the affected minerals to sink and remain in the pulp. In scheelite flotation, one of the most common depressants is sodium silicate. However, it was observed that sodium silicate can also depress scheelite when applied in high dosages. Aside from collectors and depressants, a frother is usually added into a flotation system. The role of a frother is to reduce merging of bubbles, or bubble coalescence, and thus creating a more stable froth at the top of the flotation cell where the particles are contained before being recovered. (Farrokhpay, 2011)

The mineral processing industry is continuously developing and currently demands in mineral flotation include lowering reagent costs, improving collector efficiency and selectivity under a wide range of conditions in the treatment of complex and low-grade ores, and improving recovery. (Mouat, 1996) Major difficulties in flotation include collectors that are rarely specific or selective, having a wide variety of minerals of differing surface properties in any ore, the same mineral with different crystal structures and chemical composition resulting in different surface properties, extensive atomic substitution in crystal lattice, intimate interlocking of various minerals and/or rimming of one mineral on another. Since the mineral surface chemistry is very important in creating selectivity, riming minerals can inhibit selective adsorption of collectors onto the mineral of interest. Since minerals with rims of hydrous ferric oxides (HFO) were observed to be present in the samples, this can potentially be a problem in the flotation processes.

3.4.1 Scheelite flotation schemes

The standard concentrate grade of wolframite is 65% WO3 and 60% WO3 for scheelite. The amount of penalty elements that can be tolerated may differ between contracts. Generally, the maximum allowable content of the impurities before penalty are shown in Table 6.

Table 6. Maximum allowable content of impurities

Element Max. amount (%)

Sn 1.60

As 0.20

Cu 0.10

P 0.05

Sb 0.05

Bi 0.40

S 1.00

Pb <0.001

(22)

18 Scheelite concentrate that can be used as direct addition in steel melts requires a grade of at least 70%

WO3 with low amounts of impurities. Concentrates with lower grades can be subjected to further chemical treatment to produce a marketable product. This was one of the aims of AB Stratsguvor in their novel process if the target grade of 65% WO3 was not reached, which is higher than the common 60% WO3 target for saleable scheelite.

Scheelite is a polar salt type mineral, making it hydrophilic as it actively reacts with water molecules due to the relatively strong ionic surface bonding with high free energy values at its surface (Wills &

Finch, Wills' Mineral Processing Technology - An introduction to the practical aspects of ore treatment and mineral recovery, 2016). This surface charge is determined by the Ca²⁺ and WO42- present in the aqueous solution. This becomes critical when it is the electrostatic interactions that control the collector adsorption onto the mineral surface. Scheelite is negatively charged throughout the pH range having the most commonly exposed surfaces being {1 1 2}, {1 0 1}, and {0 0 1}. (Gao, Hu, Sun, & Drelich, 2016)

The minerals commonly associated with scheelite in ore is calcite, fluorite, and apatite, to name a few.

In a normal flotation set-up, these minerals can easily contaminate the scheelite concentrate due to the similarities in their floatability, mainly brought about by the calcium cation that is present in these minerals. (Yongxin & Changgen, 1983)

According to Kupka and Rudolph (2017), the optimum pH for scheelite flotation is 9 regardless of the reagent combination, although other authors indicate the optimum pH range to pH 8-11, or further restricting it to pH 9-10. Figure 3 shows that there is a dip in calcite recovery at pH9 while scheelite shows a fairly constant recovery between pH 8-11 which can imply better separation at pH 9. This behavior shows a consistent behavior despite different collector dosages. pH modifiers commonly used in scheelite flotation operations include sodium hydroxide (NaOH), sodium carbonate (Na2CO3), and calcium oxide (CaOH). Comparing the three, sodium carbonate is said to improve the tungsten grade and recovery as well as the selectivity of scheelite against calcite, which may be due to the limited Ca²⁺

or Mg²⁺ ions introduced to the pulp (Rutledge & Anderson, 2015).

Figure 3. Flotation of scheelite and calcite at constant sodium oleate concentration as shown in (Patil & Nayak, 1985)

Selectivity between scheelite and a siliceous gangue is relatively easy. Separation of scheelite from a calcareous gangue is much more complicated due to the similarity of the surface properties of scheelite to other calcium-bearing minerals Table 7. Both scheelite and calcareous gangue contain the Ca²⁺ cation and their anions have similar sizes which invites the same reaction from collectors (Filippova, Filippov, Duverger, & Severov, 2014). Some of the factors that contribute to complication in separation between these minerals include the extent of collector adsorption on the mineral surface, dissolution of minerals and interference by released species, and the interaction of the adsorbed ions with the mineral surface.

Flotation of Yxsjöberg historical tungsten ore tailings

Flotation of Yxsjöberg historical tungsten ore tailings

Princess Rochelle O. Gan

Flotation of Yxsjöberg historical tungsten ore tailings

Princess Rochelle O. Gan

Acknowledgements

Summary

Table of Contents

List of Figures

List of Tables

1 Introduction

1.1 Yxsjöberg tungsten deposit

1.2 Tungsten

2 Research topic and purpose

3 Review of Related Literature

3.1 Geological Background

3.2 Previous operations at Yxsjoberg

3.3 Novel Process

3.4 Flotation