Okontrollerad skumning i molybdenflotation: Studier och lösningar

(1)

MASTER'S THESIS

Excessive foaming in moybdenum

flotation

- studies and remedies

Jennifer Andersson

Master of Science in Engineering Technology

Chemical Engineering Design

Luleå University of Technology

(2)

Abstract

The aim of this master thesis project was to gain knowledge of potential problems in the planned molybdenum separation plant in the concentrator at Boliden Mineral’s mine Aitik. Main focus was on the uncontrollable foaming earlier observed, that caused losses of copper to the molybdenum concentrate and huge problems in the following water treatment process.

The project had two major focuses; finding the cause of the uncontrollable foaming and removing it without affecting the molybdenum recovery negatively. The pulp in the planned molybdenum flotation process could be assumed to contain mostly copper concentrate, process water, thio anions and residuals of KAX (collector) and Nasfroth 350 (frother). It was investigated if none, some, or all of above mentioned species contributed to the foaming, possibly together with temperature and time dependencies.

To evaluate the problems with foaming, the project was divided into three different parts namely; evaluation of different analytical techniques, designed conditioning experiments and finally flotation experiments.

The analytical techniques chosen were UV-spectroscopy, analysis of total organic carbon and quantitative measurement of foaming by a simple gas-sparged column. For the designed

conditioning experiments, the software MODDE was used to create an experimental design where the three factors temperature, water quality and amount of depressant (NaSH) was set at low and high levels respectively. During the flotation experiments the influence of activated carbon on foaming and molybdenum recovery was examined.

It was concluded that over-dosage of NaSH do not depress foaming, on the contrary, it makes it worse. Also, there was nothing that suggested that KAX caused foaming. It was however difficult to assess the influence of frother on the foaming, with the available analytical techniques, since it was within the same UV-absorbance range as the much larger sulphide ion peak.

Activated carbon did decrease foaming significantly, decreased the water content in flotation products and increased selectivity of molybdenum. It was the only parameter tested that gave a significant improvement to the foaming problem. It was also concluded important to maintain a low flow rate of the flotation gas, to avoid excessive foaming.

(3)

Sammanfattning

Syftet med detta examensarbete var att erhålla kunskap gällande potentiella problem i det

planerade molybdenflotationsverket vid Bolidens gruva Aitik. Störst fokus låg på den okontrollerade skumning som observerats under tidigare försök i pilotskala, och som orsakat problem med

kopparförluster till molybdenkoncentratet samt stora hanteringsproblem i vattenreningen. Examensarbetet hade två huvudinriktningar, dels finna orsaken till den okontrollerade skumningen men också att avlägsna minimisera skumningsproblemen utan att påverka molybdenflotationen negativt. Pulpen i det planerade molybdenflotationsprocessen kunde antas bestå av främst kopparkoncentrat, processvatten, olika tiosalter samt rester av KAX (samlare) och Nasfroth 350 (skumbildare). Det undersöktes om ingen, några, eller alla av dessa bidrog till skumningen, möjligen tillsammans med temperatur och tidsaspekter.

För att undersöka skumningsproblemen utfördes den praktiska delen av examensarbetet i tre delar: utvärdering av olika analystekniker, designade experiment där konditioneringssteget simulerades, samt flotationsexperiment.

Valda analystekniker blev UV-spektrometri, analys av totalt organisk kol samt kvantitativa mätningar av skumning med ett enkelt skumpelartest. Dataprogrammet MODDE användes för de designade konditioneringsexperimenten för att skapa en experimentell design. Tre faktorer valdes: temperatur, vattenkvalité och mängd tryckare (NaSH) och de varierades i två nivåer (låg och hög). Under

flotationsexperimenten undersöktes aktivt kols inverkan på skumning och molybdenutbyte. Ingenting i resultaten indikerade att KAX orsakade skumning, det konstateras därmed orsaken till okontrollerad skumning inte kan tillskrivas samlaren. Det konstaterades även att överdosering av NaSH inte dämpade skumningen, ökande NaSH halt resulterade istället till ökande skumning. Att utvärdera skumbildarens effekt på skumning var svårt med tillgänglig analysteknik, eftersom dess spektrum ligger i samma område som det mycket större sulfidspektrumet.

Användning av aktivt kol var den enda parameter som visade sig dämpa skumningen signifikant. Det sänkte även vatteninnehållet i flotationskoncentraten och ökade selektiviteten. Att hålla en låg flödeshastighet på flotationsgasen var också viktigt för att undvika överdriven skumning.

(4)

Acknowledgments

As I complete the final pages of this master thesis project, which also means completing my engineering studies, I cannot help but remembering my sixteen year old self. The girl that truly believed that engineering studies was too difficult, and not suitable for her. Therefore, my profound gratitude must firstly go to my high school mathematics and physics teacher Björn Fagerström, who gave me the confidence I lacked, by constantly pushing, challenging and believing in me. I cannot thank you enough for being (what I thought many times back then) so hard on me, and for encouraging me to choose engineering studies. Also, I owe many thanks to Vivan Östberg, my brilliant high school chemistry and mathematics teacher who kept my interest in chemistry alive. I would also like to thank my supervisor at Boliden Mineral, Jan-Eric Sundqvist, for always having an infinite amount of ideas and time for my questions but at the same time always allowed me to work without restraints. Even more so I thank my supervisor, and teacher, at Luleå University of

Technology Dr. Bertil Pålsson for never hesitating to tell me when I am wrong, and for pushing me to excel beyond limits I did not think possible. I think we both know I have come a long way since we first met, and I owe very much of that to you.

Finally, all my heart to my families, both the one I have in my hometown and the extra one I have gained by working at the Boliden laboratory. I would not trade anyone of you for anything in the world.

Boliden August 2012

Jennifer Andersson

(5)

1. Introduction

This work was done as a master thesis project to complete a Master of Science degree in chemical engineering, with focus towards mineral processing and process metallurgy.

The aim of this project was to gain knowledge of potential problems in the planned molybdenum separation plant in the concentrator at Boliden Mineral’s mine Aitik. Main focus was on the uncontrollable foaming earlier observed, that caused losses of copper to the molybdenum concentrate and huge problems in the following water treatment process. Different process parameters’ effect on foaming was investigated together with evaluation of possible actions that could be undertaken to prevent or minimize unmanageable foaming.

2. Background

2.1 Molybdenum, general production and uses

Molybdenum is a metallic element, widely used as alloying element in steel production due to its ability to enhance properties such as strength, corrosion resistance, hardenability and weldability. Molybdenum is also a common component in oils and lubricants. It is commonly found in nature as its natural mineral molybdenite (MoS2), often associated with other sulphide minerals commonly

containing iron and copper. This is also the only mineral of commercial interest for recovering molybdenum. Grades in ore bodies typically range from 0,01 % to 0,50 %.

Figure 1 and 2. Pure molybdenum (left picture) and one of many products that contain molybdenum.

Although molybdenum probably has been known and used already by the ancient Greeks, it was not until 1778 the Swedish chemist Carl Wilhelm Scheele was able to isolate molybdenum. Previously, it was likely confused with lead or galena since its name means “lead like” in Greek. In the 19th

century, the first attempts using molybdenum as an alloying metal was made. It was observed that even though molybdenum has almost half the density of tungsten, it could efficiently replace tungsten as alloying metal in some steels. This proved advantageous when World War I caused depletion in tungsten supplies and demands for molybdenum quickly increased. At the end of the war, demands of molybdenum again decreased, and focus was shifted towards finding new areas of use. In the 1930’s use of molybdenum as alloying metal in high speed steels and construction steels

(8)

2

defined molybdenum as an important alloying metal in itself, not only as a substitute for tungsten. [1]

Table 1. Uses of molybdenum globally.

Primary consumption sectors by end-use (source IOMA)

Stainless steels and super alloys 30 %

Low alloy steels 30 %

Chemicals and Mo Metals 20 %

Tool and high speed steels 10 %

Foundry 10 %

2.2 Boliden Mineral AB

The basis for the mining and smelting company Boliden was established in 1924 by the discovery of what would turn out to be Europe’s, at that time, richest and largest gold deposit. Due to the deficiency in metals, as a consequence of World War I, exploration work was intensified during the 1920’s, which led a group of diamond drillers, supervised by Oscar Falkman, to their astonishing discovery. At Fågelmyran approximately 30 km west of Skellefteå in the northern part of Sweden a December day in 1924 the diamond drillers found an ore body that carried on average 15 g gold per tonne of ore, but also significant amounts of copper and silver. Mining operations quickly started along with the building of the town of Boliden close to the mine site. Falkman would remain an important person for the newly formed company, as he was president until 1941.

In 1930, processing of the Boliden ore commenced at the newly built smelter Rönnskär close to Skellefteå. This is still in operation today, processing concentrates from the Aitik open pit mine close to Gällivare and mines within the Skellefteå districts as well as electronic scrap for production of metals from secondary raw material. The Boliden mine closed in 1967, however, the concentrator on-site remained, being a large provider of work for the people living in the town of Boliden. So far, 30 mines have been opened and closed by Boliden within the Skellefteå district only.

Today, Boliden is a modern mining and smelting company with 4400 employees divided into two areas: mining and smelting respectively. The Boliden Mineral AB, which is the mining area, operate the mines Renström, Kristineberg, Maurliden, Maurliden Östra and Kankberg within the Skellefteå district, the large open pit copper mine Aitik close to Gällivare, Swedens oldest still active mine Garpenberg in Dalarna and the zinc mine Tara on Ireland. The annual production of concentrates was 524 Kt of zinc, 76 Kt of lead and 329 Kt of copper concentrate respectively (2011).

The smelting area includes the previously mentioned Rönnskär smelter, together with the Harjavalta and Kokkola smelters in Finland and the Odda smelter in Norway. The Boliden share is listed on the NASDAQ OMX Stockholm exchange as well as the Toronto stock exchange.

Production of the base metals zinc and copper are the main focus of the company. However, production of lead, gold, silver and, soon tellurium is of vast significance for the profitability. Globally, Boliden is a small company regarding the production of copper but are one of the ten largest producers of zinc and in Europe Boliden is one of the biggest producers of all above mentioned metals. [2] [3]

(9)

3

2.3 Aitik open pit mine

Currently, the Aitik copper mine is northern Europe’s largest open pit mine, and Sweden’s by far largest sulphide mine. The area comprises the original open pit together with the new open pit Salmijärvi. The original pit is today over 3000 m long, 1000 m wide and 345 m deep, and the ore zone stretches to a depth of 800 m in some areas of the pit.

During the last years of the 2000´s, the Aitik 36 Mt expansion project was started, increasing annual capacity from 18 Mt to 36 Mt. To meet the increased capacity demands, a brand new concentrator, where copper is recovered by flotation, was built and taken into production in 2011. [4]

Today, annual throughput is approximately 32 Mt of ore, with average grades of 0,24 % Cu, 0,14 g/tone Au and 2,15 g/tone Ag (2011). This places Aitik amongst the top producers of above

mentioned metals in Europe. During 2012, production will reach 36 Mt annually and during the next decade the aim is to increase production even further. [3] [5]

(10)

4

2.4 Geology of Aitik

The Aitik mine is located within an area of metamorphosed plutonic and volcano-sedimentary rocks of Precambrian age (1,9 – 1,8 Ga). Since the host rock has been subjected to metamorphosis, it is highly deformed and the ore mineral grains are finely dispersed within the rock, which results in a low average grade of all valuable metals within the ore body. In order to make a profit out of low-grade ores, open pit mining is the natural choice, together with means to produce and process large quantities of ore.

Within the ore zone, the rock is dominated by two types of schists: garnet-bearing biotite schist in the center of the ore zone and quartz-muscovite-(sericite) schist in areas closer to the hanging wall. A quartz monzodioritic intrusion in the southeastern part of the deposit also belongs to the ore zone. The dominant copper mineral is chalcopyrite (CuFeS2), which accounts for over 98 % of the

total copper bearing minerals present. Chalcocite (Cu2S) and bornite (Cu5FeS4) can be found in trace

amounts. Other ore minerals present comprise pyrite (FeS2, most abundant), magnetite (Fe3O4),

pyrrhotite (FeS1-x), ilmenite (FeTiO3) and molybdenite (MoSs). It has been observed that chalcopyrite

is more abundant in the biotite schist, while the quartz-muscovite-(sericite) schist contains more pyrite.

The ore minerals may be found in several configurations within the ore zone: disseminated, as pure sulphide veins, in veins together with other minerals such as: quartz, amphibole, garnet and

magnetite to name a few, and as patches and clots. Gold is present in its native state, as amalgam (AuHg), and as electrum (AuAg). Several generations of pegmatite dykes are present within the deposit area. Those that occur in the ore zone often contain significant amounts of molybdenite. Molybdenite also occurs disseminated in the host rock schists and in the intrusive unit. [4]

(11)

5

2.5 Flotation theory

The technique of recovering minerals by means of froth flotation has been applied since the early 1900’s and is one of the most important beneficiation processes existing today. It is a versatile, efficient process which enabled treatment of ores that previously would have been considered un-economical due to their complex nature and/or low metal grade.

Flotation is a common separation process, situated after comminution, where valuable minerals are separated from gangue minerals due to their different surface properties when exposed to water and air bubbles. These properties may be natural, or altered by chemicals to create the desired surface properties and enhance the separation. A typical flotation cell consists of a tank with baffles and an agitator to provide both adequate stirring and air bubbles to the cell. As air is introduced to the cell, minerals with hydrophobic surface properties will attach to the bubbles and form a froth layer on top of the pulp. The froth layer may be scraped off and the hydrophilic minerals remain in the pulp. Industrially, several flotation banks are always used to upgrade the mineral froth to desired grade.

Figure 6. Principle of froth flotation.

Although flotation has been practiced for over a century, and the mechanisms are known on a macroscopic level, much is still to be understood regarding the complex nature of the actual flotation process on molecular basis. The process can be considered to consist of three different phenomena namely selective attachment, entrainment and physical entrapment.

Selective attachment, also referred to as true flotation, is the mechanism where the particles attach to the air bubbles and reports to the froth. The other two is different entrainment mechanisms, water carrying hydrophilic mineral particles being trapped in the froth layer or hydrophilic particles being trapped between larger hydrophobic particles. These three mechanisms all contribute to the overall separation, though selective flotation is the dominating. [6]

(12)

6

As previously stated, different chemicals are added to the flotation process in order to achieve the necessary separation. These can be categorized into modifying reagents, collectors and frothers. Collectors are a wide group of different organic compounds that are used for their ability to enhance hydrophobicity on mineral surfaces. Different collectors are used depending on the mineral, and it is important that the collector is selective enough, so it primarily attaches to the correct mineral. The polar end attaches to the mineral and the carbon chain creates a hydrophobic layer surrounding the mineral particle. This attracts the mineral to the air bubble. Collectors can be either anionic, cationic and in some, more un-common, cases non-ionic. Choosing the proper collector depend upon the pulp chemistry and the properties of the mineral.

For flotation of sulphide minerals, which is the minerals of focus in this project, xantathes are most often used. These are anionic collectors and are selective towards sulphide minerals. Sodium and potassium are two most common xanthate salts used and the length of the non-polar carbon chain depends on what selectivity that is desired, however, it rarely exceeds five carbon atoms. In Aitik, the potassium amyl xanthate, referred to as KAX hereby on, is used as collector.

Figure 7. Molecular structrure of potassium amyl xanthate (KAX).

Modifying reagents are classified into pH-controlling reagents, activators and depressants. The role of the modifying agents is to create a pulp environment that provides the best separation between value minerals and gangue minerals. Keeping pH in an optimum range is crucial for achieving optimum separation.

Activators and depressants are chemicals that may enable collectors to attach more easily to the mineral or not attach at all, depending upon what is desired. Lime (CaOH2) is used both as

pH-controlling reagent and depressant of pyrite in the Aitik flotation plant. [7]

Frothers are added mainly to stabilize the froth, but are also believed to have other impacts such as lowering the water surface tension and making bubbles smaller and stronger. There are a wide range of frothers on the market, both natural and synthetically derived. Nasfroth 350, which is a

(13)

7

2.6 Separation of molybdenite from copper concentrate

s

In the Aitik ore, molybdenum is present as molybdenite. Due to the laminar crystal structure of molybdenite after grinding, it is naturally hydrophobic. Thus, molybdenite and chalcopyrite may be efficiently separated from each other by flotation even with low molybdenite grades, if chalcopyrite may be sufficiently depressed.

When recovering molybdenite from copper sulphide concentrates, where chalcopyrite is the dominating copper bearing mineral, it is necessary to depress copper and iron sulphides. This is commonly achieved with some type of depressing agent, cyanides being the most effective, especially for depressing copper sulphides, but scarcely selected due to environmental and safety hazards. Instead, sodium sulphides are the dominating reagents used, sodium hydrogen sulphide, NaSH, being the most utilized reagent industrially. [8] [9]

The mechanism is not quite fully understood, but the depression is believed to be pH and redox potential dependent, starting below -500 mV and may involve desorption of xanthate from the mineral surface according to

2CuX(s) + HS-  Cu2S(s) + 2X- + H+ (1) [10]

The challenges associated with achieving good separation between chalcopyrite and molybdenite includes unwanted oxidation of NaSH, which promotes less depressing of chalcopyrite, and losses of chalcopyrite to the molybdenum concentrate due to excessive water in the froth.

To avoid unnecessary oxidation, nitrogen gas is used as flotation gas rather than air. This prevents oxidation of NaSH into other sulphide species such as polysulphides, thiosulphates and sulphates which will not contribute to the depressing of chalcopyrite. It is also preferable to have a low gas flow during the roughing stages to avoid excessive foaming and losses of copper.

After rougher flotation, regrinding and several cleaning stages often follow to enhance the grade of the molybdenum concentrate. Typically, molybdenum concentrates with molybdenite grades ranging from 70 to 90 % may be produced. [11] [12]

(14)

8

2.7 Beneficiation of the Aitik ore

After primary crushing stages in the open pit, the ore is transported to the beneficiation plant where it is subjected to two stages of autogenous grinding operating in closed circuit with a screw classifier. Prior to flotation the pulp is again classified by hydrocyclones.

The Aitik ore consist of sulphide minerals, thus the flotation process in optimized to recover copper containing sulphide minerals. The two dominating minerals in the Aitik ore are, as previously mentioned, chalcopyrite and pyrite. In the first rougher flotation stages, chalcopyrite is floated from pyrite to a rougher concentrate by addition of KAX and lime (CaOH2). Lime is acting as a depressant

for pyrite.

Figure 8. The flotation plant at the Aitik concentrator.

Middlings from the rougher flotation are treated in the de-pyritization circuit, where the aim is to separate tailings with high and low sulphur grades for separate disposal. Pyrite is therefore floated, by addition of more KAX. The copper rougher concentrate is subsequently up-graded in several cleaning stages by further addition of lime and KAX, producing a final copper concentrate with about 25 % copper,but also gold, silver and molybdenum. Presently, the copper concentrate is sent to the Rönnskär smelter without any separation of copper and molybdenum. [3] [13] [14]

(15)

9

2.7-1 Planned molybdenum recovery process in Aitik

As previously mentioned, there is no current copper-molybdenum separation process in Aitik. However, plans have been made regarding what units that should be included and how the process should be operated. A simplified flow scheme is presented in figure 9 below.

After de-watering of the primary copper concentrate, the pulp is subjected to conditioning with NaSH in two stages before the first rougher molybdenum flotation stages. The concentrate is transferred to a cleaning stage followed by re-grinding and further cleaning; producing after dewatering, a final molybdenum concentrate ready for smelting. The rougher tailings are subjected to a scavenger stage, where the middlings are considered to be the secondary copper concentrate. [15]

Figure 9. Principle scheme of the planned molybdenum separation circuit at the Aitik concentrator. Copper concentrate

with Mo from Cu/pyrite separation circuit Thickener

65 % solids Overflow

Final copper conc thickener Pre-conditioner NaHS stock tank 45% solids Mo-Cu separation flotation circuit

Final copper conc filtration Overflow

Final Mo conc thickener

Final Molly- conc filtration Overflow Make-up water To smelter To smelter Filtrate Filtrate 65 % solids

Air or N2 or recycled air as flotation gas

Process water bleed to the pyrite tailings pond

(16)

10

3. Problem definition

It is of interest to recover the molybdenum present in the Aitik ore as a separate concentrate. A molybdenite separation plant was therefore planned as an extension of the existing flotation plant in Aitik. However, pilot scale test revealed unmanageable foaming in several parts of the process, including both flotation and water treatment stages. This caused multiple problems: The water content in the concentrate was high, resulting in unnecessary losses of copper to the molybdenum concentrate and also putting larger demands on the cleaner stages.

The excessive foaming also caused problems in subsequent water treatment. It was desirable to use air as oxidizing reagent when removing sulphides from the water before discharging it to the tailings pond, but since uncontrollable foaming started as soon as air was introduced; it proved to be more difficult than what was previously thought. If foaming cannot be solved the use of much more expensive chemicals, such as hydrogen peroxide (H2O2), for the same purpose, may be required. [16]

The project thus has two major focuses; finding the cause of the uncontrollable foaming and removing it without affecting the molybdenum recovery negatively. It is also considered important to have gained sufficient knowledge regarding the behavior of molybdenum flotation of the Aitik ore, thus enabling proper education of future process operators.

4. Problem solving strategy

The pulp in the molybdenum flotation process is primarily containing copper concentrate, process water, thio anions and residuals of KAX and Nasfroth 350. None, some, or all of above mentioned species may contribute to the foaming, possibly together with temperature and time dependencies. To evaluate the problems described in the section above, the project was divided into three

different parts namely; evaluation of different analytical techniques, designed conditioning experiments and finally flotation experiments.

A major part of this thesis project is to evaluate plausible analytical method for detecting and quantifying the presence of xantathes and its degradation products, sulphides and frothers. This is considered to be of utter importance since correlation between changes in foaming and

concentration of above mentioned species might provide useful information when determining causes of foaming.

Analytical techniques were chosen after a literature survey of different plausible techniques. Techniques chosen was initially subjected to series of tests with synthetic solution of the chemicals mentioned to evaluate relevance, and if appropriate used as responses during the subsequent designed conditioning and flotation experiments.

Conditioning of copper concentrate was planned to be the first unit process in the molybdenum flotation process and also where NaSH together with copper concentrate would be mixed. Therefore, this stage was investigated by series of fishing experiments, with the aim of observing and gaining knowledge regarding how the foaming changes with alterations in process conditions. Focus was on finding the process condition that would result in minimal problems with foaming.

(17)

11

Results from the designed experiments was subsequently used to design optimal flotation tests, and to determine whether optimal process conditions regarding minimizing foaming affected the

following molybdenum flotation in any way. This was investigated by performing three rougher flotation tests with likely the best process conditions regarding minimizing of foaming.

5. Evaluation of analytical techniques

Three analytical techniques are evaluated in this section: UV-spectroscopy for detecting species in solution, total organic carbon analysis by cuvette tests and quantitative determination of foaming degree by a column test. The techniques were believed to provide information that would

contribute to gaining good knowledge regarding the foaming problems.

5.1 UV-spectroscopy

A UV-VIS spectrophotometer is a device capable of detecting the amount of light absorbed at every wavelength within the UV and visible regions, from 180 to 800 nm. Solutions are placed in

transparent cuvettes during the analysis. In this work, 1 cm quartz cells were used.

The device used within the scope of this project is a Hach Lange DR 5000. Initially, a cell containing appropriate standard solution is used as blank. When the sample subsequently is inserted, the spectrophotometer may display the absorbance of the sample as a function of wavelength.

Most often it is the wavelength of maximum absorption, called λmax, which is of particulate interest.

Many species have maximum absorption at a wave length, which is shared with few or no other species. Therefore the λmax can be used as a direct indication that a certain species is present. Adding

to that, knowledge of λmax for samples with known concentrations provide means of creating

standard curves, which may subsequently be used for determining unknown concentrations of the same sample.[17] [18]

(18)

12

5.1-1 UV-spectroscopy for detection of xanthates and evaluation of interference from other species

Determination of xanthates in solution can be readily done by UV-spectrometry, due to xanthates having a sharp absorbance maximum at 301 nm, see figure 11. Woodcock and Jones [19] has done thorough experiments regarding this.

By constructing calibration curves for absorption of xanthates in water at different concentrations the concentration may be determined in an un-known sample. Depending on other species present, Woodcock and Jones propose different analytical methods for assaying xanthates.

Figure 11. UV spectra of KAX in de-ionized water.

If no other substances are present that contributes significantly to the absorbance at 301 nm, xanthate concentration may be determined by direct measurement of the solution (method 1). However, if other, un-known, species are present, both dissolved and dispersed, it is recommended that the pH of the solution is adjusted to between 6,5 and 6,6, measured for absorbance at 301 nm, and then acidified below a pH value of 2 which decomposes xanthate. pH is thereafter re-adjusted to pH 6,5-6,8 and absorbance measured again. The xanthate concentration is the difference between the two measurements (method 2). [19]

The KAX used at the Boliden development laboratory had a maximum absorbance at 300 nm, therefore this was the wave length considered. To determine the possibilities of using spectrometry within this work, synthetic solutions of KAX was prepared and analyzed in the UV-spectrometer. This was done with both de-ionized water and process water collected in Aitik, by method 1 described above. The choice of method 1 was due to difficulties to correctly adjust pH and also due to formation of H2S-gas upon acidification.

0 0,5 1 1,5 150 200 250 300 350 400 A b sor b an ce Wavelength (nm)

(19)

13

Figure 12 shows the calibrations curves for KAX in de-ionized water. As can be seen, there is a linear connection between absorbance and concentration if no interfering components are present.

Figure 12. Calibration curve of KAX.

However, the solutions produced in this report do not only contain KAX but also, thio anions,

Nasfroth (frother) and possible finely dispersed particles. In order to correctly assay presence of KAX, it must be investigated if and to what extent the other species present may interfere with the

absorbance at 300 nm.

Woodcock and Jones investigated the effect of a wide range of frothers to the absorbance at 301 nm. Depending on the frother used interference may be observed on the absorbance at 300 nm. [20]

Synthetic solutions of 1 g/l Nasfroth 350, the frother used in Aitik, were prepared and assayed in the spectrophotometer. As can be seen in figure 13, Nasfroth 350 contributes almost nothing to the absorbance at 300 nm, even at high concentrations. Therefore it can be concluded that the presence of frother do not affect the maximum absorption at 300 nm in any significant way.

(20)

14

Figure 13. UV spectrum of Nasfroth 350 in de-ionized water.

To evaluate whether addition of sulphides interfere with the absorbance at 300 nm, synthetic solutions with different concentrations of Na2S were prepared and assayed in the

spectrophotometer. Figure 14 shows the standard curve resulting from the tests. As can be seen, the contribution from sulphides at 300 nm increased with the concentration.

Figure 14. Absorbance at 300 nm for different concentrations of Na2S.

As was concluded by Woodcock and Jones, detecting KAX by UV-spectrometry can be readily done. Therefore, it can be concluded that this is an appropriate and easy method for detecting presence of xanthates. The linear relationship was observed below 0,3 absorbance units approximately. If higher absorbances are found, dilution of the samples should be made prior to analyzing. [21]

In presence of sulphides determining accurate amounts of KAX may be difficult. For later

experiments, it was decided to partly compensate for this error by withdrawing the contribution of the baseline absorbance at 300 nm.

(21)

15

5.2 TOC

Total Organic Carbon (TOC) is analyzed by a simple cuvette test provided by Hach Lange. The sample is placed in a cuvette and acidified before being introduced, without lid, to a shaker for 5 minutes. This removes the inorganic carbon (TIC). The sample cuvette is subsequently connected by a membrane to an indicator cuvette containing solution that changes colour upon alterations in pH. The assembly is placed in a heating block for 2 hours at 100 degrees, causing organic carbon to oxidize into carbon dioxide, travel trough the membrane and into the indicator cuvette. The change in colour is measured in a Hach Lange DR 2800 spectrometer and corresponding TOC is calculated automatically. [22]

Hach Lange confirms that the test is sensitive to the following contaminants:

Table 2. Summary of ions determined to affect the TOC analysis. Concentration (mg/l) Substance

2000 Ca2+, Mg2+

1000 Cl

-250 TIC (total in-organic carbon)

200 NH4-N

Hach Lange has not done any sensitivity analyses regarding presence of thio anions in the sample. In the scope of this work, it was chosen to not remove TIC before placing the sample in the heating block. This was mainly due to occupational safety reasons as sulphide containing solutions forms H2S

gas upon acidification, because earlier observations at the Boliden development lab indicated that this procedure removed some of the organic carbon and finally, TIC content was not considered significant.

5.2-1 Determination of single chemicals effect on the TOC analysis

The possibility to correctly assay the carbon content in KAX and Nasfroth with and without presence of sulphides was evaluated by preparing synthetic solutions which were subjected to the Hach Lange tests. The different water types were also analyzed with TOC.

Results obtained were compared with theoretical TOC values for KAX and Nasfroth, values that were determined by considering the chemicals formulas and molecular masses see figures 15, 16 and table 3. From this information theoretical carbon content can be assayed.

Figure 15 and 16. Molecular formulas of KAX and Nasfroth.

(22)

16

Table 3. Molecular weights and carbon content of KAX and Nasfroth. Species Molecular weight

(g/mole) Molecular weight carbon (g/mole) Carbon content (%) KAX 202,38 72,06 35,6 Nasfroth 350 App. 350 218,46 62,6

The theoretical carbon content in KAX and Nasfroth 350 are 35,6 and 62,6 % respectively. Table 4 displays the carbon content assayed by the Hach Lange tests. The three first tests were done with different water qualities only, and it can be seen that the process water in Aitik contain small amounts of carbon probably due to residuals of collectors and frothers present.

Test 4 and 5 shows that the Hach Lange cuvette test accurately assays the carbon content of the two compounds when the solution is prepared with de-ionized water. Would process water be used, an increase of TOC should of course be expected.

Test 6, with only Na2S and de-ionized water is unfortunately also contributing to the TOC value. This

is likely due to limitations in the Hach Lange test, which use change in colour in the pH-sensitive solution as a measure of TOC. However, since H2S-gas will form upon acidification of a solution

containing sulphides, it is likely that the H2S reporting to the pH sensitive solution will cause errors in

the TOC analysis.

Table 4. Results from TOC analysis of different solutions.

Test Type of water Type of chemical Concentration (mg/l) TOC (mg/l) measured TOC (mg/l) theoretical 1 De-ionized - - 0 0 2 Process - - 9,98 - 3 Cu-conc - - 13,6 - 4 De-ionized KAX 1000 340 356 5 De-ionized Nasfroth 350 1000 635 626 6 De-ionized Na2S 5000 33 -

(23)

17

5.3 Height of foam in column

A straight forward, quantitative and user friendly method for easy evaluation of foaming is

measuring the height of foam in a column, made out of a cylindrical glass shell with a glass filter in the bottom. Solution is placed in the column and as gas is introduced from the bottom it ascends through the filter and into the solution. The height of foam is considered corresponding to the maximum height of the foam pillar in the column, and is measured in centimeters. [23]

Figure 17. The column used for the foaming tests.

For tests within the scope of this work nitrogen gas was used, since it was the flotation gas used in subsequent flotation experiments. Analyses by the column were done by two column tests referred to as height of foam 1 and 2 respectively in the report. Height of foam 1 is the average value of three consecutive tests, with fresh rinsing and fresh solution (from the same batch) between each test, whereas height of foam 2 corresponds to the height measured in the second test only.

5.3-1 Determination of single chemicals effect on foaming

In the flotation process KAX, Nasfroth and sulphides may be considered the dominating chemicals present. Therefore, the single effect of each was evaluated by creating synthetic solutions with different concentrations, Nasfroth being excluded since it was considered unnecessary. Water used was de-ionized water, process water from Aitik and water from copper concentrate.

The height of foam was measured one time per solution, the results displayed in figures 18, 19 and in table 5.

(24)

18

Figure 18. Results from height of foam tests on synthetic solutions without sulphides.

Synthetic solution with de-ionized water and KAX show no foaming properties, therefore it appears as if KAX alone cannot be the cause of observed foaming. However, process water and water from the copper concentrate have foaming properties, which is most likely due to residuals of frothers and possible other species present.

Figure 19. Results from height of foam tests on synthetic solutions with sulphides.

Freshly prepared sulphide solutions do not appear to cause foaming either. However, since

(25)

19

Table 5. Summary of results from height of foam tests.

Test Water Na2S (g/l) KAX (mg/l) N2 (l/min) Height of foam (cm)

1 De-ionized 0 0 5 0 2 Process 0 0 5 43 3 Cu-conc 0 0 5 55 4 De-ionized 5 0 5 0 5 Process 5 0 5 43 6 Cu-conc 5 0 5 55 7 De-ionized 0 100 5 0 8 Process 0 100 5 40 9 Cu-conc 0 100 5 55 10 De-ionized 5 100 5 0 11 Process 5 100 5 30 12 Cu-conc 5 100 5 51

It was decided that the three analytical techniques evaluated in the sections above should be used as responses in subsequent conditioning experiments, even though the TOC analysis was

(26)

20

6. Conditioning experiments

Based upon the results in the previous section, it was decided that a limited amount of qualitative experiments should be made with focus on the conditioning stage. For this purpose the software MODDE was used to create an experimental design, which could detect the correlation between chosen factors and responses.

A full factorial screening design was chosen. This is commonly used in the early stages of an experimental procedure, for gaining information regarding whether the factors have influence on the responses or not. Three factors, with two levels, were chosen together with four responses. With three replicate experiments at the center point, this resulted in 11 experiments totally. [24]

6.1 Chosen factors and responses

Before factors and responses were selected, a series of conditioning tests were made to determine whether it was necessary to include a time aspect. Figure 20 displays the absorbance at 300 nm during different time intervals of a conditioning test with 5 g/l Na2S. Already after a couple of

minutes the maximum absorbance is reached, indicating that desorbing of xanthates occurs immediately. Therefore, it is concluded that conditioning time is not an important parameter

regarding degree of desorption of KAX from the mineral surface and that it should suffice with only a few minutes if necessary.

(27)

21

Instead, the temperature of the conditioning (ᵒC), amount of NaSH added (kg/tonne concentrate) and quality of process water (ranging from de-ionized water to process water) was chosen as factors, while absorbance at 300 nm, height of foam from two tests and TOC represented the responses. Table 6 displays within which ranges the factors were selected.

Table 6. High and low levels on chosen factors.

Factor Unit Low level High level

Temperature ᵒC 5 25

NaSH kg/tonne 2 8

Process water % of total water 0 100

The level of the factors were set in such manners that low and high values would represent extreme conditions in the processing plant, making it possible to more easily detect dependency of responses upon factors.

Data resulting from the experiments were fitted with multiple linear regressions (MLR) and evaluated in MODDE by various tools/graphs resulting in a model for each response.

(28)

22

6.2 Sample preparation

Copper concentrate from Aitik was collected from the sampling point from which pulp is sent to the x-ray equipment for online assays. It was kept in a container under water to prevent oxidation of mineral surfaces. Prior to each experiment, the desired amount of concentrate was subtracted from the container during agitation. After filtration the concentrate was subjected to two minutes of comminution with stainless steel rods as grinding media, this was done to ensure fresh mineral surfaces and increase reproducibility. Grinding was always done with the same water as used in subsequent experiment. The conditioning was made at approximately 40 % solids by weight. Process water was collected directly at the plant and kept in plastic containers.

Instead of NaSH, Na2S*3H2O was used as depressant, mainly due to lack of fresh NaSH and thus lack

of control regarding the degree of oxidation of the available NaSH. Therefore Na2S*3H2O was used,

but added in quantities corresponding to equivalent amounts of NaSH calculated on sulphur basis.

6.3 Method

The run order provided by MODDE was followed. Correct amounts of Na2S*2H2O was dissolved in

1,5 dm3_{of the water quality specified and placed in a tank fitted with baffles, stirrer, water bath for}

heating and a cooler. The temperature of the solution was set with the water bath/cooler during purging of N2 gas to avoid oxidation of Na2S. When steady state was reached, copper concentrate

that had been prepared according to previous section, was added. Conditioning time was 30 minutes. Flow of N2 gas during conditioning was 7 dm3/min.

After 30 minutes, a 40 ml sample was withdrawn during stirring, filtered using Munktell micro-glass fibre paper, grade MGA, and analyzed by UV-spectrometry and TOC. The remaining solution was filtered and used for foaming tests in the column.

Figure 22. The experimental setup. To the far left is the reactor, equipped with stirrer. In the middle can thewater bath and cooler be seen and the nitrogen gas is to the far right of the picture.

(29)

23

6.4 Results and discussion

In table 23, the results are displayed as entered into MODDE. The model was fitted using MLR and various graphs were used to display the results and evaluate the factor influences on selected responses. Interpretation of different graphs is done individually and a summary of results together with summary of conclusions may be found at end of the section

Figure 23. Results from the conditioning experiments as inserted in MODDE.

Generally, it can be observed that the values for the responses Height of foam 1, 2 and Abs 300 nm respectively are within a quite narrow range, which is important to remember when interpreting graphs in later sections.

6.4-1 Coefficient plots

This section displays coefficient plots, useful graphs that provide information regarding both effects of factors and experimental noise. The coefficients are scaled and centered. The first three

coefficients provides information regarding each parameter’s main effect upon the responses, whereas the three latter displays interactions between the factors. Experimental noise is calculated from the value spread for centre point and shown by the confidence intervals, uncertainty increased with increased size of the confidence intervals.

(30)

24

6.4-1.1 Coefficient plot, Height of foam 1

Figure 24. Coefficient plot for the response height of foam 1.

In figure 24, the coefficient plot for response Height of foam 1 is displayed. Noticeable are the large confidence intervals, indicating high levels of experimental noise. All factors are non-significant, exception being process water which is strongly negatively correlated to height of foam 1. The other main effects show tendencies towards weak positive correlations between temperature and amount NaSH, but since it is tendencies no major conclusions should be made.

6.4-1.2 Coefficient plot, TOC

(31)

25

In figure 25 above, the factors effect on the TOC response is shown. Again, note the large confidence intervals. The only significant factor appears to be NaSH, which is positively correlated to the

amount of TOC. This is the most important result from this graph, and supports the observations made during the initial experiments, i.e. that sulphides contribute to the TOC values measured. No other factors are significant, since the confidence intervals pass through zero, however tendencies towards negative correlation can be seen. Especially for water quality, this may seem peculiar since process water contain more carbon than de-ionized water. However, due to the extent of the confidence intervals, this should not be over interpreted.

6.4-1.3 Coefficient plot, 300 nm

Figure 26. Coefficient plot for the response Abs 300 nm.

Coefficient plot for the response 300 nm are displayed in figure 26. All confidence intervals are large and pass through zero, therefore results displayed here is interpreted as tendencies. Noticeable are the tendencies towards negative correlation for main factors temperature and water quality, which in combination display positive tendencies.

What may be concluded from this graph is that Abs at 300 nm is not a suitable response for the chosen factors, or vice versa.

(32)

26

6.4-1.4 Coefficient plot, Height of foam 2

Figure 27. Coefficient plot for the response height of foam 2.

In figure 27, the coefficient plot for response Height of foam 2 is displayed. Generally, this response displays the smallest confidence intervals, indicating the least level of experimental noise of all the responses. Only temperature is non-significant, and the remaining factors are fairly or strongly significant. This concludes that Height of foam 2 was the response best explained by the chosen factors.

The strongest factor is evidently quality of process water which is strongly negatively correlated to the height of foam. Amount of NaSH is weakly positively correlated, whilst all the interaction parameters are inversely correlated to height of foam 2.

(33)

27

6.4-2 Main effects plots

Main effect plots are useful when investigating how one factor, when changing from its high value to its low value, influences the responses when all other factors are kept at their average values. The blue dots represent the center points. In this section, only the significant plots will be displayed, the remaining plots may be found in appendix.

6.4-2.1 Main effects, temperature

In this section, the main effect of temperature on the responses is displayed. The main effect of increased temperature seems to result in lower values of the TOC response. Regarding the TOC response it may be explained by increased evaporation, resulting in increased losses of sulphides with vapor.

Figure 28. Main effect plots of the factor response for the responses TOC and height of foam 2.

Regarding the change in temperatures effect on the response height of foam 2, it can be observed that an increase in temperature seem to result in slightly higher foam values compared to low temperatures. It appears as if foaming is beneficiated from higher temperatures, which may be due to more species with foaming properties releasing from the mineral surfaces during conditioning at higher temperatures compared to lower.

6.4-2.2 Main effects, NaSH

The main effect of changing amount of sulphide from low to high values appears to be increased values of both the height of foam 2 and TOC values. The connection is more evident for the TOC response, again supporting previous observations that presence of sulphide ions contribute to increased TOC values, when using the Hach Lange test. The tendency for the height of foam value

(34)

28

yet again supports what has been observed earlier, that increased NaSH content may worsen foaming.

Figure 29. Main effects plot of the factor NaSH for the responses TOC and height of foam 2.

6.4-2.3 Main effects, Processwater

Investigating the main effect of process water to the responses it can be seen in figure 30 that the strongest effect of increased process water content is decreased foaming. It is evident that foaming properties are weakened with dirtier water quality. The response TOC show lower values at high levels of process water, which may indicate that less carbon containing species are released to the solution in this case.

(35)

29

6.4-3 Cross correlation plots

With cross correlation graphs, it is possible to evaluate the effect of combined factors on the

responses. If the lines cross, the interaction of the two factors may be significant for the value of the response investigated. In this section, selected graphs are evaluated; the rest may be viewed in appendix.

6.4-3.1 Cross correlations, temperature*NaSH

The interaction effect between temperature and sulphide was significant for the response height of foam 2 only. Since the value was decreased with higher temperatures at high levels of sulphides it may indicate that something that enhances foaming, perhaps colloidal sulphur, is consumed or evaporated.

Figure 31. Interaction plots for the combined factors temperature and NaSH for the responses Abs 300 nm and height of foam 2.

As figure 31 also displays, there are no significant interaction effects between temperature, sulphide content and absorbance at 30 nm. What can be seen is a decrease in absorbance at 300 nm with temperature, possibly due to consumption and/or evaporation of sulphide ions. This is true for both high and low levels of NaSH.

(36)

30

6.4-3.2 Cross correlations, NaSH*processwater

There are no significant interaction affects for amount of sulphides and process water on the responses, only addition effects. Figure 32 illustrate this for the two responses abs 300 nm and height of foam 2.

Figure 32.Interaction plots for the combined factors NaSH and process water for the responses Abs 300 nm and height

of foam 2.

6.4-3.3 Cross correlations, temperature*process water

For the response TOC, the interaction effect between temperature and process water appear significant, for some still unexplained reason.

Figure 33. Interaction plots for the combined factors temperature and process water for the responses height of foam 1 and TOC.

(37)

31

6.4-4 Contour plots

Since height of foam 2 appeared to be the most significant response, it is the only contour plots shown in this section, the remaining can be seen in appendix. Contour plots are similar to interaction plots in displaying two factors influence on the chosen response, while keeping the third at different levels. Illustrated in figure 34 are the contour plots of height of foam 2, when the sulphide

concentration and temperature are varied from low to high values, at the three different levels of the process water parameter.

Figure 34. Contour plot for the response height of foam 2.

As can be seen, it is preferable to maintain a low level of sulphides together with lower temperature. Therefore, if problems with foaming arise in the actual plant, there is no reason to add even more depressant; foaming might even be due to excessive amounts of depressant. Also, as was discussed before, the use of process water does not seem to increase foaming; therefore, there is no reason why process water from the plant should not be used.

(38)

32

7. Flotation experiments

The main focus of this work is, as previously stated, to investigate different process

parameters/process conditions effect on the extensive foaming. It is, however, also important to determine whether actions towards lowering foaming could affect the recovery and selectivity of molybdenite in any way. This was checked by three flotation experiments, the procedure based on results from previous sections and earlier work by Boliden. [25]

It was decided to include activated carbon as a parameter, since it had been observed that activated carbon may enhance separation between molybdenite and copper sulphide minerals during

flotation. Also, it can be readily obtained by Boliden since activated carbon is already used in the cyanide leach plant within the Boliden area. [26]

The aim was to investigate whether choice of “optimal” conditions based on results from the designed experiments together with use of activated carbon would affect the recovery of

molybdenum, and amount of water in concentrates during rougher flotation. The amount of water in concentrates was considered important because it might give an estimate whether losses of copper to the concentrates are transported by the water (entrainment) or by the gas bubbles (true flotation) to the concentrate.

From the conditioning experiments, the following process conditions were chosen:

 Flotation was done with process water at ambient temperature and nitrogen as flotation gas.

 Amount of NaSH was chosen at its center point, i.e. 5 g/tonne concentrate.

 Conditioning time 30 minutes prior to flotation.

 Analyses done were assay of elements in all products, including mp, together with analysis by UV spectrometry and height of foam on the solution of above mentioned products. TOC analysis was excluded.

7.1 Material and sample preparation

Copper concentrate was collected at the same point as previously, i.e. at the sample point of the X-ray equipment. The first test was considered an orientation test; the material was therefore collected one day prior to the material used for test 2 and 3. Since the flotation experiments were made in the laboratory on site in Aitik, the material could be used while fresh, however, two minutes of mild steel grinding using rods was still done since it had been done in all other tests. Process water was used during comminution.

(39)

33

7.2 Method

After comminution and decantation, the concentrate was placed in the flotation cell. Process water was added, together with correct amount of Na2S*3H2O and 2 g of activated carbon in two of three

cases. Stirring was started, but no nitrogen flow due to the flotation cell overflowing. pH and redox potential was noted. After 30 minutes the flotation started by enabling gas flow carefully. The pulls were made with flotation times 30, 60 and 90 seconds in each stage respectively.

All products, including middling, was weighed wet and then filtered. The material was covered and dried in an oven over night, while the solutions were assayed with UV-spectrometry and the height of foam column. The day after, all dry products were weighed and representative samples sent for analysis of elements by ICP/mass spectrometry.

Table 7. Experimental setup.

Test Sample collected NaSH (g/tonne) Carbon added (g) Temperature (ᵒC) Redox potential (mV) 1, carbon 2012-06-19 5 2 19 -613 2, no carbon 2012-06-20 5 19 -610 3, carbon 2012-06-20 5 2 19 -623

7.3 Results, flotation experiments

In this section, the results from the flotation experiments are displayed, which comprise results from the analytical devices and ICP/mass spectrometry together with a water balance. The results are summarized and further discussed at the end of the section.

7.3-1 Results, UV-spectrometry and height of foam in column

Test 1 and 3, which were done with addition of activated carbon during conditioning showed significantly less foaming properties during the foam tests. Foaming was to a large extent reduced in the middling product, and significantly decreased compared to test 2, which was done with no activated carbon. This result appears promising since it is the first parameter observed that actually has a positive influence on the foaming. As figures 35, 36 and table 8 also shows, both foaming and the absorbance at 300 nm is decreased significantly in the presence of activated carbon.

(40)

34

Figure 35. Results from the height of foam tests on all products from the flotation experiments.

(41)

35

Table 8. Summary of results.

Test Height of foam 1 (cm) Height of foam 2 (cm) Abs 300 nm Test 1, Conc 1 10 9 0,55 Test 1, Conc 2 16 15 0,55 Test 1, Conc 3 13 10 0,40 Test 1, mp 4 3 0,40 Test 2, Conc 1 49 40 2,8 Test 2, Conc 2 >55 >55 2,0 Test 2, Conc 3 55 47 2,05 Test 2, mp 22 23 2,5 Test 3, Conc 1 24 22 0,9 Test 3, Conc 2 20 23 0,75 Test 3, Conc 3 18 17 0,55 Test 3, mp 12 19 0,6

7.3-2 Results, distribution of solids and water in products

The content of water in products tended to decrease slightly during the flotation experiments where activated carbon was used. This is in agreement with activated carbon decreasing foaming. If carbon decreases foaming, it should also decrease water content in product, which is showed here. Test 1 reveals the least water content in the first two concentrate when comparing to the other test with carbon, however, since the sample was collected one day prior to the other this behavior may have other explanations such as different ore/chemical additions etc.

What should be focused on here is that use of activated carbon has certainly not any negative effect on the water distribution in products. It is also important to recognize that a low flow rate of

nitrogen gas is important to keep foaming at a manageable level

Figure 37. Distribution of water in products.

0 10 20 30 40 50 60 70 80 90 100

Conc 1 Conc 2 Conc 3 mp Total

D istr ib u tion (% )

Distribution of water in products

Test 1, carbon (orienting) Test 2, no carbon Test 3, with carbon

(42)

36

Table 9 shows the resulting weights from the flotation experiments.

Table 9. Summary of weights and water distribution of all products.

Test Weight solids (g) Weight water (g) Distribution water (%)

Test 1, Conc 1 6,47 106,35 4,2 Test 1, Conc 2 7,62 178,02 6,9 Test 1, Conc 3 8,36 320,44 12,5 Mp 507,39 1956,66 76,3 Test 2, Conc 1 10,94 165,27 6,1 Test 2, Conc 2 18,11 241,08 8,9 Test 2, Conc 3 17,60 331,59 12,2 Test 2 mp 525, 61 1971,79 72,8 Test 3 Conc 1 14,66 161,94 6,0 Test 3 Conc 2 11,89 221,92 8,2 Test 3 Conc 3 8,94 299,87 11,1 Test 3, mp 525,92 2018,48 76,7

7.3-3 Results, metal assays and selectivity

Metal assays presented in table 11 were provided by ALS Chemex in Öjebyn, Sweden. Full mass balance, together with grade/recovery and selectivity diagrams may be found in appendix.

It appears as is presence of activated carbon is not affecting the flotation of molybdenum negatively. On the contrary, as table 10 displays, the presence of activated carbon improves selectivity towards molybdenum. It may be due to less copper reporting to the concentrate by entrainment in the excessive foam.

Table 10. Selectivity index, calculations based on the sum of concentrate 1 and 2.

Test Selectivity index

1 89,6

2 17,5

(43)

37

Table 11. Metal assays provided by ALS Chemex.

Test Weight (g) Assays Content (g) Distribution (%)

Mo (ppm) Cu (%) Mo Cu Mo Cu Test 1, Conc 1 6,47 360 000 5,47 2,329 0,354 43,4 0,5 Test 1, Conc 2 7,62 335 000 6,85 2,553 0,522 47,6 0,8 Test 1, Conc 3 8,36 32000 14,87 0,268 1,243 5,0 1,6 Test 1, mp 507,39 427 12,49 0,217 63,373 4,0 96,8 Test 2, Conc 1 10,94 9170 25,09 0,010 2,745 0,5 2,2 Test 2, Conc 2 18,11 25400 24,13 0,460 4,370 22,6 3,4 Test 2, Conc 3 17,60 80600 21,58 1,419 3,798 69,8 3,0 Test 2, mp 525,61 272 22,1 0,143 116,160 7,0 91,4 Test 3, Conc 1 14,66 100 90 20,44 1,479 2,997 72,4 2,4 Test 3, Conc 2 11,89 27800 24,28 0,331 2,887 16,2 2,3 Test 3, Conc 3 8,94 11500 25,11 0,103 2,245 5,0 2,8 Test 3, mp 525,92 250 22,1 0,131 116,228 6,4 93,5

8. Final conclusions

It can be concluded that there are uncertainties regarding most of the responses used during the designed experiments. The most useful and reproducible appears to be the Height of foam value 2, provided by the column test. This is positive, since the test is cheap, user friendly and can readily be done within minutes. It may be a quick mean for operators to confirm or reject suspicions of excessive foaming in any stream on-site.

Using UV-spectrometry for detecting xanthates can readily be applied, but determining the exact concentration is more difficult due to sulphides contributing to the absorbance at 300 nm. Therefore using absorbance at 300 nm as means of assaying amount of xanthates in solution might be

inappropriate considering the results obtained. It may be due to sulphides disturbing, and/or because the extraction of KAX from the mineral surface is not dependent on the chosen factors. However, since there is no evidence supporting KAX being the main cause of foaming, it is definitely questionable whether absorbance at 300 nm is worth considering at all.

The frother used, Nasfroth 350 cannot be detected by UV-spectrometry in this case, since the peak is within the same area as the much larger absorbance spectrum of sulphides. Therefore, it is difficult to assess the change in frother concentration during the designed experiments.

The Hach Lange cuvette tests should not be used on solutions used in this project, which is evident from all results displayed. The concentration of sulphides present in the solutions used is too large and the results cannot be trusted. Dilution of the sample prior to analysis is not preferable since it lower the contribution of carbon containing species to such limits that it is barely detectable by the spectrometer.

It can also be concluded that neither of fresh Na2S or KAX, alone or together contribute to foaming.

(44)

38

Nasfroth cannot be detected by UV-spectrometry when sulphides are present it cannot be evaluated whether it is the frother released from mineral surfaces that contribute to the foaming observed. However, process water and water from copper concentrate have foaming properties, likely due to residuals of frother and/or other foaming species present. Surprisingly, increased amount of process water decreased foaming properties of the filtered water after conditioning. There are several plausible reasons for this, which will only be speculated on here. There might have been something erroneous with the quality of the de-ionized water, i.e., it was not completely ion free. Another cause may be the low ionic strength of de-ionized water, possibly causing more extraction of species with foaming properties from the mineral surfaces.

Nevertheless, the results indicate that there is no clear change to the better if de-ionized water is used. Therefore nothing is, at this point, indicating that process water cannot be used in case of installation of an actual molybdenum separation plant. Also, evaluation of temperatures effect on foaming indicates that colder temperatures might be slightly more beneficial than warmer but the difference is not of great magnitude.

Regarding the NaSH content, the most interesting conclusion is that foaming increase with NaSH addition. In practice, this means that there is no benefit in adding more NaSH to the conditioning stage, if problems with foaming arise. On the contrary, the results indicate the opposite. This information may be useful when dealing with actual process control at the future plant, as operators might easily believe higher dosage of depressant hinders foaming as well.

Activated carbon was the only species proven effective enough to lower the foaming in a significant manner. It also appears to increase selectivity of molybdenum. Since it can be readily obtained, the only draw-back might be difficulties in implementing use of activated carbon to the process in an efficient and economical way. A low flow rate of the flotation gas was also concluded of vast importance. It should be further evaluated exactly how low it must be, but from the experience gained during the flotation experiments done within this report it can be concluded that it is much lower than the flow rates used in any standard copper sulphide flotation test.

The main conclusions can be summarized as follows:

 Over-dosage of NaSH does not depress foaming, it rather makes it worse.

 There is no evidence supporting KAX being the cause of foaming.

 Nothing indicates that changing from process water to de-ionized water depress foaming.

 The Hach Lange method for measuring TOC should not be used on solutions high in sulphides.

 Use of activated carbon depress foaming and do not appear to influence molybdenum recovery negatively.

(45)

39

9. Future work

 Further flotation experiments in laboratory scale to investigate if molybdenite recovery could be retained, and foaming even further decreased if NaSH was added to a redox potential of -520 mV instead of below -600 mV, with presence of activated carbon.

 Flotation experiments to investigate appropriate gas flows to minimize excessive foaming.

 Determination of the amounts and retention times of activated carbon necessary to depress foaming and to what extent the carbon could be re-used within the process.

 Additional condition experiments, to fully understand the un-expected correlation between decreased foaming and increased amount of process water.

Okontrollerad skumning i molybdenflotation: Studier och lösningar

MASTER'S THESIS

Excessive foaming in moybdenum

flotation

- studies and remedies

Jennifer Andersson

Master of Science in Engineering Technology

Chemical Engineering Design

Abstract

Sammanfattning

Acknowledgments

Table of Contents

1. Introduction

2. Background

2.1 Molybdenum, general production and uses

2.2 Boliden Mineral AB

2.3 Aitik open pit mine

2.4 Geology of Aitik

2.5 Flotation theory

2.6 Separation of molybdenite from copper concentrate

s

2.7 Beneficiation of the Aitik ore

3. Problem definition

4. Problem solving strategy

5. Evaluation of analytical techniques

5.1 UV-spectroscopy

5.2 TOC

5.3 Height of foam in column

6. Conditioning experiments

6.1 Chosen factors and responses

6.2 Sample preparation

6.3 Method

6.4 Results and discussion

7. Flotation experiments

7.1 Material and sample preparation

7.2 Method

7.3 Results, flotation experiments

Distribution of water in products

8. Final conclusions

9. Future work