Pretreatment Methods for Manganese Containing Anode Sludge

(1)

Pretreatment Methods for

Manganese Containing Anode Sludge

Johan Stenman

Sustainable Process Engineering, master's level 2017

Luleå University of Technology

Department of Civil, Environmental and Natural Resources Engineering

(2)

i

Acknowledgements

I would like to thank my MEFOS supervisors Marcel Magnusson, for help with the report and analysis, Tobias Lindbäck, for being invaluable in the planning and execution of furnace experiments, and Ulf Sjöström, for assistance with agglomeration work and pellets analysis. I would also like to thank the rest of the staff at Swerea MEFOS for their companionship and assistance, in particular Fredrik Skemark for his help in preparing and performing the furnace experiments. Additional thanks to my supervisor Associate Professor Fredrik Engström and the rest of the Process Metallurgy staff at LTU for always being very helpful with calculations and sample analysis. I would also like to extend my thanks to my Boliden supervisor Justin Salminen, as well as Aija Rytioja at Boliden Kokkola, for providing chemical analysis and information regarding their operations.

And finally many thank you’s to my friends, family and my fiancée Maria Eriksson for supporting me and helping me keep my sanity when things got stressful.

(3)

ii

Abstract

This master thesis work examines whether it is possible to separate lead from an electrolytic sludge rich in manganese using pyrometallurgical treatment, and also attempts to determine the optimum process parameters for such a treatment. It also includes a theoretical study of the possible

applications for lead and manganese, as well as thermodynamic calculations predicting the behaviour of the material at increasing temperatures.

The experimental work completed includes characterisation and agglomeration of the raw material, as well as tests in a chamber furnace and a rotary furnace. The anode sludge was characterised using chemical analysis, XRD, SEM and PSD. The anode sludge was agglomerated into pellets using either both bentonite and water, or only water as binder. The smaller scale tests in the chamber furnace examined the impact of several variables on lead removal. These variables included type and amount of reduction agent used, temperature, and whether the anode sludge was added as untreated material or pellets. The most promising of these results were further tested in the rotary furnace at a slightly larger scale. The variables used for the rotary furnace tests were amount of reduction agent added, whether the anode sludge was added as untreated material or pellets, and whether the reduction agent was added at the start of or during the experiment. All samples were sent for chemical analysis, and selected samples were further analysed using XRD and SEM.

The conclusions drawn from the results of the thermodynamic calculations and experimental work are as follows:

• In the untreated anode sludge the primary phases are MnO2, CaSO4, and (Pb,Sr)SO4.

• The anode sludge can be agglomerated into pellets, with or without added binder.

• After treatment the primary phases present are MnO and (Ca,Sr)2SiO4. Lead is present as small separate grains.

• The most effective treatment method should adhere to the following parameters:

o Use of a rotating furnace.

o Anode sludge added in the form of pellets, to simplify materials handling.

o Temperature of 1400-1500 °C.

o Reduction agent added in batches after initial smoke formation has stopped.

o Total addition of reduction agent should be 10 wt% of anode sludge.

• Significant weight loss occurs during treatment.

• The amount of reduction agent added to the anode sludge has the greatest effect on the removal of lead and zinc.

• Charcoal is a potential candidate for a renewable reduction agent, but leads to increased weight loss.

• It is possible to separate lead from the manganese anode sludge using pyrometallurgical treatment, down to 100 ppm. Zinc can also be separated, down to 600 ppm.

(4)

iii

Sammanfattning

Detta examensarbete undersöker huruvida det är möjligt att separera bly från ett anodslam rikt i mangan med hjälp av pyrometallurgiska behandlingsmetoder, och försöker även avgöra de optimala processparametrarna för en sådan behandling. Arbetet inkluderar även en teoristudie av möjliga användningsområden för bly och mangan, samt termodynamiska beräkningar som förutser materialets beteende vid ökande temperaturer.

Det experimentella arbetet som utförts inkluderar karakterisering och agglomerering av råmaterialet, samt försök i en kammarugn och en rullugn. Anodslammet karakteriserades med hjälp av kemisk analys, XRD, SEM, och partikelstorleksfördelning. Anodslammet agglomererades till pellets med antingen bentonit och vatten eller bara vatten som bindemedel. Försöken i mindre skala i

kammarugnen undersökte hur flera variabler påverkade blyavdrivningen. Dessa variabler inkluderade typ och mängd av reduktionsmedel som tillsattes, temperatur, och huruvida anodslammet som användes var obehandlat material eller pellets. De mest lovande av dessa resultat användes för vidare försök i rullugnen i något större skala. Variablerna som undersöktes vid rullugnsförsöken var mängd reduktionsmedel som tillsattes, huruvida anodslammet som användes var obehandlat material eller pellets, samt huruvida reduktionsmedlet tillsattes vid start eller under försökets gång.

Alla prover skickades för kemisk analys, och utvalda prover analyserades ytterligare med XRD och SEM.

Slutsatserna som dragits utifrån resultaten av de termodynamiska beräkningarna och det experimentella arbetet är som följande:

• I det obehandlade anodslammet är de primära faserna MnO2, CaSO4, och (Pb,Sr)SO4.

• Anodslammet kan agglomereras till pellets.

• Efter behandling är de primära faserna i materialet MnO och (Ca,Sr)2SiO4. Kvarvarande bly är närvarande som små separata korn.

• Den mest effektiva behandlingsmetoden bör använda följande parametrar:

o Användning av en roterande ugn.

o Anodslam bör tillsättas i form av pellets för att underlätta materialhantering.

o Temperatur mellan 1400-1500 °C.

o Reduktionsmedel tillsatt i omgångar efter att initial rökbildningen avstannat.

o Total tillsats av reduktionsmedel bör vara 10 vikt% av anodslammets vikt.

• Signifikanta viktförluster under behandling.

• Mängden tillsatt reduktionsmedel är den faktor som har störst effekt på avlägsnandet av bly och zink.

• Träkol är en potentiell kandidat för ett förnyelsebart reduktionsmedel, men orsakar ökade viktförluster.

• Det är möjligt att separera bly från mangan med hjälp av pyrometallurgiska metoder, ned till 100 ppm bly.

(5)

iv

Abbreviations

Table 1. Abbreviations of analysis methods.

Abbreviation Analysis method

XRD X-Ray Diffraction

SEM-EDS Scanning Electron Microscope with Energy-Dispersive x-ray Spectroscopy PSD Particle Size Distribution

Table 2. Abbreviations of test and sample names.

Abbreviation Test/Sample

ASC Anode Sludge Characterisation ASCF Anode Sludge Chamber Furnace ASRF Anode Sludge Rotary Furnace

(8)

1

1 Introduction 1.1 Boliden

The origins of the Boliden Group can be traced back to the 1920s, when gold ore was found at the Fågelmyran mine outside Skellefteå. The mining companies responsible for the mine merged in 1931 to form Boliden Gruvaktiebolag, which would go on to become the Boliden group we know today.

Since then, Boliden has grown to include mining and smelting facilities in Norway, Finland and Ireland in addition to those in Sweden. The main products from Boliden’s smelters are copper cathodes, zinc ingots, lead ingots, silver granules, and gold as both ingots and granules.

1.2 Zinc production

Zinc is one of the most common metals to be found in today’s society. As one of the base metals, it has a number of different uses, of which the primary one is corrosion protection, e.g. the

galvanization of steel (Sinclair, 2005). As it is one of the essential elements needed by living

organisms, it also finds uses in medicine, such as sunscreen, ointments and dietary supplements, as well as in fertilisers. Additional areas where zinc is important include battery and rubber production (Emsley, 2001).

The global production of zinc metal in 2016 was 13.7 Mton (ILZSG, 2017). According to the U.S.

Geological Survey, the average annual SHG zinc price at the London Metal Exchange in 2014 was

$2167 per metric ton (U.S.G.S., 2014 Minerals Yearbook: Zinc [Advance Release], 2016). Most

modern zinc production facilities use the electrolytic process. This has three primary stages: roasting, leaching and electrowinning (Sinclair, 2005).

Boliden Kokkola is a zinc producer located in northern Finland. They have an annual production of 315 kilotonnes of pure zinc and various zinc alloys, which makes them the second largest zinc plant in Europe. A simplified schematic of the process is shown in Figure 1 (New Boliden, 2017).

Figure 1. Simplified schematic of Boliden Kokkola’s zinc production process.

The metal concentrate, primarily from Boliden’s own mines, containing approximately 50 % zinc, is first roasted in order to remove sulphur (New Boliden, 2017). Zinc sulfide reacts with injected

oxygen, forming zinc oxide (i). Some of this zinc oxide will react with iron in the concentrate and form zinc ferrite (ii). These reactions are shown below (Sandström, 2016):

ZnS(s) +³

2 O₂(g) → ZnO(s) + SO₂(g) (i)

ZnO(s) + Fe₂O₃→ ZnO ∙ Fe₂O₃ (s) (ii)

Concentrate Roasting Leaching Purification Electrowinning Casting

(9)

2 The resulting ZnO(s) is known as calcine, and the sulphur dioxide created is used to produce sulphuric acid. In the next stage, the calcine is leached with sulphuric acid, producing a zinc sulfate solution while iron is precipitated as jarosite, goethite, paragoethite or hematite depending on the process.

(New Boliden, 2017). In Boliden Kokkola two different leaching methods are used; a neutral leaching step where zinc oxide dissolves, and a hot acid leaching step where the zinc ferrite is dissolved. The calcine is fed to the neutral leaching, and the residue from this stage is fed to the hot acid leaching.

The second leaching line, called direct leaching, leaches ZnS containing concentrate directly by means of oxygen, ferric/ferrous redox pairs, and sulphuric acid. The iron is precipitated as jarosite in the direct leaching stage. Displayed below are the reactions for neutral leaching (iii), hot acid leaching (iv), and jarosite precipitation (v), respectively (Sandström, 2016).

ZnO(s) + H₂SO₄(l) → ZnSO₄(aq) + H₂O(l) (iii)

ZnO ∙ Fe₂O₃ (s) + 4 H₂SO₄(l) → ZnSO₄(aq) + Fe₂(SO₄)₃(aq) + 4 H₂O(l) (iv) 3 Fe₂(SO₄)₃(aq) + 5 ZnO(s) + 2 NH₃(aq) + 7 H₂O(l) → (v)

2 NH₄Fe₃(SO₄)₂(OH)₆(s) + 5 ZnSO₄(aq)

The solution is purified in a three-stage process to remove and recover trace amounts of copper, nickel, cobalt and cadmium. This is achieved by cementation with zinc powder (Sandström, 2016).

Finally, the zinc is recovered by electrowinning as cathodes with a zinc content of 99.995 %. These are cast into slabs or ingots for sale to customers (New Boliden, 2017).

Figure 2 shows a simplified flowsheet of the electrolysis process. The system consists of lead-silver anodes and aluminium cathodes, submerged in a zinc sulfate electrolyte.

Figure 2. Zinc electrolysis process.

At the anode water reacts according to

2 H₂O = O₂+ 4H⁺+ 4e⁻ (vi)

(10)

3 The increased acidity of the electrolyte is utilized by returning it to the leaching stage. At the cathode zinc is precipitated according to the following reaction

Zn²⁺+ 2e⁻= Zn (s) (vii)

However, if the electrolyte solution contains manganese, this will also react at the anode, forming a layer of manganese dioxide according to

Mn²⁺+ 2H₂O = MnO₂(s) + 4H⁺+ 2e⁻ (viii)

This protective layer is an important factor in reducing anode corrosion. When the anode is new, reaction (viii) is rapid, leading to excessive formation of MnO2, creating anode sludge which sinks to the bottom of the tank. In addition, before the protective layer has formed on the anodes, some corrosion will occur, resulting in increased lead ion content in the electrolyte. In order to counteract this and keep the lead content in the cathodes low, strontium carbonate is added (Sinclair, 2005).

Strontium carbonate disassociates and reacts with the electrolyte according to the reaction SrCO₃(aq) + H₂SO₄(l) → SrSO₄(s) + CO₂(g) + H₂O(l) (ix) Lead is then incorporated into the growing strontium sulfate crystal to form a solid solution of strontium sulfate and lead sulfate.

The anode sludge produced at Boliden Kokkola is currently sent to landfill. However, since the sludge is rich in manganese, it could instead become a valuable by-product, provided a sufficient amount of impurities can be removed. The main impurity which needs to be separated from the anode sludge is lead, which in turn could be a valuable by-product in and of itself. An overview of the different applications for manganese and lead is provided below, in chapters 1.3 and 1.4, respectively.

1.3 Manganese and it’s applications

Manganese is an element which is utilised in a number of different areas, from steelmaking to agriculture. It is an essential element for all species, although the exact role of manganese in the human metabolism is still unknown. The daily requirements for a human are low enough that

sufficient manganese is supplied through a normal diet. However, excessive ingestion or inhalation of manganese can lead to adverse health effects, ranging from fatigue and impotence to hallucinations and increased aggressiveness (Emsley, 2001), (Kogel, Trivedi, Barker, & Krukowski, 2006). The global production of manganese in 2014 was 17.8 Mton (U.S.G.S., 2017). The average price of manganese metal in April 2016 was $1984 per metric ton (metalprices.com, 2017).

1.3.1 Metallurgy

The majority of all manganese is consumed by the steelmaking industry (Corathers, 2016). First ferro- or silicomanganese is produced, which is then used as a deoxidizing or desulphurization agent, or as alloying material. Several criteria must be fulfilled by the raw material for ferromanganese

production. These include 47% manganese content, with a Mn:Fe ratio of at least 8:1, low contents of alumina and phosphorous, and only a few ppm of other trace metals. During ferromanganese production, the manganese losses to slag are relatively high, making it possible to utilise the slag as a feed material to silicomanganese production (Kogel, Trivedi, Barker, & Krukowski, 2006).

(11)

4 The reactions for deoxidation (x) and desulfurization (xi) using ferromanganese are as follows

(excluding non-Mn content of ferromanganese):

FeO + Mn → Fe + MnO (slag or gas) (x)

FeS + Mn → Fe + MnS (slag) (xi)

Manganese as an alloying element can confer several useful traits on the steel being produced. In some nickel steels, manganese is added as carbide stabiliser, forming Mn3C, in order to increase the carbon content. Certain austenitic stainless-steel types also replace nickel with manganese in order to reduce costs. Recently manganese has also been added to ordinary carbon steels, as a way to improve the impact resistance and ductility, although this forces a reduction in carbon content. The creation of high-manganese steel, also known as Hadfield steel after its inventor, took place in 1882.

Steel using the same composition and heat-treatment is still being produced today. Hadfield steel is relatively soft, but extremely tough and resistant to shock and mechanical wear. The reason for these properties can be found in the structure of the steel; as the component is subjected to mechanical disturbances, the surface hardens, resulting in a very hard shell around a softer core. While the hardened surface gives good wear resistance, the soft core improves shock resistance. Despite being difficult to machine into shape, there is currently no known substitute for Hadfield steel in its primary areas of use. These include e.g. railroad crossings, crushing machinery and other applications where very high toughness and wear resistance is required (Higgins, 1993), (Emsley, 2001).

Smaller amounts of manganese are also alloyed to several other metals, such as copper, aluminium and magnesium. High-tensile copper obtains its increased strength from added manganese and a few other metals. When added to aluminium or magnesium, manganese’s primary function is to increase the corrosion resistance, but it also acts as a grain refiner (Higgins, 1993), (Emsley, 2001).

1.3.2 Batteries

One of the major non-metallurgical uses for manganese is as a component in dry-cell batteries. The first dry-cell battery was invented in the 1880s by Carl Gassner (American Chemical Society, 2005).

Also known as a zinc-carbon battery, modern batteries of this type consist of a zinc container, carbon electrode, a cathode mixture and electrolyte.

The carbon electrode runs through the centre of the battery, connecting to the positive metal end cap. The zinc container acts as anode. The cathode mixture consists of manganese dioxide and carbon in order to increase the conductivity. The electrolyte is a solution of either ammonium chloride or zinc chloride (Davidson, 2015).

Another common type of battery which contains manganese is the alkaline battery. This has similar construction as the zinc-carbon battery, but reverses the positions of anode and cathode, so that the manganese cathode mixture is closer to the shell and a zinc powder-gel anode is in the centre. The outer zinc container is replaced by one made of steel, and the carbon electrode is removed. The most significant difference, and the one for which the battery is named, is the replacement of the acidic electrolyte with an alkaline one. Alkaline batteries have increased capacity and can operate under more demanding conditions (United States Patentnr 2960558, 1960).

(12)

5 1.3.3 Other applications

As stated earlier, manganese is an essential element for both flora and fauna. In some nutrient- starved or high pH soils, crops and plants may suffer from manganese deficiency. This can affect the photosynthesis, lead to leaf necrosis and renders the plant more vulnerable to fungal diseases (Birkelund-Schmidt, Jensen, & Husted, 2016). For this reason, manganese is sometimes added to fertilizers to improve the content in the soil.

Manganese can also be used in chemical analysis, or for removal of organic impurities from water or waste gases (Emsley, 2001).

1.4 Lead and it’s applications

Lead is a heavy metal that has been used by humans for thousands of years. It has had many applications throughout history, from medicine and plumbing to bullets and radiation shielding.

However, lead is also a well-known health hazard which accumulates in the body over time, in the form of lead phosphate in the bones. High levels of lead in the body inhibit the production of haemoglobin, thus causing a number of unpleasant symptoms such as infertility and anaemia (Emsley, 2001).

Due to the health hazards of lead, it has been replaced in many applications. The reduced demand has meant that the ore-based production has not increased significantly; rather it is the recycling of lead scrap that has caused increased consumption (Björkman, 1998). Almost half of the worlds lead production is from secondary materials. This makes lead one of the worlds most recycled metals. A large portion of the recycled materials is in the form of lead-acid batteries from cars and other vehicles. Over 85% of lead-acid batteries are recycled in Western Europe, as shown in Figure 3 (Jolly

& Rhin, 1994).

Figure 3. Recycling data for Western Europe.

1.4.1 Lead-acid batteries

Production of lead-acid batteries is the largest modern application for lead (Björkman, 1998). The lead-acid battery was originally invented in 1859 by Gaston Planté. Today, lead-acid batteries are used in a wide array of applications, among them automotive batteries and back-up power for hospitals. The modern automotive battery uses sulphuric acid as an electrolyte, and has a sealed casing consisting primarily of polypropylene. The lead forms the electrodes of the battery, in the shape of a lead grid with rectangular holes filled by a paste of lead oxide (Dell & Rand, 2001) .

Recycling in Western Europe (%)

(13)

6 1.4.2 Other applications

Two large applications from the 20^th century were as an additive in gasoline to raise the octane level, and in pipes for plumbing. Both of these have been phased out due to the health hazards of lead. In modern times, several different applications for lead exist, such as bullets, weights, and radiation shielding for hospital personnel operating X-ray equipment.

1.5 Separation of manganese and lead

In order to obtain viable by-products from the anode sludge, manganese and lead first have to be separated, preferably by removing the maximum amount of lead possible while minimising the loss of manganese. There have been previous investigations in regards to treating manganese slimes from zinc production, however, these have been focused on using hydrometallurgical methods.

(Huang, Bi, Mu, & Zhang, 2014) developed a purification treatment that enables the anode sludge to be used as raw material for battery manganese production. This involved repeatedly washing the material with acid and water to remove soluble materials, followed by grinding. The material then went to roasting at 700 °C and finally leaching with 2 mol/L acetic acid, leaving mostly Mn2O3 in the leaching residue and removing most of the lead, down to 0.1 %.

(Ayala & Fernández, 2013) took a different approach, using sulphuric acid to treat the material at high temperature (>400°C), and then water to leach the manganese itself instead of the impurities.

The aqueous solution was purified using precipitation with sodium sulfide, in order to remove zinc, copper, nickel and cobalt. In the next stage, electrolytic manganese of high purity (>99.9 %) was produced from the manganese sulfate solution. Lead could potentially be recovered from the remaining solid residue.

Another potential option for the separation of lead and manganese would be by vaporising one of the compounds while leaving the other intact. The temperatures, in Kelvin, required to reach certain partial pressures for lead, manganese and strontium are displayed in Table 3. As can be clearly seen, Manganese requires the highest temperatures and will thus be the last to evaporate. Pure zinc should be vaporised before the other two metals.

Table 3. Temperatures required to reach specific partial pressures of Mn, Pb and Zn. (Kaye & Laby, 2005)

Pressure (Pa) Manganese Lead Zinc

10 1980 K 1670 K 991 K

20 2080 K 1760 K 1040 K

50 2240 K 1905 K 1120 K

101325 2390 K 2030 K 1185 K

Some preliminary tests were performed by Swerea MEFOS prior to the start of this project. These took place in a small chamber furnace, and the results indicated that a pyrometallurgical treatment could be a viable option for the separation of lead from manganese-rich anode sludge. This formed the starting point for this master thesis work.

(14)

7

2 Aim

The aim of this thesis is to examine whether it is possible to separate lead from an electrolytic sludge rich in manganese by using pyrometallurgical treatment, as well as determining optimum process parameters for such a treatment. The final goal is to obtain a profitable by-product.

The specific goal of the pyrometallurgical treatment, consisting of high temperature reduction using carbon, is to obtain the lowest possible distribution of lead to the solid phase, while keeping the distribution of manganese to the solid phase as high as possible.

(15)

8

3 Materials and Experimental Procedure 3.1 Materials

A total of 160 kg of anode sludge was supplied by Boliden Kokkola for use in this thesis work. The material was supplied wet and divided between 8 separate barrels. The material was known to primarily consist of manganese, with significant amounts of lead, sulfur, and strontium, along with several trace elements.

The carbon added during some tests was in one of 3 forms; either as coke gravel with a size of 0-6 mm, pulverized coke gravel, or pulverised charcoal.

3.2 Methods

The different methods of analysis used during this master thesis work are described below.

3.3.1 Drying

In order to determine the moisture content of the supplied material, the accumulated water layer in the barrel of material was first decanted and weighed. The remaining material was excavated in layers, each of which was weighed prior to drying for 24 hours at 105°C. An illustration of the different layers excavated from the barrel is shown in Figure 4.

Figure 4. Excavated layers for drying.

After drying the material was weighed once again and collected in a bucket. In addition, the weight of the empty drying trays was recorded before and after the drying process. Finally, a mass balance was calculated in order to determine the moisture content.

3.2.2 Characterisation

Samples were sent to Boliden Kokkola for chemical analysis using an ICP mass spectrometer, where the solid sample material was first dissolved and the analysed.

Samples were sent to Luleå University of Technology for determination of density using a pycnometer and particle size distribution (PSD) using laser diffraction (Malvern 2000).

3.2.3 XRD

The sample was filled in a back-filling sample holder and placed in the XRD. PANAlytical Empyrean equipment with a copper tube, Pixel3D detector and a pyrolytic graphite monochromator was used to collect data. The phases present in the sample were identified using the software PANAlytical High Score Plus and Crystallography Open Database (COD).

(16)

9 3.2.4 SEM-EDS

The sample was cast in an epoxy mould, which was polished to obtain a smooth surface. The sample was then carbon coated to prevent charging of the surface during the SEM analysis. The phases in the sample were analysed in a Zeiss Merlin SEM at Luleå University of Technology equipped with an Oxford Xmax 50mm EDS detector. This was done by doing point analysis on different sites on each sample.

3.2.5 Agglomeration

3 kg of dried anode sludge, in some cases together with 45g of bentonite to act as binder, was added to an Eirich mixer. After starting the mixer, water was added during as appropriate until granules had formed, as determined by ocular inspection.

Moisture was measured through a quick test immediately after the completed batch, and by drying of approximately 1 kg of material over 24 hours at 105°C. The dried material was split down to approximately 200g and sieved, using the sieves shown in Table 4.

Table 4. Sieve sizes used for PSD.

Sieve size (µm) 7100 5600 4000 2800 2000 1400 1000 710 500

3.2.6 Chamber furnace testing

Small scale tests were performed in a chamber furnace, with an inert nitrogen atmosphere. Figure 5 shows the inside of the chamber furnace, containing a control thermocouple high up in the furnace chamber and a “charge temperature” thermocouple that is located close to the sample.

Figure 5. Interior of the chamber furnace.

(17)

10 Sample holders consisted of sintered magnesia bricks which had been hollowed to form a crucible, as shown in Figure 6 below.

Figure 6. Magnesia brick sample holder.

Dried anode sludge and carbon was added to sample holders and placed in the furnace, which was then filled with an inert nitrogen atmosphere and heated to target temperature with a programmed rate of 20 °C/s. The total time for one treatment cycle, excluding heating and cooling, was two hours for all tests. During each test two samples were treated simultaneously, designated A and B.

After two test cycles, one full cycle was run while the furnace was empty. For this cycle, the furnace had an oxidizing atmosphere, in order to restore the protective oxide layer of the Kanthal heating elements.

3.2.7 Rotary furnace testing

Further tests at a slightly larger scale were performed in a rotary furnace lined with magnesia bricks, with heat supplied by a burner in one end. Figure 7 shows the rotary furnace during operation.

Figure 7. Rotary furnace during operation.

(18)

11 For each experiment, the furnace rotation was initialised and the furnace heated to target

temperature. 10 kg of sample material, plus the appropriate amount of reduction agent, was then charged to the furnace. Sampling was performed with 20-minute intervals. After 2 hours, the furnace was tapped and prepared for the next test. All tests were conducted with a rolling time of 2 hours, and the target temperature was 1400 °C. Carbon was added in the form of coke gravel with a size of 0-6 mm.

3.3 Execution

3.3.1 Drying

Two barrels of anode sludge were emptied and dried according to the method described in chapter 3.2.1. For the remaining barrels, the drying procedure followed the same method, but only the final total weights of dried material were recorded.

3.3.2 Characterisation of anode sludge

Small samples were taken from the first three barrels delivered for use in the characterization of the raw material. These samples were mixed together, dried and ground using mortar and pestle. This combined sample material was then used during the characterization work. Chemical composition, PSD, and density were obtained by sending samples for analysis, as described in chapter 3.2.2. Two samples each were prepared for XRD and SEM analysis, as described in chapters 3.2.3 and 3.2.4, respectively.

3.3.3 Agglomeration

Four initial batches of material were granulated, exactly as described in chapter 3.2.5, with addition of bentonite. Four additional batches were granulated using the same method, except no bentonite was added.

3.3.4 Chamber furnace testing

Five tests were run in the chamber furnace, using the method described in chapter 3.2.6. The experimental plan is shown in Table 5. The different variables used were carbon addition (10, 8, 6, or 4%), temperature (1400 or 1450°C), and material type (powder or agglomerate, additionally one test was run using organic carbon material). The carbon added was in the form of pulverised coke gravel, except in Test 5A and 5B, which instead used pulverised charcoal.

Table 5. Experimental plan for chamber furnace tests.

Test Carbon (wt%) Temperature (°C) Time (h) Material

1A 10 1400 2 Powder

1B 10 1400 2 Agglomerate

2A 8 1400 2 Powder

2B 8 1400 2 Agglomerate

3A 6 1400 2 Powder

3B 4 1400 2 Powder

4A 10 1450 2 Powder

4B 8 1450 2 Powder

5A 10 1400 2 Powder, Charcoal

5B 8 1400 2 Powder, Charcoal

(19)

12 Test 1A was used as the baseline for comparison with the other tests. This used the conditions which had achieved the best results in the preliminary tests performed by MEFOS. Tests 2A, 3A and 3B compared whether sufficient lead removal could still be achieved with lower amounts of carbon added. In tests 1B and 2B, agglomerated material was used in order to see if this yielded any

advantages. Tests 4A and 4B served to determine the effects of higher temperature on lead removal and material properties. Test 5A and 5B used the same conditions as tests 1A and 2A, respectively, but replaced the coke gravel with charcoal. The purpose of this was to compare whether the same results could be obtained while using a renewable source of carbon.

Samples from all tests were sent for chemical analysis according to chapter 3.2.2, and based upon the returned results 6 samples were further analysed using XRD and SEM, as described in chapters 3.2.3 and 3.2.4, respectively. Table 6 displays the selected samples and the reason for their selection.

Table 6. Sample selection for XRD and SEM.

Sample Reason for selection

ASCF1A Good Pb removal, used as baseline for comparison

ASCF1B Comparison of agglomerated material, lower Pb removal but less weight loss ASCF2A Comparison of lower carbon addition, lower Pb removal but less weight loss ASCF3A Comparison of lower carbon addition, material melted completely

ASCF4A Comparison of higher temperature treatment results ASCF5B Comparison of charcoal and coke, high mass loss

3.3.5 Rotary furnace testing

5 tests were conducted in the rotary furnace, as described in chapter 3.2.7. The conditions used for these tests were based on the results from the chamber furnace tests. Table 7 displays the

experimental plan. The variables used were material type (powder, agglomerate or agglomerate with bentonite), amount of carbon added (10 or 8 %) and timing of carbon addition (at start together with anode sludge, or added during the test).

Table 7. Experimental plan for rotary furnace tests.

Test Material Carbon (wt%)

Timing of carbon addition ASRF1 Powder 10 + 2 At start

ASRF2 Powder 8 + 2 At start ASRF3 Agglomerate 10 + 2 At start

ASRF4 Powder 10 Incrementally during test ASRF5 Agglomerate w.

Bentonite 10 Incrementally during test

Test 1 served as the baseline test, using the best result from the chamber furnace test. Test 2 examined whether the same lead removal could be achieved with lower carbon addition, something that was not possible in the chamber furnace. Test 3 served to compare agglomerated material with untreated anode sludge. For these tests, 10 wt% carbon was added together with the sample material at the start of the experiment. An additional 2 wt% of carbon was added after the material

(20)

13 had stopped emitting smoke, in an attempt to ensure complete removal of all lead. At the start of Test 1 and 2, a metal plate was placed at the gas outlet in an attempt to condensate and sample the lead-containing gas phase leaving the furnace.

Test 4 had no carbon added together with anode sludge. Instead, carbon was added in two batches of 5 % of the weight of anode sludge used, adding up to a total of 10 wt%. The additions took place as soon as the material had stopped emitting smoke. This had the goal of determining whether it is more effective to vaporize some lead prior to feeding carbon into the furnace.

Finally, Test 5 was not part of the original experimental plan, but was added since there were complications during Test 4 and extra time was available. Similarly to Test 4, Test 5 had no carbon added at the start of the experiment, but rather in two separate batches after the material stopped emitting smoke. The total carbon addition was 10 wt%. The sample material used in Test 5 was 8 kg of agglomerated anode sludge containing bentonite as a binder.

After all tests were complete, the furnace was opened and the material which had stuck to the walls was removed using hammer and chisel. Figure 8 shows the inside of the furnace prior to removal of stuck material. The image has a slightly blue tint due to the camera flash used.

Figure 8. Interior of rotary furnace after testing.

Samples from all tests were sent for chemical analysis as described in chapter 3.2.2, and based upon the mass balance 5 samples were further analysed using XRD and SEM according to chapters 3.2.3 and 3.2.4. Table 8 displays the selected samples and the reason for their selection.

Table 8. Sample selection for XRD and SEM.

Sample Reason for selection

ASRF1D Good results during ASCF testing, used as baseline for comparison ASRF1S Comparison of material before and after additional carbon was added ASRF2D Comparison of lower carbon addition

ASRF3D Comparison of agglomerated material

ASRF5C Comparison of agglomerated material with no carbon addition

(21)

14

4 Thermodynamic calculations

To help confirm the experimental results, the theoretical behaviour of the anode sludge at increasing temperatures and with different amounts of carbon added was simulated using the software

FactSage 7.0. The temperature range simulated was 500-1450°C in steps of 50 °C, and the pressure was set to 1 atm. The databases used were ideal gas phases, pure solids, FToxid-SLAGA, FToxid- SPINB, and FToxid-MeO_A. The composition used was the average composition of the anode sludge, and the carbon addition ranged from 4-10 wt%. The results data was divided into 4 groups:

Manganese, Lead, Zinc and a final shared group for Calcium and Strontium. Diagrams for all thermodynamic calculations can be found in Appendix A.

4.1 Manganese

Larger carbon addition (8-10 wt%) leads to large amount of solid MnO phase, but there is also a molten slag phase present, which consists of MnO, MnS and Mn2O3. At 6 wt% carbon addition, increasing temperature causes most of the MnO(s) to morph into a Mn3O4 spinel phase, as shown in Figure 9. The slag phase is still present.

Figure 9. Behaviour of Mn at 500-1450 °C using 6 wt% carbon.

The primary manganese phase at low carbon addition (4 wt%) and temperatures below 900°C is Mn2O3, with some MnSO4 also present. Above 900 ° almost all manganese changes into a Mn3O4 spinel phase. A small amount is also present as a slag phase.

4.2 Lead

The behaviour of lead can be separated into 2 distinct variants. At 4 and 6 wt% carbon added, the lead is first present as PbSO4, and then changes into other solid lead sulfate phases before forming a PbO melt. Finally, at very high temperatures the lead evaporates as PbO gas. When using larger carbon additions (8 or 10 wt% added), the lead is initially in the form of a solid PbS phase, which at approximately 675 °C transitions into a melt primarily composed of PbS, but with some PbO present

(22)

15 as well. Further increasing the temperature causes the lead to escape as a gas mixture of Pb(g) and PbS(g). Figure 10 shows a diagram of the material’s behaviour at 8 wt% carbon addition.

Figure 10. Behaviour of Pb at 500-1450 °C using 8 wt% carbon.

4.3 Zinc

At 4 wt% carbon addition, zinc transitions from Zn2SiO4 to a ZnO melt at 1100C°, and then finally forms a ZnO gas phase. Increasing the carbon addition to 6 wt% causes all zinc to already have formed ZnO at 500 °C. Some of the zinc oxide forms a ZnO melt above 900 °C. For 8 and 10 wt%

carbon addition, at 500 °C there is a ZnS solid phase that is in the process of being converted to solid ZnO. This continues until approximately 675 °C, where remaining ZnS(s) melts rapidly while the ZnO remains as a solid phase. At 10 wt % carbon addition a portion of the ZnO also melts, as shown in Figure 11. Above 1100°C zinc is driven off in the form of Zn gas.

Figure 11. Behaviour of Zn at 500-1450 °C using 10 wt% carbon.

(23)

16

4.4 Calcium and Strontium

Calcium and strontium were examined together due to their similar properties. Both elements initially exist as sulfates, except at 10 wt% carbon addition where carbonates are dominant instead.

As the temperature increases, calcium transitions through Ca2SiO4 to finally become part solid CaO and part molten calcium slag phase. Strontium changes from sulfate to Sr2SiO4, except at 4 wt%

carbon addition, where it only changes to a different sulfate phase. The higher the carbon addition, the lower the temperature where these changes take place. Figure 12 shows the behaviour of the material at 8 wt% carbon addition.

Figure 12. Behaviour of Ca and Sr at 500-1450 °C using 8 wt% carbon.

(24)

17

5 Results and Discussion

5.1 Characterisation of anode sludge

The typical chemical composition of the raw material is displayed in Table 9. The dominant element by far was manganese, with a few percent each of the remaining main components. Additionally, several minor elements were found in the material.

Table 9. Composition of anode sludge.

Main

components Mn

Pb Sr Zn S Ca

The two samples analysed showed very similar diffractograms. Figure 13 shows the diffractogram of sample ASC1.

Figure 13. Diffractogram of sample ASC1.

The main phases found in the XRD analysis are shown in table 10. Both samples analysed were taken from the same material. All samples contained manganese dioxide along with various sulfate

compounds. The lead-strontium sulfide detected is most likely also a sulfate, rather than a sulfide, as the process conditions favour the generation of sulfates. Zinc sulfide was only found in the second sample. Full XRD analysis data for all samples is displayed in Appendix C.

(25)

18

Table 10.Phases found during XRD analysis of anode sludge.

Phase Sample ASC1

Sample ASC2

MnO2 X X

(PbSr)S X X

SrSO4 X X

CaSO4 X X

ZnS X

An example image of the structure of the anode sludge is displayed in Figure 14. Three main types of structure existed in the samples. The first of these were light grey particles (green arrows), these contained mainly manganese and oxygen, in such proportions that they likely consisted of

manganese dioxide with a few impurities. The second were pure white particles (red arrows), these consisted of different lead-containing phases, such as PbO and PbSO4. The third were dark grey areas (teal arrows), forming string-like shapes between the other phases. This was primarily composed of calcium, sulfur and oxygen, probably forming calcium sulfate, but small amounts of manganese oxides were also present in some places. Strontium was mainly found in lead and calcium sulfate, but occasionally together with manganese phases.

Figure 14. SEM image of anode sludge structure.

Full SEM analysis data for all samples is displayed in Appendix D.

Figure 15 displays the particle size distribution of the raw material, as determined by laser diffraction. The p80 of the untreated sludge is 55 µm.

(26)

19

Figure 15. PSD of the dried anode sludge.

5.2 Agglomeration

Figure 16 shows the cumulative undersize distribution for all 8 batches of pellets. Batch 1-4 were made using bentonite as a binder in addition to water, while Batch 5-8 were made using only water.

The majority of the batches had a p80 between 2 and 3 mm. Batch 1 had smaller pellets with a p80 close to 1 mm, while Batch 3 had larger pellets with a p80 of approximately 3,5 mm. In addition, Batch 3 had a larger variation in grain size compared to the other batches, something that was also observed visually during the agglomeration process.

Figure 16. PSD of agglomerated anode sludge.

0,00 20,00 40,00 60,00 80,00 100,00 120,00

100 1000 10000

Cumulative undersize (wt%)

Particle size (µm)

PSD Agglomerated anode sludge

Batch 1 Batch 2 Batch 3 Batch 4 Batch 5 Batch 6 Batch 7 Batch 8

(27)

20 Table 11 shows a summary of the amounts of material input and received out from the

agglomeration process, as well as the moisture content. For Batches 4, 6, and 8, the total mass of material out from the process was higher than the material input for that batch.

Table 11. Mass balance and moisture for agglomeration.

Batch 1 2 3 4 5 6 7 8

Anode sludge (g) 3 000 3 000 3 000 3 000 3 000 3 000 3 000 3 000

Bentonite (g) 45 45 45 45 0 0 0 0

Water via system (g) 500 450 500 500 550 600 575 575

Spray water (g) 65 128 86 107 279 53 100 105

TOTAL IN (g) 3 610 3 623 3 631 3 652 3 829 3 653 3 675 3 680

Granules out, moisture testing (g)

975 950 1 067 1 030 1 024 982 971 962

Granules out (g) 2 360 2 645 2 505 2 610 2 350 2 755 2 450 2 255

Samples (g) 10 10 10 10 10 10 10 10

TOTAL OUT (g) 3 345 3 605 3 582 3 700 3 384 3 747 3 431 4 051

Balance (%) 7.3 0.5 1.4 -1.3 11.6 -2.6 6.6 -10.1

Moisture content, quick test (%, wet base)

14.2 15.1 15.9 15.7 18.8 17.9 17.2 17.3

Moisture content, 1 kg (%, wet base)

14.4 15.3 16.1 15.7 19.3 19.6 17.3 17.5

Table 12 shows the measured density of the pellets along with the calculated porosity and the porosity required to fit all water into pores.

Table 12. Density and porosity of pellets.

Batch 1 2 3 4 5 6 7 8

Density wet

granules (gr/cm³) 2.9 2.8 2.9 2.9 2.6 2.7 2.8 2.7 Moisture content

Quick sample (%, wet base)

12.8 13.9 15.1 15.8 19.1 18.2 16.7 16.7

Material density, calc. (gr/cm

3

) 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0

Total porosity (%) 16.4 19.7 20.0 20.2 29.5 29.1 24.2 25.6 Porosity, water

(%) 42.4 43.7 46.2 44.8 50.8 52.2 47.8 47.6

(28)

21

5.3 Chamber furnace testing

A comparison of the samples’ visual appearance before and after the chamber furnace tests 2 and 3 are shown in Figures 17 and 18. In each image, the sample in the left and right sample holders are designated A and B, respectively. The corresponding images for the remaining chamber furnace tests can be found in Appendix B.

All samples using dried anode sludge showed sintering, as displayed by sample ASCF2A. The samples of granulated anode sludge showed both sintering as well as a degree of discolouration, as shown by sample ASCF2B.

Both samples in test no. 3 diverged significantly in their behaviour, as they melted; ASCF3A forming a smooth smelt and ASCF3B forming a chunkier smelt. These samples could not be removed from the sample holders without breaking them.

Figure 17. Material before and after tests ASCF2A and ASCF2B.

Figure 18. Material before and after tests ASCF3A and ASCF3B.

Before After

3A 3B 3A 3B

2A 2B 2A 2B

(29)

22 In addition, after all tests discolouration of the sample holders was observed. This was especially pronounced after test 5. Figures 19 and 20 show the sample holder after test 5B, as viewed from the side and a cross-section, respectively. The cross-section shows that a large portion of the magnesia brick had been discoloured, as outlined in red. Cross-section images of remaining sample holders are displayed in Appendix B. The discolouration appeared less prominent in the samples using smaller carbon addition.

Figure 19. Discolouration of sample holder ASCF5B, side view.

Figure 20. Discolouration of sample holder ASCF5B, cross-section.

(30)

23 The measured weights of the sample holders, anode sludge, carbon, and the total before and after testing are displayed in Table 13.

Table 13. Weights before and after ASCF testing.

Sample Sample holder before (g)

Anode sludge (g)

Carbon (g)

Total before (g)

Total after (g)

Sample holder after

(g)

ASCF1A 1175 201.4 20.1 1396 1304 1207

ASCF1B 1759 200.0 20.0 1979 1891 1790

ASCF2A 1680 200.1 16.1 1896 1803 1688

ASCF2B 1725 200.0 16.1 1941 1860 1753

ASCF3A 1210 200.1 12.1 1422 1339 1339

ASCF3B 1264 200.0 8.10 1472 1403 1403

ASCF4A 1771 200.0 20.0 1990 1897 1810

ASCF4B 1839 200.0 16.0 2055 1962 1862

ASCF5A 1918 200.0 26.6 2146 2047 1985

ASCF5B 1983 200.8 21.3 2206 2110 2031

The calculated weights of the sample before and after testing, along with the weight loss, are shown in table 14. The final sample weight of ASCF3A and ASCF3B was calculated using the weights of the sample holder before and after the test, since the entire sample melted and thus stuck to the sample holder. Thus, the mass balance for these tests could potentially be somewhat inaccurate.

Table 14. Mass balance for ASCF testing.

Sample

Initial sample weight (g)

Final sample weight (g)

Difference (Initial-Final)

(g)

Weight loss (%)

ASCF1A 221 97 124 56

ASCF1B 220 101 119 54

ASCF2A 216 115 101 47

ASCF2B 216 107 109 50

ASCF3A 212 129 83 39

ASCF3B 208 139 69 33

ASCF4A 219 87 132 60

ASCF4B 216 100 116 54

ASCF5A 228 62 166 73

ASCF5B 223 79 144 65

(31)

24 Table 15 shows the chemical analysis of each sample after completed testing. The lowest lead

content was found in samples ASCF1A, ASCF4A, ASCF5A and ASCF5B.

Table 15. Chemical analysis of anode sludge after ASCF testing.

Sample Mn (%)

Zn (%)

Pb (%) ASCF1A 58 0.04 0.01 ASCF1B 67 0.28 0.44 ASCF2A 56 0.18 0.03 ASCF2B 63 0.56 0.48 ASCF3A 39 0.80 0.37 ASCF3B 45 1.1 0.66 ASCF4A 59 0.01 0.01 ASCF4B 56 0.19 0.02 ASCF5A 60 0.02 0.01 ASCF5B 58 0.03 0.01

Figure 21 displays the distribution of Mn, Zn, and Pb to the solid phase for all tests. The most desirable outcome is a high distribution of manganese to the solid phase, but a very low distribution of lead, zinc and other elements to the solid phase. In the following diagrams (Figure 21-25),

Distribution always refers to the distribution to the solid phase, calculated using the formula below:

Distribution to solid phase = 100 ∗ (Metal content in sample)∗(Sample weight after test) (Metal content in raw material)∗(Sample weight before test)

Figure 21. Distribution to solid phase for Mn, Zn and Pb.

0 10 20 30 40 50 60 70 80

Solids distribution

Distribution to solid phase (%)

Mn Zn Pb

(32)

25 The effect of varying carbon addition is compared in Figure 22. There was a clear trend in regards to the distribution of both Zn and Pb to decrease with higher carbon addition. However, no such distinct pattern existed for Mn.

Figure 22. Distribution to solid phase for varying carbon addition.

The difference in distribution for Mn, Zn and Pb when using dried sludge or pellets is displayed in Figure 23. All elements had higher solids distribution when using pellets.

Figure 23. Distribution to solid phase for dried sludge and pellets.

0 10 20 30 40 50 60 70 80

4 6 8 10

Solids distribution (%)

Carbon addition (wt%)

Distribution vs Carbon addition

Mn Zn Pb

0 10 20 30 40 50 60 70 80

Sludge, 10 wt% C

Pellets, 10 wt% C

Sludge, 8 wt% C

Pellets, 8 wt% C

Distribution vs Material

Mn Zn Pb

(33)

26 Figure 24 compares the effect of temperature on the distribution for Mn, Zn and Pb with both 10 wt% and 8 wt% of added carbon. The distribution of Pb is plotted on the secondary axis. The distribution was lower for all elements at 1450 than at 1400 °C, but numerically speaking the difference was small.

Figure 24. Distribution to solid phase for different temperatures, with Pb distribution plotted on secondary axis.

Figure 25 shows the distribution for Mn, Zn and Pb when using coke and charcoal with either 10 or 8 wt% added. The distribution of Pb is plotted on the secondary axis. The distribution for all elements was lower when using charcoal.

Figure 25. Distribution to solid phase using coke or charcoal, with Pb distribution plotted on secondary axis.

0 0,05 0,1 0,15 0,2 0,25 0,3

0 10 20 30 40 50 60 70 80

1400°C, 10 wt% C

1450°C, 10 wt% C

1400°C, 8 wt% C

1450°C, 8 wt% C

Pb distribution(%)

Distribution vs Temperature

Mn Zn Pb

0 0,05 0,1 0,15 0,2 0,25 0,3

0 10 20 30 40 50 60 70 80

Coke, 10% Charcoal, 10%

Coke, 8% Charcoal, 8%

Pb distribution(%)

Distribution vs Reduction Agent

Mn Zn Pb

(34)

27 5.3.1 XRD analysis

Figure 26 shows the diffractogram of sample ASCF1A, with the most significant peaks labelled. Most of the other samples had very similar diffractograms. The primary peaks in all samples corresponded to different forms of manganese oxide. MnO was dominant in all analysed samples except ASCF3A, where Mn2O3 was the most common phase.

Figure 26. Diffractogram of sample ASCF1A.

Table 16 shows the different phases found in the samples from chamber furnace testing.

Table 16. Phases found during XRD analysis after ASCF testing.

Phase ASCF1A ASCF1B ASCF2A ASCF3A ASCF4A ASCF5B

MnO X X X X X X

Mn2O3 X X X X X X

Mn3O4 X X X X X X

Ca2SiO4 X X X

MnS X

CaMn2O4 X X

For full XRD analysis data, please check Appendix A.

5.3.2 SEM analysis

Most of the analysed samples, specifically ASCF 01A, 01B, 04A and 05B, showed similar composition, consisting of three primary structures. Figure 27 shows an example from sample ASCF1A. Most of the sample was made up of rounded grey grains (green arrow), consisting of mainly manganese and oxygen with some Ca contamination. The ratio between manganese and oxygen is such that this phase was most likely MnO. The second phase was smaller dark grey grains (blue arrow), which consisted of calcium, strontium, silicon and oxygen, in such proportions that the most probable composition was (Ca,Sr)2SiO4, i.e. calcium silicate with some strontium substitution. These phases would also agree with the results of the XRD analysis. The final major phase found was lighter grey areas between the other grains (yellow arrow). This appeared to be a non-crystalline residue left over during the formation of the other faces. The composition of this residual phase varied, but commonly contained varying amounts of Mn, Ca, Sr, S and O. In addition to the primary phases,

(35)

28 separate grains of metals or metal oxides/sulfides of Pb, Fe, Cu or Zn were found in some samples.

Figure 27 contains 2 such grains, one of copper (orange arrow) and one of a lead containing phase (red arrow).

Figure 27. SEM image of anode sludge structure after test ASCF1A.

Sample ASCF2A, which had lower carbon addition, displayed a different composition. Figure 28 shows an example of the structures present. Here, the main manganese phase appeared to be Mn2O3

with some Ca contamination forming grey grains (yellow arrow), while MnO was present as a lighter grey residue (green arrow) between the grains. The calcium-strontium silicate phase formed larger clumps (blue arrow) instead of dispersed smaller particles, as in sample ASCF1A.

(36)

29 The most divergent sample was ASCF3A. Figure 29 displays an example of the structure. The grain boundaries were not as clear as in the other samples. What appeared to be Mn2O3 with varying Ca contamination was present as slightly darker areas (yellow arrow) inside grains of what was most likely MnO (green arrow). The calcium-strontium silicate phase was present as distinct lighter grey particles (blue arrow), much fewer of which were found in this sample.

Further examples of the structrure of sample ASCF3A are shown in Figure 30. In this area, only a single occurrence of MnO could be found (green arrow). A phase containing large amounts of Fe and Cr could also be seen (orange arrow). The rest of the material appeared to be Mn2O3 with varying Ca contamination, which here displayed a tendency of forming large crystalline structures (yellow arrows).

Figure 30. SEM image of large crystalline structures after test ASCF3A.

(37)

30 Finally, an anomalous particle was found in sample ASCF1B. Figure 31 shows an overview of said particle. The composition of this particle clearly deviated from the remainder of this sample, the structure of which can be seen in the top left of the image. The anomalous particle had a much denser structure, with an increased amount of lead phase inclusions. This may indicate that the composition of the samples after, and possibly before, treatment was not homogenous. This in turn could have affected the different analyses performed on the material, e.g. if a sample sent for chemical analysis contained such a particle, it may show a higher lead content than expected.

Figure 31. SEM image of anomalous particle after test ASCF1B.

Figure 32 shows a magnified image of this structure. The dark grey “needle” structures (yellow arrow) appeared to consist of Mn2O3 with varying Ca contamination. The white areas (red arrows) were phases with high lead content. The remaining material (green arrow) was most likely MnO.

Figure 32. Zoomed in image of anomalous particle after test ASCF1B.

Full SEM analysis data for all samples is displayed in Appendix D.

Pretreatment Methods for Manganese Containing Anode Sludge