The main waste products of ironmaking at Tata Steel IJmuiden are CO2¬(g) and a solid waste stream consisting of mainly carbon and iron-bearing compounds. This solid waste is generally called Zn rich filter cake which contains too much Zn to be reintroduced into the iron making process. Besides Zn, Pb and Cd are also present. These filter cakes originate from the flue dusts that are expelled from the top of the blast furnace. Currently the solid waste is separated based on size distribution which allows the recycling of most of the materials and concentrates the heavy metals into the Zn rich filter cake. Previous research has shown that these cakes are resistant to regular leaching methods for selectively dissolving the contained heavy metals. This is due to their sulfide form. The methods that do work (such as acid leaching) also dissolve additional material such as iron, which is the main component of the waste stream. Therefore Tata Steel Europe is investigating whether alternative options are available to selectively extract these heavy metals and reintroduce the remaining material as secondary ore back to ironmaking.

In this research the option of adding oxidants with and without additional microwave treatment is investigated. Based on what is applicable in reality, literature and thermodynamic evaluations several oxidants were selected to oxidize the metal-sulfides to metal-oxides which can be readily leached in ammonia based media. Furthermore the material was characterized and an investigation was done on where the sulfur originates and whether this stream can be reduced or avoided. The quality of the results was verified by comparing the heavy metals dissolved into the leaching solution by ICP analysis and by determining the heavy metal content in the solid material before and after leaching with XRF. It was found that 98% of the sulfur originates from the blast furnace and only 2% from the addition of sodium sulfide at the water cleaning department.

Leaching experiments were done with a 40 °C leaching solution, a 5:1 liquid to solid ratio. Leaching solutions were all ammonia based due to its selectivity to heavy metals. Leaching the sample material without microwaving or addition of oxidants the highest Zn, Pb and Cd recovery was 54%, 0.5% and 24% respectively. Potassium permanganate as the oxidant resulted in the highest extraction of Zn at 67% in a 2M ammoniumsulfate, 9M ammonia leaching solution. Ferric chloride in a 2M ammoniumsulfate, 9M ammonia leach solution reached extractions of Zn and Cd at 64% and 45% respectively for the IJmuiden filter cake. For Pb the best extraction was 5% in a 2M ammoniumcarbonate, 9M ammonia leach solution with the same oxidants. Microwaving the sample during leaching resulted in a higher extraction for both Pb and Cd, whilst matching the non-microwave oxidant leaching extraction for Zn. The highest extractions were 61%, 42%, 83% for Zn, Pb and Cd respectively. Aging has also been proven to have an detrimental effect on the extraction of heavy metals, however the aging effect is negated when using microwave treatment or oxidants. Extractions for lab reactor leaching of fresh samples without added oxidants or microwaving with an ammonia sulfate solution, resulted in extractions of 62%, 0%, 40% for Zn, Pb and Cd respectively.

It has been verified that most of the sulfur comes from the blast furnace. Further it has been reaffirmed that the metal sulfides this sulfur forms are difficult to leach out. One study states that metal sulfides can and will form in the blast furnace. The current study supports that, even though no direct study of the blast furnace itself was done. Oxidants and microwave assisted leaching seem to work well for Pb and Cd, but are less effective on extracting a higher percentage of Zn. It is recommended to use this method in combination with another process, a different ironmaking process such as HIsarna would be preferable due to the reduction of ZnS to ZnO which is readily leached with an ammonia leaching solution.

The rapid degradation of the Thorn-sandstone could be explained by the presence of organics or minerals such as swelling glauconite, oxidized glauconite or non-swelling clays sensitive to hydric variations despite being non-swelling. A variety of tests and observations have been carried out to identify the most likely process of degradation. Conventional optical microscopic observations, XRF and XRD analyses, Methylene Blue tests on rock powder, Methylene Blue staining of thin-sections, and Loss-on-Ignition have been conducted as well as more advanced observations using X-ray micro-computed tomography (micro-CT) and Scanning Electronic Microscopy (SEM). The micro-CT scans allowed the visualization of the grain arrangement in 3D along with the network of micro-fissures spread around within the material. The SEM combined with EDEX measurements allowed the visualization of grains in 2D in great detail along with determining chemical elements at the sub-grain scale. Combined results showed that the Thorn-sandstone consists of quartz grains along with some grains of hematite, rutile or ilmenite. Porosity (excluding fissures) is very low. The grains are tightly packed and surrounded by a very heterogeneous fine-grained brown material. This brown material consists of a mixture of minerals, possibly clays, chlorite , micas, weathered glauconite and feldspar as well as small grains of rutile, hematite or ilmenite. The aforementioned Micro-fissures run through the brown material. The sensitivity of the brown clayey material to water, when exposed to wetting and drying cycles, is likely to explain the degradation of the Thorn-sandstone. Other hypotheses could be formulated to explain the rapid deterioration of the Thorn sandstone.

The metal cobalt is essential in creating a sustainable society, as it is an important component in lithium- ion batteries and wind energy turbines. About half of the Earth’s land-based cobalt reserves can be found in the Democratic Republic of Congo (DRC). This study focusses on characterising cobalt-bearing copper ore from a DRC based mining operation and linking characteristic ore properties to the comminution behaviour. Portable X-ray Fluorescence (pXRF) is used to determine the geochemistry, while QEMSCAN, CT scan and InfraRed (IR) spectroscopy are used to assess mineralogy on different length scales. In terms of geotechnical properties, the surface hardness of the ore is determined using a rebound hardness tester. The comminution behaviour is assessed by crushing selected ore samples below 3.35 mm and by grinding the crushed samples for 21 minutes using a ball mill. The energy consumption required for different ore types is quantified using the crushing and sieving simulation package Bruno.

The chemical characterisation of ore samples showed significant variation in cobalt content, as the cobalt grade measured in the samples using pXRF varied from 0% to 33.5%. By applying QEMSCAN analysis to samples coming from different locations within the deposit, two different mineralogical zones have been identified: a weathered, and a sulphide zone, with large differences in modal mineralogy. In the weathered zone kolwezite and heterogenite are the principal cobalt bearing minerals, in the sulphide zone carrollite is the only mineral containing cobalt. The CT scan confirmed the unique mineralogy of certain samples, as the mineralogy of pure carrollite nodules found using QEMSCAN showed to continue in 3D. Both QEMSCAN and CT scan were able to detect differences in mineral texture of the studied samples.
In terms of geotechnical properties, differences between surface hardness of the samples were observed using the rebound hardness tester. An empirical formula has been derived to adjust the measured surface hardness of drill core samples with a mass below 300 grams. This formula aided in identifying the differences in surface hardness for the distinct mineral textures of the sulphide zone. Influence of mineralogy on the sur- face hardness has been observed, as areas containing only carrollite showed to have a lower surface hardness than areas with a mixture of Mg silicates, quartz and carrollite.

Fourier Transformed IR (FTIR) spectroscopy was used to identify a range of gangue minerals within the samples. Carrollite could be identified if it was present in significant proportions and no other IR active minerals were present. FTIR spectroscopy also allowed to refine mineralogical data, as it was possible to differentiate between talc or serpentine and muscovite or illite in the studied samples, which was not possible using QEMSCAN. In total 8 minerals were identified using IR spectroscopy. The relative proportions of the modal mineralogy showed to be in agreement with the mineralogy found through QEMSCAN.

The comminution experiments led to the identification of three different comminution classes. These three classes can again be linked to the identified mineral texture types. The surface hardness showed to be correlated to the p80 of samples with the same mineral texture. After 21 minutes of milling, the samples with a higher surface hardness had a coarser p80. Simulation showed that the crushing energy requirements differ up to a maximum of 27% for different comminution classes. This research provides a foundation towards identification of geometallurgical domains in the sulphide section of the investigated cobalt-bearing copper deposit.

This MSc thesis was carried out at Camborne School of Mines (University of Exeter) as part of the NERC funded Cobalt: Geology, Geomicrobiology and Geometallurgy (CoG3) project.