Project topics and materials
Get Complete Project Material File(s) Now! »
THEORY AND LITERATURE REVIEW
ALUMINIUM PRODUCTION – THE HALL-HEROULT PROCESS
As the name indicates, the invention of the method to produce aluminium metal by electrochemical reduction of alumina (Al2O3) is attributed both to Charles Hall and Paul Héroult. Although major changes in the operation, materials and understanding of this process have taken place since the invention in 1886, the fundamental principle remains the same. In essence, alumina is dissolved in the molten electrolyte bath (mainly cryolite or Na3AlF6) at temperatures around 960°C, an electric current is passed through the electrolyte reducing the aluminium oxide into liquid aluminium and oxygen in the form of negatively charged oxyfluoride complexes, reacting with the carbon anodes to form carbon dioxide. The main electrochemical reaction that takes place can be described as follows (reaction 3.1.): 2Al2O3(diss) + 3C(s) -> 4Al(l) +3CO2(g) /3.1./
The theoretical reaction voltage (or decomposition potential) is approximately 1.17 V at an operating temperature around 960°C and with a ‘normal’ bath chemistry; however, the total cell voltage lies around 4.0 V, when the polarisation and overvoltage terms are included [20]. Liquid aluminium has a higher density than the electrolyte bath and is therefore deposited at the bottom of the electrochemical cell. Some aluminium remains in solution (or may be dissolved due to process fluctuations or disturbances) and can react with CO2 (formed in reaction 3.1.) to form Al2O3 and carbon monoxide, according to reaction 3.2.: 2Al(diss) + 3CO2(g) -> Al2O3(diss) + 3CO(g) /3.2./
This ‘back-reaction’ is undesirable as it, in addition to forming toxic CO gas, decreases the current efficiency of the cell and increases the carbon consumption.
Although alternative ways of producing aluminium exists, the Hall-Heroult process is the only commercially used method at this time. It should also be noted that two variations of the process exist, namely: the prebake and the Soederberg technology. The Soederberg technology uses a continuously created anode; made by addition of a mixture of petroleum coke and coal tar pitch (25-28 wt-%) to the top of the anode (the heat from the smelting process ‘bakes’ the pitch). As the name implies, in the prebake technology the anodes are made in a separate process where they are ‘baked’ in large ovens at high temperatures before being used in the cell. The prebake anode technology now dominates the industry. A schematic representation of a prebake cell can be seen in figure 3.1. a) and a cross section in figure 3.1. b) (figures reproduced from [21]). Typical cell dimensions range from 3-5 meters in width and 9-18 meters in length. The electrical current enters the cell via the busbar, goes through the anodes and into the bath and metal (sodium ions are the main current carrier in the bath), after which it passes through the cathode and into the collector bars. The collector bars carry current through an external busbar network to the next cell in line. Several hundreds of cells may be connected in series, forming a ‘potline’. The amount and dimensions of the anodes vary between smelters, but typically around 20 anodes are used in each cell. The trend in the industry has been for higher electrical currents and larger cells (today testpots are being trialled with capabilities up to 500 kA [22]), as this facilitates energy savings and increases production rates.
The Many Roles of Alumina
As may be seen in table 3.1. Metallurgical Grade Alumina is not only used as a feedstock. Alumina quality requirements such as: “ability to capture fluoride” and “desirable crust formation” obviously relate to other processes or operations in the aluminium smelter. Unless primary alumina is used in the reduction cell, which is the case when high purity metal is produced, the alumina is used to adsorb the fluoride containing cell off-gases before it is fed to the cell (note that CO2, CO, SO2, PAHs, PFCs and HFCs may also be produced in the process and are not as efficiently absorbed by the alumina and thus requires other means for their removal and/or reduction). This is carried out in so called ‘dry-scrubbers’, in order to reduce harmful gaseous and particulate emissions and to recover important bath chemicals (fluorides). In addition, alumina is also used as a component of cell cover to protect the anodes from airburn and to prevent fugitive emissions from the cell. For these reasons maintaining a proper anode cover, in terms of integrity and thermal and mechanical properties, is critical. This follows from having the right mixture and granulometry of the alumina and crushed bath components of which the cover material is comprised. The main uses of alumina may be summarised as:
• Primary feedstock
• Anode and cell cover material
• Scrubbing medium
The many and different roles or uses of alumina in the smelter puts a lot of demand on the properties. There are also additional requirements, not directly related to any particular process step, such as: low impurity content, high particle strength and low dusting tendency, critical for an efficient operation and a safe and desirable working place. In the following, the main uses (listed above) will be discussed separately, and the alumina properties most closely tied to the performance criteria highlighted.
Alumina Feeding and Dissolution
The stoichiometry of reaction 3.1. dictates that for every mass unit of aluminium metal that is produced (e.g. 1 kg Al) approximately one third of that mass of carbon (0.33 kg C) and almost twice the mass of alumina (~1.92 kg Al2O3) is needed (in practice the carbon consumption is significantly higher than the theoretical value due to the back reactions, discussed earlier, and other process inefficiencies). The capacity for dissolution of alumina in the bath is, however, limited and controlling the alumina concentration in the cell therefore becomes critical for maintaining a stable operation. Feeding too much alumina (‘over-feeding’) results in the accumulation of undissolved alumina on the cathode surface (called ‘sludge’ or ‘muck’), with negative impacts on the operation. Conversely, not feeding enough alumina (‘under-feeding’) can, unless detected, result in a preferential reduction of the bath components inducing an ‘anode effect’, which results in a direct loss of efficiency and is often associated with increased emissions.
Alumina dissolution is a multistep process that involves: an initial agglomeration and a freezing of bath on the agglomerates, followed by mineralization of transition alumina (to alpha alumina) and melting of the freeze and finally dissolution by formation of aluminium-oxyfluoride complexes. Apart from having a sufficient dissolution power (often referred to as ‘superheat’, which is the temperature above the melting point of the bath) providing a higher surface area for the dissolution process/reaction is also beneficial. One way to achieve a higher surface area for the dissolution is through improved dispersion of the alumina on, and in, the bath. The bubble induced bath flow is important for this (and hence also the placement and design of slots in the anodes), but more so, how the alumina enters the cell. For this two main methods are used: ‘point feeders’ or ‘bar breakers’. Point feeding is becoming more widely used, particularly in modern cell designs and as part of retro-fitting older cells. In the case of using point feeders, the alumina is fed into the cell via a number (typically 3-5) of point feeders located along the centre channel in the cell. A point feeder will feed up to 5 kg of alumina every 1-3 minutes. Using the Bar Breaker method, the crust is broken into the molten bath to replenish the alumina content; this is then followed by dumping alumina over the broken surface to repair the crust. This method of feeding is carried out less frequently, with feeding cycles typically between 10 and 60 minutes.
The dispersion and dissolution has also been found to be influenced by the alumina properties [15, 17, 24-26]. In the studies by Liu et al. and Welch et al. visualisation of the dissolution process was made possible through the use of video recording equipment in laboratory scale electrolysis cells and the results were correlated against simultaneous measurements of the electrolyte temperature and the alumina concentration in the bath [17, 25]. Both studies indicated that transition aluminas dissolve faster than alpha alumina rich samples. This was partly attributed to the tendency of the alpha alumina rich particles to aggregate. Keller et al. showed that if the size and density of the agglomerates are high enough these may sink through the bath and metal to form deposits on the cathode surface (‘sludge’) [26]. Welch et al. also point out that the exothermic nature of the transition alumina to alpha alumina transformation may aid with the bath temperature whereas dissolution of alpha alumina is more dependent on the superheat in the bath. Furthermore, Liu et al. argues that the shattering of transition alumina rich particles (possibly due to a volume change and/or evolution of vapour) in the melt increases the surface area and thus the dissolution rate.
Østbø attributed differences in the dissolution characteristics for two aluminas with similar surface areas to differences in the nano-structure and pore structure (based on X-ray and porosimetry studies) [15]. The results by Østbø also indicated that the transition aluminas have a greater tendency to aggregate, which the author attributed to the formation of platelets (in the mineralization process) which would interlink the grains. Dando et al. also presented a similar mechanism and observed a formation of ‘rafts’ of alumina aggregates on the surface of the bath [24]. Using electron microscopy these rafts were shown to be composed of alumina particles interlinked by platelets, which would support the findings by Østbø. Studies have also shown that the presence of fluoride species on and in the alumina, as well as the moisture (adsorbed water) and residual hydroxyl content and the presence of un-calcined material (gibbsite, Al(OH)3) influences the dissolution behaviour as well as the formation of HF [15, 24, 27, 28]. Residual hydroxyls have been observed to be present in the transition aluminas (where they are thought to be incorporated in the crystal structures) due to the incomplete conversion to alpha alumina in the calcination stage in the Bayer process (discussed briefly in section 3.2. and in further detail in section 3.3).
The Role of Alumina in the Cell Cover
The anode cover, or more aptly “cell cover” material, plays important roles for maintaining a stable and productive operation. Today it is widely accepted that alumina is a key component in controlling the properties of this material.
Traditionally the anode cover had mainly an insulating role (i.e. minimising the heat losses). However, with higher line currents being used the anode temperatures also increased and protecting the anodes against airburn (which is the oxidation of the anode carbon by reaction with oxygen or CO2) became increasingly important. With high amperage operation a need to dissipate more heat through the anode cover also became apparent. In fact, adjusting the thermal properties of the anode cover (through composition and granulometry as well as thickness) is one of only two ways to control the heat flow out of the cell without changing the cell design [19] (the other being through adjusting the liquid levels and hence the heat loss through the sidewall).
The environmental and health aspects are increasingly important research areas in aluminium production. In this regard the cell cover is critically important as it also acts as a primary prevention mechanism in emissions control by capturing volatile fluorides (such as NaAlF4, AlF3 and HF). It is easy to see that the cover therefore also plays an important role in the bath level and bath chemistry control [29] as well as in metal purity [30].
The requirements on the anode cover material goes beyond the mechanical and thermal properties needed to control the heat balance and maintain the integrity of the cover during the routine operations (such as feeding, changing anodes etc). The ability to capture volatile fluorides arises from the high surface area of the transition alumina phases, and is intricately associated with the formation of interlocking alpha alumina platelets which bond the crushed bath components together. This is however a delicate balance as alpha alumina (also known as corundum) is one of the hardest materials known to science (having a hardness of 9 on the Moh’s scale). As the spent anodes are recycled the associated bath and cover material is also recovered to make fresh cover material. In the presence of significant amounts of alpha alumina the processing of this material may pose mechanical challenges. The properties also have to be balanced with the correct thermal properties, as the risk otherwise is that the cover starts to melt, which undermines the mechanical properties ultimately resulting in a collapse of the cover.
For these reasons the control of the cover composition is extremely important. Although it is widely appreciated that alumina is needed in an optimal cover material, researchers are only now beginning to understand the impact of alumina properties on the integrity and evolution of the cover and crust material and properties.
Alumina Properties and Fluoride Emissions
Fluoride emissions are regulated and therefore mechanisms for their reduction are an essential part of the aluminium production process. The two main technologies for fluoride removal are: Wet Scrubbing and Dry Scrubbing. As the name implies, Wet Scrubbing makes use of an aqueous spray to remove soluble species which are collected as a liquor and then further treated. Dry Scrubbing makes use of the adsorbing capability of alumina to reduce the emissions. Both technologies have the added benefit of allowing the bath chemicals (fluorides) to be returned to the cell. The major drawbacks with wet scrubbing are the corrosion problems and handling and treatment difficulties related to the highly acidic nature of the scrubbing solution. Today, wet scrubbers are therefore mainly used for sulphur dioxide removal.
The modern, injection type, dry scrubbing system comprises of a reactor (where alumina is injected and comes in contact with the collected cell off-gases) and a separate filtrations system (where the reacted alumina and other particulates are separated from the gas stream). The reactors are designed to ensure good mixing between gas and solids. These systems can be operated at very high efficiencies, and more than 99 % of the gaseous and particulate fluorides may be recovered. With good pot hood collection efficiency, and the dry scrubber recovering and recycling most of the particulate and gaseous fluorides, the operation can be regarded as a closed loop between the cells and the scrubber [31]. As the high particulate collection efficiency also includes those impurities lost to the duct (elements forming volatile fluoride compounds), the recycling of the dry scrubber alumina (secondary or reacted alumina) naturally leads to increased impurity levels in the bath, and thus also in the metal [32].
The chemistry and kinetics of the dry scrubbing process (i.e. the adsorption and reaction of HF with alumina) was explored in depth by Gillespie [13]. Gillespie argued that since the fluoride adsorption capacity is related to both the specific surface area of the alumina and the relative humidity during the adsorption, the reactions must occur in a surface process. The author presented a mechanism which involved several steps, starting with the adsorption of water on the surface of the alumina followed by HF adsorption and acidification of this surface layer which would dissolve the alumina surface and form AlO2- and hydroxide species and finally cause the precipitation of oxy- and hydroxyfluorides. Most importantly the reaction was shown to be irreversible under dry scrubbing conditions (temperature and atmospheric), an important finding which enabled new developments of the dry scrubbing process and control strategies to be made.
Although the mechanisms for the reaction between HF and alumina (in the dry scrubbing process) are relatively well understood, the role of alumina microstructure and porosity is less clear. There is a direct relationship between the surface area available for adsorption and the adsorption capacity [4], however it could be expected that the pore size distribution also plays an important role in providing access to internal reaction sites. The porosity and surface area arises from the incomplete conversion of gibbsite to alpha alumina in the calcination stage and is therefore associated with the transition alumina phases.
As mentioned earlier, the residual hydroxyls are an integral part of the transition alumina structures, but are also the main source of HF formation in the electrolyte bath. In the cell HF is formed when the released water (or OH) reacts with the electrolyte. The surface adsorbed water and more loosely bound hydroxyls (represented by the MOI values) on the other hand are rapidly flashed off when the alumina is fed into the electrolytic cell [28]. This may actually be beneficial for the dissolution process as it helps with the dispersion of the material [27].
It is then easy to see that there is a conflicting relationship between generation of HF from residual hydroxide and capturing the HF by increasing the surface area (which inevitably results in more structural hydroxyls and increased HF formation in the first place). Understanding how the development of alumina microstructure is influenced by the calcination conditions and precursor material, and in particular understanding the structures that arise during the transformation reactions, is key to improve the efficiency of both the calciners and the aluminium smelting process through finding an optimum balance, or compromise, between the critical properties.
TABLE OF CONTENTS
ABSTRACT
ACKNOWLEDGEMENTS
TABLE OF CONTENTS
LIST OF FIGURES
LIST OF TABLES
1. GENERAL INTRODUCTION
2. FOCUS AND STRUCTURE OF THESIS
3. THEORY AND LITERATURE REVIEW
3.1. ALUMINIUM PRODUCTION – THE HALL-HEROULT PROCESS
3.2. ALUMINA PRODUCTION – THE BAYER PROCESS
3.3. GIBBSITE CALCINATION AND STRUCTURAL EVOLUTION
4. INVESTIGATIONS INTO THE EFFECTS OF CALCINER TECHNOLOGY AND OPERATION ON THE DEVELOPMENT OF ALUMINA MICROSTRUCTURE
4.1. DEVELOPMENT OF POROSITY AND MICROSTRUCTURE
5. PHASE TRANSFORMATION PATHWAYS AND PHASE DISTRIBUTION IN METALLURGICAL GRADE ALUMINA
5.1. PHASE ANALYSIS OF METALLURGICAL GRADE ALUMINA
5.2. GIBBSITE CALCINATION PATHWAYS
5.3. EFFECT OF HEATING RATE AND IMPURITIES ON ALUMINA PHASE DISTRIBUTION
6. BEYOND THE MICROSTRUCTURE – EVALUATIONS OF NANOSCALE ORDERING IN METALLURGICAL GRADE ALUMINA
6.1. SITE OCCUPANCY AND SHORT RANGE ORDER IN ALUMINA
7. CONCLUDING REMARKS AND SUGGESTIONS FOR FUTURE WORK
8. DETERMINISTIC METHODS
Equipment for Short-term Calcinations
Nitrogen Porosimetry
Electron Microscopy
Time of Flight SIMS
Thermal Analysis
X-Ray Diffraction
Particle Size Distribution Measurements
27Al SS MAS NMR
X-ray Absorption Spectroscopy (XAS)
APPENDICES
REFERENCES
GET THE COMPLETE PROJECT
EVOLUTION OF NANO- AND MICROSTRUCTURE DURING THE CALCINATION OF BAYER GIBBSITE TO PRODUCE ALUMINA
