Question 7 of 7: Aluminum Production (20 marks)

Question 7 — Aluminum Production (20 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Primary aluminium production is a two-stage industry, and the question follows that division: the Bayer process refines bauxite to pure alumina by hydrometallurgy, and the Hall–Héroult process reduces that alumina to metal by molten-salt electrolysis. Roughly four tonnes of bauxite give two tonnes of alumina and one tonne of aluminium.

(a) The Bayer process. The Bayer process exploits a single fact: aluminium hydroxide is amphoteric and dissolves in hot concentrated caustic soda, whereas the iron, titanium and silicon compounds that accompany it in bauxite very largely do not. The separation is achieved by dissolving the alumina, filtering off the insoluble residue, and then reversing the dissolution.

Preparation. Bauxite — a mixture of gibbsite Al(OH)₃, boehmite and diaspore AlO(OH), with iron oxides, silica and titania — is crushed, ground and slurried in spent caustic liquor returning from the precipitation stage.

Digestion. The slurry is heated with sodium hydroxide in a pressure vessel, dissolving the aluminium as sodium aluminate:

Conditions are set by the mineralogy: gibbsitic bauxite digests at 140–150 °C, while the boehmitic and diasporic ores need 200–250 °C and correspondingly higher caustic concentration and pressure. Reactive silica dissolves as well and then reprecipitates as sodium aluminosilicate — “desilication product” — which consumes both soda and alumina and is the reason a high reactive-silica bauxite may be uneconomic regardless of its alumina grade.

Clarification. The pregnant liquor is flashed down in pressure, and the undissolved residue is separated by settling in thickeners with a flocculant, followed by washing in a counter-current decantation train to recover entrained caustic, and finally polish filtration. The residue is the red mud of part (b).

Precipitation. The clarified liquor is cooled to 50–75 °C and seeded with fine aluminium hydroxide crystals, which reverses the digestion reaction and crystallises out gibbsite over 20–50 hours:

Seeding controls both the yield and the particle size distribution, which matters downstream because the smelter needs a free-flowing sandy alumina that dissolves readily in cryolite and does not dust.

Classification and calcination. The hydrate is classified, the coarse fraction becoming product and the fines returning as seed. The product is washed and calcined at 1000–1100 °C in a rotary kiln or fluidised-bed calciner:

The spent liquor, now depleted in alumina but rich in caustic, is concentrated by evaporation and returned to digestion. This closed caustic loop is the economic heart of the process; soda losses to the red mud and to desilication product are the principal reagent cost.

(b) What is red mud? Red mud is the insoluble residue left after digestion — the fraction of the bauxite that the caustic did not dissolve. Its colour comes from the iron oxides, principally haematite and goethite, that make up 30–60 % of it; the balance is titania, quartz and unreacted alumina, together with the sodium aluminosilicate desilication product. It is strongly alkaline, typically pH 10–13, from entrained caustic that washing never entirely recovers, and one to two tonnes are produced per tonne of alumina. That alkalinity and the sheer volume make its disposal the industry's largest environmental liability, managed by dry stacking or in engineered impoundments; the 2010 Ajka dam failure in Hungary is the reference case for what happens when such an impoundment is not. Modern practice presses toward neutralisation with seawater or carbon dioxide, dry stacking to reduce the liquid inventory, and recovery of the iron, titanium and residual soda — but no use has yet been found that consumes it at anything approaching the rate it is produced.

(c) The Hall–Héroult process. Alumina melts at 2054 °C and, being an ionic solid, cannot be reduced by carbon to metal in any practical furnace. Charles Hall and Paul Héroult independently solved the problem in 1886 with the same insight: dissolve the alumina in molten cryolite, Na₃AlF₆, and electrolyse the solution at a temperature the plant can sustain.

The electrolyte. The bath is cryolite carrying 2–6 % dissolved alumina, with aluminium fluoride and calcium fluoride added to depress the liquidus, and it operates at 950–970 °C. Cryolite is the essential ingredient because it dissolves alumina readily, conducts well, is less dense than aluminium so that the metal settles beneath it, and is not itself decomposed at the cell voltage.

The cell. A steel casing is lined with carbon, which serves as the cathode connection; the working cathode surface, however, is the pad of molten aluminium that lies on the floor of the cell. Prebaked carbon anodes, made from petroleum coke and pitch, are suspended into the bath from above. A frozen crust of bath and alumina forms at the sides and insulates the cell, and fresh alumina is fed through it by point feeders.

The electrochemistry. Aluminium is deposited at the metal pad, $\mathrm{Al^{3+}+3\,e^-\rightarrow Al}$, and oxygen is discharged at the anode, where it immediately attacks the carbon rather than evolving as a gas. The overall cell reaction is therefore:

The carbon anode is a reactant, not an inert electrode: about 0.4–0.5 kg of carbon is consumed per kilogram of aluminium, and anode replacement on a fixed cycle is a major part of potroom labour. Consuming the carbon also lowers the required cell voltage substantially compared with liberating free oxygen, so the anode consumption is a deliberate design choice, not merely a nuisance.

Operation. Modern cells run at 150–400 kA and about 4.2 V, of which only 1.2 V is the thermodynamic decomposition potential — the remainder covers the anode overvoltage and the ohmic drop through bath, anodes and busbars, and that resistive loss supplies the heat that keeps the bath molten. Current efficiency is 92–96 %, the shortfall arising mainly from dissolved metal being re-oxidised by CO₂ in the bath. Energy consumption is 13–15 kWh per kilogram, which is why smelters are sited at cheap hydroelectric power — the reason the Canadian industry is concentrated in Québec and at Kitimat, British Columbia. Metal is siphoned from the pad every one to two days at 99.7–99.9 % purity and cast, and cells are connected in series in potlines of 100–300 units.

(d) What is the anode effect? The anode effect is a cell malfunction that occurs when the dissolved alumina concentration falls too low, typically below about 1–2 %. Starved of oxide ions to discharge, the anode begins to discharge fluoride instead, generating a film of gas — principally the perfluorocarbons CF₄ and C₂F₆ — that no longer wets the carbon. The film insulates the anode surface, so cell resistance rises sharply, the voltage jumps from about 4.2 V to 20–50 V, and the cell begins to arc visibly at the anode with a large release of waste heat. The effect is undesirable on three counts: it wastes a great deal of energy, it disturbs the thermal balance of the cell, and the perfluorocarbons released are extremely potent greenhouse gases, with CF₄ having a global-warming potential several thousand times that of carbon dioxide. Anode-effect frequency is consequently a headline environmental key performance indicator for a smelter and is reported under Canada's greenhouse gas reporting programme. It is cleared by lowering the anode momentarily or by mechanically stirring the bath to break the gas film, and it is prevented by computer-controlled point feeding that holds the alumina concentration within a narrow band — which is why modern potlines report anode-effect frequencies below 0.1 events per cell-day, against several per day in the 1980s.