The Bayer Process

What the Bayer process does

Bauxite is not a mineral but a rock — a weathered laterite that is mostly aluminium hydroxides mixed with iron oxides, silica, and titania. It looks like red-brown dirt because of that iron. You cannot smelt aluminium out of it directly: the impurities would contaminate the metal and the electrolysis cell. The Bayer process is the chemical refining step that turns dirty ore into a clean white oxide, and it hinges on one property of aluminium chemistry — amphoterism.

Aluminium hydroxide reacts with acids and with bases. The iron and titanium oxides that give bauxite its colour react with neither strong base. So if you soak crushed bauxite in hot concentrated sodium hydroxide, the aluminium dissolves away as a clear, colourless sodium aluminate solution, and everything else is left behind as an insoluble red sludge. Filter, cool the liquor, and reverse the reaction to drop the aluminium back out — now pure. That selective dissolve-and-reprecipitate is the whole trick.

  bauxite  ──NaOH, hot──►  clear NaAl(OH)₄ liquor   +   red mud (Fe₂O₃, TiO₂, SiO₂)
                                    │ cool, dilute, seed
                                    ▼
                              pure Al(OH)₃ crystals
                                    │ calcine ~1050 °C
                                    ▼
                              Al₂O₃  (smelter-grade alumina)

Step 1 — Digestion: dissolving the aluminium

Crushed bauxite is fed into a slurry of hot, strong caustic soda inside pressurized digesters. The amphoteric aluminium hydroxide behaves as an acid toward the excess hydroxide and dissolves as the tetrahedral aluminate ion, [Al(OH)₄]⁻ (equivalently written as sodium aluminate, NaAlO₂ + water). The exact reaction depends on which aluminium mineral is present:

Gibbsite  (γ-Al(OH)₃):   Al(OH)₃ + OH⁻ (aq)  ─►  [Al(OH)₄]⁻ (aq)
Boehmite  (γ-AlOOH):     AlO(OH) + OH⁻ + H₂O ─►  [Al(OH)₄]⁻ (aq)
Diaspore  (α-AlOOH):     AlO(OH) + OH⁻ + H₂O ─►  [Al(OH)₄]⁻ (aq)

net (as sodium aluminate):
   Al₂O₃·3H₂O  +  2 NaOH  ─►  2 NaAl(OH)₄

Which mineral dominates sets the conditions. Gibbsitic bauxites (typical of West Africa and Brazil) dissolve readily at ~140-150 °C and a few bar. Boehmitic and diasporic bauxites (common in the Mediterranean, China, and India) are far more stubborn: diaspore needs ~250-270 °C and 30+ bar, and often extra lime (CaO) to break its dense crystal lattice. The caustic liquor is strong — typically 120-350 g L⁻¹ Na₂O equivalent — and the whole digestion is run under pressure precisely so the water can be superheated above 100 °C without boiling.

Step 2 — Clarification: leaving the iron behind

After digestion you have a hot slurry of dissolved aluminate liquor plus suspended solid impurities. Those solids — red mud — are the parts of the ore that would not react:

• Iron(III) oxide, Fe₂O₃ (hematite) and FeO(OH) (goethite) — a basic oxide, chemically inert to NaOH, and the source of the red colour.
• Titanium dioxide, TiO₂ — insoluble in caustic under these conditions.
• Silica, SiO₂ — the troublemaker. Reactive clay silica does dissolve, then reprecipitates as insoluble sodium aluminosilicate ("desilication product", a sodalite-like phase). This is bad twice over: it consumes both caustic and dissolved alumina. Every reactive-silica atom in the ore locks up soda and lost yield, which is why low-silica bauxite commands a price premium.

The liquor is cooled slightly, flocculant is added, and the mud settles in large thickeners and is washed on filters to recover entrained caustic (worth far too much to throw away). The overflow is a clear, supersaturated green liquor of sodium aluminate, sent forward; the underflow red mud is pumped to residue storage.

Step 3 — Precipitation: reversing the reaction

Here is the elegant part. Digestion and precipitation are the same equilibrium pushed in opposite directions by temperature and concentration:

        hot, concentrated NaOH  →  dissolution
   Al(OH)₃ (s)  +  OH⁻ (aq)   ⇌   [Al(OH)₄]⁻ (aq)
        cool, dilute, seeded    →  precipitation

The clarified liquor is cooled to about 60-75 °C and diluted, which shifts Le Chatelier's balance back toward solid Al(OH)₃. But the liquor is metastable: it can stay supersaturated for hours because gibbsite nucleates spontaneously only very slowly. So the plant seeds it — dumping in tonnes of fine gibbsite crystals recycled from a previous batch. The dissolved aluminate grows onto that ready surface instead of forming new nuclei, and over 30-60 hours in tall precipitator tanks, up to about half the dissolved alumina crystallizes out as pure aluminium hydroxide.

Seeding controls crystal size, not just yield. Coarse, well-formed gibbsite (~50-100 µm) filters, washes, and calcines cleanly; fine or dendritic crystals clog filters and carry caustic. The spent liquor — now weaker caustic — is reheated, re-concentrated in evaporators, and pumped back to digestion. The caustic soda is a recycled reagent, not a consumable; the only NaOH the plant buys is to top up losses to red mud and desilication product.

Step 4 — Calcination: driving off the water

The washed gibbsite is still hydroxide — Al(OH)₃ — and a smelter wants the anhydrous oxide. Heating to ~1000-1100 °C in a rotary kiln or, in modern plants, an energy-efficient gas-suspension (flash) calciner drives off the chemically bound water:

   2 Al(OH)₃  ──►  Al₂O₃  +  3 H₂O↑        (ΔH ≈ +180 kJ per mol Al₂O₃, endothermic)

   intermediate: gibbsite → boehmite/χ-alumina → γ-Al₂O₃ → α-Al₂O₃ (above ~1100 °C)

The product is smelter-grade alumina: a white, free-flowing powder over 99 % Al₂O₃. Its phase mix and particle size are tuned deliberately — smelters want mostly γ- and transition-aluminas (not fully-sintered corundum) because they dissolve fast in the molten cryolite bath and adsorb fluoride emissions in the cell off-gas. That alumina then goes to the electrolytic Hall-Héroult step to become metal.

Bayer process vs Hall-Héroult vs older routes

	Bayer process	Hall-Héroult process	Deville (pre-Bayer, obsolete)
What it does	Refines bauxite → pure Al₂O₃	Reduces Al₂O₃ → Al metal	Reduced AlCl₃ with sodium
Type of step	Chemical (acid-base + crystallization)	Electrochemical (electrolysis)	Chemical reduction
Key reagent / input	Recycled NaOH, heat, pressure	Molten cryolite, ~15 kWh per kg Al	Metallic Na (very expensive)
Product	Al₂O₃ (smelter-grade alumina)	Aluminium metal (~99.7 %)	Aluminium metal (impure, costly)
Year	1888 (Bayer)	1886 (Hall & Héroult, independently)	1854 (Deville)
Waste	Red mud (alkaline residue)	Spent pot lining, CO₂, PFCs	Sodium chloride salts
Where it sits	First half of the supply chain	Second half of the supply chain	Made aluminium a precious metal

The two modern processes were invented within two years of each other and are almost always run in sequence: Bayer removes the impurities, Hall-Héroult removes the oxygen. Together they collapsed the price of aluminium roughly a thousandfold — from a metal more precious than gold in the 1850s (the capstone of the Washington Monument is cast aluminium) to a throwaway drinks can.

Worked example: mass balance for one tonne of alumina

Follow the numbers through a typical gibbsitic plant to see why bauxite grade matters so much.

• Ore. A good bauxite is ~50 % available alumina (Al₂O₃). To make 1 t of alumina you must dissolve 1 t of Al₂O₃, so you feed roughly 2 t of high-grade bauxite (more, 4-5 t, for lean or high-silica ore).
• Digestion. 1 t Al₂O₃ (≈ 9.8 kmol on an Al₂O₃ basis) reacts with 2 mol NaOH per mol Al₂O₃; the caustic is recycled, so the plant only buys the fraction lost — typically 50-100 kg NaOH per t alumina, mostly consumed by silica forming desilication product.
• Red mud. The insoluble balance of the ore leaves as residue — about 1-1.5 t red mud per t alumina, still alkaline at pH 12-13.
• Calcination. Precipitated Al(OH)₃ is 2 × 78 g mol⁻¹ = 156 g losing 54 g water per 102 g Al₂O₃ formed — so you crystallize about 1.53 t of gibbsite to calcine down to 1 t of Al₂O₃, driving off ~0.53 t of water.
• Energy. Calcination is endothermic (~180 kJ per mol Al₂O₃ for the dehydration itself, and far more once the solids are heated to ~1000 °C and the water is boiled off) and digestion needs high-pressure steam, so a modern refinery spends on the order of 10-14 GJ of thermal energy per tonne of alumina.

Multiply out and the reason for the industry's geography is obvious: you need cheap energy, cheap caustic, and somewhere to put a mountain of alkaline mud. That is why alumina refineries cluster near bauxite mines and cheap gas.

The red mud problem

Selectivity has a cost: everything you did not dissolve has to go somewhere. Red mud (bauxite residue) is the single largest by-product tonnage in the whole of chemical industry — over 150 million tonnes generated a year, with several billion tonnes already in storage. It is a slurry of iron oxides, titania, aluminosilicate desilication product, and leftover caustic, sitting at pH 12-13.

• Disposal. Historically pumped wet into vast diked ponds. Modern practice is "dry stacking" — thickened to a paste and dried in layers so it stores more safely and re-vegetates.
• The Ajka disaster (2010). A residue dam at Ajka, Hungary failed and released ~1 million m³ of caustic red mud, killing 10 people and burning survivors with alkali. It is the definitive cautionary tale for wet storage.
• Neutralization. Treating the residue with seawater precipitates Mg/Ca hydroxides and carbonates, or bubbling flue-gas CO₂ through it carbonates the free caustic, dropping the pH to ~8-9 and locking away CO₂ at the same time.
• Re-use. Small fractions go into cement clinker, as an iron source, in bricks, and for rare-earth/scandium and iron recovery. But re-use still consumes only a few percent of what is produced — the mismatch between output and demand is the unsolved part.

History: Bayer, Le Chatelier, and the aluminium age

Carl Josef Bayer, an Austrian chemist working in Saint Petersburg, patented the two key insights in 1887-1889. His 1888 patent covered the crucial discovery that a sodium aluminate solution, if seeded with aluminium hydroxide and left to stand, will spontaneously precipitate coarse, easily-filtered gibbsite — and that the spent liquor can then be reused. His 1892 patent added the digestion side: that bauxite could be dissolved directly under pressure in the recycled caustic. Together those closed the loop into the cyclic process we still use.

He was building on earlier work by Louis Le Chatelier (father of the Le Chatelier of the equilibrium principle), who in 1855 had shown bauxite could be sintered with sodium carbonate and leached — but that route wasted soda and gave fine, unmanageable precipitate. Bayer's genius was the seeded, closed-loop crystallization. His process arrived at exactly the right moment: Hall and Héroult had independently invented cheap electrolytic smelting in 1886, and the two together made aluminium an industrial metal for the first time. In over 130 years the chemistry has not changed — only the scale, the pressure vessels, and the energy efficiency.

Industrial and safety notes

• Caustic burns. The liquor is hot concentrated NaOH under pressure — among the more hazardous streams in heavy industry. Leaks cause severe alkali burns; the whole plant is engineered around containing a boiling caustic slurry.
• Scaling. Sodium aluminosilicate and other scales plate out on heat-exchanger and digester walls, throttling heat transfer. Descaling (acid washing or mechanical) is a constant maintenance battle and a major operating cost.
• Grade economics. Available-alumina and reactive-silica content set the caustic consumption and mud volume, so bauxite is priced on chemistry, not just tonnage. A high-silica ore can consume so much soda and alumina that it is uneconomic despite a high total-aluminium assay.
• Energy intensity. Alumina refining is one of the more energy-hungry chemical processes; combined with the electricity-hungry Hall-Héroult step, aluminium's total embodied energy is why recycling scrap aluminium (which skips both) saves ~95 % of the energy of primary metal.
• Cyclic reagent. The most under-appreciated feature: the NaOH is regenerated in every loop. The plant is really a giant continuously-operated crystallizer that shuttles aluminium between the dissolved and solid states, not a one-shot reaction vessel.