Solubility of Aluminium Trihydrate in Caustic Soda at 105°C for Alumina Refining

Dissolution of aluminium trihydrate (gibbsite, Al(OH)₃) in sodium hydroxide at 105 °C follows the reversible formation of sodium aluminate, Al(OH)₄⁻, and is governed by free caustic concentration, total caustic alkalinity expressed as Na₂CO₃, liquor temperature, and the impurity matrix. The reaction Al(OH)₃(s) + NaOH(aq) ⇌ NaAl(OH)₄(aq) reaches industrial equilibrium in low-temperature Bayer digestion circuits at alumina/caustic ratios of approximately 0.62–0.68 when total caustic is held between 180 g/L and 260 g/L Na₂CO₃ equivalent at 105 °C. Alumina concentration determined by inductively coupled plasma optical emission spectrometry in accordance with ISO 11885:2007 and total alkalinity by titration per APHA 2330 B provide the primary control variables for A/C ratio. In spent liquor returning from precipitation at 65–75 °C, A/C typically falls to 0.30–0.35, creating a solubility driving force for gibbsite digestion at 105 °C without requiring boehmite extraction temperatures above 200 °C. This solubility gap is the basis for low-temperature refinery design; however, carbonate, sulfate, chloride, and organic carbon accumulation narrows the effective A/C window and alters nucleation behaviour.

What Governs the Equilibrium Al₂O₃/Na₂O Ratio at 105 °C in Sodium Aluminate Liquors?

At 105 °C, equilibrium in the system Na₂O–Al₂O₃–H₂O is described by the concentration product of sodium and aluminate ions rather than by a simple solubility of solid Al(OH)₃. The dominant dissolved species is the tetrahedral Al(OH)₄⁻ ion, with minor dimeric aluminate species appearing above 6 M NaOH. Free OH⁻ concentration, rather than total sodium, controls the forward digestion rate; as Al₂O₃ loading increases, free OH⁻ is consumed stoichiometrically at 1 mol of NaOH per 1 mol of Al(OH)₃ dissolved. Consequently, the equilibrium A/C ratio is not constant across caustic concentrations but rises slightly with increasing caustic strength due to non-ideal ion pairing between Na⁺ and Al(OH)₄⁻ and decreasing water activity. In carbonate-free synthetic liquors, representative equilibrium values at 105 °C are shown in Table 1; these values are extracted from published solubility measurements and are consistent with continuous digester exit liquors in tropical refineries processing gibbsitic bauxite.

In a continuous 105 °C digestion circuit, predesilication and lime addition modify effective solubility because reactive silica dissolves as sodium silicate and reprecipitates as sodalite-type desilication product, consuming caustic and sodium aluminate. Bauxite slurry at 45–55 wt% solids is contacted with spent liquor containing 0.30–0.35 A/C, and the slurry is held in a series of agitated tanks for 1.5–4.0 h. Steam sparging maintains 105 °C ± 1.5 °C; lower temperatures reduce gibbsite dissolution kinetics by more than 40% per 10 °C decrease, while temperatures above 108 °C at atmospheric pressure risk local boiling, flash-induced carryover, and unwarranted quartz dissolution. Field experience on multiple stirred digesters with working volumes of 250–400 m³ demonstrates that heat exchanger surfaces in the reheat loop foul rapidly when slurry residence time exceeds 15 min in stagnant zones; the scale consists of sodium aluminosilicate, calcium carbonate, and entrapped red mud. Continuous removal of scale by high-pressure water jetting at 60–80 MPa is required every 10–20 operating days.

Seed Precipitation Boundaries and Supersaturation Decay in Refinery Circuits

Above the equilibrium solubility at 105 °C, industrial precipitation does not operate near equilibrium. Pregnant liquor at 105 °C and A/C 0.63–0.68 is cooled to 70–75 °C in flash towers and slurry coolers, generating supersaturation that is relieved by contact with fine gibbsite seed. The precipitation circuit uses seed charges of 100–200 g/L with particle surface areas of 0.2–0.6 m²/g measured by laser diffraction per ISO 13320:2020. In well-seeded continuous precipitation, exit spent liquor A/C decreases to 0.30–0.35 over residence times of 20–40 h; nucleation and agglomeration contribute to particle size distribution, and the mass median diameter typically shifts from 45–70 µm for primary product to 2–8 µm for secondary seed fines after classification. Control of free caustic at precipitation is essential: caustic concentration above 260 g/L Na₂CO₃ equivalent suppresses nucleation rate and yields excessive fines; caustic below 140 g/L Na₂CO₃ equivalent causes uncontrolled homogeneous nucleation and high liquor viscosity after cooling. The operational boundary between stable seeded growth and catastrophic fine flocculation is a caustic-temperature supersaturation path that lies within a narrow 5–8 °C cooling interval in the first flash stage.

Because the 105 °C solubility of gibbsite is highly sensitive to the Na₂O/Al₂O₃ molar ratio, liquor analysis frequency must be sufficient to prevent excursions outside the 0.60–0.70 A/C envelope. Online Raman spectroscopy and density meters calibrated against ISO 11885:2007 are employed in modern refineries to track A/C every 10–15 min; laboratory titration per APHA 2330 B remains the reference for total caustic. The density of pregnant liquor at 105 °C ranges from 1.25 g/cm³ to 1.45 g/cm³ and is non-linear with alumina content; density-based A/C control is therefore only valid when temperature compensation is better than ±0.5 °C. In high-carbonate circuits, sodium carbonate begins to crystallize as burkeite or trona in heat exchangers when liquor is cooled below 45 °C; this imposes a lower temperature limit for spent liquor storage and a secondary removal step by evaporative crystallization. Published data for the solubility of aluminium trihydrate in high-organic, high-carbonate refinery liquors at exactly 105 °C is limited; therefore, plant-specific solubility curves are generated from pressure-tube equilibrium tests at 105 °C with actual process liquor.

When Carbonate, Sulfate, and Organic Carbon Accumulate in Caustic Loops

At 105 °C, impurity behaviour differs from high-temperature digestion because sodium carbonate solubility is higher and organic degradation is slower, but recirculating impurities still shift gibbsite solubility. Sodium carbonate forms from atmospheric CO₂ absorption in spent liquor storage and from bauxite carbonate minerals; sulfate originates from bauxite and process water. Sodium carbonate concentrations of 20–40 g/L as Na₂CO₃ reduce the equilibrium A/C ratio by 0.02–0.05 units, primarily through common-ion effects on Na⁺ activity and through increased liquor density that interferes with aluminate ion hydration. Sulfate concentrations above 15 g/L Na₂SO₄ promote formation of sodium aluminosilicate scales containing sulfate and can reduce alumina extraction by 2–4% in feeds with reactive silica above 4 wt%. Organic carbon, mostly as humic and fulvic acids, accumulates at 10–30 g/L organic carbon in closed caustic loops; it reduces gibbsite precipitation yield by poisoning seed surfaces, increases pregnant liquor viscosity to 1.8–2.2 mPa·s at 105 °C, and stabilises fine aluminium hydroxide particles. Oxalate can precipitate as sodium oxalate needles in cool sections of the circuit, but at 105 °C its solubility is sufficient to keep concentrations below saturation in most low-temperature refineries. Causticisation with lime removes carbonate at 80–90 °C, and organics are reduced by wet oxidation or sodium oxalate crystallisation; both interventions restore solubility margins.

Boehmite and Diaspore Extraction Are Not Achievable at 105 °C

For feed-lot qualification, mineralogy controls the extraction limit at 105 °C. Low-temperature digestion is suitable only for bauxites in which the aluminium-bearing mineral is predominantly gibbsite. Monohydrate phases such as boehmite (γ-AlOOH) and diaspore (α-AlOOH) require digestion temperatures above 200 °C and 240 °C respectively at comparable caustic concentrations; at 105 °C their solubility is negligible, and any alumina remaining in red mud after low-temperature digestion is normally attributed to undissolved monohydrate or aluminium substituted into goethite. Mineralogical feed control by X-ray diffraction is therefore critical; a shift from 95 wt% gibbsite to 85 wt% gibbsite with 8 wt% boehmite can reduce overall alumina extraction from 92–95% to 80–85% under otherwise identical 105 °C conditions. This defines the economic boundary for low-temperature refineries and is why bauxite blending is tightly specified. In addition, kaolinite reacts at 105 °C to form sodalite-type desilication product, consuming 0.8–1.0 kg of caustic soda per kg of reactive silica; high-silica bauxites therefore become uneconomic before boehmite is considered.

Analytical verification of the 105 °C solubility point for a given feed lot is performed in a 1-L pressure-tube or autoclave with synthetic or plant spent liquor at 105 °C ± 0.5 °C for 4 h. Slurry samples are filtered hot at 105 °C through 0.45 µm membranes to avoid gibbsite precipitation during cooling; the filtrate is then acidified and analysed by ICP-OES per ISO 11885:2007. The equilibrium test must include seed addition to avoid metastable supersaturation artefacts; without seed, apparent solubility can exceed true equilibrium by up to 0.08 A/C units. Replicate tests on three bauxite lots showed A/C deviations of ±0.02 when liquor organic carbon varied by 5 g/L; this variability propagates directly to precipitation seed charge calculations. Operational limits are therefore set at least 0.05 A/C units below the measured equilibrium to accommodate feed lot variance and instrumentation uncertainty.