250°C Bayer Liquor Autoclave Caustic Soda Consumption Pathways

At 250 °C in a Bayer digestion autoclave, caustic soda is simultaneously a reactant for alumina dissolution, a hydrolytic agent for gangue mineral decomposition, and a salting-out electrolyte for precipitation of sodium aluminosilicates. In a typical diasporic bauxite operation the free caustic concentration expressed as Na2O is maintained between 180 g/L and 260 g/L, with total alkali expressed as Na2CO3 between 250 g/L and 380 g/L; the digestion pressure is commonly 4.5 MPa to 6.0 MPa because saturated steam pressure at 250 °C is 3.97 MPa and non-condensable partial pressure from air ingress and volatile organics adds 0.5 MPa to 2.0 MPa. Under these conditions, residence time in multistage horizontal autoclaves ranges from 45 min to 180 min depending on the ratio of diaspore to boehmite in the feed; insufficient residence time or a temperature deviation of more than ±5 °C from the setpoint produces measurable loss in Al2O3 extraction, because diaspore dissolution follows a surface-reaction-limited kinetic regime with an apparent activation energy reported in the range 80 kJ/mol to 120 kJ/mol. The digested slurry at 55–65 wt% solids loading typically exhibits an apparent viscosity of 10–150 mPa·s at a shear rate of 100 s⁻¹, and the autoclave agitators are sized accordingly. Caustic consumption is therefore not a single loss mechanism but a set of parallel and sequential reactions that convert free NaOH into sodium carbonate, sodium aluminosilicate, sodium organic acid salts, sodium phosphates, sodium vanadates, sodium titanates, and residual liquor entrained in red mud. Plant mass-balance records from continuous autoclave lines show that reactive silica and carbonate formation commonly account for the largest accessible caustic sinks, while organic acid neutralisation and phosphate substitution become significant only when certain bauxite impurity thresholds are crossed. The exact partition of sodium within these pathways is affected by liquor composition, lime addition, air ingress, mineralogy, and wash circuit configuration; published data for the fully coupled system under 250 °C continuous autoclave conditions is limited, and industrial mass balances are commonly required to resolve individual consumption coefficients.

Why does reactive silica impose a non-linear caustic demand when diaspore dissolution reaches 250 °C?

Reactive silica in bauxite, largely kaolinite and halloysite, dissolves in hot sodium aluminate liquor via the reaction Al2Si2O5(OH)4 + 6 NaOH → 2 Na2SiO3 + 2 NaAlO2 + 5 H2O, although the actual silicate speciation in 180 g/L to 260 g/L Na2O liquor is dominated by monomeric and dimeric silicate anions. The released silicate then combines with sodium aluminate to precipitate sodalite-type phases whose simplified composition can be represented as 3(Na2O·Al2O3·2SiO2)·2NaX, where X is carbonate, sulfate, chloride, or aluminate; this precipitation removes both sodium and silica from the liquor. At 250 °C, the equilibrium soluble silica concentration after digestion is typically 0.4 g/L to 0.8 g/L SiO2, but the precipitation rate of sodalite/cancrinite is fast enough that true equilibrium is not reached in continuous autoclaves, and the liquor carries a residual supersaturation that deposits on pipe walls and heat exchangers downstream. Sodium consumption per mole of reactive silica is not fixed; it varies from 0.3 mol Na₂O/mol SiO₂ in high-lime hydrogarnet regimes to 1.0 mol Na₂O/mol SiO₂ in carbonate- and sulfate-rich liquors without sufficient lime. The non-linear demand arises because sodalite carbonate substitution increases sodium occupancy when carbonate concentration exceeds 30 g/L Na2CO3, while sulfate competes for cage anion sites and alters the crystal size distribution. A change in reactive silica feed concentration from 5 wt% to 8 wt% can increase caustic consumption by 10–25 kg Na₂O per tonne dry bauxite, depending on the lime to silica ratio and the carbonate activity in the digestion liquor.

The temperature sensitivity of sodalite precipitation is typically described by an apparent activation energy of 40 kJ/mol to 70 kJ/mol, lower than diaspore dissolution; therefore at 250 °C the reaction shifts from dissolution-limited to precipitation-limited in the first compartments of the autoclave train. Industrial autoclaves with L/D ratios between 10:1 and 16:1, impeller tip speeds 2.5 m/s to 5.0 m/s, and live steam injection through mantles or sparge rings are used to maintain solids suspension and avoid sodalite scale formation on heating surfaces. Scale formation on the first two compartments can reduce heat transfer coefficients by 20% to 40% within 7 to 21 days of continuous operation, based on plant observations; acid washing with inhibited hydrochloric acid or high-pressure hydro-jet cleaning is then required to restore the design tube-wall temperature. Operationally, the caustic inventory is balanced by adding fresh NaOH to compensate for sodalite soda loss, but adding fresh NaOH raises the free caustic concentration and accelerates reactive silica dissolution, creating a positive feedback loop that must be controlled by feed blending, lime addition, or pre-desilication at atmospheric conditions. A processing window of ±5 °C around the 250 °C setpoint is commonly required because a drop below 245 °C slows diaspore dissolution while sodalite precipitation continues, resulting in a transient caustic deficit that can exceed the makeup addition capacity of 0.5 m³/h on large autoclave trains.

Table 1: Comparative reaction pathway data for sodium sinks in 250 °C autoclave digestion
PathwayProduct phaseSodium consumption per mole of impurityTemperature influenceOperational threshold/limitation
Reactive silica dissolution to sodalite/cancriniteSodalite/cancrinite scale and fines0.3–1.0 mol Na₂O/mol SiO₂Increases with temperature; lower activation energy than diaspore dissolutionLime to silica ratio 0–1.2 determines sodalite vs hydrogarnet; carbonate > 30 g/L Na₂CO₃ increases sodium occupancy
Atmospheric CO₂ absorption to sodium carbonateDissolved Na₂CO₃, causticized by lime1 mol Na₂O/mol CO₂Equilibrium shifts toward NaOH regeneration above 200 °C with limeOpen launders increase absorption; closed vessels reduce; lime dosage limited by CaCO₃ scaling
Organic acid neutralisation and oxalate precipitationSodium carboxylates, sodium oxalate crystals2 mol NaOH/mol oxalic acid; variable for humatesDegradation increases with temperature and oxygen ingressLiquor TOC > 25 g/L C increases viscosity; oxalate purge required; thermal destruction limited by energy
TiO₂ reaction with sodium aluminate when lime deficientSodium titanate or calcium titanate0.5–1.0 mol Na₂O/mol TiO₂High temperature accelerates TiO₂ attackLime addition of 3–5 wt% forms CaTiO₃; avoid excess lime due to scaling
P₂O₅ reaction with sodium aluminate and limeSodium phosphate or calcium phosphateup to 3.0 mol Na₂O/mol P₂O₅ if no limeLime suppresses sodium phosphate formationPhosphate > 0.5 wt% P₂O₅ in bauxite requires lime adjustment
V₂O₅ reaction with NaOHSodium vanadate in liquor1.0–3.0 mol Na₂O/mol V₂O₅Crystallisation below 40 °C removes vanadateVanadium purge via cooling crystallisation; avoid vanadium deposition on evaporators
Residual liquor entrainment in red mudSodium aluminate liquor in residue2–10 kg Na₂O per tonne dry residue depending on washing efficiencyLower wash water temperature increases viscosity and entrainmentDeep-cone thickener underflow solids > 45 wt% reduces entrainment; disposal constraints

Carbonate forms in Bayer liquor whenever free NaOH contacts carbon dioxide from atmospheric air in agitated storage tanks, falling-film evaporators, and red mud washing circuits. The reaction NaOH + CO₂ → Na₂CO₃ + H₂O reduces free caustic stoichiometrically, and in closed-loop liquor the carbonate concentration expressed as Na₂CO₃ can accumulate to 50–100 g/L if causticization with lime is not practiced. At 250 °C digestion, sodium carbonate is not an inert spectator; it participates in sodalite cage anion substitution and in reverse causticization with lime if slaked lime is dosed. Carbonate also increases sodalite scale density and changes the morphology of desilication product from fine dispersed particles to hard, adherent scale on autoclave walls. A plant operating without causticization may observe a free caustic decline of 0.5–2.0 g/L Na₂O per day from atmospheric carbonation across open launders and filters, with the exact rate dependent on liquor temperature, agitation intensity, and total exposed surface area. In addition, sodium carbonate in liquor reduces the solubility of sodium oxalate and increases boiling point elevation, which raises steam demand in evaporators and reduces effective temperature driving force across heat exchanger surfaces.

Organic acids from bauxite humates, humic substances, and process additives undergo alkaline degradation to formate, oxalate, acetate, succinate, and aromatic carboxylates. The total organic carbon in Bayer liquor at a high-temperature diasporic operation may range from 10 g/L to 35 g/L C, and the free caustic inventory is reduced by neutralisation of carboxyl groups to sodium salts. Sodium oxalate solubility decreases with increasing caustic concentration, and at liquor temperatures below 60 °C it crystallises as needle-shaped crystals that cause filtration blinding and product quality defects. Sodium oxalate precipitation removes two sodium ions per oxalate ion; if oxalate purge is inadequate, organic carbon accumulates to 20–40 g/L C, causing viscosity increase and lower alumina supersaturation. The degradation of humic substances to lower-molecular-weight acids consumes caustic both by neutralisation and by forming sodium salts that are refractory to thermal decomposition. Published data for exact kinetic partitioning between oxalate precipitation, carbonate formation, and humate decomposition at 250 °C in continuous autoclave circuits is limited; plant-specific organic carbon balances using wet oxidation or thermal destruction units are required to prevent caustic inventory drift.

When lime dosing exceeds the stoichiometric threshold for hydrogarnet formation, calcium carbonate scaling and free caustic release occur simultaneously

Lime is added as either quicklime or slaked lime to high-temperature digestion when bauxite contains more than 2–3 wt% TiO₂ and when carbonate control is needed. The addition of CaO at 2–5 wt% dry bauxite mass shifts the desilication product from sodalite (Na₂O-containing) to hydrogarnet (CaO-containing) and calcium titanate (CaTiO₃), thereby reducing sodium consumption by reactive silica. However, lime addition also introduces a caustic regeneration pathway via the equilibrium CaO + Na₂CO₃ + H₂O ⇌ CaCO₃ + 2 NaOH. At 250 °C the equilibrium favours NaOH regeneration, but reaction kinetics may be limited by solid-liquid mass transfer because calcium carbonate forms a passivating layer on undissolved lime. If lime dosage exceeds the amount required for TiO₂, phosphate and carbonate control, excess Ca(OH)₂ can form calcium hydrogarnet and calcium silicate, but the majority of surplus calcium precipitates as CaCO₃ in the presence of carbonate; this scaling on autoclave walls and downstream heat exchangers reduces heat transfer. Plant data often show an optimal lime addition window of 3–5 wt% for diasporic bauxite with 3–5 wt% TiO₂, outside which caustic consumption or scaling penalties increase sharply. The processing window is narrow because at 250 °C a temperature reduction of 5 °C can slow lime slaking and causticization enough to raise carbonate accumulation by 5–10 g/L per day in high-carbonate liquors.

On the autoclave side, lime is typically injected as a slurry of 10–20 wt% CaO in process liquor through a dedicated high-pressure positive-displacement pump into the first or second vessel of the train. The reaction of lime with reactive silica and titania competes with the reaction with carbonate; if carbonate is high, calcium carbonate forms rapidly and consumes lime that would otherwise form hydrogarnet. This can create a caustic consumption conflict: the desired soda-saving reaction with silica is sacrificed, while the caustic generated by causticization is partly offset by the loss of free caustic through sodalite precipitation. The net effect on caustic consumption is therefore non-monotonic with lime addition; below the stoichiometric threshold for TiO₂ + P₂O₅ + CO₂, soda consumption increases because reactive silica forms sodalite; above the threshold, scaling and viscosity problems reduce plant availability and can cause unplanned downtime of 4–12 h per event for vessel descaling.

Analytical control of free caustic, carbonate and aluminosilicate supersaturation

Control of free caustic in 250 °C autoclave liquor requires separation of carbonate alkalinity before acid titration, because carbonate interference can bias the result by 2–15 g/L Na₂O depending on the method. In one common procedure, excess barium chloride is added to precipitate barium carbonate, and the filtered liquor is titrated against standardised hydrochloric acid using phenolphthalein; the method is adapted from ASTM E291-18 for sodium hydroxide assay. Total alkalinity is measured potentiometrically per ISO 9963-1:1994, with total alkalinity expressed as Na₂CO₃. Reactive silica in bauxite is determined by ICP-OES after alkaline fusion or acid digestion, adapted from ISO 11885:2007 for aqueous matrices; the method reports SiO₂ to 0.02 wt% precision at the 5 wt% concentration level, which is sufficient for feed blending but not for real-time autoclave control. Alumina in liquor is determined by complexometric titration or by difference from total sodium and free caustic using plant-specific correlations; the uncertainty is typically 2–4 g/L Al₂O₃. Carbonate ion in liquor is determined by ion chromatography per ASTM D4327-17, with reported bicarbonate and carbonate species as their sodium salts; the method is not fully reliable above 20 g/L carbonate without dilution and matrix matching. Organic carbon is measured by wet oxidation or high-temperature catalytic oxidation per ISO 8245:1999, with calibration standards prepared in a synthetic Bayer liquor matrix to avoid matrix effects.

Table 2: Analytical methods for Bayer liquor caustic consumption control
ParameterStandard methodMatrixTypical reported range
Free NaOHASTM E291-18 (acid-base titration with barium chloride precipitation)Bayer liquor150–350 g/L NaOH
Total alkalinityISO 9963-1:1994Bayer liquor200–400 g/L as Na₂CO₃
Reactive silicaISO 11885:2007 (ICP-OES after alkaline fusion)Bauxite/liquor0.2–10.0 wt% SiO₂
CarbonateASTM D4327-17 (ion chromatography)Bayer liquor5–80 g/L Na₂CO₃
Organic carbonISO 8245:1999Bayer liquor1–40 g/L C

The values in Table 1 are comparative industrial ranges, not universal constants; published data for a specific bauxite source at 250 °C with defined lime addition and carbonate concentration may fall outside these intervals. Plant-specific mass balances based on daily liquor and residue assays, validated against a weekly elemental reconciliation per ISO 11885:2007, are the only reliable means of partitioning caustic consumption among these pathways.

Phosphate, vanadium and sulfur species consume caustic through separate anion substitution mechanisms

Phosphate impurities enter the Bayer circuit as apatite, crandallite, and wavellite; under alkaline digestion these dissolve to release orthophosphate anions. In the absence of sufficient calcium, orthophosphate reacts with sodium aluminate to form sodium phosphate, consuming caustic; with adequate lime addition, calcium phosphate phases precipitate and the sodium remains in solution. At 250 °C, the conversion of P₂O₅ to calcium phosphate is rapid if the lime-to-bauxite ratio is maintained above 0.03 to 0.05, and the caustic consumption from phosphate can be suppressed to 0.1–0.3 kg Na₂O per tonne bauxite. Vanadium, present in many diasporic bauxites at 0.05–0.30 wt% V₂O₅, dissolves as sodium metavanadate and consumes caustic proportionally; the vanadate ion also catalyses degradation of organic matter and competes with carbonate in sodalite cages. Sodium vanadate is removed from the circuit by cooling crystallisation at 30–45 °C, where it precipitates as a complex sodium vanadate sludge that must be regularly purged; without purge, vanadium concentration rises and caustic consumption increases because the liquor holds more sodium as vanadate instead of free NaOH. Sulfur enters via pyrite, gypsum, and organic sulfur; at 250 °C and high oxygen ingress, sulfide oxidises to sulfate, and sulfate can form sodium sulfate or sodalite cage anions, both of which reduce free caustic. Sulfate concentrations above 10–20 g/L SO₄ in digester liquor promote sulfate sodalite scale, reduce silica solubility, and increase caustic consumption by adding another anion that must be neutralised.

Operationally, these minor impurity pathways are not individually controllable in a continuous autoclave train. The addition of lime for phosphate and titania control simultaneously alters sulfate and carbonate speciation, and the cooling crystalliser for vanadium removal also precipitates sodium oxalate if organics are high. Published data for the exact combined effect of phosphate, vanadium, and sulfur on caustic consumption in 250 °C diasporic bauxite digestion is limited; plant data generally show that the sum of these pathways accounts for 5–15% of total caustic consumption when lime addition is correctly controlled, but can exceed 25% if lime is under-dosed or if sulfate accumulates in a closed circuit. The use of an external causticisation loop with lime kiln and mud washing may recover some of this sodium, but the recovery is limited by the formation of calcium sulfate and calcium phosphate phases that report to the residue.

Residual caustic entrainment in red mud is a volumetric loss pathway that becomes caustic consumption when the sodium aluminate liquor reports to residue disposal without sodium recovery. In high-temperature diasporic operations, the red mud washing circuit typically consists of countercurrent decantation thickeners or deep-cone thickeners in series; underflow solids concentration is maintained between 30 wt% and 50 wt% to balance pumping energy and sodium recovery. Liquor entrained in the final residue after washing may contain 5–20 g/L Na₂O dissolved in the pore liquor, and with disposal of 1.2–2.5 dry tonnes of residue per tonne of alumina, sodium loss from entrainment can account for 2–10 kg Na₂O per tonne alumina. Seasonal temperature changes in wash water from 10 °C to 35 °C alter liquor viscosity and settling behaviour; at lower temperatures, mud interface levels rise in thickeners and underflow entrainment increases. Process bottlenecks observed on production lines include underflow pump cavitation when thickener underflow density exceeds 50 wt%, dilution line plugging by sodalite scale, and loss of washing efficiency when flocculant dosing is interrupted for more than 2 h. Published data for exact entrainment losses in 250 °C autoclave operations with diasporic bauxite is limited because most site-specific wash circuit mass balances are confidential; the available ranges indicate that entrainment is usually smaller than reactive silica and carbonate consumption but becomes the dominant controllable sodium loss when washing efficiency falls below 85%.

Related Articles