Uncertainty-Aware Techno-Economic Assessment of Calcium-Based Treatment and Caustic Recovery from Refinery Spent Caustic

A computational screening study. No refinery or experimental data are used; all input ranges are generic placeholders, the PHREEQC results are model calculations and not experiments, and the economic and greenhouse-gas results are conditional screening estimates, not forecasts. Prepared for submission to Chemical Engineering Research and Design (Elsevier). Electrolyte-thermodynamic calculations (PHREEQC 3.7.3, pitzer.dat) show that gypsum precipitation vanishes once the final equilibrium hydroxide exceeds about 0.08–0.23 mol/kgw (25–90 °C, 0.05–0.5 mol/kgw sulfate), so a regenerated caustic liquor cannot co-produce gypsum; recovering hydroxide and removing sulfate as a solid therefore need separate treatment environments. A Monte Carlo model (10,000 draws, 720 to 1,320 design points, two feed archetypes) compares staged oxidation + lime causticisation + liquor recycle with neutralisation + gypsum precipitation and with third-party disposal, with chance-constrained design selection, an oxygen-mass-transfer feasibility check (the kinetics-only optimum is an upper bound; oxygen-enriched, transfer-limited designs are reported alongside), a sulfate-limited recycle loop with salt-saturation margins, decision-swing analysis and a preliminary greenhouse-gas screen. Version note (1.1.0 / preprint v2): revised after a readiness review; adds the oxygen-transfer analysis, recycle operability, strengthened thermodynamic analysis, convergence and decision-swing analysis, and five figures. The headline Route B result is now reported as conditional on oxygen supply. Code, data and results: https://github.com/sandlerleon/spent-caustic-recovery-tea, archived at 10.5281/zenodo.23406668.