COMPUTER-AIDED IDENTIFICATION OF CHEMICAL-REACTION HAZARDS

Many of the thermal runaway accidents that occur in the pharmaceutical industry are due to exothermic secondary reactions. A good understand ing of the runaway (e.g., for dynamic modeling purposes) requires know ledge of the stoichiometries of these reactions. A screening test to e lucidate tile stoichiometries of secondary reactions is presented. It eliminates implausible stoichiometries from a candidate set due to the ir thermodynamic infeasibillty. We use computational chemistry coupled with statistical thermodynamics to estimate the thermodynamic propert ies of the unfamiliar species typically involved in secondary reaction s. The real life runaway of a process to manufacture tetrahydrofurfury l benzenesulfonate is treated as a case study. Hartree-Fock and Densit y Functional Theory calculations are performed to test for thermodynam ic feasibility, leading to several postulated stoichiometries correspo nding to hazardous thermal decomposition.