A mixed integer nonlinear programming (MINLP) model is presented for synthesizing reactive distillation columns when chemical reaction equilibrium cannot be assured. The MINLP minimizes the total annual cost subject to a rigorous tray-by-tray model. The solution of this MINLP yields the optimal number of trays, the optimal feed rates, and the optimal feed tray locations. The liquid holdup per tray, the reflux ratio, and the temperature and composition profiles within the column are optimized as well. The MINLP is solved using generalized Benders decomposition, and the technique is illustrated with ethylene glycol synthesis.