This work presents a systematic optimization-based methodology for the synthesis of reactive absorption columns. A novel Generalized Modular Framework (GMF), previously developed by the authors is extended, and coupled with Orthogonal Collocation (OC), a model order-reduction technique. This integration allows avoiding the combinatorial issues that accompany stage-wise physical models. Yet the GMF/OC process representation still exhibits a "stage-wise character" due to polynomial approximations being applied at specific points of the represented columns for the main process variables. An example problem concerning the synthesis of a NO, reactive absorption process is used to illustrate the applicability and efficiency of the proposed methodology. (c) 2005 Elsevier B.V. All rights reserved.