The cost of dilute ethanol purification is a significant component of the overall cost for fuel grade ethanol production through fermentation or other biological routes. In this study, we consider bioethanol purification to the fuel grade in a closed system of photobioreactors (PBRs) that produce a dilute (0.5-5 wt %) ethanol solution, and where the water is recycled back to the reactors after separation. To reduce the energy consumption and hence the cost of dilute ethanol purification, a reverse osmosis (RO) membrane process is introduced as a potential pretreatment in order to purify ethanol water mixtures to an intermediate concentration. The pretreated mixture is separated across the azeotropic point to the fuel grade by a hybrid distillation-membrane-pervaporation (D-PV) process. A superstructure of the overall separation process is formulated in the General Algebraic Modeling System (GAMS) environment as a mixed-integer nonlinear programming (MINLP) problem and optimized to minimize the total separation cost. The problem is posed with a constraint on the minimum ethanol recovery by the overall system. For dilute feeds, the steam used in the distillation reboiler dominates both the energy and the cost of the overall system. Installing the RO system reduces the steam usage at the expense of larger capital investment in membrane modules, which has been found to be optimal for dilute feeds below 3 wt % ethanol. The optimal number of membrane stages and the feed location of individual RO modules change at different feed concentrations and ethanol recoveries. The optimal process design results in a higher ethanol recovery by the main hybrid separation unit compared to the RO membrane process due to the high capital cost of an RO system to achieve a high recovery. Developing a cheaper RO membrane may further reduce the separation cost and may expand an optimal dilute feed concentration range.