Fabrication of nanofiltration membranes with tunable separation characteristics using methods of uniform design and regression analysis

To facilitate the determination of the optimized casting conditions for the fabrication of the nanofiltration (NF) membrane from a selected polyamidesulfonamide (PASA), the uniform design (UD) method was employed. With the aid of the chemometric method, the influences of one qualitative and four quantitative variables of the fabrication conditions of membranes on the membrane performances can be elucidated with a limited number of experiments. Guided by the ultrafiltration (UF) method, 24 sets of represented treatment points from the studied experimental domain, incorporating all five variables, were selected to prepare the NF membranes. The permeation flux and retention ability of duplicate NF membranes prepared at each of the selected 24 experimental points toward the permeation of four carbohydrate reference solutions, namely, glucose, sucrose, raffinose and beta-cyclodextrin, were determined. The response data collected from each of the selected 24 experimental points were then regressed with the studied casting conditions by using the SAS software. Acceptable regression equations were obtained in all cases studied. Based on the predictions of the regression equations, several optimized NF membranes targeted at the retention of glucose, sucrose, raffinose and beta-cyclodextrin, respectively, were prepared. Over 70% of the predicted values of the response of the membranes can be validated by the real measurements of the NF membranes. (C) 2002 Elsevier Science B.V. All rights reserved.